text
stringlengths 66
377k
|
|---|
The health-economy trade-off during the Covid-19 pandemic: Communication matters
How do people balance concerns for general health and economic outcomes during a pandemic? And, how does the communication of this trade-off affect individual preferences? We address these questions using a field experiment involving around 2000 students enrolled in a large university in Italy. We design four treatments where the trade-off is communicated using different combinations of a positive framing that focuses on protective strategies and a negative framing which refers to potential costs. We find that positive framing on the health side induces students to give greater relevance to the health dimension. The effect is sizeable and highly effective among many different audiences, especially females. Importantly, this triggers a higher level of intention to adhere to social distancing and precautionary behaviors. Moreover, irrespective of the framing, we find a large heterogeneity in students' preferences over the trade-off. Economics students and students who have directly experienced the economic impact of the pandemic are found to give greater value to economic outcomes.PLOS ONEPLOS ONE | https://doi.org/10.1371/journal.pone.
# Introduction
The Covid-19 health emergency prompted governments around the World to adopt unprecedented measures in order to control the spread of the infection. These measures involved the curtailment of basic individual freedoms, which ranged from the total lock-down of economic and social activities to the adoption of precautionary behaviors such as the wearing of masks and the maintaining of interpersonal distance. However, it is quite evident that the efficacy of these measures is strictly related to people's compliance, which in turn depends on whether or to what extent people agree with them. Self-isolation, social distancing and other precautionary individual behaviors are extremely difficult to impose without people's willingness to cooperate.
In the absence of financial incentives, a key role for the enforcement of Covid-19 related measures might be played by communication and persuasion policies as "individual compliance has collective benefits, but full enforcement is costly and controversial". In this paper, we investigate how the framing of the communication of Covid-19 related issues to the public affects individual preferences and, ultimately, their intention to adhere to precautionary behaviors.
We focus on a key aspect of Covid-19 related measures: the inherent health-economy trade-off that they have to deal with. By restricting production and consumption in some sectors, government interventions aimed at reducing the spread of the virus (such as social distancing, lockdowns) produce an immediate negative effect on the economy. On the other hand, in their absence the pandemic could be harsher in terms of a higher number of sick people and, also, more negative effects on the economy. Quantifying costs and benefits of different scenarios, in a situation characterized by a very high level of uncertainty, such as that experienced during the Covid-19 emergency, is particularly complex. Then, in absence of a well-defined counterfactual, it is difficult to assess how much of the economic costs of the pandemic can be considered as indirect costs to be imputed to the policy interventions. Even if it is true that policies characterized by high short-term economic costs can lead to lower costs in the future and, as a consequence, less restrictive policies might not improve but weaken the economy, in the short run individuals perceive especially the economic costs deriving from government restrictions. This is well documented by "reopening protests", decrying the economic costs of business closure and social distancing, that have been held in many countries from the beginning of the pandemic until today.
In fact, at least in the short run, policy interventions aimed at saving lives tend to produce negative consequences on the economic activity. Quasi-complete lockdown policies implemented by many governments worldwide have mitigated the extent of the spread of the contagion but have also given rise to very considerable short-term economic costs. As regards Italy, the estimated cost for each week of closure of all non-essential activities is a reduction in GDP of 0.5-0.75%. The less costly alternatives, e.g. targeted lockdown policies-which often characterize the re-opening phase-make even more explicit the tradeoff between health and economic outcomes and such a dilemma can only be expected to persist until a medical response (full vaccination and medical treatments) becomes available. As the health-vs-economy trade-off confronting the social planner is ultimately based on the value that society puts on population health versus short-term economic gains, it is therefore essential to understand how people evaluate health versus economic outcomes and to identify which factors influence their preferences.
We investigate how the framing of communication over both sides of the trade-off affects individual preferences over the policies to be implemented by using a survey field experiment which involved around 2000 students enrolled in a major university in Italy in April 2020. A key element of our research is that of focusing on young individuals. This is not casual. Indeed, while having virtually the same probability of getting the infection, young individuals have also -by far-less severe consequences from it . These two elements (high prevalence but low health consequences) suggest that the behaviors of young individuals are crucial in order to control the diffusion of the pandemic. While data on the violations of Covid-19 measures are not available by age of the offender, the report of the Police department shows that violations happened mostly in areas of 'movida', i.e. restaurants, pizzerias, fast-food, pubs and bars that are usually frequented by young individuals.
Students involved in our experiment live in the South of Italy, an area that, despite experiencing lower rates of contamination and death, is much weaker from an economic point of view and is endowed with a very poor health system . The key question of the survey asked students to reveal their preferences of policies that gave a different weight to the health and economic outcomes of the pandemic. We manipulate the framing of the introductory text of the question associated with the two elements of the trade-off comparing a positive framing which focuses on the protection of health/economic outcomes with a negative framing that presents one or both elements of the trade-off in terms of costs.
We induce the positive framing for health and economic outcomes of the pandemic using the word protection and the negative framing using the word costs. The choice of the framing -in particular the positive one-is inspired by the language used by politicians and the media during the lockdown in Italy in which expressions such as "health protection" or "health safeguard" have been principally used. For instance, on February 26 th (during the initial phase of the epidemic), the Italian Government's bill including the closure of schools and many economic activities was presented by the ministry of Health as "Actions for the protection of community health" . "La Repubblica", one of the most widely read Italian newspapers, published 29 articles mentioning the phrase "health protection" during the lockdown phase.
Based on the existing evidence related to the effects of framing, we hypothesize that students are more likely to choose health-(economy) oriented policies when health (economy) is framed positively instead of negatively. In fact, it is generally found that, when the framing is positive, subjects view the outcomes as gains (showing risk aversion) while, when the framing is negative outcomes are perceived as losses (leading to risk seeking). A similar hypothesis applies if we consider the link between the two elements of the trade-off. In fact, the worsening of health or economy in the trade-off can be perceived either as a cost or as a loss. The dead-loss effectposits that an individual's subjective state can be improved by framing negative outcomes as costs rather than as losses. In our context, this means that students should be more willing to sustain the costs of the worsening of health or of the economic situation if such a payment is seen as the cost for the protection of the other element of the trade-off, instead of an uncompensated loss. Therefore, our hypothesis is that communicating the trade-off by using a positive framing (protection) for one element and a negative framing (cost) for the other, instead of negative framing for both, will shift preferences towards the positively framed element and will motivate students to be more willing to sacrifice the negatively framed element.
We find that preferences over the trade-off are related to how the trade-off is communicated. Compared with the framing where both health and economic concerns are expressed as costs, when the trade-off is framed as economic costs to be paid in order to protect against a worsening of health, a large majority of students weigh more the health dimension, deciding to care less about the economic costs to be sustained in return for health protection. Under this framing, 47.36% of students responded that they would prefer policies that consider "extremely" or "very much" the protection of health and "not much" or only a "little bit" the costs for the worsening of the economic situation. On the other hand, under the negative framing, 34.15% of respondents answered that they would prefer policies that consider "extremely" or "very much" the costs for the worsening of health and "not much" or a "little bit" the costs for the worsening of the economic situation. This is consistent with prospect theory predictions-as the positive framing induces risk aversion-and with the dead-loss effect. This also supports a large body of empirical evidence that shows that the adoption of healthy behaviors is strongly influenced by the framing strategy.
Combining health with a positive framing seems to be a low-cost but highly effective communication strategy. In fact, when digging deeper to see whether only audiences with particular characteristics are affected by the positively framed communication, we find that the effect is quite homogeneous, even if women and trustworthy students are found to be particularly reactive.
Regardless of framing, students' preferences over the health-economy trade-off are highly influenced by several individual characteristics. The field of study (i.e. studying economics) and a pandemic-induced difficult household economic situation affect preferences increasing the weight being given to the economic dimension of the pandemic. On the other hand, students with more highly educated parents, those with altruistic feelings and those who feel more anxious assign more weight to the health dimension. Lastly, when looking at the intention to comply with official advice for self-isolation and precautionary behaviors, we find that intended compliance is higher among students who position themselves more on the health side of the trade-off. Using an instrumental variables approach that exploits the random assignment to the treatments in the framing experiment, we find support for the causality of the relationship that runs from the perception of the trade-off to the compliance with prescribed behaviors.
This paper contributes to the research on framing effects. Beginning with, it has been shown that different framings affect the perceived domain of the outcomes thus leading to different choices. The relevance of framing in influencing individual behavioral patterns has been widely documented in a variety of contexts, e.g. in public-sector decision making, in health decisionsand consumer choices. We complement this literature by analyzing framing effects in a new and previously unexplored setting, i.e. in the midst of an emergency involving two key dimensions of individual well-being such as health and economic outcomes.
We also contribute to the literature in political science that has started to apply behavioral economics insights to the study of political processes (see, for a survey). A number of papers has considered the importance of framing in decision-making and applied prospect theory to explain the behavior of governments and leaders in crisis situations.show that the decision whom to vote for is strongly influenced by the way policy programs are described. Other works highlight how the political supply side can use some well-known biases, such as loss aversion or the status quo bias, in order to manipulate the evaluation of alternatives.show that whether labor market policies are presented as aiming at lower unemployment or higher employment makes a great difference for public opinion. Other works show that when the outcome of a policy is perceived as a loss, the propensity to take risks to mitigate the situation increases, while when a policy creates benefits that are also perceived as gains, the willingness to take risks to achieve even better results diminishes. Our paper contributes to this literature by investigating a special setting in which-in absence of a direct political competition-citizen's biased decision-making might be primarily exploited by policy-makers in order to spread pro-social behaviors and thus support the crisis management.
Our paper also speaks to a growing stream of Covid-19 economics literature that is investigating individual perceptions over the health-economy trade-off and compliance with recommended behaviors. For instance,show that how people evaluate health versus short run economic outcomes and compliance with prescribed behaviors depends on the information they receive. They assess public preferences over this trade-off by randomizing information provision on economic and health costs of the pandemic and find that people strongly prioritize health over economy, but these priorities seem to change in predictable ways as the experience of death and income loss unfolds. More importantly, they also find that individuals choosing the maximum valuation of health over economy are more likely to comply with recommended behaviors. Likewise,study the role of cost-benefit considerations in shaping support for mandatory social distancing and stay-at-home measures by varying information on perceived economic costs and health benefits in an experimental setting. However, to the best of our knowledge, no previous paper so far has focused on the role of communication in shaping individual preferences on this trade-off.
Our result that a positive framing on the health side of the trade-off encourages people to worry and care more about health offers valuable insights to public authorities on how to tailor communication after the end of lockdown measures. Framing the policies adopted in terms of "protecting" health could stimulate individuals to place more weight on health concerns and might also positively affect their compliance with precautionary behaviors, helping to contain the spread of the virus. In turn, this could allow policymakers to concentrate more on the economic consequences of the pandemic.
A very preliminary version of this work appeared inand the main insights of our research have been popularized also by .
The remainder of the paper is structured as follows. In Section 2 we describe the experimental design, data and balance checks. In Section 3 we discuss our main results. Section 4 is devoted to explore heterogeneous effects across different groups, while Section 5 examines the relationship between health-economy preferences and precautionary behaviors. Section 6 offers some concluding remarks.
## Experimental design, data and balance checks
We study the effect of different communication strategies on individual preferences regarding the trade-off between health and economic outcomes by collecting survey data through a field experiment (randomized controlled trial, RCT). The University of Calabria, in the person of the Rector, approved the study and gave the authorization to collect and analyze data from students anonymously. The invitation to the survey was sent by the administrative office via the institutional email address. Prior to participating, subjects were asked to read a statement about the study, as reported in S1 Appendix. After this statement, they were asked to give or deny consent. By clicking on the consent agreement, they proceeded on to the survey questions.
The survey was submitted on April 20 th -and remained open until April 25 th -to about 10,000 students regularly enrolled at the 2 nd and 3 rd year of the different First Level Degrees, 1 st year of the Second Level Degrees and all years of "Lauree a Ciclo Unico" offered by the University of Calabria (61% of them are female; on average they are 22 years old; 29% of them belong to the Department of Social Sciences, 20% to Engineering, 18% to Humanities and 33% to Sciences). Students were randomly assigned to four treatment groups on the basis of their matriculation number. We have firstly divided students into two groups: those with an even matriculation number and those with an odd matriculation number. Then, within each group, we have randomly created two subgroups of equal dimension. Participation in the survey was voluntary and data were collected anonymously. The response rate to our survey was 17.5%.
The four treatment groups were created by manipulating the framing associated with the two elements of the trade-off, thus enabling comparisons between a positive framing which focuses on the protection of health/economic conditions with a negative framing that presents one or both elements of the trade-off in terms of costs. The survey question which was used to induce treatment conditions was the following: "The government is planning the reopening after the temporary self-isolation measures introduced to deal with the coronavirus emergency. At this stage, it is necessary to consider the consequences that each decision has in terms of protection (costs for the worsening) of health-number of infections-and protection (costs for the worsening) of the economic situation. If you were the head of the government, which strategy would you choose?". Respondents could choose from the following five alternatives: "I would consider extremely the protection (costs for the worsening) of health and not much the protection (costs for the worsening) of the economic situation"; "I would consider very much the protection (costs for the worsening) of health and a little bit the protection (costs for the worsening) of the economic situation"; "I would take into account enough the protection (costs for the worsening) of health and enough the protection (costs for the worsening) of the economic situation"; "I would consider a little bit the protection (costs for the worsening) of health and very much the protection (costs for the worsening) of the economic situation"; "I would consider not much the protection (costs for the worsening) of health and extremely the protection (costs for the worsening) of the economic situation".
Thus, we design four treatments in a between-subjects design. In the first treatment, HealthCosts-EconomyCosts (HC-EC, hereafter), participants are framed the trade-off in terms of costs both for health and for the worsening of the economic condition. In the second treatment, HealthProtection-EconomyCosts (HP-EC hereafter), participants are framed the tradeoff in terms of protection of health and costs for the worsening of the economic condition. In the third treatment, HealthProtection-EconomyProtection (HP-EP, hereafter), both elements of the trade-off are framed in terms of protection while in the fourth treatment, HealthCosts-EconomyProtection (HC-EP, hereafter), the choice is between the costs for health and the protection of the economic situation.
Inwe describe the question asking how students would balance health and economic concerns after the end of lockdown measures. In order to make the framing (and thus our treatments) more salient, the same wording is used both in the text of the question and in the text of the possible alternatives among which the students could choose. We report the percentage of students choosing each option under the four different treatments. The HP-EC treatment shifts individual preferences toward policies focusing on health concerns, while under the HC-EP treatment, the option of equally considering both health and economic concerns records the highest percentage of preferences compared to all the other treatments.
We use responses to the question on how students evaluate the health-economy trade-off to create our dependent variable, Health-Economy Trade-off, which is an ordinal variable taking values ranging from 0 (for participants who selected "I would consider not much the protection (costs for the worsening) of health and extremely the protection (costs for the worsening) of the economic situation") to 4 (for participants who selected "I would consider extremely the protection (costs for the worsening) of health and not much the protection (costs for the worsening) of the economic situation"). Thus, the variable is increasing in terms of the importance given to health concerns.
## Plos one
The health-economy trade-off during the Covid-19 pandemic
show that we can reject the null hypothesis of independence whenever we compare the HP-EC treatment with the other treatments (p-value = 0.001). Comparisons between the remaining pairs of treatments always fail to reject the null hypothesis.reports descriptive statistics of our variables both in the full sample and separately by treatments. Health-Economy Trade-off is on average 2.43 in the full sample. It takes on average the value of 2.4 in the HC-EC and HP-EP treatments, the value of about 2.6 in the HP-EC treatment and a lower value (2.3) in the HC-EP treatment.
In order to collect information on students' baseline preferences towards health and economic concerns, we posed the following question before introducing the treatment (see S1 Appendix for a translation of the survey questions): "Some research shows that the closure of non-essential activities was accompanied in Italy by a reduction of R t (an indicator of the spread of the epidemic) from 8.2 to 0.4. However, each week of non-essential business closures seems to reduce a country's income and profits by 0.75%. If you were the head of government and the following scenarios were proposed to you for the next two months, which one would you choose: a) No closure, R t = 8.2, Reduction of gross domestic product = 0%; b) Closes ¼ of non-essential activities, R t = 6.15, Gross domestic product reduction = 1.5%; c) Half of nonessential activities closed, R t = 4.1, Gross domestic product reduction = 3%; d) All non-essential activities closed, R t = 0.4, Gross domestic product reduction = 6%".
The variable Baseline Health-Economy Trade-off takes values from 0 (for respondent choosing the option "a") to 3 (for respondents choosing the option "d"), increasing in the importance given to the health side of the trade-off. It allows us to have a baseline measure of individual preferences that helps to investigate whether treatment effects are homogeneous or are dependent on ex-ante preferences. The average value of the variable in our sample is 2.
## Plos one
The health-economy trade-off during the Covid-19 pandemic Baseline and post-treatment preferences for the health-economy trade-off are positively correlated (corr = 0.18, p-value = 0.000).
We have also obtained information on personal characteristics (gender, age, studies, family background, and residence), personality traits, well-being and intention to adhere to social distancing and precautionary behaviors. In, we report descriptive statistics of each variable both overall and separately by treatment groups. When looking at predetermined characteristics, we see that students are on average 22 years old and about 71% of them are female. As regards their family background, parents have studied on average for 12 years.
As an indicator of students' personality traits, we included in the survey a question asking students how much they see themselves as a person who is Altruistic (21% of the sample), Trustworthy (29%), Extroverted (6%), Open to experience (22%) and Neurotic (13%). Students could choose among 7 alternatives: completely disagree; very much disagree; somewhat disagree; neither agree nor disagree; somewhat agree; very much agree; completely agree. The variables are dummies taking the value of 1 when the answer is "completely agree" and 0 otherwise.
We also collected information on Covid-19 health and economic implications. About 13% of the respondents state that they know someone (relatives, friends or even themselves) who tested positive for the diagnosis of Covid-19 and, for about 28% of students, both parents became unemployed because of the Covid-19 emergency. We also measure students' psychological conditions including in our survey two modules of the Patient Health Questionnaire (PHQ, a diagnostic tool for mental health disorders used by health care professionals,. On the basis of students' answers to a depression module and an anxiety module, we build a depression and an anxiety severity scale, respectively. The depression severity scale (calculated by assigning scores of 0, 1, 2, and 3, to the response categories of not at all, several days, more than half the days, and nearly every day, respectively) takes values from 0 to 24 and has an average value of 9.4, while the anxiety severity scale (calculated by assigning scores of 0, 1, 2, 3 and 4 to the response categories it doesn't match at all; it doesn't match; neither matches nor does not match; it matches; it matches completely, respectively) takes values from 0 to 20 with an average value of 13.20.
Finally, we asked students to report on a 0-100 range their willingness to comply with the following recommended behaviors: stay at home as much as possible; do not attend social events; wear face mask; stay at least two meters from other people; wash hands frequently; stay at home with symptoms of coronavirus; avoid hugs and handshakes. Using responses to these questions, we built two measures of compliance to these behaviors. First, we create a variable -named Compliance PCA-through a Principal Component Analysis of each of the seven questions on prescribed behaviors. As an alternative variable, we construct a "count" measure of compliance, summing up the values of the seven variables, and obtaining an indicator that ranges between 0 (when all the seven variables take the value of 0) and 700 (when all the seven variables take the value of 100). We adopt this approach as, in practice, the incidence of compliance is highly correlated across the different behaviors. For instance, the correlation
## Plos one
The health-economy trade-off during the Covid-19 pandemic between the intention to "Stay at home when sick" and "Wash your hands frequently" is equal to 0.557, p-value = 0.000, while the correlation between "Avoid hugs and handshakes" and "Stay at least 2 meters from other people" is equal to 0.563, p-value = 0.000. The average value of the variable is 648, it ranges from 641 in the HC-EP treatment to 655 in the HP-EC treatment. These high levels of compliance with recommended behaviors are consistent with findings reported bywho rely on a representative survey of Italian adults conducted during the initial phase of the Covid-19 pandemic.
To investigate the effects that the four treatments produce on individual outcomes we need four comparable groups. The last column ofreports p-values of tests of equality of variables' means among treatments. Treatment groups are evenly balanced on a large number of covariates (with the exclusion of Parents Unemployed Covid-19) and data regarding predetermined characteristics show that we are unable to reject the hypothesis that the randomization was successful in creating comparable treatment groups in respect of observable characteristics in the subsample of students submitting their responses to the survey questions. We have also tested the equality of variables means for each possible pair of treatments. We find that treatments are always equally balanced in terms of age and gender but sometimes they present differences in the distribution of the field of study. For this reason, in our estimates we control for field of study dummies. Also, if we compare predetermined characteristics of respondents with those of the average student population we find that our sample is quite representative of the student population, along the dimensions of age and field of study while, due to a higher response rate, women are slightly over-represented (61% of students included in the survey are female).
## Communication and preferences for health and economy: main results
In this section we carry out an econometric analysis to investigate whether being assigned to the four different framings adopted in our experiment induces students to balance differently health and economic outcomes.
We estimate several specifications of the following simple model:
[formula] Health À Economy Trade À off i ¼ b 0 þ b 1 ðHP À ECÞ i þ b 2 ðHC À EPÞ i þ b 3 ðHP À EPÞ i þ b 4 X i þ b 5 F i þ þb 6 BaselineHealth À EconomyTrade À off i þ b 7 W i þ b 8 Z i þ u ið1Þ [/formula]
where the vector X i includes individual pre-determined characteristics (gender, age, field of study, etc.), F i includes family background variables (parents' education, etc.), W i includes controls for Covid-19 health and economic implications (parents' employment, experience with Covid-19, psychological conditions), Z i is a set of variables measuring current personality traits, and u i is the error term.
In this setting, β 1 is the difference between HP-EC and HC-EC (that is the treatment effect of framing health in terms of protection instead of costs) in the propensity to favor policies that give greater weight to health concerns arising from the spread of Covid-19. Positive values of β 1 suggest that, in the management of the reopening after the lockdown measures, communicating the trade-off using for health a positive framing which focuses on protective strategies -instead of a negative framing based on costs-increases students' concerns for the health consequences of the pandemic. A similar interpretation holds for β 2 and β 3 that represent the effect induced by the other two treatments, HC-EP and HP-EP, respectively, with respect to the framing HC-EC.
Our hypotheses are the following:
- H1: β 1 >0, that is, the use of a positive framing (protection) for health induces students to associate a greater weight to health in the trade-off, being both more risk averse on this domain and more inclined to bear higher economic costs as they are seen as a payment needed in order to protect health;
- H2: β 2 <0, that is, the use of a positive framing (protection) for economic outcomes increases the weight of economic outcomes in the trade-off for the same reasons as above;
- H3: β 3 �0, that is when both elements of the trade-off are framed in terms of protection; either they should carry the same weight or, given the strong health concerns under a pandemic, the protection of health may carry more weight.
Inwe report estimation results of several specifications of model. We estimate an Ordered Probit Model to study the effect of the assigned treatment condition on the probability of students giving greater consideration to health concerns in policy decisions. Since the dependent variable increases with the importance associated with health concerns, positive coefficients suggest the likelihood of preferences being more shifted toward health concerns. In all the regressions, standard errors (corrected for heteroscedasticity) are reported in parentheses.
As shown in column (1), where we do not include controls, we find that, compared with the HC-EC treatment, the HP-EC framing induces students to choose a policy that gives greater consideration to health issues. Thus, our data fail to reject hypothesis H1: when the trade-off is communicated as health protection versus costs for the worsening of the economic situation, instead of framing both health and economy as costs, respondents perceive the worsening of the economic situation as a cost allowing for protection against the worsening of health, instead of as an uncompensated loss, and are therefore more willing to sustain it. This is confirmed by the test on the coefficient reported at the bottom of the showing that we can reject the hypothesis that the coefficient is smaller than or equal to zero at a 0.01 significance level. The shift in preferences that favor policies that mainly focus on health issues produced by the HP-EC treatment is statistically significant also when compared with the other different types of framing used in our experiment, as shown in Panel A.
As regards hypothesis H2, we find evidence for a negative effect of the positive framing associated with economic outcomes on the preference for health-oriented policies. The test of our hypothesis reported in Panel B shows that, in all estimates but the first one, we can reject the hypothesis that the coefficient is greater than or equal to zero at a 0.10 significance level. Finally, when looking at the HP-EP treatment (H3), we find a positive but not statistically significant from zero coefficient. This is consistent with the test reported in Panel B and would suggest that framing both elements of the trade-off in terms of protection is the same as using the framing "costs" and, even under a pandemic, the protection of health does not carry significantly more weight when joined with the protection of the economic situation. When comparing HP-EP with the remaining treatments (Panel A) we find that the HP-EC treatment generates a significantly bigger shift in preferences that favor policies that mainly focus on health while the effect of the HP-EP treatment is significantly different from the HC-EP treatment (which induces a shift towards economy centered policies) in all but the first specification.
These results remain qualitatively unchanged when we add controls for age, gender and field of study (column 2) and when we also add controls for family background and province of residence fixed effects (column 3). In column (4) we include among controls our measure Baseline Health-Economy Trade-off, which is positively correlated with preferences for a health-centered policy, but does not affect the influence produced by our treatment conditions. No relevant changes are found also when we add, among regressors, proxies for individual exposure to the Covid-19 emergency both in terms of health and economic outcomes (column 5) and when we control for individual personality traits (column 6). The impact of the HP-EC treatment is sizeable. When looking at average marginal effects for the specification including all the control variables (column 6) we find that when the tradeoff is expressed in terms of protection of health and costs for the worsening of the economic situation-instead of in terms of costs for both health and the economy-students are about 0.45% less likely to choose the policy giving the greatest weight to the economic situation; about 0.78% less likely to choose the policy considering a little bit health and very much economic outcomes; 11.8% less likely to choose the intermediate policy; 7.7% more likely to choose the policy considering very much health and a little bit economic outcomes and about 5.2% more likely to choose the policy that gives greatest weight to health concerns.
As regards control variables, we find that the field of study reveals different preferences and that students enrolled in scientific disciplines tend to prioritize health concerns compared with students enrolled in economics and social sciences and engineering. There is also an important difference in terms of socio-economic background; students who have more highly educated parents and who live in larger houses show a preference for policies that tend to favor health protection. Since both parental education and floor space per person are usually associated with the economic conditions of the family, the result shows that those who come from contexts of greater economic distress tend to give greater weight to the economic costs of the pandemic while for students having better-off families the trade-off may be less salient. This is also confirmed by the fact that students with parents who lost their jobs due to the emergency tend to express themselves more favorably towards a compromise that takes due account of the economic costs of the crisis. On the other hand, students who are particularly anxious, due to the Covid-19 emergency, are more favorable to policies more focused on health issues. Finally, those who describe themselves as altruistic also tend to prefer health-centered policies.
In S2 Appendix, we show the robustness of our results to Ordered Logit and Multinomial Logit estimates. As suggested by the distribution in, Multinomial Logit estimates show that the effect of the HP-EC treatment is statistically significant for the two categories referring to more health centered policies.
To check the robustness of our results, we have also created, as an outcome variable, a dummy taking the value of 1 for students who report preferences for policies that give 'very
## Plos one
The health-economy trade-off during the Covid-19 pandemic less' or 'less' relevance to the economic costs of the crisis and 0 otherwise. Probit estimates are qualitatively very similar to those discussed above. The only difference concerns the HC-EP coefficient that now is more precisely estimated but still typically not statistically significant at conventional levels. In addition, we have created another ordinal variable taking values ranging from 0 to 2 where 0 and 2 are participants selecting the two extreme options (respectively 0 = participants who selected "I would consider not much the protection (costs for the worsening) of health and extremely the protection (costs for the worsening) of the economic situation"; 2 = participants who selected "I would consider extremely the protection (costs for the worsening) of health and not much the protection (costs for the worsening) of the economic situation") while the value of 1 is assigned to all participants selecting the intermediate options.
Finally, since the middle option is the "neutral" response (which could be interpreted as a "don't know" answer for many respondents) we have also tested our results to the exclusion of such respondents from the sample. Our evidence of a positive and statistically significant effect of the HP-EC treatment holds in both cases. Results are not reported but are available upon request.
## Heterogeneous impact of framing
In the previous sections we have seen that a simple communication strategy, that positively frames the health side of the health-economy trade-off arising from the current health emergency, impacts on students' preferences towards the trade-off. Nonetheless, communication takes place in different contexts and is directed to different audiences, who might be more or less reactive to how messages are framed. Then, in this section, we investigate if individual characteristics, such as gender, economic and social background, personality traits, experiences and beliefs, can amplify or nullify the impact of framing.
With this aim, we analyze whether our treatments have produced heterogeneous effects across the three sets of controls that we have considered in the previous analysis (predetermined characteristics and background; personality traits; Covid-19 health and economic implications) and whether the impact is related to individual baseline preferences. This would suggest in which circumstances framing can be effectively used to try to build up consensus towards certain types of policies. For each control that we consider, we report bar graphs showing the difference in the level of our indicator Health-Economy Trade-off between each of our treatments and the reference treatment HC-EC, separately by category, and 95% confidence levels, based on the estimates of a seemingly unrelated regression equations (SURE) model reported in detail in Tables C1-C4 in S3 Appendix. For robustness, in Tables C5-C8 in S3 Appendix, we also report the estimates of a model with interaction terms and p-values obtained applying the Sidak's and the Holm's adjustment for multiple testing. Albeit the two models lead to very similar conclusions, in what follows, we will refer mostly to the interaction model to assess the statistical significance of the heterogeneous analyses.
Inlook at two predetermined characteristics and an indicator of family background.focuses on gender and reports mean values of Health-Economy Trade-off for each treatment, separately for men and woman. We can see that for both genders the variable Health-Economy Trade-off has the highest mean value when health is associated with a positive framing; however, females show a large and statistically significant shift in preferences over the health-economy trade-off towards health-oriented policies when health is framed using the positive word "protection". Differences across gender are statistically significant for what concerns the comparison between HP-EC treatment but not statistically significant when considering the other treatmentsColumn 1 in S3 Appendix). This evidence of females' higher sensibility to the positive framing for health is in line withwho find that on average females are more likely to perceive Covid-19 as a very serious health problem and to agree and comply with precautionary behaviors.
Inwe look at Age and split the sample into students with an age higher than (or equal to) 22 (the average) and students younger than 22 years old. Again, for both categories it emerges that there is a positive effect of the HP-EC treatment; nonetheless, the differences are not statistically significant across age groups for all treatments, Column 2 in S3 Appendix). Inwe split our sample according to parents' education (above and below the median). We find that the HP-EC treatment shifts the preferences towards health for both groups in a significant way. However, differences across groups in the effect of the treatment are not statistically significant when using Sidak-Holm adjustment, Column 3). Interestingly, compared with the HC-EC treatments, students with more highly educated parents when exposed to HP-EP treatment increase their favor towards policies focusing on health concerns quite substantially, albeit in a not statistically significant way, Column 3 in S3 Appendix) and this can be due to a high variance in the responses. Heterogeneity across the other dimensions, such as the number of squared meters available for each person in the house, does not produce any significant effect.
Ininvestigate whether students' reactions to how communication is framed are related to their self-reported personality traits. We here consider only altruism, trustworthiness and extroversion because students claiming to be opened to experience and neurotic have preferences very close to those not self-defining likewise.looks at altruistic students who, according to estimates shown in, tend to prefer policies that focus on the health side. The graph clearly shows a significantly stronger effect of the HP-EC treatment on these students compared to others who did not see themselves as particularly altruistic, albeit this difference is not statistically significant across the two groups, column 1 in S3 Appendix). Likewise, when looking at Trustworthy studentswe find that they report
## Plos one
The health-economy trade-off during the Covid-19 pandemic preferences significantly shifted towards health-oriented policies (higher values of Health-Economy Trade-off) in the HP-EC treatment compared with the HC-EC treatment. Compared with other students, they also seem to favor more health in all treatments except for the HC-EC treatment and this difference is also statistically significant for HP-EC and HC-EP, column 2 in S3 Appendix). This result is important since it shows that trust can be an important lever to influence preferences during an emergency situation and build up consensus towards certain types of policies. Finally, when looking at extroversion in, it emerges that students who consider themselves as extrovert are more influenced by the HP-EC treatment than their not extrovert counterparts but in a not statistically significant way, column 3 in S3 Appendix).
We have also examined the role of students' exposure to the Covid-19 emergency in terms of both health and economic implications. Firstly, we have considered whether students more directly exposed to Covid-19, because relatives or friends tested positive to the virus, are less or more influenced by framing.shows that preferences are overall similar across students for what concerns HP-EC treatment. The only relevant difference is that individuals who more closely experienced the epidemic show preferences more shifted towards health policies across all treatments (in comparison with HC-EC), while students who have not closely experienced the epidemic seem to have preferences more in favor of policies that tend to limit the impact of the crisis on the economy when not exposed to HP-EC treatment. Differences across these groups are, however, not statistically significant , column 1 in S3 Appendix). Then, to consider personal exposure to the economic crisis, we have split the sample
## Plos one
The health-economy trade-off during the Covid-19 pandemic according to whether or not students' parents have lost their jobs due to the pandemic. We find similar preferences in all the treatments.
Additionally, as the way in which individuals react to the communication messages they receive also depends on their psychological conditions, inand 4(D) we split our sample considering the depression and anxiety severity scales we have described in Section 3. The figures confirm the effect of a positive framing for health concerns. In particular, we find that for students who feel more depressed, the increase in preferences for more health-oriented policies when the trade-off is expressed in terms of protection of health and costs for the worsening of the economic situation-instead of in terms of costs for both health and the economy -is almost twice as large as the effect found for students in better psychological conditions. Differences across these groups are anyway not statistically significant when using Sidak-Holm adjustment , columns 3 and 4 in S3 Appendix).
As a last heterogeneous effect analysis, we also look at whether the framing has a differential effect among students with different baseline preferences over the health-economy trade-off.the average value of Health-Economy Trade-off for students who, before the treatment, indicated to having economic oriented, middle or health-oriented preferences, respectively. The graph shows that albeit baseline and ex-post preferences are positively correlated, the treatment is able to shift preferences towards Health-Centered Preferences quite importantly. Indeed, when the positive framing is used for the health side of the trade-off, even those with ex-ante more economic oriented preferences shift towards policies that assign higher value to health and indeed these students are those experiencing the highest treatment effect as compared with the HC-EC treatment. We do not find a specular effect when using the positive framing for the economic side of the trade-off. Differences across these groups are
## Plos one
The health-economy trade-off during the Covid-19 pandemic anyway not statistically significant when using Sidak-Holm adjustment.
All in all, our analysis shows that, a simple and zero cost communication strategy that associates a different framing to the two sides of the trade-off has a widespread effect with very few statistically significant differences across different audiences. The only noteworthy exceptions are represented by females and trustworthy students who seem to react significantly more to the HP-EC treatment with respect to their counterparts.
## Communication and compliance with prescribed behaviors
Our analysis so far has shown that communication style affects individual preferences over the health-economy trade-off of Covid-19 related policies. In this section, we take a further step by looking at how preferences over the trade-off correlate with intentions to adhere to behaviors that have been suggested as useful tools to limit the spread of the epidemic (i.e. wash hands; avoid touching eyes, nose, mouth; stay at least two meters from other persons, stay at home with symptoms of coronavirus). These are measured by the two proxies described in Section 2, Compliance PCA and Compliance. Even if, as discussed by, self-reported measures of compliance with prescribed behaviors might suffer of desirability bias, this would not qualitatively affect our results if the desirability bias is not differentiated in relation to the variable of interest, Health-Economy Trade-off, and ultimately in relation to our treatments (which, as discussed below, are used as instruments to deal with potential endogeneity in preferences over the health-economy trade-off).
Inwe report OLS estimation results investigating the relationship between intention to adhere to prescribed behaviors and preferences for health centered policies (Health-Economy Trade-off). In odd columns we consider as the outcome variable Compliance PCA, while in even ones we use Compliance. As shown in columns (1) and (2), without controls, we find that Health-Economy Trade-off is positively and significantly correlated with both measures of compliance. The same results hold true when we add among the regressors individual and family characteristics and our measure of baseline preferences (columns 3 and 4) and the full set of controls (columns 5 and 6). As regards control variables, we find that women are more likely to follow precautionary behaviors. A positive correlation is found also for Baseline Health-Economy Trade-off, Parents' education, Anxiety severity index, Trustworthy.
The positive correlation between Health-Economy Trade-off and individual compliance shown inindicates that students with preferences for health-centered policies are generally more likely to have an intention to adhere with prescribed behaviors. However, this does not imply causation since it is possible that unobserved factors associated with both the perceived trade-off and compliance cause a spurious correlation between these two variables. Thus, in order to gain a better understanding of the extent to which preferences regarding the health-economy trade-off causally affect adherence with prescribed behaviors, following, we adopt an instrumental variable approach. This strategy strongly relies on the availability of a valid instrument, that is a variable that (1) affects the endogenous variable (relevance) butshows no independent association with the outcome variable for reasons beyond its effect on the endogenous regressor (exclusion restriction) and (3) does not share common causes with the outcome variable (independence). Finding instruments satisfying these conditions is typically difficult, however, thanks to our experiment, we can exploit the exogenous variation in the perceived trade-off induced by our treatments. More precisely, we take advantage of the fact that assignment to the treatments is random and can be used as an exogenous instrument predicting preferences for health-centered policies in the first-stage. In fact, given the random assignment of individuals to treatment conditions, exposure to our treatments is uncorrelated with the error term, then satisfying a crucial condition for a valid instrument. In addition, as shown in the previous section, the HP-EC and the HC-EP treatments produce a significant effect on the suspected endogenous variable Health-Economy Trade-off. As regards the
## Plos one
The health-economy trade-off during the Covid-19 pandemic exclusion restriction assumption, we rely on the idea that assignment to treatment does not produce any direct effect on compliance and all the impact on the outcome variable is mediated by the impact produced on individual preferences. Inwe report reduced form estimates. We estimate specifications (5) and (6) reported in, including the full set of controls and using alternatively the two outcome variables Compliance PCA and Compliance. In the first two columns, we add to the independent variables the HP-EC treatment, which produces a positive and statistically significant effect on both measures of compliance. In columns (3) and (4) we add also the HC-EP treatment. The effect of the HP-EC treatment is still positive and statistically significant while the HC-EP treatment does not produce any impact. Finally, in the last two columns we include all the treatments. We still find a positive impact of the HP-EC treatment on the outcome variables, but estimates are less precise.
Based also on the results shown in the previous table, in our preferred 2SLS specification we only use the HP-EC treatment as an instrument for the endogenous variable. However, we also experiment with alternative specifications adding as instruments the other treatments. Results of this analysis are reported in. We again estimate specifications (5) and (6) reported in, including the full set of controls and using alternatively the two outcome variables Compliance and Compliance PCA. In the first two columns, we use as an instrument the HP-EC treatment. First-stage regression results (Panel B) confirm a strong and significant effect of the treatment HP-EC on the perceived trade-off. First stage F-test statistics (32.17) is well above the common threshold of 10 used to detect weak instruments. Importantly, second stage regressions (Panel A) show a positive and statistically significant effect of the perceived trade-off on both measures of compliance.
In columns (3) and (4) we use as instruments both HP-EC and HC-EP-the two treatments producing an effect on the endogenous variable-and find results in line with those discussed above. The Hansen test with a p-value of 0.260 and 0.123 in the two specifications indicates that the overidentifying restrictions are not rejected. F-statistics for the test that the coefficients of the instruments in the First Stage are jointly zero are again above the rule of thumb threshold suggested by(18.07). Finally, in columns (5) andin order to exploit all the variability in our data we use the all the treatments. The F-test is equal to 12.25 and the Hansen tests has a p-value of 0.272 and 0.146 in the two specifications. Again, results are qualitatively similar to those discussed above. It worthwhile to notice that the magnitude of the impact produced by individual preferences (Health-Economy Trade-off) on the intention to adhere to prescribed behaviors changes in relation to the instrument used. This suggests that they are likely to generate different sets of compliers and, when we rely on all the sets of possible instruments, the estimated Local Average Treatment Effect (LATE) applies to a less specific group of individuals.
These results taken together further confirm that the type of communication we have analyzed in this paper affects intentions to adhere to prescribed behaviors through a switch in preferences over the health-economy trade-off.
# Conclusions
The management of the health emergency by Covid-19 represents a great public challenge that requires a massive effort in terms of individual cooperation in order to limit the diffusion of the epidemic. The role of public communication, especially in the absence of financial incentives, has been recognized by several studies as decisive in order to ensure individual compliance with recommended behaviors. In particular, a key issue to be addressed concerns the management of the trade-off between public health and economic outcomes.
In this paper, we study how young individuals balance this trade-off during the pandemic and how the communication strategy over this trade-off affects their preferences for policies aimed at managing the restart of economic and social activities, and, ultimately their intention to adhere to prescribed behaviors. We investigate this issue in Italy-one of the country most affected by the outbreak-using a field experiment involving around 2000 students who took part in a survey administered during the period 20 th April -25 th April, i.e. at the beginning of the Covid-19 pandemic and before the end of the first lockdown period. In our analysis we
## Plos one
The health-economy trade-off during the Covid-19 pandemic compare a positive framing which focuses on protective strategies ("protection") with a negative framing focusing on potential losses ("costs").
Our results show that a policy focusing on the protection of health and the costs for the worsening of the economic situation induces students to give more weight to health issues than when the trade-off is articulated in terms of costs for both health and economy. The effect is substantial, highly effective across different typologies of audiences, especially females, and associated with a higher intention to comply with precautionary behaviors. We find that 47.36% of students responded that they would consider 'extremely' or 'very much' health when framed as protection versus economic costs, while this share reduces to 34.15% in the group having both elements of the trade-off framed as costs.
These results pertaining to a specific group of individuals cannot be extended to the whole population. Nonetheless, as the behavior of young individuals can be crucial in order to control the spread of the infection, they have important policy implications. They suggest that the communication strategy during an emergency-such as that originating from the diffusion of the Covid-19 pandemic-plays a critical role and that a positive framing that focuses on the "protection" of the health conditions is likely to significantly affect individual preferences over the health dimension of the crisis. Under the assumption that self-reported preferences do not deviate significantly from real preferences, we may speculate that such a communication is likely to increase political consensus and may represent a costless strategy to ensure higher compliance with recommendations in the phases following the end of lockdown measures.
Being able to shape individual preferences over the health-economy trade-off, especially with cost-effective measures, is particularly important as such preferences affect individual decision to comply with behaviors that have been strongly recommended by doctors and specialists since the onset of the emergency in order to limit the spread of the virus. Exploiting the random assignment to the treatments in an instrumental variables framework, we provide causal evidence of a positive effect of health-oriented preferences on compliance. This suggests that an effective communication strategy may be a way to induce an otherwise non-incentivized active role in the defeat of the epidemic.
Moreover, our paper shows that characteristics such as personal attitudes, specific knowledge (i.e. the field of study) and state-dependent conditions affect preferences for the healtheconomy trade-off during the Covid-19 pandemic, regardless of the framing of the communication. Among these, the differences due to socio-economic background may pose important policy concerns. In many countries, current political debate is dominated by very polarized positions over the priorities to be given to the management of the reopening phase. Our paper suggests that the asymmetric economic consequences of the pandemic might explain these differences. One implication of this result is that the decision to financially help people who faced large economic shocks may also be supported as a way to strengthen social cohesion and preferences alignment over the management of the Covid-19 pandemic.
Finally, we find an interesting gender differential in the impact of the framing of the communication on preferences over the trade-off that might deserve further exploration. Despite the fact that when the survey was administered the health consequences of the Covid-19 virus seemed to be less pronounced among women, we find that they are significantly more affected by a positive framing focusing on the protection of the health conditions. Whether this depends on gender specific attitudes or on the role model of the male breadwinner might be a nice area of future research.
## Supporting information
|
Health Literacy Regarding Infectious Disease Predicts COVID-19 Preventive Behaviors: A Pathway Analysis
Health literacy has been identified as one vital determinant of public health and healthy behaviors, but very few studies regarding infectious disease prevention have been found. This descriptive cross-sectional study aimed to validate the pathway of infectious disease-specific health literacy (IDSHL), COVID-19 (coronavirus disease 2019) preventive behaviors, and their determinants. A sample of 1459 casino workers in Macao was eligible for analysis. The concept model was verified with a comparative fit index of 0.937 and goodness-of-fit index of 0.971. Government responses was a significant determinant of situational factors (helpfulness of health information, resource accessibility, and organizational training adequacy), while situational factors showed a direct effect on COVID-19 preventive behaviors. Education and organization training adequacy was the strongest influencing factor of IDSHL, which should be a key target of intervention programs for COVID-19.
## What we already know
- - The 2019 novel coronavirus disease (COVID-19) has created a major public health threat and emergency worldwide. - - Health literacy was suggested to be a very important investment in our society for better prevention of pandemics such as COVID-19. - - The theory model between health literacy and prevention behaviors in pandemic has not been reported in original studies.
## What this article adds
- - This article verified a systemic theory model of health literacy and preventive behavior for the first time regarding infectious disease in the COVID-19 pandemic. - - Infectious-disease-specific health literacy should be a key target of intervention programs for COVID prevention, especially at the community or organization level. - - Government policies can promote individual preventive behaviors by creating a supportive atmosphere, especially in terms of effective health information delivery.
# Introduction
The 2019 novel coronavirus disease (COVID-19) has created a major public health threat and emergency worldwide. [bib_ref] Unique epidemiological and clinical features of the emerging 2019 novel coronavirus pneumonia..., Wang [/bib_ref] To mitigate the spread of the virus and alleviate the loss caused by the pandemic, significant shifts in daily behavior are highly recommended by the relevant guidelines.At the same time, to react more effectively to this "extremely serious incident," a national emergency response, such as restrictive border control and strict social distancing rules, has been introduced and implemented in many countries and regions, [bib_ref] An investigation of transmission control measures during the first 50 days of..., Tian [/bib_ref] including Macao, China. Macao, famous for its gambling industry, is the most densely populated region in the world but the pandemic has been well controlled,without new cases confirmed for more than 9 months. The successful experience for pandemic management is an informative reference for other countries and regions. Health literacy has been identified as one vital determinant of public health at the population level and has become a heated topic of research since its introduction in the 1970s. [bib_ref] HLS-EU) Consortium Health Literacy Project European. Health literacy and public health: a..., Sørensen [/bib_ref] Recently, many researchers have highlighted the importance of health literacy in the prevention of COVID-19. [bib_ref] Readability of online patient education material for the novel coronavirus disease (COVID-19):..., Szmuda [/bib_ref] [bib_ref] COVID-19: health literacy is an underestimated problem, Paakkari [/bib_ref] However, very few studies have been found regarding infectious disease prevention.
Sørensen et al [bib_ref] HLS-EU) Consortium Health Literacy Project European. Health literacy and public health: a..., Sørensen [/bib_ref] have composed an integrated definition and conceptual model of health literacy based on systemic literature review. In this model, both distal factors and proximal factors significantly influence health literacy. Societal and environmental predictors are particular distal factors (eg, culture, political forces, and societal systems). Proximal factors include personal demographic predictors and situational predictors (eg, media use and social support). These distal and proximal factors can also have a significant effect on each other and healthy behaviors. [bib_ref] HLS-EU) Consortium Health Literacy Project European. Health literacy and public health: a..., Sørensen [/bib_ref] Although this comprehensive model has been proposed based on a systematic review, it needs more statistical validation, similar to other models. [bib_ref] Determinants of health literacy and health behavior regarding infectious respiratory diseases: a..., Sun [/bib_ref] The purpose of this study was to verify the conceptual model incorporated by Sørensen et al 7 in the COVID-19 prevention, as shown in [fig_ref] Figure 1: Conceptual model of health literacy and COVID-19 [/fig_ref] , with the hope of providing significant references for future responses to the current pandemic as well as other new infectious disease outbreaks in the future.
# Methods
## Participants and data collection
This descriptive study was approved by the institutional review board at Kiang Wu Nursing College of Macau (#2020FEB02). The participants were recruited using quota sampling,sampled from all 6 concessionaries that operated casinos in Macao between April 20 and May 22, 2020, approximately 1 to 2 months after the casino reopening amid the pandemic. The inclusion criteria were full-time adult employees in casinos who could answer Chinese e-questionnaires online and consent to participate in this study. The minimum sample size was 1067. Considering the efficiency rates of the questionnaire, we recruited at least 1200 participants.
Written consent was confirmed at first by the directors of 6 concessionaries. Then, middle managers related to human resource management were contacted for the sample recruitment process after a target sample size of occupation distribution was provided. All questionnaires with a written consent form were input into an online survey platform, and a poster with a QR code and website link was created. Either emails or printed posters were used to conveniently collect data in the different concessionaries. To achieve a representative occupation composition, the survey ended when the minimum sample size was achieved. Middle managers were contacted to send more invitations to candidates with poorly sampled occupations. The survey ended on the day when the sample size reached 1500, which was approximately 25% higher than planned.
## Measurements
Infectious-Disease-Specific Health Literacy (IDSHL) Scale. After thoroughly examining the items of each IDSHL scale available, we decided to use the questionnaire validated by Tian et al, [bib_ref] Study on the development of an infectious disease-specific health literacy scale in..., Tian [/bib_ref] which seems more stable to examine health literacy regarding infectious diseases during a pandemic.
The IDSHL scale was divided into 2 subscales 12 : skills to prevent infectious disease (22 items, with an overall Cronbach's coefficient of 0.839) and cognitive ability (6 items requiring reading and comprehension). The first subscale contained 4 domains: infectious-disease-related knowledge (7 items), disease prevention (7 items), management or treatment (4 items), and pathogen and infection source identification (4 items). The subscale score was calculated using a weight coefficient by difficulty. The total subscale score was 38.62, and the participant was considered to have adequate IDSHL at the cutoff point of 16.74. In our study, only the first subscale was applied to make the survey more user friendly. [bib_ref] Study on the development of an infectious disease-specific health literacy scale in..., Tian [/bib_ref] The overall Cronbach's coefficient of the subscale in our study was 0.815.
## Covid-19 preventive behavior questionnaire (pbq).
A selfmade questionnaire measuring COVID-19 preventive behaviors was composed according to the latest guidelines published by the World Health Organization, Chinese Center for Disease Control and Prevention, and Health Bureau of Macao.The instrument consisted of 17 items: 3 regarding social distance, 5 related to prevention skills, 4 concerning the environment, and 5 about health monitoring and promotion. 2,3,13 Each item was rated using a Likert-type 5-point scale (from none to always). As a result, the total score of the COVID-19 PBQ was 85. Six public health experts were asked to evaluate the content validity of this self-made questionnaire. The universal agreement of the scale was 0.9. The item-level content validity index ranged from 0.83 to 1.00, and the average CVI (Content Validity Index) of the scale was 0.98. A pretest was used to test the face validity of the questionnaire, and 1 item was reworded to clarify its meaning. During the pilot study in 131 participants, the overall internal consistency reliability was sufficient, [bib_ref] How reliable are measurement scales? External factors with indirect influence on reliability..., Ursachi [/bib_ref] with a Cronbach's α coefficient of 0.898, and subscale Cronbach's coefficients ranged from 0.656 to 0.823.
Predicting Factor Questionnaire. The predictors of IDSHL and COVID-19 preventive behaviors were measured using a self-made questionnaire based on the concept model of health literacy created by Sørensen et al 7 and literature review. The distal factors in this study were related to government responses to the pandemic. 7,15, [bib_ref] Infectious disease policy: towards the production of health, Porter [/bib_ref] We asked participants to choose the 3 (out of 8) most important policies made by the Macao government, including providing masks, regulations about wearing masks in public, information about prevention, and so forth. A number rated scale (NRS) 0 to 10 was applied to determine the helpfulness of these policies: 0 = totally unhelpful to 10 = extremely helpful.
The proximal predicting factors in the questionnaire included demographic information (age, gender, marriage, education, living area, identification in Macao, occupation, and working experience) and situational predictors (media use, resource accessibility, and organizational training adequacy). [bib_ref] Emerging infectious disease prevention: where should we invest our resources and efforts?, Ellwanger [/bib_ref] [bib_ref] Infectious disease policy: towards the production of health, Porter [/bib_ref] [bib_ref] A study on consumer health information of Taiwan government agencies of health..., Cheng [/bib_ref] [bib_ref] Public health literacy defined, Freedman [/bib_ref] Media use included categories of the most important health information related to COVID-19 prevention and the helpfulness of this information (using NRS 0-10). Three reverse questions regarding resource accessibility (with a Cronbach's coefficient of 0.821) were asked: difficulty in buying resources (masks, disinfection supplies, and living resources) with a 5-point Likert-type scale (0 = none at all, 5 = completely difficult). Organizational training adequacy was rated on a 5-point Likert-type scale (0 = none, 5 = completely enough).
# Statistical analysis
Data were exported from the online questionnaire platform and imported into IBM SPSS statistics 22. Descriptive statistics were applied to demonstrate the demographic characteristics, COVID-19 preventive behaviors, and IDSHL. One-way ANOVA (analysis of variance) was adopted to compare the differences among demographic groups, and the Student-Newman-Keuls method was applied to compare the differences among subgroups if a significant difference was found. The Pearson correlation coefficient was used to examine the relationships among IDSHL, COVID-19 preventive behaviors, and their predictors, which are continuous variables.
Path analysis was implemented by IBM SPSS Amos 26 Graphics. Maximum likelihood estimation was conducted to evaluate the parameters with the covariance matrix. The path model was modified according to the P regression weight, modification indicators, and goodness-of-fit indexes (GFIs). Only variables with significant P values (less than .05) remained in the model. A comparative fit index ≥0.90 and root mean square error of approximation <0.06 means that the fit was acceptable. 19
# Results
# Descriptive analysis
A sample of 1513 casino workers responded to our study, 1459 of which (96.4%) were eligible for analysis. The average age of the participants was 37.85 ± 9.27 (range from 21 to 66) years, and 65.5% of them were female [fig_ref] Table 1: Differences of IDSHL and PBQ Among Different Demographic Groups [/fig_ref]. The greatest proportion of occupations was dealers (36.1%), followed by clerks (25.9%) and managers (19.1%).
The average IDSHL score was 22.87 ± 6.81 (ranging from 0 to 38.62), and 83.1% of participants were identified with adequate IDSHL. In terms of the COVID-19 PBQ, the average total score was 71.06 ± 9.64 (ranging from 21 to 85), and the mean item score was 4.18 ± 0.57 (4 = often, 5 = always).
The most important government responses and health information about COVID-19 were demonstrated in Supplemental Tables 1 and 2, available online. Mask supply is the top-ranked government policy, followed by the enactment of policy to wear facemasks in public areas. COVID-19 prevention knowledge and skills were the most popular health information. [fig_ref] Table 1: Differences of IDSHL and PBQ Among Different Demographic Groups [/fig_ref] demonstrates the differences among different social demographic groups on the IDSHL scale and COVID-19 PBQ. Gender, marriage, and occupation were correlated with IDSHL but not with the COVID-19 PBQ. Age, the helpfulness of government policies, the helpfulness of health information, and organizational training adequacy were all significantly associated with IDSHL and the COVID-19 PBQ (P < .01; Supplemental , available online), while education was only related to IDSHL, not to preventive behaviors.
# Univariate analysis
# Pathway analysis
A pathway analysis was performed according to the conceptual model proposed in [fig_ref] Figure 1: Conceptual model of health literacy and COVID-19 [/fig_ref]. The final path model was verified with satisfying fit indexes, [bib_ref] Emerging infectious disease prevention: where should we invest our resources and efforts?, Ellwanger [/bib_ref] as shown in [fig_ref] Figure 2: Verified conceptual model of health literacy and COVID-19 preventive behavior [/fig_ref] , with a CFI of 0.937 and root mean square error of approximation of 0.058. Other fit indexes were also satisfying or acceptable [bib_ref] Emerging infectious disease prevention: where should we invest our resources and efforts?, Ellwanger [/bib_ref] : the GFI = 0.971; GFI adjusted for degrees of freedom = 0.951; root mean square residual (RMR) = 0.386; chi-square = 230.565, df = 39, P < .001, chi-square minimum/df (CMIN/df) = 5.912; and standardized root mean square residual = 0.0496.
The relationships among the COVID-19 PBQ, IDSHL, and their determinants are demonstrated in [fig_ref] Figure 2: Verified conceptual model of health literacy and COVID-19 preventive behavior [/fig_ref]. The usefulness of government policies directly affects proximal situational predictors, with the largest loading factor on health information (0.85), followed by organizational training adequacy (0.23) and resource accessibility (0.15). These situational predictors are important determinants of the COVID-19 PBQ, as well as IDSHL. Demographic data (gender, marriage, education, identification in Macao, and years working in casinos) influenced COVID-19 PBQ with a mediating effect of IDSHL. A total of 20% of the variance in the COVID-19 PBQ was explained by the verified model.
# Discussion
This study verified a conceptual model demonstrating the relationships among health literacy regarding infectious disease and preventive behaviors during the COVID-19 outbreak, as well as the distal and proximal predictors. Abbreviations: IDSHL, infectious disease specific health literacy; PBQ, preventive behavior questionnaire. ***P < .001, **P < .01, *P < .05. The Student-Newman-Keuls method was applied to control the total α value. There were significant differences between § , # , and £ groups, but no significant differences within each group.
Demographic variables only affect the COVID-19 PBQ indirectly, with a mediating effect of IDSHL. Education is the most important determinant of IDSHL. The government response was acting as an indirect influencing factor on COVID-19 PBQ through situational social factors, including media usage or health information, organizational training, and prevention resource. IDSHL was an important mediator in the relationship between background data and preventive behaviors, which should be considered seriously when composing public interventions. According to a study in Hong Kong during SARS (severe acute respiratory syndrome), the population's high preventive behaviors did not necessarily increase their IDSHL. [bib_ref] Impacts of SARS on health-seeking behaviors in general population in Hong Kong, Lau [/bib_ref] Even very basic skills such as wearing face masks and hand hygiene may require more interventions in the general population. According to a survey in Hong Kong during the nonepidemic period, none of the participants performed all the required steps when wearing a face mask, and hand hygiene was not taken seriously by the majority of the participants before putting on (91.5%), taking off (97.3%) or after disposing (91.5%) of face masks. [bib_ref] Practice and technique of using face mask amongst adults in the community:..., Lee [/bib_ref] Improving IDSHL is a very urgent task in the field of community public health management, especially in this society with the increasing presence of emerging infectious diseases.
The most prominent influence of government response on social factors was that on media usage or health information.
How to obtain accurate health information is part of health literacy and, thus, should be taken seriously. In our study, the most important health information included COVID-19related knowledge and skills for prevention and governmental policies, and similar results were found in Greece. [bib_ref] Achieving a covid-19 free country: citizens preventive measures and communication pathways, Kamenidou [/bib_ref] According to a study by Basch et al, [bib_ref] The role of YouTube and the entertainment industry in saving lives by..., Basch [/bib_ref] videos about practices to alleviate community transmission on YouTube were viewed by a tremendous portion of the population, but fewer than half of these videos contained any of the prevention behaviors recommended by the US Centers for Disease Control and Prevention. And educational information online was too difficult to understand for the public. [bib_ref] Readability of online patient education material for the novel coronavirus disease (COVID-19):..., Szmuda [/bib_ref] Therefore, improving health literacy is not only an individual issue but also a governmental and professional issue. How to convey health information effectively and promptly will affect the population's health literacy and public health in society. In our study, the most important health information was identified as prevention-related knowledge and skills, and government-published policies. To avoid the irrational panic brought about by myths and misinformation, daily press conferences have been broadcast to provide the latest epidemicrelated information and preventive measures.Many videos, leaflets, and announcements were available in public areas. Additionally, the government uses many different languages to disseminate health information to accommodate the multilanguage culture in Macao.These types of dissemination is believed to improve individual compliance with preventive behaviors by creating a supportive atmosphere.
The relationship between demographic variables and health literacy have been discussed in recent studies. For instance, education is the most important determinant of IDSHL in our study, which is consistent with a previous study, [bib_ref] Determinants of health literacy and health behavior regarding infectious respiratory diseases: a..., Sun [/bib_ref] [bib_ref] Study on the development of an infectious disease-specific health literacy scale in..., Tian [/bib_ref] [bib_ref] Health literacy in Beijing: an assessment of adults' knowledge and skills regarding..., Zhang [/bib_ref] because education can not only affect health literacy directly but also demonstrate an indirect effect with a mediating effect of prior knowledge. [bib_ref] Determinants of health literacy and health behavior regarding infectious respiratory diseases: a..., Sun [/bib_ref] In the present study, education only indirectly affected preventive behaviors by mediating the effect of IDSHL. Lower educational status was a social determinant of the spread of COVID-19 within the community, 26 due to the lower knowledge level and poorer preventive practices of such individuals. [bib_ref] Knowledge, attitudes, risk perceptions, and practices of adults toward COVID-19: a population..., Honarvar [/bib_ref] During the H1N1 2009 pandemic, education was also positively associated with perception and compliance with preventive behaviors in the Chinese population, and knowledge about the disease also influenced risk perception and prevention, [bib_ref] Knowledge, attitudes and practices (KAP) related to the pandemic (H1N1) 2009 among..., Lin [/bib_ref] which was consistent with another study showing that knowledge had a mediating effect on the relationship between education and health literacy. [bib_ref] Determinants of health literacy and health behavior regarding infectious respiratory diseases: a..., Sun [/bib_ref] Organizational training adequacy and the helpfulness of health information were also found to be the 2 most significant factors influencing the COVID-19 PBQ, which were also prominently influenced by government response. Many industry-specific prevention recommendations were published by the Health Bureau in Macao, which supported organizational training. Information exchange may play an important role in organizational training, a type of targeted health information that is more appropriate and easier to apply during work and life. Nazir et al 29 indicated that social media exposure influenced preventive behavior through an indirect path with the mediators of awareness and information exchange rather than a direct effect. Similarly, the willingness to share information and sufficient knowledge on epidemic prevention can promote preventive behaviors during COVID-19 in Taiwan participants. [bib_ref] How to defend COVID-19 in Taiwan? Talk about people's disease awareness, attitudes,..., Hsu [/bib_ref] The accessibility of prevention-related resources and supplies directly influenced preventive behaviors and has been dramatically important during COVID-19, especially when there was a global lack of medical resources. Many people have had to find their alternatives, but Macao has succeeded in supplying masks to residents.In our study, 90.7% of participants felt no difficulty or little difficulty in buying facemasks.
# Limitations
We aimed to verify a theoretical model of health literacy related to infectious disease in the current COVID-19 pandemic, which is the first study, so we only measured those variables using self-made questionnaires. Additionally, only casino workers were recruited in the research; thus, the theory model needs to be examined in other populations. However, organizational training was highlighted in this homogeneous sample, which provide very valuable for reference in other occupations.
# Conclusions
In summary, the proposed conceptual model was well verified in our study. IDSHL could be a key target of intervention programs for COVID-19 and future pandemic prevention. This research also highlighted the function of government response on individual COVID-19 prevention behaviors by creating a supportive environment in the community or organization, which may have explained the secret of the success of pandemic prevention in Macao and China.
[fig] Figure 1: Conceptual model of health literacy and COVID-19 (coronavirus disease 2019) preventive behavior. [/fig]
[fig] Figure 2: Verified conceptual model of health literacy and COVID-19 preventive behavior. [/fig]
[table] Table 1: Differences of IDSHL and PBQ Among Different Demographic Groups. [/table]
|
Electromuscular Incapacitating Devices Discharge and Risk of Severe Bradycardia
Electromuscular incapacitating devices (EMDs) are highvoltage, low-current stimulators causing involuntary muscle contractions and sensory response. Existing evidence about cardiac effects of EMD remains inconclusive. The aim of our study was to analyze electrocardiographic, echocardiographic, and microvolt T-wave alternans (MTWA) changes induced by EMD discharge.We examined 26 volunteers (22 men; median age 30 years) who underwent single standard 5-second duration exposure to TASER X26 under continuous echocardiographic and electrocardiographic monitoring. Microvolt T-wave alternans testing was performed at baseline (MTWA-1), as well as immediately and 60 minutes after EMD exposure (MTWA-2 and MTWA-3, respectively).Mean heart rate (HR) increased significantly from 88 ± 17 beats per minute before to 129 ± 17 beats per minute after exposure (P < 0.001). However, in 2 individuals, an abrupt decrease in HR was observed. In one of them, interval between two consecutive beats increased up to 1.7 seconds during the discharge. New onset of supraventricular premature beats was observed after discharge in 1 patient. Results of MTWA-1, MTWA-2, and MTWA-3 tests were positive in one of the subjects, each time in a different case.Standard EMD exposure can be associated with a nonuniform reaction of HR and followed by heart rhythm disturbances. New MTWA positivity can reflect either the effect of EMD exposure or a potential false positivity of MTWA assessments.
E lectromuscular incapacitating devices (EMDs) are frequently used by law-enforcement officers or private persons worldwide. They are high-voltage (peak voltage, 1200 V), low-current (peak current, 3 A) stimulators that cause involuntary muscle contractions in combination with sensory response. The electrical stimuli are in the form of high short-duration (10-100 μseconds) and repetitive (10-19 per second) pulses.Existing data concerning adverse cardiac events of EMD including incidental deaths are still inconclusive. A number of animal 2 and human 3-7 studies have reported relative safety of EMD exposure. In contrast, some reports of severe adverse events including ventricular fibrillation [bib_ref] Ventricular fibrillation after stun-gun discharge, Kim [/bib_ref] [bib_ref] Ventricular fibrillation and death after TASER injury, Sadhu [/bib_ref] [bib_ref] Presenting rhythm in sudden deaths temporally proximate to discharge of TASER conducted..., Swerdlow [/bib_ref] [bib_ref] Successful resuscitation of a patient in asystole after a Taser injury using..., Schwarz [/bib_ref] [bib_ref] Ventricular fibrillation in a man shot with a Taser, Naunheim [/bib_ref] [bib_ref] Sudden cardiac arrest and death following application of shocks from a TASER..., Zipes [/bib_ref] or other effects on myocardium [bib_ref] Fortuitous therapeutic effect of Taser shock for a patient in atrial fibrillation, Richards [/bib_ref] from EMD application exist.
Heart rate (HR) acceleration before and after TASER exposure has been repeatedly described. [bib_ref] Physiological effects of a conducted electrical weapon on human subjects, Vilke [/bib_ref] [bib_ref] Cardiovascular and physiologic effects of conducted electrical weapon discharge in resting adults, Ho [/bib_ref] [bib_ref] Twelve-lead electrocardiogram monitoring of subjects before and after voluntary exposure to the..., Vilke [/bib_ref] [bib_ref] Human cardiovascular effects of a new generation conducted electrical weapon, Ho [/bib_ref] However, it is difficult to determine exact HR dynamics during exposure and whether the EMD application caused cardiac electrical capture owing to the fact that EMD discharge causes substantial electrical interference with the electrocardiographic (ECG) recording. Although clearly visible abnormalities in ECG after EMD application have not been described, [bib_ref] Twelve-lead electrocardiogram monitoring of subjects before and after voluntary exposure to the..., Vilke [/bib_ref] very little is known about subtle repolarization changes in relation to EMD. Such changes could be revealed by using microvolt T-wave alternans (MTWA). This method allows us to reveal a subtle ECG pattern in which the amplitude of T-wave varies every other beat, and it is linked to arrhythmogenic mechanism. [bib_ref] Mechanism linking T-wave alternans to the genesis of cardiac fibrillation, Pastore [/bib_ref] The aims of our study were to analyze ECG, echocardiographic, and presence of MTWA before, during, and after EMD exposure and to evaluate the potential cardiac electrical capture during TASER application using continuous echocardiography.
# Materials and methods
We analyzed ECG recordings and MTWA tests in a group of healthy volunteers who underwent single standard 5-second duration exposure to TASER X26 (Taser International, Scottsdale, Ariz) discharge. The study protocol took place in parallel with routine training of EMD user license. The EMD energy application was performed by an authorized TASER instructor who was responsible for training according to standard training protocol. Independently of the training procedures, we obtained a detailed medical history of our volunteers and performed a baseline physical examination. In addition, we recorded a standard 12-lead ECG as well as 2-dimensional and Doppler echocardiography. The study criteria excluded subjects with potentially significant clinical abnormalities as detected by these methods. However, none of the study participants had a relevant ECG and/or echocardiographic pathology. The local ethics committee gave us approval of the protocol. All subjects signed an informed consent to the study. Only healthy volunteers older than 18 years were included in this study. The following study exclusion criteria were set: family history of sudden cardiac death, personal history of structural heart disease and serious cardiac arrhythmias, history of neurological or psychiatric disease, cardiac medication use, drug or alcohol abuse, as well as pregnancy.
## Exposition to taser energy
The EMD energy was applied in supine position with arms parallel to the side of the body. The EMD probes were manually positioned on the subject's body in predefined positions: 1 electrode in the second intercostal space in the right midclavicular line and second electrode in the fifth intercostal space in the left anterior axillary line. The distance between the probes was at least 30 cm. The probes were attached to the skin with a plaster to achieve maximal (probe to skin) contact. The single standard 5-second duration exposure by TASER X26 was delivered. The subject was maintained in the lying position until sinus tachycardia, rhythm disorders, or breathing changes disappeared, at least for 3 minutes. Upon completion of the application protocol, the electrodes were removed and the site of contact was checked for possible injuries.
## The study protocol
The analysis of digital recordings of standard 12-lead ECG was performed. The recordings were monitored using Prucka Cardiolab 7000 (GE Healthcare, United Kingdom). The baseline ECG was analyzed at the beginning of all procedures during standard physical examination. Periprocedural continuous 12lead ECG measurement was recorded. The recordings were initiated 1 minute before exposure and concluded at the end of the third minute after EMD energy delivery. Another ECG measurement was recorded 60 minutes after the EMD shock delivery. Heart rhythm, heart rate (HR ECG ), conduction intervals, P wave, QRS complex, ST-T segment morphology, as well as arrhythmia occurrence were analyzed. Echocardiography (Vivid q; GE Healthcare, United Kingdom) performed to reveal potentially significant clinical pathology before including the subjects in the study. Therefore, echocardiography was used for documentation of echocardiographic HR (HR ECHO ) during TASER discharge during which frequent electrical artifacts occur and make ECG analysis impossible. Cine loops were obtained in modified apical 4-chamber or subcostal view in lying subjects. The recording was initiated 5 seconds before discharge and terminated 10 seconds after discharge. The HR ECHO was calculated as the frequency of systolic contraction of the lateral wall of the left ventricle during a 5-second period of EMD exposure. Owing to limited echocardiographic window, no other parameters were consistently documented during the discharge.
Simultaneously, MTWA testing by treadmill exercise was performed at baseline (MTWA-1), immediately, and after 60 minutes after stun gun delivery (MTWA-2 and MTWA-3, respectively). The spectral method by graded exercise protocol was used. In brief, after careful skin preparation, high-resolution ECG leads (Cambridge Heart, Inc, Tewksbury, Mass) were placed on the standard precordial positions and in an orthogonal configuration. Exercise protocol consisted of gradually increasing workload so that constant HR ECG was archived. The HR ECG was measured between 100 and 110 beats per minute for 150 seconds; then subsequently, the HR increased to zone 110 to 120 beats per minute for 90 seconds. The MTWA test was interpreted with previously described criteria. [bib_ref] Interpretation and classification of microvolt T wave alternans tests, Bloomfield [/bib_ref] The test result was classified as positive, intermediate, or negative. Positive result was defined as the presence of sustained MTWA for at least 1 minute with alternans voltage of 1.9 μV or greater, K ratio less than 3, and an onset HR ECG of 110 beats per minute or less in any orthogonal leads or in 2 adjacent precordial leads. Negative test was defined by unmet criteria for positivity in case HR ECG was greater than 105 beats per minute. In case the test did not meet the criteria for positivity or negativity, it was classified as intermediate. Aside to standard classification, any presence of MTWA criteria was recorded.
# Statistical analysis
All continuous variables were expressed as mean ± standard deviation or as median with range for abnormally distributed variables. The Wilcoxon paired test was used to compare measured values when appropriate. P values less than 0.05 was considered to be significant. All analyses were performed with STATISTICA 6.1 package (StatSoft, Inc, Tulsa, Okla).
# Results
The data were obtained from 26 healthy volunteers (22 male and 4 female, mean age of 30 ± 8; median age 30, range 19 -46 years). The baseline clinical, ECG, and demographic data are shown in [fig_ref] TABLE 1: Clinical and Demographic Data [/fig_ref].
As indicated in [fig_ref] TABLE 2: ECG Parameters at Baseline, During, and After EMD Discharge [/fig_ref] , the median of HR ECG immediately before EMD exposure increased significantly from baseline level (P < 0.01). Another considerable increase in HR ECG was noted immediately after the EMD administration (P < 0.001 in comparison with preshock values). The mean HR ECG decreased back to the baseline level in all patients within 3 minutes after the EMD exposure. The analysis of HR ECHO was available only in 19 subjects (73%). Remaining recordings were invalid because of strong muscle contraction and loss of optimal acoustic window during the EMD exposure. Values of HR ECHO identified by echocardiograph corresponded to those determined by ECG.
The HR ECG changes in all subjects are shown in [fig_ref] FIGURE 1: Heart rate profile in an individual [/fig_ref]. Out of the total cohort, 2 male subjects manifested a different pattern of HR ECG dynamics during the discharge. In those subjects, HR ECG paradoxically decreased during and immediately after the EMD exposure. The HR ECHO profile was documented [fig_ref] TABLE 2: ECG Parameters at Baseline, During, and After EMD Discharge [/fig_ref]. Frequent monomorphic ventricular premature beats (VPBs) with left bundle branch block morphology and inferior axis were noted at baseline in 1 case. Besides VPB absence at the 30-second period after EMD exposure, their frequency was not changed after the EMD shock delivery. The new onset of frequent supraventricular premature beats (SPBs) was detected after EMD exposure in 1 case. The SPBs were not accompanied by any symptoms and spontaneously disappeared within 5 minutes. No other changes in ECG parameters and morphology as well as no severe cardiac arrhythmias were found after the EMD exposure.
The MTWA protocol was performed in 21 (81%) volunteers. The detailed results are summarized in [fig_ref] TABLE 3: Microvolt T-wave Alternans Test ResultsMTWA Test [/fig_ref]. Results of MTWA tests at the baseline (MTWA-1), immediately after (MTWA-2), and after 60 minutes after (MTWA-3) TASER shot gun delivery were negative in 16 (76%), 14 (67%), and 15 (71%) cases, respectively. All results of the MTWA tests 1, 2, and 3 were positive in 1 (5%) of 21 subjects, different person on each occasion. Remaining tests (24% of all tests) were indeterminate. The reason for the indetermination was low quality of signals in 10 cases and missing criteria of full negative or positive result in 3 cases.
# Discussion
The major finding of our study is the detection of nonuniform reaction of HR in response to EMD exposure. Although HR acceleration was present in the majority of subjects, bradycardia was detected during exposure to EMD in 1 case with echocardiography. In another subject, a similar decrease in HR was visible immediately after terminating the EMD energy administration in ECG. In addition, in 1 case, we observed a series of asymptomatic frequent SPB after EMD exposure. All 3 MTWA tests had positive results in one of the tested subjects, different person on each occasion.
It is generally accepted that increased HR is triggered by physiologic stress accompanied with adrenergic stimulation owing to an anticipated physical inconvenience. [bib_ref] Physiological effects of a conducted electrical weapon on human subjects, Vilke [/bib_ref] [bib_ref] Cardiovascular and physiologic effects of conducted electrical weapon discharge in resting adults, Ho [/bib_ref] [bib_ref] Twelve-lead electrocardiogram monitoring of subjects before and after voluntary exposure to the..., Vilke [/bib_ref] [bib_ref] Human cardiovascular effects of a new generation conducted electrical weapon, Ho [/bib_ref] [bib_ref] Acidosis and catecholamine evaluation following simulated law enforcement "use of force" encounters, Ho [/bib_ref] This hypothesis is supported by previously described systolic blood pressure and minute ventilation increase imminently before stun gun discharge. [bib_ref] Twelve-lead electrocardiogram monitoring of subjects before and after voluntary exposure to the..., Vilke [/bib_ref] [bib_ref] The cardiovascular, respiratory, and metabolic effects of a long duration electronic control..., Dawes [/bib_ref] In our opinion, the onset of frequent SPB after TASER exposure in 1 case may have an analogous adrenergic mechanism. [bib_ref] Acidosis and catecholamine evaluation following simulated law enforcement "use of force" encounters, Ho [/bib_ref] The dynamics of SPB frequency and their disappearance during the subsequent observation correlated with HR decrease during postexposure observation period. Of note, another volunteer presented frequent fully asymptomatic VPB during both pre-EMD and post-EMD periods. This rhythm disorder has been classified as idiopathic ventricular ectopy of the right ventricular outflow tract origin before EMD exposure and was considered benign. Both stress and adrenergic activation with increasing HR are responsible for shift in frequency of VPB. To the best of our knowledge, only 1 case of VPB present before shock that resolved afterward has been previously reported. [bib_ref] Human cardiovascular effects of a new generation conducted electrical weapon, Ho [/bib_ref] It seems that, in individuals without any evidence of organic heart disease, the VPB and SPB dynamics observed are without clinical significance and that their occurrence is not related to increased risk of serious tachyarrhythmias after EMD application.
We speculated that the bradycardia observed in relation to EMD exposure is likely explained as a vagally mediated response induced by the profound muscular contraction (Valsalva maneuver equivalent) and resembles mechanisms of neurally mediated syncope. [bib_ref] Cerebrovascular and cardiovascular responses to graded tilt in patients with autonomic failure, Bondar [/bib_ref] We believe that the direct effect of high-frequency stimulation causing suppression of sinus node, the similar phenomenon known from the assessment of sinus node recovery time, is highly unlikely, but it cannot be definitively ruled out. Because of existing electrical interference with the ECG monitoring, Data are expressed as n (%) or median (range). *P < 0.01comparison with baseline. †P < 0.001comparison with value time point: imminent before EMD discharge. only limited data are known about cardiac capture during EMD application. Under certain experimental conditions, the cardiac capture during EMD application has been detectable in an animal model. [bib_ref] Metabolic acidosis in restraint-associated cardiac arrest: a case series, Hick [/bib_ref] [bib_ref] Acute effects of Taser X26 discharges in a swine model, Dennis [/bib_ref] [bib_ref] Cardiac electrophysiological consequences of neuromuscular incapacitating device discharges, Nanthakumar [/bib_ref] [bib_ref] Taser dart-to heart distance that causes ventricular fibrillation in pigs, Wu [/bib_ref] [bib_ref] Tests on a shocking device-the stun gun, Roy [/bib_ref] [bib_ref] Effects of cocaine intoxication on the threshold for stun gun induction of..., Lakkireddy [/bib_ref] [bib_ref] TASER X26 discharges in swine produce potentially fatal ventricular arrhythmias, Walter [/bib_ref] [bib_ref] Effect of an electronic control device exposure on a methamphetamine intoxicated animal..., Dawes [/bib_ref] Although the chest and heart of humans and pigs differ greatly, 1 case of cardiac capture phenomenon during EMD exposure has been detected by using echocardiography [bib_ref] Human cardiovascular effects of a new generation conducted electrical weapon, Ho [/bib_ref] and another case revealed cardiac capture in a patient with implanted pacemaker when the device was interrogated. [bib_ref] Taser-induced rapid ventricular myocardial capture demonstrated by pacemaker intracardiac electrograms, Cao [/bib_ref] However, the causality of sudden death and EMD use is not exactly established. The presenting rhythm in sudden cardiac death proximate to the use of EMD was asystole or pulseless electrical activity more frequently than ventricular fibrillation. [bib_ref] Presenting rhythm in sudden deaths temporally proximate to discharge of TASER conducted..., Swerdlow [/bib_ref] Pulseless electrical activity was also more likely than ventricular fibrillation in animals that died after repeated long-duration EMD exposures. [bib_ref] Survival of anesthetized Sus scrofa after cycling (7-second on/3-second off) exposures to..., Jauchem [/bib_ref] That is why any information about observed incidental bradycardia during EMD exposure may be of practical value.
In agreement with previous data in human studies, 3-7,15,16 we observed neither clinically significant ECG morphology changes nor any clinically relevant arrhythmias detected after EMD exposure. Despite the observed increase in HR, our data fail to confirm the previously described sympathetically mediated slight shortening of PR and QT intervals after the stun gun discharge. [bib_ref] Twelve-lead electrocardiogram monitoring of subjects before and after voluntary exposure to the..., Vilke [/bib_ref] The observed trends in our study did not reach statistical significance.
The MTWA testing was used in our study as a method for detecting potential subclinical repolarization changes induced by EMD discharge. The value of MTWA in predicting ventricular tachyarrhythmia was shown in populations with structural heart disease. [bib_ref] T-wave alternans negative coronary patients with low ejection and benefit from defibrillator..., Hohnloser [/bib_ref] [bib_ref] Microvolt T-wave alternans distinguishes between patients likely and patients not likely to..., Bloomfield [/bib_ref] [bib_ref] T-wave alternans and the susceptibility to ventricular arrhythmias, Narayan [/bib_ref] [bib_ref] Prognostic utility of microvolt T-wave alternans in risk stratification of patients with..., Chow [/bib_ref] [bib_ref] Does microvolt T-wave alternans testing predict ventricular tachykarrhythmias in patients with ischemic..., Chow [/bib_ref] [bib_ref] The ABCD (Alternans before cardioverter defibrillator) trial, Constantini [/bib_ref] [bib_ref] Ability of microvolt T-wave alternans to modify risk assessment of ventricular tachyarrhythmias..., Gupta [/bib_ref] [bib_ref] Clinical utility of microvolt T-wave alternans testing in identifying patients at hight..., Merchant [/bib_ref] Very few data are known about MTWA prognostic value in apparently healthy populations where positive MTWA has been sporadically detected with a prevalence range between 2% and 5%. [bib_ref] Prevalence of T wave alternans in healthy subjects, Weber [/bib_ref] Such cases resemble one of our study participants who had a positive test result before the EMD discharge already. It is not clear whether the new onset of MTWA positivity detected in our study was given only by repolarization change after EMD exposure. Observation of the new MTWA positivity in 2 of 21 subjects may merely reflect a low reproducibility of MTWA assessment. [bib_ref] The immediate reproducibility of T-wave alternans during bicycle exercise, Bloomfield [/bib_ref] On the other hand, development of repolarization abnormalities after EMD exposure could reflect an objective physiological response. Even if we exclude a potential direct impact of EMD exposure, as discussed previously, both anger and stress followed by induction of a substantial sympathetic response may have direct impact on myocardial electrophysiological properties. It has been speculated that the repolarization heterogeneity is influenced indirectly by changes in the tone of sympathetic nervous system, [bib_ref] Angerlike behavioral state potentiates myocardial ischemia-induced T-wave alternans in canines, Kovach [/bib_ref] [bib_ref] Effects of selective autonomic blockade on T-wave alternans in humans, Rashba [/bib_ref] which can be potentiated by physical resistance. [bib_ref] Acidosis and catecholamine evaluation following simulated law enforcement "use of force" encounters, Ho [/bib_ref] This hypothesis is supported by the data that show that β-blockers are able to reduce MTWA positivity during anger in experimental animal studies. [bib_ref] Effect of metoprolol and sotalol on microvolt T-wave alternans. Results of prospective,..., Klingenheben [/bib_ref] The potential deleterious effects of sympathetic overstimulation are very obvious in stress-induced Tako-Tsubo cardiomyopathy, which is an example of heart functional deterioration with potentially malignant course. [bib_ref] Tako-tsubo-like left ventricular dysfunction with ST-segment elevation: a novel cardiac syndrome mimicking..., Kurisu [/bib_ref] Stress and its consequences can persist longer than any repolarization changes induced directly by cardiac capture; therefore, it would be inappropriate to conclude that EMD exposure cannot lead to delayed adverse cardiac consequences. This may be particularly important in settings where EMD is being used in subjects under drug influence and exposed to subsequent law-enforcement procedures. This may explain the occurrence of fatal outcomes occurring in time delay of several minutes after the EMD exposure. [bib_ref] Electronic gun (Taser) injuries, Ordog [/bib_ref] However, it needs to be mentioned that MTWA testing may be unsuitable for proarrhythmogenic risk detection of EMD because this method predicts the risk of ventricular tachycardia and fibrillation not of asystole and pulseless electrical activity.
# Study limitations
There have been several limitations to our study. The extent of study population and its restriction to healthy and mostly well-trained individuals makes an extrapolation to field use of EMD very problematic. However, it is particularly difficult to get a significant group of volunteers in case of the anticipated inconvenience of the procedure. Any testing on subjects with any structural heart disease is not acceptable. Several technical difficulties should be mentioned. The character of artifacts did not allow continuous detection of ECG signals during the discharge. Our attempt to overcome this limitation by using echocardiography was limited by the restricted acoustic windows in subjects in lying position. Moreover, the muscle contraction during TASER exposure caused a misplacement of the probe in some subjects and it led to the loss of some data. The nature of the MTWA protocol leads inevitably to a delay (up to 10 minutes) between the discharge and acquisition of the first postshock recordings.
# Conclusions
Standard EMD exposure was not associated with any clinically relevant ECG changes except the significant sinus tachycardia in the majority of subjects and the new onset of frequent SPB in 1 case, which was possibly induced by stress reaction due to stun gun shock. The observation of the 2 extraordinary cases in which the EMD discharge induced a brief but profound bradycardia possibly related to vagal stimulation by holding breath and muscular contraction is of particular importance. The new MTWA positivity detected in 2 of the 21 subjects after the EMD exposure may be caused by its direct effect on the myocardium or by sympathetic activation induced by stress, pain, and anger related to the procedure but may be also caused by a potential false positivity of MTWA assessment.
[fig] FIGURE 1: Heart rate profile in an individual. Heart rate was obtained from ECG recordings in predefined time points. The HR increased in majority of cases after EMD discharge (close circles and full line). Two abnormal case profiles are highlighted (open circle and dashed line). [/fig]
[table] TABLE 1: Clinical and Demographic Data (n = 26) [/table]
[table] TABLE 2: ECG Parameters at Baseline, During, and After EMD Discharge [/table]
[table] TABLE 3: Microvolt T-wave Alternans Test ResultsMTWA Test (n = 21)Data are expressed as n (%) or median (range). [/table]
|
Microfluidic Systems for Biosensing
In the past two decades, Micro Fluidic Systems (MFS) have emerged as a powerful tool for biosensing, particularly in enriching and purifying molecules and cells in biological samples. Compared with conventional sensing techniques, distinctive advantages of using MFS for biomedicine include ultra-high sensitivity, higher throughput, in-situ monitoring and lower cost. This review aims to summarize the recent advancements in two major types of micro fluidic systems, continuous and discrete MFS, as well as their biomedical applications. The state-of-the-art of active and passive mechanisms of fluid manipulation for mixing, separation, purification and concentration will also be elaborated.Future trends of using MFS in detection at molecular or cellular level, especially in stem cell therapy, tissue engineering and regenerative medicine, are also prospected.
# Introduction
During the last two decades, the exploration of biological systems from molecules, through cells, to small multicellular organisms has explosively grown based on the advancement in microfluidic system. This enabling technology allows sensing of ever-decreasing sample volumes and target analyte concentrations in ways that are not possible using conventional testing systems. Such technology also has the benefit of scaling the dimensions that enables a range of fundamental features to accompany system miniaturization such as reduced reagent consumption, high temporal resolution due to rapid mixing, high throughput, enhanced analytical performance, less waste, low unit cost, reduced energy consumption, and reduced dimensions when compared to macroscale techniques [bib_ref] Microfluidics: fluid physics at the nanoliter scale, Squires [/bib_ref]. It is a powerful tool holding great promise to facilitate novel experiments with unprecedented performance and has already found unique applications in chemical and system biology [bib_ref] Microfluidic systems for chemical kinetics that rely on chaotic mixing in droplets, Bringer [/bib_ref] [bib_ref] Applications of microfluidics in chemical biology, Weibel [/bib_ref] [bib_ref] Microfluidics-based system biology, Breslauer [/bib_ref] , high-throughput biological screening, cell analysis and clinical diagnostic [bib_ref] Microchip-based cell analysis and clinical diagnosis system, Sato [/bib_ref] , as well as point-of-care (POC) ion analysis for biomedical and environmental monitoring [bib_ref] Lab-on-a-chip systems for biomedical and environmental monitoring, Gardeniers [/bib_ref].
Recently, significant development of bioanalysis and clinical analysis has mainly been driven by the strong demand for fast and reliable results, which are essential for early diagnosis and further medical treatment. Results concerning potential drug targets, vaccine studies and speciation of toxic substances must also be of the highest reliability. These bioanalytical challenges in many cases can be solved using specifically designed and fabricated miniaturized tools called lab-on-a-chip systems or micro total analysis systems (μTAS) [bib_ref] Micromachining a miniaturized capillary electrophoresis-based chemical analysis system on a chip, Harrison [/bib_ref]. Advances in technology have allowed chemical and biological processes to be integrated on a single platform. Adaptation of these approaches to Lab-on-a-Chip (or μTAS) formats is providing a new kind of research tools for the investigation of biochemistry and life processes.
Since this review article is a special issue mainly focused on the state-of-the-art technological development in UK, we highlights some of the most important and interesting recent developments on microfluidics mainly from UK researchers, complementary with some outstanding research findings from international communities.
## Microfluidic systems and components
## Systems
Microfluidic systems for biosensing normally consist of a set of fluidic operation units that allow different biomolecules to be detected and assayed in an easy and flexible manner. Overall, the chip-based platform which has good integration with micro/nano-fluidic components is capable of sampling, filtration, preconcentration, separation, restacking, and detection for biomolecules. [fig_ref] Figure 1: Schematic of one idealized total analysis device showing the various functions on... [/fig_ref] shows a generalized schematic of the types of functional elements used for constructing such a microfluidic chip [bib_ref] Microtechnologies and nanotechnologies for single-cell analysis, Andersson [/bib_ref]. Based on their flow type, microfluidic system can be categorized into two main types, continuous and discrete, and their details are reviewed in the following sections (2.1.1 & 2.1.2).
## Continuous microfluidic system
Continuous-flow microfluidic operation is a promising approach because it is easy to implement and less sensitive to protein fouling problems. Continuous-flow devices are adequate for many well-defined and simple biochemical applications, and for certain tasks such as chemical separation, but they are less suitable for tasks requiring a high degree of flexibility or complicated fluid manipulations. These closed-channel systems are inherently difficult to integrate and scale because the parameters that govern the flow field vary along the flow path making the fluid flow at any one location dependent on the properties of the entire system. Permanently-etched microstructures also lead to limited reconfigurability and poor fault tolerance capability.
For biosensing or diagnostic applications, the microfluids involved are biomolecules or chemicals derived from biological tissues or body fluids. For the purposes of simple and point-of-care diagnostics, that sample is most likely blood, saliva, or nasal fluid. The preparative steps, including sample collection, metering and filtration, analyte enrichment, labeling and detection, are generally required prior to the diagnostic measurements. Recently advanced microfluidic systems not only provide elegant solutions to relieve the complexity of a biosensing or diagnostic test, but also improve responsive speed and miniaturize the size of analysis equipment [bib_ref] Parallel picoliter RT-PCR assays using microfluidics, Marcus [/bib_ref].
## Discrete microfluidic system
Droplet-based microfluidic systems are currently an emerging area of microfluidic research. One of the most popular means is to inject multiple laminar streams of aqueous reagents into an immiscible carrier fluid and therefore to induce flow instability instantly for forming the droplets [bib_ref] Formation of droplets and mixing in multiphase microfluidics at low values of..., Tice [/bib_ref]. There are several distinctive advantages based on droplet-based microfluidic systems. First, the systems promise a new high-throughput technology that enables the generation of microdroplets in excess of several thousand per seconds [bib_ref] Quantitative detection of protein expression in single cells using droplet microfluidics, Huebner [/bib_ref]. In addition, parallel and serial in-vitro compartmentalization is possible with this technology. The reagents are confined inside the droplets in water-in-oil (w/o) emulsions and reagent transport occurs with no dispersion [bib_ref] Miniaturizing chemistry and biology in microdroplets, Kelly [/bib_ref]. This unique feature enables chemical reaction indexing, thereby facilitates many chemical reactions in a highly organized manner. Furthermore, fast mixing can occur within minute volumes of microdroplets (nanoliter to femtoliter range) due to the short diffusion distance and chaotic mixing within droplets with the use of twisting channel geometries by stretching, folding, and reorienting fluid [bib_ref] Microfluidic systems for chemical kinetics that rely on chaotic mixing in droplets, Bringer [/bib_ref]. Another feature is that the variation of the channel dimensions can regulate the droplet volumes and decrease volumes anything up to 10 9 times compared to the smallest assays in conventional microtiter plates [bib_ref] Miniaturizing chemistry and biology in microdroplets, Kelly [/bib_ref]. With a control of flow rate, the reagent concentrations can be modified accordingly [bib_ref] Microfluidic systems for chemical kinetics that rely on chaotic mixing in droplets, Bringer [/bib_ref]. It is the confluence of the aforementioned unique features and the ability to regulate and manipulate the droplet motions to split, merge, and sort that has revolutionized our ability to control fluid/fluid interfaces for use in fields ranging from material processing and biomaterials to chemical biology and nanomedicine.
## Components
Microfluidic system normally consists of a micropump, micromixer, valve, separator and concentrator. Among these components, micropumps and micromixers are the key components for microfluidic applications due to their actively functioning capability. The types of micropumps vary widely in terms of design and application but can be generally categorized into two main groups: mechanical and non-mechanical pumps. Conventional mechanical micropumps represent smaller versions of macrosized pumps that typically consist of a microchamber, check valves, microchannels and an active diaphragm to induce displacement for liquid transportation. Thermal bimorph, piezoelectric, electrostatic and magnetic forces, as well as shape memory mechanisms, have been utilized to actuate the diaphragm [bib_ref] A polymeric piezoelectric micropumps based on lamination technology, Truong [/bib_ref] [bib_ref] Creating, transporting, cutting, and merging liquid droplets by electrowetting-based actuation for digital..., Cho [/bib_ref] [bib_ref] Pressure injection in continuous sample flow electrophoresis microchips, Lacharme [/bib_ref] [bib_ref] Design, fabrication, and testing of an electrohydrodynamic ion-drag micropump, Darabi [/bib_ref]. These micropumps are relatively complicated, expensive, typically made by multi-wafer processes and difficult to be integrated with other systems such as integrated circuits (IC) for control and signal processing due to incompatible processes and structures [bib_ref] Liquid dielectrophoresis on the microscale, Jones [/bib_ref] [bib_ref] Electrowetting-based actuation of liquid droplets for microfluidic applications, Pollack [/bib_ref] [bib_ref] Electrowetting-based actuation of droplets for integrated microfluidics, Pollack [/bib_ref] [bib_ref] Electrostatic transportation of water droplets on superhydrophobic surfaces, Torkkeli [/bib_ref]. They generally have a large dead volume, leading to excessive waste of biosamples and reagents which are very expensive and precious in biological analysis, especially for forensic investigations. These micropumps typically have moving parts which lead to a high failure rate, low production yield in fabrication and poor reliability in operation. These technologies are based on the manipulation of continuous liquid flow through microfabricated channels. Actuation of liquid flow is implemented either by external pressure sources, external mechanical pumps, integrated mechanical micropumps, or by combinations of capillary forces and electrokinetic mechanisms [bib_ref] Electrostatic stabilization of fluid microstructures, Mugele [/bib_ref].
## Pump
Controlling fluid flow is crucial in microfluidic devices, especially for processing biochemical reactions. Such a process generally relies on active control by mechanical pressure [bib_ref] Reagent-loaded cartridges for valveless and automated fluid delivery in microfluidic devices, Linder [/bib_ref] [bib_ref] A bubble-driven microfluidic transport element for bioengineering, Marmottant [/bib_ref] , electroosmotic force [bib_ref] Micromachining a miniaturized capillary electrophoresis-based chemical analysis system on a chip, Harrison [/bib_ref] , electrowetting [bib_ref] Valveless pumping using traversing vapor bubbles in microchannels, Jun [/bib_ref] [bib_ref] Reversible switching of high-speed air-liquid two-phase flows using electrowetting-assisted flow-pattern change, Huh [/bib_ref] , and electrochemical reaction [bib_ref] Electrochemical principles for active control of liquids on submillimeter scales, Gallardo [/bib_ref]. These active manipulations enable close control in a rapid and precise manner. Electrokinetic sampling has been widely used for microfluidic chip, especially for microfluidic chip electrophoresis, because the electric field can be easily and precisely applied to the reservoirs on the chip. The popular mechanism used for these active micropumps is electrokinetic force. Based on the mechanism, various micropumps such as dielectrophoresis, asymmetric electric field, electroosmosis and electrophoresis (the latter two are considered as part of the electrohydrodynamic (EHD) phenomena) [bib_ref] Fluid flow induced by nonuniform ac electric field in electrolytes on microelectrodes:..., Green [/bib_ref] [bib_ref] Pumping of water with ac electric fields applied to asymmetric pairs of..., Brown [/bib_ref] have been developed.
Moving sample fluids and reagents on a biosensing microfluidic device requires developing a pressure difference in the flow path to direct fluid in one direction or another. Miniaturized versions of positive-displacement pump designs such as gear or peristaltic pumps have been proposed for microfluidic applications, but these all require some external power source or repetitive motion to control. It is desirable for fluidic motion in a passive microfluidic system design to be driven by a readily available force such as gravity, capillary action, absorption in porous materials, chemically induced pressures or vacuums (e.g., by a reaction of water with a drying agent), or by vacuum and pressure generated by simple manual action. Wicking and capillary action have been widely used to motivate fluids for POC diagnostics. For example, low cost lateral flow tests demonstrate the elegant and inexpensive use of wicking to drive multiple sample types through all steps of an assay.
One of the simple methods for transporting fluids on microfluidic devices is to apply pressure manually to deflect a diaphragm [bib_ref] Microfluidic tectonics platform: A colorimetric, disposable botulinum toxin enzyme-linked immunosorbent assay system, Moorthy [/bib_ref]. Diaphragm membrane pumps have been demonstrated successfully in moving fluid on a microfluidic device. However, it is not easy to control the flow rate in a reproducible way. Zhu et al. [bib_ref] Gravity-Driven Microfluidic Pump with a Steady Flow Rate, Zhu [/bib_ref] reported a gravity microfluidic pump for producing constant flow rate. This passive system employs a microchannel and a gravity-driven pump consisting of horizontally oriented reservoirs that supply fluid to the microchannel at a substantially constant rate.
The passive device may be useful for numerous microfluidic applications such as cell-size sorting [bib_ref] Gravity-Driven Microhydrodynamics-Based Cell Sorter (microHYCS) for Rapid, Inexpensive, and Efficient Cell Separation..., Huh [/bib_ref].
The pumps have been developed based on osmotic pressure as the actuation mechanism have been used in many drug-delivery applications to deliver medication over a prolonged period of time [bib_ref] A water-powered micro drug delivery system, Su [/bib_ref] [bib_ref] A water-powered osmotic microactuator, Su [/bib_ref].
The advantages of these pumps include simple construction and the absence of moving parts. Another passive system involves controlled evaporation of a liquid into a chamber with an absorption agent flow [bib_ref] An evaporating-based disposable micropump concept for continuous monitoring applications, Effenhauser [/bib_ref]. As fluid evaporates from the channel, capillary forces induce fluid flowing from reservoir to replace the evaporating fluid. This micropump has advantages of low cost, high reliability and constant flow rate over a long period of time. The major disadvantage of the evaporation micropump is the need to control environmental conditions for constant flow rates and lower flow rates.
The micropumps have also been developed by employing fluid-responsive polymers to deliver fluids [bib_ref] A valved responsive hydrogel microdispensing device with integrated pressure source, Eddington [/bib_ref]. Fluid-responsive polymers swell when exposed to certain environmental conditions, such as changes in moisture, pH, or temperature. One recent fluid-responsive pump consists of an array of responsive polymers that deforms a flexible membrane made from PDMS and produces flow rates [bib_ref] A valved responsive hydrogel microdispensing device with integrated pressure source, Eddington [/bib_ref]. The disadvantage of the pump is the requirement of pressure to inject the buffer solution in order to active the pump.
## Valve
The development of valves for microfluidic systems has been progressing rapidly in recent years. The applications of the microvalves include flow regulation, on/off switching, or sealing of biomolecules, micro or nano particles, chemical reagents, oils, water, bubbles, gases, vacuum and many others. Most of them generally can be divided into two major categories: active microvalves and passive microvalves. Most active microvalves mechanically actuate moving parts using magnetic, electric, piezoelectric, thermal or other actuation methods. However, the complexity, high cost and external supporting equipment has largely limited the application of active microvalves. Alternatively, passive microvalves are desirable due to their structure simplicity, easy integration and miniaturization of a system. Passive microvalves can normally be categorized as two categories as with and without moving parts. The passive microvalves with moving parts, also called check valves, are incorporated in inlets and outlets of micropumps as mechanical moving parts, such as flaps [bib_ref] An effervescent reaction micropump for portable microfluidic systems, Brian [/bib_ref] , membranes [bib_ref] Fabrication and characterization of a micromachined passive valve, Bien [/bib_ref] [bib_ref] Micro check valves for integration into polymeric microfluidic devices, Nguyen [/bib_ref] [bib_ref] Fabrication and characterization of a micromachined passive valve, Bien [/bib_ref] , spherical balls [bib_ref] A ball valve micropump in glass fabricated by powder blasting, Yamahata [/bib_ref] or mobile structures [bib_ref] A magnetically driven PDMS micropump with ball check-valves, Pan [/bib_ref]. The passive valves with moving parts only open to forward pressure, excising diode-like characteristics. The one-way behavior of check valves significantly affects the pumping performance of a reciprocal displacement micropump. Leakage in the check valves reduces backpressure and pumping rate in the micropump. The passive microvalves without moving parts; e.g., using nozzle [bib_ref] A valveless diffuser/nozzle-based fluid pump, Stemme [/bib_ref] [bib_ref] Microdiffusers as dynamic passive valves for micropump applications, Gerlach [/bib_ref] , diffuser [bib_ref] A valve-less diffuser micropump for microfluidic analytical systems, Andersson [/bib_ref] [bib_ref] Surface micromachined thermally driven micropump, Jang [/bib_ref] [bib_ref] The dynamic micropump driven with a screen printed PZT actuator, Koch [/bib_ref] or Tesla [bib_ref] Low-order modeling of resonance for fixed-valve micropumps based on first principles pump, Morris [/bib_ref] [bib_ref] Geometry-based macro-tool evaluation of non-moving-part valvular microchannels, Feldt [/bib_ref] elements, have been widely used in inlets and outlets of micropumps.
Another method to control fluid flow, taking advantage of the large surface-to-volume ratio in microfluidic systems, is the passive capillary microvalve which utilizes the geometries or the surface wetting properties in the microchannels [bib_ref] Nanoliter-sized liquid dispenser array for multiple biochemical analysis in microfluidic devices, Yamada [/bib_ref] [bib_ref] A liquid-triggered liquid microvalve for on-chip flow control, Melin [/bib_ref] [bib_ref] Pressure barrier of capillary stop valves in micro sample separators, Leu [/bib_ref] [bib_ref] Hydrophobic valves of plasma deposited octafluorocyclobutane in DRIE channels, Andersson [/bib_ref] [bib_ref] Micromachined filter-chamber array with passive valves for biochemical assays on beads, Andersson [/bib_ref]. The passive microvalves using capillary effects are useful for passive biosensing microfluidics since autonomous and spontaneous valving can be realized due to the geometry [bib_ref] Nanoliter-sized liquid dispenser array for multiple biochemical analysis in microfluidic devices, Yamada [/bib_ref] [bib_ref] A liquid-triggered liquid microvalve for on-chip flow control, Melin [/bib_ref] [bib_ref] Pressure barrier of capillary stop valves in micro sample separators, Leu [/bib_ref] and surface wettability properties [bib_ref] Hydrophobic valves of plasma deposited octafluorocyclobutane in DRIE channels, Andersson [/bib_ref] [bib_ref] Micromachined filter-chamber array with passive valves for biochemical assays on beads, Andersson [/bib_ref] of the microchannels. These passive capillary valves are used preferably to block and pass fluidic flows for avoiding the valve-actuation-induced interference with biofluids due to the actuation energy of microvalves.
## Mixer
Mixing is a physical process to achieve homogeneity of the different components involved in the certain process. In some cases, the mixing will be the rate determining step when the mixing time is in the same order or longer than the molecular reaction time. Because the fluid streams mainly appear naturally as laminar flow on a chip, the mixing will mainly depend on molecule diffusion. Mixing small amounts of reagents and samples in microfluidic channels or structures is a challenging task.
Likewise, mixing in passive micromixers relies mainly on molecular diffusion and chaotic advection.
To speed mixing process, the T-mixer or Y-mixer which consists of the inlets converging into a long microchannel has been developed as a simple and effective solution [bib_ref] Quantitative analysis of molecular interactive in microfluidic channel: the T-sensor, Kamholz [/bib_ref] [bib_ref] Molecular diffusive scaling laws in pressure-driven microfluidic channels: deviation from one-dimensional Einstein..., Kamholz [/bib_ref] [bib_ref] Experimental and theoretical scaling laws for transverse diffusive broadening in two-phase laminar..., Ismagilov [/bib_ref]. Other methods for fast mixing have been implemented through reducing the mixing path in a narrow mixing channel [bib_ref] Characterization method for a new diffusion mixer applicable in micro flow injection..., Veenstra [/bib_ref] and realizing parallel lamination with multiple streams [bib_ref] Microfluidic systems with on-line UV detection fabricated in photodefineable epoxy, Jackman [/bib_ref] [bib_ref] Improved characterization technique for micromixers, Koch [/bib_ref].
Besides diffusion, advection is another important form of mass transfer in flows with a low Reynolds number. However, advection is often parallel to the main flow direction, and is not useful for the transversal mixing process. The chaotic advection generated by special geometries in the mixing channel can improve mixing significantly. The basic idea is the modification of the channel shape for splitting, stretching, folding and breaking of the flow. The simplest method to get chaotic advection is to insert obstacles or structures in the mixing channel. However, it has been shown that eddies or recirculation cannot be generated in a microchannel, because of its low Reynolds number [bib_ref] Optimizing layout of obstacles for enhanced mixing in microchannels, Wang [/bib_ref]. The effective method to produce chaotic advection is to modify the wall of mixing channel with ribs, grooves and staggered-herringbone grooves. Johnson et al. [bib_ref] Rapid microfluidic mixing, Johnson [/bib_ref] were the first to investigate this phenomenon. They ablated the grooves on the bottom wall of the channel by laser. This structure allows mixing at a relatively slow velocity of 300 µm/sec. Stroock et al.investigate two different groove patterns, slanted groove and staggered. The so-called staggered herringbone mixer can work well at low Reynolds number.
## Separator
It is challenging to process complex biological samples without the sophisticated sample preconditioning capabilities for a passive microfluidic system. To answer the challenge, the H-filter system developed by Yager's group to provide an alternative solution to a conventional porous barrier filter [bib_ref] Diffusion-based extraction in a microfabricated device, Yager [/bib_ref]. The H-filter is based upon the parallel laminar flow of two or more miscible streams in contact with each other. The streams do not mix, but species can diffuse from one stream to the other with smaller species diffusing faster than larger ones. The H-filter allows continuous filtration of unwanted components or extraction of desired analytes from fluids without the need for a membrane filter or similar component that requires cleaning or replacement. Wu et al. have developed a microcapillary electrochromatography (μCEC) chip for performing a highly efficient separation of double-stranded DNA (dsDNA) fragments through vertically aligned multi-wall carbon nanotubes (MWCNTs) in a microchannel as shown in [bib_ref] Nanostructured pillars based on vertically aligned carbon nanotubes as the stationary phase..., Wu [/bib_ref]. This is the first report on the development of a novel stationary nanocolumn by directly growing homogeneous and vertically aligned MWCNTs in microchannels to enable improved analytical performance in capillary electrochromatography. This device incorporated well-arranged nanostructures to replace the non-homogenous traditional packing medium to enhance the phase ratio, reduce the flow resistance, and maintain the laminar flow pattern in a higher efficiency.
## Concentrator
The concentrator increases concentration of dissolved or dispersed substances under mild conditions to keep activity and viability. They can increase the signal strength of any interesting substance. The most general method is centrifugation. However, this method has not been applied in microfluidic system. Other methods such as chromatography, dielectrophoresis, transverse isoelectric focusing, and ultrasonic trapping have been developed. A novel concentration method employing the dual-asymmetry electrokinetic flow (DAEKF) has been developed by Wu et al. [bib_ref] Dual-Asymmetry Electrokinetic Flow (DAEKF) Focusing for Pre-concentration and Analysis of Catecholamines in..., Wu [/bib_ref] for catecholamines concentration and detection in a CEEC nanochannel. The combination of asymmetry EOF and field-effect control was introduced to carry out a two-dimensional gradient shear flow exposed to a downward trailing velocity vector for the generation of a strong downward rotational flow for sample concentration (see [fig_ref] Figure 3: The schematics of [/fig_ref]. However, the method is not suitable for a passive microfluidic system. Sharma et al. [bib_ref] Development of an evaporation-based microfluidic sample concentrator, Sharma [/bib_ref] developed a relatively simple concentrator device based on isothermal evaporation. It is capable of removing 0.8 ml of water per minute at 37 °C, and is also able to concentrate liquids at room temperature at lower evaporation rates. The evaporative concentrator can be used as a stand-alone device or integrated into various processes and analytical instruments, substantially increasing their sensitivity while decreasing processing time.
## Droplet-based microfluidics
Due to recent advances in droplet microfluidics that have provided the promise of unique and optimal solutions for the study of genomics, proteomics, cellomics, and metabolomics, many novel device functions and ingenious applications based on this technology have been demonstrated and brought to commercialization. The physics of droplet microfluidics [bib_ref] Controlled microfluidic interfaces, Atencia [/bib_ref] , their ways of generating, controlling, and manipulating, and their applications on chemical kinetics [bib_ref] Microfluidic systems for chemical kinetics that rely on chaotic mixing in droplets, Bringer [/bib_ref] , system biology [bib_ref] Microfluidics-based system biology, Breslauer [/bib_ref] , high-throughput biological screening, directed evolution of proteins and RNAs and polymerase chain reactions [bib_ref] Quantitative detection of protein expression in single cells using droplet microfluidics, Huebner [/bib_ref] [bib_ref] Microfluidic DNA amplification-a review, Zhang [/bib_ref] , and as chemical reactors in microfluidic channels [bib_ref] Reactions in droplets in microfluidic channels, Song [/bib_ref] have been extensively reviewed elsewhere in detail. This section highlights the contributions mainly from UK researchers on the physics of droplet microfluidics, the operation principles of droplet manipulations in terms of droplet generations, droplet manipulation methods, droplet fission and merging, mixing in droplets, and droplet sorting, and how they are applied to biomedical applications.
## Droplet generations
Vast amount of femtoliter droplets of typically 1 µm to 100 µm in diameter (about 0.5 fl to 0.5 nl volume) in bulk oil solutions can be easily formed by simply adding the reaction mixture to stirred mineral oil containing surfactants without the need for automation. A volume reduction up to 10 9 is possible compared to conventional microtiter-plate method and excitingly, this allows a one-step, high throughput production of as many as 10 10 reactions in a total volume of only 1 ml of emulsion [bib_ref] Quantitative detection of protein expression in single cells using droplet microfluidics, Huebner [/bib_ref]. Council, Cambridge to develop the method of linking genotype and phenotype at the molecular level by designing a system for in vitro evolution that uses man-made compartments, a technique they termed in-vitro compartmentalization (IVC) [bib_ref] Man-made cell-like compartments for molecular evolution, Tawfik [/bib_ref]. The large excess of droplets, each statistically contains a single gene on average, serve as artificial cells that contain all the ingredients for transcription and translation, and the activity of the resulting RNA or proteins. Each droplet can be used as an independent microreactor. They also applied this IVC method to select DNA catalystsand directed evolution of novel Diels-Alderase ribozymes [bib_ref] Selection of ribozymes that catalyse multiple-turnover Diels-Alder cycloadditions by using in vitro..., Agresti [/bib_ref]. The main problem that bulk microdroplet generation encounters is wide distribution of microdroplet size.
Micro and nanopipette delivery is popular among biologist and life science researchers. In addition to the commonly used micropipette in molecular biology as well as medical tests which can dispense between 1 and 1,000 µl of liquid droplet, Ying et al. from Cambridge University and Imperial College London developed nanopipettes to employ controlled deposition and local delivery of reagents and biomolecules below nanoliter scale regime based on controlled voltage-driven method for mapping of specific species [bib_ref] The scanned nanopipette: a new tool for high resolution bioimaging and controlled..., Ying [/bib_ref]. The nanopipettes can be easily fabricated from borosilicate capillary glass.
Operating in conductance solution (usually physiological buffer), one electrode is placed both in the pipette and the bath to maintain a constant ion current by adjusting the pipette-sample distance under a constant voltage between both electrodes to act as a scanning ion conductance microscopy. By applying an electrical potential to the bath electrode opposite to the charge of the sample in the pipette reservoir, the sample can be driven out from the pipette and absorbed and bound to the nearest surface.
Based on this method, the delivery of fluorophore labeled DNA from the pipette was demonstrated and it was found that on application of a voltage pulse, they can get controlled pulses of DNA from the pipette down to the level of just 20 molecules per pulse [bib_ref] Programmable delivery of DNA through a nanopipet, Ying [/bib_ref]. This proven method facilitated the controlled insertion of few alpha toxin channels locally into the cell membrane of a cardiac myocyte in a small cluster of cells to study cell-cell communication [bib_ref] The scanned nanopipette: a new tool for high resolution bioimaging and controlled..., Ying [/bib_ref]. The T-junction channels can also be combined with other operating microfluidic devices for multi-purpose applications. Niu et al. from Imperial College London integrated a T-junction channel (with an oil inlet, a sample inlet for effluent from the liquid chromatography capillary, and a droplet outlet channel) and the generated droplets were transferred in sequence to a second device which was composed of channels with pillars for evacuating oil and loading analyte droplets into the second capillary electrophoresis channel to form a fully functional droplet connector for two-dimensional separations in both time and space [bib_ref] Droplet-based compartmentalization of chemically separated components in two dimensional separations, Niu [/bib_ref]. The droplet-mediated method could become key components in 2D or multidimensional separations. The stability and monodispersity of the droplet formation based on T-junction can also be used in micro-optical fluidic system in sensing applications due to its wide tenability in the optical characteristics [bib_ref] Tunable microfluidic optical fiber, Mach [/bib_ref]. Chin et al. employed a stream of plugs formed by two immiscible liquids, i.e., immersion oil as the carrier liquid and calcium chloride solution as the dispersed liquid at a T-junction, to act as a long-period grating [bib_ref] An on-chip liquid tunable grating using multiphase droplet microfluidics, Chin [/bib_ref]. The liquid grating can be tuned easily based on the flow rates of the liquids, the refractive index, and the index variation of the core layer by using combinations of different liquids to act as a biochemical sensor and an optical tunable filter. [bib_ref] An integrated device for monitoring time-dependent in vitro expression from single gene..., Courtois [/bib_ref]. High yields of GFP were obtained and it is possible to perform protein expression from single copies of the DNA template, thereby generating monoclonal droplets.
Dielectrophoretic (DEP) force is used to pull out multiple sessile droplets from a large liquid reservoir [bib_ref] Dynamic control of DEP actuation and droplet dispensing, Wang [/bib_ref]. When a voltage is applied to the electrodes, a finger-shaped liquid rivulet with semicircular cross-section is formed which protrudes from the parent droplet reservoir and moved rapidly along the electrodes due to attraction of polarizable fluid to areas of higher electric field intensity. The rivulet stops when it gets to the end of the electrodes and remains in a stable electrohydrostatic equilibrium. When the applied voltage is removed, capillary instability takes over to form sessile droplets within 2 ms. The size and uniformity of the droplets are determined by the magnitude and frequency of the applied voltage. Another important criterion to ensure droplet uniformity is the bumps should be arranged close to the most unstable wavelength predicted by Rayleigh's cylindrical jet theory. Electrowetting-on-dielectric (EWOD) can also be used to generate droplets. The principle of EWOD is based on the change of free energy on the dielectric surface due to the electric charge accumulation when a voltage is applied, thereby creates a change in wettability on the surface and contact angle of the droplet [bib_ref] Electrowetting-based actuation of liquid droplets for microfluidic applications, Pollack [/bib_ref]. All liquid droplet movements are confined between two plates. The device not only can transfer and merge droplets but also generate droplets by cutting an elongated liquid droplet through necking. Through detailed theoretical study, it is found that smaller channel gap, a larger droplet and a larger contact angle change make droplet cutting easier to perform.
An excellent report on the comparison between DEP and EWOD can be referred elsewhere [bib_ref] Principles of droplet electrohydrodynamics for lab-on-a-chip, Zeng [/bib_ref]. It has been understood that capillary force is the prominent force in micro scale regimes. Recently, a novel microcontact printing system capable of printing tens to thousands of biological droplet reagents into an array with batch filling and parallel printing based on the capillary force has been developed as shown in [fig_ref] Figure 5: Schematic of microarray system for batch-filling and in parallel printing of multiple... [/fig_ref] [bib_ref] Rapid microarray system for passive batch-filling and in-parallel-printing protein solution, Ho [/bib_ref]. The system consists of microfilling chip that facilitate the transfer of numerous protein reagents into the microstamp chip by capillary force in seconds. The microstamp chip can then employ the capillary force again to transfer the protein solutions into the corresponding tips and finally come into contact with the substrate for biofluid array printing. More microarray printing devices that have the potential to be expanded to a high throughput system for simultaneously printing large array of biofluid spots for hundred times in minutes, a typical requirement for high throughput disease diagnosis and drug screening, has been demonstrated [bib_ref] Micro-stamp systems for batch-filling, parallel-spotting, and continuously printing of multiple biosample fluids, Ho [/bib_ref].
## Droplet manipulations
It is generally agreed that moving biological and chemical samples in isolated droplets could provide major advantages over single-phase continuous-flow microfluidic devices as mentioned above.
Several methods are currently available to manipulate droplets such as thermocapillary, surface acoustic wave/vibration, surface chemical or morphological gradient, electric field, optoelectrowetting, droplets on a surface subjected to temperature gradients can be referred elsewhere [bib_ref] Fundamental studies on micro-droplet movement by Marangoni and capillary effects, Tseng [/bib_ref]. The results
show that temperature gradients, the change of dynamic receding/advancing contact angles across the droplets, and the flow fields inside the droplet are the key parameters.
Vibration-based actuation method is also a versatile method to manipulate droplets. Brunet et al.
from University of Bristol reported an exciting vibration-induced droplet climbing phenomenon against the gravity when the droplet was placed on a vertically vibrating inclined plate [bib_ref] Vibration-induced climbing of drops, Brunet [/bib_ref]. The droplet excised an upward motion when above the threshold in vibration acceleration. The droplet motion is caused by the deformation of the drop as a result of an up or down symmetry breaking induced by the presence of the substrate. This finding allows a droplet to move along an arbitrary path in a plane without special surface treatments or localized forcing. Surface acoustic waves (SAW) were also used to actuate and process very small volumes of fluids (from 50 nL to 100 nL) on the planar surface of a piezoelectric chip [bib_ref] Acoustic manipulation of small droplets, Wixforth [/bib_ref]. The actuation force is originated from the SAW-mediated internal streaming in the fluid. Recently, a distributed pressure-control scheme has been devised that employs acoustic resonance cavities and rectification structures to translate the frequencies contained in an acoustic signal into separately addressable output pressures that can be used to control liquids in a microfluidic device [bib_ref] Acoustically driven programmable liquid motion using resonance cavities, Langelier [/bib_ref]. This method can be used to perform precise droplet positioning, merging, splitting, and sorting within open microfluidic networks and to generate acoustically tunable liquid gradients.
Surface gradients are passive methods to move droplets where the surface chemical or physical properties gradually change over a given distance. The ways for creating the surface gradients have been extensively reviewed elsewhere [bib_ref] Surface-chemical and -morphological gradients, Morgenthaler [/bib_ref] [bib_ref] Surface-bound soft matter gradients, Genzer [/bib_ref]. An example of a surface gradient based on both the chemical and morphology properties is the wedge-shaped gradients based on low hysteresis nanotextured surfaces recently demonstrated by Khoo and Tseng to facilitate spontaneous and fast motions for a wide range of water droplet volume [bib_ref] Spontaneous high-speed transport of subnanoliter water droplet on gradient nanotextured surfaces, Khoo [/bib_ref]. A 2 µL droplet velocity as high as 0.5 m/s was successfully achieved. Ascension of water droplets with all-round acclivity and a subnanoliter droplet movement were also successfully demonstrated as shown in. It is concluded that the actuation mechanism is based on the combination of surface tension gradient and nanowetting. Dielectrophoretic (DEP) droplet manipulation can be used to dispense and locate arrays of nanoliter-sized droplets loaded with DNA or protein molecules while separating these particles based on their size. The separation mechanism is based on size-dependent downward directed DEP force exerted on the moving bioparticles resulting from the non-uniform electric field. The separation process is quick, requiring only ~1 s. The method has also been used to move aqueous droplets for the formation of artificial bilayer lipid membranes (BLMs) Edinburgh [bib_ref] Demonstration of a wireless driven MEMS pond skater that uses EWOD technology, Mita [/bib_ref]. The trapped air bubbles attached to the hydrophobic surface of the device were moved by EWOD. This low voltage EWOD-driven mechanism is potentially useful for FR power transmission.
## Droplet fusion, mixing, sorting, trapping and releasing
Many droplet reactions and assays require multiple steps where new reagents are added at defined times, to scale up, start, modify or terminate a reaction. Therefore, device operations such as fusion, mixing, sorting, trapping, and releasing are essential for a fully functional high throughput biochemical device.
## For fusion method, fidalgo et al. from university of cambridge demonstrated a method for droplet
fusion based on a surface energy patterning on the walls of a microfluidic device without any active mechanical or electrical parts that allows the fusion of more than two droplets at a single point [bib_ref] Surface-induced droplet fusion in microfluidic devices, Fidalgo [/bib_ref].
Niu et al. proposed a novel method to passively merging of aqueous microdroplets within segmented flow microfluidic devices in a controlled manner [bib_ref] Pillar-induced droplet merging in microfluidic circuits, Niu [/bib_ref]. Major advantages of this design include the ability to adjust the inter-droplet distance in a facile manner and the ability to selectively merge droplets according to their size or number. The merging mechanism depends on the difference in hydrodynamic resistance of the continuous phase and the surface tension of the discrete phase through the use of pillar structures. The merging process depends on the droplet size, mass flow rate and volume ratio between the droplets and the merging chamber.
Chaotic advection can be employed to rapidly mix multiple reagents isolated in droplets on a microfluidic device by using a combination of turns and straight sections or so-called winding microfluidic channels as long as time-periodic flows are induced [bib_ref] Formation of droplets and mixing in multiphase microfluidics at low values of..., Tice [/bib_ref]. These winding channels create chaotic mixing by folding, stretching and reorienting the fluid volume. The mixing is rapid (sub-millisecond) which allows for an accurate description of fast reaction kinetics. Srisa-Art et al.
from Imperial College London has applied this method and integrated it with a confocal fluorescence detection system to monitor the real-time streptavidin-biotin binding kinetics [bib_ref] Monitoring of real-time streptavidin-biotin binding kinetics using droplet microfluidics, Srisa-Art [/bib_ref]. a design strategy to scale up microfluidics for producing monodispersed emulsions [bib_ref] Novel parallel integration of microfluidic device network for emulsion formation, Tetradis-Meris [/bib_ref]. Based on this parallelization methodology, a hydrophobic platform containing 180 devices has been successfully demonstrated to produce highly monodispersed W/O emulsions with coefficient of variation ~5%. The success of this approach can be attributed to correct assessment on the geometric array to be scaled up (ladder-type versus tree-type) and the introduction of a drainage manifold architecture which improves the operational conditions. It is anticipated that the number of operating devices can be further increased.
## Microfluidics devices for biosensing
## Genomic applications
## Microfluidic pcr chip
So far, the most successful commercialization of compartmentalization in Bio-MEMS, i.e. water-inoil droplet, is the emulsion polymerase chain reaction (ePCR). ePCR enables clonal amplification of templates from complex mixtures in a bias-free manner, thus enabling a number of emerging applications such as high-throughput sequencing. However, droplets produced from conventional bulk emulsion techniques are not uniform in size [bib_ref] Directed evolution by in vitro compartmentalization, Miller [/bib_ref]. The non-uniformity of the droplet results in the uncertainty in quantitative readout of experimental data. Another limitation of bulk emulsion droplets is that multistep processing of droplets is difficult. Therefore the application of microfluidic device for the generation of monodispersed droplets will potentially elevate the performance of the existing emulsion PCR technique, through the integration of PCR with other microfluidic operations such as fusion and real-time-analysis. Moreover, the generation of monodispersed droplets facilitates a quantitative analysis, as required for quantitative real-time PCR. The first commercially available platform using microfluidic droplets for various biosensing such as gene expression analysis, drug development, and disease marker detection has been developed by RainDance Technologies [bib_ref] The potential of microfluidic water-in-oil droplets in experimental biology, Schaerliab [/bib_ref]. A collection of primer pairs corresponding to selected genomic regions are encapsulated in microfluidic droplets and then merged with droplets containing the genomic DNA and the PCR reaction mixture, followed by off-chip thermal cycling [bib_ref] High-throughput single copy DNA amplification and cell analysis in engineered nanoliter droplets, Kumaresan [/bib_ref].
To enhance the throughput of PCR, a new design device (as shown in [fig_ref] Figure 7: A continuous-flow PCR device where microfluidic droplets move through a temperature gradient... [/fig_ref] for carrying continuous-flow microfluidic droplets in which the reaction mixture passes through zones of alternating temperature corresponding to denaturation, annealing and extension temperatures. For example, microfluidic droplets in a continuous-flow PCR device move through a temperature gradient across the radial direction. The droplets pass through the inner zone with a high temperature to ensure initial denaturation of the template and travel to the outer zone where primer annealing and template extension occur. The droplets then flow back to the centre, where the DNA is denatured and a new thermal cycle begins. Finally, the droplets exit the device after 34 cycles [bib_ref] Continuous-flow polymerase chain reaction of singlecopy DNA in microfluidic microdroplets, Schaerli [/bib_ref]. This new design avoids temperature cycling of the entire device and leads to more rapid heat transfer, and smaller droplets allow higher throughput. The scale-down from milliliter to picolitre droplets in continuous-flow microfluidic PCR may lead to higher sensitivity [bib_ref] Automated microdroplet platform for sample manipulation and polymerase chain reaction, Chabert [/bib_ref] [bib_ref] High-throughput quantitative polymerase chain reaction in picoliter droplets, Kiss [/bib_ref] [bib_ref] On-chip single-copy real-time reverse-transcription PCR in isolated picoliter droplets, Beer [/bib_ref]. More recently, microfluidic droplet-based PCR has become a versatile module that can be readily integrated with other unit operations and detection technologies of MEMS, such as fluorescence-based monitoring of PCR products.
Amplification in a microfluidic device was monitored online by recording the intensity of a fluorescence resonance energy transfer (FRET) probe for the amplicon [bib_ref] On-chip single-copy real-time reverse-transcription PCR in isolated picoliter droplets, Beer [/bib_ref]. [bib_ref] Continuous-flow polymerase chain reaction of singlecopy DNA in microfluidic microdroplets, Schaerli [/bib_ref].
## Sequencing and other applications
To achieve in vitro directed evolution, the microdroplet boundary acts like the cell wall where the microdroplet compartmentalizes genes and proteins to link the genotype (DNA or RNA) to the phenotype (binding or catalytic activity) [bib_ref] New genotype-phenotype linkages for directed evolution of functional proteins, Leemhuis [/bib_ref]. The genes, a single member of a nucleic acid library, are transcribed and translated in microdroplets by an in vitro transcription/translation (IVTT) extract.
This approach offer a complete in vitro system in which the selection environment is not limited to conditions compatible with cell survival (such as pH, temperature or solvents) [bib_ref] New genotype-phenotype linkages for directed evolution of functional proteins, Leemhuis [/bib_ref]. Microdroplets containing a desired phenotype are selected by a suitable strategy such as fluorescence detection.
Ghadessy et al. [bib_ref] Molecular breeding of polymerases for amplification of ancient DNA, Abbadie [/bib_ref] dispersed the cells into droplets together with primers and dNTPs, and the droplets are subjected to thermal cycling. The polymerase and its gene are released from the cell, allowing self-replication by PCR. Amplified genes are then re-cloned for further selection.
Mastrobattista et al. [bib_ref] High-throughput screening of enzyme libraries: in vitro evolution of a beta-galactosidase by..., Mastrobattista [/bib_ref] dispersed a library of genes into droplets in which the genes are transcribed and translated in vitro. Moreover, active enzymes convert a non-fluorescent substrate into a fluorescent product. Fluorescent droplets are separated from non-fluorescent droplets using a fluorescence-dependent sorting. An alternative strategy that takes advantage of sorting by fluorescence is microbead display [bib_ref] Microbeads display of proteins using emulsion PCR and cell-free protein synthesis, Gan [/bib_ref]. Beads carrying one gene of a library, each with an epitope tag, and antibodies against this tag are compartmentalized inside the droplets. The translated proteins can attach to the beads via the epitope tag-antibody interaction. The emulsion is broken down and the beads, displaying multiple copies of the protein, are sorted by fluorescence detection. For understanding the functions of proteins, the microfluidic droplets can also be used to address the effective analysis of enzyme-substrate reactions, protein-protein interactions, and protein modifications. Srisa-Art et al.
studied the binding kinetics of streptavidin and biotin using FRET between two fluorescent dyes [bib_ref] Monitoring of real-time streptavidin−biotin binding kinetics using droplet microfluidics, Monpichar [/bib_ref] as well as between angiogenin and an anti-angiogenin antibody [bib_ref] Analysis of protein-protein interactions by using droplet-based microfluidics, Monpichar [/bib_ref].
## Protein application
## Microfluidic chip for separation
Many researchers have exploited the formation of liquid droplets in microfluidic systems to perform a variety of analytical processes due to its distinct advantages with respective to speed, analytical throughput, reagent usage, process control, automation and operational flexibility. In particular, in vitro compartmentalization of reactions in water-in-oil droplets combines the necessary ability to carry out large numbers of experiments under controlled conditions with quantitative readout, and has recently advanced towards automation by generating droplets in microfluidic devices.
Generally, liquid droplets can be formed spontaneously through flow instabilities between two immiscible fluid layers. Each droplet is isolated from other droplets therefore each one acts as an individual reaction vessel. More importantly, liquid droplets can be generated at high frequencies (e.g., kHz), meaning that millions of individual reactions can be processed in a single experiment. In addition to droplet formation, the microfluidic technology allows the integration of different unit operations in which droplets can be divided, fused, incubated, analyzed, sorted and broken up [bib_ref] Microdroplets: A sea of applications, Huebner [/bib_ref] [bib_ref] Reactions in droplets in microfluidic channels, Song [/bib_ref] [bib_ref] Lab-on-a-chip in vitro compartmentalization technologies for protein studies, Zhu [/bib_ref]. Integration of these steps can potentially create a system for demanding biological experimentation. For example, after the droplets have been formed, they can be kept moving in microchannels or captured in traps or reservoirs. The droplets also can be incubated offline and re-injected into the device for further manipulations such as splitting or fusion. To detect the reaction within the droplets, the most frequently used system is based on the readout of fluorescence where the fluorescent droplets can be quantitatively detected and sorted from non-fluorescent droplets [bib_ref] Overview: methods and applications for droplet compartmentalization of biology, Miller [/bib_ref].
Hollfelder at University of Cambridge and deMello at Imperial College London have done extensive works on droplet-based microfluidics for high-throughput chemistry and biology [bib_ref] The potential of microfluidic water-in-oil droplets in experimental biology, Schaerliab [/bib_ref].
## Bio-functional polymer particles/microgels assist in concentration
Recently, uniformly micron-sized polymer particles which carry with functional groups are exploited for several biotechnological and biomedical applications, such as affinity chromatography, immobilization technologies, drug delivery, and cell culture [bib_ref] Functional polymer microspheres, Kawaguchi [/bib_ref] [bib_ref] Measurement of enzyme kinetics using a continuous-flow microfluidic system, Seong [/bib_ref]. Detection, immobilization, and separation of DNA, cells, and proteins require monodisperse particles carrying surface functionalities, e.g., carboxyl, hydroxyl, amine, amide and chloromethyl groups. Polymer particles within the size range of 50 nm to 2 mm can be produced by various manufacturing processes including suspension, emulsion, and dispersion polymerizations [bib_ref] Microspheres for biomedical applications: preparation of reactive and labelled microspheres, Arshady [/bib_ref] [bib_ref] Micron-size crosslinked microspheres bearing carboxyl groups via dispersion copolymerization, Zhang [/bib_ref]. However, these methods are usually time-consuming, and provide insufficient control over particle size distribution and particle compositions.
A microfluidic-based method [bib_ref] Microfluidics: from dynamic lattices to periodic arrays of polymer disks, Seo [/bib_ref] [bib_ref] Formation of biphasic janus droplets in a microfabricated channel for the synthesis..., Nisisako [/bib_ref] for the rapid continuous in-situ photopolymerization of monomer droplets emulsified in a microfluidic flow-focusing device (MFFD) has been proposed and proven to produce extremely monodispersed polymer particles. Typically, the two-dimensional (2D) MFFDs composed by PDMS (polydimethyl siloxane) microchannels on glass substrates were extensively used to produce polymer particles. However, wetting properties associated with both liquids and microchannels for current 2D MFFDs are a vital issue in the generations of monomer droplets. This indicates that the wetting of the channel surfaces, i.e., hydrophobicity and hydrophilicity, has a significant effect on the type of dispersion that can be produced [bib_ref] Microfluidics: from dynamic lattices to periodic arrays of polymer disks, Seo [/bib_ref]. To avoid wetting of the channel wall by the dispensed phase, axisymmetric flow-focusing device (AFFD) is a promising technique as the inner dispensed phase is surrounded by the continuous phase and never touches the channel wall [bib_ref] An axisymmetric flow-focusing microfluidic device, Takeuchi [/bib_ref] [bib_ref] Monodisperse double emulsions generated from a microcapillary device, Utada [/bib_ref]. Huang et al.proposed a planar 3D MFFD utilizing PDMS with a simplified fabrication process, which can produce monodispersed copolymer particles carrying surface carboxyl groups in the range of 50~200 µm prepared through in-situ UV polymerization of ethyleneglycol dimethacrylate (EGDMA) with acrylic acid (AA) [fig_ref] Figure 8: Schematic diagram of MFFD for production of copolymer particles via in-situ UV... [/fig_ref]. High efficiency of bioconjugation on carboxylated copolymer particles was successfully demonstrated by increasing the concentration of AA.
## Immunoassays and other specific sensing
Lateral Flow (LF) Immunoassays or Immunochromatographic strips (ICS) or Test strips, as they are called, are one of the most successful point-of-care diagnostics devices, commercially available for detecting various health and environmental threats. Lateral flow (LF) immunosensors which use antibodies as a biosensing element serves as rapid, handy, and disposable tools for point-of-care detection. They also can be regarded as one of the most successful microfluidic platforms for µTAS due to their simplicity, cost and integration with other systems [bib_ref] The triage cardiac panel: Cardiac markers for the triage system, Clark [/bib_ref] [bib_ref] Fabrication of a Disposable Biosensor for Escherichia coli O157:H7 Detection, Qureshi [/bib_ref] [bib_ref] Wicking Assays for the Rapid Detection of West Nile and St. Louis..., Ryan [/bib_ref] [bib_ref] Quantitative Impedimetric Immunosensor for Free and Total Prostate Specific Antigen Based on..., Fernandez-Sanchez [/bib_ref].
Examples of the applications for LF assays include test kits for detecting pregnancy, drugs of abuse, infectious diseases, cancers, and cardiovascular disorders [bib_ref] Microfluidic diagnostic technologies for global public health, Yager [/bib_ref]. The system is normally composed of porous membranes which serve to immobilize biological elements for biosensing and to transport sample reagents by capillary force. The basic operation principle of the strips is passive liquid manipulation through capillary forces within the capillaries of a fleece or a micro-structured channel to control the flow of fluids during immunoassay. The liquid samples are loaded into a sample reservoir from where they penetrate the underlying fleeces or direct capillary filling of the strip from a sample loading point for self testing applications. In general, the test results from LF assays are based on optical detection where fluorescent markers or tagged colored particles or enzymes generate signal upon the completion of specific molecular recognition. The sample fluid with proper flow speed passes through the detection zone to ensure the sample molecules tagged by the marker molecules binding to the immobilized element capture molecules/antibodies. Since the concentration of markers is quite low, they have to be collected in the detection zone to get remarkable signals. By using fluorescent markers, the fluorescence at each discrete detection zone is measured by a fluorometer, a reader with some optical components. In addition, gold or latex particles are used to produce a readable signal. The detection molecules binding with nano gold or latex particles accumulate at the detection zone and the color shows up. The use of tagged color particles can achieve cheap and fast reading of assay results; however, it can only be suitable for readout of binary signal assays, such as pregnancy tests. Besides fluorescent and colorimetric techniques, electrochemical method is also used for detection, e.g., blood glucose.
## Cellomics application
## Capillary electrophoresis microchip for cells manipulation and analysis
Cell plays a significant role as the fundamental unit of life and hence high efficient and sensitive detection of the biomolecules or subcellular substance released from single cells are desirable for the studies of physiological processes [bib_ref] Lab-on-a-chip: microfluidics in drug discovery, Ali [/bib_ref]. The single cell analysis approach not only provides complementary information of complex biological systems but also reveals the actual functional interaction of biomolecules on the cellular and tissue structural basis [bib_ref] Dynamic single-cell analysis for quantitative biology, Carlo [/bib_ref]. In fact, single cell studies are much more complicated and time-consuming than their population counterparts, though the response and interaction of individual cells within a heterogeneous population are difficult to be assessed. Various components (nucleic acids, proteins, carbohydrates) of a cell could be separated into different channels on an appropriate miniaturized platform, and then individually analyzed with the help of the respective individual 'cellomics' techniques (details see the [fig_ref] Figure 1: Schematic of one idealized total analysis device showing the various functions on... [/fig_ref] of [bib_ref] Scrapheap challenge and single cell, Rupak [/bib_ref].
There are various assays which have been developed on the capillary microfluidic platform for rapid and accurate sensing of biomarkers in clinical diagnostics and point-of-care applications. In developed countries, interest in moving to a more patient-centric point-of-care /home-testing approach arises and near-patient testing using point-of-care devices has become popular. The cardiovascular disease has become one of the predominant causes of death and disability in all developed countries.
Early detection of heart failure or heart attack significantly improves a patient's survival rates and minimizes the permanent damage to tissue and organs. Therefore, reliable, fast and accurate sensing of cardiovascular disease is becoming increasingly important in point-of-care settings.
Tweedie et al. [bib_ref] Fabrication of impedimetric sensors for label-free Point-of-Care immunoassay cardiac marker systems, with..., Tweedie [/bib_ref] at Ulster University has developed miniaturized point-of-care sensors for cardiovascular disease markers based on impedimetric sensing of cardiac enzyme capture by an antibody layer immobilized on a planar gold electrode sensor and passive microfluidic delivery. They focused on the detection of cardiac markers of heart attack and acute myocardial infarction (AMI) such as BNP, CK-MB, Myoglobin and Troponin I. The use of impedimetric sensing is advantageous as it is label-free, not requiring the use of any fluorescent reagents. Therefore, it is simpler in operation, and potentially, easier for throughput assay. Passive capillary flow input of patient whole blood, or filtered blood/serum, makes such devices simpler to use, and easier and cheaper to manufacture than other micropumped systems. Another group in UKalso proposed a point-of-care diagnostic device for a more reliable and earlier assessment of deep vein thrombosis and related blood clotting conditions. The device measures whole blood concentration of D-dimer, a recognized biomarker of increasing blood clotting activity, through immunochemical biosensor comprising antibody captured onto the surface of impedimetric analysis electrodes. The detection system lies within a disposable liquid handling cartridge, containing a microfluidic system for whole blood handling capabilities and test and data management. Its transfer system is integrated with a base unit.
There are a limited number of commercially available systems for point-of-care diagnostics, for example, Biosite Inc., USA. They developed a rapid and accurate clinical diagnostic platform, containing a disposable cartridge and a reader, for cardiovascular diseases and illegal drugs testing in 16 different immunoassays. The disposable cartridge contains passive microcapillaries to control the mixing of samples with reagents and to control the flow of fluid during immunoassay, and also contains a blood filter to trap blood cells. This platform uses dehydrated fluorescent reagents on-board with optical rather than impedimetric sensing. The high sensitivity fluorescent dyes have following characteristics: (1) can be excited with near infrared wavelengths, and therefore are usable with complex biological samples such as serum, plasma, and whole blood; (2) stable for the shelf period of the product; (3) no overlap between emission and excitation spectrums, so that detection can be performed with inexpensive solid state electronic components.
## Nanoparticles assist in subcellular analysis
In recent years, suspension arrays [bib_ref] Suspension array technology: Evolution of the flat array paradigm, Nolan [/bib_ref] , which use self-encoded microcarriers as sensing elements, are attracting increasing interest in the field of drug discovery, gene-expression analysis, and clinical diagnosis [bib_ref] Fiber Optic Array Biosensors, Walt [/bib_ref] [bib_ref] Duplexed sandwich immunoassays on a fiber-optic microarray, Rissin [/bib_ref]. Photonic crystals used in suspension arrays have been suggested as a new type of optical spectrum-encoding carrier, whose code is the characteristic reflection peak based on their periodical structure. The code of the reflection peak is very stable, and the fluorescent background is low. These properties render photonic crystals suitable for highly sensitive detection [bib_ref] Encoded porous beads for label-free multiplex detection of tumor markers, Zhao [/bib_ref]. . The scheme of the microfluidic device for the fabrication of photonic crystal beads by means of evaporation or UV polymerization to aggregate the colloidal nanoparticles that self-assembled in close-packed structures [bib_ref] Optofluidic Encapsulation of Crystalline Colloidal Arrays into Spherical Membrane, Kim [/bib_ref].
Microfluidic devices for the fabrication of colloidal crystal beads have been proposed [bib_ref] Optofluidic Encapsulation of Crystalline Colloidal Arrays into Spherical Membrane, Kim [/bib_ref]. In this method, the oil-phase flow and the aqueous-phase flow were simultaneously injected into microchannels by syringe pumps. The aqueous suspension with the colloidal nanoparticles was cut off into droplets by the oil flow. This method can fabricate colloidal crystal beads with controllable size, narrow size distribution, and good repeatability. The generated droplets were then heated or polymerized by UV irradiation to aggregate the colloidal nanoparticles that self-assembled in closepacked structures after solidification . Therefore, the beads showed photonic crystal features, whose reflection peaks were used as the coding elements of biomolecular carriers [bib_ref] Photonic crystal hydrogel beads used for multiplex biomolecular detection, Hu [/bib_ref]. Huge coding capacity could be gained by changing the volume fraction of the nanoparticles in the droplets.
Gu et al.proposed a novel type of microcarrier developed for optical encoding and fluorescence enhancement by depositing semiconductor quantum dots (QDs) on silica colloidal crystals beads (SCCBs). Monodispersed SCCBs were used as the support to deposit different-sized and different-layer QDs, thus leading to wavelength-and-intensity coding. The unique properties of QDs were well suited for multiplexed optical encoding, which could potentially yield a large number of molecular bar codes. The SCCBs possess high porosity and high surface-to-volume ratio (SVR), which can provide stronger detection signals and thus high detection sensitivity. Gu et al. also proposed the encoded porous beads for label-free multiplex of detection of tumor markers by the use of inverse-opaline photonic beads as carriers for suspension arrays [bib_ref] Encoded porous beads for label-free multiplex detection of tumor markers, Zhao [/bib_ref]. The beads showed large encoding capacity by changing their lattice constant. A label-free multiplex detection of tumor markers, Human CA125, CA19-9, and CEA, which showed great significance in early screening and clinical diagnosis of some tumor diseases including colorectal cancer, gastric cancer, and lung cancer, provides the flexibility and feasibility in emerging clinical applications.
## Single cell analysis and droplets compartmentalize cells
From a technical point of view, it is difficult process to perform on biomedical single cells, especially in advanced species. Such analysis calls for significant improvements in terms of cell manipulation, lysis, and separation of cellular constituents, detection sensitivity, and throughput.
During the past two decades, a number of novel techniques have been developed for carrying out single cell analysis [bib_ref] Recent developments in single-cell analysis, Lu [/bib_ref]. Capillary electrophoresis/electrochromatography is one of the excellent techniques for identifying and quantifying the contents of single cells [bib_ref] Recent developments in capillary electrophoresis and capillary electrochromatography of peptides, Kasicka [/bib_ref]. However, conventional capillary-based techniques lack the ease of high-throughput analysis of single cells due to even smaller sample volumes and masses found in mammalian cells.
The microfluidic devices provide a multiple platform for single cell analysis owing to their unique characteristics to overcome this problem. These devices open the possibility of integrating a variety of cellular operations on the micrometer scale, such as the positioning, trapping or lysis of single cells, as well as detection, analysis and even separation of cellular compounds [bib_ref] Analysis of single mammalian cells on-chip, Sims [/bib_ref]. Such lab-on-a-chip devices further allow the analysis with improved performance, throughput, parallelization and automation. Lysed cells lose their integrity as well as the likelihood of enabling identification of their specific phenotypes [bib_ref] Microtechnologies and nanotechnologies for single-cell analysis, Andersson [/bib_ref]. A goal of modern biology is to understand the molecular mechanisms underlying cellular function. The ability to manipulate and analyze single cells is crucial for this task.
The advancement of microengineering has provided biologists with unprecedented opportunities for cell handling and investigation on a cell-by-cell basis [bib_ref] Microfluidic devices for cellomics: A review, Andersson [/bib_ref]. As a result, lab-on-a-chip technologies are emerging as the next revolution in tools for biological discovery. Cell interactions are critical to the function of many organ systems and hence further understanding of cell-material interactions play an important role in the investigation of their physiological traits and functions so as to gain fundamental insight as well as suggest approaches that will allow the manipulation of tissue function in vitro for clinical applications [bib_ref] In vivo Osteocyte Death, Chao [/bib_ref]. Cell to cell signaling has also been investigated using microfluidic chip based methods [bib_ref] Capillary electrophoresis coupled to biosensor detection, Bossi [/bib_ref] [bib_ref] Chip based electroanalytical systems for cell Analysis, Spegel [/bib_ref].
Microfluidic droplets can also be used to compartmentalize a number of cell types, such as bacteria [bib_ref] Simultaneous Determination of Gene Expression and Enzymatic Activity in Individual Bacterial Cells..., Shim [/bib_ref] , yeast [bib_ref] Dropspots: a picoliter array in a microfluidic device, Schmitz [/bib_ref] and mammaliancells, to study the biological function of a single cell for performing drug screening with high sensitivity of detection. Cells have been shown to remain viable in several oil-surfactant mixtures and device designs. It is possible to incubate the droplets for a long-term storage up to several days offline and re-inject them into a device for analysis [bib_ref] Droplet microfluidic technology for single-cell high-throughput screening, Brouzes [/bib_ref].
Because the environmental conditions in a droplet can be well controlled, the intracellular components and secretion from cells are constrained within the droplet compartment. The small compartment size creates a high local concentration and leads to high sensitivity of detection. Therefore, cells in droplets can be interrogated by multiple optical methods, such as molecular markers including fluorescent proteins or reporter enzymes [bib_ref] Visualizing biochemical activities in living cells, Johnsson [/bib_ref] for cellular processes, and imaging of cell morphology [bib_ref] Extracting rich information from images, Carpenter [/bib_ref]. In addition, fluorescent life time imagingluminescence detection [bib_ref] Mixing crowded biological solutions in milliseconds, Liau [/bib_ref] , mass spectrometric analysis [bib_ref] Analysis of samples stored as individual plugs in a capillary by electrospray..., Pei [/bib_ref] , and surface enhanced Raman scattering detection [bib_ref] Surface-enhanced Raman scattering in nanoliter droplets: towards high-sensitivity detection of mercury (II)..., Wang [/bib_ref] have also been demonstrated.
## Prospective
As an enabling technique, microfluidic systems have provided new opportunities for the advancement in several emerging biomedical applications such as tissue engineering, drug delivery/testing, and stem cell therapy.
Tissue engineering aims to cultivate engineered tissues by harvesting cells from patient or donor and then seeding/culturing them on biomimetic scaffolds, fabricated from natural or man-made biomaterials. However, creating 3-D engineered tissue has been limited by how to effectively arrange cells and distribute nutrients into the scaffold during the formation of tissue constructs. To tackle this problem, MFS has recently been used as a tool for fabricating novel "Microfluidic Scaffolds". As such, embedding microfluidic networks directly within cell-seeded scaffolds can facilitate convective mass transfer for control of the distributions and fluxes of solutes in the bulk of the 3D culture [bib_ref] Microfluidic Scaffolds for Tissue Engineering, Choi [/bib_ref].
Microfluidic systems have also been used to develop perfusion-based micro 3-D cell culture platforms for high throughput drug testing [bib_ref] Development of perfusion-based micro 3-D culture platform and its application for high..., Wu [/bib_ref]. The microfluidic system fabricated based on soft lithography of PDMS and incorporated pneumatic-based pumping culture medium and cell-scaffold loading mechanism to form the cell culture platform has demonstrated to provide a homogenous and steady cell culture environment which will be useful for tissue engineering applications.
Microfluidic system can be easily integrated with microphotonics such as optical trapping to form micro-opto-fluidic system (MOFS) for single cell manipulation. Cultivating single cell in the MOFS can produce homogeneous daughter cells, which have a significant impact on stem cell therapy.
The method can potentially apply for the stem cell delivery or the creation of stem cell niche that is to create a microenvironment consisting of various pluripotent cells, appropriate biochemical solutions and mechanical stimuli, e.g., flow shear stress, for fostering stem cells to differentiate into desirable tissues. Recently MFS incorporated with holographic optical tweezers to form a platform which can potentially be used as a reconfigurable force sensor array with piconewton resolution to investigate chemo-mechanical processes. This new technique may provide a powerful tool for multi-cellular manipulation in the cell arrangements and stimuli of engineered tissue [bib_ref] Optical Tweezers for Single Cells, Zhang [/bib_ref]. MFS has also been used for the development of in vitro physiological systems for studying fundamental biological phenomena for tissue growth [bib_ref] Microfabrication and Microfluidics for Tissue Engineering: State of the Art and Future..., Andersson [/bib_ref]. Microfluidic chips have also provided an excellent approach for cell-based screening and detection of different toxicities [bib_ref] A Prototype Microfluidic Chip using Fluorescent Detection of Toxic Compounds, Garcia-Alonso [/bib_ref]. This technique can provide a low cost, fast speed, high throughput screening for testing different metabolic responses to drug on a cellular level and hence will be useful for in-situ tissue growth monitoring and drug testing. Microfluidic system can uniquely serve as a functional tool for studying cell mechanics which is important for the advancement of drug delivery and tissue engineering. A recent review has comprehensively highlighted its capabilities that are of significance for understanding the mechanical behaviors of cells.
[fig] Figure 1: Schematic of one idealized total analysis device showing the various functions on a micro fluidic chip[8]. [/fig]
[fig] Figure 2: (a) Photographic overhead view of the μCEC chip (dark area in the central channel consisting of MWCNTs array, 1: sample reservoir, 2: eluent buffer/running buffer reservoir, 3: sample waste, 4: eluent buffer waste, 5: running buffer waste, D: detector), (b) The schematic arrangement of MWCNTs in μCEC channels as vertically aligned nanopillars, (c) the SEM image of a cross section (a-a' plane) of the μCEC channel, (d) The SEM image of vertically-aligned MWCNTs directly grown in microchannel[65]. [/fig]
[fig] Figure 3: The schematics of (a) flow field evolution for DAEKF in a capillary electrophoresis nanochannel (solid black arrows represent analyte flow direction, andζ2>ζ1>ζ3>ζ0), (b) detailed flow fields of five different regions for the DAEKF system in a nanochannel. (Region I: a 2-D shear flow, Region II: pulling effect-asymmetric electroosmotic flow (AEOF), Regions IV: pushing effect AEOF). Fluorescence image ofRhodamine B for (c) traditional EOF in a nanochannel, (d) the restacking effect by the DAEKF system in a nanochannel[66]. [/fig]
[fig] Figure 4: Schematic diagrams of the MEMS-based 3D MFFD using three layers of SU-8 resist for formation of (a) single emulsions and (b) double emulsions. (c) SEM images of the 3D MFFD for formation of single emulsions with inset showing the flow-focusing orifice with dimensions of 50 µm(W) × 50 µm(H). (d) SEM image of the 3D MFFD configuration for formation of double emulsions. The inner fluid orifice measuring 100 µm(W) × 50 µm(H) × 100 µm(L) and the flow-focusing orifice measuring 200 µm(W) × 50 µm(H) × 100 µm(L) are coaxial, separated by a distance of 200 µm. The total height of the microchannels is 250 µm. (reprinted with permission from ref. -high-throughput assays on a very small scale is currently an emerging and powerful format and flow-focusing method is well-suited for compartmentalization of chemical or biological reactions in droplets (typically fL to nL volume) that act as discrete reaction vessels. Courtois et al. from University of Cambridge designed an integrated device allowing precisely controlled formation of W/O droplets based on flow-focusing which are stable for at least 6 h in a reservoir, an extended timescale for the study of in vitro expression of green fluorescent protein (GFP) [/fig]
[fig] Figure 5: Schematic of microarray system for batch-filling and in parallel printing of multiple proteins. (1) The micro connectors of the micro stamp chips are connected into the nozzles of the micro filling chip, and then the bio-fluids are transferred into the micro stamp simultaneously. (2) After the filling has been completed, those two chips are separated, and then PDMS stamps are used to print in parallel numerous arrays (reprinted with permission from ref. 92, ©2008 IEEE). [/fig]
[fig] Figure 6: (a) Side view images of a 2 μL surface-ascending water droplet under different inclination angles θ (40 o , 90 o , 130 o , and 180 o ) moving along the gradient with ψ = 8 o and L = 12 mm. (b) A time sequence of top view images of a self-directed subnanoliter water droplet (0.21 nL) moving on a device with an array of circulating wedge-shape gradients. The dashed line indicates the coverage area of the gradients with the detailed dimensions shown in the enlarged view (after photoresist development) of the device. The arrows indicate the position of the droplet of which is transient (reprinted with permission from ref. 101, Copyright 2009, American Institute of Physics). [/fig]
[fig] Figure 7: A continuous-flow PCR device where microfluidic droplets move through a temperature gradient toward the radial direction. The device contains an oil inlet (A) that joins an aqueous inlet channels (B) to form droplets (C). The droplets pass through the inner circles in the hot zone (D) to ensure initial denaturation of the template and travel on to the periphery where primer annealing and template extension occur (E). The droplets then flow back to the centre, where the DNA is denatured and a new cycle begins. For illustration, only 7 cycles are demonstrated, but 34 cycles have been achieved in [/fig]
[fig] Figure 8: Schematic diagram of MFFD for production of copolymer particles via in-situ UV polymerization by co-flow of aqueous (A) and comonomer (B) phases. Typical fluorescent images of the copolymer particles conjugated with IgG-Cy3 for C AA = 40 wt%.Fluid A: DI water + 2 wt% SDS, Fluid B: monomer (ethyleneglycol dimethacrylate, EGDMA) + 0~40 wt% acrylic acid (AA) + 4wt% photoinitiator (HCPK, 1hydroxycyclohexyl phenyl ketone)[140]. [/fig]
|
Increased epigenetic alterations at the promoters of transcriptional regulators following inadequate maternal gestational weight gain
Epigenetic modifications are thought to serve as a memory of exposure to in utero environments. However, few human studies have investigated the associations between maternal nutritional conditions during pregnancy and epigenetic alterations in offspring. In this study, we report genomewide methylation profiles for 33 postpartum placentas from pregnancies of normal and foetal growth restriction with various extents of maternal gestational weight gain. Epigenetic alterations accumulate in the placenta under adverse in utero environments, as shown by application of Smirnov-Grubbs' outlier test. Moreover, hypermethylation occurs frequently at the promoter regions of transcriptional regulator genes, including polycomb targets and zinc-finger genes, as shown by annotations of the genomic and functional features of loci with altered DNA methylation. Aberrant epigenetic modifications at such developmental regulator loci, if occurring in foetuses as well, will elevate the risk of developing various diseases, including metabolic and mental disorders, later in life.
Scientific RepoRts | 5:14224 | DOi: [bib_ref] Epigenetic programming of diabetes and obesity: animal models, Seki [/bib_ref].1038/srep14224 of CpGs 12 . The genome-wide DNA methylation pattern in various types of cells in the body is bimodal, with the majority of CpG sites being highly methylated (> 85%), while CpG islands (CGIs) are largely unmethylated (< 10%). After fertilisation of the oocyte by sperm, the paternal genome becomes actively demethylated in the zygote, and the maternal genome undergoes passive demethylation until the early blastocyst stage [bib_ref] Dynamic CpG island methylation landscape in oocytes and preimplantation embryos, Smallwood [/bib_ref]. DNA methylation patterns are re-established in a lineage-specific manner [bib_ref] A unique regulatory phase of DNA methylation in the early mammalian embryo, Smith [/bib_ref] [bib_ref] Epigenetic programming and reprogramming during development, Cantone [/bib_ref]. The genome in cells of the placenta remains relatively hypomethylated compared to that in somatic tissues. A genome-wide DNA methylation analysis of promoter regions in human placental tissues collected during the first, second, and third trimesters revealed that there is a significant increase in the average methylation level in autosomes from the second-trimester placenta to the full-term placenta [bib_ref] Evidence for widespread changes in promoter methylation profile in human placenta in..., Novakovic [/bib_ref]. Interindividual variations in DNA methylation levels have also been shown to increase during gestation [bib_ref] Evidence for widespread changes in promoter methylation profile in human placenta in..., Novakovic [/bib_ref]. Foetal environmental factors, such as maternal weight, maternal alcohol intake, maternal smoking, and maternal psychological stress, have recently been shown to affect DNA methylation in the human placenta However, compared to the number of studies that have examined the DNA methylation levels at repetitive sequences, such as long interspersed nuclear elements (LINEs) and short interspersed nuclear elements (SINEs), and at certain imprinted loci [bib_ref] Placental adaptations to the maternal-fetal environment: implications for fetal growth and developmental..., Sandovici [/bib_ref] [bib_ref] Birthweight, maternal weight trajectories and global DNA methylation of LINE-1 repetitive elements, Michels [/bib_ref] [bib_ref] In utero exposures, infant growth, and DNA methylation of repetitive elements and..., Wilhelm-Benartzi [/bib_ref] , studies evaluating the alterations in DNA methylation on the genome-wide scale in relation to the foetal environment have been limited [bib_ref] Infant growth restriction is associated with distinct patterns of DNA methylation in..., Banister [/bib_ref] [bib_ref] In utero exposures, infant growth, and DNA methylation of repetitive elements and..., Wilhelm-Benartzi [/bib_ref] [bib_ref] Maternal tobacco use modestly alters correlated epigenome-wide placental DNA methylation and gene..., Suter [/bib_ref] [bib_ref] Gestational diabetes mellitus epigenetically affects genes predominantly involved in metabolic diseases, Ruchat [/bib_ref] [bib_ref] The effect of genotype and in utero environment on interindividual variation in..., The [/bib_ref].
Maternal gestational weight gain (GWG) influences the foetal nutritional environment during gestation. The Japan Ministry of Health, Labour, and Welfare recommends that women with a prepregnancy body mass index (BMI) between 18.5 and 25 should gain about 7-12 kg body weight during the gestational period. Excessive GWG has been reported to be associated with increased neonatal obesity during infancy and adulthood [bib_ref] Excessive weight gain in women with a normal pre-pregnancy BMI is associated..., Josefson [/bib_ref] [bib_ref] Gestational weight gain in relation to offspring body mass index and obesity..., Schack-Nielsen [/bib_ref] [bib_ref] Effects of suboptimal or excessive gestational weight gain on childhood overweight and..., Ensenauer [/bib_ref]. On the other hand, insufficient GWG is related to increased risk of low birth weight [bib_ref] Prepregnancy weight, gestational weight gain, and risk of growth affected neonates, Simas [/bib_ref] [bib_ref] Association of second and third trimester weight gain in pregnancy with maternal..., Drehmer [/bib_ref] , which is known to be associated with metabolic syndrome, including impaired glucose tolerance, insulin resistance, and coronary heart disease, during adulthood [bib_ref] The long-term renal and cardiovascular consequences of prematurity, Abitbol [/bib_ref] [bib_ref] Developmental and environmental epigenetic programming of the endocrine pancreas: consequences for type..., Sandovici [/bib_ref] [bib_ref] Early determinants of type-2 diabetes, Berends [/bib_ref] [bib_ref] Effect of fetal and child health on kidney development and long-term risk..., Luyckx [/bib_ref]. Foetal growth restriction (FGR), which may be caused by foetal, placental, and/or maternal factors, is defined as a foetus that has not reached its growth potential (below the 10th percentile for gestational age).
In this study, we elucidate the effects of in utero environments on the human placental epigenome. To this end, we examine a collection of postpartum placentas using array-based genome-wide DNA methylation analysis and evaluate DNA methylation levels in placental tissues in relation to GWG and birth weight. We demonstrate that inadequate GWG perturbs the placental epigenome variably among subjects, and that such epigenetic alterations occur preferentially at the CGI promoters of genes encoding transcriptional factors. Therefore, our results demonstrate that epigenetic alterations accumulate in the placenta under adverse in utero environments, supporting the importance of appropriate in utero conditions and maternal health in foetal development.
# Results
Alterations in placental DNA methylation were associated with FGR and GWG. We subjected placentas from 14 births exhibiting FGR and 19 births within the normal range of birth weight to genome-wide DNA methylation analysis, and assessed whether the FGR placentas contained CpG sites that were differentially methylated compared with the placentas with a birth weight within normal range. Wilcoxon rank-sum tests 31 did not detect any CpG sites as significantly differentially methylated between two groups (significance level = Benjamini-Hochberg [BH] adjusted p-value of 0.05). Comparisons of FGR and normal placentas within subgroups depending on maternal GWG (insufficient, adequate, and excessive) also did not detect any differentially methylated CpG sites in the FGR placentas. Next, we assessed whether placentas from subjects with excessive or insufficient maternal GWG contained CpG sites that were differentially methylated compared with those in placentas from subjects with adequate GWG. Four comparisons (i.e., insufficient versus adequate and excessive versus adequate within FGR and normal categories) did not detect any significantly differentially methylated CpG sites between two subgroups. These results suggested that no specific CpG sites showed consistent changes in DNA methylation associated with the FGR phenotype or inadequate maternal GWG in this study.
Next, we considered the possibility that the FGR phenotype and/or inadequate GWG may affect the placental epigenome in different ways among individual subjects rather than showing similar effects for all individuals within a group. To evaluate this possibility, we searched for CpG sites whose methylation level differed significantly in one placenta (as compared with all of the other placenta samples) by performing Smirnov-Grubbs' outlier test with Bonferroni multiple test corrections (significant level = 0.1) for each placenta. We detected 2,983 and 1,416 CpG sites as hyper-and hypomethylated outliers, respectively, among the 33 subjects. To reduce the numbers of outliers that could have been detected spuriously due to SNPs at/near the target CpG sites, we excluded the CpG sites whose corresponding probes are annotated to contain known SNPs as described in the Methods. When 89,678 probes were regarded as potentially SNP-containing based on the Illumina probe annotation, 2,521 (85%) and 977 (69%) CpG sites remained as hyper-and hypomethylated outliers, respectively.
We subjected these remaining outliers to further data analyses. Hypomethylated outliers coincided with SNP-containing probes more often than hypermethylated outliers (439/1,416 (31%) versus 462/2,983 (15%)). The mean (standard deviation β values of the 2,521 hyper-and the 977 hypomethylated outliers were 0.24 (0.13) and 0.56 (0.19), respectively. The mean (SD) Δ β values (Δ β = the β value of the Scientific RepoRts | 5:14224 | DOi: 10.1038/srep14224 outlier-the mean β value of the other samples) of hyper-and hypomethylated outliers were 0.18 (0.11) and -0.27 (0.12), respectively.
While the numbers of outliers in the "normal_adequate" category were low and relatively consistent among subjects (ranging from 46 to 74), those in the other five categories were higher, exhibiting statistical significance (Tukey's multiple comparison test p-value < 0.001;and diverse among the subjects (ranging from 44 to 421). In normal subjects, the greater the insufficiency or excessiveness of maternal GWG, the higher the number of methylation outliers, as represented by the U-shaped appearance of the bar plots for the number of outliers in normal subjects sorted according to weight gained during pregnancy. In FGR subjects, all three subcategories (FGR_insufficient, FGR_adequate, and FGR_excessive) contained significantly higher numbers of outliers than the "normal_adequate" category. The numbers of outliers in FGR_insufficient and FGR_excessive categories were also significantly higher than those in FGR_adequate. The numbers of outliers were neither associated with C-section nor correlated with gestational weeks (Supplementary . These results suggested that both FGR and inadequate GWG conditions affected the placental epigenome independently and additively.
Next, we examined the numbers of hyper-and hypomethylated outliers in each subject. While the numbers of hypomethylated outliers were not much different among subjects, the numbers of hypermethylated outliers were significantly higher in subjects in the other five categories compared to subjects in the "normal_adequate" category (P < 0.001;. Therefore, only hypermethylated outliers occurred with FGR pregnancies and normal pregnancies with inadequate GWG. Because of the nature of the Smirnov-Grubbs' outlier test, the identified methylation outliers were all specific to individuals (deviated only in one sample among the cohort). Our results demonstrate that the adverse pregnancy conditions, FGR and inadequate GWG, affected the placental epigenome variably among individuals.
Contrasting genomic features of hyper-and hypomethylated outliers. We subsequently annotated the genomic features of 2,521 hyper-and 977 hypomethylated outliers [fig_ref] Figure 2: Genomic features of 2,521 hyper-and 977 hypomethylated outliers [/fig_ref]. Among these outliers, 2,107 (84%) and 758 (78%) CpG sites were located in genic regions (in 1,001 and 606 genes, respectively). Hypermethylated outliers were found to be predominantly located in CGIs or their shores/ shelves (94% in total) and proximal to the transcriptional start sites (defined as "pTSS" hereafter; i.e., TSS1500, TSS200, the 5′ untranslated region [UTR], and the first exon categories; 77%). In contrast, hypomethylated outliers were most frequently located outside of CGIs, shores, and shelves (open sea, 46%) and in gene bodies (66%) [fig_ref] Figure 2: Genomic features of 2,521 hyper-and 977 hypomethylated outliers [/fig_ref]. Hypermethylated outliers tended to be detected consecutively at two or more adjacent probes ("clustered"; 56%), while hypomethylated outliers did not (11%; [fig_ref] Figure 2: Genomic features of 2,521 hyper-and 977 hypomethylated outliers [/fig_ref]. These results implied that hypermethylated outliers tended to be clustered within CGI promoters. We therefore scrutinised the extent of hypermethylation and the positional distribution relative to the TSS of hypermethylated outliers by visualising β and Δ β values on the Integrative Genomics Viewer (IGV, www.broadinstitute.org/igv/home). Indeed, we found that hypermethylated outliers were often distributed in a promoter-wide manner (i.e., located consecutively and clustered around the TSS) with relatively large methylation differences, as exemplified by FOXC1, FOXL2, and HOXB7 loci [fig_ref] Figure 3: Examples of promoter-wide hypermethylation at FOXC1 [/fig_ref]. The methylation statuses in the outlier sample and a control (Normal_adequate_7) at these promoter regions were validated to be hypermethylated and unmethylated, respectively, by targeted bisulfite sequencing analyses [fig_ref] Figure 3: Examples of promoter-wide hypermethylation at FOXC1 [/fig_ref]. The appearance of both of heavily methylated and unmethylated clones in individual outlier samples may indicate the mosaic composition of normal and epimutated cells in these placentas.
The observation that hypermethylated outliers were often clustered at CGI promoters suggested that placental hypermethylation events do not occur in a purely random manner in terms of genomic location, but instead occur due to dysfunction of certain intrinsic mechanisms regulating the epigenetic status of CGI promoters under adverse in utero environments.
Hypermethylated outliers were frequently associated with genes encoding transcriptional regulators. In order to search for functional characteristics of genes containing hypermethylated outliers, we performed gene ontology (GO) analysis; 1,001 genes hosting hypermethylated outliers (as well as 606 genes hosting hypomethylated outliers for comparison) were analysed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) v6.7. The 606 genes hosting hypomethylated i.e., TSS1500, TSS200, the 5′ UTR, the first exon, the gene body, and the 3′ UTR, are regarded as genic regions, in which 2,107 (84%) and 758 (78%) CpG sites were located. The ratio of clustered and isolated outliers is shown (C).
Scientific RepoRts | 5:14224 | DOi: 10.1038/srep14224 outliers were found to be weakly enriched with only one term, "cytoskeletal protein binding", in the Molecular Function (MF) category (Benjamini's corrected Pc = 0.0025). However, the 1,001 genes hosting hypermethylated outliers were highly enriched with terms related to transcriptional regulators and neuronal differentiation in the Biological Process (BP) and MF categories (e.g., BP terms "regulation of transcription, DNA-dependent" [Pc = 1.96 × 10 −8 ] and "neuronal differentiation" [Pc = 3.16 × 10 −8 ];. We subsequently performed GO analysis for subgroups of genes: 409 genes hosting highly deviated (Δ β > 0.2) hypermethylated outliers, 709 genes hosting hypermethylated outliers in the pTSS, and 317 genes hosting two or more clustered hypermethylated outliers. These subgroups of genes were also found to be significantly enriched with terms related to transcriptional regulators. These results supported our observation that hypermethylated outliers are often distributed in a promoter-wide manner and that the genes hosting such outliers are significantly enriched with genes encoding transcriptional regulators. We further performed GO analysis for the 163 genes hosting highly deviated (Δ β > 0.2) and clustered hypermethylated outliers in the pTSS. Among those, 36 genes were assigned to the category "GO:0006355~regulation of transcription, DNA-dependent" with a statistical significance [Pc = 0.0038] and showed a higher fold enrichment value to the term than that of the entire (1,001) genes (2.25 versus 1.70, Supplementary. Importantly, in 35 out of the 36 genes encoding transcriptional regulators (97%), promoter hypermethylation was detected in the placentas from cases of inadequate GWG or FGR.
# Discussion
In this study, we demonstrated the possibility that inadequate maternal GWG enhances aberrant DNA methylation in the placenta. We initially failed to identify specific loci whose methylation was commonly altered across all subjects in each of the GWG categories. We subsequently used Smirnov-Grubbs' outlier tests, which detect the most significantly deviated outlier among subjects, for each of the CpG probes and found that hypermethylated loci accumulated in normal pregnancies with inadequate GWG and in FGR pregnancies. The results suggested that the epigenetically affected loci due to adverse in utero environments were variable among the subjects examined in this study. It should be noted that the relatively small number of the enrolled subjects (partly due to exclusion of the subjects with certain types of pregnancy complications) with various layers of heterogeneities (e.g., genetic, phenotypic, and environmental) may account for a primary cause of the absence of commonly epigenetically affected loci and the variation of affected loci among the individuals studied. While many animal studies have clearly demonstrated direct associations between in utero nutritional conditions during foetal development and epigenetic alterations (at certain loci or globally) [bib_ref] Epigenetic programming of diabetes and obesity: animal models, Seki [/bib_ref] , evidence from studies in human populations has been limited. Unlike the homogeneous genetic backgrounds of animal models and the well-controlled environmental and experimental conditions that can be easily achieved in animal studies, individuals in human studies are genetically heterogeneous and have not been exposed to identical environments throughout their lives. These unavoidable genetic and environmental heterogeneities in human subjects very likely give rise to individual variations in epigenetically affected loci, even when the subjects were exposed to similar nutritional environments for a certain period. Provided that epimutations could occur not only at common loci but at variable loci among subjects, Smirnov-Grubbs' outlier test is effective in evaluating the extent of the accumulation of the latter type of epimutations under certain disease and/or malnutrition conditions and may be applicable to a wide range of epigenetic studies in human populations. FGR is idiopathic in most cases and is generally thought to be caused by foetal, placental, maternal, and/or environmental factors. Therefore, the hypermethylation events observed with significantly high frequencies in placentas from FGR births in this study may also be explained by various factors. Unidentified genetic factors, such as foetal and/or placental chromosomal abnormalities and mutations at certain genes, if they exist, could affect the epigenomes of both the foetus and placenta, regardless of in utero conditions. Maternal and environmental factors deteriorating in utero conditions and contributing to the FGR phenotype may not have been identified in some subjects enrolled in this study. On the other hand, in normal pregnancy cases with inadequate GWG, since the body weights of the babies were within the normal range, the foetuses (and the placentas) were considered to be genetically normal. Under this assumption, promoter hypermethylation observed with higher frequencies in placentas with inadequate GWG than in those with adequate GWG can be regarded as environmentally induced epigenetic alterations.
Multiple independent studies have shown that genetic variants can cause variations in DNA methylation levels, defined as sequence-dependent allele-specific DNA methylation (ASM) [bib_ref] Allele-specific DNA methylation: beyond imprinting, Tycko [/bib_ref]. A recent methylC-Seq study of the mouse genome revealed that sequence-dependent ASMs typically exist as isolated CpG sites in intergenic and intronic regions, but are relatively depleted from proximal promoters [bib_ref] Base-resolution analyses of sequence and parent-of-origin dependent DNA methylation in the mouse..., Xie [/bib_ref]. Moreover, sequence-dependent ASMs are influenced by defined sequences nearby and they appear to have little effect on gene expression. The genomic features of hypomethylated outliers in our study were similar to those of sequence-dependent ASMs. On the other hand, the characteristics of the hypermethylated outliers, being clustered (56%) in the pTSS (77%), were distinct from those of sequence-dependent ASMs. It is generally challenging to distinguish whether differentially methylated regions among genetically heterogeneous human populations are epimutations or sequence-dependent ASMs. However, considering the above-mentioned genomic features of the hypermethylated outliers as well as their enrichment in the promoter regions of transcriptional regulator genes (which will be discussed in detail in the next paragraph), at least a portion of these outliers likely represent genuine epigenetic alterations rather than sequence-dependent changes in DNA methylation.
We initially considered that placental epimutations may have occurred randomly under aberrant in utero environments; our data subsequently revealed that hypermethylated outliers were not found completely randomly in terms of genomic location, but tend to be frequent at the promoters of genes encoding transcription factors. Considering that the promoter regions of genes encoding developmental regulators, such as homeobox proteins and other developmental transcription factors, have been reported to be mostly devoid of sequence-dependent ASMs 33 , the hypermethylated outliers located at the promoter regions of such genes identified in this studymost likely represent epigenetic alterations due to aberrant in utero environments. In a recent genome-wide DNA methylation study using reduced representation bisulphite sequencing (RRBS) in a murine model of FGR, genes hosting differentially methylated regions in the placenta upon maternal calorie restriction are significantly enriched (P < 0.05) with GO terms such as homeobox and transcription factor activity, among others [bib_ref] Intrauterine calorie restriction affects placental DNA methylation and gene expression, Chen [/bib_ref]. Notably, our own annotations for the 131 genes hosting hypermethylated regions in the placenta upon maternal gestational calorie restriction 34 using DAVID revealed that these genes were moderately enriched with genes assigned with the GO Molecular Function term "DNA binding" (17 out of the 131 genes were assigned this term). Therefore, although the statistical method used for detecting differentially methylated regions is different from that in our study, some aspects of this murine study were consistent with our findings demonstrating the enrichment of placental epimutations in transcriptional regulator genes. Our findings also suggested the possibility that certain epigenetic regulatory systems are susceptible to the disruptive effects of aberrant in utero environments. In fact, a careful analysis of the 36 genes assigned with GO terms related to transcriptional regulationrevealed that polycomb group repressive complexes (PRCs) [bib_ref] A new world of Polycombs: unexpected partnerships and emerging functions, Schwartz [/bib_ref] represent a primary candidate of such regulatory mechanisms. We found that seven out of the 36 genes (i.were included in the 653 PRC2 targets in mouse embryonic stem cells, as identified by a ChIP-on-chip analysis [bib_ref] Control of developmental regulators by Polycomb in human embryonic stem cells, Lee [/bib_ref]. Further annotations of the 36 genes using the ChIP-seq data for EZH2 and SUZ12, which are components of PRC2 35 , from a human ES cell line (H1-hESC) produced by the Encyclopedia of DNA Elements (ENCODE) Consortium (http://genome.ucsc.edu/ENCODE/) identified additional eight PRC2 targets. Consistent with our observations, epigenetic variation between twin-twin transfusion syndrome children, wherein twin foetuses occasionally exhibit striking growth differences, is most prominent at the CpG sites within the target regions of PRCs 37 . Furthermore, Wilhelm-Benartzi et al. reported significant associations of placental LINE-1 and AluYb8 methylation levels with birth weight percentile and significant differences in the methylation levels of these repetitive elements upon maternal alcohol or tobacco use during pregnancy [bib_ref] In utero exposures, infant growth, and DNA methylation of repetitive elements and..., Wilhelm-Benartzi [/bib_ref]. Interestingly, the authors also revealed the positive association of increased placental AluYb8 methylation with the average methylation levels of CpG sites in polycomb group target genes. Therefore, evidence from these previous reports and our current findings suggest the possibility that PRCs occasionally fail to recognise their targets with a stochastic nature in the placenta under improper in utero environments, leading to epigenetic switching from PRC marks (H3K27me3) to DNA methylation. Another striking feature of these 36 genes was that 15 (42%) were zinc-finger genes. Zinc-finger genes are often silenced through H3K9me3-mediated gene silencing coupled with promoter DNA methylation in toxicant-induced carcinogenesis, suggesting the existence of an unknown epigenetic mechanism through which many zinc-finger genes are coregulated [bib_ref] Coordinate H3K9 and DNA methylation silencing of ZNFs in toxicant-induced malignant transformation, Severson [/bib_ref]. This hypothetical regulatory mechanism may also be susceptible to the effects of adverse in utero environments.
In addition to the enrichment of GO terms related to transcriptional regulation, the genes hosting hypermethylated outliers were also found to be enriched with the GO term "neuron differentiation" (Supplementary.This seemingly unexpected observation is consistent with those of previous studies. In an array-based expression study that identified 7,519 genes exhibiting differential expression between human placentas sampled during the first and third trimesters, both up-and downregulated genes in the third trimester were found to be enriched with genes involved in human neurogenesis [bib_ref] Differences in gene expression between first and third trimester human placenta: a..., Sitras [/bib_ref]. The authors of the study have suggested that the brain and placenta possibly share common developmental routes. In the above-mentioned RRBS study of the murine model of intrauterine malnutrition 34 , GO terms found to be enriched in genes hosting altered placental DNA methylation upon maternal caloric restriction were shown to contain neuron-related terms [bib_ref] Intrauterine calorie restriction affects placental DNA methylation and gene expression, Chen [/bib_ref]. Additionally, several neural factors, such as BDNF [bib_ref] Differential expression of human placental neurotrophic factors in preterm and term deliveries, Dhobale [/bib_ref] , NGF 41 , and serotonin 42 , have been shown to be secreted from the placenta. Among these factors, BDNF has also been shown to potentiate placental development and play an important role in cytotrophoblast differentiation [bib_ref] Brain-derived neurotrophic factor promotes implantation and subsequent placental development by stimulating trophoblast..., Kawamura [/bib_ref] [bib_ref] Brain-derived neurotrophic factor/tyrosine kinase B signaling regulates human trophoblast growth in an..., Kawamura [/bib_ref]. Furthermore, placental BDNF expression has been reported to be significantly correlated with neonatal birth weight 40 and to be decreased upon maternal malnutrition in rats [bib_ref] Placental BDNF/TrkB signaling system is modulated by fetal growth disturbances in rat..., Mayeur [/bib_ref]. Because of the functional significance of a subset of genes in both the placenta and brain, it is tempting to speculate that the foetuses may have gained epigenetic alteration patterns that are similar to those observed in the placenta in pregnancies with inadequate GWG. Hypermethylation at the promoter regions of genes encoding developmental regulators (PRC2 targets) and neuronal regulators at early embryonic stages would reduce their expression levels when these genes are expressed in a spatio-temporal manner, and such aberrant expression of critical developmental regulators may elevate the risk of developing various diseases, including metabolic and mental disorders, later in life.
In this study, we demonstrated that loci with alterations in the placental DNA methylation under inadequate GWG were not common among subjects but were instead distributed in an individual-specific manner. Furthermore, such epigenetic alterations under the adverse pregnancy condition were found to occur preferentially at the CGI promoters of genes encoding transcriptional factors. Our novel findings support the necessity of large-scale epigenomic studies of placental tissues and samples (e.g., cord blood) from newborns for pregnancies under normal and malnutrition conditions, together with follow-up studies when the newborns reach adulthood in order to elucidate the epigenetic mechanisms underlying developmental programming in humans and their roles in health and disease in later life.
# Materials and methods
Study design. The present study was approved by the Ethics Committee of the National Center of Child Health and Development (NCCHD), Japan and by the Human Study Committee of the Hokkaido University Hospital, Japan. Informed consent was obtained from all subjects. Pregnant Japanese women who did not have pregnancy complications of gestational diabetes, pre-eclampsia, or pregnancy-induced hypertension were enrolled. All enrolled subjects did not smoke or drink alcohol, and did not exhibit hypertension or proteinuria during pregnancy. Subjects (n = 33) were categorised into six categories according to GWG and newborn birth weight: FGR_adequate, FGR_insufficient, FGR_excessive, nor-mal_adequate, normal_insufficient, and normal_excessive, consisting of 5, 5, 4, 9, 5, and 5 placentas, respectively. Prepregnancy BMIs were similar among all groups. The characteristics of each group are shown in . BMI, body weight, GWG, and additional clinical information (maternal complication, gestational week, delivery method, and newborn's gender) for each of the subjects are provided as Supplementary . Although the Institute of Medicine of the United States recommends that pregnant women whose prepregnancy BMI is in the normal range (18.5-24.9) should gain 11.3-15.9 kg during pregnancy, we defined adequate GWG as gaining 7-12 kg in this study in accordance with the recommendations of the Japan Ministry of Health, Labour, and Welfare 46 . This difference is also consistent with the different average BMIs of Japanese and US women (21.14 ± 3.28 47 versus 27.05 ± 0.35, respectively).
Genomic DNA extraction and DNA methylation profiling. Full-term placental samples were obtained from normal caesarean sections or vaginal deliveries. Chorionic villous tissue was obtained from the foetal side of the placenta. Genomic DNA was purified from the tissue using a QIAamp DNA Mini kit (Qiagen, Valencia, CA, USA). Genomic DNA (1.5 μ g) was bisulphite converted using an EpiTect Plus DNA Bisulfite Kit (Qiagen). After determining the concentration of bisulphited DNA, 300 ng of bisulphite DNA from each sample was subjected to Illumina Infinium HumanMethylation450 BeadChip analysis using the manufacturer's standard protocol.
Data processing. To calculate the DNA methylation levels of more than 480,000 CpG sites assayed on the HumanMethylation450 BeadChip (Illumina), the signal intensity data (.idat files), produced by the Illumina iSCAN system, were processed using Illumina GenomeStudio Methylation Analysis Module v1.9.0 with background subtraction and control normalisation options. The methylation levels were calculated as β values ranging from 0 (completely unmethylated) to 1 (completely methylated; β value = intensity of the methylated allele/[intensity of the unmethylated allele + intensity of the methylated allele + 100]). The obtained data have been deposited in NCBI's Gene Expression Omnibus and are accessible through GEO accession number GSE62733. From 485,577 probes on the BeadChip array, the following probes were excluded: the probes on sex chromosomes, the probes for 65 random SNPs (which assay highly-polymorphic SNPs rather than DNA methylation), and the probes whose detection p-value was higher than 0.01 or whose β value was missing in one or more samples. The β values (methylation levels) of the remaining 449,848 probes were corrected by an Empirical Bayes method, ComBat 49 , to remove the array-batch effect, and subjected to statistical tests.
To detect differentially methylated CpG sites between groups, the Illumina Methylation Analyzer (IMA) [bib_ref] IMA: an R package for high-throughput analysis of Illumina's 450K Infinium methylation..., Wang [/bib_ref] was run using the Wilcoxon rank-sum test for inference of differences between categorical groups. The BH procedure was used for multiple testing corrections, and the cut-off for the adjusted p-values was set to 0.05. Smirnov-Grubbs' outlier test with Bonferroni multiple test corrections was performed using the R Package 'outliers' (http://cran.r-project.org/web/packages/outliers/outliers.pdf) and custom R scripts to detect outlying CpG sites, and the cut-off for the corrected p-values was set to 0.1.
The Illumina-provided probe annotation, HumanMethylation450_15017482_v.1.1.csv, was used to sort out the outlying CpG sites whose β value could possibly have been affected by sequence variation within the corresponding probe sequence. This table lists 89,678 probes as SNP-containing in its "probe_SNPs" and "probe_SNPs_10" columns based on the information of NCBI dbSNP Build 131. The refSNP information registered in dbSNP Build 142 was also tested for the same purpose of SNP filtering [fig_ref] Figure 2: Genomic features of 2,521 hyper-and 977 hypomethylated outliers [/fig_ref].
When a single CpG site was assigned to multiple gene symbols or gene features in the Illumina probe annotation, only the lead-off gene symbol or feature was used for gene ontology and genome feature annotations.
Targeted bisulfite sequencing. Bisulfite sequencing analysis was performed as described previously 50 using bisulfite-PCR primers designed by the MethPrimer website 51 . The forward and reverse primer sequences, and the genomic interval (hg19) of the amplicon are: 5′ -GAGAGGTTGGGGTAATTTTAG-3′ , 5′ -AAAAACTTCTAAACTTTTAAACATCC-3′ and chr6:1609671-1610171 (501bp) for the FOXC1 locus; 5′ -GGGGTAGTTGGTTATTATGATAAAGT-3′ , 5′ -ACTCCCCATAACCAAAAACTAAACT-3′ , and chr3:138665547-138665794 (248 bp) for the FOXL2 locus; 5′ -AGTTTTGTGGATTGGGGTTG-3′ ,
[fig] Figure 1: (A) Heatmap visualisation of the β value of methylation outliers detected by Smirnov-Grubbs' outlier tests. The numbers of outliers detected in each placenta are indicated above the heatmap. The colour scale represents the β value from 0 to 1. The left and right panels represent hypermethylated and hypomethylated outliers, respectively (B,D,F). Box plots showing the distribution of the numbers of outliers in each of six placental categories (***, Tukey's multiple comparison test P-value < 0.001). ins, insufficient; ad, adequate; ex, excessive (C,E,F). Bar plots for the numbers of outliers in FGR and normal subjects sorted according to weight gained during pregnancy. Plots for all outliers (B,C), hypermethylated outliers only (D,E), and hypomethylated outliers only (F,G) are shown. Red and green bars represent the numbers of hyper-and hypomethylated outliers, respectively (C,E,F). [/fig]
[fig] Figure 2: Genomic features of 2,521 hyper-and 977 hypomethylated outliers. Distribution of outliers in relation to CGIs and their shores and shelves (A) and to gene feature groups (B). Six gene feature categories, [/fig]
[fig] Figure 3: Examples of promoter-wide hypermethylation at FOXC1 (A), FOXL2 (B), and HOXB7 (C) loci. The β value of the outlier, the mean of the β values of samples other than the outlier, and the Δ β are shown together with Refseq gene and UCSC-defined CGIs using IGV at the left side in each panel. The data range of 0 to 0.5 (or 0 to 1.0) is shown for β and Δ β values. The outlier samples for the three loci are Normal_insufficient_3 (A), Normal_insufficient_1 (B), and FGR_adequate_3 (C). DNA methylation status of these promoter regions were validated by targeted bisulfite sequencing (BS). The black horizontal bar at the bottom in each panel shows the interval of the bisulfite-PCR amplicon. The BS results for the outlier sample and a control (Normal_adequate_7) are shown at the right side in each panel. Open and closed circles represent unmethylated and methylated CpG sites, respectively. Each row of circles corresponds to an individual clone sequenced. The overall methylation rate (%) is shown underneath each panel of the BS results. [/fig]
|
Acupuncture-Point Stimulation for Postoperative Pain Control: A Systematic Review and Meta-Analysis of Randomized Controlled Trials
The purpose of this study was to evaluate the effectiveness of Acupuncture-point stimulation (APS) in postoperative pain control compared with sham/placebo acupuncture or standard treatments (usual care or no treatment). Only randomized controlled trials (RCTs) were included. Meta-analysis results indicated that APS interventions improved VAS scores significantly and also reduced total morphine consumption. No serious APS-related adverse effects (AEs) were reported. There is Level I evidence for the effectiveness of body points plaster therapy and Level II evidence for body points electroacupuncture (EA), body points acupressure, body points APS for abdominal surgery patients, auricular points seed embedding, manual auricular acupuncture, and auricular EA. We obtained Level III evidence for body points APS in patients who underwent cardiac surgery and cesarean section and for auricular-point stimulation in patients who underwent abdominal surgery. There is insufficient evidence to conclude that APS is an effective postoperative pain therapy in surgical patients, although the evidence does support the conclusion that APS can reduce analgesic requirements without AEs. The best level of evidence was not adequate in most subgroups. Some limitations of this study may have affected the results, possibly leading to an overestimation of APS effects.
# Introduction
Nearly 86% of surgery patients experience moderate to severe postoperative pain [bib_ref] Complications occurring in the postanesthesia care unit: a survey, Hines [/bib_ref]. Depending on surgery type, as many as half of these patients go on to experience chronic postoperative pain [bib_ref] Chronic postoperative pain. Epidemiology and psychological risk factors, Dimova [/bib_ref]. Unsatisfactory pain control can limit patients' physical activities, prolong recovery time, and contribute to poor quality of life [bib_ref] The effect of acupuncture on relieving pain after inguinal surgeries, Taghavi [/bib_ref] [bib_ref] Strategies for postoperative pain management, Brown [/bib_ref]. Pain may also increase postoperative complications, such as postoperative morbidity, and may extend the length of hospitalization, increasing health care costs [bib_ref] The prevalence and perception of pain amongst hospital in-patients, Yates [/bib_ref] [bib_ref] Auricular acupressure for managing postoperative pain and knee motion in patients with..., Chang [/bib_ref].
Administration of standard analgesics, which are considered generally safe and effective, remains the primary approach to postoperative pain management [bib_ref] Beyond opioid patient-controlled analgesia: a systematic review of analgesia after major spine..., Sharma [/bib_ref]. However, systemic analgesic administration can induce some adverse effects (AEs), such as nausea, vomiting, depressive symptoms, pruritus, urinary retention, gastrointestinal motility, and ileus [bib_ref] Patient-controlled analgesia in the management of postoperative pain, Momeni [/bib_ref] [bib_ref] Auriculotherapy for pain management: a systematic review and meta-analysis of randomized controlled..., Asher [/bib_ref]. AEs can impair physical and psychological wellbeing and, more seriously, may result in significant morbidity or even mortality [bib_ref] Patient-controlled analgesia in the management of postoperative pain, Momeni [/bib_ref] [bib_ref] Auriculotherapy for pain management: a systematic review and meta-analysis of randomized controlled..., Asher [/bib_ref] [bib_ref] Effectiveness of acute postoperative pain management: I. Evidence from published data, Dolin [/bib_ref]. To achieve better postoperative pain relief and reduce the requirement for analgesic medication, various nonpharmacological approaches, including educational intervention, relaxation, and acupuncture-point stimulation (APS), have been employed. APS has been lauded as a promising alternative method for achieving postoperative pain relief [bib_ref] Acupuncture and related interventions for smoking cessation, White [/bib_ref].
APS is a widely used component of traditional Chinese medicine (TCM) together with full-body and auricular approaches . In addition to the most popular methods of manual acupuncture and acupressure, APS can be achieved using modalities such as electrical acupuncture (EA) or laser stimulation and acupoint massage [bib_ref] Acupuncture and related interventions for smoking cessation, White [/bib_ref]. According to TCM philosophy, the stimulation of target acupoints along meridians produces positive effects by rebalancing qi circulation in the body [bib_ref] The meridian system and mechanism of acupuncture: a comparative review, Chang [/bib_ref]. However, the existence of the qi meridian system, as described in TCM, has never been demonstrated scientifically [bib_ref] The meridian system and mechanism of acupuncture: a comparative review, Chang [/bib_ref]. Nevertheless, the management of various forms of pain remains a key purported benefit of APS [bib_ref] Analgesic effect of electroacupuncture in postthoracotomy pain: a prospective randomized Trial, Wong [/bib_ref] [bib_ref] Specifying the nonspecific components of acupuncture analgesia, Vase [/bib_ref].
Many animal experiments and clinical studies have examined the therapeutic effects of APS [bib_ref] Advancing acupuncture research, Ahn [/bib_ref]. Early studies showed that APS provided postoperative pain relief in comparison with control groups [bib_ref] The treatment of postoperative pain with the use of semipermanent auricular needles, Mastroianni [/bib_ref] [bib_ref] Comparative study of the analgesic effect of transcutaneous nerve stimulation, Martelete [/bib_ref]. Recently, several small trials [bib_ref] The effect of acupuncture on relieving pain after inguinal surgeries, Taghavi [/bib_ref] [bib_ref] A randomized controlled trial of auricular transcutaneous electrical nerve stimulation for managing..., Tsang [/bib_ref] demonstrated that APS can relieve pain and reduce analgesic requirements associated with hysterectomy and inguinal surgery. However, Sakurai et al. [bib_ref] Minute sphere acupressure does not reduce postoperative pain or morphine consumption, Sakurai [/bib_ref] failed to identify any significant change in pain intensity or morphine requirement in surgical patients undergoing acupressure. A prior systematic review found that acupuncture and related techniques aided postoperative pain control, but the quality of evidence was low due to the quality and quantity of included trials, and no subgroup analysis according to acupuncture type was performed [bib_ref] Acupuncture and related techniques for postoperative pain: a systematic review of randomized..., Sun [/bib_ref]. The results of another systematic review conducted by Usichenko et al. [bib_ref] Auricular acupuncture for postoperative pain control: a systematic review of randomised clinical..., Usichenko [/bib_ref] suggested that auricular acupuncture was a promising method of postoperative pain reduction, but the heterogeneity of primary studies precluded data synthesis and the evidence was insufficient to draw a definitive conclusion about the treatment's effectiveness. Following the 2008 publication of these reviews, several clinical trials were conducted to evaluate the efficacy of APS for postoperative pain management, generating new evidence on this topic [bib_ref] Capsicum plaster at the hegu point reduces postoperative analgesic requirement after orthognathic..., Kim [/bib_ref] [bib_ref] Auricular acupressure for analgesia in perioperative period of total knee arthroplasty, He [/bib_ref] [bib_ref] Integrative acupoint stimulation to alleviate postoperative pain and morphine-related side effects: a..., Chung [/bib_ref].
The present systematic review and meta-analysis was conducted to evaluate the effectiveness of APS for pain control following surgical procedures. Therefore, in this study, current evidence generated by randomized controlled trials (RCTs) on the use of APS interventions for postoperative pain management was reviewed and analyzed. Data from patients receiving APS were compared with those from control groups receiving sham/placebo acupuncture, usual care, or no treatment. Compared with the previous literature, this systematic review and meta-analysis incorporates new evidence not previously synthesized and distinguishes between multiple types of APS for postoperative pain control. [fig_ref] Figure 1: Flow chart of the study selection process [/fig_ref] We also used PubMed's "related articles" function to identify additional potentially relevant studies. The electronic search had no language restriction. In the case that there were multiple publications from the same RCT, overlapping results were extracted from one publication.
# Methods
## Study selection. as summarized in
Two reviewers assessed all potentially relevant articles independently. Disagreements regarding study selection were resolved by discussion, with strict adherence to the inclusion criteria. Studies were selected for inclusion based on the following criteria: (1) RCT; (2) adult (age ≥ 18 years) participants with pain following any surgical procedure; (3) APS intervention (including full-body or auricular manual acupuncture or EA, acupressure, seed embedding, and plaster therapy) conducted by an acupuncturist, TCM practitioner, or other health care providers with qualification and/or training in acupuncture therapy; (4) control group receiving standard treatment (e.g., active pain control approach normally provided to surgical patients, including analgesia medication, nursing guidance, and other usual cares), sham/placebo APS (faked APS intervention), or no treatment (provision of usual postoperative care not involving active analgesic interventions); (5) primary outcome of pain intensity, measured by a valid self-reported instrument such as a visual analog scale (VAS), numerical rating scale (NRS), or verbal reporting; and (6) secondary outcomes of analgesic consumption and APS-related AEs (i.e., any adverse events resulting from APS intervention, minor (e.g., needling site pain), intermediate (e.g., bleeding and hematoma), or serious (pneumothorax and cardiac tamponade)).
In the study selection process, acupuncture was defined as the stimulation of specific acupuncture points along the skin of the body by using thin needles, with or without the application of heat, pressure, or laser light to these same points [bib_ref] Acupuncture and related interventions for smoking cessation, White [/bib_ref]. EA is similar to acupuncture but involves the use of devices (e.g., a wristwatch-like device and surface electrodes attached to a transcutaneous electrical nerve stimulation device) on acupoint [bib_ref] Analgesic effect of electroacupuncture in postthoracotomy pain: a prospective randomized Trial, Wong [/bib_ref]. Seed embedding was defined as an auricular acupressure process involving the embedding of magnetic beads or other seeds within skin-colored adhesive tape, which is placed on the auricular acupoints and retained in situ for several days [bib_ref] Auricular acupressure for managing postoperative pain and knee motion in patients with..., Chang [/bib_ref]. In this systematic review, plaster therapy mainly referred to the use of capsicum plaster as an alternative to acupuncture [bib_ref] Capsicum plaster at the hegu point reduces postoperative analgesic requirement after orthognathic..., Kim [/bib_ref].
## Quality assessment.
Two reviewers conducted independent assessments of the methodological quality and risk of bias of each RCT using Cochrane Collaboration's risk of bias tool. This tool provides for the assessment of seven domains: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and "other issues." Items were scored as showing low, high, or unclear risk of bias. All disagreements on scoring were resolved by discussion. When a sufficient number of studies were available and a meaningful assessment of publication bias could be carried out, a funnel plot was constructed.
Adequate allocation concealment and blinding of outcome assessors were designated as key domains for this assessment, where key domains are not only more likely to influence bias magnitude and direction but also more likely to impact study results. Domain-based evaluation was employed as described in the Cochrane Handbook for Systematic Reviews of Interventions 5.0.2 (updated September 2009). The overall risk of bias was categorized as follows. An overall low risk of bias (plausible bias unlikely to alter the results) was identified when all key domains were assessed as having a low risk of bias. An overall unclear risk of bias (plausible bias that raises some doubt about the results) was identified when one or more key domains were assessed as having an unclear risk of bias. An overall high risk of bias (plausible bias that seriously weakens confidence in the results) was identified when one or more key domains were assessed as having a high risk of bias. Small studies have been shown to overestimate treatment effects and to be at increased risk of bias, allowing critical criteria such as blinding to be compromised. Studies were considered to be at low risk of bias if they had at least 200 participants, at unknown risk if they had 50 to 200 participants, and at high risk if they had fewer than 50 participants.
## Data extraction and management.
Two reviewers independently extracted the following data from included studies using a predefined form: authors, study design, setting, population and participant demographics, intervention characteristics (e.g., acupuncture type, acupoints used, and treatment duration), comparators, outcome measures and instruments, follow-up, and some numeric data (mainly the results of pain intensity and analgesic consumption). We contacted RCT authors by email to obtain data necessary for effect size estimation when such data were missing from publications (e.g., due to aggregated data reporting). When authors did not reply, outcome data presented only in figures and/or graphs were extracted when possible; these data were included in the analysis only when the two reviewers independently obtained the same results.
When a study reported multiple group comparisons (e.g., high electrical stimulation versus low electrical stimulation or usual care and preoperative acupuncture versus postoperative acupuncture or usual care), only data from the treatment group that received the more intensive and comprehensive postoperative intervention were included in the analysis. These data were compared with those from the control group.
# Subgroup analysis.
When data were sufficient, subgroup analyses of different types of APS, surgery, and control groups were conducted. Analyses of APS type compared the use of acupoints on the body (EA, manual acupuncture, acupressure, and plaster therapy) and/or auricular points (EA, manual acupuncture, and seed embedding). A subgroup analysis of EA studies compared the use of different devices (e.g., a wristwatch-like device and surface electrodes attached to a transcutaneous electrical nerve stimulation device). Analyses of control group types compared APS with standard treatment (usual care and no treatment) or placebo/sham therapies. On the basis of whether body acupoints, auricular 4 Evidence-Based Complementary and Alternative Medicine In the presence of significant heterogeneity ( 2 test, < 0.1), random-effects model was used. Otherwise, fixed-effects model was applied. Descriptive analysis was used when data could not be converted or pooled. Potential sources of heterogeneity in the outcomes examined are differences in the tool used to measure pain, population differences (e.g., surgery type, age, and sex), and differences in the comparator used (e.g., sham/placebo acupuncture, usual care, and no treatment). We assessed heterogeneity using the 2 statistic, which describes the percentage of total variation across trials (low, 0-40%; moderate, 30-60%; substantial, 50-90%; considerable, 75-100%; Chapter 9: Analysing Data and Undertaking Meta-Analyses; The Cochrane Collaboration 2011, available from http://www .cochrane-handbook.org/). To assess which RCTs affected the overall results, sensitivity analyses were performed for the entire sample and subgroups with significant heterogeneity. If heterogeneity was considerable, even with the random-effects model, best evidence synthesis was also used. The evidence was synthesized based on each subgroup. We employed a qualitative modified approach to grading of evidence, as summarized in [fig_ref] Table 1: Qualitative modified approach to grading of evidence [/fig_ref]. [fig_ref] Table 2: Characteristics of RCTs examining body acupoint stimulation included in the meta-analysis [/fig_ref]. Nine publications were in Chinese, and the remaining 50 publications were in English. The included studies were conducted in mainland China, Hong Kong, Taiwan, the United States, Germany, Austria, South Korea, Japan, Iran, the United Kingdom, Brazil, Sweden, Singapore, Italy, and Turkey. Sixteen of these RCTs had three arms and three trials had four arms.
# Results
## Characteristics of included
RCTs included in the analysis involved a total of 4,578 randomized patients, 4,402 of whom completed the respective studies (APS groups, = 2, 097; control groups, = 2, 305; 96.16% completion rate). The average sample size was 73 (range, . Standardized anesthetic and postoperative analgesia regimens were used in all studies. Follow-up duration ranged widely from 7 days [bib_ref] Evaluation of acupuncture for pain control after oral surgery: a placebo-controlled trial, Lao [/bib_ref] to 4 months [bib_ref] Acupuncture and arthroscopic acromioplasty, Gilbertson [/bib_ref]. Twenty-one RCTs stated that an intention-to-treat analysis was used.
Five types of APS were used: low-and/or high-intensity EA, manual needle acupuncture, seed embedding, acupressure, and plaster therapy. Preoperative and postoperative APS were used in two RCTs, and sham/placebo control was used in 36 studies. Chinese herbs were used as a control in a single study.
Acupuncture points on the body and/or auricular points were stimulated. Commonly used body points included Hegu (LI4), Sanyinjiao (SP6), Zusanli (ST36), and Nei guan (P6); commonly used auricular points included Shen Men (TF4), Stomach (CO4), and Lung (CO14).
## Methodological quality and risk of bias of included
Trials. The methodological quality of included studies is characterized in [fig_ref] Figure 2: Methodological quality of included studies [/fig_ref]. Eighteen (30.5%) publications specifically stated that the outcome assessor was blinded and 23 (38.98%) studies used adequate allocation concealment together with full methodological description. Twelve (20.33%) studies were rated highly in both of these domains and were deemed to be at low risk of overall bias. Forty-two studies (71.19%) were deemed to be at unclear risk of overall bias. Finally, five studies (8.5%) were deemed to be at high overall risk of bias. Allocation concealment was not reported or was described poorly in 24 (40.67%) studies. According to the number of participants, no studies were considered to be at low risk of bias (≥200 participants), 42 studies (71.19%) were at unknown risk of bias , and 17 studies (28.81%) were at high risk of bias Evidence-Based Complementary and Alternative Medicine 5 14 Evidence-Based Complementary and Alternative Medicine (<50 participants). Visual inspection of funnel plots revealed some substantial asymmetry in comparisons [fig_ref] Figure 4: Funnel plot. [/fig_ref].
## Meta-analysis and descriptive analysis of outcomes
## Postoperative pain.
The results of the meta-analysis are reported in [fig_ref] Table 5: Summary of meta-analysis results [/fig_ref] [bib_ref] Randomized sham-controlled trial of acupuncture for postoperative pain control after stapled haemorrhoidopexy, Langenbach [/bib_ref] [bib_ref] Acupuncture for pain and dysfunction after neck dissection: results of a randomized..., Pfister [/bib_ref] that used NRS scores reported a significant difference between groups, but these data could not be included in the meta-analysis due to clinical and statistical heterogeneity. One of two RCTs [bib_ref] Pain reduction of acupoint electrical stimulation for patients with spinal surgery: a..., Yeh [/bib_ref] [bib_ref] Randomized controlled trial of a special acupuncture technique for pain after thoracotomy, Deng [/bib_ref] that used BPI scores reported a significant difference between the intervention and control groups, but these data could not be included in the meta-analysis due to different data modes. Three trials [bib_ref] Preoperative intradermal acupuncture reduces postoperative pain, nausea and vomiting, analgesic requirement, and..., Kotani [/bib_ref] [bib_ref] Evaluation of acupuncture for pain control after oral surgery: a placebo-controlled trial, Lao [/bib_ref] using a four-point scale reported that the intervention reduced pain intensity, but one of these studies [bib_ref] Preoperative intradermal acupuncture reduces postoperative pain, nausea and vomiting, analgesic requirement, and..., Kotani [/bib_ref] reported that total or partial pain relief did not differ significantly between the groups. Two trials [bib_ref] Evaluation of acupuncture for pain control after oral surgery: a placebo-controlled trial, Lao [/bib_ref] reported that the APS intervention increased the duration of postoperative painfree status compared with that of the control groups. Thirty-eight RCTs used body points for stimulation. Subgroup analyses according to control treatment and meta-analysis of 20 RCTs indicated that APS interventions improved VAS scores significantly in comparison with standard treatment and sham/placebo control (both < 0.00001; [fig_ref] Table 5: Summary of meta-analysis results [/fig_ref]. Similarly, pooled results from 24 trials showed that body APS significantly improved VAS scores in comparison with all control groups ( < 0.00001), and subgroup analyses revealed similar improvement compared with standard treatment ( < 0.00001) and sham/placebo control ( < 0.0001; [fig_ref] Table 5: Summary of meta-analysis results [/fig_ref]. The evidence for body points APS reducing postoperative pain intensity in surgery patients was determined to be of Level I quality based on six overall high quality RCTs [bib_ref] Analgesic effect of electroacupuncture in postthoracotomy pain: a prospective randomized Trial, Wong [/bib_ref] [bib_ref] Capsicum plaster at the hegu point reduces postoperative analgesic requirement after orthognathic..., Kim [/bib_ref] [bib_ref] The analgesic effects of capsicum plaster at the Zusanli point after abdominal..., Kim [/bib_ref] [bib_ref] Prevention of postoperative sore throat using capsicum plaster applied at the Korean..., Park [/bib_ref] [bib_ref] Objectifying acupuncture effects by lung function and numeric rating scale in patients..., Maimer [/bib_ref] [bib_ref] Preoperative intradermal acupuncture reduces postoperative pain, nausea and vomiting, analgesic requirement, and..., Kotani [/bib_ref]. A meta-analysis of pooled results and subgroup analyses of body EA, as well as invasive and noninvasive forms of this treatment, yielded similar results [fig_ref] Table 5: Summary of meta-analysis results [/fig_ref]. The evidence for body points EA reducing postoperative pain intensity in surgery patients was determined to be of Level II quality based on one overall high quality RCT [bib_ref] Analgesic effect of electroacupuncture in postthoracotomy pain: a prospective randomized Trial, Wong [/bib_ref]. High-frequency EA was found to be more effective than low-frequency EA [bib_ref] The effect of high and low frequency electroacupuncture in pain after lower..., Lin [/bib_ref] [bib_ref] Effect of the intensity of transcutaneous acupoint electrical stimulation on the postoperative..., Wang [/bib_ref]. Pooled results from three RCTs examining acupressure [bib_ref] Effect of acupressure on nausea, vomiting, anxiety and pain among post-cesarean section..., Chen [/bib_ref] [bib_ref] Pressure on acupoints decreases postoperative pain, Felhendler [/bib_ref] [bib_ref] Effect of acupressure of Ex-Le7 point on pain, nausea and vomiting after..., Adib-Hajbaghery [/bib_ref] showed a significant difference in VAS scores between intervention and control groups ( = 0.01; [fig_ref] Table 5: Summary of meta-analysis results [/fig_ref] , although a fourth study [bib_ref] Minute sphere acupressure does not reduce postoperative pain or morphine consumption, Sakurai [/bib_ref] not included in the meta-analysis showed no such difference. The evidence for body points acupressure reducing postoperative pain intensity in surgery patients was determined to be of Level II quality based on three moderate quality RCTs [bib_ref] Minute sphere acupressure does not reduce postoperative pain or morphine consumption, Sakurai [/bib_ref] [bib_ref] Effect of acupressure on nausea, vomiting, anxiety and pain among post-cesarean section..., Chen [/bib_ref] [bib_ref] Pressure on acupoints decreases postoperative pain, Felhendler [/bib_ref] and one low quality RCT [bib_ref] Effect of acupressure of Ex-Le7 point on pain, nausea and vomiting after..., Adib-Hajbaghery [/bib_ref]. Synthesis of data from two RCTs [bib_ref] Capsicum plaster at the hegu point reduces postoperative analgesic requirement after orthognathic..., Kim [/bib_ref] [bib_ref] The analgesic effects of capsicum plaster at the Zusanli point after abdominal..., Kim [/bib_ref] examining plaster therapy showed a significant reduction in pain intensity compared with standard treatment ( < 0.00001) and sham controls ( < 0.0001; [fig_ref] Table 5: Summary of meta-analysis results [/fig_ref] , and one other study [bib_ref] Prevention of postoperative sore throat using capsicum plaster applied at the Korean..., Park [/bib_ref] examining this treatment obtained similar results. The evidence for body points plaster therapy reducing postoperative pain intensity in surgery patients was determined to be of Level I quality based on three overall high quality RCTs [bib_ref] Capsicum plaster at the hegu point reduces postoperative analgesic requirement after orthognathic..., Kim [/bib_ref] [bib_ref] The analgesic effects of capsicum plaster at the Zusanli point after abdominal..., Kim [/bib_ref] [bib_ref] Prevention of postoperative sore throat using capsicum plaster applied at the Korean..., Park [/bib_ref].
In contrast, meta-analysis including three studies [bib_ref] Effects of acupuncture on post-cesarean section pain, Wu [/bib_ref] [bib_ref] Additional use of acupuncture to NSAID effectively reduces post-tonsillectomy pain, Sertel [/bib_ref] [bib_ref] Effects of acupuncture and sham acupuncture in addition to physiotherapy in patients..., Tsang [/bib_ref] revealed no significant effect of manual acupuncture on VAS score. Four [bib_ref] Effects of acupuncture on post-cesarean section pain, Wu [/bib_ref] [bib_ref] Effects of acupuncture and sham acupuncture in addition to physiotherapy in patients..., Tsang [/bib_ref] [bib_ref] The effect of pre-emptive acupuncture treatment on analgesic requirements after day-case knee..., Gupta [/bib_ref] [bib_ref] Evaluation of acupuncture for pain control after oral surgery: a placebo-controlled trial, Lao [/bib_ref] of 10 RCTs examining manual acupuncture reported no difference in pain score between the intervention and control groups, whereas the remaining six studies found that this treatment reduced postoperative pain intensity ( < 0.05).
Twelve RCTs used body point stimulation for patients with abdominal surgery. The pooled results from eight trials [bib_ref] Effect of transcutaneous electrical stimulation of Zusanli (ST 36) and Liangqiu (ST..., Yin [/bib_ref] [bib_ref] Effect of low frequency current acupoint stimulation on postoperative analgesia following gastrectomy..., Deng [/bib_ref] [bib_ref] Effect of previous analgesia of scalp acupuncture on post-operative epidural morphine analgesia..., He [/bib_ref] [bib_ref] The effect of high and low frequency electroacupuncture in pain after lower..., Lin [/bib_ref] [bib_ref] Effects of electroacupuncture on intraoperative and postoperative analgesic requirement, Sim [/bib_ref] [bib_ref] Effect of the intensity of transcutaneous acupoint electrical stimulation on the postoperative..., Wang [/bib_ref] [bib_ref] The analgesic effects of capsicum plaster at the Zusanli point after abdominal..., Kim [/bib_ref] [bib_ref] Effect of acupressure of Ex-Le7 point on pain, nausea and vomiting after..., Adib-Hajbaghery [/bib_ref] showed that body APS significantly improved VAS scores in these patients ( = 0.0006). The evidence for body points APS reducing postoperative pain intensity in patients who had undergone abdominal surgery was determined to be of Level II quality based on one overall high quality RCT [bib_ref] The analgesic effects of capsicum plaster at the Zusanli point after abdominal..., Kim [/bib_ref]. Five [bib_ref] Transcutaneous electrical nerve stimulation on acupoints reduces fentanyl requirement for postoperative pain..., Lan [/bib_ref] [bib_ref] The application of electroacupuncture to postoperative rehabilitation of total knee replacement, Chen [/bib_ref] [bib_ref] Pressure on acupoints decreases postoperative pain, Felhendler [/bib_ref] [bib_ref] Effects of acupuncture and sham acupuncture in addition to physiotherapy in patients..., Tsang [/bib_ref] [bib_ref] The effect of pre-emptive acupuncture treatment on analgesic requirements after day-case knee..., Gupta [/bib_ref] RCTs used body point stimulation for patients with knee surgery. Pooled results from four trials [bib_ref] Transcutaneous electrical nerve stimulation on acupoints reduces fentanyl requirement for postoperative pain..., Lan [/bib_ref] [bib_ref] The application of electroacupuncture to postoperative rehabilitation of total knee replacement, Chen [/bib_ref] [bib_ref] Pressure on acupoints decreases postoperative pain, Felhendler [/bib_ref] [bib_ref] Effects of acupuncture and sham acupuncture in addition to physiotherapy in patients..., Tsang [/bib_ref] showed that body APS did not significantly improve VAS scores for these patients ( = 0.16). Each of two RCTs used body point stimulation for patients with oral surgery [bib_ref] Evaluation of acupuncture for pain control after oral surgery: a placebo-controlled trial, Lao [/bib_ref] , cardiac surgery [bib_ref] Randomised, controlled study of preoperative eletroacupuncture for postoperative pain control after cardiac..., Coura [/bib_ref] [bib_ref] Postoperative pain and respiratory function in patients treated with electroacupuncture following coronary..., Colak [/bib_ref] , hemorrhoid operation [bib_ref] Randomized sham-controlled trial of acupuncture for postoperative pain control after stapled haemorrhoidopexy, Langenbach [/bib_ref] , or cesarean section [bib_ref] Effect of acupressure on nausea, vomiting, anxiety and pain among post-cesarean section..., Chen [/bib_ref] [bib_ref] Effects of acupuncture on post-cesarean section pain, Wu [/bib_ref]. Pooled results from two trials showed that body APS significantly improved VAS scores for patients undergoing cardiac surgery [bib_ref] Randomised, controlled study of preoperative eletroacupuncture for postoperative pain control after cardiac..., Coura [/bib_ref] [bib_ref] Postoperative pain and respiratory function in patients treated with electroacupuncture following coronary..., Colak [/bib_ref] ( = 0.002) or cesarean section [bib_ref] Effect of acupressure on nausea, vomiting, anxiety and pain among post-cesarean section..., Chen [/bib_ref] [bib_ref] Effects of acupuncture on post-cesarean section pain, Wu [/bib_ref] ( < 0.00001). The evidence for body points APS reducing pain intensity in patients who underwent cardiac surgery and cesarean section surgery was determined to be of Level III quality based on two moderate quality RCTs [bib_ref] Randomised, controlled study of preoperative eletroacupuncture for postoperative pain control after cardiac..., Coura [/bib_ref] [bib_ref] Postoperative pain and respiratory function in patients treated with electroacupuncture following coronary..., Colak [/bib_ref]. Other studies could not be included in the meta-analyses due to insufficient data and the different types of surgery.
Fourteen [bib_ref] Auricular acupressure for managing postoperative pain and knee motion in patients with..., Chang [/bib_ref] [bib_ref] A randomized controlled trial of auricular transcutaneous electrical nerve stimulation for managing..., Tsang [/bib_ref] [bib_ref] Auricular acupressure for analgesia in perioperative period of total knee arthroplasty, He [/bib_ref] [bib_ref] Influence of auricular point sticking on incidence of nausea and vomiting and..., Zhang [/bib_ref] [bib_ref] Acupoint electrical stimulation reduces acute postoperative pain in surgical patients with patient-controlled..., Yeh [/bib_ref] [bib_ref] Electrical punctual stimulation (P-STIM) with ear acupuncture following tonsillectomy, a randomised, controlled..., Kager [/bib_ref] [bib_ref] Perioperative auricular electroacupuncture has no effect on pain and analgesic consumption after..., Michalek-Sauberer [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after ambulatory knee surgery: a randomized trial, Usichenko [/bib_ref] [bib_ref] Auricular electro-acupuncture as an additional perioperative analgesic method during oocyte aspiration in..., Sator-Katzenschlager [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after ambulatory knee arthroscopy-a pilot study, Usichenko [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after total hip arthroplasty-a randomized controlled study, Usichenko [/bib_ref] [bib_ref] Relieving effects of Chinese herbs, ear-acupuncture and epidural morphing on postoperative pain..., Li [/bib_ref] RCTs used auricular points for stimulation. Data synthesis from 12 studies showed significantly lower VAS scores in intervention groups than in all types of control group ( = 0.001), and similar results were obtained in comparison with standard treatment ( = 0.04) and sham/placebo control ( = 0.02) groups [fig_ref] Table 5: Summary of meta-analysis results [/fig_ref]. The evidence for auricular points APS reducing postoperative pain intensity was determined to be of Level I quality based on six overall high quality RCTs [bib_ref] Auricular acupressure for managing postoperative pain and knee motion in patients with..., Chang [/bib_ref] [bib_ref] Auricular acupressure for analgesia in perioperative period of total knee arthroplasty, He [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after ambulatory knee surgery: a randomized trial, Usichenko [/bib_ref] [bib_ref] Auricular electro-acupuncture as an additional perioperative analgesic method during oocyte aspiration in..., Sator-Katzenschlager [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after total hip arthroplasty-a randomized controlled study, Usichenko [/bib_ref]. Metaanalysis of data from five studies [bib_ref] Auricular acupressure for managing postoperative pain and knee motion in patients with..., Chang [/bib_ref] [bib_ref] Auricular acupressure for analgesia in perioperative period of total knee arthroplasty, He [/bib_ref] [bib_ref] Influence of auricular point sticking on incidence of nausea and vomiting and..., Zhang [/bib_ref] [bib_ref] Acupoint electrical stimulation reduces acute postoperative pain in surgical patients with patient-controlled..., Yeh [/bib_ref] examining seed embedding also showed a significant effect on VAS score in comparison with all control groups ( = 0.02; [fig_ref] Table 5: Summary of meta-analysis results [/fig_ref]. The evidence for auricular points seed embedding reducing postoperative pain intensity was determined to be of Level II quality based on two overall high quality RCTs [bib_ref] Auricular acupressure for managing postoperative pain and knee motion in patients with..., Chang [/bib_ref] [bib_ref] Auricular acupressure for analgesia in perioperative period of total knee arthroplasty, He [/bib_ref]. Two studies [bib_ref] Auricular acupressure for managing postoperative pain and knee motion in patients with..., Chang [/bib_ref] [bib_ref] Acupoint electrical stimulation reduces acute postoperative pain in surgical patients with patient-controlled..., Yeh [/bib_ref] of this auricular APS technique found a gradual reduction in pain, but no significant difference, according to VAS and Short-Form McGill Pain Questionnaire scores. One studyof manual auricular acupuncture data reported a significant difference in VAS score, and another study [bib_ref] Relieving effects of Chinese herbs, ear-acupuncture and epidural morphing on postoperative pain..., Li [/bib_ref] showed a positive trend toward less pain in the intervention group, but meta-analysis of data from four studies [bib_ref] Auricular acupuncture for pain relief after ambulatory knee surgery: a randomized trial, Usichenko [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after ambulatory knee arthroscopy-a pilot study, Usichenko [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after total hip arthroplasty-a randomized controlled study, Usichenko [/bib_ref] showed that this auricular APS type was not associated with significant pain reduction. The evidence for manual auricular acupuncture reducing postoperative pain intensity in surgery patients was determined to be of Level II quality based on two overall high quality RCTs [bib_ref] Auricular acupuncture for pain relief after ambulatory knee surgery: a randomized trial, Usichenko [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after total hip arthroplasty-a randomized controlled study, Usichenko [/bib_ref]. Metaanalysis of auricular EA data from three studies [bib_ref] A randomized controlled trial of auricular transcutaneous electrical nerve stimulation for managing..., Tsang [/bib_ref] [bib_ref] Electrical punctual stimulation (P-STIM) with ear acupuncture following tonsillectomy, a randomised, controlled..., Kager [/bib_ref] [bib_ref] Auricular electro-acupuncture as an additional perioperative analgesic method during oocyte aspiration in..., Sator-Katzenschlager [/bib_ref] showed a significant reduction in VAS scores (including those reflecting pain at rest and on huffing and coughing; < 0.0001), although two of the four RCTs examining this treatment found no significant difference due to low pain intensity in intervention groups. The evidence for auricular EA reducing postoperative pain intensity in surgery patients was determined to be of Level II quality based on one overall high quality RCT [bib_ref] Auricular electro-acupuncture as an additional perioperative analgesic method during oocyte aspiration in..., Sator-Katzenschlager [/bib_ref].
Five [bib_ref] Auricular acupressure for managing postoperative pain and knee motion in patients with..., Chang [/bib_ref] [bib_ref] Auricular acupressure for analgesia in perioperative period of total knee arthroplasty, He [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after ambulatory knee surgery: a randomized trial, Usichenko [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after ambulatory knee arthroscopy-a pilot study, Usichenko [/bib_ref] RCTs used auricular point stimulation for patients with knee surgery. Pooled results from five trials [bib_ref] Auricular acupressure for managing postoperative pain and knee motion in patients with..., Chang [/bib_ref] [bib_ref] Auricular acupressure for analgesia in perioperative period of total knee arthroplasty, He [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after ambulatory knee surgery: a randomized trial, Usichenko [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after ambulatory knee arthroscopy-a pilot study, Usichenko [/bib_ref] showed that auricular point APS did not significantly improve VAS scores for these patients ( = 0.20). Two [bib_ref] A randomized controlled trial of auricular transcutaneous electrical nerve stimulation for managing..., Tsang [/bib_ref] [bib_ref] Influence of auricular point sticking on incidence of nausea and vomiting and..., Zhang [/bib_ref] RCTs used auricular point stimulation for patients with abdominal surgery. Pooled results from both trials [bib_ref] A randomized controlled trial of auricular transcutaneous electrical nerve stimulation for managing..., Tsang [/bib_ref] [bib_ref] Influence of auricular point sticking on incidence of nausea and vomiting and..., Zhang [/bib_ref] showed that auricular point APS significantly improved VAS scores for these patients ( = 0.01). The evidence for auricular point stimulation reducing postoperative pain intensity in abdominal surgery patients was determined to be of Level III quality based on two moderate quality RCTs [bib_ref] A randomized controlled trial of auricular transcutaneous electrical nerve stimulation for managing..., Tsang [/bib_ref] [bib_ref] Influence of auricular point sticking on incidence of nausea and vomiting and..., Zhang [/bib_ref].
Seven RCTs [bib_ref] Integrative acupoint stimulation to alleviate postoperative pain and morphine-related side effects: a..., Chung [/bib_ref] [bib_ref] Effects of electroacupuncture on local anaesthesia for inguinal hernia repair: a randomised..., Dias [/bib_ref] [bib_ref] Acupuncture for pain and dysfunction after neck dissection: results of a randomized..., Pfister [/bib_ref] [bib_ref] Randomized controlled trial of a special acupuncture technique for pain after thoracotomy, Deng [/bib_ref] [bib_ref] Symptom management with massage and acupuncture in postoperative cancer patients: a randomized..., Mehling [/bib_ref] [bib_ref] Electro-acupuncture efficacy on pain control after mandibular third molar surgery, Tavares [/bib_ref] [bib_ref] Increased postoperative pain and consumption of analgesics following acupuncture, Ekblom [/bib_ref] used integrative APS (combined stimulation of body and auricular points) and evaluated postoperative pain relief using VAS ( = 4) and NRS ( = 3) scores. This meta-analysis showed a significant effect of integrative APS on pain intensity based on pooled VAS and NRS scores ( = 0.03; [fig_ref] Table 5: Summary of meta-analysis results [/fig_ref] [bib_ref] Effects of electroacupuncture on local anaesthesia for inguinal hernia repair: a randomised..., Dias [/bib_ref] [bib_ref] Acupuncture for pain and dysfunction after neck dissection: results of a randomized..., Pfister [/bib_ref] [bib_ref] Randomized controlled trial of a special acupuncture technique for pain after thoracotomy, Deng [/bib_ref] [bib_ref] Symptom management with massage and acupuncture in postoperative cancer patients: a randomized..., Mehling [/bib_ref] [bib_ref] Electro-acupuncture efficacy on pain control after mandibular third molar surgery, Tavares [/bib_ref] [bib_ref] Increased postoperative pain and consumption of analgesics following acupuncture, Ekblom [/bib_ref]. The evidence for integrative APS reducing postoperative pain in surgery patients was determined to be of Level II quality based on five moderate quality [bib_ref] Integrative acupoint stimulation to alleviate postoperative pain and morphine-related side effects: a..., Chung [/bib_ref] [bib_ref] Effects of electroacupuncture on local anaesthesia for inguinal hernia repair: a randomised..., Dias [/bib_ref] [bib_ref] Symptom management with massage and acupuncture in postoperative cancer patients: a randomized..., Mehling [/bib_ref] [bib_ref] Electro-acupuncture efficacy on pain control after mandibular third molar surgery, Tavares [/bib_ref] [bib_ref] Increased postoperative pain and consumption of analgesics following acupuncture, Ekblom [/bib_ref] and two low quality [bib_ref] Acupuncture for pain and dysfunction after neck dissection: results of a randomized..., Pfister [/bib_ref] [bib_ref] Randomized controlled trial of a special acupuncture technique for pain after thoracotomy, Deng [/bib_ref] RCTs. Two [bib_ref] Electro-acupuncture efficacy on pain control after mandibular third molar surgery, Tavares [/bib_ref] [bib_ref] Increased postoperative pain and consumption of analgesics following acupuncture, Ekblom [/bib_ref] RCTs used integrative APS for patients with oral surgery. Pooled results from both trials [bib_ref] Electro-acupuncture efficacy on pain control after mandibular third molar surgery, Tavares [/bib_ref] [bib_ref] Increased postoperative pain and consumption of analgesics following acupuncture, Ekblom [/bib_ref] showed that integrative APS did not significantly improve the VAS scores for these patients ( = 0.34).
## Analgesic requirement.
Forty-three RCTs measured analgesic use, and most studies documented a lesser analgesic requirement in APS intervention groups than in control groups. Meta-analysis of data from six RCTs [bib_ref] The effect of acupuncture on relieving pain after inguinal surgeries, Taghavi [/bib_ref] [bib_ref] Integrative acupoint stimulation to alleviate postoperative pain and morphine-related side effects: a..., Chung [/bib_ref] [bib_ref] Pain reduction of acupoint electrical stimulation for patients with spinal surgery: a..., Yeh [/bib_ref] [bib_ref] The effect of high and low frequency electroacupuncture in pain after lower..., Lin [/bib_ref] [bib_ref] Effects of electroacupuncture on intraoperative and postoperative analgesic requirement, Sim [/bib_ref] [bib_ref] Effects of acupuncture on post-cesarean section pain, Wu [/bib_ref] showed a significant reduction in total morphine consumption in intervention groups compared to the control groups ( = 0.0001). Similar results were obtained in the comparison of intervention and sham/placebo control groups ( < 0.00001; [fig_ref] Table 5: Summary of meta-analysis results [/fig_ref]. In addition, Lin et al. [bib_ref] The effect of high and low frequency electroacupuncture in pain after lower..., Lin [/bib_ref] reported that the morphine requirement after high-frequency EA was decreased by 31% compared with that after low-frequency EA. The evidence for APS reducing analgesic requirement in surgery patients was determined to be of Level I quality based on multiple overall high quality RCTs.
## Aes.
No serious AEs were associated with APS, and patients were reported to tolerate the intervention well in the 21 RCTs that reported on this outcome. Reported minor AEs included temporary increased pain [bib_ref] Acupuncture for pain and dysfunction after neck dissection: results of a randomized..., Pfister [/bib_ref] , localized pain or discomfort at insertion sites [bib_ref] Evaluation of acupuncture for pain control after oral surgery: a placebo-controlled trial, Lao [/bib_ref] [bib_ref] Auricular acupuncture for pain relief after total hip arthroplasty-a randomized controlled study, Usichenko [/bib_ref] , minor bruising or bleeding [bib_ref] Acupuncture for pain and dysfunction after neck dissection: results of a randomized..., Pfister [/bib_ref] , constitutional symptoms [bib_ref] Acupuncture for pain and dysfunction after neck dissection: results of a randomized..., Pfister [/bib_ref] , and a mild burning sensation with erythema [bib_ref] Capsicum plaster at the hegu point reduces postoperative analgesic requirement after orthognathic..., Kim [/bib_ref] [bib_ref] The analgesic effects of capsicum plaster at the Zusanli point after abdominal..., Kim [/bib_ref] [bib_ref] Prevention of postoperative sore throat using capsicum plaster applied at the Korean..., Park [/bib_ref]. Michalek-Sauberer et al. [bib_ref] Perioperative auricular electroacupuncture has no effect on pain and analgesic consumption after..., Michalek-Sauberer [/bib_ref] stated that 38% of patients reported minimal side effects of acupuncture, most commonly fatigue (16%) and ear pain (10%).
## Sensitivity and heterogeneity.
Given the detection of obvious heterogeneity ( 2 > 50%) in meta-analyses, we conducted a sensitivity analysis to remove studies with a greater risk of bias. The results are presented in [fig_ref] Table 6: Sensitivity analysis results [/fig_ref]. 2 values were decreased substantially by the removal of such trials in most comparisons.
# Discussion
In this review, it was determined that there is insufficient evidence thus far to conclude that APS is an effective nonpharmacological approach to the reduction of postoperative pain intensity for surgery patients, although the evidence did show a reduced analgesic requirement with no significant adverse effects in surgery patients. The results may have been affected by some limitations of this study, such as the wide variability of interventions and participants, absence of follow-up evaluation in most included trials, and the often mediocre methodological quality of the included studies. These factors contributed to the high heterogeneity of the data, which limits the strength of the evidence. No studies were considered to be at low risk of bias (≥200 participants) based on the number of participants. These factors may have led to overestimations of APS efficacy. Given the intensity of surgical trauma, postoperative pain is inevitable and it is deemed to be a serious problem. If this pain is not managed effectively, it can contribute to several clinical risks and affect patients' physical and psychological wellbeing; potential effects include emotional distress, infection, increased myocardial oxygen consumption, and prolonged hospitalization. Associated pathological changes can harm organs and lead to abnormal function [bib_ref] Transcutaneous electrical nerve stimulation on acupoints reduces fentanyl requirement for postoperative pain..., Lan [/bib_ref] [bib_ref] Randomised, controlled study of preoperative eletroacupuncture for postoperative pain control after cardiac..., Coura [/bib_ref]. Reduction of postoperative pain is therefore essential.
Our meta-analysis of overall effects from 39 trials showed that interventions involving stimulation of body or auricular points significantly reduced postoperative pain, as measured by VAS scores. Data from studies using integrative APS or manual acupuncture showed uncertain outcomes or no significant change. In one of these studies, Deng et al. [bib_ref] Randomized controlled trial of a special acupuncture technique for pain after thoracotomy, Deng [/bib_ref] suggested that these results may be due to the insufficient strength of APS to produce analgesic effects.
Among body APS studies, the largest subgroup analyzed, all intervention types except manual acupuncture significantly reduced postoperative pain. The precise analgesic mechanism of body APS remains unclear. However, it has been found to facilitate central nervous system release of met-enkephalin and dynorphins into the spinal fluid, causing synergistic pain relief with exogenous opioid medication and production of pain-producing substances, such as potassium and lactic acid [bib_ref] Transcutaneous electrical nerve stimulation on acupoints reduces fentanyl requirement for postoperative pain..., Lan [/bib_ref] [bib_ref] Pain reduction of acupoint electrical stimulation for patients with spinal surgery: a..., Yeh [/bib_ref] [bib_ref] Effect of intraoperative electroacupuncture on postoperative pain, analgesic requirements, nausea and sedation:..., El-Rakshy [/bib_ref] [bib_ref] Acupuncture and endorphins, Han [/bib_ref]. The finding that highfrequency EA at body points was more effective than lowfrequency EA may be due to differences in opioid peptide release [bib_ref] Randomised, controlled study of preoperative eletroacupuncture for postoperative pain control after cardiac..., Coura [/bib_ref].
Similarly, auricular APS therapies were found to significantly reduce postoperative pain, with the exception of manual acupuncture. The most commonly used auricular point is Shen Men, which generates analgesic, sedative, and anti-inflammatory effects [bib_ref] Auricular acupressure for managing postoperative pain and knee motion in patients with..., Chang [/bib_ref]. It also increases endorphin secretion and serotonin production, thereby suppressing the transmission of pain messages and thus pain perception [bib_ref] Acupuncture analgesia: areas of consensus and controversy, Han [/bib_ref]. The results for integrative (auricular and body) APS are less clear; this treatment was found to significantly reduce NRS and VAS scores. Thus, the existing evidence neither supports nor refutes the effectiveness of integrative APS for postoperative pain control.
We also undertook subgroup analyses of surgery types, including abdominal, knee, oral, cesarean, and cardiac surgeries. The meta-analysis results showed that body point acupuncture stimulation and auricular therapy had no significant change on VAS scores for patients undergoing knee surgery. The same trend was observed for patients receiving integrative acupoint stimulation and undergoing oral surgery. Short-term APS stimulation may have been insufficient to reduce patients' pain intensity after knee or oral surgery, or the postoperative rehabilitation program may have affected the results of APS interventions [bib_ref] Effects of acupuncture and sham acupuncture in addition to physiotherapy in patients..., Tsang [/bib_ref]. Rigorously designed large-scale RCTs are needed to identify the effects of APS for these kinds of patients.
This analysis also showed that APS significantly reduces patients' postoperative analgesic requirement. Given the dose-response relationship between analgesics and related adverse effects [bib_ref] Reduction in opioid-related adverse events and improvement in function with parecoxib followed..., Langford [/bib_ref] , any nonpharmacological method that reduces the use of analgesic medication is likely to be beneficial. Lin et al. 's [bib_ref] The effect of high and low frequency electroacupuncture in pain after lower..., Lin [/bib_ref] finding of reduced morphine requirement after high-frequency EA compared with that after lowfrequency EA demonstrates the existence of a dose-response relationship in this treatment as well. However, analgesic requirements are controlled by the health care staff and directly affected by the surgery type and patient's economic condition. Thus, analgesic medication use is not a particularly reliable indicator for the effects of APS.
No APS study reported the occurrence of a serious adverse event, although some minor (mild and transient) side effects were reported. To prevent such effects, APS should be carried out by experienced, well-trained health care professionals who understand the theories underlying this therapy and take necessary precautions.
APS may produce strong placebo effects; for example, sham acupuncture did not affect analgesic-related side effects but did exert a moderate pain-relieving effect [bib_ref] The effect of high and low frequency electroacupuncture in pain after lower..., Lin [/bib_ref]. The use of sham/placebo control groups, as in 36 of the examined RCTs, enables clear distinction between true and placebo effects. This meta-analysis showed that the true effects of APS were much stronger than placebo effects. Short-term APS and placebo interventions have shown similar effects, but long-term APS treatment causes beneficial changes in specific brain areas [bib_ref] Do the neural correlates of acupuncture and placebo effects differ, Dhond [/bib_ref].
A small sample size can distort the results of metaanalyses, by overestimating treatment effects, probably due to methodological weaknesses [bib_ref] Small study effects in meta-analyses of osteoarthritis trials: meta-epidemiological study, Nüesch [/bib_ref]. In our review, no studies were considered to be at low risk of bias (≥200 participants) on the basis of sample size. Forty-two studies were at an unknown risk of bias , and 17 studies (28.81%) were at a high risk of bias (<50 participants).
In this review, statistical heterogeneity was considerable, even with use of the random-effects model. The best level of evidence was not found for most forms of APS, suggesting that there is, thus far, insufficient evidence to conclude that APS is an effective method for reducing pain intensity in postoperative patients. Within the available body of evidence, there is Level I evidence supporting the effectiveness of body points plaster therapy. Additionally, there is Level II evidence supporting the use of body points EA, body points acupressure, and body points APS in abdominal surgery patients specifically, as well as Level II evidence supporting the use of auricular points seed embedding, manual auricular acupuncture, and auricular EA in surgery patients. Meanwhile, there is only Level III evidence for the use of body points APS in patients who have undergone cardiac surgery and a cesarean section and Level III evidence for the use of auricular point stimulation for pain reduction after abdominal surgery. The main reason that better levels of evidence were not achieved was the methodological quality of the included studies, with only 13 (22.03%) studies meeting at least five of the seven Cochrane review criteria and only 12 (20.33%) studies that were rated highly in key domains being considered at low risk of overall bias.
Two systematic reviews [bib_ref] Acupuncture and related techniques for postoperative pain: a systematic review of randomized..., Sun [/bib_ref] [bib_ref] Auricular acupuncture for postoperative pain control: a systematic review of randomised clinical..., Usichenko [/bib_ref] with objectives similar to those of the present study were published in 2008, but overall they produced low quality evidence due to the insufficient quality of included trials. A number of the clinical trials included in the present analysis also had some methodological problems that may have affected their efficacy results. However, we examined all types of APS, with combined and separate analyses of body, auricular, and integrative APS. Rigorously designed large-scale RCTs are needed to identify an optimal standard APS program.
# Study limitations.
Some limitations of this study may have affected the results. For example, the wide variability in APS and surgery types, populations, intervention durations, and timing of outcome measurement may be the main factors underlying the observed heterogeneity, which limits the strength of the study results. The small samples and absence of follow-up evaluation in most included trials may have led to overestimation of the effects of APS. Methods of randomization, blinding, and allocation concealment were not reported or were poorly described in some trials, making quality assessment difficult. In addition, visual inspection 22 Evidence-Based Complementary and Alternative Medicine of the funnel plots revealed some substantial asymmetry in comparisons; thus, the possibility of publication bias (i.e., preference for publication of significant over nonsignificant results) cannot be excluded. In addition, the end-points of included studies varied. End-points in Gilbertson et al. [bib_ref] Acupuncture and arthroscopic acromioplasty, Gilbertson [/bib_ref] and Chen et al. [bib_ref] The application of electroacupuncture to postoperative rehabilitation of total knee replacement, Chen [/bib_ref] were 4 and 3 months, respectively. When removing these two studies, the 2 values were decreased markedly [fig_ref] Table 6: Sensitivity analysis results [/fig_ref]. Therefore, the end-points of included studies have important biases in this review. Future studies of APS should be designed rigorously to ensure a high level of methodological quality.
## Implications for practice and research.
The major advantages of APS are related to its clinical safety, favorable effects in postoperative pain relief, and low complication rate following surgery [bib_ref] Preoperative intradermal acupuncture reduces postoperative pain, nausea and vomiting, analgesic requirement, and..., Kotani [/bib_ref]. Clinical nurses and other health care providers should thus be encouraged to learn and implement this simple, convenient, and economical method of postoperative pain control in routine clinical care [bib_ref] Pain reduction of acupoint electrical stimulation for patients with spinal surgery: a..., Yeh [/bib_ref].
Our findings have implications for research on the precise mechanism of APS in postoperative pain relief. Optimal acupoint selection, session duration, stimulation intensity, and application frequency have not been established. A standardized APS program for postoperative pain management should be designed using an evidence-based method. Because available evidence for integrative APS and manual acupuncture is inconclusive, further studies should focus on further assessing the effects of these treatments on postoperative pain control. Moreover, the best APS type for the reduction or elimination of long-term opioid use and the long-term effects of APS therapies remain unknown. Thus, large-scale multicenter RCTs with long-term followup periods should be conducted to verify the short-and long-term effects of APS on postoperative pain control. Furthermore, more attention should be paid to the economic effects of APS in health care systems.
In conclusion, this study indicates that, thus far, there is still insufficient evidence to conclude that APS is an effective method for controlling postoperative pain in surgery patients, although the evidence does suggest that APS can reduce patients' analgesic requirement with no significant adverse effects. The best level of evidence was not adequate in most subgroups. Some limitations of this study may have affected the results, leading to an overestimation of the effects of APS. Rigorously designed large-scale RCTs are needed to identify the effects of APS.
# Appendix
## A. searching strategies
A. [bib_ref] Complications occurring in the postanesthesia care unit: a survey, Hines [/bib_ref]
[fig] Figure 1: Flow chart of the study selection process. CINAHL, Cumulative Index to Nursing and Allied Health Literature; AMED, Allied and Complementary Medicine Database; CBMdisc, Chinese Biological Medical Literature Database; CNKI, China National Knowledge Infrastructure. [/fig]
[fig] Figure 2: Methodological quality of included studies. Each methodological quality item was qualitatively assessed and is presented as a percentage across all included studies. [/fig]
[fig] Figure 3: sequence generation (selection bias) Allocation concealment (selection bias) Blinding of participants and personnel (performance bias) Blinding of outcome assessment (detection bias) Incomplete outcome data (attrition bias) Selective reporting (reporting bias) Other biases Adib-Hajbaghery and Etri, 2013 An et al., 2013 Chang et al., 2012 Chen et al., 2012 Chen et al., 2005 Chen et al., 1998 Chung et al., 2014 Colak et al., 2010 Coura et al., 2011 Deng et al., 2008 Deng et al., 2010 Dias et al., 2010 Ekblom et al., 1991 El-Rakshy et al., 2009 Felhendler and Lisander, 1996 Gilbertson et al., 2003 Gupta et al., 1999 He et al., 2013 He et al., 2007 Kager et al., 2009 Kim and Nam, 2006 Kim et al., 2009 Kong, 2012 Kotani et al., 2001 Lan et al., 2012 Langenbach et al., 2012 Lao et al., 1995 Lao et al., 1999 Larson et al., 2010 Lin et al., 2002 Li et al., 1994 Maimer et al., 2013 Marra et al., 2011 Masuda et al., 1986 Mehling et al., 2007 Michalek-Sauberer et al., 2007 Park et al., 2004 Pfister et al., 2010 Sahmeddini et al., 2010 Sakurai et al., 2003 Sator-Katzenschlager Sertel at el., 2009 Sim et al., 2002 Taghavi et al., 2013 Tavares et al., 2007 Tsang et al., 2011 Tsang et al., 2007 Usichenko et al., 2005A Usichenko et al., 2005B Usichenko et al., 2007 Wang et al., 1997 Wang et al., 2012 Wong et al. 2006 Wu et al., 2009 Yeh et al., 2010 Yeh et al., 2011 Yin et al.Risk of bias in the included studies. Each bias item was qualitatively assessed. [/fig]
[fig] Figure 4: Funnel plot. [/fig]
[table] Table 1: Qualitative modified approach to grading of evidence. When the same continuous outcome was assessed using different instruments, the standardized mean difference was calculated. For dichotomous outcomes, effect size variables, such as the relative risk (RR), were calculated. [/table]
[table] Table 2: Characteristics of RCTs examining body acupoint stimulation included in the meta-analysis. [/table]
[table] Table 3: Characteristics of RCTs examining auricular acupoint stimulation included in the meta-analysis. [/table]
[table] Table 4: Characteristics of RCTs examining integrative acupoint stimulation included in the meta-analysis. [/table]
[table] Table 5: Summary of meta-analysis results. [/table]
[table] Table 6: Sensitivity analysis results. [/table]
[table] . PubMed: #1 "acupuncture"[MeSH Terms] OR "acupuncture therapy"[MeSH Terms] OR "acupuncture analgesia"[MeSH Terms] OR "acupuncture points"[MeSH Terms] OR "acupressure"[MeSH Terms] OR "auriculotherapy"[MeSH Terms] OR "acupuncture, ear"[MeSH Terms] acupunctur * [Title/Abstract]) OR acupoin * [Title/Abstract]) OR acupressur * [Title/ Abstract]) OR auriculotherap * [Title/Abstract]) OR (auricu * [Title/Abstract] AND poin * [Title/ Abstract])) OR (ear[Title/Abstract] AND poin * [Title/Abstract])) OR (auricu * [Title/Abstract] AND acupoin * [Title/Abstract])) OR (ear[Title/ Abstract] AND acupoin * [Title/Abstract])) OR (auricu * [Title/Abstract] AND plaster * [Title/ Abstract])) OR (massag * [Title/Abstract] AND ear[Title/Abstract])) OR (ear[Title/Abstract] AND plaster * [Title/Abstract])) OR (massag * [Title/ Abstract] AND auricu * [Title/Abstract])) OR (magnetic[Title/Abstract] AND ear[Title/Abstract])) OR (magnetic[Title/Abstract] AND auricu * [Title/ Abstract])) OR otopoin * [Title/Abstract] OR vaccaria * [Title/Abstract] #3 #1 OR #2 #4 "perioperative period"[MeSH Terms] OR "postoperative period"[MeSH Terms] OR "preoperative period"[MeSH Terms] OR "intraoperative period"[MeSH Terms] #5 (((((perioperati * [Title/Abstract]) OR surger * [Title/ Abstract]) OR preoperati * [Title/Abstract]) OR intraoperati * ) OR postoperati * [Title/Abstract]) OR operati * #6 #4 OR #5 #7 ((((pain[Title/Abstract]) OR ache[Title/Abstract])) OR ("pain"[MeSH Terms] OR "acute pain"[MeSH Terms] OR "pain management"[MeSH Terms] OR "chronic pain"[MeSH Terms]) OR "analgesia"[MeSH Terms]) #8 #3 AND #6 AND #7 #9 (((((((("randomized controlled trial"[Publication Type]) OR "controlled clinical trial"[Publication Type]) OR "ramdomized"[Title/Abstract]) OR "ramdomised"[Title/Abstract]) OR "placebo"[Title/ Abstract]) OR "sham"[Title/Abstract]) OR "randomly"[Title/Abstract]) OR "trial"[Title/Abstract]) OR "groups"[Title/Abstract] #10 (animals[MeSH Terms] NOT (humans[MeSH Terms] AND animals[MeSH Terms])) #4 acupunctur * :ab,ti OR acupoin * :ab,ti OR acupressur * : ab,ti OR auriculotherap * :ab,ti OR (auricu * NEAR/3 acupunctur * ):ab,ti OR (auricu * NEAR/3 acupressur * ):ab,ti OR (auricu * NEAR/3 poin * ):ab,ti OR 'auricular plaster':ab,ti OR (ear NEAR/3 plaster * ): ab,ti OR (ear NEAR/3 poin * ):ab,ti OR (ear NEAR/3 acupoint * ):ab,ti OR otopoin * :ab,ti OR earhole * :ab,ti OR (vaccaria * NEAR/15 ear):ab,ti OR (vaccaria * NEAR/15 auricu * ):ab,ti OR (massag * NEAR/3 auricu * ):ab,ti OR (massag * NEAR/3 ear):ab,ti OR (cowherb NEAR/15 ear):ab,ti OR (cowherb NEAR/15 auricu * ):ab,ti OR (seed * NEAR/15 auricu * ):ab,ti OR (seed * NEAR/15 ear):ab,ti OR (magnetic NEAR/15 ear):ab,ti OR (magnetic NEAR/15 auricu * ):ab,ti OR erxue * :ab,ti #5 #1 OR #2 OR #3 OR #4 #6 'perioperative period'/exp #7 'postoperative period'/exp #8 'preoperative period'/exp #9 'intraoperative period'/exp #10 'surgery'/exp #11 'perioperative period':ab,ti OR operative:ab,ti OR surgery:ab,ti OR (peri NEAR/3 operative):ab,ti OR (post * NEAR/5 operative):ab,ti OR (pre * NEAR/5 operative):ab,ti OR (intra * NEAR/5 operative):ab,ti #12 #6 OR #7 OR #8 OR #9 OR #10 OR #11 #13 'pain'/exp #14 'analgesia'/exp #15 pain * :ab,ti OR analgesia:ab,ti OR ache * :ab,ti OR (pain NEAR/3 management):ab,ti OR (pain NEAR/3 control):ab,ti #16 #13 OR #14 OR #15 #17 #5 AND #12 AND #16 #18 'controlled clinical trial'/exp OR 'single blind procedure'/exp OR 'double-blind procedure'/exp OR 'crossover procedure'/exp #19 random * :ab,ti OR crossover * :ab,ti OR (cross NEAR/3 over * ):ab,ti OR placebo:ab,ti OR (doubl * NEAR/3 blind * ):ab,ti OR (doubl * NEAR/3 mask * ):ab,ti OR (singl * NEAR/3 blind * ):ab,ti OR (singl * NEAR/3 mask * ):ab,ti OR (trebl * NEAR/3 blind * ):ab,ti OR (trebl * NEAR/3 mask * ):ab,ti OR (tripl * NEAR/3 blind * ):ab,ti OR (tripl * NEAR/3 mask * ):ab,ti OR assign * :ab,ti OR allocat * :ab,ti OR volunteer * :ab,ti #20 #18 OR #19 #21 'animal'/exp OR 'nonhuman'/exp OR 'animal experiment'/exp #1 MeSH descriptor: [Acupuncture] explode all trees #2 MeSH descriptor: [Auriculotherapy] explode all trees #3 MeSH descriptor: [Acupressure] explode all trees #4 MeSH descriptor: [Acupuncture Analgesia] explode all trees #5 MeSH descriptor: [Acupuncture, Ear] explode all trees #6 acupunctur * or acupressur * or acupoin * or auriculotherap * or (auricu * near/3 poin * ) or (ear near/3 poin * ) or (ear near/3 plaster * ) or (auricu * near/3 plaster * ) or (ear near/3 acupoint * ) or otopoint * or earhole * or (vaccaria * near/15 ear) or (vaccaria * near/ 15 auricu * ) or (cowherb near/15 ear) or (cowherb near/15 auricu * ) or (magnetic near/15 ear) or (magnetic near/15 auricu * ) or (massag * near/3 ear) or (massag * near/3 auricu * ) or erxue * :ti,ab,kw (Word variations have been searched) #7 #1 OR #2 OR #3 OR #4 OR #5 OR #6 #8 MeSH descriptor: [General Surgery] explode all trees #9 perioperative period or operati * or surger * or (peri * near/5 operati * ) or (post * near/5 operati * ) or (pre * near/5 operati * ) or (intra * near/5 operati * ):ti,ab,kw (Word variations have been searched) #10 #8 OR #9 #11 MeSH descriptor: [Pain] explode all trees #12 MeSH descriptor: [Analgesia] explode all trees #13 pain * or analgesia or ache * or (pain near/3 management) OR (pain near/3 control):ti,ab,kw (Word variations have been searched) #14 #11 OR #12 OR #13 #15 #7 AND #10 AND #14 #16 #15 in Trials A.4. Cumulative Index to Nursing and Allied Health Literature (CINAHL) #1 MM Acupuncture OR MM Auriculotherapy OR MM Acupuncture, Ear OR MM acupressure OR MM acupuncture analgesia OR MM acupuncture points #2 TI acupunctur * OR acupoin * OR acupressur * OR auriculotherap * OR (auricu * N3 acupunctur * ) OR (ear N3 poin * ) OR (ear N3 plaster * ) OR (auricu * N3 plaster * ) OR (ear N5 acupoint * ) OR otopoint * OR earhole * OR (vaccaria * N15 ear) OR (vaccaria * N15 auricu * ) OR (magnetic N15 ear) OR (magne * N15 auricu * ) OR (massag * N3 ear) OR (massag * N3 auricu * ) #3 AB acupunctur * OR acupoin * OR acupressur * OR auriculotherap * OR (auricu * N3 acupunctur * ) OR (ear N3 poin * ) OR (ear N3 plaster * ) OR (auricu * Evidence-Based Complementary and Alternative Medicine N3 plaster * ) OR (ear N5 acupoint * ) OR otopoint * OR earhole * OR (vaccaria * N15 ear) OR (vaccaria * N15 auricu * ) OR (magnetic N15 ear) OR (magne * N15 auricu * ) OR (massag * N3 ear) OR (massag * N3 auricu * ) #4 #1 OR #2 OR #3 #5 MM perioperative period OR MM postoperative period OR MM preoperative period OR MM intraoperative period [/table]
|
Wearable sensors for clinical applications in epilepsy, Parkinson’s disease, and stroke: a mixed-methods systematic review
## Quantitative studies searched in pubmed
## Qualitative studies searched in scopus population
Search Strategy: epilepsy OR parkinson OR "Parkinson Disease" OR stroke OR seizure*
## Intervention
Search Strategy: TITLE-ABS-KEY ( accelerometer ) OR TITLE-ABS-KEY ( "wearable sensors" OR "wireless sensors" OR "body worn sensors" AND wearable OR smart AND cloth ) OR TITLE-ABS-KEY ( "wearable sensors" OR "wireless sensors" OR "body sensors" ) AND NOT mobile Comparison N/A Outcome Search Strategy: ( "focus group interview" OR "focus group discussion" OR interviews OR survey OR questionnaires OR qualitative OR "prefe rence" OR compliance OR acceptance OR needs OR satisfaction OR barriers OR f acilitator OR "User-centered design" ) Search strategy ( ( epilepsy OR parkinson OR "Parkinson Disease" OR stroke OR seizure* ) AND ( TITLE-ABS-KEY ( accelerometer ) OR TITLE-ABS-KEY ( "wearable sensors" OR "wireless sensors" OR "body worn sensors" AND wearable OR smart AND cloth ) OR TITLE-ABS-KEY ( "wearable sensors" OR "wireless sensors" OR "body sensors" ) AND NOT mobile ) AND ( "focus group interview" OR "focus group discussion" OR interviews OR survey OR questionnaires OR qualitative OR "prefe rence" OR compliance OR acceptance OR needs OR satisfaction OR barriers OR f acilitator OR "User-centered design" ) ) AND ( EXCLUDE ( PUBYEAR , 2018 ) OR EXCLUDE ( PUBYEAR , 2 017 ) OR EXCLUDE AND ( EXCLUDE ( SUBJAREA , "ENGI" ) OR EXCLUDE ( SUBJAREA , "CO MP" ) OR EXCLUDE ( SUBJAREA , "BIOC" ) OR EXCLUDE ( SUBJAREA , "AG RI" ) OR EXCLUDE ( SUBJAREA , "PHYS" ) OR EXCLUDE ( SUBJAREA , "MA TE" ) OR EXCLUDE ( SUBJAREA , "MATH" ) OR EXCLUDE ( SUBJAREA , "C HEM" ) OR EXCLUDE ( SUBJAREA , "CENG" ) OR EXCLUDE ( SUBJAREA , " ENVI" ) OR EXCLUDE ( SUBJAREA , "ARTS" ) OR EXCLUDE ( SUBJAREA , " EART" ) OR EXCLUDE ( SUBJAREA , "IMMU" ) OR EXCLUDE ( SUBJAREA , " PHAR" ) OR EXCLUDE ( SUBJAREA , "DECI" ) OR EXCLUDE ( SUBJAREA , " BUSI" ) OR EXCLUDE ( SUBJAREA , "ENER" ) OR EXCLUDE ( SUBJAREA , " ECON" ) OR EXCLUDE ( SUBJAREA , "VETE" ) ) AND ( LIMIT-TO ( LANGUAGE , "English" ) ) AND ( LIMIT-TO ( DOCTYPE , "ar" ) ) AND ( LIMIT-TO ( SRCTYPE , "j" ) ) |
Evaluation of an E-Learning Training Program to Support Implementation of a Group-Based, Theory-Driven, Self-Management Intervention For Osteoarthritis and Low-Back Pain: Pre-Post Study
Background: By adaptation of the face-to-face physiotherapist-training program previously used in the Self-management of Osteoarthritis and Low back pain through Activity and Skills (SOLAS) feasibility trial, an asynchronous, interactive, Web-based, e-learning training program (E-SOLAS) underpinned by behavior and learning theories was developed.Objective: This study investigated the effect of the E-SOLAS training program on relevant outcomes of effective training and implementation.Methods: Thirteen physiotherapists from across Ireland were trained via E-SOLAS by using mixed methods, and seven physiotherapists progressed to implementation of the 6-week group-based SOLAS intervention. The effectiveness of E-SOLAS was evaluated using the Kirkpatrick model at the levels of reaction (physiotherapist engagement and satisfaction with E-SOLAS training methods and content), learning (pre-to posttraining changes in physiotherapists' confidence and knowledge in delivering SOLAS content and self-determination theory-based communication strategies, administered via a SurveyMonkey questionnaire), and behavior (fidelity to delivery of SOLAS content using physiotherapist-completed weekly checklists). During implementation, five physiotherapists audio recorded delivery of one class, and the communication between physiotherapists and clients was assessed using the Health Care Climate Questionnaire (HCCQ), the Controlling Coach Behaviour Scale (CCBS), and an intervention-specific measure (ISM; 7-point Likert scale). A range of implementation outcomes were evaluated during training and delivery (ie, acceptability, appropriateness, feasibility, fidelity, and sustainability of E-SOLAS) using a posttraining feedback questionnaire and individual semistructured telephone interviews.Results:With regard to their reaction, physiotherapists (n=13) were very satisfied with E-SOLAS posttraining (median 5.0; interquartile range 1.0; min-max 4.0-5.0) and completed training within 3-4 weeks. With regard to learning, there were significant increases in physiotherapists' confidence and knowledge in delivery of all SOLAS intervention components (P<.05).RenderXPhysiotherapists' confidence in 7 of 10 self-determination theory-based communication strategies increased (P<.05), whereas physiotherapists' knowledge of self-determination theory-based strategies remained high posttraining (P>.05). In terms of behavior, physiotherapists delivered SOLAS in a needs supportive manner (HCCQ: median 5.2, interquartile range 1.3, min-max 3.7-5.8; CCBS: median 6.6, interquartile range 1.0, min-max 5.6-7.0; ISM: median 4.5, interquartile range 1.2, min-max 2.8-4.8). Fidelity scores were high for SOLAS content delivery (total %mean fidelity score 93.5%; SD 4.9%). The posttraining questionnaire and postdelivery qualitative interviews showed that physiotherapists found E-SOLAS acceptable, appropriate, feasible, and sustainable within primary care services to support the implementation of the SOLAS intervention.Conclusions:This study provides preliminary evidence of the effectiveness, acceptability, and feasibility of an e-learning program to train physiotherapists to deliver a group-based self-management complex intervention in primary care settings, which is equivalent to face-to-face training outcomes and would support inclusion of physiotherapists in a definitive trial of SOLAS.
# Introduction
International clinical guidelines for osteoarthritis and low-back pain endorse self-management, exercise, and physical activity as key components of health care interventions [bib_ref] EULAR recommendations for the non-pharmacological core management of hip and knee osteoarthritis, Fernandes [/bib_ref] [bib_ref] American College of Rheumatology 2012 recommendations for the use of nonpharmacologic and..., Hochberg [/bib_ref] , but the evidence for their effectiveness is weak and of low quality [bib_ref] Self-management programs for chronic musculoskeletal pain conditions: a systematic review and meta-analysis, Du [/bib_ref] [bib_ref] Effectiveness of self-management of low back pain: systematic review with meta-analysis, Oliveira [/bib_ref] [bib_ref] The effectiveness of physiotherapist-delivered group education and exercise interventions to promote self-management..., Toomey [/bib_ref]. The Self-management of Osteoarthritis and Low back pain through Activity and Skills (SOLAS) intervention is an evidence-supported group treatment approach developed through intervention mapping [bib_ref] Using intervention mapping to develop a theory-driven, group-based complex intervention to support..., Hurley [/bib_ref] , which is a logical six-step process for the development and evaluation of theory-driven and evidence-based interventions that takes into account stakeholder needs and the practicalities of implementation. SOLAS was evaluated for its acceptability and preliminary effects in comparison with individual physiotherapy in a feasibility trial (trial registration: ISRCTN49875385) set in Dublin, Ireland, between September 2014 and June 2016 [bib_ref] Theory-driven group-based complex intervention to support self-management of osteoarthritis and low back..., Hurley [/bib_ref]. Intervention physiotherapists who participated in the trial were trained using brief interactive lectures, videos, role play, and practical skills to deliver the SOLAS intervention using communication skills underpinned by self-determination theory. This theory proposes that people have basic psychological needs for autonomy, competence, and relatedness, which if met, for example, by the needs supportive communication style of a health care practitioner (HCP), will increase an individual's autonomous motivation and engagement in health behaviors such as self-management [bib_ref] Self-Determination Theory Applied to Health Contexts: A Meta-Analysis, Ng [/bib_ref]. The Medical Research Council guidelines recommend that complex behavior-change programs train their intervention deliverers to ensure implementation with high fidelity [bib_ref] Developing and evaluating complex interventions: the new Medical Research Council guidance, Craig [/bib_ref]. Hence, the Kirkpatrick model was used to evaluate training at the levels of Reaction, Learning, and Behavior, which showed that physiotherapists were satisfied with face-to-face training and their confidence in the self-determination theory-based communication strategies. Knowledge of the intervention content significantly increased, and the physiotherapists delivered SOLAS in a needs supportive manner with high fidelity to the intervention content [bib_ref] Using mixed methods to assess fidelity of delivery and its influencing factors..., Toomey [/bib_ref] [bib_ref] Feasibility of Training Physical Therapists to Deliver the Theory-Based Self-Management of Osteoarthritis..., Keogh [/bib_ref]. Upscaling to a definitive national trial would render the face-to-face training impractical for physiotherapists due to significant time, travel, and costs constraints [bib_ref] Using mixed methods evaluation to assess the feasibility of online clinical training..., Richmond [/bib_ref]. Therefore, we subsequently developed an asynchronous, interactive, Web-based, e-learning training program for SOLAS (E-SOLAS) to prepare physiotherapists to deliver the SOLAS intervention. If successful, the program would reduce the time needed to move to a definitive trial. Furthermore, E-SOLAS has the potential to increase the competencies of physiotherapists with regard to self-management behavior-change skills in line with the shared strategic priority of Ireland's public health service and higher education institutions to train and prepare future health care graduates with the skills necessary to support lifestyle behavior change in their patients, making the intervention more accessible to physiotherapists for long-term sustainability.
Despite the increased availability of e-learning training for HCPs internationally, there is limited formal evaluation of such training programs. Current evidence, which predominantly involves undergraduate HCP students [bib_ref] The Physiotherapy eSkills Training Online resource improves performance of practical skills: a..., Preston [/bib_ref] [bib_ref] Physiotherapy students' perspectives of online e-learning for interdisciplinary management of chronic health..., Gardner [/bib_ref] , suggests that e-learning shows similar effectiveness to traditional methods for knowledge acquisition [bib_ref] A systematic review and meta-analysis of online versus alternative methods for training..., Richmond [/bib_ref] and user satisfaction [bib_ref] Internet-based learning in the health professions: a meta-analysis, Cook [/bib_ref] , but further research regarding the effectiveness of e-learning on HCP behavior change and the translation of learning to clinical practice has been advocated [bib_ref] The effectiveness of Internet-based e-learning on clinician behaviour and patient outcomes: A..., Sinclair [/bib_ref]. Hence, we evaluated E-SOLAS in the same way as our face-to-face training. In addition, a range of World Health Organization-recommended implementation outcomes were included for evaluation, including the acceptability, appropriateness, feasibility, fidelity, and sustainability of E-SOLAS, in a range of primary care physiotherapist settings across Ireland [bib_ref] Implementation research: what it is and how to do it, Peters [/bib_ref] [bib_ref] Evaluating technology-enhanced learning: A comprehensive framework, Cook [/bib_ref] in order to understand the contextual elements of e-learning [bib_ref] Using mixed methods evaluation to assess the feasibility of online clinical training..., Richmond [/bib_ref].
The study objectives were to evaluate the effect of the E-SOLAS training program on physiotherapists' reaction, learning, and delivery of the SOLAS intervention as intended and to assess the acceptability, appropriateness, feasibility, fidelity, and sustainability of E-SOLAS to aid the implementation of the SOLAS intervention in primary care settings.
# Methods
## Study design and research ethics
This was a single-group, pre-post study. Ethical approval was granted by the UCD Human Subject (Sciences) Ethics Committee in two phases: in
## Participants and procedure
Physiotherapy managers from 10 primary care areas across Ireland who had not participated in the SOLAS feasibility trial were sent a study information leaflet for screening based on their service facilities and staffing capabilities. Seven physiotherapy managers fulfilled the criteria for inclusion, provided letters of support, and nominated two staff members to undertake E-SOLAS training. Nominated physiotherapist staff were sent the study information leaflet and consent form. Consenting participants were required to possess a device that could connect to the internet and were given password-protected access to the social learning platform Curatrthat hosted the E-SOLAS training program. Participants were encouraged to complete the training over a 4-week period by working at their own pace and at times that were convenient for them. During training, they had access to ongoing technical support from the research team and were requested to keep a log of the time spent on each aspect of training.
At the end of the training period, participants were invited to set up and deliver the SOLAS intervention according to the treatment protocol [bib_ref] Theory-driven group-based complex intervention to support self-management of osteoarthritis and low back..., Hurley [/bib_ref] in each of their primary care areas. Physiotherapists had ongoing access to E-SOLAS during implementation and were provided with any additional intervention materials required to deliver the intervention by the research team (ie, intervention PowerPoint [Microsoft Corp, Redmond, WA] slide deck on a universal serial bus, pedometers, and relaxation CDs for each client). Following completion of the 6-week delivery phase, each physiotherapist was invited to participate in an individual semistructured telephone interview to explore their views of E-SOLAS as a tool to support implementation of the intervention.
## E-solas training program
E-SOLAS is a Web-based e-learning training program designed to train physiotherapists to deliver a group-based education and exercise intervention for patients with osteoarthritis and chronic low-back pain. The content is based on the face-to-face training program developed for the SOLAS feasibility trial [bib_ref] Theory-driven group-based complex intervention to support self-management of osteoarthritis and low back..., Hurley [/bib_ref] [bib_ref] Feasibility of Training Physical Therapists to Deliver the Theory-Based Self-Management of Osteoarthritis..., Keogh [/bib_ref]. The E-SOLAS program is hosted on Curatr, an online social learning platform that creates a collaborative learning environment and uses gamification principles. The development process for the E-SOLAS program is outlined in Multimedia Appendix 1 [bib_ref] Self-Regulated Learning and Academic Achievement: An Overview, Zimmerman [/bib_ref] [bib_ref] Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being, Ryan [/bib_ref] [bib_ref] Validation of the theoretical domains framework for use in behaviour change and..., Cane [/bib_ref] [bib_ref] The behavior change technique taxonomy (v1) of 93 hierarchically clustered techniques: building..., Michie [/bib_ref].
## E-solas program structure and content
The E-SOLAS program contained six hierarchical linear levels, whereby the user was required to finish each level before progressing to the next level (Multimedia Appendix 2, [fig_ref] Figure 1: E-SOLAS home page screenshot [/fig_ref].
Briefly, the program begins in Level 1 with an overview of the training program and the SOLAS intervention. Level 2 describes the education content for each week of the SOLAS intervention (eg, the key learning points and the materials required; [fig_ref] Figure 2: E-SOLAS program content screenshot [/fig_ref]. At Level 3, the self-determination theory-based communication strategies that physiotherapists use as part of the intervention are introduced [fig_ref] Figure 3: E-SOLAS theory screenshot [/fig_ref] , and in Level 4, they are given the opportunity to role play these strategies. Level 5 highlights the exercises and their mode of delivery, and finally, Level 6 concludes the program by highlighting the next steps for intervention delivery. [fig_ref] Figure 4: E-SOLAS video-based program activity screenshot [/fig_ref] , videos from the research team, and a peer role model explaining certain elements of the intervention; short "in level" activities and self-reflection opportunities; and end-of-level "gate" assessments with varying levels and modes of feedback depending on the activity.
## Outcome measures
The effectiveness of the E-SOLAS training program was assessed using the Kirkpatrick model of evaluation at the levels of reaction, learning, and behavior. Furthermore, a range of implementation outcomes were evaluated during the training and intervention-delivery phases. The measurement tools used to assess learning and implementation outcomes are described in detail in Multimedia Appendix 3 and briefly outlined below.
## Training outcomes
## Reaction
To measure physiotherapists' reaction to E-SOLAS, a researcher-devised feedback measure was developed by adapting the face-to-face training feedback measure [bib_ref] Feasibility of Training Physical Therapists to Deliver the Theory-Based Self-Management of Osteoarthritis..., Keogh [/bib_ref] and incorporating factors related to the evaluation of technology enhanced learning [bib_ref] Evaluating technology-enhanced learning: A comprehensive framework, Cook [/bib_ref]. It was administered following posttraining assessment via SurveyMonkey and included items related to participant satisfaction, engagement, accessibility, and quality of E-SOLAS as well as several implementation outcomes as detailed below (Multimedia Appendix 4). Physiotherapists' engagement with E-SOLAS was further evaluated using Curatr analyticsand a self-reported activity log completed by each physiotherapist during training.
## Learning
Learning was assessed by evaluating physiotherapists' perceptions of self-reported knowledge and confidence pre-and posttraining (Multimedia Appendix 5) and their use of skills during training by using a range of measures.
## Behavior
Physiotherapists' behavior was assessed during delivery of the SOLAS intervention to evaluate fidelity to the intervention content and self-determination theory-based communication strategies using previously validated checklists [bib_ref] Using mixed methods to assess fidelity of delivery and its influencing factors..., Toomey [/bib_ref] and audio recordings [bib_ref] Feasibility of Training Physical Therapists to Deliver the Theory-Based Self-Management of Osteoarthritis..., Keogh [/bib_ref]. In line with fidelity guidelines [bib_ref] The Assessment, Monitoring, and Enhancement of Treatment Fidelity In Public Health Clinical..., Borrelli [/bib_ref] , each audio recording was coded by one blinded expert rater (AK) to assess physiotherapists' communication style [bib_ref] A self-determination theory approach to understanding the antecedents of teachers' motivational strategies..., Taylor [/bib_ref] , and three audio recordings were coded by a second expert rater (JM). The Health Care Climate Questionnaire (HCCQ) [bib_ref] Internalization of biopsychosocial values by medical students: a test of self-determination theory, Williams [/bib_ref] was the primary measure to assess provider delivery of the self-determination theory-based communication style, with an adapted version of the Controlling Coach Behaviour Scale (CCBS) [bib_ref] The controlling interpersonal style in a coaching context: development and initial validation..., Bartholomew [/bib_ref] and an intervention-specific SOLAS scale used as secondary measures [bib_ref] Feasibility of Training Physical Therapists to Deliver the Theory-Based Self-Management of Osteoarthritis..., Keogh [/bib_ref].
## Implementation outcomes
Implementation outcomes were measured through specific items on the feedback measure related to the acceptability, appropriateness, feasibility, and sustainability of E-SOLAS (Multimedia Appendix 4) and an individual semistructured telephone interview of the physiotherapists conducted by an experienced qualitative researcher (SG) within 2 weeks of completing group class delivery. A topic guide was developed for the participant interviews with specific questions and probes related to their views of E-SOLAS as a model of training in order to support physiotherapists in delivering the SOLAS intervention in primary care settings. All interviews were audio recorded.
# Data analysis
Data from all outcome measures were analyzed using Excel (version 14.2.3, Microsoft Corp) and a statistical software package (SPSS Statistics, version 20, IBM Corp, Armonk, NY) following checks for errors in data entry.
## Training outcomes
## Reaction
In order to assess physiotherapists' views of their satisfaction, accessibility, and quality of the E-SOLAS program and their engagement with the e-learning training, descriptive statistics were used to analyze quantitative data, and thematic analysis was used to analyze free-text answers.
## Learning
# Analytical methods
Descriptive statistics were used to calculate scores pre-and posttraining for overall confidence in delivering SOLAS content, the specific SOLAS intervention components, and the use of each self-determination theory-based communication strategy. Differences between pre-and posttraining were calculated using the Wilcoxon signed-rank tests and adjusted for multiplicity using Bonferroni corrections (0.05/n tests).
## Knowledge
Descriptive statistics were used to calculate the level of SOLAS intervention knowledge, and pre-and posttraining differences were calculated using the Wilcoxon signed-rank tests. Following discussion between raters, there was excellent agreement (100%) in the coding of physiotherapists' narrative case studies. The number of self-determination theory-based communication strategies used by each physiotherapists and the percentage of physiotherapists who used each strategy was calculated, with differences in the rate of use of all strategies and each strategy pre-and posttraining determined using McNemar tests. All results were adjusted for multiplicity using Bonferroni corrections.
## Skills
Each role-play audio recording was rated for the use of self-determination theory-based communications strategies on a 7-point Likert scale ranging from "1 -not at all well" to "7very well," with values at or above the mid-point of the Likert scale (4/7) defined as demonstrating skills that could be considered acceptable in terms of competence [bib_ref] Feasibility of Training Physical Therapists to Deliver the Theory-Based Self-Management of Osteoarthritis..., Keogh [/bib_ref].
## Behavior
The mean fidelity levels to SOLAS intervention content and fidelity levels according to physiotherapist, site, session, and session category were obtained by calculating total actual scores as a percentage of the total possible score using checklists. Fidelity of duration was established by calculating the difference between the actual and the intended session durations using a one-sample Wilcoxon test. Levels of fidelity were interpreted as previously reported in the literature [bib_ref] Development of a Feasible Implementation Fidelity Protocol Within a Complex Physical Therapy-Led..., Toomey [/bib_ref]. A review of the raters' scores for the audio recordings of physiotherapists' delivery of SOLAS session 4 demonstrated excellent agreement (90%). To establish physiotherapists' competence in the self-determination theory-based communication style, a median result for each of the three outcome measures was calculated separately. For the SOLAS scale, a median score per construct subsection (eg, autonomy), subcomponent strategy (eg, positive feedback), and class component (eg, education) was also calculated.
## Implementation outcomes
Descriptive statistics were used to analyze quantitative data related to physiotherapists' views of the acceptability, appropriateness, feasibility, and sustainability of E-SOLAS. Qualitative data from the physiotherapists' interviews were transcribed verbatim and analyzed using inductive thematic analysis [bib_ref] Using thematic analysis in psychology, Braun [/bib_ref]. A coding frame was developed from a review of provisional themes, which were then reexamined and refined (DMcA). The reliability of the identified themes was established by a second researcher (DAH) who independently coded a random sample of 25% of each dataset using the coding frame, with 70% agreement taken as the minimum cut-off rate [bib_ref] Theory-driven group-based complex intervention to support self-management of osteoarthritis and low back..., Hurley [/bib_ref]. The level of agreement between raters was 85%.
# Results
## Principal findings
Thirteen physiotherapists from seven primary care areas completed the E-SOLAS training, of which 12 were invited to participate in the implementation study (ie, delivery of the SOLAS intervention). Nine physiotherapists consented to participate, and seven progressed to deliver SOLAS. The profile of physiotherapists in each study phase is provided in , and the flow of participants through Phase 2 is outlined in [fig_ref] Figure 5: Participant flow through the study [/fig_ref]. The training and delivery groups were comparable for the majority of descriptive variables, apart from the median years qualified, which was lower in the delivery group.
## Training outcomes
## Reaction
Physiotherapists (n=13) were very satisfied with E-SOLAS training posttraining and found it enjoyable and engaging, with all participants completing the program within the 4 weeks available (Multimedia Appendix 4). Physiotherapists reported that they spent a mean of 9.1 (SD 3.3) hours (min-max 4.1-16.1) over 16.3 (SD 6.0) days to complete E-SOLAS, which was not statistically different from the duration of training recorded by Curatr analytics (mean difference -1.69; SD 4.37; t=-1.397; df=12; P=. [bib_ref] The Physiotherapy eSkills Training Online resource improves performance of practical skills: a..., Preston [/bib_ref] ; Multimedia Appendix 6). All physiotherapists successfully completed all-level gate assessments and the required three uploads and made at least one online posting to the group discussion. The majority of physiotherapists reported completing E-SOLAS outside work hours and spent 1-2 hours at any one time on training. The most commonly cited positive features of E-SOLAS were the range of brief video clips (46.2%; n=6) and focus on communication skills and client motivation (23.1%; n=3). Nine of the 13 participants experienced some difficulties during training; the most common difficulty was related to accessing online materials (46.2%, n=6), completing gate assessments (38.5%, n=5), and computer access at work (30.8%, n=4). Although the median ratings for working independently and not having access to other therapists were very positive, four physiotherapists required support from the University College Dublin team during training for accessing resources (n=3), logging into E-SOLAS via work email (n=2), or uploading audio files (n=1). Nonetheless, the majority of physiotherapists highly rated the quality of the training program and format.
## Learning
## Knowledge
Physiotherapists used all nine self-determination theory-based communication strategies in their responses pretraining; the most commonly used strategies were collaborative goal setting and action planning and building relationships, with no significant change in the rate of use of individual strategies posttraining [fig_ref] Table 2: Change in physiotherapists' confidence and knowledge of self-determination theory-based communication strategies [/fig_ref]. Knowledge of the SOLAS intervention content and structure improved overall as well as in nine of the 10 intervention components. The use of pain modalities significantly increased posttraining, of which knowledge of content, structure, and group-based exercise programs remained significant following Bonferroni corrections [fig_ref] Table 3: Change in physiotherapists' confidence and knowledge of the Self-management of Osteoarthritis and... [/fig_ref].
## Confidence
Physiotherapists' confidence significantly increased posttraining overall and for 7 of the 10 individual self-determination theory strategies; set clear expectations and provide direction remained significant after Bonferroni correction [fig_ref] Table 2: Change in physiotherapists' confidence and knowledge of self-determination theory-based communication strategies [/fig_ref].
Similarly, physiotherapists' confidence in delivery of the SOLAS content overall and all 10 intervention components significantly increased posttraining; five components remained significant after Bonferroni correction [fig_ref] Table 3: Change in physiotherapists' confidence and knowledge of the Self-management of Osteoarthritis and... [/fig_ref].
## Skills
The majority of physiotherapists demonstrated acceptable use of the self-determination theory skill scores during training (median 5.0; interquartile range 1.3; min-max 2.0-6.0), with only two physiotherapists scoring <4.
## Behavior
Of the six primary care sites that agreed to implement the intervention, five completed delivery and one site ceased delivery after Session 3 due to poor client attendance. Physiotherapists delivered SOLAS to a median of 4.0 (interquartile range 4; range 3-8) participants per class. The total mean %fidelity score (93.5%; SD 4.9%) and the overall fidelity scores were high (≥80%) . The difference between the actual and intended duration of all sessions was not statistically significant, apart from the education component of Session 1, which was significantly longer than the protocol (P=.03, Z=-2.23).
Physiotherapists delivered SOLAS in a needs supportive manner consistent with a self-determination theory-based communication style (HCCQ: median 5.2, interquartile range 1.3, min-max 3.7-5.8; CCBS: median 6.6, interquartile range 1.0, min-max 5.6-7.0; [fig_ref] Table 4: Physiotherapists' use of the self-determination theory-based communication strategies during implementation of Session... [/fig_ref]. The SOLAS scale results demonstrated that physiotherapists implemented the intervention overall with acceptable competence (median 4.5, interquartile range 1.2, min-max 2.8-4.8; [fig_ref] Table 4: Physiotherapists' use of the self-determination theory-based communication strategies during implementation of Session... [/fig_ref]. The median scores of only 2 of the 15 self-determination theory strategies were delivered below the competence level during both the education and exercise components of the intervention (ie, use support and encouragement rather than pressurising behaviors and acknowledge patient's feelings and perspectives).
## Implementation outcomes
## Posttraining feedback questionnaire
The median scores for physiotherapists' ratings of the acceptability, appropriateness, and sustainability of E-SOLAS training to support delivery of the SOLAS intervention were high (Multimedia Appendix 4). All physiotherapists reported that E-SOLAS could be used as a training method in primary care, with 100% of respondents (n=13) recommending it to other primary care physiotherapists and the majority expressing a preference for e-learning alone (69.2%, n=9) over blended learning (30.8%, n=4).
## Postdelivery qualitative interviews
Five of the seven physiotherapists who delivered the SOLAS intervention were interviewed within 2 weeks of program completion. Ten themes were identified from the analysis of participant interview data (Multimedia Appendix 8).
## Acceptability of e-solas
Physiotherapists reported that they had a very positive experience with E-SOLAS training and felt that it was an acceptable and valuable method of training. A number of physiotherapists emphasized the convenience and flexibility of e-learning as a method of training as compared to face-to-face training. The format of training with gate-level assessments and the resource materials contained in E-SOLAS were also viewed positively.
## Appropriateness of e-solas
All five physiotherapists were positive about the appropriateness of the E-SOLAS content and resources in meeting their practical needs and the needs of their clients in preparing them to deliver SOLAS using needs supportive communication. One physiotherapist also reported gaining greater confidence in managing clients beyond the class setting, whereas another physiotherapist added that the e-learning format had the advantage of allowing her to reflect on learning new skills in relation to the autonomy-supporting style of delivery.
## Feasibility of e-solas
## Demand
Although all physiotherapists reported spending additional time reviewing the E-SOLAS content and resources in preparation for delivery of the intervention, they felt the additional time was important for the first delivery of any new program and would reduce with subsequent deliveries (Multimedia Appendix 9).
## Adaptation
None of the physiotherapists reported deviating from the training specifications; however, all physiotherapists made recommendations for future adaptations to either the E-SOLAS content or training format. These included providing additional resources to guide physiotherapists in educating clients about the health risks associated with the overuse of pain medications, healthy eating guidelines, a wider range of exercise options, and the provision of outcome measures for clinicians to evaluate the intervention independently. Proposed adaptations to the training format included giving participants an estimate of the time required to complete each level and additional e-learning training and blended learning (ie, small-group face-to-face coaching alongside E-SOLAS) to support delivery of the self-determination theory-based communication strategies during goal setting and action-planning activities.
## Fidelity to e-solas
All physiotherapists aimed to deliver the intervention content and self-determination theory-based needs supportive communication with high fidelity.
## Sustainability of e-solas
Overall, physiotherapists were positive about the potential for integration of E-SOLAS into existing primary care settings to support the sustained use of the SOLAS intervention as a treatment and reported plans to continue implementation in their service area. One physiotherapist proposed training a designated clinician in each primary care area through face-to-face training, who would act as a peer mentor to support colleagues who completed E-SOLAS to specifically deliver the self-determination theory-based communication strategies.
# Discussion
## Overview
The overall aim of this study was to develop and evaluate an e-learning training program to support physiotherapists to deliver the SOLAS intervention in a primary care setting. The effectiveness of E-SOLAS on physiotherapists' knowledge, skills, and delivery of the SOLAS intervention was assessed alongside the acceptability and feasibility of the training program. Specifically, results indicated that physiotherapists' knowledge and confidence increased from pretraining to posttraining assessment and physiotherapists' behavior was positively influenced by E-SOLAS training, as the SOLAS intervention content and theory-based communication style were delivered with high fidelity. Finally, implementation outcomes posttraining and from the qualitative interviews were overtly positive with regard to the acceptability, appropriateness, feasibility, and potential for future integration of E-SOLAS into existing primary care health services.
## Effectiveness of the e-solas training program
Physiotherapists' confidence in the SOLAS intervention content and self-determination theory-based communication style increased posttraining, which is important because it can indicate how likely a learner (in this case, physiotherapist) is to engage in the required behavior [bib_ref] Self-efficacy: toward a unifying theory of behavioral change, Bandura [/bib_ref] [bib_ref] The theory of planned behaviour: reactions and reflections, Ajzen [/bib_ref]. Knowledge also increased for the SOLAS intervention content, but there were limited changes in knowledge of the self-determination theory-based communication strategies. This may be explained by physiotherapists' high pretraining knowledge levels, suggesting a ceiling effect, as they were highly experienced and the majority had undertaken communication-based training previously, thus limiting their potential for future improvement [bib_ref] A Meta-analysis of the Effectiveness of Intervention Programs Designed to Support Autonomy, Su [/bib_ref]. These findings mirror our previously published evaluation of SOLAS face-to-face training [bib_ref] Feasibility of Training Physical Therapists to Deliver the Theory-Based Self-Management of Osteoarthritis..., Keogh [/bib_ref]. The majority of physiotherapists also demonstrated acceptable competence in relation to their skills; however, two physiotherapists were rated below the competence level. Review of the audio recordings revealed that one recording was very short (<2 minutes), and thus, it was difficult to assess it in a meaningful way. Although guidelines were provided on how to conduct the role play, no guidance was given on its duration. Despite the difficulty in prescribing a set amount of time, a minimum time period could have been set to ensure a meaningful assessment, which could be applied for future iterations of E-SOLAS.
In terms of behavior, physiotherapists delivered the intervention as intended, adhering to the intervention content and delivery in a manner consistent with the self-determination theory-based communication style. The mean high fidelity to intervention content based on physiotherapists' self-reported checklists was 93%, which aligns with the findings of the previous feasibility trial [bib_ref] Using mixed methods to assess fidelity of delivery and its influencing factors..., Toomey [/bib_ref]. For assessment of the self-determination theory-based communication style, the physiotherapists' scores on the two global measures aligned closely with face-to-face training. More specifically, for the HCCQ, the median score was 5.2 in this study and 5.3 (on a 7-point Likert scale) in the face-to-face training study [bib_ref] Feasibility of Training Physical Therapists to Deliver the Theory-Based Self-Management of Osteoarthritis..., Keogh [/bib_ref]. Scores on the CCBS were consistent for both studies. However, an intervention-specific measure of needs support enables a more focused look at contextual elements [bib_ref] Promoting physical activity: development and testing of self-determination theory-based interventions, Fortier [/bib_ref]. Here, there was some divergence between e-training and face-to-face training, with median scores of 4.5 and 4.0 on a 7-point Likert scale, respectively, favoring E-SOLAS [bib_ref] Feasibility of Training Physical Therapists to Deliver the Theory-Based Self-Management of Osteoarthritis..., Keogh [/bib_ref].
In this study, two self-determination theory-based communication strategies (use support and encouragement rather than pressurising behaviours and acknowledge patients' feelings and perspectives) were delivered with low competence across both the education and exercise components of the class, highlighting the need for further training or adaption to E-SOLAS to further support these strategies. Interestingly, the communication strategies related to goal setting, action planning, and problem solving were delivered to a higher level of competence than the face-to-face training [bib_ref] Feasibility of Training Physical Therapists to Deliver the Theory-Based Self-Management of Osteoarthritis..., Keogh [/bib_ref]. This may have been due to the additional interactive elements added to E-SOLAS to address the concerns identified by the physiotherapists during the development phase. Furthermore, this improvement in goal setting-related strategies may have inadvertently reduced competence in the communication strategy use support and encouragement rather than pressurising behaviours, as emphasis was placed on physiotherapists being more directive with clients regarding goal setting in situations where clients were unable to articulate or formulate a goal themselves. Recent research has highlighted the difficulty in applying effective goal setting in clinical settings [bib_ref] Shared decision making within goal setting in rehabilitation settings: A systematic review, Rose [/bib_ref] , and future training programs need to consider these strategies carefully. Overall, E-SOLAS training seems at least as effective as face-to-face training in developing physiotherapists' knowledge, confidence, and ability to deliver the intervention as intended [bib_ref] Internet-based learning in the health professions: a meta-analysis, Cook [/bib_ref].
## Implementation outcomes for the e-solas training program
Physiotherapists were very positive about E-SOLAS following training and delivery and believed it was an acceptable, appropriate, feasible, and sustainable method of training in primary care. Participants spent a mean of 9 hours completing the training over 16 days while working at their own pace and predominantly in their own time, which has clear advantages over the 12-hour face-to-face training time in addition to travel, cost, and time off work experienced by physiotherapists in our previous feasibility trial. E-SOLAS participants demonstrated high levels of engagement with training, including a 100% completion rate within the specified timeframe. This may reflect the physiotherapist-recognized importance of group-based self-management programs for busy primary care settings as well as the emphasis HCPs now place on a client-centered communication style and the acquisition of behavior-change skills. Furthermore, these high levels of physiotherapists' satisfaction and engagement could also reflect the systematic and inclusive process used to develop the E-SOLAS training program according to the recommendations of the Medical Research Council [bib_ref] Developing and evaluating complex interventions: the new Medical Research Council guidance, Craig [/bib_ref] [bib_ref] Process evaluation of complex interventions: Medical Research Council guidance, Moore [/bib_ref].
In terms of feasibility, technical difficulties can sometimes hamper the success of e-learning with HCPs [bib_ref] The status of training and education in information and computer technology of..., Eley [/bib_ref]. Six of the 13 physiotherapists reported difficulty accessing online materials. Therefore, it is important to ensure that technical support is in place to maintain user engagement. One of the main advantages of e-learning is flexibility and control of the time and location for program completion [bib_ref] Evaluation of a Web-Based E-Learning Platform for Brief Motivational Interviewing by Nurses..., Fontaine [/bib_ref] , as demonstrated in this study, wherein the majority of participants completed E-SOLAS outside work.
Although physiotherapists were satisfied with the program overall, there were some adaptations suggested, including provision of further information to support the delivery of certain education components, inclusion of details of the estimated time to complete training, and the use of blended learning. These suggestions are in line with the general recommendations for e-learning programs should be tailored to HCPs' particular knowledge and experience [bib_ref] Instructional design variations in internet-based learning for health professions education: a systematic..., Cook [/bib_ref]. For example, in the context of E-SOLAS, one physiotherapist may want more information on pain medication, whereas another might like additional videos of communication strategies [bib_ref] Evaluation of a Web-Based E-Learning Platform for Brief Motivational Interviewing by Nurses..., Fontaine [/bib_ref]. Such individualized learning pathways may lead to not only a more engaged learner with enhanced knowledge but also more effective delivery of the intervention.
Despite the high rate of planned implementation of the SOLAS intervention posttraining, the program was fully delivered by six physiotherapists at five sites across four primary care areas. The main reasons for nonimplementation were beyond the scope of the study and were related to the nonavailability of staff. Of the five sites with full implementation, there was an equal mix of sole and shared delivery, in contrast to the previous feasibility trial where all physiotherapists delivered the intervention independently [bib_ref] Feasibility of Training Physical Therapists to Deliver the Theory-Based Self-Management of Osteoarthritis..., Keogh [/bib_ref]. Physiotherapists who delivered the intervention implemented it with high fidelity, apart from the education component of Session 1, which is consistent with the findings of face-to-face training [bib_ref] Using mixed methods to assess fidelity of delivery and its influencing factors..., Toomey [/bib_ref]. Although the qualitative interview findings did not suggest any significant barriers to future implementation by a sole practitioner following training, the suggestion of a local peer mentor and the development of blended learning may be warranted to overcome this potential obstacle.
# Strengths and limitations
The major strengths of this study are its focus on program development and evaluation within a group of experienced physiotherapists who received e-learning training while working within their primary care setting. Specifically, E-SOLAS was developed and underpinned by theory, with a clear rationale about how the intervention components were developed and adapted. The use of a formal evaluation modelallowed for a more comprehensive understanding of the effectiveness of E-SOLAS training, including the objective evaluation of physiotherapists' behavior during training, which is frequently absent from assessments [bib_ref] The effectiveness of Internet-based e-learning on clinician behaviour and patient outcomes: A..., Sinclair [/bib_ref] [bib_ref] Evaluation of Technology-Enhanced Learning Programs for Health Care Professionals: Systematic Review, Nicoll [/bib_ref] [bib_ref] An E-learning Module on Chronic Low Back Pain in Older Adults: Effect..., Jacobs [/bib_ref]. Furthermore, the application of the World Health Organization's implementation outcomes using mixed methods enabled a comprehensive assessment of the feasibility of implementation of this e-learning training program and required adaptations to increase acceptability [bib_ref] Implementation research: what it is and how to do it, Peters [/bib_ref]. Finally, the assessment of fidelity of intervention delivery using validated measures following e-learning has been rarely reported in the literature and is one of the novel aspects of this study.
A few limitations of this study should be acknowledged. Owing to the relatively small sample size, particularly for the delivery phase of the study, further investigation in a larger sample is warranted. Although a nonvalidated feedback measure was used to evaluate some training and implementation outcomes, its components were informed by a framework for the evaluation of technology-enhanced learning [bib_ref] Evaluating technology-enhanced learning: A comprehensive framework, Cook [/bib_ref] and our face-to-face training feedback measure [bib_ref] Feasibility of Training Physical Therapists to Deliver the Theory-Based Self-Management of Osteoarthritis..., Keogh [/bib_ref]. Although physiotherapists' competence to deliver the SOLAS intervention was assessed posttraining, there was no pretraining assessment of their skills, which should be included in future studies [bib_ref] Instructional design variations in internet-based learning for health professions education: a systematic..., Cook [/bib_ref]. Future studies should also incorporate some form of client measurement to more clearly understand the efficacy of this training approach. Self-report checklists were used to assess the fidelity to intervention content, which is less robust than other methods such as independently rated audio recordings [bib_ref] Using mixed methods to assess fidelity of delivery and its influencing factors..., Toomey [/bib_ref]. Any future research evaluating a new program should apply robust fidelity-assessment methods to all parts of the intervention [bib_ref] Development of a Feasible Implementation Fidelity Protocol Within a Complex Physical Therapy-Led..., Toomey [/bib_ref] [bib_ref] Treatment Fidelity Workgroup of the NIH Behavior Change Consortium. Enhancing treatment fidelity..., Bellg [/bib_ref]. Finally, the role-play activities were an important part of E-SOLAS training; however, they were designed as one-on-one interactions (ie, between the physiotherapist and one client). Therefore, physiotherapists did not get an opportunity to practice their delivery of the intervention in a group setting prior to implementation. Future programs should try to ensure that all elements of the intervention are accurately reflected in the training program.
# Conclusions
The comprehensive evaluation reported in this study provides preliminary evidence of the effectiveness, acceptability, and feasibility of an e-learning program to train physiotherapists to deliver a group-based self-management intervention in a primary care setting that is equivalent to face-to-face training. These findings will inform the development and implementation of a definitive trial and support its scalability to the wider primary care system.
[fig] Figure 1: E-SOLAS home page screenshot. E-SOLAS: E-learning training program for Self-management of Osteoarthritis and Low back pain through Activity and Skills. [/fig]
[fig] Figure 2: E-SOLAS program content screenshot. E-SOLAS: E-learning training program for Self-management of Osteoarthritis and Low back pain through Activity and Skills. [/fig]
[fig] Figure 3: E-SOLAS theory screenshot. E-SOLAS: E-learning training program for Self-management of Osteoarthritis and Low back pain through Activity and Skills. [/fig]
[fig] Figure 4: E-SOLAS video-based program activity screenshot. E-SOLAS: E-learning training program for Self-management of Osteoarthritis and Low back pain through Activity and Skills. Throughout each level, there are lectures with voice-overs, video examples of good and poor practice [/fig]
[fig] 1: Baseline all 6 sessions at Site 1; ID 3 ceased delivery in Site 6 after session 3. e ID 10 delivered all 6 sessions at Site 3, ID 10 and 11 delivered 3 sessions each at Site 5. J Med Internet Res 2019 | vol. 21 | iss. 3 | e11123 | p. 8 https://www.jmir.org/2019/3/e11123/ (page number not for citation purposes) [/fig]
[fig] Figure 5: Participant flow through the study. E-SOLAS: E-learning training program for Self-management of Osteoarthritis and Low back pain through Activity and Skills; ID: participant identification number. [/fig]
[table] Table 2: Change in physiotherapists' confidence and knowledge of self-determination theory-based communication strategies. [/table]
[table] Table 3: Change in physiotherapists' confidence and knowledge of the Self-management of Osteoarthritis and Low back pain through Activity and Skills intervention content. [/table]
[table] Table 4: Physiotherapists' use of the self-determination theory-based communication strategies during implementation of Session 4 of the Self-management of Osteoarthritis and Low back pain through Activity and Skills. [/table]
|
Feasibility, efficacy and safety of early lens extraction in patients with pseudoexfoliation glaucoma: a feasibility and pilot study
[bib_ref] Ocular and systemic pseudoexfoliation syndrome, Schlotzer-Schrehardt [/bib_ref] [bib_ref] The relationship between glaucoma and pseudoexfoliation: the Blue Mountains Eye Study, Mitchell [/bib_ref] [bib_ref] Exfoliation syndrome and exfoliation glaucoma, Vesti [/bib_ref] [bib_ref] Pseudoexfoliation, intraocular pressure, and senile lens changes in a population-based survey, Hiller [/bib_ref] [bib_ref] Prevalence and risk factors of lens opacities in the elderly in Finland...., Hirvela [/bib_ref] [bib_ref] Pseudoexfoliation in South India, Arvind [/bib_ref] [bib_ref] Clinical features of capsular glaucoma in comparison with primary open-angle glaucoma in..., Futa [/bib_ref] [bib_ref] Effectiveness of early lens extraction for the treatment of primary angle-closure glaucoma..., Azuara-Blanco [/bib_ref] [bib_ref] Long-term effect of phacoemulsification on intraocular pressure using phakic fellow eye as..., Chang [/bib_ref] [bib_ref] Intraocular pressure after phacoemulsification in patients with uncontrolled primary open angle glaucoma, Iancu [/bib_ref] [bib_ref] Effect of phacoemulsification surgery on various parameters in patients with glaucoma, Bhandari [/bib_ref] [bib_ref] Lowering of Intraocular Pressure After Phacoemulsification in Primary Open-Angle and Angle-Closure Glaucoma:..., Thomas [/bib_ref] [bib_ref] Intraocular pressure after phacoemulsification in eyes with pseudoexfoliation, Altan-Yaycioglu [/bib_ref] [bib_ref] Intraocular pressure decrease after phacoemulsification in patients with pseudoexfoliation syndrome, Merkur [/bib_ref] [bib_ref] Intraocular pressure following phacoemulsification in patients with and without exfoliation syndrome: a..., Damji [/bib_ref] [bib_ref] Surgical and laser interventions for pseudoexfoliation glaucoma systematic review of randomized controlled..., Pose-Bazarra [/bib_ref]
# Methods objective
The main objective of this pilot study is to inform a definite trial to determine whether clear or early lens extraction surgery is a clinically effective and safe intervention for the treatment of PXFG or ocular hypertension (OHT) and pseudoexfoliation syndrome compared to standard medical treatment.
## Study design
This pilot trial was carried out with patients recruited from the public health care system (SERGAS) in the Ophthalmology Units of the Hospital University Complex of Ferrol and Santiago de Compostela (Galicia, Spain) from May 2019 to February 2020. These hospitals serve a population of 182.751 and 450.136 patients, respectively. In this region there is a high prevalence of PXF [bib_ref] Prevalence of pseudoexfoliation syndrome among adult Spanish in the Salnes eye Study, Viso [/bib_ref]. Patients were randomized in two groups in equal proportion (1:1 ratio) by using sealed envelopes.
The pilot study was prospectively registered at ClinicalTrials.gov (NCT03494465, 03/21/2018) and it is described in line with CONSORT (Consolidated Standards of Reporting Trials) checklist for guidance for pilot and feasibility studies [bib_ref] statement: extension to randomised pilot and feasibility trials, Eldridge [/bib_ref].
Based on the recommendations of published studies on sample size calculations for pilot studies [bib_ref] Estimating the sample size for a pilot randomised trial to minimise the..., Whitehead [/bib_ref] [bib_ref] The size of a pilot study for a clinical trial should be..., Sim [/bib_ref] , we estimated our sample size of 40 patients in order to obtain an accurate and concrete view of the objectives we aim to achieve.
## Inclusion criteria
1. Mild or moderate PXFG or OHT over 25 mmHg + PXF syndrome.
Glaucoma was defined as: reproducible defects (two or more contiguous points with a loss of P < 0.01, three or more contiguous points with a loss of P < 0.05 or greater) in the visual field (VF) or detectable damage to the optic nerve (cup-to-disc ratio ≥ 0.7 and/or focal thinning of the optic nerve rim and/or asymmetry of cup disc ratio ≥ 0.2 between both eyes). Mild glaucoma was defined as MD < −6 dB, and moderate glaucoma was defined as MD < −12 dB. PXF was defined clinically by the presence of pseudoexfoliative material in the anterior capsule of the lens. 2. Presence of early cataract with some visual symptoms that justify the intervention, with best corrected visual acuity (BCVA) between range >0.4 and <0.7. 3. Naive diagnosed patients. 4. Age: patients > 60 years. 5. Willingness to participate in the trial and to defer cataract surgery for 12 months if randomized to initial medical treatment.
## Exclusion criteria
1. Advanced glaucoma. It is defined as an average deviation > −12 dBand/or threat of fixation (paracentral point with sensitivity of 0 dB), and/or cup-to-disc ratio > 0.9.
2. Corneal edema, corneal opacity or any other known corneal factor that may increase the risk of complications during surgery. 3. Previous cataract surgery in the eye considered for the study. 4. Axial length < 20 mm. 5. Estimated IOL power > 30 dioptres. 6. Signs of zonular weakness: phacodonesis, iridodonesis, lens subluxation, asymmetry of anterior chamber depth apparent on clinical exam. 7. Pupillary dilation < 5 mm. 8. Advanced cataract, with vision worse than 0.4.
## Primary and secondary outcomes
The primary outcome was IOP at 12 months post-randomization.
Secondary outcomes included clinical, patient reported outcomes and safety, specifically: VF according to the global VF index (VFI), retinal nerve fibre layer (RNFL) global thickness by optical coherence tomography (OCT), need for glaucoma surgery, BCVA, quality of life using a visual function quality questionnaire: National Eye Institute Visual Function Questionnaire (NEI-VFQ39), number of antiglaucoma drugs, adverse events and rate of recruitment.
Investigators evaluating outcomes were masked for IOP measurements and interpretation of VF tests results. Patients were not masked.
## Measurements and interventions
Target IOP was used in both groups to inform the need of escalation of therapy; in patients with OHT the target IOP was set at 21 mmHg. In patients with mild glaucoma target IOP was 18 mmHg, and in patients with moderate glaucoma the target IOP was set at 15 mmHg [bib_ref] Target IOP Workshop participants. Canadian perspectives in glaucoma management: setting target intraocular..., Damji [/bib_ref]. Patients in both groups were treated in a staggered manner: 1. single topical medication with prostaglandin analogues; 2. double topical therapy; 3. triple topical therapy. If necessary, glaucoma surgical treatment was offered.
In lens extraction arm, patients underwent standard lens phacoemulsification with intraocular lens implant (IOL) within 60 days after randomization. Medical treatment was initiated at the time of diagnosis. If additional treatment was required, the stepped sequence of therapy described above was used.
Patients allocated to medical treatment were offered lens extraction surgery after 12 months post-randomization. Two ophthalmologists (SPB and MJVL), one per centre, performed the lens extraction surgeries. The calculation of the biometric power was made with an IOL Master (Carl Zeiss Meditec AG, Jena, Germany), and the formula SRK-T was used. IOL implanted was Clareon ® Aspheric Hydrophobic Acrylic IOL, Model SY60WF (Alcon Laboratories, Inc., Fort Worth, TX). IOP measurement, with Perkins applanation tonometry, was masked by using a second observer. Humphrey Perimeter (SITA Standard 24-2) (Carl Zeiss Meditec AG, Jena, Germany) was used for VFs testing; RNFL determinations were carried out with Spectralis ® spectraldomain OCT (Heidelberg Engineering, Inc., Heidelberg, Germany). The BCVA was taken on a Snellen scale. Complications related to cataract surgery and adverse events were collected.
The visit schedule included a complete examination at baseline and at the end of the study , including visits at 4 and 8 months for IOP, BCVA, number of treatments, OCT RNFL and possible complications. Visual Function Questionnaire was completed at baseline and at the end of the study.
# Statistical analysis
Statistical comparison of patient baseline characteristics between treatment arms was carried out using the exact Wilcoxon-Mann-Whitney (WMW) test or Fisher's exact test, according to the variable type. Changes in variables describing visual status were computed by subtracting final (i.e., at 12 months) measurements from baseline values. For these variables, 95% confidence intervals (95%CI) for the median change in each of both arms were computed using Bauer's procedure [bib_ref] Constructing confidence sets using rank statistics, Bauer [/bib_ref] , based on the Wilcoxon signed rank statistic. Comparison of the change in these variables between treatment arms was performed using the exact WMW test.
For statistical computing and graphs, Rand the R packages coin [bib_ref] Implementing a class of permutation tests: the coin package, Hothorn [/bib_ref] , exactRankTestsand ggplot2were used.
# Ethical-legal aspects
The development of the project was carried out respecting the Declaration of Helsinki of the World Medical Association 1964 and successive ratifications. The present study was approved by the Research Ethics Committee (CEI). Any clinical data collected from the subjects was separated from the personal identification data, guaranteeing the confidentiality of the participants in the research and respecting the Law of Protection of Personal Data.
This study was supported in kind by the researchers' organizations.
## Survey
An anonymous online survey was developed using a commercial website (SurveyMonkey, Palo Alto, California, USA) and was distributed to the members of the UK and Eire Glaucoma Society (UKEGS) (http:// www.glaucoma-societyuke.org) and the Spanish Glaucoma Society (SEG) (https://www.sociedadglaucoma.com) in October 2019. A second e-mail survey was sent 4 weeks after as a reminder. Data were analyzed using surveymonkey.com analysis tools and Microsoft Excel (Microsoft Corporation, Redmond, WA, USA). At the time of survey distribution, there were 116 glaucoma subspecialty consultants registered as UKEGS members, and 429 registered as SEG members.
Participants were asked questions about percentage of PXF patients treated in daily consultation, their current practices in the timing for phaco [fig_ref] Table 1: Online survey for glaucoma experts [/fig_ref].
# Results
From May 2019 to February 2020, 19 patients were assessed for eligibility in the study. Recruitment was halted due to the COVID-19 pandemic. The number of patients declining to participate in the trial was 7 (39%). Finally, 12 patients were randomized to either phaco-intervention group (n = 6) or control group (n = 6) and none were lost to follow-up. Participants in both arms were comparable at baseline and no statistically differences at the 0.05 significance level were found between groups. [fig_ref] Table 2: Characteristics of the patients at baseline by treatment [/fig_ref] provides a detailed description of baseline patient characteristics. [fig_ref] Table 3: Comparison of visual status variables between treatment groups at 12 months [/fig_ref] shows summaries of the variables characterizing the visual status of the patients at 12 months from randomization in both treatment groups. In [fig_ref] Table 3: Comparison of visual status variables between treatment groups at 12 months [/fig_ref] the measurements are also expressed in terms of the change from baseline. [fig_ref] Figure 1: Post-treatment IOP outcome [/fig_ref] illustrates the time profile of IOP at baseline, 4, 8 and 12 months by treatment. Median IOP decreased in both control and intervention groups, from 29.5 mmHg baseline to 15.0 mmHg and 16.0 mmHg at 12 months, respectively (95% CI's for the median change from baseline and , for control and intervention groups, respectively). At the 0.05 significance level, no location shift in the distribution of the change of IOP between treatment arms was found (p value = 0.893).
## Primary outcome
## Secondary outcomes
For BCVA, the time profile at baseline, 4, 8 and 12 months by treatment is displayed in [fig_ref] Figure 2: Post-treatment BCVA outcome [/fig_ref] by means of a grouped box (and dot) plot. Median BCVA decreased in standard medical treatment from 0.65 at baseline to 0.5 at 12 months (95% CI for the median change: [0.1; 0.3]), while it improved in the surgery arm from 0.5 to 1 (95% CI for the median change:
[formula] [−0.5, −0.2]). [/formula]
The evolution of the number of antiglaucoma treatments (NT) along time is shown with the conditional bar chart displayed in [fig_ref] Figure 3: Changes in NT [/fig_ref]. Median NT remained stable in the control group, but in the lens extraction group it was reduced from 1.5 at baseline to 0 at 12 months (no 95% CI for the median change is given, because the technique cannot be reliably applied due to extreme data discreteness).
As for the VFQ-39 questionnaire, [fig_ref] Figure 4: Post-treatment VFQ-39 outcome [/fig_ref] shows box (and dot) plots of the scores by time and treatment group. The median VQF-39 score showed a worsening in the visual status of the patients treated with standard therapy, decreasing from 90. For all of BCVA, NT and VQF-39, the distribution of the change from baseline to 12 months showed a significant location shift between treatment groups at the 0.05 level (p value = 0.002, 0.013 and 0.009, respectively). On the other hand, at the same significance level, no between-group differences were found for any of the other secondary outcomes (OCT parameters, VF index and mean deviation and endothelial cell count). In [fig_ref] Table 2: Characteristics of the patients at baseline by treatment [/fig_ref] , 95% Eye CI's for the median change of these variables between visits can be found, both for control and intervention groups.
## Adverse events
No adverse events were encountered in either arm of the trial. No complications occurred during cataract surgery. No patients required additional glaucoma surgery during the study follow-up period.
## Survey
A total of 72 glaucoma experts from Spain, UK and Ireland completed the survey representing a 14% response rate from eligible experts. Most respondents (94%) were glaucoma specialists. All survey questions were completed.
PXFG is a common disease in the daily practice of glaucoma clinicians of both societies. The majority (over 50%) see several patients with PXFG every week.
Overall, glaucoma experts considered phaco to be an effective lowering IOP procedure, but discrepancies were seen between answers in both societies. In UK/Eire, nearly half of respondents (46%), considered that phaco was an effective option and an adequate initial procedure while a similar proportion considered that the efficacy was limited. In Spain, less (30%) of survey respondents thought that phaco was sufficient to control de IOP at least in some cases, while the majority (63%) thought that it could control IOP in certain cases but usually it did not provide a satisfactory control.
UK/Eire and Spain respondents agreed (79% and 88% respectively) to early cataract extraction in order to avoid possible complications of denser cataracts and weaker zonules.
On the question about the possibility of changing practice if there was evidence from a large trial proving that early lens extraction is safe and effective for IOP control in patients with PXFG, 50% of the specialists surveyed would be willing to offer this option, as is already their current practice. The rest of the participants (49%) would be keen to do it if the results show efficacy with a good safety profile.
The majority of respondents (71%) expressed an interest in participating in a trial evaluating the efficacy and safety of lens extraction in patients with PXFG.
# Discussion
The results of this randomized pilot and feasibility study provide preliminary confirmation that a large trial of early lens extraction for PXFG would be feasible as it would have support from glaucoma experts from Spain, UK and Ireland, although some challenges remain.
Similar to previous studies, in patients with mild to moderate PXFG controlled with 1 or 2 medications, phacoemulsification has resulted in a moderate decrease in IOP (20%) and in the number of medications required after surgery (35%) [bib_ref] The Effect of Phacoemulsification on Intraocular Pressure in Glaucoma Patients: A Report..., Chen [/bib_ref]. This IOP reduction seems to be maintained over time, even 7 years after cataract surgery [bib_ref] Effect of phacoemulsification on intraocular pressure in eyes with pseudoexfoliation: singlesurgeon series, Shingleton [/bib_ref]. In a very recent study, the results of phacoemulsification alone are comparable in terms of IOP decrease versus viscocanalostomy in patients with mild to moderate PXFG. Patients with advanced stages of PXFG were excluded from the study due to uncertainty about the efficacy and safety of this intervention. In a future trial it will be useful to stratify the population according to the severity of disease and IOP level.
The sample size was lesser than expected and was halted because of the COVID pandemic. Of 19 eligible patients, 12 were included, achieving a recruitment rate of over 50% (65%). The most frequent reason for declining participation was refusal of surgery given the insufficient visual limitations they presented. This was surprising considering that patients had cataract and some degree of vision loss.
The mean age of the patients included in the trial was 77 years, influenced by our inclusion criteria, which used a visual acuity cut-off point. The indication for cataract surgery in clinical practice is individualized and typically based on the presence of a visual symptoms and limitations, without any fixed visual acuity criterium. Participants enrolled in our pilot study had early or moderate cataracts but were willing to defer surgery. Actually, some patients who had reduced vision due to early cataract and fulfilled other inclusion criteria declined to participate in the study because they were asymptomatic and would not consider surgery. In our study, a BCVA between 0.4 and 0.7 on the Snellen scale was used, avoiding dense cataracts. In a future RCT it will be interesting to target patients with an earlier stage of lens opacity and better visual acuity. However, the possibly of including patients with clear lens and perfect vision will raise clinical and ethical challenges that will need to be addressed.
Lens removal in patients with PXF is a procedure with higher complication rates due to its smaller pupillary diameter and its greater zonular weakness [bib_ref] Outcome of phacoemulsification and intraocular lens implantion in eyes with pseudoexfoliation and..., Shingleton [/bib_ref]. However earlier cataract surgery in people with PXF may decrease the risk of surgical complications [bib_ref] Intraocular lens dislocation in pseudoexfoliation: a systematic review and meta-analysis, Vazquez-Ferreiro [/bib_ref] if zonules are not yet compromised. In our small sample, we have not found any adverse event related to the surgery or a lower endothelial cell count after the intervention. A prospective study has been published [bib_ref] Corneal endothelial cell loss following cataract surgery in patients with pseudoexfoliation syndrome:..., Kristianslund [/bib_ref] analyzing endothelial damage secondary to cataract surgery in pseudoexfoliative patients compared to a non-PXF control group. The results show that there are no significant differences at 2 years between groups in the corneal endothelial cell loss. However, even though our pilot study was not associated with serious adverse events we acknowledge that our study and even a larger randomized trial may not have sufficient power to quantify the incidence of uncommon severe complications. Many of the complications may be late and would require a longer follow-up than 1 year to be observed and larger sample size.
Our survey found consistent agreement among glaucoma specialists about the importance of an early cataract surgery in patients with PXF in order to avoid possible difficulties associated denser cataracts, weaker zonular support and poor pupillary dilation. Early cataract surgery is a feasible option and chosen by many surgeons to improve control IOP while improving the visual and life quality of their patients. The majority of the respondents would be willing to participate in a multicentre study evaluating the efficacy of early lens extraction for IOP control in PXFG patients without cataract.
The following are the limitations of the study. First, our original target for our pilot study was not met, in part due to the COVID pandemic. However, we consider our study has been able to answer key questions regarding the feasibility of a larger, definite trial.
Another limitation of this study is the relatively low response rate to the survey, lower than other online glaucoma surveys in the United Kingdom [bib_ref] Survey of glaucoma surgical preferences and post-operative care in the United Kingdom, Rodriguez-Una [/bib_ref] , but in common with other previously published online surveys of glaucoma surgical preferences [bib_ref] Practice Preferences for Glaucoma Surgery: A Survey of the American Glaucoma Society, Vinod [/bib_ref]. Since it has been conducted in different countries and we have achieved similar opinions, we can have more confidence in the generalizability of our findings.
The main methodological strength of this trial is its prospective randomized design and the rigorous methodology, with masked investigators evaluating the primary outcome. Despite the small sample size, we obtained statistically significant results in relation to quality of life, number of treatments and BVCA, and confirmed the efficacy of the intervention up to 12 months. In addition, this pilot study will serve to calculate the sample size of a larger study that will allow us to analyze the hypothesis in depth.
To the best of our knowledge, this is the first randomized clinical trial (RCT) investigating the feasibility and preliminary effect of early cataract surgery as a treatment for PXFG. The evidence from RCT on surgical interventions for PXFG is still very scarce. Our study supports the need for a definitive larger RCT to evaluate early or clear lens extraction in patients with PXFG.
## Summary
## What was known before
- Lens removal has been demonstrated by multiple retrospective studies as an effective hypotensive treatment in almost all types of glaucoma.
- Only in cases of narrow angles has it been further demonstrated by prospective well-designed trials.
What this study adds - First RCT to demonstrate the feasibility of lens extraction for the treatment of glaucoma or hypertension with pseudoexfoliation syndrome.
- Survey of glaucoma experts supporting the hypotheses put forward in the study.
## Data availability
Anonymised datasets generated and analyzed during this study are available from the corresponding author upon reasonable request.
[fig] 2: at the baseline to 89.2 at the final visit (95% CI for the median change: [−4.5; 13.5]). On the contrary, it showed an improvement in the surgery group, increasing from 83.4 to 95.8 (95% CI for the median change: [−20.1; −2.4]). [/fig]
[fig] Figure 1: Post-treatment IOP outcome. Combined box and dot plots of IOP by time and treatment group. [/fig]
[fig] Figure 3: Changes in NT. Bar chart of NT by time and treatment group. [/fig]
[fig] Figure 4: Post-treatment VFQ-39 outcome. Combined box and dot plots of VFQ-39 by time and treatment group. S. Pose-Bazarra et al. [/fig]
[table] Table 1: Online survey for glaucoma experts: multiple-choice questionnaire. 1. Do you see many patients with PXFG or PXF in your practice? • Yes, I see several patients every week • Not often, only a few patients every month • PXF is rare in my area 2. Do you think phaco + IOL improves IOP control in patients with PXFG or PXF and high IOP? • Yes, in my experience this is a common finding • Sometimes but typically the change in IOP is minor • No 3. Do you modify the timing of phaco in patients with cataract and PXFG? • Yes, I offer earlier phaco in patients with PXFG and early caratact, as I think surgery will be more difficult with denser cataract or it may help control the IOP • Occasionally • No 4. If there were evidence from a large trial that early lens extraction is safe and effective for IOP control in patients with PXFG, would you be willing to offer this option to your patients? • Yes, and this is already my current practice • Possibly, if the results of the trial are definite • No 5. Would you be willing to participate in a trial evaluating the efficacy and safety of lens extraction in patients with PXFG? • Yes • Possibly • No 6. What is your position? (e.g., consultant, ophthalmologist, hospital optometrist, trainee) S. Pose-Bazarra et al. [/table]
[table] Table 3: Comparison of visual status variables between treatment groups at 12 months. [/table]
|
Phase Transitions in Paradigm Shift Models
Two general models for paradigm shifts, deterministic propagation model (DM) and stochastic propagation model (SM), are proposed to describe paradigm shifts and the adoption of new technological levels. By defining the order parameter m based on the diversity of ideas, D, it is studied when and how the phase transition or the disappearance of a dominant paradigm occurs as a cost C in DM or an innovation probability a in SM increases. In addition, we also investigate how the propagation processes affect the transition nature. From analytical calculations and numerical simulations m is shown to satisfy the scaling relation m~1{f (C=N) for DM with the number of agents N. In contrast, m in SM scales as m~1{f (a a N).
# Introduction
Transitions are ubiquitous in human history and in scientific activities as well as in physical systems. Human history of civilizations has qualitatively distinguishable periods from stoneage to contemporary civilizations, which depend on dominating themes such as philosophy, art, and technology. In scientific activities such dominating themes correspond to disparate prevailing ideas or concepts such as chaos, complexity, nanophysics, and string theory, which are generally called as paradigms. Tomas Kuhn said that the successive transition from one paradigm to another via revolution is the usual developmental pattern of mature science. This paradigm shift is also very similar to the adoption of a new discrete technology level. Examples of such technological levels are operating system versions as Linux distributions and versions of recently-popular smart phones.
To describe the appearance and disappearance of those paradigms, various models [bib_ref] Statistical Physics of social dynamics, Castellano [/bib_ref] [bib_ref] Reinforced communication and social navigation generate groups in model networks, Rosvall [/bib_ref] [bib_ref] Opinion evolution in closed community, Sznajd-Weron [/bib_ref] [bib_ref] Generalization to square lattice of Sznajd sociophysics model, Stauffer [/bib_ref] [bib_ref] Dynamics of Majority Rule in Two-State Interacting Spin Sys-tems, Krapivsky [/bib_ref] [bib_ref] Statistical methods applied to study of opinion formation models: a brief overview..., Bordogna [/bib_ref] were suggested. In appearance of a paradigm, the propagation of an idea through the social interaction between individuals is essential. To study how the information flow affects the formation of a group sharing common interest in social networks, communication-navigation model with local memory was studied [bib_ref] Reinforced communication and social navigation generate groups in model networks, Rosvall [/bib_ref]. However this model [bib_ref] Reinforced communication and social navigation generate groups in model networks, Rosvall [/bib_ref] did not consider the innovation process in which a new paradigm appears and old paradigms disappears. On the other hand, due to its simplicity, two-state interacting spin systems were also widely used to investigate how an existing idea evolves into a dominating theme through the interactions between agents [bib_ref] Opinion evolution in closed community, Sznajd-Weron [/bib_ref] [bib_ref] Generalization to square lattice of Sznajd sociophysics model, Stauffer [/bib_ref] [bib_ref] Dynamics of Majority Rule in Two-State Interacting Spin Sys-tems, Krapivsky [/bib_ref] [bib_ref] Statistical methods applied to study of opinion formation models: a brief overview..., Bordogna [/bib_ref]. Those models were generally focused on the emergence of a single paradigm through social consensus. But, they did not correctly capture the complex dynamical features of paradigm shifts, such as invention of new ideas, competition between ideas, propagation of ideas against the competition to form a new global paradigm, and decline of already existing ones.
Recently, an interesting model was suggested by to explain such complex dynamical properties of the paradigm shift [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref]. In the Bornholdt model (BM) [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref] , two essential processes for the paradigm shift were suggested. The two essential processes were never considered simultaneously in previous studies [bib_ref] Statistical Physics of social dynamics, Castellano [/bib_ref] [bib_ref] Reinforced communication and social navigation generate groups in model networks, Rosvall [/bib_ref] [bib_ref] Opinion evolution in closed community, Sznajd-Weron [/bib_ref] [bib_ref] Generalization to square lattice of Sznajd sociophysics model, Stauffer [/bib_ref] [bib_ref] Dynamics of Majority Rule in Two-State Interacting Spin Sys-tems, Krapivsky [/bib_ref] [bib_ref] Statistical methods applied to study of opinion formation models: a brief overview..., Bordogna [/bib_ref]. In BM each agent i resides on a node of a graph and is assigned an integer r i . The number plays the role of a particular idea or concept. Then, at any time step the two essential processes are attempted: (i) With probability a a randomly selected agent k is assigned a new random integer which does not appear anywhere else in the system. Thus a represents the ''innovation'' rate. (ii) An agent i is randomly chosen. Then one of the nearest neighbors j to the agent i is randomly selected. Denoting by N the total number of agents in the system and by n j the total number of agents with integer value equal to that of j, the integer value of the agent i is changed into that of its neighbor j with probability n j =N, provided that i never assumed that particular integer value before. In case it had, then no update is made. The process (i) is the innovation process, in which a new idea or a new paradigm to the whole system is introduced successively. Thus the number of ideas is not limited in BM. The process (ii) is the propagation process. In the propagation process, the memory effect that any agent does not accept any idea experienced before is imposed. This memory effect was also an essential feature of BM. The memory effect was argued to be originated from the part of cultural or scientific activity where people are on an ongoing hunt for new ideas and ideally never return to exactly their old positions [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref]. By the numerical study of BM on a square lattice Bornhodlt et al. showed the existence of the ordered phase with a globally dominant paradigm for the small innovation probability a [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref]. In this ordered phase the pattern of sudden emergence and slow decline of a dominant paradigm repeats again and again. The epochal things of BM [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref] are the innovation process and the memory effect.
Even though Bornhodlt et al. showed the existence of a dominant paradigm for small innovation rate a, it is still an open fundamental question when and how this ordered phase disappears as a gets larger or approaches to 1. The clear understanding of the transition nature provides more profound physical insight to understand fundamental properties of the system. Furthermore, the propagation of an idea generally occurs successively and continuously or has avalanches as can be seen from the spread of an idea through community networks, social network services and mass communications. Nevertheless the propagation of a paradigm in the process (ii) of BM [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref] was only considered to occur locally without avalanche. In addition, the propagation can occur deterministically as the difference (or the gap) of ideas (or technological levels) between two interacting agents grows [bib_ref] Self-organized evolution in a socioeconomic environment, Arenas [/bib_ref] [bib_ref] Modeling diffusion of innovations in a social network, Guardiola [/bib_ref] [bib_ref] Morphology of technological levels in an innovation propagation model, Kim [/bib_ref] , whereas the propagation in BM was only considered probabilistically and stochastically.
Therefore, to answer the question when and how the transition occurs from the ordered phase in which a dominant paradigm exists to the disordered phase without any dominant paradigm, and to investigate how the details of propagation process affect the paradigm shifts, we provide two realistic and generalized models for paradigm shifts, deterministic propagation model (DM) and stochastic propagation model (SM). In our models, DM and SM, the innovation process is identical to the process (i) of BM [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref]. DM and SM also have the same memory effect as BM. The essential difference between our models and BM is in the details of the propagation process. The details of the propagation are very important in two senses. The first is that the propagation process is the essential mechanism to decide the pattern of sudden emergence and slow decline of a globally dominant paradigm in the system. The second is that the propagation process in a model must reflect the real propagation process in the existing system. The real propagation process should have successive and continuous propagations, i.e., the avalanche. In our models, DM and SM, the propagation process has the avalanche, whereas BM [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref] has no avalanche. The real propagation process should also be decided either by the difference of ideas or probabilistically. Therefore we consider two models in this paper. In DM the propagation of an idea between the interacting pair of agents, i and j, occurs only if the difference of ideas Dr i {r j D §C. In SM the propagation of an idea between the interacting pair occurs probabilistically and stochastically as in BM.
By defining the order parameter, m, based on the diversity of ideas, D, we analytically show that the disappearance of a dominant paradigm can be mapped into the thermal orderdisorder transition in physical systems. In DM it is shown that m satisfies the scaling relation m~1{f (C=N). In contrast, m in SM is shown to follow the relation m~1{f (a a N), where a is the innovation probability. Here f (x) is a scaling function satisfying f (x)*x b for x%1 and f (x)~1 for x&1. m in BM is also proved to satisfy the same scaling relation as m in SM. Therefore, the transition threshold C à in DM scales as C Ã^N , whereas the transition probability a à in both SM and BM scales as a à *N {1=a . The exponents a and b depend both on the propagation mechanism and on the underlying interaction topology of agents. Therefore, from this work, we first provide a standard theoretical framework to understand phase transitions and related phenomena in the paradigm shifts.
# Analysis
To be specific, let's assume that each agent resides on a node of a certain graph and a pair of nodes connected by a link in the graph is an interacting pair of agents. At a given time t each agent i has a non-negative integer r i (t), which represents a particular idea or a technological level. In the innovation models of references [bib_ref] Self-organized evolution in a socioeconomic environment, Arenas [/bib_ref] [bib_ref] Modeling diffusion of innovations in a social network, Guardiola [/bib_ref] [bib_ref] Morphology of technological levels in an innovation propagation model, Kim [/bib_ref] , the technological level changes continuously or takes rational number. In contrast ideas in this paper take only non-negative integers as in BM [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref]. Initially all agents in the system are assumed to have no idea, i.e., r i (0)~0 for any i. Then at time tz1, a randomly selected agent i takes an innovation process with the probability a or propagates his idea to other agents with the probability 1{a. In the process of the paradigm shift, innovation naturally occurs occasionally, whereas propagation occurs frequently and rapidly. Thus, the innovation probability a should naturally be very small. In the innovation process at t, r i (t) of a randomly-chosen agent i takes a discrete jump to be the smallest positive integer which has not been experienced by any agent in the whole system until the time t [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref]. The propagation process can be a deterministic and rational process or a stochastic and contingent process.
To analyze phase transitions from the ordered phase to the disordered phase of paradigm shift models, we should first understand the model with a~1, which we call the random innovation model (RIM). Since innovation processes in DM and SM are the same, DM and SM are reduced to RIM at a~1. RIM, in which only innovation processes occur without propagation process, cannot have a dominant paradigm any time and is always in the disordered phase. In RIM one can exactly calculate the diversity D(t), which is defined as D(t):
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi vr 2 (t)w{vr(t)w 2 p , where vr u (t)w:v½ P i r u i (t)=Nw and v:::w means the average over all possible configurations of fr 1 (t),r 2 (t),:::,r N (t)g. In RIM, a randomly selected agent i at the time t changes his idea into t as r i (t)~t. Let's denote p:1=N and q:1{1=N, where p is the selection probability of a particular agent among N agents. Then the probability P t (r) that an agent has the idea r at t is written as P t (r)~pq t{r for 0vrƒt and P t (0)~q t . Thus we get
[formula] D 2 (t)~1 {q 2t p 2 { 1{2q 2t zq t z2tq t p zq t z2tq t {q 2 t:ð1Þ [/formula]
In the limit N??,
[formula] D(t)~N ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi 1{e {2t=N {2(t=N)e {t=N q :ð2Þ [/formula]
Equation (2) has been confirmed by numerical simulation. In the steady state (or t??), D(t??):D(?)~N. D(?)~N corresponds to the disordered phase for a?1 for paradigm shift models. Thus we take the order parameter m for the phase transition of the paradigm shift models in the steady state as m:1{D(?)=N. Then m~0 for the disordered phase and m~1 for completely ordered phase with D(?)~0, in which all the agents have one same idea. We now consider two different paradigm shift models based on specifics of propagation process.
# Results
## Deterministic propagation model
When a new idea (or a new technological level) is created, one normally decides to adopt the new idea by comparing the new idea with his present idea. If the difference between the new idea and the present idea is small, the adoption of the new idea hardly happens. The larger the difference becomes, the more easily one adopts the idea. Therefore the propagation process can depend on the difference in the ideas or the cost [bib_ref] Self-organized evolution in a socioeconomic environment, Arenas [/bib_ref] [bib_ref] Modeling diffusion of innovations in a social network, Guardiola [/bib_ref] [bib_ref] Morphology of technological levels in an innovation propagation model, Kim [/bib_ref]. In this sense, the deterministic propagation model (DM) in which the propagation Phase Transitions in Paradigm Shift Models PLOS ONE | www.plosone.org process is deterministically controlled by the cost is defined in the following way. In the propagation process of DM, a randomly selected agent i propagates his idea r i (t) to each nearest neighbor j, i.e., r j (tz1)~r i (t) at the time tz1, only if r i (t){r j (t) §C. Here C is a constant which represents a propagation cost to adopt a new paradigm. Then the propagation process triggers an avalanche; i.e., if r j (tz1) is updated, then repeat the same propagation process for all nearest neighbors of j. This propagation process is repeated until all the nearest neighbor pairs satisfy the inequality
[formula] Dr j (t){r i (t)DvC. [/formula]
In DM, D(?) depends only on C for small a as shown in [fig_ref] Figure 1 B: Figure 2B, C and D [/fig_ref] A, because a controls only the time t s taken for the system to arrive the steady state as t s^1 =a. This result physically means that the system is in the steady state if the mean number of innovations, at, satisfies at&C and the physical properties of the steady state depend only on C.
First we consider DM on the complete graph (CG). Each agent on CG is a nearest neighbor of all the other agents. Therefore one propagation process from a randomly-selected node makes propagation tries to all the other agents. Let's think a steady state configuration that ideas in the system spread in an integer set fr min ,r min z1,:::,r max (~')g when the '-th innovation process happens. In the average sense r max~'~a t if the '-th innovation happens at t. If a is small enough, there should exist a propagation process initiated from an agent with r~' among the many propagation processes before the ('z1)-th innovation process occurs. The propagation process from the agent with r~' makes the configuration with r i
By applying the recursion relations (3) C times, we obtain
[formula] P ' (r)~pq '{r zq C P ('{C) (r{C) pq '{r zpq '{rzC zq 2C P ('{2C) (r{2C) :::~P ½r=C i~0 pq '{rziC~p q '{r 1{q (½r=Cz1)C 1{q C ,ð4Þ [/formula]
where ½r=C is an maximal integer which is not greater than r=C.
In the limit t?? or '??, P ' (r)~pq '{r =(1{q C ) and D(?) is written as
[formula] D(?)~N ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi 1{ 1 N { C N 2 q C (1{q C ) 2 sð5Þ [/formula]
In the large N limit, m thus satisfies
[formula] m~1{g(C=N) g(x)~ffi ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1{ x 2 cosech x 2 h i 2 r ! :ð6Þ [/formula]
Even though Equation (6) was derived under the physical assumption that a is very small, Equation (6) agrees very well with the simulation results for quite large a or for aƒ0:5 as shown in [fig_ref] Figure 2: Analytic and simulation results of DM [/fig_ref] A. The ordered state of DM on CG has a peculiar physical property. Since P ' (r)^(1{('{r)=N)=C^1=C for C%N, there doesn't exist a unique dominant idea, but C ideas are nearly equally probable in the steady state. This peculiar ordered state comes from the global connectivity of CG. In the sense that DM naturally regards ideas within the difference drvC as the same one, the ordered state on CG is physically plausible and understandable.
In contrast, there exists a unique dominating idea in DM on other graphs with local connectivity for C%N as shown in Now we want to show how the configuration with the (kz1)-th dominating idea r d (kz1) happens. As shown in [fig_ref] Figure 1 B: Figure 2B, C and D [/fig_ref] B, the nodes (or sites) with r d k form a macroscopic percolation cluster through the links (or bonds) of the graph and the nodes with r=r d k form only isolated microscopic clusters. Thus the propagation process which changes the dominating idea happens only through the macroscopic percolation cluster. Therefore the configuration with the r d (kz1) does not happen until the idea r~r d k zC appears in the system. After the idea r~r d k zC appears, subsequent propagation processes through the macroscopic cluster which make the configuration with r d (kz1) (~r d k zC) appear before the next innovation process happens. The configurations with r d (kz1)~r d k zCz1,r d k zCz2 and … are also possible, but the probabilities that these exceptional configurations happen are nearly negligible if a is small and N is large. So we neglect these exceptional configurations in the subsequent calculations. In the configuration with r d (kz1) (~r d k zC), the ideas in the system spread in the set fr d (kz1) {Cz1,r d (kz1) {Cz2,:::,r d (kz1) g. Then before the configuration with r d (kz2) appears, the configuration of the system can evolve into one in which the ideas spread in the set fr d (kz1) {Cz1,r d (kz1) {Cz2,:::,r d (kz1) ,:::,r d (kz1) zn I g with n I vC. Here n I is the number of the innovations which occur before the configuration with r d (kz2) appears. Generally the system in the steady state has a configuration with the ideas spread in the set fr d {Cz1,r d {Cz2,:::,r d ,:::,r d zn I g.
Now we consider the probability P(r) that an agent has an idea r in the steady state. Clearly P(r)~0 for rƒr d {C and rwr d zn I . Furthermore, in the large N limit P(r) is expected to satisfy P(r)^p~1=N for r[fr d {Cz1,r d {Cz2,:::,r d {1,r d z1,:::, r d zn I g, because an idea in this set is originated from an innovation process. From the continuum limit vr u wÐ drr u P(r), we get D 2 nI as
Since n I is equally probable to be any integer in the set f0,1,:::,C{1g,
[formula] D 2 (?)~1 C ð C 0 1 3N (C 3 zn 3 I ){ 1 4N 2 (n 4 I zC 4 {2n 2 I C 2 ) dn Ĩ N 2 5 12 C N 3 { 2 15 C N 4 " # :ð8Þ [/formula]
Therefore, m for C%N satisfies m~1{ ffiffiffiffiffiffiffiffiffiffi 5=12 p (C=N) 3=2 . For C&N, DM reduces to RIM and m~0. Thus m satisfies the scaling relation
[formula] m~1{f (C=N),ð9Þ [/formula]
where f (x)*x b with b~3=2 for x%1 and f (x)~1 for x&1. On CG the same scaling relation with b~1 holds for m.
To confirm the scaling relation (9) on the graphs with local connectivity, DM is studied by simulations on various graphs. The graphs used in this paper are a scale-free network with the degree exponent c~2:5 [bib_ref] Universal Behavior of Load Distribution in Scale-Free Networks, Goh [/bib_ref] , and an Erdös-Rényi type random network, and a two-dimensional square lattice. To accord with the square lattice, the mean degree vkw of the scale-free and random networks is set as vkw~4. The simulation data of m on each graph in [fig_ref] Figure 2: Analytic and simulation results of DM [/fig_ref] are obtained by averaging over at least 1000 realizations. The scaling relation of m with b~3=2 or Equation (9) is confirmed by simulations on the random network and the square lattice as shown in [fig_ref] Figure 2: Analytic and simulation results of DM [/fig_ref] In contrast, on a scale-free network with degree exponent c~2:5, the scaling relation with b~1:20(3) is obtained [fig_ref] Figure 2: Analytic and simulation results of DM [/fig_ref]. The deviation of the exponent b from 3/2 on the scale-free network is probably explained from the hub effect of the scale-free networks with cv3, which provides an aspect of global connectivity. Thus C at which the phase transition occurs, C Ã , scales as C Ã^N on arbitrary graph.
Even though the scaling relationwas derived under the physical assumption that a is very small, we have confirmed that Equation (9) agrees very well with simulation results for quite large a or for aƒ0:5 on the square lattice and the random network as on CG [fig_ref] Figure 2: Analytic and simulation results of DM [/fig_ref].
## Stochastic propagation model
We now consider the stochastic propagation model (SM) in which the propagation process occurs probabilistically and stochastically. In SM, the feature that a minority idea is more difficult to be adopted than a more widespread idea [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref] is considered. Therefore, the propagation process in SM is defined in the following way. If a propagation try is taken at a given time t with the probability 1{a, first a site i with the idea r i (t) is randomly selected. Then with the probability n i =N the propagation process starting from the site i occurs. Here n i is the number of agents in the system which have the same idea with r i (t). If the propagation process happens, then the ideas of all nearest neighbors of i are simultaneously made to be equal to r i (t), except the ideas of neighbors who have experienced r i (t) before. In addition, all the neighbors whose ideas are changed also propagate the idea r i (t) to all of their nearest neighbors in the same manner with the updated probability n i =N, because n i increases as propagations continue. The propagations continue until the propagations are terminated by the probability (1{n i =N) or all the agents in the system are tried to be propagated. Therefore, the propagation process of SM also has the avalanche and an idea r can spread to the whole system for one time step. Moreover, as we shall see, the scaling properties of SM on graphs with local connectivity are the same as those of BM [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref].
m of SM on CG is analytically calculable, because an idea propagates to the whole system by single propagation process. For the calculation one should understand the time evolution of configurations in SM on CG. The schematic diagram for the evolution is shown in [fig_ref] Figure 3: Schematic diagram for the evolution of configurations in SM on CG [/fig_ref]. For the explanation of the evolution, let's define the maximal r, r max k , appeared in the system until the time t k at which the k-th dominant idea r d k appears. A typical configuration at t(t (k{1) vtvt k ) is one with r i [fr d (k{1) ,r max (k{1) z1,r max (k{1) z2,:::g (see [fig_ref] Figure 3: Schematic diagram for the evolution of configurations in SM on CG [/fig_ref]. Then the next propagation process drives this configuration into one with all r i~r d k , where r d k [fr max (k{1) z1,r max (k{1) z2,:::,r max k g (see the process (a) in [fig_ref] Figure 3: Schematic diagram for the evolution of configurations in SM on CG [/fig_ref]. Then successive innovation processes make the configuration with r i [fr d k ,r max k z1,r max k z2,:::g (see the process (b) in [fig_ref] Figure 3: Schematic diagram for the evolution of configurations in SM on CG [/fig_ref]. Note that the propagation process which cannot be executed by the probability (1{a) [fig_ref] Figure 1 B: Figure 2B, C and D [/fig_ref] does not change the configuration of system (see the process (c) in [fig_ref] Figure 3: Schematic diagram for the evolution of configurations in SM on CG [/fig_ref]. Then the (kz1)-th propagation process drives this configuration into one with all r i~r d (kz1) with r d (kz1) [fr max k z1,r max k z2,:::,r max (kz1) g (see the process (d) in [fig_ref] Figure 3: Schematic diagram for the evolution of configurations in SM on CG [/fig_ref]. Then an innovation process drives the configuration with r i [fr d (kz1) ,r max (kz1) z1g (see the process (e) in [fig_ref] Figure 3: Schematic diagram for the evolution of configurations in SM on CG [/fig_ref]. In the steady state of SM, this evolution pattern is repeated again and again. Thus now we analytically calculate D(?) or m of SM based on this evolution pattern. Now we consider the probability P (t'zt k ) (r) that an agent has the idea r at t~(t'zt k ) with 0ƒt'vt (kz1) {t k . In the average sense, the number of innovations occurring from t~t k to t~(t'zt k ) is at'. Then at t~(t'zt k ), similar to RIM, P (t'zt k ) (r) is written as
[formula] P (t 0 zt k ) (r)~q [/formula]
Thus from vr u w~(r d k ) u P (t'ztk) (r d k )z P r max k zat' r~r max k z1 r u P (t'ztk) (r), we get D(t'zt k ) of the configuration with A is a configuration with r d (k{1)~7 and r max (k{1)~9 . The next propagation process (a), (or the k-th propagation) at t~t k , changes A into B with all r i~r d k~1 0 and r max k~1 3. Successive innovation processes (b) change B into C with r i [fr d k ,r max k z1,r max k z 2,r max k z3,r max k z4g with r d k~1 0 and r max k~1 3. The propagation process (c), which cannot be executed by the probability (1{a)(1{1=N), leaves C as it is. The propagation process (d) at t (kz1) initiated from an agent with r~15 drives C into D with all r i~r d (kz1)~1 5 and r max (kz1)~1 7. An innovation process (e) drives D into E with r i [fr d (kz1) ,r max (kz1) z1g (r d (kz1)~1 5 and r max (kz1)~1 7). doi:10.1371/journal.pone.0070928.g003 r i [fr d k ,r max k z1,:::,r max k zat'g for large N as
[formula] D(t'zt k )~N 1{ exp ({at'=N) exp ({at'=N){2(at'=N) ð ½ z2dr(at'{(N{1)(1{ exp ({at'=N)))=N zdr 2 (1{ exp ({at'=N))=N 2 ÁÃ1 2 ,ð11Þ [/formula]
with q at'~( 1{1=N) at'~e xp ({at'=N) and dr:r max k {r d k . D(?) is thus written as
[formula] D(?)~P ? dr~0 P(dr) P ? t'~0 S(t')D(t'zt k ) P ? t'~0 S(t') ,ð12Þ [/formula]
where S(t') is the probability that no propagation processes happen from t k until (t'zt k ) and P(dr) with dr~r max k {r d k is the probability that a configuration with r i [fr d k ,r max k z1g occurs at the very next innovation process after t k or at (t k za {1 ) in the average sense (see [fig_ref] Figure 3: Schematic diagram for the evolution of configurations in SM on CG [/fig_ref]. Now we calculate S(t'). The probability that a propagation process at (t'zt k ) can be executed is ½(1{a)n i =N P rwr max k
[formula] P (t'zt k ) (r)~½(1{a)=N½1{P (t'zt k ) (r d k ) (1{a)(1{q at' )=N with n i~1 and P (t'zt k ) (r d k )~q at' . Then S(t')~S(t'{1) 1{(1{a)(1{ exp ({at'=N))=N ½ [/formula]
in the large N limit. By taking the continuum time limit,
[formula] dS(t') dt'~{ S(t') 1{a N (1{ exp ({at'=N)) ! :ð13Þ [/formula]
Thus we get
[formula] S(t')~exp { 1{a N t'z 1{a a (1{ exp ({at'=N)) !ð14Þ [/formula]
from S(0)~1. Now we want to calculate P(dr). At (t k {1),
[formula] P (t k {1) (r d k ) [/formula]
that an agent has the idea r d k ([fr max (k{1) z1,:::,r max k g) is
[formula] P (t k {1) (r d k )~pq r max k {r d k~p q dr . [/formula]
Then the propagation process to make r d k the k-th dominant idea occurs with the probability ½(1{a)=NP (t k {1) (r d k )(~(1{a)p 2 q dr ). Since the probability for
[formula] r max k {(r max (k{1) z1) §dr is P a(t k {t (k{1) )wdr S(t k {t (k{1) {1) [/formula]
, P(dr) can be written as
[formula] P(dr)~(1{a)p 2 q dr X ? t' §dr=a S(t'):ð15Þ [/formula]
From Equations (11), [bib_ref] Morphology of technological levels in an innovation propagation model, Kim [/bib_ref] , (14) and (15), m(~1{D(?)=N) on CG can be calculated through exact enumeration. The results of the exact enumerations are shown in [fig_ref] Figure 4: Analytic and simulation results of SM [/fig_ref] A. For small innovation probability a, D(?) or m can be analytically calculable. For at'%N, dr=N%1. Thus
[formula] D(t'zt k )*N ffiffiffiffiffiffiffi ffi 1=3 p (at'=N) 3=2 ,ð16ÞS(t')^exp ({t' 2 a(1{a)=2N 2 ),ð17ÞÐ ? 0 S(t')D(t'zt k )dt' Ð ? 0 S(t')dt' *Na 3=4ð19Þ [/formula]
for at'%N, and
[formula] m(a)~1{a 3=4 :ð20Þ [/formula]
We also confirm that Equation (20) holds even for quite large a or for a close to 1 by comparing Equation (20) with the results of the exact enumerations as shown in [fig_ref] Figure 4: Analytic and simulation results of SM [/fig_ref] A. This result means that there always exists a dominating idea or the global paradigm on CG if av1.
On the graphs only with local connectivity, the analytic approach as on CG to SM is hardly possible. Instead simulations are carried out. The simulation results on various graphs with local connectivity show that D(?) satisfies the scaling ansatz D(?)~h(a d N) very well. As shown in [fig_ref] Figure 4: Analytic and simulation results of SM [/fig_ref] , m satisfies the scaling function similar to that of DM as increases as the global connectivity decreases. The scaling behavior of SM on the random network is nearly equal to that on the square lattice. This result means that the scaling behavior hardly depends on the dimensionality of the graph, but depends on the connectivity. We also study m of BM [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref]. In BM, a randomly selected agent i tries to propagate his idea to a randomly chosen nearest neighbor with the probability n i =N. No further propagation process is attempted in BM or BM does not allow the avalanche in a propagation process. Since the propagation in BM is local, it is difficult to treat the model analytically even on CG. Thus BM is studied numerically. From the simulations we confirm the same scaling behavior m~1{f (a a N) with a~1:10(2) and b~1:12(3) on any graph, especially on CG. The scaling behavior of BM on any graph is the same as those of SM on the square lattice. BM has only local propagation process on any graph and does not use the connectivity of large scale or the global connectivity, even on CG. Therefore the scaling properties of BM are irrelevant to the dimensionality or the connectivity of the graph. SM on the square lattice physically has only local avalanches, and thus the scaling properties of SM on the square lattice are the same as those of BM. a à of BM also scales as a à *N {1=a with a^1:1.
# Conclusion and discussion
We introduce two paradigm shift mechanisms as the deterministic propagation model (DM) and the stochastic propagation model (SM). Both models have the memory effect that an agent never returns to any paradigm experienced before by any process as BM. In both models there commonly exists the innovation process, which occurs with the probability a. With 1{a the propagation process occurs. Both DM and SM have the avalanche in the propagation process. In DM, the propagation process is controlled by the cost C, which represents the idea difference or resistance to make one adopt a new idea. In contrast, the propagation process of SM occurs probabilistically and stochastically by considering the feature that that a minority idea has more difficulty for adoption than a more widespread idea.
To analyze phase transitions from the ordered phase with a dominant paradigm to the disordered phase in paradigm shift models, the disordered phase is exactly defined by using the random innovation model (RIM) in which the diversity of ideas D~N. By defining the order parameter, m as m:1{D(?)=N, we first provide a novel theoretical framework in which transition in paradigm shift models is analyzed quantitatively by applying the scaling theory of statistical physics for the analysis of the traditional thermal order-disorder transition. In DM m of the steady state satisfies the scaling relation m~1{f (C=N) on any graphs. In contrast, m in SM follows the scaling relation m~1{f (a a N).
Here f (x) is a common scaling function satisfying f (x)*x b for x%1 and f (x)~1 for x&1. m of BM [bib_ref] Emergence and Decline of Scientific Paradigms, Bornholdt [/bib_ref] on any graph is also proved to satisfy the same scaling relation as m of SM on the square lattice. Thus, in DM the transition threshold C à scales as C Ã^N and the transition probability in both SM and BM scales as a à *N {1=a . The exponents a and b depend both on the models and on the underlying interaction topologies.
Thus this paper suggests a novel theoretical method based on the scaling theory of the statistical physics to understand the phase transitions in social systems such as paradigm shifts quantitatively. The resultant scaling relations in DM and SM also quantitatively and exactly show that there cannot exist a dominant paradigm if innovations happen too frequently or the resistance to make one adopt a new idea becomes large in the systems with finite N. The deterministic and stochastic propagations coexist in real world. Thus, it would be an interesting open question how the nature of the phase transition and the dynamical properties in paradigm shifts are affected by the coexistence of two processes. Furthermore, it would also be very interesting to apply the paradigm shift models (DM, SM, and BM) to the analysis of the real data for the paradigm shifts or the technological level shifts. One of the such real data should be the adoption patterns of operating system versions or versions of recently-popular smart phones. Another interesting future study would be to investigate modified versions of the innovation models in which the technological level changes continuously or takes rational number [bib_ref] Self-organized evolution in a socioeconomic environment, Arenas [/bib_ref] [bib_ref] Modeling diffusion of innovations in a social network, Guardiola [/bib_ref] [bib_ref] Morphology of technological levels in an innovation propagation model, Kim [/bib_ref]. Such innovation models [bib_ref] Self-organized evolution in a socioeconomic environment, Arenas [/bib_ref] [bib_ref] Modeling diffusion of innovations in a social network, Guardiola [/bib_ref] [bib_ref] Morphology of technological levels in an innovation propagation model, Kim [/bib_ref] only considers the deterministic propagation process. As emphasized previously, any propagation should have both stochastic and deterministic aspects. Therefore it would also be very interesting to investigate the innovation model with continuously varying technological levels and the combination of deterministic and stochastic propagation processes.
# Author contributions
[fig] Figure 1 B: Figure 2B, C and D. Thus we now want to analytically show the existence of the ordered state with a dominating idea on the graphs with local connectivity. In DM any nearest neighbor pair vijw of agents should satisfy the condition Dr i {r j DvC after a propagation process. Let's first think about the configuration with the k-th dominating macroscopic idea r d k [fr min ,r min z1,:::,r max g. [/fig]
[fig] Figure 1: Scaling plot of D(t) and a snapshot in DM. (A) Scaling plot of D(t) against at of DM on a square lattice with N~2 12 and C~82. Inset: plot of D(t) against t. (B) A snapshot of a steady state configuration of DM on the square lattice with the size 32|32. Black dots denote agents with a dominant idea r d . White dots denotes those with ideas different from r d . doi:10.1371/journal.pone.0070928.g001 [/fig]
[fig] Figure 2: Analytic and simulation results of DM. Scaling plots of m against C=N of DM (A) on the complete graph with N~8:0|10 3 , (B) on a scale-free network (C) on a random network and (D) on a square lattice. Curves in the figures show the analytic results Equation (6) and Equation (9). All the simulation data in B, C and D are obtained by use of a~10 {2 . Inset of D shows the plots of m for various N against C. doi:10.1371/journal.pone.0070928.g002 Phase Transitions in Paradigm Shift Models PLOS ONE | www.plosone.org [/fig]
[fig] 0: {(r{r max k ) for r[fr max k z1,r max k z2,:::,r max k zat [/fig]
[fig] Figure 3: Schematic diagram for the evolution of configurations in SM on CG. [/fig]
[fig] Figure 4: Analytic and simulation results of SM. (A) Plot of m against a of SM on the complete graph. The data both from the exact enumerations and simulations are shown. The curve represents the analytic result m~1{a 3=4 . (B-D) Scaling plots of the simulation data for m of SM against a a N on the scale-free network (B), on the random network (C) and on the square lattice (D [/fig]
|
Cognitive Spare Capacity and Speech Communication: A Narrative Overview
Background noise can make speech communication tiring and cognitively taxing, especially for individuals with hearing impairment. It is now well established that better working memory capacity is associated with better ability to understand speech under adverse conditions as well as better ability to benefit from the advanced signal processing in modern hearing aids. Recent work has shown that although such processing cannot overcome hearing handicap, it can increase cognitive spare capacity, that is, the ability to engage in higher level processing of speech. This paper surveys recent work on cognitive spare capacity and suggests new avenues of investigation.
# Introduction
Speech is the main mode of communication for most people. If speech understanding is compromised by noise or hearing impairment, communication may become harder, leading to limitations in social participation. Technical compensation is available in the form of hearing aids. However, although the amplification provided by hearing aids can improve speech understanding in quiet, persons with hearing impairment still have disproportionately large difficulties understanding speech in noise. One of the reasons for this may be that when the cognitive resources required for speech comprehension are engaged in the lower level processes of deciphering the signal, fewer resources may be available for higher level language processing. In other words, cognitive spare capacity is reduced.
## Speech comprehension.
Speech comprehension requires the auditory ability to hear the signal and the cognitive ability to relate this information to the existing knowledge stored in semantic long-term memory [bib_ref] Candidature for and delivery of audiological services: special needs of older people, Kiessling [/bib_ref] [bib_ref] Effects of age on auditory and cognitive processing: implications for hearing aid..., Pichora-Fuller [/bib_ref]. The role of cognition in speech comprehension is reflected in the hierarchical nature of its cortical representation [bib_ref] The hemispheric lateralization of speech processing depends on what "speech" is: a..., Peelle [/bib_ref] [bib_ref] The neuroanatomical and functional organization of speech perception, Scott [/bib_ref].
Speech processing engages a clearly defined cortical network involving the classical language areas in the left inferior frontal cortex and superior temporal gyrus [bib_ref] The hemispheric lateralization of speech processing depends on what "speech" is: a..., Peelle [/bib_ref] [bib_ref] The neuroanatomical and functional organization of speech perception, Scott [/bib_ref]. The primary auditory cortex is sensitive to most sounds and is the first cortical region to be activated during speech perception [bib_ref] The neuroanatomical and functional organization of speech perception, Scott [/bib_ref]. Listening to words activates the middle and superior temporal gyri bilaterally and listening to sentences engages regions involved in processing semantics and syntax in the left prefrontal cortex [bib_ref] The hemispheric lateralization of speech processing depends on what "speech" is: a..., Peelle [/bib_ref]. It has been possible to trace the pathways linking these regions by using animal models [bib_ref] The neuroanatomical and functional organization of speech perception, Scott [/bib_ref] [bib_ref] Subdivisions of auditory cortex and processing streams in primates, Kaas [/bib_ref] [bib_ref] Maps and streams in the auditory cortex: nonhuman primates illuminate human speech..., Rauschecker [/bib_ref]. These pathways represent different functional streams that take either a ventral route through superior temporal regions to ventrolateral prefrontal cortex or a dorsal route through posterior parietal cortex and dorsolateral prefrontal cortex [bib_ref] Maps and streams in the auditory cortex: nonhuman primates illuminate human speech..., Rauschecker [/bib_ref] [bib_ref] The cortical organization of speech processing, Hickok [/bib_ref]. One ventral route seems to deal more with conceptual or semantic processing, while there is a dorsal route that is more related to phonological or articulatory processing [bib_ref] Maps and streams in the auditory cortex: nonhuman primates illuminate human speech..., Rauschecker [/bib_ref] [bib_ref] The cortical organization of speech processing, Hickok [/bib_ref]. Ventral and dorsal routes for syntactic processing have also been proposed [bib_ref] The language network, Friederici [/bib_ref].
2 BioMed Research International 1.2. Hearing Impairment. Around 25% of the population in developed countries has a hearing impairment severe enough to interfere with speech communication. Hearing sensitivity decreases with age such that although only about 2% of individuals in their early twenties have a hearing loss, the prevalence of significant hearing impairment is 40-45% in persons over the age of 65 and exceeds 83% in persons over the age of 70 [bib_ref] Epidemiology of aging and hearing loss related to other chronic illnesses, Davis [/bib_ref] [bib_ref] Hearing loss and aging: new research findings and clinical implications, Gordon-Salant [/bib_ref]. Hearing difficulties are associated with long-term absence from work in the working age population [bib_ref] Hearing ability in working life and its relationship with sick leave and..., Nachtegaal [/bib_ref] [bib_ref] Self-reported hearing difficulties, main income sources, and socio-economic status; a cross-sectional population-based..., Pierre [/bib_ref] and loneliness in the older population [bib_ref] Hearing status in older persons: a significant determinant of depression and loneliness?..., Pronk [/bib_ref]. Further, individuals with better cognitive abilities report more hearing difficulties [bib_ref] Relationships between self-report and cognitive measures of hearing aid outcome, Ng [/bib_ref] [bib_ref] Cognitive abilities relate to self-reported hearing disability, Zekveld [/bib_ref] , possibly because they have higher expectations of their communication. Even moderate degrees of hearing impairment lead to decrease in neural activity during speech processing and may contribute to grey matter loss in primary auditory cortex [bib_ref] Auditory cortex signs of age-related hearing loss, Eckert [/bib_ref] [bib_ref] Hearing loss in older adults affects neural systems supporting speech comprehension, Peelle [/bib_ref].
Types of hearing loss are traditionally categorized according to site of lesion: impairment of sound transmission in the external or middle ear is referred to as conductive hearing loss, while other types of hearing loss are referred to as sensorineural. Sensorineural hearing loss can be further subdivided into sensory loss, resulting from impairment of cochlear function, retrocochlear loss, resulting from impairments relating to conduction in the auditory nerve or brainstem, and central losses, resulting from impairments in cortical processing of the auditory signal. Sensorineural hearing loss is the major diagnostic category and includes age-related hearing loss or presbyacusis. These categories are relatively coarse and it has been suggested that they may be inadequate for pinpointing the contribution of hearing loss to communication difficulties under adverse listening conditions [bib_ref] The Signal-Cognition interface: interactions between degraded auditory signals and cognitive processes, Stenfelt [/bib_ref].
The primary diagnostic tool in audiology is the pure tone audiogram. This method of determining frequencyspecific hearing thresholds is based on delivering sine waves of different intensities to each ear and asking the patients to respond by pressing a button each time they hear a sound. The resulting resolution is poor, and since this procedure requires the processes of intention and attention that characterize listening as opposed to simply hearing and thus tap into cognitive processes that may also be declining with age, diagnosis may be confounded. Other diagnostic tools include measures of auditory brainstem response and otoacoustic emissions which may be more independent of high-level cognitive contribution, although it has recently been shown that cognitive load influences brainstem responses [bib_ref] Working memory capacity and visual-verbal cognitive load modulate auditorysensory gating in the..., Sörqvist [/bib_ref] and otoacoustic emissions may also be influenced by attention through efferent innervation [bib_ref] Effects of cross-modal selective attention on the sensory periphery: cochlear sensitivity is..., Srinivasan [/bib_ref]. Assessment of speech intelligibility in quiet and in noise is also part of hearing evaluation.
## Hearing aids.
The most important objective for hearing aid signal processing is to make speech audible [bib_ref] Cognition and hearing aids, Lunner [/bib_ref]. This is not a trivial problem. Over 30 years ago, Plomp [bib_ref] Auditory handicap of hearing impairment and the limited benefit of hearing aids, Plomp [/bib_ref] proposed a model of hearing aid benefit that classed hearing impairment in terms of attenuation and distortion showing that while the hearing aids of the day could compensate well for the former by providing amplification, they were poorer at tackling the latter. As distortion is a characteristic of even the mildest hearing losses, it is important that hearing aids address this issue and the industry has taken on this challenge [bib_ref] The future of hearing aid technology, Edwards [/bib_ref]. Distortion can be simply characterized as a decrease in the ability to distinguish speech from noise. It is not only due to decreased frequency and temporal resolution, as well as impaired ability to discriminate pitch and localize sound sources, but also due to abnormal growth of loudness [bib_ref] Perceptual consequences of cochlear hearing loss and their implications for the design..., Moore [/bib_ref] , such that if all sounds are amplified the same way, some may become uncomfortably loud. Thus, modern digital hearing aids include technologies that tackle some of these problems. Wide dynamic range compression systems restore audibility by amplifying weaker sounds more than loud sounds to compensate for the abnormal growth of loudness. The regulation of the compression system may be fast (syllabic) or slow (automatic volume control). Fast-acting wide dynamic range compression (fast WDRC) provides different gain-frequency responses for adjacent speech sounds with different short-term spectra on a syllabic level. On the assumption that communication partners look at each other, directional microphones may be used to attenuate sounds not coming from the front. Of course, if the attended signal does not come from the front, directional microphones may make communication harder. Single-channel noise reduction schemes (NR) may reduce background sounds by identifying portions of the signal as nonspeech and attenuating these. This does not improve speech intelligibility per se, but it may reduce the annoyance from background sounds. Notwithstanding the benefits of signal processing, there is no getting away from the fact that it may also degrade the auditory signal, which may make listening harder. This applies in particular to aggressive signal processing algorithms that may be used experimentally but are not generally prescribed to patients. Aggressive processing is characterized by substantial spectral alteration of the signal within the space of a few milliseconds. For example, some aggressive NR algorithms generate audible artifacts [bib_ref] Speech perception of noise with binary gains, Wang [/bib_ref] and WDRC distorts individual speech sounds in ways that influence the phonological or sublexical structure of the incoming speech signal [bib_ref] Cognition and hearing aids, Lunner [/bib_ref] [bib_ref] Quantifying the effect of compression hearing aid release time on speech acoustics..., Jenstad [/bib_ref] [bib_ref] Multichannel compression: effects of reduced spectral contrast on vowel identification, Bor [/bib_ref] [bib_ref] Working memory supports listening in noise for persons with hearing impairment, Rudner [/bib_ref].
## Noise.
Acoustic noise impacting speech perception can be categorized as signal degradation, energetic masking, and informational masking [bib_ref] Speech recognition in adverse conditions: a review, Mattys [/bib_ref]. Signal degradation reduces the amount of information in the signal. As we have seen, this is the result of hearing aid signal processing. Other examples relate to processing for data transmission. Energetic masking is a competing signal that partially obscures the target signal. Air conditioning fans are a good example. Informational masking also obscures the target signal but in addition has a fluctuating structure that in some circumstances may distract the listener but in others may allow the listener to systematically glimpse parts of the signal. An informational masker may consist of tonal patterns, for example, or one or more competing speakers. As regards the neural networks underpinning speech comprehension in noise, a pattern is starting to emerge involving widespread frontal and parietal activation as well as increased temporal activation [bib_ref] The neural processing of masked speech, Scott [/bib_ref]. There is also some evidence that the brain tracks target and competing speech streams in a manner that is modulated by attention [bib_ref] Mechanisms underlying selective neuronal tracking of attended speech at a 'cocktail party, Zion Golumbic [/bib_ref] with selective attention networks for pitch and location [bib_ref] Auditory selective attention reveals preparatory activity in different cortical regions for selection..., Lee [/bib_ref].
Persons with hearing impairment have particular difficulties listening in noise which may be reflected in recruitment of neural networks supporting compensatory processing [bib_ref] Neurocognitive aging and the compensation hypothesis, Reuter-Lorenz [/bib_ref] [bib_ref] Aging and cortical mechanisms of speech perception in noise, Wong [/bib_ref] whereas persons with normal hearing are generally better at coping with informational than energetic masking [bib_ref] Effect of a single interfering noise or speech source upon the binaural..., Duquesnoy [/bib_ref] ; the same may not always be true for persons with hearing impairment [bib_ref] Effects of fluctuating noise and interfering speech on the speech-reception threshold for..., Festen [/bib_ref] [bib_ref] Auditory and nonauditory factors affecting speech reception in noise by older listeners, George [/bib_ref] [bib_ref] Speech perception problems of the hearing impaired reflect inability to use temporal..., Lorenzi [/bib_ref]. An informational masker includes cues in terms of pitch or temporal fine structure that may help segregation and dips in the masker may reveal portions of the target signal. This may result in the listener perceiving fragments of a target signal that need to be pieced together to achieve understanding. An informational masker may also include semantic information that distracts the listener from the target signal and thus needs to be inhibited. Such processes rely on cognitive functions.
## Working memory and speech comprehension
## The role of cognition in listening.
Cognitive processes are required to focus on the speech signal and match its contents to stored knowledge [bib_ref] Candidature for and delivery of audiological services: special needs of older people, Kiessling [/bib_ref] [bib_ref] Effects of age on auditory and cognitive processing: implications for hearing aid..., Pichora-Fuller [/bib_ref]. When listening takes place in adverse conditions, for example, when there is background noise or the listener has a hearing impairment, high-level cognitive functions such as working memory and executive processes are implicated [bib_ref] The Ease of Language Understanding (ELU) model: theory, data, and clinical implications, Rönnberg [/bib_ref] [bib_ref] The role of working memory capacity in auditory distraction: a review, Sörqvist [/bib_ref]. Working memory (WM) is the capacity to perform task-relevant processing of information kept in mind [bib_ref] The role of working memory capacity in auditory distraction: a review, Sörqvist [/bib_ref] [bib_ref] Working memory, Baddeley [/bib_ref] and is supported by a frontoparietal network [bib_ref] Imaging cognition II: an empirical review of 275 PET and fMRI studies, Cabeza [/bib_ref] [bib_ref] Working memory: a view from neuroimaging, Smith [/bib_ref] that is sensitive to stimulus quality and memory load [bib_ref] Dissociating working memory from task difficulty in human prefrontal cortex, Barch [/bib_ref] [bib_ref] A parametric study of prefrontal cortex involvement in human working memory, Braver [/bib_ref]. Many different models of WM have been proposed, and one of the most influential of them is the component model originating in the seminal 1974 paper by Baddeley and Hitch [bib_ref] Working memory, Baddeley [/bib_ref]. This model was characterized by a central executive controlling two slave buffers for processing verbal and visuospatial information, respectively. It elegantly accounted for a host of empirical data from dual task paradigms, that is, tasks requiring two different kinds of processing at the same time. However, it could not easily account for evidence of multimodal information binding, for example, use of visual cues during speech understanding. A new generation of WM models including an episodic buffer filling just such a function saw the light of day around the turn of the 21st century. These include an updated version of the original component model [bib_ref] The episodic buffer: a new component of working memory?, Baddeley [/bib_ref] and a model specifically describing the role of WM in language understanding: the WM model for ease of language understanding (ELU) [bib_ref] The Ease of Language Understanding (ELU) model: theory, data, and clinical implications, Rönnberg [/bib_ref] [bib_ref] Cognition in the hearing impaired and deaf as a bridge between signal..., Rönnberg [/bib_ref]. Although early work placed the episodic buffer among executive functions organized in the frontal lobes [bib_ref] Integration of diverse information in working memory within the frontal lobe, Prabhakaran [/bib_ref] , later work has shown that multimodal information binding does not necessarily load on executive functions. For example, visual binding has been shown to take place without executive involvement [bib_ref] Is the binding of visual features in working memory resource-demanding?, Allen [/bib_ref] and multimodal semantic binding has been shown to have its locus in the temporal lobes [bib_ref] Neural representation of binding lexical signs and words in the episodic buffer..., Rudner [/bib_ref] [bib_ref] The role of the episodic buffer in working memory for language processing, Rudner [/bib_ref]. The ELU model [bib_ref] The Ease of Language Understanding (ELU) model: theory, data, and clinical implications, Rönnberg [/bib_ref] links in with a parallel line of conceptual development represented by the individual differences approach to WM. This approach focuses on the large variance in individual ability to perform WM tasks rather than characterizing different components of WM [bib_ref] A capacity theory of comprehension: individual differences in working memory, Just [/bib_ref] [bib_ref] Individual differences in integrating information between and within sentences, Daneman [/bib_ref] [bib_ref] Individual differences in working memory and reading, Daneman [/bib_ref]. According to the ELU model [bib_ref] The Ease of Language Understanding (ELU) model: theory, data, and clinical implications, Rönnberg [/bib_ref] , language understanding proceeds rapidly and smoothly under optimal listening conditions, facilitated by an episodic buffer which matches phonological information in the incoming speech stream with the existing representations stored in long-term memory. Because this buffer deals with the rapid, automatic multimodal binding of phonology, it is known by the acronym RAMBPHO. Adverse listening conditions hinder RAMBPHO processing. This may result in a mismatch between auditory signal and information in the mental lexicon in long-term memory. Under such circumstances, explicit or conscious processing resources need to be brought into play to unlock the lexicon. The ELU model proposes that this occurs in a slow processing loop. Processing in the slow loop may include executive functions such as shifting, updating, and inhibition [bib_ref] The unity and diversity of executive functions and their contributions to complex..., Miyake [/bib_ref]. Inhibition may be required to suppress irrelevant interpretations, while updating may bring new information into the buffer at the expense of discarding older information. Shifting may come into play to realign expectations [bib_ref] Working memory supports listening in noise for persons with hearing impairment, Rudner [/bib_ref] [bib_ref] Cognitive spare capacity as a measure of listening effort, Rudner [/bib_ref]. All these functions are linked to the frontal lobes [bib_ref] Imaging cognition II: an empirical review of 275 PET and fMRI studies, Cabeza [/bib_ref] and there is evidence that they are supported by anatomically distinct substrates [bib_ref] Frontal lobes and human memory insights from functional neuroimaging, Fletcher [/bib_ref]. Their role in speech communication under adverse conditions may be bringing together ambiguous signal fragments with relevant contextual information. There is a constant interplay between predictive kinds of priming of what is to come in a dialogue and postdictive reconstructions of what was missed through mismatches with the lexicon in semantic long-term memory [bib_ref] The Ease of Language Understanding (ELU) model: theory, data, and clinical implications, Rönnberg [/bib_ref]. There is no doubt that such processing is effortful and increases cognitive load [bib_ref] Working memory capacity may influence perceived effort during aided speech recognition in..., Rudner [/bib_ref] [bib_ref] Cognitive load during speech perception in noise: the influence of age, hearing..., Zekveld [/bib_ref] and modulates the neural networks involved in speech processing under adverse conditions [bib_ref] Effortful listening: the processing of degraded speech depends critically on attention, Wild [/bib_ref]. From an individual difference perspective, it makes sense that individuals with high WM capacity would perform better on tasks requiring speech understanding under adverse conditions, and this is indeed the case [bib_ref] Cognitive function in relation to hearing aid use, Lunner [/bib_ref] [bib_ref] Are individual differences in speech reception related to individual differences in cognitive..., Akeroyd [/bib_ref] [bib_ref] How linguistic closure and verbal working memory relate to speech recognition in..., Besser [/bib_ref].
More than a decade ago, it was established that there is a relation between cognitive ability, in particular WM capacity, and the benefit obtained from hearing aid signal processing [bib_ref] Cognitive function in relation to hearing aid use, Lunner [/bib_ref] [bib_ref] How cognition might influence hearing aid design, fitting, and outcomes, Pichora-Fuller [/bib_ref] [bib_ref] Benefits from hearing aids in relation to the interaction between the user..., Gatehouse [/bib_ref] [bib_ref] Cognitive spare capacity as a window on hearing aid benefit, Rudner [/bib_ref]. In particular, it was shown that any benefit of fast-acting WDRC in terms of the ability to understand speech in noise was contingent on cognitive ability [bib_ref] Cognitive function in relation to hearing aid use, Lunner [/bib_ref] [bib_ref] Benefits from hearing aids in relation to the interaction between the user..., Gatehouse [/bib_ref]. Since then, it has been shown that this relationship is influenced by type of background noise [bib_ref] Recognition of speech in noise with new hearing instrument compression release settings..., Foo [/bib_ref] [bib_ref] Interactions between cognition, compression, and listening conditions: effects on speech-in-noise performance in..., Lunner [/bib_ref] [bib_ref] Phonological mismatch and explicit cognitive processing in a sample of 102 hearing-aid..., Rudner [/bib_ref] and the type of target speech material [bib_ref] Working memory supports listening in noise for persons with hearing impairment, Rudner [/bib_ref] [bib_ref] Recognition of speech in noise with new hearing instrument compression release settings..., Foo [/bib_ref] [bib_ref] Cognition and aided speech recognition in noise: specific role for cognitive factors..., Rudner [/bib_ref]. Cognitive resources are especially important when modulated noise is combined with fast-acting WDRC [bib_ref] Working memory supports listening in noise for persons with hearing impairment, Rudner [/bib_ref] [bib_ref] Working memory capacity may influence perceived effort during aided speech recognition in..., Rudner [/bib_ref] [bib_ref] Interactions between cognition, compression, and listening conditions: effects on speech-in-noise performance in..., Lunner [/bib_ref] [bib_ref] Phonological mismatch and explicit cognitive processing in a sample of 102 hearing-aid..., Rudner [/bib_ref] [bib_ref] Cognition and aided speech recognition in noise: specific role for cognitive factors..., Rudner [/bib_ref] above all when the target speech is unpredictable [bib_ref] Working memory supports listening in noise for persons with hearing impairment, Rudner [/bib_ref]. These complex relations change over time [bib_ref] Working memory supports listening in noise for persons with hearing impairment, Rudner [/bib_ref] [bib_ref] Cognition and aided speech recognition in noise: specific role for cognitive factors..., Rudner [/bib_ref] [bib_ref] Short and long compression release times: speech understanding, real-world preferences, and association..., Cox [/bib_ref].
The capacity of WM can be increased by training, suggesting an inherent plasticity in the system [bib_ref] Transfer of learning after updating training mediated by the striatum, Dahlin [/bib_ref] [bib_ref] Computerized training of working memory in children with ADHD-a randomized, controlled trial, Klingberg [/bib_ref]. Training effects may generalise to similar nontrained tasks, for example, a different WM task [bib_ref] Transfer of learning after updating training mediated by the striatum, Dahlin [/bib_ref]. This is known as near transfer. However, generalization to other cognitive abilities, known as far transfer, has been elusive [bib_ref] Putting brain training to the test, Owen [/bib_ref]. Recent work, however, has shown that for older adults, cognitive training requiring multitasking can result in sustained reduction in multitasking costs and improvement in WM [bib_ref] Video game training enhances cognitive control in older adults, Anguera [/bib_ref]. As we have noted, WM is about simultaneous storage and processing, in other words a form of multitasking. The results of Anguera et al. [bib_ref] Video game training enhances cognitive control in older adults, Anguera [/bib_ref] suggest that in order to improve WM, it may be more efficient to target multitasking abilities as such. Since WM capacity is related to the ability to understand speech in noise, it is tempting to speculate that increasing WM capacity may also improve the ability to understand speech in noise. However, published evidence for the efficacy of individual computer-based auditory training for adults with hearing loss is not robust [bib_ref] Efficacy of individual computer-based auditory training for people with hearing loss: a..., Henshaw [/bib_ref]. We suggest that cognitive training that targets the multitasking abilities inherent in speech understanding under adverse conditions may improve WM capacity and result in better speech understanding in adverse conditions. This is an important avenue for future research.
## Cognitive spare capacity for communication
## Cognitive spare capacity.
When listening takes place in adverse conditions, it is clear that the cognitive resources available for higher level processing of speech will be reduced. In other words, the listener has less cognitive spare capacity (CSC) [bib_ref] Cognitive spare capacity as a measure of listening effort, Rudner [/bib_ref] [bib_ref] Cognitive spare capacity as a window on hearing aid benefit, Rudner [/bib_ref] [bib_ref] Speech understanding and cognitive spare capacity, Mishra [/bib_ref] [bib_ref] Seeing the talker's face supports executive processing of speech in steady state..., Mishra [/bib_ref]. CSC is closely related to WM in that it is concerned with short-term maintenance and processing of information [bib_ref] Cognitive spare capacity as a measure of listening effort, Rudner [/bib_ref]. Work to date suggests that the storage functions of CSC and WM are similar [bib_ref] Effects of noise and working memory capacity on memory processing of speech..., Ng [/bib_ref] but that once executive processing demands are introduced, there no longer seems to be a simple relationship between the two concepts [bib_ref] Cognitive spare capacity as a window on hearing aid benefit, Rudner [/bib_ref] [bib_ref] Seeing the talker's face supports executive processing of speech in steady state..., Mishra [/bib_ref] [bib_ref] Visual information can hinder working memory processing of speech, Mishra [/bib_ref]. Thus, in order to understand the role of cognition in speech understanding under adverse conditions, it is important to measure not only WM capacity but also CSC. The concept of CSC is related to, although distinct from, other concepts in the literature. For example, differences in susceptibility to functional impairment as a result of brain damage have been explained in terms of "cognitive reserve, " that is, individual differences in cognitive function [bib_ref] Efficiency, capacity, compensation, maintenance, plasticity: emerging concepts in cognitive reserve, Barulli [/bib_ref] , or "brain reserve, " that is, individual differences in brain size [bib_ref] Brain and cognitive reserve: mediator(s) and construct validity, a critique, Satz [/bib_ref]. CSC is similar to these concepts in that it is based on individual differences in cognitive function and may explain differences in speech communication and underlying mechanisms that may be related to functional changes at any level of the auditory system [bib_ref] Cognitive spare capacity as a window on hearing aid benefit, Rudner [/bib_ref] [bib_ref] Speech understanding and cognitive spare capacity, Mishra [/bib_ref].
Recent work has shown that noise reduction (NR) in hearing aids can enhance CSC by improving retention of heard speech [bib_ref] Effects of noise and working memory capacity on memory processing of speech..., Ng [/bib_ref] [bib_ref] Objective measures of listening effort: effects of background noise and noise reduction, Sarampalis [/bib_ref]. This applies to both adults with normal hearing thresholds [bib_ref] Objective measures of listening effort: effects of background noise and noise reduction, Sarampalis [/bib_ref] and adults with sensorineural hearing impairment [bib_ref] Effects of noise and working memory capacity on memory processing of speech..., Ng [/bib_ref]. In the study by Ng et al. [bib_ref] Effects of noise and working memory capacity on memory processing of speech..., Ng [/bib_ref] , experienced hearing aid users listened to sets of highly intelligible, ecologically valid sentences from the Swedish hearing in noise test (HINT) [bib_ref] A Swedish version of the Hearing In Noise Test (HINT) for measurement..., Hällgren [/bib_ref] [bib_ref] Development of the hearing in noise test for the measurement of speech..., Nilsson [/bib_ref]. The HINT sentences were presented in noise and the participants were asked to memorize the final word of each sentence. The participants repeated all the target words to ensure that they were intelligible. At the end of each set, participants were prompted to recall all the sentence-final words. Although they were capable of repeating the sentence-final words, irrespective of the presence of background noise, noise did disrupt recall performance [bib_ref] Effects of noise and working memory capacity on memory processing of speech..., Ng [/bib_ref]. Being able to retain heard information is an integral part of speech communication. Thus, the findings of Ng et al. [bib_ref] Effects of noise and working memory capacity on memory processing of speech..., Ng [/bib_ref] demonstrate that, for individuals with hearing impairment, background noise reduces the cognitive resources available for performing the kind of cognitive processing involved in communication. This is in line with the work showing that extra effort expended simply in order to hear comes at the cost of processing resources that might otherwise be available for encoding the speech content in memory [bib_ref] Hearing loss and perceptual effort: downstream effects on older adults' memory for..., Mccoy [/bib_ref] [bib_ref] Comparing the effects of aging and background noise on shortterm memory performance, Murphy [/bib_ref]. However, when NR was implemented, the negative effect of noise on recall was reduced, even though the ability to repeat sentencefinal words remained the same [bib_ref] Effects of noise and working memory capacity on memory processing of speech..., Ng [/bib_ref]. This demonstrates that hearing aid signal processing can enhance memory processes underpinning speech communication. Informational masking was more disruptive of memory processing than energetic masking and was also more susceptible to the positive effect of NR [bib_ref] Effects of noise and working memory capacity on memory processing of speech..., Ng [/bib_ref]. However, it remains to be determined whether it is the semantic content or phonological structure of the informational masker that interacts with the ability of NR to improve memory for highly intelligible speech.
Speech communication under adverse conditions is likely to draw on cognitive functions other than simply memory retention [bib_ref] Working memory supports listening in noise for persons with hearing impairment, Rudner [/bib_ref] [bib_ref] Cognitive spare capacity as a measure of listening effort, Rudner [/bib_ref]. In order to investigate the ability to perform executive processing of heard speech at different memory loads, the cognitive spare capacity test (CSCT) [bib_ref] Seeing the talker's face supports executive processing of speech in steady state..., Mishra [/bib_ref] [bib_ref] Visual information can hinder working memory processing of speech, Mishra [/bib_ref] was developed. In the CSCT, sets of spoken twodigit numbers are presented and the participant reports back certain numbers according to instructions. Two executive functions are targeted at two different memory loads. The executive functions in question are updating and inhibition, both of which are likely to be engaged during speech understanding in adverse conditions. Updating ability may be required to strategically replace the contents of WM with relevant material while inhibition ability may be brought into play to keep irrelevant information out of WM. Memory load depends on how many numbers need to be reported. In everyday communication, seeing the face of your communication partner can enhance speech perception by several dB [bib_ref] Visual contribution to speech intelligibility in noise, Sumby [/bib_ref]. Thus, in order to determine how visual cues influence CSC, the CSCT manipulates availability of visual cues. The CSCT can be administered in quiet or in noise and other manipulations introducing different kinds of signal processing are also possible.
Across three different studies including persons with and without hearing loss, an interesting pattern of results has emerged [bib_ref] Cognitive spare capacity as a window on hearing aid benefit, Rudner [/bib_ref] [bib_ref] Seeing the talker's face supports executive processing of speech in steady state..., Mishra [/bib_ref] [bib_ref] Visual information can hinder working memory processing of speech, Mishra [/bib_ref] [bib_ref] Cognitive spare capacity in older adults with hearing loss, Mishra [/bib_ref]. Adults with normal hearing who perform the CSCT in quiet conditions have lower scores when they see the talker's face [bib_ref] Seeing the talker's face supports executive processing of speech in steady state..., Mishra [/bib_ref] [bib_ref] Visual information can hinder working memory processing of speech, Mishra [/bib_ref]. This is probably because when target information is highly intelligible, visual cues provide superfluous information that causes distraction during performance of the executive tasks [bib_ref] Seeing the talker's face supports executive processing of speech in steady state..., Mishra [/bib_ref] [bib_ref] Visual information can hinder working memory processing of speech, Mishra [/bib_ref]. Although this finding is contrary to the literature on speech perception, which demonstrates better performance in noise when the talker's face is visible, for individuals with normal hearing [bib_ref] Interaction of audition and vision in the recognition of oral speech stimuli, Erber [/bib_ref] and individuals with hearing impairment [bib_ref] Auditory and auditoryvisual intelligibility of speech in fluctuating maskers for normalhearing and..., Bernstein [/bib_ref] [bib_ref] Auditory-visual speech recognition by hearing-impaired subjects: consonant recognition, sentence recognition, and auditory-visual..., Grant [/bib_ref] [bib_ref] The use of visible speech cues for improving auditory detection of spoken..., Grant [/bib_ref] , it is in line with other lines of evidence showing that visual cues may increase listening effort [bib_ref] Evaluating the effort expended to understand speech in noise using a dual-task..., Fraser [/bib_ref] [bib_ref] Older adults expend more listening effort than young adults recognizing audiovisual speech..., Gosselin [/bib_ref]. In particular, dual task performance is lower for audiovisual compared to auditory stimuli when intelligibility is equated across modalities [bib_ref] Evaluating the effort expended to understand speech in noise using a dual-task..., Fraser [/bib_ref] [bib_ref] Older adults expend more listening effort than young adults recognizing audiovisual speech..., Gosselin [/bib_ref].
Adults with normal hearing who perform CSCT in noisy conditions do not show this pattern [bib_ref] Seeing the talker's face supports executive processing of speech in steady state..., Mishra [/bib_ref] and nor do older adults with raised hearing thresholds, even in quiet [bib_ref] Cognitive spare capacity in older adults with hearing loss, Mishra [/bib_ref]. In these conditions, visual cues probably help segregate the target signal from internal or external noise, resulting in richer cognitive representations [bib_ref] Seeing the talker's face supports executive processing of speech in steady state..., Mishra [/bib_ref] [bib_ref] ERP Evidence that auditory-visual speech facilitates working memory in younger and older..., Frtusova [/bib_ref]. Older adults with hearing loss demonstrate lower CSC than young adults, even with better SNR, adapted to provide high intelligibility [bib_ref] How young and old adults listen to and remember speech in noise, Pichora-Fuller [/bib_ref] and individualised amplification, and this effect is most notable in noise and when memory load is high [bib_ref] Cognitive spare capacity as a window on hearing aid benefit, Rudner [/bib_ref]. Although CSC and WM do not seem to be strongly related, there is evidence that age-related differences in WM and executive function do influence CSC [bib_ref] Cognitive spare capacity as a window on hearing aid benefit, Rudner [/bib_ref] [bib_ref] Cognitive spare capacity in older adults with hearing loss, Mishra [/bib_ref]. It remains to be seen how different kinds of hearing aid signal processing will interact with executive processing of speech with and without visual cues and whether training CSC can counteract agerelated decline in its capacity or even improve CSC. Adaptive training based on CSCT processing may provide a means of improving the ability to understand speech under adverse conditions.
## Phonological
Representation. The ELU model describes the way in which the mapping of phonological structure of target speech onto phonological representations in the mental lexicon [bib_ref] Recognizing spoken words: the neighborhood activation model, Luce [/bib_ref] is mediated by WM during speech understanding under adverse conditions [bib_ref] The Ease of Language Understanding (ELU) model: theory, data, and clinical implications, Rönnberg [/bib_ref]. We have seen that fast-acting WDRC distorts the speech signal in a way that may influence its phonological characteristics [bib_ref] Cognition and hearing aids, Lunner [/bib_ref] [bib_ref] Quantifying the effect of compression hearing aid release time on speech acoustics..., Jenstad [/bib_ref] [bib_ref] Multichannel compression: effects of reduced spectral contrast on vowel identification, Bor [/bib_ref] [bib_ref] Working memory supports listening in noise for persons with hearing impairment, Rudner [/bib_ref]. In the short term, this may make it harder to match speech to representations, thus requiring more cognitive engagement to achieve speech understanding [bib_ref] The Ease of Language Understanding (ELU) model: theory, data, and clinical implications, Rönnberg [/bib_ref] [bib_ref] Recognition of speech in noise with new hearing instrument compression release settings..., Foo [/bib_ref] [bib_ref] Cognition and aided speech recognition in noise: specific role for cognitive factors..., Rudner [/bib_ref]. However, in the long term, when hearing aid users have had the opportunity to become accustomed to the way in which speech sounds different, phonological representations may alter to match incoming information. Some evidence of this has been found in cochlear implantees [bib_ref] Phonological processing in post-lingual deafness and cochlear implant outcome, Lazard [/bib_ref] and hearing aid users [bib_ref] Working memory supports listening in noise for persons with hearing impairment, Rudner [/bib_ref]. It is even possible that the new phonological representations based on processed speech may be more mutually distinct than the representations they replace based on less appropriate signal processing. The neural correlates of such changes in phonological representation due to habitual use of WRDC have yet to be investigated.
Lexical access is faster when phonological representations are easier to distinguish from each other [bib_ref] Recognizing spoken words: the neighborhood activation model, Luce [/bib_ref] [bib_ref] Gated audiovisual speech identification in silence vs. noise: effects on time and..., Moradi [/bib_ref]. However, long-term severe acquired hearing impairment may lead to less distinct phonological representations [bib_ref] Phonological processing in post-lingual deafness and cochlear implant outcome, Lazard [/bib_ref]. This makes it harder to determine whether printed words rhyme with each other [bib_ref] Deterioration of the phonological processing skills in adults with an acquired severe..., Andersson [/bib_ref] , especially when orthography is misleading [bib_ref] Working memory compensates for hearing related phonological processing deficit, Classon [/bib_ref]. For example, individuals with poor phonological representations due to severe long-term hearing impairment may be more unsure than their peers with normal hearing whether "pint" rhymes with "lint" or whether "blue" rhymes with "through. " However, good WM capacity can compensate for this deficit, albeit at the cost of long-term memory representations [bib_ref] Working memory compensates for hearing related phonological processing deficit, Classon [/bib_ref]. Compensatory processing by individuals with hearing impairment during visual rhyme judgment is associated with larger amplitude of the N2 component [bib_ref] Early ERP signature of hearing impairment in visual rhyme judgment, Classon [/bib_ref] , indicating use of a compensatory strategy, possibly involving increased reliance on explicit mechanisms such as articulatory recoding and grapheme-to-phoneme conversion.
In summary, phonological structure of target speech material is not only influenced by speaker characteristics but also by distortion due to hearing aid signal processing. Phonological representations in the mental lexicon may be influenced by long-term effects of both hearing impairment and signal processing. Further, both of these may have distinct neural signatures. Measures designed to improve phonological distinctiveness of both target speech and phonological representations are likely to enhance CSC and support speech communication under adverse conditions. This deserves further investigation.
## Semantic context.
Provision of semantic context can facilitate speech understanding under adverse conditions. This process engages language networks in left posterior inferior temporal cortex and inferior frontal gyri bilaterally [bib_ref] The neural mechanisms of speech comprehension: fMRI studies of semantic ambiguity, Rodd [/bib_ref]. Studies investigating the role of WM capacity in the benefit obtained from WDRC have indicated that the semantic content of the materials delivered for speech recognition may influence this relationship. For example, Rudner et al. [bib_ref] Working memory supports listening in noise for persons with hearing impairment, Rudner [/bib_ref] found that WM capacity was associated with speech understanding for individuals with hearing impairment using WDRC listening to matrix-type sentences [bib_ref] Efficient adaptive methods for measuring speech reception threshold in quiet and in..., Hagerman [/bib_ref] [bib_ref] Sentences for testing speech intelligibility in noise, Hagerman [/bib_ref] , but not Swedish HINT sentences [bib_ref] A Swedish version of the Hearing In Noise Test (HINT) for measurement..., Hällgren [/bib_ref] [bib_ref] Development of the hearing in noise test for the measurement of speech..., Nilsson [/bib_ref]. The Hagerman sentences are semantically coherent, but the fiveword syntactic structure is always the same and each word comes from a closed set of ten appropriate items. Thus none of the items can be accurately predicted. The HINT sentences, by contrast, are diverse in length, syntactic structure and semantic coherence. It is likely that the constrained structure and content of the Hagerman sentences make guessing harder and thus increase reliance on the bottom-up information provided by the speech signal. However, it has been found that the benefit of having access to the temporal fine structure of the speech signal was greater for open set materials than for closed-set materials [bib_ref] Effect of speech material on the benefit of temporal fine structure information..., Lunner [/bib_ref] , indicating that the regular structure and closed set of matrix-like sentences can facilitate guessing . Future work should systematically investigate the interaction between the semantic coherence of the speech signal, hearing aid signal processing, and individual cognitive characteristics such as WM and CSC.
Text cues can facilitate speech understanding in noise when they match the semantic content of the auditory signal [bib_ref] The effects of working memory capacity and semantic cues on the intelligibility..., Zekveld [/bib_ref] [bib_ref] Predictive top-down integration of prior knowledge during speech perception, Sohoglu [/bib_ref] [bib_ref] Audiovisual perception of speech in noise and masked written text, Zekveld [/bib_ref] [bib_ref] Behavioral and fMRI evidence that cognitive ability modulates the effect of semantic..., Zekveld [/bib_ref] [bib_ref] The influence of semantically related and unrelated text cues on the intelligibility..., Zekveld [/bib_ref] and inhibit it when they are misleading [bib_ref] The influence of semantically related and unrelated text cues on the intelligibility..., Zekveld [/bib_ref]. Cue integration is supported by language networks including the inferior frontal gyrus and temporal regions [bib_ref] Behavioral and fMRI evidence that cognitive ability modulates the effect of semantic..., Zekveld [/bib_ref]. Matching text cues also enhance the perceived clarity of degraded speech [bib_ref] Effortful listening: the processing of degraded speech depends critically on attention, Wild [/bib_ref] and recently it was shown that this effect may be modulated by both lexical access speed and WM capacity [bib_ref] Lexical access speed determines the role of working memory in popout, Signoret [/bib_ref]. WM capacity modulates the activation of networks involved in semantic processing [bib_ref] Behavioral and fMRI evidence that cognitive ability modulates the effect of semantic..., Zekveld [/bib_ref] and also predicts the ability to inhibit misleading text cues during speech understanding in steady state noise [bib_ref] The influence of semantically related and unrelated text cues on the intelligibility..., Zekveld [/bib_ref] as well as the facilitation of speech understanding against a single talker background [bib_ref] The effects of working memory capacity and semantic cues on the intelligibility..., Zekveld [/bib_ref]. Recently, it has been shown that coherence and cues can have separate facilitatory effects on perceived clarity of degraded speech [bib_ref] Lexical access speed determines the role of working memory in popout, Signoret [/bib_ref]. Future work should focus on determining the benefit of providing text cues for hearing aid users, for example, using automatic speech recognition [bib_ref] Audiovisual perception of speech in noise and masked written text, Zekveld [/bib_ref] and how this interacts with the semantic coherence of the target speech, the availability of semantic content in the noise background, and individual cognitive skills. Imaging studies are likely to provide important information about the neurocognitive systems supporting these complex interactions.
## Aging and communication.
Sensory and cognitive functions decline with age [bib_ref] Memory aging and brain maintenance, Nyberg [/bib_ref] [bib_ref] Emergence of a powerful connection between sensory and cognitive functions across the..., Baltes [/bib_ref]. Sensory decline can be traced to physiological change, but the mechanisms behind cognitive change are more elusive, although both genetic and lifestyle factors have been implicated [bib_ref] Memory aging and brain maintenance, Nyberg [/bib_ref]. Several different theories attempt to explain the relation between sensory and cognitive decline. The common cause hypothesis [bib_ref] Emergence of a powerful connection between sensory and cognitive functions across the..., Baltes [/bib_ref] proposes that a general reduction in processing efficiency drives both phenomena. The information degradation hypothesis [bib_ref] Listening in aging adults: from discourse comprehension to psychoacoustics, Schneider [/bib_ref] , on the other hand, claims than when sensory input is degraded, cognitive processing becomes less efficient as a result. Reserve theories suggest that the ability to cope with brain damage is related to premorbid brain size or cognitive ability [bib_ref] Brain and cognitive reserve: mediator(s) and construct validity, a critique, Satz [/bib_ref]. The compensation-related utilization of neural circuits hypothesis [bib_ref] Neurocognitive aging and the compensation hypothesis, Reuter-Lorenz [/bib_ref] suggests that older adults compensate for less efficient processing by engaging more neural resources than younger adults when task load is still relatively low while brain maintenance theory [bib_ref] Memory aging and brain maintenance, Nyberg [/bib_ref] proposes that individual differences in the manifestation of age-related brain changes and pathology allow some people to show little or no age-related cognitive decline. All these theories are more or less sophisticated in their attempts to capture the relationship between physiological, sensory, and cognitive function in an aging perspective. The relations they describe suggest that keeping the brain healthy and providing it with better sensory input will facilitate speech understanding for individuals of advancing age. The theories that focus on a special role for cognition suggest that lowering cognitive load and enhancing CSC during speech communication may have special importance in later adulthood and even allow some older adults to function communicatively just as successfully as their younger counterparts.
Recent work has shown that older adults show less activation in auditory cortex than younger adults while listening to speech in noise, especially at poor signal to noise ratios and compensate by recruiting prefrontal and parietal areas associated with WM [bib_ref] Aging and cortical mechanisms of speech perception in noise, Wong [/bib_ref]. Epidemiological studies show that individuals with hearing loss are at increased risk of cognitive impairment and that rate of cognitive decline and risk of cognitive impairment are associated with severity of hearing loss [bib_ref] Hearing loss and cognitive decline in older adults, Lin [/bib_ref]. Thus, hearing loss may result in decreasing CSC. No study has yet specifically addressed this issue. However, analysis of data from the Betula study of cognitive aging [bib_ref] The betula prospective cohort study: memory, health, and aging, Nilsson [/bib_ref] demonstrated that hearing aid users with poorer hearing also had poorer long-term memory [bib_ref] Hearing loss is negatively related to episodic and semantic long-term memory but..., Rönnberg [/bib_ref]. This applied even when the long-term memory task had no auditory component. However, degree of hearing loss was not associated with decline in WM. Importantly, there was no significant association between loss of vision and cognitive function. These results suggest that although hearing loss and cognitive decline are related, even in hearing aid users, the association may not apply across all cognitive domains. The challenge is to uncover the specific mechanisms behind age-related sensory and cognitive decline so that speech communication can be preserved into old age by optimizing cognitive capacity. This may involve a range of different interventions that target hearing through appropriate hearing aid fitting, enhance the role of other sensory modalities that can be exploited in communication, and capitalize on cognitive abilities by seeking to maintain and extend them.
# Conclusion
Speech communication in adverse conditions makes specific demands on cognitive resources. In particular, WM capacity and executive function are engaged in unravelling the speech signal. This depletes CSC and leaving fewer resources for higher level processing of speech. CSC is influenced by cognitive load, noise, visual cues, and aging and can be enhanced by appropriate hearing aid signal processing. The phonological structure and semantic content of speech influence processing mechanisms and engagement of cognitive resources. Optimizing CSC is an important aim for preserving speech communication into old age. We have reviewed evidence suggesting that CSC may be enhanced by a number of means including cognitive training and providing the optimal balance between visual, phonological, and semantic information. Future research should focus on finding ways to optimize CSC.
Conflict of InterestsThe authors declare that there is no conflict of interests regarding the publication of this paper.AcknowledgmentThe authors would like to thank Jerker Rönnberg for comments and suggestions on a previous version of the paper. |
The Acid–Base Balance and Gender in Inflammation: A Mini-Review
Abbreviations: TNF-α, tumor necrosis factor-α; IL-1β, interleukin-1β; IL-6, interleukin-6; GLUT 2, glucose transporter 2; NF-κB, nuclear factor kappa B; BCL6, B-cell lymphoma 6 protein; G-CSF, granulocyte colony-stimulating factor.Casimir et al.
In humans, acid-base balance is crucial to cell homeostasis. Acidosis is observed in numerous inflammatory processes, primarily acute conditions such as sepsis, trauma, or acute respiratory distress where females tend to exhibit better prognosis compared with males. The mechanisms underlying these gender-dependent differences are multiple, probably involving hormonal and genetic factors, particularly the X chromosome. Although pH influences multiple immunological functions, gender differences in acid-base balance have been poorly investigated. In this review, we provide an update on gender differences in human susceptibility to inflammatory diseases. We additionally discuss the potential impact of acid-base balance on the gender bias of the inflammatory response in view of our recent observation that girls present higher neutrophilic inflammation and lower pH with a trend toward better prognosis in severe sepsis. We also highlight the potent role played by endothelial cells in gender differences of inflammation through activation of proton-sensing G protein-coupled receptors.
Keywords: inflammation, homeostatic balance, neutrophils, monocytes, endothelial cells, gender differences, acid-base balance, mechanisms of inflammatory cascade inTRODUCTiOn In both humans and animals, physiology and metabolism need homeostatic mechanisms [bib_ref] Homeostasis, inflammation, and disease susceptibility, Kotas [/bib_ref] to maintain the stability of not only the intracellular but also the extracellular milieu and plasma, thereby guaranteeing long-term survival of multicellular organisms. Various factors including basic metabolism, diet, physical activity, and environmental aggressions are daily life triggers of many physiological imbalances. Inflammation is the main active response designed to avoid dramatic stress challenges to homeostasis in the setting of infections, tissue injuries, cancers, or large burns. To maintain a permanent cell homeostasis, different receptor types are capable of sensing short and significant fluctuations in certain variables, along with multiple messengers that likely modify cell recruitment (especially neutrophils from vessels and bone marrow), diapedesis of monocytes and macrophages, vascular permeability, and production of protein mediators like cytokines and chemokines. Neutrophils, the major inflammatory effector cells at the site of acute injury (abscess, surgical trauma, pleural effusion, etc.), are capable to live in extremely challenging conditions such as severe acidosis or lack of oxygen. Chronic inflammation and diseases with a similar response profile can be considered as a dysregulation of the defense mechanisms, causing deleterious tissue damage and increased morbidity and mortality. In these cases, different parameters may be chronically modified by inflammation, thereby exceeding normal ranges, without any possibility of returning to normal patterns when the system remains "locked. "
Mechanisms that control homeostasis could differ between males and females.
## Role of acid-base imbalance as trigger of inflammatory responses
During acute inflammatory processes, particularly infections and even more dramatically sepsis, the acid-base balance is usually severely challenged. The extracellular milieu's pH interferes with a wide range of immunological functions [bib_ref] Extracellular acidosis induces neutrophil activation by a mecha nism dependent on activation..., Martinez [/bib_ref]. The role of acid as trigger of cytokine production was already described in 1997. In vitro studies have emphasized, among others, the following associations: increased inflammatory cytokines, such as interleukin (IL)-1β, IL-6, or tumor necrosis factor-α (TNF-α), produced by mononuclear cells; neutrophil activation with upregulation of cluster of differentiation (CD)18 expression and hydrogen peroxide production; and maturation of human dendritic cells. More recently, in critically ill patients, a positive correlation was found between a strong anion gap and the concentrations of IL-6, IL-8, IL-10, and TNF-α [bib_ref] The effect of lactic acid on mononuclear cell secretion of proinflammatory cytokines..., Steele [/bib_ref].
In the initial response to acute infection, the primarily phagocyte-based innate immune system likely plays a crucial role in managing the inflammatory process. A critical modulation of the early inflammatory process by metabolic acidosis may impact the prognosis of the sepsis. In a blunt trauma setting, upon admission, patients exhibit a significant base deficit that is associated with differential immune/inflammatory pathways, which may subsequently predispose patients to a more complicated clinical course [bib_ref] Elevated admission base deficit is associated with a complex dynamic network of..., Abdul-Malik [/bib_ref]. It could be postulated that the proton concentration that must be drastically controlled in life could represent a unifying signal inducing inflammation. Proton-sensing G protein-coupled receptors, including OGR1, GPR4, and TDAG8, were reported to prove highly significant for physiological pH homeostasis and inflammation control [bib_ref] Fluid resuscitation and hyperchloremic acidosis in experimental sepsis: improved short-term survival and..., Kellum [/bib_ref]. Patients suffering from inflammatory bowel disease have been shown to express higher levels of these proton-sensing receptors in the mucosa compared to controls. It is interesting to note that proton pump inhibitors, which block gastric acid secretion, were shown in vitro to selectively inhibit TNF-α and IL-1β secretion by TLR-receptor-activated human monocytes, without any cellular toxic effects. They are thus considered as promising agents targeting severe inflammation (8), but might also account for increased susceptibility to infections in these patients [bib_ref] Proton pump inhibitors protect mice from acute systemic inflammation and induce long-term..., Balza [/bib_ref]. Proton pump inhibitors can also enhance the risk of Clostridium difficile infections (10).
GPR4, a proton-sensing receptor expressed in endothelial cells and other cell types, is fully activated by acidic extracellular pH. However, this product exhibits less activity at the physiological pH 7.4 and only minimal activity at a more alkaline pH [bib_ref] The risk of Clostridium difficile infection in patients with pernicious anaemia: a..., Othman [/bib_ref]. When varying GPR4 expression in human umbilical vein endothelial cells, it proves possible to induce a substantially increased expression of numerous inflammatory genes, such as chemokines, cytokines, adhesion molecules, nuclear factor kappa B (NF-κB) pathway genes, prostaglandin-endoperoxide synthase 2, and stress response genes. This also applies to human lung microvascular endothelial cells and pulmonary artery endothelial cells. While acidosis-induced GPR4 activation stimulates the expression of numerous inflammatory genes in endothelial cells, it has been possible to suppress this inflammatory response by small molecule inhibitors of GPR4, which suggest a potential therapeutic value of such agents.
Although numerous mediators have been shown to increase neutrophil levels when injected into experimental animals (leukotriene B4, complement C5a), the initiation process (particularly monocytes-macrophages and endothelial cells) and triggering factor could, however, differ depending on the disease's origin. Regardless of the disease type, these cells may be generated by the same unifying physiological mechanism that induces the inflammatory cascade. To identify and assess targeted interventions, there is a pressing need to better understand inflammatory signaling along with the cascade of specific mechanisms [bib_ref] Acidosis activation of the proton-sensing GPR4 receptor stimulates vascular endothelial cell inflammatory..., Dong [/bib_ref] and steps pertaining to the inflammatory process. Therefore, studying pro-inflammatory stimuli that elicit rapid transcriptional responses via transduced signals with the aim to master regulatory transcription factors proves determinant for apprehending the response sequences. TNF-α can induce a rapid global redistribution of chromatin activators to massive de novo clustered enhancer domains, with endothelial cells likely to play a major role in this process. Several endothelial dysfunction markers such as plasma endocan, a proteoglycan excreted by the endothelial cells, could be employed to monitor (13) the endothelial response to aggression. In the future, antioxidant enzymes, such as catalase and superoxide dysmutase, could represent (14) a strategy designed to protect organs and tissues from inflammation and oxidative stress. In addition, a recently published paper emphasized the role of monocyte subtypes Ly6C low , with these cells routinely patrolling the endothelial wall under steady-state conditions [bib_ref] Endothelial targeting of nanocarriers loaded with antioxidant enzymes for protection against vascular..., Hood [/bib_ref]. These cells were shown to precede neutrophil arrival and orchestrate cell extravasation in response to TLR7/8-mediated vascular inflammation. The relative roles of monocytes and endothelial cells, along with their respective production of cytokines and chemokines, have yet to be clarified.
## Gender differences in the inflammatory process
Sexual dimorphism is observed in inflammatory conditions all along the life course. While estrogens and androgens are sexual hormones known to modulate inflammation, their fluctuant levels in males and females of any age cannot account for the gender differences of the inflammation observed in humans and animals from birth to death. No uniform concept covering all inflammatory conditions could be found because of highly variable responses of the immune system to sexual hormones [bib_ref] Toll-like receptors elicit different recruitment kinetics of monocytes and neutrophils in mouse..., Imhof [/bib_ref]. Although estrogens are clearly known to modulate the immune response, in terms of cytokine production, receptors, and clinical outcome, these observations cannot fully explain the universal gender differences observed in acute inflammation, found across all age groups, from premature infants to geriatric patients.
In acute inflammatory conditions, male gender is associated with a higher risk of morbidity and mortality, with females at any age exhibiting better prognosis. However, in chronic inflammatory processes, less frequent than acute, females display worse prognosis and higher mortality, probably because of collateral tissue damages caused by higher inflammation [bib_ref] Toll-like receptor signaling, Akira [/bib_ref]. The longer the inflammation of a tissue lasts, the more damage is done. This could account for the higher mortality observed in females suffering from cystic fibrosis [bib_ref] The complex role of estrogens in inflammation, Straub [/bib_ref] and chronic obstructive pulmonary diseases [bib_ref] Sex and inflammation in respiratory diseases: a clinical viewpoint, Casimir [/bib_ref]. In chronic inflammation, we also reported worse prognosis in girls suffering from chronic asthma, cystic fibrosis, or sickle cell anemia (SCA) [bib_ref] Sex gap in cystic fibrosis mortality, Rosenfeld [/bib_ref].
These observations have triggered gene analysis on the X chromosome, as well as investigating the potential influence of sexual steroids on inflammatory responses [bib_ref] Female smokers beyond the perimenopausal period are at increased risk of chronic..., Gan [/bib_ref]. In females, one of the X chromosomes is randomly silenced during X chromosome inactivation in the early stage of female embryogenesis, whereas the pseudoautosomal region of the X chromosome escapes inactivation. This process results in female cellular mosaicism, with half of the cells expressing genes derived from the maternal X chromosome and the other half expressing those derived from the paternal X chromosome. Moreover, spreading the inactivation signal on the pseudoautosomal region of the X chromosome may cause partial silencing of genes on the border, explaining higher gene expression of certain genes of the pseudoautosomal regions in males. The diversity in females is further increased because, if disadvantageous mutations occur in an X chromosome-linked gene, this will result in the functional loss of the respective protein in all cells of a male, but only in half of the cells in a female, resulting in differing regulatory responses and capacities. Finally, many mechanisms can hypothetically explain the better prognosis for females in acute inflammation (as infections), such as the expression of genes located on the non-recombining regions of the Y chromosome, sex hormone-mediated effects, differences in X-linked gene expressions of maternal or paternal origin, gene-dosage effects of sex chromosome-linked genes (namely, those genes that escape X chromosome inactivation or are reactivated), non-random X chromosome inactivation, and, finally, cellular mosaicism of females. The genes encoding some protein members of the TLR signaling pathway are linked to the X chromosome, such as IL-1 receptor-associated kinase 1, NF-κB essential modulator, and Bruton's tyrosine kinase, with the [fig_ref] FiGURe 1 |: Protein kinases [/fig_ref] adapted from the study by Akira et al.
Prepubescent children, displaying very low levels of sexual hormones, prove to represent a good model to evaluate the inflammatory response and clinical course of acute or chronic inflammatory diseases. We previously showed that in acute inflammation caused by pneumonia, pyelonephritis or bronchiolitis inflammatory markers (C-reactive protein or erythrocyte sedimentation rate) and neutrophil count were higher in females [bib_ref] Chronic inflammatory diseases in children are more severe in girls, Casimir [/bib_ref] , suggesting better inflammatory recruitment in females during acute inflammation. This sexual dimorphism has been observed in all major disease categories, except for diseases of the musculoskeletal system and connective tissue in children younger than 20 years, with a higher mortality reported in males compared to females [bib_ref] The X chromosome in immune functions: when a chromosome makes the difference, Libert [/bib_ref].
As similar observations were made by others in adults [bib_ref] Gender differences in inflammatory markers in children, Casimir [/bib_ref] [bib_ref] Pediatric mortality in males versus females in the United States, Balsara [/bib_ref] [bib_ref] Race and sex differences in C-reactive protein levels, Khera [/bib_ref] , this could explain the high number of studies reporting higher infections rates in males of any age, along with worse prognosis [bib_ref] The erythrocyte sedimentation rate, Olshaker [/bib_ref] [bib_ref] Influence of age and sex on the 28-joint Disease Activity Score (DAS..., Radovits [/bib_ref]. In septic shock, gender differences have commonly been reported, indicating that the males belong to most at-risk group (especially black in the United States) [bib_ref] Bronchiolitis in US emergency departments 1992 to 2000: epidemiology and practice variation, Mansbach [/bib_ref] [bib_ref] The influence of age and sex on the population-based incidence of communityacquired..., Gutiérrez [/bib_ref]. In other studies, women with sepsis exhibited lower age-specific incidence and mortality rates (32), being less frequently affected than males although with variable prognosis. This may perhaps be accounted for by underlying conditions, such as chronic respiratory failure, diabetes, or metastatic cancer, or by infection sites [bib_ref] Influence of sex on the outcome of severe sepsis: a reappraisal, Adrie [/bib_ref] [bib_ref] Epidemiology of severe sepsis in the United States: analysis of incidence, outcome,..., Angus [/bib_ref] [bib_ref] Systematic review of sex-dependant outcomes in sepsis, Papathanassoglou [/bib_ref].
In children suffering from severe sepsis, we recently showed that girls tended to exhibit higher neutrophilic inflammation, longer fever duration, and lower pH on admission [bib_ref] Sex differences in inflammatory response and acid-base balance in prepubertal children with..., Lefevre [/bib_ref]. In this study, the difference in neutrophil counts became significant on the third day after admission corresponding to the mean generation time of myelocytes. This observation points toward the origin of the difference being in the bone marrow rather than the marginated pool of neutrophils. In a previous work, we showed a different kinetic between males and females in terms of inflammatory cytokine production in whole blood stimulated with endotoxin, which could account for this gender difference becoming significant only on the third day from the beginning of the sepsis. The higher level of circulating neutrophils in girls could also contribute to the better pathogen clearance in girls during the early inflammatory response and consequently their better survival of sepsis. This is the first description of genderrelated differences in the acid-base balance of children with sepsis. We have observed a significantly lower pH associated with a higher base deficit in girls at admission to the PICU. This difference could enhance the inflammatory response in females by increasing the expression of adhesion molecules and production of pro-inflammatory cytokines.
In adult patients with SCA, metabolic acidosis was likewise found with a much higher prevalence in women (52 versus 27% in men; p < 0.001) [bib_ref] Systematic review of sex-dependant outcomes in sepsis, Papathanassoglou [/bib_ref]. Such acidosis, associated with several hemolytic markers and impaired ammonium availability, might contribute to the higher frequency of vasoocclusive crises and acute chest syndromes in girls with SCA (37).
## Genes implicated in the inflammatory process
Inflammation is controlled by a highly coordinated gene expression program (37), involving numerous transcription factors, being potentially influenced by specific variables of the internal milieu like acid-base imbalance. This essential defense mechanism has developed early in metazoan evolution, as indicated by typical inflammatory responses to wounds in invertebrates like the starfish. This mechanism protects and organizes the symbiotic life of various cells in multicellular animals. While inflammation is a strong determinant for restoring the homeostatic balance in the body, it can, however, exceed its usual goals and cause major tissue damage such as in the event of shock sepsis. Such excessive inflammatory processes have accounted for the development of chronic inflammatory and autoimmune diseases. Therefore, an active knowledge and understanding of inflammatory processes appear essential, not only for improving the mechanisms' efficiency in several acute diseases but also for preventing deleterious complications in a chronic setting.
In the presence of harmful agents that likely alter the organism's integrity, a highly complex response is set in motion to restore the organism's homeostasis [bib_ref] Origin and physiological roles of inflammation, Medzhitov [/bib_ref]. Genes involved in environmental and inflammatory responses have been shown to display an unusually high rate of duplication and loss during evolution [bib_ref] The functional repertoires of metazoan genomes, Ponting [/bib_ref]. Recent technological advancements provide a clearer picture of the organizational principles underlying inflammatory gene expression. During the inflammatory response triggered upon stimulation (infection, burn, surgical procedure, etc.), several hundred genes are activated in a kinetically complex manner, either synchronously or only after many hours for some of them. Inducible recruitment of target genes, such as NF-κB, can apparently be influenced by a pre-existing chromatin state [bib_ref] Selective transcription in response to an inflammatory stimulus, Smale [/bib_ref]. The requirement for a chromatin-remodeling step at inflammatory genes has been shown to cause slower activation kinetics, while imposing the presence of additional transcription factors that induce the initial remodeling step. Many of these transcription factors involved in inflammatory processes can be selectively stimulated by a specific inflammation trigger, which lays the groundwork for stimulus specificity in genetic inflammation expression.
It seems now clear that genes activated by an identical trigger may differ extensively depending on cell types, even when the same cytokines are involved. The genomic regions that are active as enhancers in different cell types show only a slight overlap [bib_ref] Histone modifications at human enhancers reflect global cell-type-specific gene expression, Heintzman [/bib_ref]. Macrophages contain at least 35,000-45,000 identifiable genomic regions that are presently classified as enhancers. Regulatory mechanisms that control the inflammatory process designed to lessen potential tissue damage must have been positively selected in the course of evolution. A central role is played by the B-cell lymphoma 6 protein (BCL6), a sequence-specific transcriptional repressor known for its role in both B-cell differentiation and B-cell lymphomas. The BCL6 has been shown to prevent excessive production of a large fraction of lipopolysaccharide-inducible genes [bib_ref] Bcl-6 and NF-κB cistromes mediate opposing regulation of the innate immune response, Barish [/bib_ref]. Currently, the entire set of players involved in the inflammatory response is still incompletely defined and thus largely unknown.
In immune challenges, especially if acute, females exhibit a better prognosis and survival than males at any age. In inflammatory processes, the immune response relies on the heterogeneity of immune cells, along with their ability to respond to pathogen challenges [bib_ref] Epigenetic control of immunity, Busslinger [/bib_ref] , with lymphocytes displaying a highly diverse antigen receptor repertoire that matches pathogen diversity. In addition, the inflammatory response is under the influence of epigenetic regulation, which requires flexible adaptation to diverse environmental challenges like pH variations.
## Concluding remarks
There is evidence favoring the existence of links between acidbase balance and cytokine concentrations, with acidosis as potential unifying factor for the trigger threshold of the inflammatory response. Gender differences in the inflammatory response could be linked to the acid-base balance of the cellular environment that influences the expression of genes related in particular to the X chromosome. Endothelial cells may play a fundamental role in this process by sensing acid-base fluctuations. Further understanding of their potent role in the initiation of the inflammatory cascade could help design new strategies to interfere with the inflammatory process.
# Author contributions
The review results from the discussion and the consensus of all authors listed (GC, NL, FC, JD, and MC). Literature review on the topic was analyzed and produced by GC, NL, and MC. The review was written by GC.
## References
[fig] FiGURe 1 |: Protein kinases (circled in red) encoded by X-linked genes and involved in the TLR4 signaling pathway, adapted from the study byAkira et al. (22). An example of TLR4 signaling pathways is shown to highlight the implication of key X chromosome-linked kinases in the triggering of the inflammatory response. AP1, activator protein 1; BTK, Bruton's tyrosine kinase; c/EBPβ, CCAAT/enhancer-binding protein β; ECSIT, evolutionary-conserved signaling intermediate in Toll pathway; ERK, extracellular signal-regulated MAP kinase; IKK, IκB kinase; IκB, inhibitor kappa B; IRAK, interleukin-1 receptor-associated kinase; JNK, c-Jun N-terminal kinase; MAL, myelin and lymphocyte; MAP, mitogen-activated protein; MD2, lymphocyte 96 antigen; MEKK, MAP/ERK kinase kinase; MKK, MAP kinase kinase; MSK, mitogen-and stress-activated kinase; MYD88, myeloid differentiation primary response 88; NEMO, NF-κB essential modulator; NF-κB, nuclear factor kappa B; p38, p38 MAP kinase; p50, NF-κB subunit 1; p65, NF-κB subunit 3 or RELA, v-rel avian reticuloendotheliosis viral oncogene homolog A; RIP2, receptor interacting protein-2; TAB, TAK1-binding protein; TAK, TGF-β-activated kinase; TIR, Toll/interleukin-1 receptor; TRAF, tumor necrosis factor receptor-associated factor; TRAM, translocating chain-associating membrane; TRIAD3/RNF216, ring finger protein 216; TRIF, TIR-domain-containing adapter-inducing interferon-β; TOLLIP, Toll interacting protein; UBC13, ubiquitin-conjugating enzyme 13; UEV1A, ubiquitin-conjugating enzyme variant 1A. [/fig]
|
Caspase-1 is active since the early phase of rheumatoid arthritis
# Introduction
Rheumatoid arthritis (RA) is a chronic autoimmune disease characterized by joint destruction. We have previously reported an increase in interleukin (IL)-1β levels in the serum of very early RA and in the synovial fluid (SF) of RA patients. Therefore, the molecular mechanisms that regulate IL-1β activation, such as caspase-1, might play a relevant role and have potential as therapeutic targets in RA.
## Aim
The main goal of this study was to evaluate the state of activation of caspase-1 in early polyarthritis and established RA patients.
# Patients and methods
Blood samples were collected from 12 untreated early polyarthritis patients with less than 12 months of disease duration, 12 methotrexate (MTX) treated established RA patients and from 10 healthy controls. Caspase-1 activity was assessed using the Caspase Assay kit.
# Results
We found that both early polyarthritis and established RA patients have higher levels of active caspase-1 than healthy controls. Surprisingly, caspase-1 activation levels in patient cells remain unchanged when cells were stimulated with E. coli and ATP for 4h and 24h, whereas in healthy control cells caspase-1 levels are significantly increased after stimulation. We observed no significant differences when comparing early polyarthritis with established RA samples. However, in RA samples from patients with higher disease activity (DAS28>5.2) the active caspase-1 ratio is higher.
# Conclusions
Our results suggest that leukocytes in both early polyarthritis and established RA patients have high baseline levels of activated caspase-1, which cannot be significantly increased by bacterial and ATP stimulation. We also show that MTX treatment does not affect caspase-1 activation. Altogether, our data supports the hypothesis that caspase-1 is already activated in the early phase of RA. |
Grand scale genome manipulation via chromosome swapping in Escherichia coli programmed by three one megabase chromosomes
In bacterial synthetic biology, whole genome transplantation has been achieved only in mycoplasmas that contain a small genome and are competent for foreign genome uptake. In this study, we developed Escherichia coli strains programmed by three 1-megabase (Mb) chromosomes by splitting the 3-Mb chromosome of a genome-reduced strain. The first split-chromosome retains the original replication origin (oriC) and partitioning (par) system. The second one has an oriC and the par locus from the F plasmid, while the third one has the ori and par locus of the Vibrio tubiashii secondary chromosome. The tripartite-genome cells maintained the rod-shaped form and grew only twice as slowly as their parent, allowing their further genetic engineering. A proportion of these 1-Mb chromosomes were purified as covalently closed supercoiled molecules with a conventional alkaline lysis method and anion exchange columns. Furthermore, the second and third chromosomes could be individually electroporated into competent cells. In contrast, the first splitchromosome was not able to coexist with another chromosome carrying the same origin region. However, it was exchangeable via conjugation between tripartite-genome strains by using different selection markers. We believe that this E. coli-based technology has the potential to greatly accelerate synthetic biology and synthetic genomics.GRAPHICAL ABSTRACT
# Introduction
Building and booting-up synthetic genomes is a powerful bottom-up approach for understanding and engineering living systems [bib_ref] Technological challenges and milestones for writing genomes, Ostrov [/bib_ref] [bib_ref] Synthetic genomes, Zhang [/bib_ref] [bib_ref] Building genomes to understand biology, Coradini [/bib_ref] [bib_ref] Budding yeast as a factory to engineer partial and complete microbial genomes, Vashee [/bib_ref]. To construct bacteria programmed by synthetic genomes, the J. Craig Venter Institute built circular chromosomes in Saccharomyces cerevisiae by assembling DNA fragments [bib_ref] Complete chemical synthesis, assembly, and cloning of a Mycoplasma genitalium genome, Gibson [/bib_ref] [bib_ref] One-step assembly in yeast of 25 overlapping DNA fragments to form a..., Gibson [/bib_ref] [bib_ref] Design and synthesis of a minimal bacterial genome, Hutchison [/bib_ref] , and then installed those synthetic genomes into cells of a recipient bacterium that was phylogenetically similar to the species on which the synthetic chromosome was based [bib_ref] Design and synthesis of a minimal bacterial genome, Hutchison [/bib_ref] [bib_ref] Creating bacterial strains from genomes that have been cloned and engineered in..., Lartigue [/bib_ref]. To date this genome transplantation (GT) approach has only worked for a small closely related group of Mycoplasma species [bib_ref] Budding yeast as a factory to engineer partial and complete microbial genomes, Vashee [/bib_ref]. In those instances, Mycoplasma capricolum was converted to a related bacterium such as Mesoplasma florum [bib_ref] Creating bacterial strains from genomes that have been cloned and engineered in..., Lartigue [/bib_ref] [bib_ref] Impact of donor-recipient phylogenetic distance on bacterial genome transplantation, Labroussaa [/bib_ref] [bib_ref] Cloning and transplantation of the Mesoplasma florum genome, Baby [/bib_ref]. M. capricolum is the only bacterium known to have an extraordinarily high competence for foreign genome uptake [bib_ref] Creating bacterial strains from genomes that have been cloned and engineered in..., Lartigue [/bib_ref] [bib_ref] Impact of donor-recipient phylogenetic distance on bacterial genome transplantation, Labroussaa [/bib_ref] [bib_ref] Genome transplantation in bacteria: changing one species to another, Lartigue [/bib_ref].
On the other hand, S. cerevisiae has the potential to maintain a bacterial chromosome as large as 1.8 Mb [bib_ref] Budding yeast as a factory to engineer partial and complete microbial genomes, Vashee [/bib_ref]. Whole genomes of mycoplasmas (up to 1.8 Mb) [bib_ref] Impact of donor-recipient phylogenetic distance on bacterial genome transplantation, Labroussaa [/bib_ref] , Acholeplasma laidlawii (1.5 Mb) [bib_ref] Cloning the Acholeplasma laidlawii PG-8A genome in Saccharomyces cerevisiae as a yeast..., Karas [/bib_ref] , Prochlorococcus marinus (1.6 Mb) [bib_ref] Sequence analysis of a complete 1.66 Mb Prochlorococcus marinus MED4 genome cloned..., Tagwerker [/bib_ref] , and Haemophilus influenzae (1.8 Mb) [bib_ref] Direct transfer of whole genomes from bacteria to yeast, Karas [/bib_ref] were cloned into yeast via a cell-fusion method [bib_ref] Direct transfer of whole genomes from bacteria to yeast, Karas [/bib_ref]. Besides, essential genomes of Escherichia coli [bib_ref] CasHRA (Cas9-facilitated Homologous Recombination Assembly) method of constructing megabase-sized DNA, Zhou [/bib_ref] and Caulobacter crescentus [bib_ref] Chemical synthesis rewriting of a bacterial genome to achieve design flexibility and..., Venetz [/bib_ref] were designed and assembled in yeast. Apart from yeast, the 3.5-Mb whole genome of Synechocystis PCC6803 was assembled and cloned in the Bacillus subtilis chromosome [bib_ref] Combining two genomes in one cell: Stable cloning of the Synechocystis PCC6803..., Itaya [/bib_ref]. These cloned genomes might be rebooted after their transfer to proper host cells, which should be identical with or closely related to the bacterial species from which the cloned genomes derived. Unfortunately, the straightforward GT approach has not yet been achieved in conventional bacteria such as E. coli.
The E. coli genome has been engineered by genome segment swap methods [bib_ref] Design, synthesis, and testing toward a 57-codon genome, Ostrov [/bib_ref] [bib_ref] Defining synonymous codon compression schemes by genome recoding, Wang [/bib_ref] [bib_ref] Total synthesis of Escherichia coli with a recoded genome, Fredens [/bib_ref] [bib_ref] Simultaneous non-contiguous deletions using large synthetic DNA and site-specific recombinases, Krishnakumar [/bib_ref] [bib_ref] Large-scale recoding of a bacterial genome by iterative recombineering of synthetic DNA, Lau [/bib_ref]. Instead of transforming E. coli directly with a synthetic genome, synthetic segments as large as 100 kb replaced the original genomic regions via recombination and were assembled into a fully synthetic genome via conjugation [bib_ref] Total synthesis of Escherichia coli with a recoded genome, Fredens [/bib_ref] [bib_ref] Simultaneous non-contiguous deletions using large synthetic DNA and site-specific recombinases, Krishnakumar [/bib_ref]. These sequential and hierarchical assembly methods successfully produced E. coli strains having a different genetic code [bib_ref] Total synthesis of Escherichia coli with a recoded genome, Fredens [/bib_ref]. One drawback of this approach is that the genomic structures remained unchanged from their template genomes. To achieve more flexible genome design, Wang et al. developed a refined method for chromosome fission and fusion and demonstrated the transfer of a genomic region to another position in the genome [bib_ref] Programmed chromosome fission and fusion enable precise large-scale genome rearrangement and assembly, Wang [/bib_ref]. Despite these innovations, there is a need for more straightforward methods for flexible genome design and for rebooting a heterologous genome in E. coli [bib_ref] CasHRA (Cas9-facilitated Homologous Recombination Assembly) method of constructing megabase-sized DNA, Zhou [/bib_ref].
One approach is the downsizing of the E. coli genome from 4.6 Mb [bib_ref] A guideline and challenges toward the minimization of bacterial and eukaryotic genomes, Kurasawa [/bib_ref] [bib_ref] Emergent Properties of Reduced-Genome Escherichia coli, Pósfai [/bib_ref] [bib_ref] Oxidative stress sensitivity of engineered Escherichia coli cells with a reduced genome, Iwadate [/bib_ref] [bib_ref] Genetic manipulations restored the growth fitness of reduced-genome Escherichia coli, Hirokawa [/bib_ref]. Among the several genome-reduced strains of E. coli, DGF-298W (29) carries the second smallest genome (2.98 Mb), grows vigorously, and accepts various types of genetic engineering. However, the 3-Mb genome is still three times larger than the Mycoplasma mycoides genome (1.08 Mb) that is used in GT reactions [bib_ref] Creation of a bacterial cell controlled by a chemically synthesized genome, Gibson [/bib_ref]. A compromising way of downsizing is to split the single chromosome genome. Several strains of E. coli with two chromosomes have been established by using an additional pair of replication origins (ori) and partitioning loci (par) for the secondary chromosome [bib_ref] Programmed chromosome fission and fusion enable precise large-scale genome rearrangement and assembly, Wang [/bib_ref] [bib_ref] ) migS, a cis-acting site that affects bipolar positioning of oriC on..., Yamaichi [/bib_ref] [bib_ref] Escherichia coli with two linear chromosomes, Liang [/bib_ref] [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref]. During the establishment of the two-chromosome strains of DGF-298W (2 Mb and 1 Mb) (32), we realized that chromosomes of up to 1-Mb in size can be handled in vitro as covalently closed supercoiled molecules and that 0.5-Mb supercoiled chromosomes can be even purified using a commercial DNA extraction kit [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref]. Furthermore, a previous study reported that up to 0.5-Mb synthetic chromosomes can be electroporated into E. coli competent cells [bib_ref] Assembly of eukaryotic algal chromosomes in yeast, Karas [/bib_ref]. These findings tempted us to establish a DGF-298W-derived strain consisting of three 1-Mb split-chromosomes [fig_ref] Figure 1: Design principle of tripartite genomes [/fig_ref] and to assess whether each split-chromosome can be electroporated into E. coli cells.
In the present study, we first identified a good combination of ori-par systems that enabled stable maintenance of three split-chromosomes in E. coli. The 3-Mb chromosome of DGF-298W was safely split into three 1-Mb chromosomes (1.12, 0.84 and 1.02 Mb). Next, the transferability of each split-chromosome was examined via electroporation or via conjugation. The 0.84-Mb and 1.02-Mb chromosomes were purified, electroporated into commercial competent cells of E. coli HST08, and extracted again from the transformed cells. The 1.12-Mb chromosome was successfully purified but was not introduced via electroporation. Upon conjugation, the 1.12-Mb chromosome transferred from the donor strain replaced the 1.12-Mb chromosome in the recipient strain. Thus, we demonstrate chromosome swapping via conjugation [bib_ref] Programmed chromosome fission and fusion enable precise large-scale genome rearrangement and assembly, Wang [/bib_ref] and chromosome implantation via electroporation in the 1/3 genome-scale in E. coli.
# Materials and methods
## Agarose gel electrophoresis analyses
The pulsed-field gel electrophoresis (PFGE) analysis of chromosomes was performed in the same manner as before [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref]. In short, the CHEF-DR II and III Pulsed Field Electrophoresis Systems (Bio-Rad) were used for resolving DNA molecules embedded in agarose plugs. The DNA size markers used were S. cerevisiae Chromosomes and Hansenula wingei Chromosomes (Bio-Rad). The canonical agarose gel electrophoresis analysis of chromosomes was performed using 0.5% gels and 0.5× TBE buffer in the same manner as before [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref]. The size marker used was Marker 3 (NIPPON GENE). In both cases, the gels were stained with dsGreen, a SYBR GREEN-I analogue, (Funakoshi) and scanned by using a Typhoon FLA 9500 (GE Healthcare).
## Bac purification and electroporation
Chromosomes were purified by using NucleoBond Xtra BAC kit (Takara Bio) essentially as described before [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref] but with a few modifications. Cells were harvested from 750 Nucleic Acids ml of LB cultures incubated overnight at 37 - C using an IN-NOVA 42 (Eppendorf). Cell extracts from cells corresponding to OD 600 = 1500 at maximum were applied to each column of anion exchange resin. After elution of DNA from a column by using 15 ml of Buffer ELU-BAC, the eluate was transferred to six 5-ml Eppendorf tubes and subjected to isopropanol precipitation by using 1 l of Ethachinmate (NIPPON GENE) as a carrier solution. After the isopropanol precipitation step, the DNA pellets were air-dried for 15 min and dissolved with 83 l of TE buffer (pH 8.0) (NIPPON GENE) overnight in the refrigerator. The eluate was transferred to a new Lo-bind tube by using a wide bore tip and stored in the refrigerator. For electroporation, a 2 l aliquot of purified chromosome solutions was mixed with 50 l of E. coli HST08 Premium Electro-Cells (Takara Bio) by pipetting five times using a wide bore tip. After incubation on ice for 1 minute, the cell-DNA mixture was transferred into a 0.1 cm gap electroporation cuvette (Bio-Rad), and electroporated by using an ELEPO21 (NEPA GENE). The condition of electroporation was set as follows; poring pulse (voltage: 1600 V, pulse width: 3.5 ms, pulse interval: 50.0 ms, number of pulses: 1, polarity: +) and transfer pulse (voltage: 100 V, pulse width: 50.0 ms, pulse interval: 50.0 ms, number of pulses: 3, polarity: +/-). Four vials of HST08 competent cells were used for electroporating each chromosome to obtain enough number of positive colonies.
## Plasmid construction and cell engineering
Plasmids were developed by using PrimeStar Max (Takara Bio) and KOD One Master Mix (TOYOBO) for PCR, In-Fusion (Clontech) for in vitro assembly, and chemical competent cells of HST08/Stellar cells (Takara/Clontech) for transformation. Some of the plasmids and BAC vectors were developed previously [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref] , and the others were newly developed by local modifications (see [fig_ref] Table 1: Plasmids, BAC vectors and chromosomes used in this study name Description pMW118-Aba... [/fig_ref]. Some of the DNA cassettes used for the Red recombination were developed previously [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref] , while the others were newly prepared. The information for new DNAs is provided as supplementary. The plasmid and chromosome maps and sequences were made using SnapGene software. The sequence files of 1-Mb chromosomes were made using the nearlycomplete draft genome sequence data of DGF-298W (32). The Red recombination was performed essentially as described previously (32) but with a new helper plasmid mediating the recombination. The new helper plasmid encodes the recA gene in addition to the bet-exo genes on pMW118 under control of the arabinose promoter, because recAdeficient derivatives of DGF-298W were used in this study. The antibiotic concentrations were 100 g/ml for carbenicillin, 15, 25, 30 g/ml for kanamycin, 17, 25, 34 g/ml for chloramphenicol, 100 g/ml for spectinomycin, Zeocin, and Blasticidin S, 5, 7, 10 g/ml for tetracycline, 3.5 or 7 g/ml for gentamicin, and 50 or 100 g/ml for hygromycin. The colony-direct PCR was performed by using GoTaq Green Master Mix (Promega). The Flp-POP cloning was performed as described previously [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref] , but with a choice of two helper plasmids carrying either of the spectinomycin and gentamicin selection markers. L-Arabinose was added to growth media to induce the expression of flippase and HK022 phage integrase, as previously described [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref]. For the self-circularization of the 1.12-Mb region, two FRT cassettes were inserted between the polB gene and the leuD gene and between the lysS gene and the yqeF gene. The conjugal transfer of chromosomes was performed in the same manner as before by using the helper plasmid pBAD-traRP4min [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref]. L-Arabinose was added to growth media to induce the expression of the conjugation apparatus, as previously described [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref]. A strain list is provided as [fig_ref] Table 2: E [/fig_ref].
## Phenotyping of tripartite-genome strains
The growth rates of the wildtype DGF-298W strain and two lines of tripartite-genome strains RGF138 and RGF140 in antibiotic-free LB media in L-shaped tubes were measured in an automatic manner by using a compact rocking incubator TVS062CA (Advantec). The growth condition was set as follows; temperature: 37 - C, shaking speed: 60 rpm, measurement while shaking, waiting time for measurement: 10 s, measurement interval: 10 min. The fluorescence microscope observation of the three strains was performed by using a ZEISS Axio Observer (ZEISS) in the dark room. The cells were cultured in antibiotic-free LB media in test tubes at 37 - C until reaching OD 600 = 0.2. Samples were prepared by mixing 5 l of the fresh cell cultures and 5 l of 10 g/ml DAPI in methanol and incubated for 4 min in the dark. Samples were observed by an agar pad method. A pad of 1.5% STAR Agar L-grade 01 (RIKAKEN) was made on a slide glass by sandwiching between two slide glasses using vinyl tape as spacers. A 0.5 l aliquot of samples was placed on the agar pad, incubated for 5 min in the dark at room temperature, and covered with a cover glass. The sample images were taken under 100× Oil objective for TL Phase and DAPI, and overlaid using NIH ImageJ 1.52k. Cell lengths were measured using ImageJ with a set scale (distance in pixels: 1, known distance: 0.063, Pixel aspect ratio: 1.0, unit of length: m). The bacterial viability test was performed using SYTO 9 and propidium iodide (PI). DGF-298W, RGF138, and RGF140 cells were cultured in antibiotic-free LB media in test tubes with vigorous shaking at 37 - C. Cells in the exponential growth phase (OD 600 = 0.3) were fixed and stained for live/dead microscopy.
# Results
## Design principle of tripartite genomes
For splitting the DGF-298W genome, we needed three oripar systems that can be used simultaneously for the replication and partitioning of three split-chromosomes. While the E. coli genome carries a typical origin of replication (oriC), the chromosome partitioning mechanism is not yet fully understood but is different from other bacteria utilizing the typical parABS system [bib_ref] ParB of Pseudomonas aeruginosa: interactions with its partner ParA and its target..., Bartosik [/bib_ref]. For generating a subchromosome in E. coli, two types of par systems have been used. One is the par locus of the E. coli F plasmid [bib_ref] Mini-F plasmid genes that couple host cell division to plasmid proliferation, Ogura [/bib_ref] , or sopABC, while the other is the parABS2 system derived from the ori and par locus of the Vibrio secondary chromosome (the ori2-par2 locus) [bib_ref] Escherichia coli with two linear chromosomes, Liang [/bib_ref]. Recently, we demonstrated that a 1.02-Mb region popped-out from the 2.98-Mb genome of DGF-298W can be maintained when fused with either a sopABC-oriC pair or the ori2-par2 system of Chr LR (pPKOZ-attB-spec) purified from RGF152 or YST01 The cloned regions: from leuD to ybgL through rrnH and from yqeF to menF through rrnG. The selection markers: kanamycin, spectinomycin, and tetracycline Chr Ter (HF053)
Chr Ter (pVtu9xF) purified from HF053 The cloned region: from gltA to glpC through dif The selection marker: kanamycin Chr Ori (HF054) Chr Ori purified from HF054 The cloned region: from lysS to polB through oriC with endA::tet The selection markers: tetracycline, hygromycin, chloramphenicol, and Zeocin * These plasmids were reported previously (32). Vibrio tubiashii [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref]. This meant that we have at least two reliable par systems and that the 1.96-Mb chromosome was controlled by the native chromosome partitioning mechanism. Another thing to be considered is chromosome structuring (36) such as the macrodomain structure and the chromosome symmetry. It is known that the genome chromosome of E. coli is organized into multiple macrodomains. The Ori macrodomain includes the oriC locus, most of the ribosomal RNA operons, and the migS (30) and maoSP loci [bib_ref] The MaoP/maoS site-specific system organizes the Ori region of the E. coli..., Valens [/bib_ref] , which are considered as important for the partitioning and organization of this macrodomain. The Ter macrodomain is organized by the MatP/matS system [bib_ref] The MatP/matS site-specific system organizes the terminus region of the E. coli..., Mercier [/bib_ref] and includes the chromosome dimer resolution site (dif) [bib_ref] Post-replicative pairing of sister ter regions in Escherichia coli involves multiple activities..., Crozat [/bib_ref]. For simplicity, the two other macrodomains located in the left and right sides of the genome were defined as Left and Right macrodomains in this study, respectively. The directions of highly transcribed genes and regulatory DNA motifs (such as KOPS and ter motifs) along the Ori-Ter axis are also important for replication fork processivity and for chromosome segregation [fig_ref] Figure 1: Design principle of tripartite genomes [/fig_ref] [bib_ref] Chromosome structuring limits genome plasticity in Escherichia coli, Esnault [/bib_ref]. In particular, the position and direction of Tus-ter replication fork traps is critical [bib_ref] Programmed chromosome fission and fusion enable precise large-scale genome rearrangement and assembly, Wang [/bib_ref] [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref] [bib_ref] Chromosome structuring limits genome plasticity in Escherichia coli, Esnault [/bib_ref] [bib_ref] Escherichia coli Tus protein acts to arrest the progression of DNA replication..., Hill [/bib_ref]. Considering all these things, a safe design principle of tripartite genomes may be individual circularization of the Ori and Ter macrodomains and joining of the Left and Right macrodomains [fig_ref] Figure 1: Design principle of tripartite genomes [/fig_ref]. In this study, we define the Ori, Ter, and Left+Right chromosomes Chr Ori , Chr Ter and Chr LR , respectively.
Nucleic Acids Circularization of the Ori macrodomain A 1.12-Mb region which roughly corresponds to the Ori macrodomain was circularized to assess whether this region alone can be stably maintained by the native chromosome partitioning mechanism. For the pop-out of the remaining 1.86 region, our Flp-POP cloning method (32) was employed [fig_ref] Figure 2: Circularization of the Ori macrodomain by Flp-POP cloning [/fig_ref]. In short, a genomic region flanked by two FRT recombination sites was excised using flippase and fused with a bacterial artificial chromosome (BAC) vector by a phage integrase in a site-specific manner. The flippase-FRT recombination regenerates a full-length chloramphenicol acetyltransferase gene (cat) and thus confers chloramphenicol (Cm) resistance to cells. The pVtu9xT or pPKOZ-attB BAC vectors employ the V. tubiashii ori2-par2 system and the oriC-sop pair, respectively. After induction of the flippase and the phage integrase, recombinant cells were selected on Cm-containing agar plates. Two Cm-resistant colonies for each BAC vector were analysed by pulsed-field gel electrophoresis (PFGE) and confirmed to have two chromosomes of 1.12-Mb (Chr Ori ) and 1.86-Mb (Chr LR+Ter ) [fig_ref] Figure 2: Circularization of the Ori macrodomain by Flp-POP cloning [/fig_ref]. Thus, two types of bipartitegenome strains, RGF123 and RGF124, were established [fig_ref] Table 2: E [/fig_ref]. This result showed that the 1.12-Mb Chr Ori was stably maintained without the aid of any additional par system.
## Developing tripartite-genome strains
Tripartite-genome strains were developed from bipartitegenome strains by repeatedly performing the Flp-POP cloning [fig_ref] Figure 3: Development and establishment of tripartite-genome strains [/fig_ref]. In our previous work (32), we developed bipartite-genome strains carrying two chromosomes of 1.96-Mb (Chr Ori+LR ) and 1.02-Mb (Chr Ter ). The 1.02-Mb region including the Ter macrodomain had been popped-out from the 2.98-Mb genome by using either of the pVtu9xT and pPKOZ-attB BAC vectors to establish RGF093 and RGF094, respectively [fig_ref] Table 2: E [/fig_ref]. To perform the second Flp-POP step, the cat marker generated by the first Flp-POP step was removed from the 1.96-Mb Chr Ori+LR by Red recombination using a tetracycline resistance gene (tet) [fig_ref] Figure 3: Development and establishment of tripartite-genome strains [/fig_ref]. Then, the split cat cassettes were newly inserted into the Chr Ori+LR by Red recombination [fig_ref] Figure 3: Development and establishment of tripartite-genome strains [/fig_ref]. The BAC vectors for the second Flp-POP cloning, pPKOZ-attB-spec and pVtu9xT-PEM7spec, encoded a spectinomycin resistance gene (spec) in addition to the kan gene or instead of the kan gene, respectively. From the 1.96-Mb Chr Ori+LR , the pop-out of the 0.84-Mb region corresponding to the Left and Right macrodomains was attempted by using three combinations of BAC vectors [fig_ref] Figure 3: Development and establishment of tripartite-genome strains [/fig_ref]. First, ori2-par2 and oriC-sop were used for Chr Ter and Chr LR , respectively. Second, oriCsop and ori2-par2 were used for Chr Ter and Chr LR , respectively. Lastly, oriC-sop was repeatedly used for both Chr Ter and Chr LR chromosomes. Two Cm-resistant colonies for each BAC vector combination were analysed by PFGE [fig_ref] Figure 3: Development and establishment of tripartite-genome strains [/fig_ref]. Both colonies had three split-chromosomes when their Chr Ter and Chr LR were powered by ori2-par2 and oriC-sop, respectively. This tripartite-genome strain was named RGF138. When the opposite BAC combination was used, one of the two colonies had three splitchromosomes, while the other had Chr Ori and Chr LR+Ter , which may have emerged by the spontaneous fusion of the Chr Ter and the excised 0.84-Mb region. The former tripartite-genome strain was named RGF140. In contrast, both colonies had Chr Ori and Chr LR+Ter when the oriCsop type BAC vectors were used repeatedly, suggesting that two different types of BAC vectors should be used. Next, the stability of the three-chromosome configurations of the RGF138 and RGF140 strains was investigated by performing another PFGE analysis after >100 generations without antibiotics [fig_ref] Figure 3: Development and establishment of tripartite-genome strains [/fig_ref]. Each of the four sub-strains examined for each strain maintained the three-chromosome configuration [fig_ref] Figure 3: Development and establishment of tripartite-genome strains [/fig_ref] , leading to the conclusion that two kinds of tripartite-genome strains were successfully established.
## Properties of the tripartite-genome strains
The growth profiles and cell shapes of RGF138 and RGF140 were investigated. In an LB growth medium in an L-shaped tube with a vigorous shaking, RGF138 and RGF140 grew about 2 times and 1.5 times more slowly than their parent DGF-298W, respectively [fig_ref] Figure 4: Phenotypes of tripartite-genome strains [/fig_ref]. This result is expected since these strains lack any sophisticated mechanism of regulating and synchronizing the replication and segregation of three split-chromosomes. Next, the cell shapes of RGF138, RGF140 and DGF-298W were observed using microscopy [fig_ref] Figure 4: Phenotypes of tripartite-genome strains [/fig_ref]. Interestingly, DGF-298W and RGF138 tend to have a typical rod-shaped morphology, while a large proportion of RGF140 cells were filamentous. Filamentous cells were often observed in the population of our bipartite-genome strains, probably due to the uncontrolled replication and segregation of the chromosomes. A live/dead microscopy assay indicated that the viabilities of DGF-298W and RGF138 in the exponential growth phase were not significantly different (80% versus 78%), whereas only 53% of the RGF140 cells or filaments were viable.
RGF138 and RGF140 were subjected to further genetic engineering [fig_ref] Figure 4: Phenotypes of tripartite-genome strains [/fig_ref] and [fig_ref] Table 2: E [/fig_ref]. First, they were engineered to eliminate the cat gene on Chr Ori by Red recombination using two copies of Blasticidin S deaminase genes (bsd-bsd), to develop RGF147 and RGF148, respectively. By an unknown reason, RGF140 was highly resistant to the genetic engineering, while RGF138 accepted it. Altogether, RGF138 appeared to have retained wildtype phenotypes in terms of the usability, at the expense of growth rate, and was chosen as the tripartite-genome strain used for further studies. Derivatives of the RGF138 strain was developed by repeating Red recombination [fig_ref] Figure 4: Phenotypes of tripartite-genome strains [/fig_ref]. The endA gene on the Chr Ori of RGF147 was replaced by a Zeocin resistance gene (zeo) together with an oriT sequence, to develop RGF152. Thus, in RGF152, the EndA endonuclease, which degrades DNA during purification [bib_ref] Degradation of RNA during lysis of Escherichia coli cells in agarose plugs..., Khan [/bib_ref] , is missing, and the Chr Ori and Chr Ter are each equipped with the oriT sequence of the RP4 plasmid for enabling their individual conjugal transfer using a helper plasmid pBAD-traRP4min [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref]. Furthermore, the dif sequence on the Chr Ter of RGF152 was replaced by a cat gene together with a dif variant sequence dif2N2 [fig_ref] Table 2: E [/fig_ref] , and the zeo gene at the endA locus was replaced by a hygromycin resistance gene (hyg), to develop RGF159.
## Purification and electroporation of split-chromosomes
Chr LR . Chromosomes were extracted from RGF152 cells by using the NucleoBond Xtra BAC kit (Takara Bio) essentially according to the manufacturer instruction. To increase DNA concentration, DNA pellets from the isopropanol precipitation step were collected into a single tube and later dissolved with TE buffer. As a control and as a size marker, a 530-kb chromosome [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref] was purified in the same manner. The supercoiled forms of extracted chromosomes were resolved by a conventional agarose gel electrophoresis method using a 0.5% agarose gel [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref]. In the gel, we observed a single high molecular weight band (>530 kb) for sample RGF152 [fig_ref] Figure 5: Purification and electroporation of 1-Mb chromosomes [/fig_ref]. This indicated that any or all of the three 1-Mb chromosomes (1.12, 0.84 and 1.02 Mb) were purified in the supercoiled form. Meanwhile, a significant proportion of DNA molecules sheared during the purification process were migrated in the gel to the compression zone of large linear DNA. The purified chromosomes were directly electroporated into E. coli HST08 Premium Electro-Cells (Takara Bio), typical commercial competent cells with a competency of >1 × 10 9 transformants/l pUC19 plasmid. Note that the Chr Ori should confer resistance to Zeocin and Blasticidin S, the Chr LR to kanamycin, spectinomycin, and tetracycline, and the Chr Ter only to kanamycin. Ten colonies showed resistance to kanamycin and tetracycline among hundreds of kanamycin-resistance colonies. No Blasticidin S-resistant colony was obtained, indicating the absence of the Chr Ori . The colony-direct PCR analysis of six Kan R -Tet R colonies and ninety-six Kan R colonies suggested that three Kan R -Tet R colonies might have a whole Chr LR . In contrast, none of the ninety-six colonies had a whole Chr Ter . By using the NucleoBond Xtra BAC kit, the whole Chr LR was purified from cultures of one of the three Kan R -Tet R strains, YST01 [fig_ref] Figure 5: Purification and electroporation of 1-Mb chromosomes [/fig_ref] , while the size of the Chr LR chromosome of YST01 was confirmed by a PFGE analysis (data not shown). These results showed that the 0.84-Mb Chr LR chromosome was purified and elec-troporated. It is likely that the placement of the kan and tet genes at the opposite poles of Chr LR [fig_ref] Figure 5: Purification and electroporation of 1-Mb chromosomes [/fig_ref] facilitated the selection of transformants.
Chr Ter and Chr Ori . To purify Chr Ter and Chr Ori chromosomes individually, we used bipartite-genome strains rather than tripartite-genome strains. Bipartite-genome strains HF053 and HF054 were newly developed via Flp-POP cloning by using a smaller BAC vector pVtu9xF. HF053 has Chr Ori+LR and Chr Ter , while HF054 has Chr Ori and Chr LR+Ter . The 1.02-Mb Chr Ter and the 1.12-Mb Chr Ori purified from HF053 and HF054 cells, respectively, migrated slightly more slowly than the 0.84-Mb Chr LR in an agarose gel on agarose gel electrophoresis [fig_ref] Figure 5: Purification and electroporation of 1-Mb chromosomes [/fig_ref]. Interestingly, the Chr Ori+LR and Chr LR+Ter were not co-purified with the Chr Ter and Chr Ori , probably because the 1.86-Mb and 1.96-Mb chromosomes were too large and fragile. This result demonstrated that circular chromosomes of up to 1.1 Mb in size can be purified by using the NucleoBond Xtra BAC kit, which employs alkaline lysis method, anion exchange columns, and isopropanol precipitation. Next, HST08 competent cells were electroporated using the purified Chr Ter and Chr Ori chromosomes. Again, no colony contained the Chr Ori . On the other hand, 6 colonies among 110 Kan R colonies were indicated to carry a whole Chr Ter according to their colony direct PCR analysis. A PFGE analysis of cultures of the 6 colonies (or YST03 strains) confirmed the existence and the size of the whole Chr Ter [fig_ref] Figure 5: Purification and electroporation of 1-Mb chromosomes [/fig_ref]. This result clearly proved that circular chromosomes of up to 1.0 Mb in size can be electroporated into E. coli competent cells.
## Partial genome swap via conjugation
The remaining question was whether the 1.1-Mb Chr Ori was too large for electroporation or just unacceptable for the cell. It is very likely that the replication and partitioning systems of the 1.1-Mb Chr Ori would interfere with that of the Nucleic Acids (C) PFGE analysis of the three-chromosome configurations after >100 generations. Descendants of both RGF138 and RGF140 strains maintained their tripartite-genome configuration even after >100 generations without antibiotics. The chromosomes were resolved by using 0.8% gel and 1× TAE buffer (lower gel). The absence of fused chromosomes means that the three-chromosome configuration was maintained. genome chromosome of the host cell and that the duplication of the highly-transcribed genetic loci such as the rrnD operon (42) would be a significant burden to the host cell. To assess this hypothesis, an extra Chr Ori was transferred via oriT-mediated conjugation between tripartite-genome strains from RGF152 to RGF159 . Remember that the Chr Ori of RGF152 has the zeo gene, while the Chr Ori of RGF159 has the hyg gene. The zeo gene is the last gene to be transferred upon conjugation. In addition, the Chr Ter of the recipient strain RGF159 has the cat gene. After conjugation, no Zeo R -Hyg R -Cm R colony was obtained, while Zeo R -Cm R colonies were obtained. This indicated that the Chr Ori of RGF152 was successfully transferred to RGF159 via conjugation and then replaced the original Chr Ori . On the other hand, preliminary attempts of the conjugal transfer of Chr Ori to HST08-derived cells were not successful (data not shown), implying that the Chr Ori may have been unacceptable for the wildtype E. coli. Furthermore, we performed the swapping of the Chr Ter in RGF152 via conjugation by the Chr Ter of RGF159 containing the cat gene . The incompatibility between the two ori2-par2 chromosomes may have driven the Chr Ter swapping.
# Discussion
We showed that once the chromosomes are down to 1 Mb in size they can be purified and directly electroporated into E. coli competent cells. It is likely that the physical compactness of the supercoiled 1-Mb chromosomes may have enabled their purification with columns and their transfer into E. coli cells via electroporation. On the other hand, large circular DNA molecules extracted in a traditional manner from cells digested in agarose plugs are often nicked/gapped and relaxed [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref]. These findings may explain why E. coli had not shown such competency for genome-size DNA uptake. In other words, supercoiling of bacterial chromosomes is a key technology for synthetic biology and synthetic genomics. Previously, we have reconstituted the chromosome-replication cycle reaction (RCR) system of E. coli [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref] [bib_ref] Exponential propagation of large circular DNA by reconstitution of a chromosome-replication cycle, Su'etsugu [/bib_ref]. The RCR system has The split-chromosomes of RGF152 were purified with NucleoBond Xtra BAC kit, analysed by agarose gel electrophoresis, electroporated into HST08. A 127.5 ng aliquot and 4 × 51 ng aliquots of purified DNA (total 21 g) were used for electrophoresis and electroporation, respectively. By using four vials of commercial electrocompetent cells of HST08, two positive colonies were obtained. The 0.84-Mb Chr LR chromosome was purified from YST01, one of the Chr LR -transformed HST08 strains. The agarose gel electrophoresis analyses were performed using 0.5% agarose, 0.5× TBE, at 60 V for 65 min. The DNA size markers used were Marker 3 (/HindIII + /EcoRI digest mixture) and a 530-kb chromosome purified in the same manner. (B) The 1-Mb split-chromosomes of HF053 and HF054 were purified and analysed in the same manner as in (A). (C) PFGE analysis of the Chr Ter chromosomes of six strains of YST03, HST08 transformed with the Chr Ter chromosome purified in (B). been used for the in vitro amplification of oriC-containing chromosomes and for the supercoiling of circular chromosomes [bib_ref] Overcoming the challenges of megabase-sized plasmid construction in Escherichia coli, Mukai [/bib_ref] [bib_ref] Exponential propagation of large circular DNA by reconstitution of a chromosome-replication cycle, Su'etsugu [/bib_ref]. Thus, our methods may be applicable to circular chromosomes assembled in vitro, purified from yeast, or derived from other bacteria.
It was implied that the combination/choice of the BAC vector backbones affected the replication and partitioning of the split-chromosomes. RGF138 and its derivative strains were relatively healthy, whereas RGF140 cells became filamentous and sick. We assume that this is in part due to a difference in replication timing between the ori2-and oriC-BAC vectors. The initiation at oriC is regulated to occur synchronously in E. coli cells bearing multiple oriC sites [bib_ref] Regulation of the replication cycle: conserved and diverse regulatory systems for DnaA..., Katayama [/bib_ref] , whereas the ori2 initiation occurs after the oriC (ori1) initiation in the two-chromosome carrying Vibrio species [bib_ref] Synchronous termination of replication of the two chromosomes is an evolutionary selected..., Kemter [/bib_ref] [bib_ref] A checkpoint control orchestrates the replication of the two chromosomes of Vibrio..., Val [/bib_ref]. RGF138 carries the Ter region on the late replicating ori2-BAC vector, while the Ter region is on the early replicating oriC-BAC vector in RGF140. The earlier timing of the Ter region replication might disturb cell cycle and induce the filamentous and sick phenotype [bib_ref] Synchronous termination of replication of the two chromosomes is an evolutionary selected..., Kemter [/bib_ref] [bib_ref] The dif resolvase locus of the Escherichia coli chromosome can be replaced..., Tecklenburg [/bib_ref]. It would be important to optimize the arrangement of genes and regulatory loci in the multipartite genome, because the copy numbers, intracellular locations, and local topologies of each genomic loci may be different from those in the wildtype E. coli.
There are many technological challenges towards multistep implantation of split-chromosomes into the same E. coli cells. The size and number of split-chromosomes should be reduced. It is ideal if each of the split-chromosomes could be easily transferred to cloning cells and stably maintained in the cells. For example, further engineering of the Chr Ori chromosome, such as addition of a strong par system and elimination of a few ribosomal genes, would produce a more portable Chr Ori . Furthermore, the electroporation efficiency of chromosomes must be improved. Although we . Partial genome swapping via conjugation. The donor strain carries a helper plasmid for conjugation, pBAD-traRP4min, indicated with a red circle. The Chr Ori and Chr Ter chromosomes contain an oriT sequence whose direction of transfer is indicated with red arrows. (A) The Chr Ori of RGF159 was swapped by the Chr Ori of RGF156 (RGF152 with the helper plasmid). Note that the zeo marker is the last gene transferred via conjugation from the oriT sequence. (B) The Chr Ter of RGF152 was swapped by the Chr Ter of RGF161 (RGF159 with the helper plasmid). The resultant strains RGF160 (A) and RGF162 (B) shared the same genotype. Selection marker gene loci were confirmed by colony PCR analyses as shown in the bottom panels (A, B). obtained 36 Chr LR -carrying colonies by using four vials of HST08 electrocompetent cells, we have not yet succeeded in the electroporation of any 1-Mb chromosome using selfmade competent cells of a DGF-298-derived strain. Increasing the concentration and purity of the supercoiled form of chromosomes would help increase the efficiency.
[fig] Figure 1: Design principle of tripartite genomes. In the zoom-in view, the ribosome RNA operons (rrnABCDEHG), oriC and dif, the ter sequences, and the eight genes at the borders are indicated. [/fig]
[fig] Figure 2: Circularization of the Ori macrodomain by Flp-POP cloning. (A) Scheme for the pop-out of Chr LR+Ter by Flp-POP cloning and the concurrent generation of Chr Ori . The resultant bipartite-genome strains gained resistance to chloramphenicol via the fused resistance gene. Two types of BAC vectors, pVtu9xT and pPKOZ-attB, were used to develop bipartite-genome strains RGF123 and RGF124, respectively. (B) PFGE analysis of the chromosomes of two colonies of RGF123 [Chr Ori & Chr LR+Ter (pVtu9xT)] and two colonies of RGF124 [Chr Ori & Chr LR+Ter (pPKOZ-attB)]. The DNA size markers used for PFGE are Hansenula wingei chromosomes. [/fig]
[fig] Figure 3: Development and establishment of tripartite-genome strains. (A) Scheme for the development of tripartite-genome strains from bipartite-genome strains. (B) PFGE analyses of split-chromosomes from tripartite-genome strains. For each combination of BAC vectors shown in the table, two Cm-resistant colonies were chosen and analysed by PFGE. The same samples were analysed in two different PFGE conditions; the chromosomes were resolved by using 0.5% gel and 0.5x TBE buffer to separate DNA bands around 1-Mb (upper left gel) or 0.8% gel and 1× TAE buffer to separate DNA bands around 2-Mb (upper right gel). RGF093 with 1.02-Mb Chr Ter and 1.96-Mb Chr Ori+LR and RGF123 with 1.12-Mb Chr Ori and 1.86-Mb Chr LR+Ter were used as controls. [/fig]
[fig] Figure 4: Phenotypes of tripartite-genome strains. (A) Growth of tripartite-genome strains. Cells were incubated at 37 • C with shaking (60 rpm) in antibioticfree LB media using L-shaped tubes. Doubling times were computed from the growth curve data. (B) Overlay of phase contrast and DAPI stained microscopy images of tripartite-genome strains (upper panels) and distribution of their cell lengths (lower panels). Exponential phase cells growing at 37 • C in antibiotic-free LB media were observed and counted. The distribution of cell lengths from 210 DGF-298W cells, 239 RGF138 cells, and 203 RGF140 cells are shown below each picture. (C) A diagram depicting the genetic modifications to RGF138 and RGF140. [/fig]
[fig] Figure 5: Purification and electroporation of 1-Mb chromosomes. (A) [/fig]
[table] Table 1: Plasmids, BAC vectors and chromosomes used in this study name Description pMW118-Aba Helper plasmid for -red recombination with recA and bet-exo cloned under araC-ParaBAD pMW118spec-flp-int * Helper plasmid for Flp-POP cloning with flippase and HK022 int cloned under araC-ParaBAD pMW118gent-flp-int pMW118spec-flp-int derivative with a gent gene pBAD-traRP4min * Helper plasmid for RP4 tra expression, oriT pPKOZ-attB * BAC vector composed of oriC, sopABC, kan, attB pPKOZ-attB-spec BAC vector derived from pPKOZ-attB with a spec gene pVtu9xT * BAC vector composed of V. tubiashii ori2-par2, kan, attB, oriT pVtu9xT-PEM7-spec BAC vector derived from pVtu9xT with a replacement of kan by spec pVtu9xF BAC vector composed of V. tubiashii ori2-par2, kan, attB, a remnant FRT Chr LR (RGF152/YST01) [/table]
[table] Table 2: E. coli strains used in this study RGF008C-derived, [Chr Ori+LR & Chr Ter (pVtu9xT)] RGF094 * RGF008C-derived, [Chr Ori+LR & Chr Ter (pPKOZ-attB)] RGF123 RGF008C-derived, [Chr Ori & Chr LR+Ter (pVtu9xT)] RGF124 RGF008C-derived, [Chr Ori & Chr LR+Ter (pPKOZ-attB)] RGF108 RGF093 derivative with hyg-cat-zeo::dif2v2-tet on Chr Ori+LR RGF109 RGF094 derivative with hyg-cat-zeo::dif2v2-tet on Chr Ori+LR RGF138 RGF108-derived, [Chr Ori & Chr LR (pPKOZ-attB-spec) & Chr Ter ] RGF140 RGF109-derived, [Chr Ori & Chr LR (pVtu9xT-PEM7-spec) & Chr Ter ] RGF147 RGF138 derivative with hyg-cat-zeo::dif201-bsd-bsd on Chr Ori RGF152 RGF147 derivative with endA::oriT-zeo on Chr Ori RGF156 RGF152 transformed with pBAD-traRP4min RGF159 RGF152 derivative with zeo::hyg on Chr Ori and with dif::dif2N2-cat on Chr Ter [/table]
|
Spontaneous extraskeletal osteosarcoma in the neck skeletal muscle of a Crlj:CD1 (ICR) mouse
Extraskeletal osteosarcoma is extremely rare in humans and animals, especially in rodents. This is the first case report on spontaneous extraskeletal osteosarcoma in the neck skeletal muscle of a Crlj:CD1 (ICR) mouse (36 weeks, dead). Necropsy revealed a solid white mass located in the neck skeletal muscle (scalenus muscle). Histological examination showed that the tumor consisted of atypical polygonal cells, a small osteoid clump, and bone tissue. Mitotic figures were observed. Serial sections showed that neoplastic cells lacked clear invasive proliferation to adjacent normal skeletal muscle and continuity with normal bone tissue. Immunohistochemical analysis showed that the neoplastic cells were positive for osteocalcin, osterix, vimentin, and S-100. Based on these results, the tumor was diagnosed as extraskeletal osteosarcoma in the neck skeletal muscle.
Osteosarcoma is a malignant tumor that normally develops in the bone tissue; however, there have been few reports on extraskeletal osteosarcoma (EO) in rats and mice [bib_ref] Nonproliferative and proliferative lesions of the rat and mouse skeletal tissues (bones,..., Fossey [/bib_ref] [bib_ref] Spontaneous extraskeletal osteosarcoma with various histological growth patterns in the abdominal wall..., Ito [/bib_ref] [bib_ref] Spontaneous extraskeletal osteosarcoma in the duodenum of a Crlj:CD1 (ICR) mouse, Ando [/bib_ref] [bib_ref] An extraskeletal osteosarcoma in the auricle of a Wistar Hannover rat, Shiga [/bib_ref] [bib_ref] Extraskeletal osteosarcoma with pulmonary metastasis in a female f344 rat, Nagaike [/bib_ref] [bib_ref] Spontaneous extraskeletal osteosarcoma in the stomach of an aged f344 rat, Okazaki [/bib_ref] [bib_ref] Extraskeletal osteosarcoma with cystic appearance in an aged Sprague-Dawley rat, Yoshizawa [/bib_ref] [bib_ref] An extraskeletal osteosarcoma in an aged rat, Minato [/bib_ref]. Here, we describe the first reported case of spontaneous EO in the neck skeletal muscle of a mouse. The animal was a male Crlj:CD-1 (ICR) mouse (Charles River Laboratories Japan Inc., Kanagawa, Japan) and allocated to the intact group for the toxicological study. In that study, the animals were individually housed in stainless mesh cages; moreover, they were maintained at a temperature of 23 ± 3°C and relative humidity of 50 ± 20% with air ventilation at 17 times/h and 12 h of illumination (0700 h to 1900 h). The animals had free access to a pellet diet (irradiation-sterilized CRF-1 LID 30, Oriental Yeast Co., Ltd., Tokyo, Japan) and water. The experiment was conducted in compliance with the Act on Welfare and Management of Animals The animal died at 36 weeks of age. Visual inspection did not show any abnormal changes. Necropsy revealed a solid white mass (5 × 4 × 4 mm) in the neck skeletal muscle (scalenus muscle) [fig_ref] Figure 1: Macroscopic findings [/fig_ref]. There was no continuity with the normal bone (spine) at necropsy and histological examination using serial sections [fig_ref] Figure 1: Macroscopic findings [/fig_ref]. No other body mass was observed.
The observed mass and adjacent neck skeletal muscle were removed and fixed in 10% neutral-buffered formalin, embedded in paraffin wax, and sectioned at a thickness of approximately 4 µm. Subsequently, the sections were stained with hematoxylin and eosin (H&E). Other mass sections were subjected to immunohistochemistry and phosphotungstic acid hematoxylin (PTAH) staining. The primary antibodies used for immunohistochemistry were as follows: osteocalcin (polyclonal, diluted 1:200, Bioss Antibodies, Woburn, MA, USA), osterix (polyclonal, diluted 1:1500, Abcam, Cambridge, UK), cytokeratin AE1/AE3 (monoclonal, diluted 1:250, Abcam), vimentin (polyclonal, diluted 1:500, Abcam), S-100 (monoclonal, 1:2000, Abcam), α-smooth muscle actin (α-SMA) (monoclonal, 1:2000, Abcam), myogenin (monoclonal, 1:100, Abcam), and PCNA (polyclonal, 1:600; DakoCytomation, Osaka, Japan).
Histologically, the mass was located on the skeletal muscle surface. The neoplastic cells had atypical polygonal nuclei and scant-to-moderate polygonal eosinophilic cytoplasm with a solid growth pattern. Mitotic figures were observed. The proliferative region showed small clumps of osteoid-like eosinophilic matrix and highly differentiated bone tissue [fig_ref] Figure 2: Microscopic findings [/fig_ref]. The mass was not clearly encapsulated; however, there was no clear invasive proliferation of the adjacent skeletal muscle; moreover, lymphocyte infiltration was noted in the boundaries [fig_ref] Figure 2: Microscopic findings [/fig_ref]. No vessel invasion was observed.
Immunohistochemically, the neoplastic cells were posi- tive for osteocalcin, osterix, vimentin, S-100, and PCNA but negative for cytokeratin AE1/AE3, α-SMA, and myogenin [fig_ref] Figure 3: Microscopic findings [/fig_ref]. PTAH staining showed no cross-striations in neoplastic cells [fig_ref] Figure 3: Microscopic findings [/fig_ref]. H&E staining indicated that the tumor was an osteosarcoma, specifically EO. However, based on the affected organ and pleomorphic proliferative pattern, rhabdomyosarcoma was considered a differential diagnosis. Therefore, additional immunohistochemistry and PTAH staining were performed. Osteocalcin and osterix are reliable markers for osteosarcoma, while myogenin is a marker for rhabdomyosarcoma. Some EO cases have been shown to be positive for vimentin, S-100, and α-SMA 2-7 . PCNA was used to determine proliferation activity. Moreover, PTAH staining was used to check for muscle fiber cross-striations in the neoplastic cells of rhabdomyosarcoma. Based on these results, the tumor was diagnosed as an EO in the skeletal muscle of the neck. There are various osteosarcoma subtypes, including eburnating (osteoplastic), chondroblastic, osteoclastic, anaplastic, osteoblastic, fibroblastic, telangiectatic (vascular), and compound (mixed) types [bib_ref] Skeletal system and Teeth, Ernst [/bib_ref] [bib_ref] Bones, joints, and synovia, Long [/bib_ref]. In this study, the tumor was considered the osteoblastic type because it was composed of neoplastic osteoblasts and highly differentiated with varying amounts of osteoid and bone tissues. In conclusion, this report presents the first case of spontaneous EO in the neck skeletal muscle of a Crlj:CD1 (ICR) mouse and describes the morphological and immunohistochemical characteristics of the tumor.
## Disclosure of potential conflicts of interest:
The authors declare that they have no conflicts of interest.
# Acknowledgments:
The authors declare that they have no conflicts of interest.
[fig] Figure 1: Macroscopic findings. A, B. Solid white mass in the neck skeletal muscle (scalenus muscle) (5 × 4 × 4 mm). C. The mass showed continuity only to the neck skeletal muscle, with no mass observed in other organs. [/fig]
[fig] Figure 2: Microscopic findings (HE). A. The neoplastic cells showed a solid growth pattern with bone tissue. Bar = 200 μm. B. The mass consisted of atypical polygonal cells with scant-to-moderate cytoplasm, and formation of small clumps of osteoid-like eosinophilic matrix and highly differentiated bone tissue was observed. Bar = 50 μm. C. Neoplastic cells and lymphocytes were observed near the boundaries without clear invasive proliferation of the adjacent normal skeletal muscle. Bar = 50 μm. [/fig]
[fig] Figure 3: Microscopic findings (immunohistochemistry and special staining). A. Most neoplastic cells were positive for osteocalcin (cytoplasm). Bar = 50 μm. B. Some neoplastic cells were positive for osterix (nucleus). Bar = 50 μm. C. Most neoplastic cells were positive for vimentin (cytoplasm). Bar = 50 μm. D. Most neoplastic cells were positive for S-100 (nucleus and cytoplasm). Bar = 50 μm. E. Some neoplastic cells were positive for PCNA (nucleus). Bar = 50 μm. F. All neoplastic cells showed no muscle fiber cross-striations and were negative for PTAH (inner box: positive control). Bar = 50 μm. [/fig]
|
Barriers towards insulin therapy in type 2 diabetic patients: results of an observational longitudinal study
Background: The course of barriers towards insulin therapy was analysed in three different groups of type 2 diabetic patients. This observational longitudinal study surveyed a three-month follow-up.Methods: Participants in this study totalled 130 type 2 diabetic patients. The first subgroup was on insulin therapy at baseline (group 1: n = 57, age 55.6 ± 8.7 yrs, disease duration 12.7 ± 7.2 yrs, HbA1c 8.5 ± 1.6%) and remained on insulin at follow-up. Of an initial 73 insulin-naïve patients, 44 were switched to insulin therapy (group 2: age 58.1 ± 6.8 yrs, disease duration 7.7 ± 5.0 yrs, HbA1c 9.1 ± 1.7%) and 29 patients remained on an oral regimen (group 3: age 52.7 ± 10.7 yrs, disease duration 5.3 ± 4.6 yrs, HbA1c 8.3 ± 1.4%). Barriers towards insulin therapy were measured using the Insulin Treatment Appraisal Scale (ITAS). As generic instruments of health related quality of life patients completed also the Problem Areas of Diabetes Questionnaire (PAID), the WHO-5 Well-Being Scale (WHO-5), the Centre for Epidemiologic Studies Depression Scale (CES-D) and the Trait Version of the State Trait Anxiety Inventory (STAI) at baseline and at three-month follow-up. Results: At the three-month follow-up, HbA1c had improved in all three groups (7.7 ± 1.2% vs. 7.1 ± 1.1% vs. 6.7 ± 0.8%). The course of negative appraisal of insulin therapy was significantly different in the three groups (p > .003): the ITAS score increased in patients remained on oral antidiabetic drugs (51.2 ± 12.2 to 53.6 ± 12.3), whereas it decreased in patients switched to insulin therapy (49.2 ± 9.8 to 46.2 ± 9.9) or remained on insulin treatment (45.8 ± 8.3 to 44.5 ± 8.0). Diabetes-related distress, trait anxiety, and well-being, showed a similar course in all three groups. The depression score improved significantly in patients switched to insulin treatment compared with patients remaining on insulin therapy. Conclusions: In summary, this study suggests that a negative appraisal of insulin treatment is modifiable by the initiation of insulin therapy. This finding indicates that barriers to insulin are a rather temporary than a stable phenomenon.
# Background
Poor glycaemic control is a risk factor for the development of diabetes-specific complications in diabetic patients. Many type 2 diabetic patients require insulin therapy after several years of disease duration in order to maintain good glycaemic control and prevent complications. But many type 2 diabetic patients do not receive insulin therapy in a timely manner because of a negative appraisal of this treatment option [bib_ref] Concerns about insulin therapy in patients with type 2 diabetes (Abstract), Kulzer [/bib_ref].
Patients' negative attitudes towards starting insulin therapy are based on their beliefs that the need for insulin therapy indicates a greater severity of the disease and proves their failure to self-manage the diabetes adequately. Worries about painful injections and the risk of severe side effects such as hypoglycaemia are also very common among type 2 diabetic patients. Some type 2 diabetic patients fear that insulin therapy is too difficult to manage in everyday life and is associated with social stigma [bib_ref] Development and validation of the insulin treatment appraisal scale (ITAS) in patients..., Snoek [/bib_ref]. In the literature the above-described phenomenon is called psychological insulin resistance [bib_ref] Psychological insulin resistance in patients with type 2 diabetes: the scope of..., Polonsky [/bib_ref] [bib_ref] Resistance to insulin therapy among patients and providers: results of the cross-national..., Peyrot [/bib_ref].
Cross-sectional data demonstrate that barriers to insulin therapy are higher in insulin-naïve type 2 diabetic patients than in insulin-treated type 2 diabetic patients [bib_ref] Development and validation of the insulin treatment appraisal scale (ITAS) in patients..., Snoek [/bib_ref] [bib_ref] Development and validation of a new measure to evaluate psychological resistance to..., Petrak [/bib_ref]. However, from cross-sectional analysis it is difficult to decide whether barriers to insulin therapy exists mainly temporary and reduce by the experience of insulin therapy or whether it is a selection factor caused by the circumstance that only type 2 diabetic patients with a more positive appraisal of insulin therapy will accept this treatment option. From a clinical perspective, each explanation would require a different therapeutic approach. In the first case, diabetes education or better instructions about insulin treatment could be powerful tools to help type 2 diabetic patients coping better with the challenges of insulin therapy and changing their negative appraisal of this treatment option. In the latter case, new types of insulin or different forms of insulin applications (e.g. inhaled insulin) might be helpful to reduce barriers to insulin therapy and encourage a greater proportion of type 2 diabetic patients to use this powerful treatment option. Thus, clearly, longitudinal data on the course of barriers towards insulin therapy after the initiation of insulin therapy are needed to address these issues.
It can also be expected that negative appraisal of insulin therapy in combination with a need for initiation of insulin therapy has a negative impact on other quality of life aspects such as diabetes-related distress, symptoms of depression and anxiety, and psychological well-being among diabetic patients [bib_ref] Development and validation of the insulin treatment appraisal scale (ITAS) in patients..., Snoek [/bib_ref] [bib_ref] Development and validation of a new measure to evaluate psychological resistance to..., Petrak [/bib_ref].
In this observational study we compared the barriers to insulin therapy and more generic quality of life aspects among insulin-treated and insulin-naïve type 2 diabetic patients who had poor glycaemic control at baseline and three months after an intensification of diabetes treatment. In a subgroup of insulin-naïve type 2 diabetic patients, insulin therapy was initiated.
# Methods
Participants in this observational longitudinal study were type 2 diabetic patients who were referred by general practitioners to practices of diabetologists or to the Diabetes Centre Mergentheim. The main reason for referral was unsatisfactory glycaemic control according to the guideline of the German Diabetes Association [bib_ref] Medical antihyperglycaemic treatment of type 2 diabetes mellitus: update of the evidence-based..., Matthaei [/bib_ref].
## Inclusion criteria of the study
- Poor glycaemic control (HbA1c > 6.5% or blood glucose excursions) - Age 18-75 years - Ability to understand the German language - Informed consent given
## Exclusion criteria
- Severe life-threatening disease according to the judgement of a diabetologist - Guardianship
The study was approved by an ethics committee. All patients declared informed consent to participate in this observational study.
## Measures
Appraisal of insulin therapy was measured by the Insulin Treatment Appraisal Scale (ITAS) [bib_ref] Development and validation of the insulin treatment appraisal scale (ITAS) in patients..., Snoek [/bib_ref]. The ITAS, which was designed to assess attitudes towards insulin treatment in type 2 diabetic patients, consists of 20 items. Subjects are requested to indicate on a 5-point Likert scale to what extent they agree with each statement, from "strongly disagree" to "strongly agree". The ITAS is a two-dimensional instrument, with "appraisal of insulin therapy" as a single underlying construct. The instrument permits the calculation of a total score and two subscale scores that measure positive (4 items) and negative (16 items) attitudes towards insulin treatment. Example items representing a positive attitudes towards insulin therapy are: "Taking insulin helps to prevent complications of diabetes" or "Taking insulin helps to maintain good control of blood glucose". Items like "Taking insulin means I have failed to manage my diabetes with diet and tablets" or "Managing insulin injections takes a lot of time and energy" are examples representing a negative attitude towards insulin therapy. The psychometric properties of the ITAS were confirmed recently. The reliability of the total scale was Cronbach's α = .89 (negative appraisal scale Cronbach's α = .90, positive appraisal scale Cronbach's α = .68) and can be regarded as highly satisfactory [bib_ref] Development and validation of the insulin treatment appraisal scale (ITAS) in patients..., Snoek [/bib_ref].
As generic measures of emotional well-being and emotional distress, established depression-, anxiety-, and well-being scales were used. Patients completed the German version of the Centre for Epidemiologic Studies Depression Scale (CES-D) [bib_ref] The CES-D scale: A self report depression scale for research in the..., Radloff [/bib_ref] to assess depressive symptoms. The CES-D has a scale range between 0 and 60; higher scores indicate higher levels of depressive symptoms. Anxiety symptoms were measured by the trait version of the State-Trait-Anxiety-Inventory (STAI). This questionnaire has a scale range between 20 and 80. Higher scores on the Trait-STAI represent higher levels of trait anxiety symptoms. The WHO-5 well-being scale was used to measure well-being. The WHO-5 contains five items, which are all positively worded. A maximum score of "25" indicates optimal well-being, whereas a score of "0" indicates minimal well-being. Diabetes-specific distress was assessed by the German version of the Problem Areas in Diabetes Scale (PAID) [bib_ref] How to screen for depression and emotional problems in patients with diabetes:..., Hermanns [/bib_ref]. The PAID questionnaire consists of 20 items. Each item can be rated on a 5-point Likert scale ranging from "0" (no problem) to "4" (serious problem). According to the recommendation of the test's authors, the PAID scores were transformed to a 0-100 scale, with higher scores indicating more serious emotional problems. The original scale has proved its validity and reliability [bib_ref] The Problem Areas in Diabetes Scale. An evaluation of its clinical utility, Welch [/bib_ref].
Glycaemic control was measured by HbA1c through use of high pressure liquid chromatography. The normal range of HbA1c is 4.1% -6.1%.
# Statistical analysis
Continuous data were analysed by use of parametric methods, and categorical data were analysed by Chi-Square tests. For comparisons of three groups, analyses of variance were used. The analysis of differences between baseline and follow-up measurements was adjusted for baseline values through use of covariance analysis (ANCOVA). Post hoc tests were performed in case of an overall significant effect.
# Results
## Baseline characteristics
Participants in this study totalled 130 type 2 diabetic patients, of whom 73 were insulin-naïve and 57 were insulin-treated. As shown in table 1 insulin-treated and insulin-naïve patients were of a comparable relatively young age and had a similar gender distribution, with fewer women than men. Glycaemic control was rather poor in both groups, as could be expected from the inclusion criteria. Insulin-naïve patients demonstrated a significantly shorter disease duration and had significantly fewer diabetes-associated complications. As could be expected from previous findings, insulin-naïve patients had a significantly more negative appraisal of insulin therapy than patients who already were being treated with insulin. The insulin-treated patients demonstrated a tendency towards higher depression scores (p < .10). On the remaining scales, insulin-treated type 2 diabetic patients had higher scores, a result that indicates a lower health related quality of life. However, these differences failed to reach significance level.
Of the patients receiving insulin therapy, 27 (47.4%) had an insulin monotherapy and 30 patients received a combination therapy of insulin and oral antidiabetic medication. The two subgroups did not differ with regard to baseline HbA1c (8.6 ± 1.9% vs. 8.5 ± 1.5%, p = .74). [fig_ref] Table 2: Baseline characteristics of patients switched to insulin vs [/fig_ref] shows the baseline characteristics of patients switched to insulin vs. patients who remained on treatment with oral antidiabetic agents. Patients who were switched to insulin therapy were significantly older, had a significantly longer diabetes duration, and had more diabetes complications than patients who remained on an oral diabetes regimen. Patients who were switched to insulin therapy had a higher HbA1c level than patients who remained on treatment with antidiabetic drugs. As reported in table 3 there were no significant differences on the ITAS total score between patients who remained on an oral regimen and patients who were switched to insulin therapy (pairwise comparisons between these two groups p = .674).
# Follow up results
All patients who were on insulin therapy at baseline remained on this treatment. The HbA1c level in the patients remaining on insulin treatment fell significantly. Interestingly, this outcome was achieved by increasing the proportion of patients who received a combination treatment of oral antidiabetic agents and insulin. The Values are means (± SD); 1 adjusted for baseline; a = p < .05 between "patients who remained insulin-naïve" and "patients switched to insulin treatment"; b = p < .05 between "patients switched to insulin treatment" and "patients who remained on insulin therapy"; c = p < .05 between "patients who remained insulin-naïve" and "patients who remained on insulin therapy"; * significant within comparison (p < .05), ns non-significant within comparison (p > .05)
proportion of patients on combination therapy with oral antidiabetic agents and insulin rose from 47.4% to 71.9% (McNemar test p < .05). The mean daily insulin dose decreased from 0.88 ± 0.62 IU/kg to 0.63 ± 0.44 IU/kg (p < .01). Body weight was also slightly, but significantly reduced.
Of the patients on oral antidiabetic medication at baseline, 44 (60.3%) were switched to insulin therapy. The decision was made by the treating diabetologist, based on clinical judgement. No patient in this study rejected the insulin therapy option.
The patients switched to insulin therapy received, on average, 0.28 ± 0.23 IU/kg and injected insulin 1.9 ± 1.4 times a day on average. HbA1c improved significantly in these patients. Interestingly, BMI was also significantly reduced (see [fig_ref] Table 3: Baseline and follow-up results in the three different treatment groups [/fig_ref].
In the patients who remained on oral antidiabetic medication, HbA1c also improved significantly.
Although in all three groups there was a significant within effect on HbA1c in the follow-up period, the improvement of glycaemic control was significantly greater in patients who were switched to insulin therapy or remained on an oral regimen than in patients who remained on insulin therapy.
Barriers towards insulin therapy developed significantly different among the three groups. In patients who remained on an oral regimen, the negative appraisal of insulin therapy increased, whereas in patients who were switched to insulin therapy the negative appraisal of insulin therapy was reduced to the level of patients who remained on insulin therapy.
The same pattern of change was present regarding the subscale of negative appraisal of insulin treatment. In the subscale of positive appraisal of insulin therapy the scores were rather stable in all three patient groups.
In patients on an oral regimen at baseline, there was a remarkable improvement of diabetes-related distress regardless of whether those patients remained on an oral regimen or were switched to insulin treatment. Patients who remained on insulin treatment improved slightly. However, there was no significant overall change in diabetes-related distress in the three treatment groups.
General well-being improved slightly in all three groups, but there was no significant difference among these patients groups.
Depressive symptoms were significantly reduced in patients who were switched to insulin therapy compared with patients who remained on insulin therapy.
There was no significant effect on trait anxiety during the follow-up period.
# Discussion
At baseline, more barriers to insulin therapy were demonstrated by insulin-naïve type 2 diabetic patients compared with insulin-treated type 2 diabetic patients. This result is in line with previous findings of cross-sectional studies [bib_ref] Development and validation of the insulin treatment appraisal scale (ITAS) in patients..., Snoek [/bib_ref] [bib_ref] Psychological insulin resistance in patients with type 2 diabetes: the scope of..., Polonsky [/bib_ref] [bib_ref] Resistance to insulin therapy among patients and providers: results of the cross-national..., Peyrot [/bib_ref] [bib_ref] Development and validation of a new measure to evaluate psychological resistance to..., Petrak [/bib_ref]. However, cross-sectional data are difficult to interpret. It is difficult to decide whether the lower level of negative appraisal of insulin therapy among insulin-treated type 2 diabetic patients is a consequence of adaptation to the demands of insulin treatment or whether a selection bias is mainly responsible for this finding. Patients who have a less negative appraisal of insulin treatment might be more likely to accept insulin treatment than patients who have a more objections against this treatment option.
This study provides longitudinal data about the course of negative appraisal of insulin in insulin-naïve and insulin-treated type 2 diabetic patients. Of the insulin-naïve patients, 60% were switched to insulin treatment. At baseline, those type 2 diabetic patients who were switched to insulin therapy and those patients who remained on an oral regimen did not differ with regard to their appraisal of insulin treatment. Thus a selection bias, meaning that only patients who had lower barriers to insulin therapy were switched to insulin treatment, seems unlikely.
At the three-month follow up, it could be demonstrated clearly that barriers to insulin therapy increased in patients who remained on an oral regimen, whereas negative appraisal of insulin treatment was reduced in patients who were switched to insulin therapy. The negative appraisal of insulin treatment in patients who were switched to insulin therapy was reduced to the level of patients already treated with insulin.
Therefore, it seems reasonable to assume that patients who are exposed to insulin therapy acquire new skills regarding how to handle insulin and change their appraisal of this treatment option. These patients may accommodate to this treatment alternative and reduce their barriers to insulin therapy.
One strength of this study is that in addition to negative appraisal of insulin treatment, a broad assessment of more generic psychological variables such as wellbeing, diabetes-related distress, anxiety, and depressive symptoms were longitudinally assessed. Except for a significant effect on depression, there was no specific impact of the subsequent diabetes treatment on anxiety symptoms, diabetes-related distress, or psychological well-being. These findings might indicate that negative attitudes regarding insulin treatment is a rather specific barrier to this treatment option and is not strongly associated with general aspects of health related quality of life. This idea is corroborated by the finding that patients who remained on an oral regimen had the lowest depression scores at baseline as well as at follow-up, although their negative appraisal of insulin treatment increased.
Patients who remained on insulin treatment and patients who were switched to insulin treatment had more late complications than insulin-naïve patients. It is well known that complications of diabetes are associated with depression [bib_ref] Rates and risks for co-morbid depression in patients with type 2 diabetes..., Pouwer [/bib_ref] [bib_ref] Association of depression and diabetes complications: a meta-analysis, De Groot [/bib_ref]. Most patients who were switched to insulin treatment experienced remarkably improved glycaemic control. This outcome might have had a specific antidepressive effect, which could explain the significantly greater reduction of depressive symptoms in this group.
There are also some limitations of this study. The study is observational, meaning that the groups were not randomised. The decision who remained on an oral regimen and who was switched to insulin was at the discretion of the clinicians. Although this clinical judgement proved to be effective with regard to the glycaemic control achieved during follow-up, a selection bias cannot be excluded.
The sample size is rather small; thus a lack of power could be responsible for the fact that the effect of the diabetes treatment at follow-up could not be assessed with respect to more generic variables such as anxiety, diabetes-related distress, and well-being.
The follow-up period is rather short; longer follow-up period is needed to evaluate if reducing barriers to insulin therapy is maintained, further reduced or increased over time.
# Conclusions
Nevertheless, in summary, this study demonstrates that negative appraisal of insulin treatment is modifiable by the initiation of insulin therapy. This finding indicates that barriers to insulin treatment is a benign, temporary phenomenon instead of an unvarying patient characteristic. Future studies should address if identifying and addressing patients concerns about insulin therapy can help to improve long-term adaptation to insulin therapy.
[table] Table 1: Baseline characteristics [/table]
[table] Table 2: Baseline characteristics of patients switched to insulin vs. patients remained insulin naïve. [/table]
[table] Table 3: Baseline and follow-up results in the three different treatment groups [/table]
|
Ions, the Movement of Water and the Apoptotic Volume Decrease
The movement of water across the cell membrane is a natural biological process that occurs during growth, cell division, and cell death. Many cells are known to regulate changes in their cell volume through inherent compensatory regulatory mechanisms. Cells can sense an increase or decrease in their cell volume, and compensate through mechanisms known as a regulatory volume increase (RVI) or decrease (RVD) response, respectively. The transport of sodium, potassium along with other ions and osmolytes allows the movement of water in and out of the cell. These compensatory volume regulatory mechanisms maintain a cell at near constant volume. A hallmark of the physiological cell death process known as apoptosis is the loss of cell volume or cell shrinkage. This loss of cell volume is in stark contrast to what occurs during the accidental cell death process known as necrosis. During necrosis, cells swell or gain water, eventually resulting in cell lysis. Thus, whether a cell gains or loses water after injury is a defining feature of the specific mode of cell death. Cell shrinkage or the loss of cell volume during apoptosis has been termed apoptotic volume decrease or AVD. Over the years, this distinguishing feature of apoptosis has been largely ignored and thought to be a passive occurrence or simply a consequence of the cell death process. However, studies on AVD have defined an underlying movement of ions that result in not only the loss of cell volume, but also the activation and execution of the apoptotic process. This review explores the role ions play in controlling not only the movement of water, but the regulation of apoptosis. We will focus on what is known about specific ion channels and transporters identified to be involved in AVD, and how the movement of ions and water change the intracellular environment leading to stages of cell shrinkage and associated apoptotic characteristics. Finally, we will discuss these concepts as they apply to different cell types such as neurons, cardiomyocytes, and corneal epithelial cells.
# Introduction
Cell survival depends on maintaining cellular stability from altered environmental conditions that occur from both inside and outside the cell. Cellular stability is accomplished through numerous homeostatic processes which allows cells to self-regulate and/or adjust various biological systems providing an unvarying environment to thrive and flourish. Examples of biological systems that cells maintain include glucose levels, acid-base balance, calcium levels, and fluid volume. Cellular stress results in the activation of a variety of Intracellular mechanisms including the DNA damage response, the unfolded protein response, cell senescence, and regulated cell death. What drives the homeostatic balance of many biological systems is the movement of monovalent ions that results in a change in water content to alter the concentration of glucose, acids/bases, and calcium. Therefore, ionic fluidity and the movement of water via this mechanism or through specific water channels have a dramatic impact on cell viability. Failure of these homeostatic processes can signal the cell to die. Interestingly, even in death, cells attempt to maintain some sense of biological homeostasis by undergoing a programmed cell death process known as apoptosis. As such, activation of apoptosis is the body's attempt to remove unwanted or dying cells without affecting neighboring healthy cells.
## Necrosis vs. apoptosis
Over centuries of scientific discovery, the analysis of dying cells has not been a field of rigorous study. Cells die due to injury, accident damage, or "old age" after fulfilling their purpose to a point where they are no longer needed. Up until the early 1970s, necrosis was the term used to described dying cells defined as an accidental cell death process characterized by cell swelling followed by eventual cell lysis. The release of intracellular products of the dying cell into the extracellular space results in an inflammatory response leading to further damage in the surrounding tissue. As the corpse of a dead cell does not lend one a great deal of substance to explore, attention focused on understanding the inflammatory response, the attraction of leukocytes, and the removal of the dead cell material. However, observations byled to the understanding of a distinctive type of necrosis termed "shrinkage necrosis". From these early studies, it was evident that the loss of cell volume or cell shrinkage was a distinguishing feature of this controlled cell deletion process. Further study defined this event as a vital, active, and inherently programmed biological process known as apoptosis.
There are many discriminating features when comparing necrosis to apoptosis. Necrosis is initiated from external factors that results in detrimental effects on the cell including ATP depletion, cell swelling, membrane disruption, and eventual lysis culminating in an inflammatory response. In contrast, apoptosis is considered a physiological mode of cell death initiated by inherent mechanisms that results in chromatin condensation, cell shrinkage, membrane blebbing leading to the formation of apoptotic bodies that are engulfed by neighboring cells or macrophages. Therefore, apoptosis culminates in a silent process with no noticeable symptoms. Given the distinct cellular events surrounding death by necrosis vs. death by apoptosis, the change in cellular morphology is the most visible characteristic that can easily discriminate between these two diverse modes of cell death.
While necrosis and apoptosis exemplify the extreme modes of cell death, many other cell death processes have been identified and defined. Necroptosis is an inflammatory form of regulated (programmed0 necrotic cell death considered a viral defense mechanism that lacks caspase activation resulting in leakage of the cellular contents into the extracellular space. Similarly, pyroptosis is a highly inflammatory form of programmed cell death that occurs frequently upon infection with intracellular pathogens and is characterized by the formation of the inflammasome (pyrotosome). While caspasedependent, pyroptosis uses a distinct set of proteolytic enzymes (caspases 1, 4, and 5) then apoptosis and the activation of poreforming proteins known as geasermins results in water influx and cell membrane rupture. Thus, similar to necroptosis, pyroptosis is not considered immunologically silent. Additionally, ferroptosis is an iron-dependent programmed cell death characterized by the accumulation of lipid peroxides triggered by the failure of the glutathione-dependent antioxidant defense mechanism. Cells undergoing ferroptosis typically contract, then swell, releasing their intracellular contents. The most similar mode of cell death that mimics apoptosis is the death of red blood cells known as eryptosis. Insults such as hyperosmolarity, oxidative stress, and heavy metal exposure can result in erythrocytes undergoing cell death characterized by cell shrinkage, membrane blebbing, activation of proteases, and externalization of phosphatidylserine. A comparison of the characteristics that define these modes of cell death is shown in.
In total, there have been 34 different modes of cell death described in the literature. This includes the orderly degradation and recycling of cellular components known as autophagy; an ischemic cell death resulting from ATP depletion known as oncosis; death of anchorage-dependent cells that detach from the surrounding extracellular matrix known as anoikis; and a programmed mode of necrotic cell death in fibroblasts known as nemosis. While many of these modes of cell death are similar in nature, they can provide a unique characterization of the physiology in a clinical or pathological setting. For example, mitotic catastrophe that occurs due to premature or inappropriate entry of cells into mitosis is the most common mode of cell death in cancer cells exposed to various chemotherapeutic treatments. Thus, the use of mitotic catastrophe has a very relevant connotation in this clinical setting. In this review, we will focus on the classical physiological mode of cell death, apoptosis; and examine cell death in several cell type model systems in regards to ion and water movement that results in the loss of cell volume.
## Maintaining fluid volume homeostasis
Alterations in cell morphology are key in distinguishing necrosis and apoptosis, thus variations in cellular water content must occur suggesting that maintaining water balance is critical for cell survival. Thus, inherent cellular mechanisms have developed to combat changes in the extracellular environment that impacts a cells hydration state. In general, a sudden change in solute concentration surrounding a cell results in an osmotic stress, also described as an osmotic shock. When the extracellular solute concentration is low (hypo-osmotic stress), cells can rapidly gain water. In contrast, when the extracellular solute concentration is high (hyper-osmotic stress), a rapid loss of water occurs from cells. Simply noted, water will flow in the direction of higher solute concentration, signifying solute flux as a central determinant of water movement. As cells have a defined perimeter and limited capacity to either contract or expand, most cells respond to changes in these environmental conditions with rapid ionic fluxes that alter their intracellular environment to adjust to the change in the extracellular environment. The gain in water that occurs when cells encounter a hypo-osmotic environment (a decrease in external osmolarity) is immediately countered with an active recovery process known as regulatory volume decrease (RVD). This inherent adaptation process involves the flux of ions, mainly potassium and chloride, along with various organic osmolytes from the cell. Potassium is the most abundant monovalent ion in the cell and permeates from the cell through various channels including voltage-gated, Ca 2++activated, inwardly rectifying, and two-pore-domain potassium channels (Pasantes-. The precise potassium channels activated during RVD appears to be both cell-type and stimulus-specific. Along with potassium channels, voltagesensitive chloride channels also have a key role in RVD in maintaining an overall electrically neutral ionic state. The voltage-regulated anion channel (VRAC) has been a channel of intense interest as VRAC is also permeate to large molecules such as gluconate and glutamate that can further facilitate a restoration in cell volume.
Cells encountering a hyper-osmotic environment (an increase in external osmolarity) immediately shrink and activate a regulatory volume increase (RVI). During this process, various mechanisms are activated to increase the concentration of intracellular osmolytes. Sodium is the most abundant ion outside the cell, and sodium enters the cell through various electroneutral cotransporters and exchangers including the Na + -Cl − cotransporter (NCC), the Na + /K + /2Cl − cotransporter (NKCC), and the Na + -H + exchanger coupled to the Cl − /HCO 3 − exchanger (Pasantes-. Additionally, the initial increase in intracellular sodium that occurs during RVI is alleviated through the activation of the Na + -K + -ATPase, that resets the initial intracellular ionic environment. Similar to the activation of RVD, the precise transporters and exchangers activated during RVI is not completely understood and occurs in both a cell-type and stimulus-specific manner.
The importance of inherent volume regulatory mechanisms in protecting cells from adverse changes in the extracellular environment was illustrated when T-cells, that lack a normal RVI, were subject to hypertonic stress. In the absence of an inherent RVI, these T-cells shrank and underwent a rapid and systematic cell death that was shown to be apoptosis, while cells such as COS-7, L-cells, and PC12 cells, that can regulate their volume via RIV, were resistant to hypertonic-induced stress and survived. Increased tonicity augmented serum-deprived induced apoptosis in vascular smooth muscle cells. In support of the concept of volume regulatory mechanisms protecting cells from death, inhibition of hypertonicity-induced cation channels was shown to sensitize HeLa cells to undergo apoptosis when the extracellular osmolality was increased . Interestingly, while hyperosmolarity did not initially induce apoptosis in rat hepatocytes that respond with an RVI,showed this condition did target the CD95 receptor to the plasma membrane sensitizing the cells to Fas ligandinduced apoptosis. While in the clinical setting, mannitol therapy has been widely used for acute and subacute reduction in brain edemas resulting from closed-head trauma, and ischemic brain swelling to improve cerebral blood flow,pointed out potential deleterious effect of hyperosmotic treatment on the vascular endothelium due to mannitol's ability to induce apoptosis. This suggests caution should be exercised for the clinical use of osmotic diuretics such as mannitol to avoid detrimental and often lethal effects to surrounding cells and tissue.
## The advent of avd
A defining characteristic of cell death as outlined earlier is a change in cell volume. This simple and straightforward visual cue allows one to immediately categorize the two most common modes of cell death that has or is occurring. The loss of cell volume that occurs during apoptosis and gain in cell volume as observed during necrosis both occur in the absence of osmotic changes in the extracellular environment. Consequently, the term necrotic volume increase (NVI) was proposed to describe the influx of sodium, lactate, and other osmolytes into cells leading to cell swelling during this accidental or necrotic cell death process. NVI is thought to be due in part to a dysregulation of the inherent RVD, specifically an impairment of volume-sensitive Cl − channelsand subsequently, an acid-sensitive outwardly rectifying (ASOR) anion channel .
In contrast, the term apoptotic volume decrease (AVD) has been applied to describe the loss of cell volume or cell shrinkage during the physiological or apoptotic cell death process. What became apparent from early studies investigating the loss of cell volume during apoptosis was that AVD most likely was not a novel volume mechanism, but occurred via sharing or commandeering inherent RVD channels/transporters for a new purpose during the programmed process. Of particular note is VRAC; the volume-activated anion channel that is essential to the apoptotic death machinery. VRAC is activated by a change in intracellular ionic strength, increased intracellular calcium, ROS, and phosphorylation, however a complete understanding these signaling cascades is not known. The exact nature of this channel remained unknown until the recent identification of LRRC8 proteins as a key component of VRAC. While LRRC8 isoforms have been shown to contribute to RVD and sense changes in ionic strength, it has been suggested that this volume-regulated anion channel may not be essential for AVD. Moreover, it is important to note that the reprogramming of RVD to AVD during apoptosis must likely involves the inactivation of RVI that would normally compensate for a loss of cell volume.
As ion flux was known to underlie cell volume regulatory processes, ion channels became a central focus for AVD. Two-pore domain K + (K(2P)) channels were suggested to underlie potassium efflux during AVD in mouse embryos. AVD was shown to be accelerated upon staurosporine-induced apoptosis of COS-7 and pulmonary artery smooth muscle cells (PASMC) overexpressing a delayedrectifier voltage-gated K + channel. Studies using a calcium-induced apoptosis lymphocyte/thymocyte model showed that AVD could be blocked with inhibitors of IKCa1, preventing the externalization of phosphatidylserine and cell death. In endothelial cells challenged with staurosporine, AVD was inhibited with the chloride channel blocker phloretin, again preserving cell viability.
Early X-ray microanalysis showed an increase in intracellular sodium coupled with a decrease in intracellular potassium within 3 h of oxidized low-density lipoprotein exposure in monocyte-macrophages. Subsequent X-ray microanalysis studies confirmed a two-phase change in intracellular using staurosporine-treated U937 cells . These authors showed during late stage apoptosis the potassium concentration continued to decrease, while chloride increased along with an increase in sodium. In a follow-up study, this group also showed that the initial stage of apoptosis in UV-induced U937 cells were characterized by a decrease in potassium and chloride, with the largest decrease occurring from the mitochondria. An earlier study byhad showed an increase in intracellular sodium in anti-Fas treated Jurket cells, that was in part to be due to the inhibition of the Na + /K + -ATPase, as direct inhibition of this ionic pump with ouabain enhanced apoptosis. This study also that cell shrinkage could be uncoupled from apoptosis, which was later confirmed in a study comparing staurosporine-and etoposide-induced apoptosis in U937 cells.
Additionally, several other studies also eluded to an increase in intracellular sodium that accompanied the initial loss of intracellular potassium during AVD. What was clear from these early studies on AVD was while no single ionic channel or pathway could account for the loss of cell volume during apoptosis, the movement of ions was a critical part of the cell death process. More recent studies have solidified this relationship between AVD and ion flux, as a computational study on the redistribution of ions and water underlying AVD in staurosporine treated human lymphoma cells (U937) concluded that along with a significant increase in chloride and potassium permeability coupled with a decreased permeability of sodium, there was also the progressive decrease in the Na + /K + activity.
In some model systems, it has been suggested that apoptosis can occur in the absence of a loss of cell volume. A modest decrease in cell volume in serum-deprived vascular smooth muscle cells does not trigger the apoptotic machinery. Interestingly,reported swelling of the whole cell prior to its entry into apoptosis; a phenomenon also noted in an earlier study using time-lapse, dual-image surface reconstruction of staurosporine treated vascular smooth muscle cells. These studies illustrate that more refined changes can occur during AVD, that typically may go unnoticed in many apoptotic model systems.
## Stages of avd
Following initial studies describing various ion flux pathways and transporters that were involved in and defined AVD, the overall nature of this process was examined which focused on both water and ion movement. Radiation-induced changes in cell size of rat thymocytes was shown to occur in two distinct stages. An early reversal of intracellular ions was observed which defined an initial or primary stage of apoptosis in lymphocytes. Of interest was this primary stage of AVD occurred during both intrinsic and extrinsic apoptosis. During the primary stage of AVD, an increase in intracellular sodium, coupled to a decrease in intracellular potassium occurred that resulted in a 20-40% decrease in cell volume. It was hypothesized that this reversal of intracellular ions was the cell's attempt to compensate for the loss of one ion (potassium) for another (sodium), however, in total, an overall decrease in cell volume occurred. During the secondary stage of AVD, both intracellular sodium and potassium were lost resulting in an 80-85% decrease in cell volume. Additionally, this secondary stage was shown to be prevented upon disruption of the actin cytoskeleton.described three distinct stages of AVD in cisplatin-induced Ehrlich ascites tumor cells; an early, transitional, and secondary stage. The early and secondary stages were defined with a loss of ions, specifically potassium, sodium, and chloride, that resulted in a 30% loss of water (early stage), and a further reduction in water in the secondary stage. The transitional stage was defined solely with an increase in sodium and chloride. Interestingly in both aforementioned studies, the increase in intracellular sodium suggests a counter or protective response even as a cell is dying. Interestingly, a protective response was illustrated bywhere U937 cells under hypertonic stress initially responded with an RVI prior to AVD. Whether RVI and AVD are independently activated and AVD is observed only apparent at a later time; or RVI becomes inhibited or fails signaling AVD is unclear.
The use of live fluorescence and transmission-throughdye microscopy showed two morphologically distinct stages of volume changes. Using actinomycin-D treated HeLa cells, the first stage was defined by extensive blebbing with a temporary volume increase, while the second stage showed a 40% decrease in cell volume and was considered to represent AVD. Interestingly in this model system, both stages had an increase in intracellular sodium, which again suggests a protective cellular response even in the act of dying. In a latter study, correlative light and cryo-scanning transmission electron microscopy (cryo-STEM) was employed to study stage-specific changes in water and ion movement in actinomycin D treated HeLa cells. This technology allowed the authors to observe changes of water and ions not only in the cell as a whole, but also in various cellular compartments. Overall, the authors observed a loss of potassium throughout the entire cell death process. Since previous studies suggested an early increase in sodium and loss of chloride, followed by a decrease in sodium coupled to a further decrease in chloride, the latter appeared to be restricted to the mitochondria. Finally, during the late stage of cell death, an increase in sodium and chloride was noted that may ensue due to a loss of membrane integrity. Overall, these studies illustrate a complex and multifaceted nature of volume dynamics during cell death.
Many studies, including the one outlined above, have shown that the concept of AVD or cell shrinkage during apoptosis is not a simple and straight-forward process. Of particular note is the idea that AVD is distinct and independent from the loss of cell volume that occurs upon separation of apoptotic bodies. Overall, the volume dynamics that reflect AVD is a direct consequence of ion flux. The observation of whether a cell shrinks or not during the cell death process appears to be both cell-type and stimulus specific. Therefore, water content of non-apoptotic and apoptotic cells may not be of much consequence, as it is not the change in cell volume that is critical, but the flux of ions that has a greater impact on the cell death program.
## The relevance of avd to other apoptotic events
Prior to AVD becoming an acknowledged scientific concept, intracellular ions were known to play a critical role in water loss during apoptosis. Furthermore, and as mentioned above, many studies support the idea of ion flux having a critical role in the apoptotic program, with the change in cell volume a byproduct of this ion movement. However, the question remained as to what consequence these ion fluxes have on apoptosis? Early it was shown that DNA degradation during apoptosis in anti-Fas treated Jurkat cells correlated with the shrunken population of cells, and inhibition of cell shrinkage via high extracellular potassium prevented this characteristic . Interestingly, anti-Fas treatment of Jurkat cells initially placed in hypotonic medium to swell and activate an RVD, thus lowering the intracellular potassium and chloride concentration, resulted in enhanced cell death, suggesting that a change in ions was having a greater effect on the cell death program than the actual change in cell size. In a follow-up study, normal intracellular levels of potassium were shown to inhibit both apoptotic DNA fragmentation and caspase-3 activation, however, once the caspase was activated, the level of intracellular potassium had no consequence ). An early study examining calcium signaling during apoptosis showed that anti-Fas treated Jurkat cells under complete calcium-free conditions resulted in only the inhibition of DNA fragmentation, suggesting nuclear activity was the only component of the apoptotic machinery that was sensitive to changes in intracellular calcium. Additionally, it was shown that elevated extracellular potassium prevented phosphatidylserine externalization, mitochondrial depolarization, and cytochrome c release, along with caspase activation upon both chemical and death receptor induced apoptosis. This same group also showed that physiological concentrations of potassium inhibited cytochrome c-dependent apoptosome formation, thus preventing the activation of caspase-9 . In a more recent study,examined the relationship between AVD and translocation of phosphatidylserine on the cell surface and membrane blebbing using time-lapsed imaging coupled with scanning ion conductance microscopy. Here, the authors using staurosporine-treated neurons reported that the loss of cell volume occurred prior to the externalization of phosphatidylserine and membrane blebbing, suggesting that these morphological events may be independent of AVD and the concurrent ion flux during apoptosis.
Since our initial understanding of the important role ion fluxes play in regulating the apoptotic machinery, considerable attention has focused on the specific channels and transports involved in this process, that may or may not contribute to AVD. In an early study using human leukemia cells (HL-60), apoptotic change in intracellular ions was prevented upon inhibition of the Na + , K + -ATPase pump or the Ca 2+ -dependent K + channel. Caspase-dependent stimulation of voltage-gated potassium (Kv1.3) channels was shown in Fas-ligand treated Jurkat cells to result in potassium efflux, cell shrinkage, and apoptosis. While potassium is a major intracellular ion whose loss would result in a reduction of cell volume, it is not the only ion to consider in understanding the relevance of AVD. The early and rapid increase in intracellular sodium shown to occur during anti-Fas induced apoptosis resulted in a depolarization of the plasma membrane . The depolarization of the plasma membrane had previously been shown to occur in part via the inhibition of the Na + /K + -ATPase. Thus, over the past 2 plus decades, many studies have solidified the critical role for ion flux in regulating the apoptosis process (reviewed in.
Conversely,suggested that a decrease in intracellular potassium concentration is not obligatory for apoptosis. In an oocyte model treated with staurosporine, a loss of intracellular potassium was observed along with the activation of caspase-3. Interestingly, when oocytes densely expressed Shaker voltage-gated potassium channels, a loss of potassium was not observed, implying that the dense Kv channel expression makes oocytes resistant to apoptosis. However, caspase-3 activity was still observed. While the authors concluded that a decrease in intracellular potassium concentration is not required for apoptosis, other ions known to have a role in decreasing the overall ionic strength was not explored. Largely, these afore-mentioned studies illustrated the importance of ions in the programmed cell death process and suggested that an overall decrease in intracellular ionic strength permits the activation of the apoptotic machinery.
## Linking apoptosis and water movement: aquaporins
Water channels, or selective water pores known as aquaporins (AQPs) play a critical role in mediating cellular water flow, and are crucial for the regulation of water homeostasis. Aquaporins provide a mechanism for the rapid movement of water across diverse membranes having a major regulatory effect in regard to changes in cell volume. In the early 2000's,reported that salivary acinar cells deficient in aquaporin 5 (AQP5) had a decrease in water permeability in response to hypertonicity-induced cell shrinkage and hypotonicity-induced cell swelling. While it is unclear exactly how aquaporins regulate this movement of water, it has been proposed that a critical factor is the number of channels expressed on the cell membrane. AQP5 was shown to have a dose-responsive decrease in response to a hypotonic stimulus. This study also showed that inhibition of the cation channel transient receptor potential vanilloid (TRPV) 4 prevented the reduction of AQP5. In a latter study, a rapid translocation of AQP1 was shown to occur under hypotonic conditions coupled with an increase in intracellular calcium. As these studies involved different aquaporins, both studies indicated that alterations in calcium were required for the change or translocation of the channels while illustrating their response in the regulatory volume response. Additionally, these studies indicate that ions, particularly calcium, may have an important regulatory role in water channel function, as opposed to regulation at the level of the channel itself. Furthermore, while initially thought to facilitate only the movement of water, it is now understood that aquaporins can also permeate other small solutes such as anions, urea, and glycerol, that would also have a role in the volume regulatory response.
Aquaporins have been shown to play a vital role during cell death as inhibition of aquaporin 1 (APQ1) was shown to prevent AVD and the subsequent downstream apoptotic events such as cell shrinkage, mitochondrial membrane permeability, caspase 3 activation, and DNA degradation. Additionally, decreased expression of APQ8 and APQ9 correlated with the lack of water movement in apoptotic-resistant tumor cells. Overexpression of AQP3 and AQP9 in human melanoma cells significantly increase the chemoresistance of these cells to arsenite via down-regulation of p53 and up-regulation of Bcl-2 and XIAP. Thus, like the overall mechanism of apoptosis, the role aquaporins play during cell death process appears to be cell-type and stimulus specific.
Since the mid 1990's, an increasing number of studies have focused on aquaporins and the movement of water, with many focusing on specific physiological systems or pathological conditions. For example, in the central nervous system, aquaporin 4 (AQP4) is noted as the main water channel. Lactacystin (proteosomalinhibition) induced apoptosis in cortical neurons showed AQP4 was highly downregulated suggesting other AQPs may be involved during programmed cell death. Interestingly, this study showed that APQ8 and APQ9 were highly upregulated upon lactacystin treatment. However, when staurosporine was used to induce apoptosis, both APQ8 and APQ9 were down regulated, suggesting aquaporins expression and function during apoptosis is stimulus specific. In a myocardial infarction model, cardiomyocytes deficient in AQP1 showed a reduced level of apoptosis, suggesting AQP1 has a positive role in the execution of the cell death process. And as early as 2004, aquaporins, specifically AQ1 was shown to be decreased in human cornea endothelial disease, and in mouse corneas subjected to corneal endothelial injury. Thus, as is the case for ion channels and transporters, aquaporins are expressed in a range of cells and show a variety of roles in regards to apoptosis.
## Ion movement and avd in unique model systems
Lymphoid cells have been a favorite model system to study AVD and apoptosis. However, over the past decade, other cell type such as neuronal cells, cardiomyocytes, and corneal epithelial cells have been increasingly studied due to their unique characteristics and prevalence in various human diseases. While the brain as a whole shows an RVI upon hypertonic perturbations, cultured and acutely isolated neurons do not. Similarly, reports of RVI in cardiomyocytes are limited to murine atrial cardiac myocyte cultured cell lines such as HL-1. As both neurons and cardiomyocytes have an RVD, it is of interest to determine their relationship to AVD upon apoptosis. While corneal epithelial cells can regulate their cell volume under both hypertonic or hypotonic conditions, the cornea is shielded by a protective tear film that makes this a model of interest in understanding ion flux and fluid movement upon cell death.
## Neuronal cells and cell death
Neuronal cells offer an attractive model to study the movement of water, ions and apoptosis. Early on it was shown that mouse neocortical neurons treated with staurosporine or serum deprivation resulted in an early enhancement of delayed rectifier (IK) currents and a loss of intracellular potassium resulting in apoptosis. Apoptosis was reduced upon addition of the potassium channel blocker TEA or elevated extracellular potassium. In turn, exposure of neuronal cells to the K + ionophore valinomycin or the K + -channel opener cromakalin induced apoptosis, suggesting that neuronal cell death may follow a similar series of ion flux as observed in more classical model systems. Shortly after this, NMDA receptor-mediated potassium flux was shown to contribute to neuronal apoptosis during brain ischemia, signifying an expansion of ion flux beyond classical voltage-gated ion channels. In a second study from this group, inhibition of potassium channels with clofilium attenuated C2-ceramide induced neuronal apoptosis, as well as hypoxia-and ischemia-induced neuronal death, both in vitro and in vivo. These early studies illustrating the critical role for potassium during neuronal cell death set the stage for further scientific investigation of neuronal cell death.
Neurons, like every other cell in the body, can also be subjected to changes in their extracellular environment. Upon encountering a condition of decreased osmolality, neurons will undergo RVD to achieve a homeostatic balance of water and ions. This RVD occurs via classical ionic channels and transport mechanisms similar to other cell types, and is observed in many neuronal cells including peripheral sympathetic neurons, cerebellar granular cells, along with numerous neuronal cultured cell lines. It was suggested that AVD in neurons appears to occur by similar ionic mechanisms to those activated during hypoosmotic-induced RVD (Pasantes-. Cation-chloride cotransporters (CCC) such as the chloride-importing Na-K-2Cl cotransporter (NKCC1) and the chloride-exporting potassium-chloride cotransporter (KCC2) have a significant role in the regulation of neuronal cell volume, along with their role in neurotransmission in the nervous system. These transporters are oppositely regulated via serinethreonine phosphorylation that inhibits NKCC1, but activates KCC2, upon dephosphorylation possibly through the WNK2 kinase . The dephosphorylation of these transporters promotes the efflux of ions, specifically potassium and chloride from the cell resulting in loss of water. Interestingly, numerous studies involving neurons (both primary and cultured) failed to demonstrate a classical RVI response upon hyperosmotic exposure. Additionally, a lack of RVI was also observed in most studies involving cultured astrocytes (reviewed in. A sound hypothesis for the absence of RVI in various neuronal cells has yet to be proposed, although it has been suggested that cultured neuronal cells may not have the required transmembrane ionic gradients that favor RVI.
During development, proper formation of synapses between neurons in the brain is known to occur via apoptosis. Furthermore, apoptosis is also the most common mode of cell death in various neurodegenerative diseases with increased apoptosis associated with Alzheimer's and Parkinson's disease, suggesting that apoptosis plays a variety of roles within the central nervous system. As mentioned earlier, and similar to other apoptotic model systems, the loss of intracellular potassium is known to have a critical impact in coordinating the cell death program. Early studies showing the enhancement of delayed rectifier (IK) potassium currents during apoptosis in neocortical neurons, were followed by studies on cultured cortical neurons treated with a variety of apoptotic inducers that showed the involvement of ion flux. Inhibition of chloride channels prevented cell shrinkage, but had no significant effect on caspase activation or DNA fragmentation; . In contrast to chloride channel inhibition, inhibition of potassium channels prevented cell shrinkage, caspase activation, and DNA fragmentation . This study suggests that potassium and chloride have distinct roles during apoptosis, with inhibition of potassium flux exhibiting a greater neuroprotective effect. Furthermore, studies such as these highlight the critical role potassium plays in regulating the apoptotic machinery, again in line with more classical apoptotic model systems.
Over the years, numerous potassium channels have been identified in neuronal cells that have a function during apoptosis including voltage-gated K + channels, inwardly rectifying K + channels, and background channels such as tandem pore domain TWIK and TASK. Many of these studies have relied on channel inhibition to show their contribution in the cell death process. Of the various potassium channels identified, the delayed rectifier current mediated by Kv2.1 channel has a critical role in apoptogentic potassium efflux in several types of neuronal cells including cortical, nigral, and hippocampal neurons.showed that potassium efflux during neuronal cell death involved newly inserted Kv2.1 channels into the cell membrane. As Kv2.1 channels are known to form clusters in the soma and proximal dendrites,discovered FIGURE 1 | Neuronal AVD. Mechanisms similar for classical RVD are engaged during neuronal AVD. Ionic cotransporters and cotransporters, mainly involving the flux of chloride are activated to counter the imbalance of intracellular water due to hypotonic conditions. For example, conventional ionic transport mechanisms such as NKCC1 and KCC2 are oppositely-regulated via serine-threonine phosphorylation such that dephosphorylation results in the inhibition of NKCC1, while simultaneously activating KCC2. The net result is the loss of both intracellular potassium and chloride with the parallel decrease in water. Additionally, individual potassium and chloride channels have also been shown to have a role during neuronal AVD. Interestingly, potassium channel activation was shown to result in AVD, caspase activation, and DNA fragmentation, while chloride channel activation resulted in only AVD. that disruption of these clusters prevented the apoptogenic increase in potassium currents, thus increasing neuronal viability upon exposure to oxidative stress. Recently in primary neurons, Kv2.1 was shown to be substrate for the aspartyl protease BACE2, and upon channel cleavage prevented the outward potassium efflux resulting in reduced apoptosis. Furthermore, methamphetamine results in pro-apoptotic effects in primary hippocampal neurons that are abrogated upon inhibition or knockdown of Kv2.1, suggesting potassium channel inactivation without the use of drugs that specifically block the channel can regulate the apoptotic program. Additionally, this study showed that p38 mitogen-activated protein kinase (MAPK) was also involved as inhibition of this kinase attenuated methamphetamine-induced pro-apoptotic effects and the upregulation of Kv2.1. Moreover, phosphorylation of what are considered pro-apoptotic residues on the N-and C-terminus of Kv2.1 via Src and p38 enhanced the insertion of this potassium channel in the plasma membrane, thus increasing the loss of intracellular potassium during apoptosis.
Like many other apoptotic model systems, potassium has been the cationic ion of emphasis, however, as mentioned earlier sodium also has a criterial role in the cell death program, and has been an ion of focus of numerous studies involving neuronal apoptosis.showed flufenamic acid and mefenaminc acid were neuroprotective by inhibiting voltagegated sodium channels in a glutamate-induced apoptotic model using neuroblast-like SH-SY5Y cells. In a model resembling epilepsy, hippocampal neurons from rats injected with kainic acid (an NMDA type glutamate receptor agonist) underwent apoptosis that was attenuated upon inclusion of various voltagegated sodium channel blockers. An increase in intracellular sodium following abnormal hyperexcitation can result in death of neurons, and a recent study probing the mechanism of this model suggested this death occurs via sodium accumulation and/or concomitant potassium loss that impairs mitochondrial function. Genistein, a primary isoflavone found in soybeans, was shown to inhibit cell death in an in vitro model of primary neurons under hypoxicischemia (oxygen-glucose deprivation) conditions in part by reversing the classic increase in potassium efflux and decreasing the sodium influx. Additionally, this group in a latter study employing a different hypoxic-ischemia model (sodium dithionite and glucose deprivation) in cultured rat primary neurons showed that a specific zinc-chelator, TPEN, suppressed apoptosis in primary neurons, however whether this was a direct inhibition of ion channels or a change in channel function due to reduced levels of zinc was not determined.
Potassium channel activation is not always congruent with triggering of apoptosis, especially in neurons. Interestingly, potassium channels are common targets for neuroprotective molecules. Activation of mitochondrial ATP-sensitive potassium channels (mKATP) prevented neuronal cell death after ischemia in neonatal rats, essentially mimicking the protective effects mediated by the preconditioning phenomenon. In an oxygen-glucose deprivation model of cell death in primary rat cortical neurons, diazoxide, a potassium channel activator attenuated neuronal cell death. Of interest in this study was that a mitochondrial ATP-sensitive potassium blocker (5-hydroxydecanoate) abolished this protective effect, while a non-selective KATP channel blocker (glibenclamide) did not, suggesting the reliance of mKATP for this resistance. Since this early deduction of neuroprotection via opening of mKATP, numerous studies have expanded on this discovery. Neuroprotection via mitochondrial ATP-sensitive potassium channels enhance cell survival against oxidative stress. In a chronic morphine (CM) preconditioning study of ischemia/reperfusion hippocampal CA1 neurons, inhibition of mKATP channels with 5-hydroxydecanoate (5-HD) significantly increased apoptosis, suggesting that CM preconditioning obstructs apoptosis via activation of mKATP channels. Various studies have suggested that the protection afforded by the activation of potassium channels may be due to membrane potential hyperpolarization, and/or increased repolarization speed effectively reducing the level of calcium entry and ATP consumption; both considered pro-apoptotic events. Thus, pharmacological modulation of mKATP has become a promising new therapeutic approach for the treatment of neurodegenerative diseases such as Alzheimer's, along with the treatment of cardiovascular and various oncological diseases.
## Cardiomyocytes and avd
While cardiomyocytes are known to respond to hypoosmotic stress with an RVD, the presence of an RVI under hyperosmotic stress has not been well documented in these cells. It has been shown that HL-1 cardiac myocytes can regulate their volume to hypertonic stress with a classical RVI, however, at present, no study has documented the presence of an RVI in primary cardiomyocytes. In the absence of this inherent RVI response, cardiac myocytes respond with a rapid and robust apoptotic response upon hyperosmotic stress.showed that this rapid and pronounced apoptosis in response to hyperosmotic stress in cardiomyocytes could be attenuated by treatment with insulin-like growth factor (IGF-1), setting in motion a series of calcium-related phosphorylation events. More recently, hyperosmotic stress was shown to induce cell death in adult rat cardiomyocytes via mechanism promoting endoplasmic reticulum stress (ERS;. In this study, cardiomyocytes placed in a hyperosmotic environment resulted in increased expression of various ERS markers, along with an increase in caspase-3 expression and the loss of cell viability. Furthermore, 4-BPA (an inhibitor of ERS), chelating calcium using EGTA, and inhibition of CaMKII prevented hyperosmotic-induced cell death. Interestingly,suggest that in the absence of the protective nature of an RVI response, cardiomyocytes invoke an autophagic pathway to provide a cardioprotective strategy in response to hyperosmotic stress. The complex nature of volume regulatory process in cardiomyocytes was illustrated by. These authors showed that hypo-osmotic stress-induced increase in intracellular calcium, thus activating RVD, however, this also resulted in an increase ROS . This masking of RVD via the increased ROS resulted in necrotic blebs and cell death. Interestingly, the overexpression of catalase, lowered ROS content, and restored RVD .
In the mid 1990s, studies investigating ischemia reperfusion injury showed that cell death of cardiomyocytes occurred predominantly through the programmed cell death process or apoptosis. Furthermore, myocardial ischemia can result in osmotic stress on cardiomyocytes that affects the overall functioning of the heart. Subsequent studies examining cardiomyocyte apoptosis and the loss of cell volume showed that volume-sensitive ion channels played a role in AVD (D' . Specifically, volume-sensitive chloride channels (I Cl,vol ) known to play a role in the regulation of cell volume (RVD), were also shown to be active during AVD Additionally, volume-sensitive outwardly rectifying (VSOR) chloride channels were shown to be involved in staurosporine treated primary mouse ventricular myocytes and neonatal rat cardiomyocytesundergoing apoptosis. Recently, inhibition of (VOSR) chloride channels were shown to improve cardiac contractility and survivability in a turnicamycin-induced cardiomyocyte ER stress model. In a follow-up study,showed during high glucose-induced apoptosis, chloride channel blockers DIDS and DCPIB prevented activation of (VSOR) chloride channels and improved the viability of cardiomyocytes.
What is noteworthy about the volume-sensitive channels during AVD is their activation in the absence of cell swelling. While the precise nature of the isotonic activation of volumesensitive channels is unknown,hypothesized several mechanisms that may permit channel activity including a lower threshold for the channel volume set point in apoptotic cells, the presence of reactive oxygen species, kinase activation and/or changes in the level of ATP as plausible mechanisms for channel activation. Subsequently,showed that PI3K/Akt played a major role in the activation of (VSOR) FIGURE 2 | RVD and AVD in cardiomyocytes. Dual role of calcium for RVD in cardiomyocytes is illustrated as a hypotonic-induced increase in intracellular calcium activates RVD. However, a simultaneous increase in ROS masks and/or prevents RVD. RVD can be restored via overexpression of catalase which lowers the ROS concentration. Activation of volume-sensitive chloride channels via kinases, ROS, and/or changes in the level of ATP were shown to have a major effect on cardiomyocyte AVD, which can be prevented upon addition of specific chloride channel blockers. Additionally, reverse mode of the Na-Ca exchanger can also result in AVD as sodium-free conditions prevents this loss of cell volume.
chloride channels during staurosporine-induced cardiomyocyte death . Additionally,suggested the involvement of the C/EBP homologous protein CHOP and Wnt inactivation upon ER stress in cardiomyocytes. Regardless of the precise mechanism, it is well-established that volume-sensitive ion channels become activated during cardiomyocyte apoptosis and result in the concomitant egress of water from the cells and persistent cell shrinkage defined as AVD.
Interestingly, beside studies examining the role of chloride channels during cardiomyocyte AVD, little information exists on other ionic flux that may be involved during apoptosis.
In an oxidative stress model, cardiomyocytes treated with H 2 O 2 showed a marked increase in intracellular sodium and calcium via reverse mode of the Na-Ca exchanger resulting in apoptosis ; . Interestingly, apoptosis was completely prevented under sodium-free conditions, but not calcium-free. Additionally, apoptosis occurred when a sodium ionophore cocktail in calcium-free medium was used instead of H 2 O 2 , suggesting the increase in intracellular sodium alone can signal the programmed cell death process in cardiomyocytes.
Glucocorticoids have been a classical apoptotic agent from the very first reports defining this physiological programmed cell death process. Intriguingly, treatment of cardiomyocytes with the synthetic glucocorticoid dexamethasone resulted in cardiac hypertrophy, and protected these cells from both serum deprivation and TNFα-induced apoptosis. Earlier studies showed that dexamethasone inhibited serum deprivation-and UV-C-induced apoptosis in rat hepatoma cells. Interestingly in the latter study, dexamethasone did not prevent anti-Fas-induced apoptosis in hepatoma cells, suggesting the protective nature of glucocorticoids may be specific to agents that activate the intrinsic, and not the extrinsic apoptotic process.
## Corneal epithelial cells and avd
The corneal epithelium, the outermost layer of the cornea composed of several layers of non-keratinized, stratified squamous epithelia cells that covers the front of the cornea, is shielded by a protective tear film consisting of an electrolyteand protein-rich aqueous-mucus layer. The cornea functions to protect the intraocular contents of the eye along with serving as the principal optical element allowing formation of an image on the retina. Precise maintenance of electrolytes forming the osmotic gradient between the tear film and the ocular surface epithelia is important for cellular function and homeostasis. The ionic composition of tear film has been established in situ along with showing a critical role for aquaporins (water channels) in maintaining osmotically driven water transport across the cornea and tear film layer. An imbalance of electrolytes in the tear film layer is a hallmark of many ocular diseases, most notably dry eye. Thus, the corneal epithelium is the main cellular barrier between the external environment and the operative components of the visual system.
As the osmotic gradient between the tear film and the ocular surface cell epithelia is vital, corneal epithelial cells are known to have various inherent volume regulatory mechanisms. Similar to other cells in the body, volume regulatory mechanisms in corneal epithelial cells are comprised of analogous channels and transporters namely volume-regulated anion channels, potassium channels, the K-Cl and Na-K-Cl cotransporters, and the Na-K-ATPase. Additionally, channels such as the nonselective cation channel (Transient Receptor Potential Vanilloid 4; TRPV4) have been shown to have a role in RVD, as a decrease in TRPV4 expression and activity in corneal epithelial cells suppresses RVD.
Along with various channels and transporters involved in volume regulatory responses, kinase signaling pathways have also been shown to have a critical role. In a swellinginduced model of rabbit corneal epithelial cells, activation of extracellular signal-regulated kinase (ERK) and stress-activated protein kinase/c-Jun N-terminal kinase (SAPK/JNK) preceded both chloride and potassium channel activity in the RVD response . During hyperosmotic stress-induced corneal epithelial cell death (an in vitro dry eye model), activation of Polo-like kinase 3 (Plk3) and c-Jun were observed to promote the cell death program. In a recent study, KIOM-2015EW (a hot water extract of maple leaves) was shown it could protect the ocular surface from hyperosmolar stress in part by inhibiting MAPK kinase signaling.
While corneal epithelial cells have inherent cell volume regulatory mechanisms, they are still at risk of environmental insults. From exposure to fresh water while swimming, use of hypotonic eye drops, or pathological conditions such as dry eye disease, corneal epithelial cells can experience periods of persistent hypotonic or hypertonic stress that can lead to cell death. Additionally, other events can also have a negative impact on the eyes ranging from not eating a wellbalanced diet to too much time in front of a computer and/or cellphone screen. Furthermore, while most people are aware of the harmful effects of UV light on the skin, few consider the equaling damaging effects excessive UV light has on the eyes. The corneal epithelium is continually exposed to ambient outdoor UVB and UVA. Therefore, it is not surprising that many diseases of the eye can be traced back to excessive UV light including cataracts, macular degeneration, and cancer of the eye.
It has been known for over 20 years that UV exposure results in corneal epithelial cell apoptosis . What also became apparent very early was the role potassium channels played in corneal epithelial cell death. UV irradiation resulted in corneal epithelial cell apoptosis through the hyperactivation of potassium channels in the cell membrane, and inhibition of potassium channels with specific potassium channel blockers resulted in a reduction of numerous apoptotic characteristics including caspase activity, and DNA degradation. The opening of potassium channels leads to a rapid loss of intracellular potassium, fast cell shrinkage, and consequentially the activation of stressrelated signaling pathways including the c-Jun N-terminal kinase cascade and p53 activation. Not surprisingly, elevated extracellular potassium was shown to prevent apoptosis in UV-B exposed corneal epithelial cells, including reducing caspase activity. Additionally, in this study the authors showed that caspase-9 had little activation, while caspase-8 was activated upon UV-B exposure, suggesting a major route of apoptotic induction was through the extrinsic pathway. However, a latter study bysuggested that the intrinsic apoptotic pathway is the major contributor to UVB-induced apoptosis in human corneal limbal epithelial cells, but TNF-R1 and FADD pathway still played an integral part in the UVB-induced potassium efflux. What was clear from these and other studies where high extracellular potassium was shown to prevent the deleterious effects of UV exposure on corneal epithelial cellsis that the relatively high concentration of potassium in tears works to suppress the loss of potassium from corneal epithelial cells in response to UV exposure. Therefore, tears function as a defensive measure in protecting the ocular surface not only from changes in extracellular osmolality, but also from ambient UV radiation.
Finally, elevated tear osmolarity exposes corneal epithelial cells to extracellular osmotic stress; a key pathological factor in dry eye disease. To combat the loss of cell volume or cell shrinkage as a result of this hypertonic stress on corneal epithelial cells, organic osmolytes such as betaine that can act as an osmo-protectant have been studied. The presence of betaine in hypertonic-stressed human corneal limbal epithelial cells reduced the loss of cell volume from 27 to 11%, reduced caspase-8,-9, and -3/7 activity, and increased cell viability, suggesting agents such as betaine stabilized corneal epithelial cell volume under hyperosmotic stress, thus limiting the extent of apoptosis.
## Perspectives
The morphological loss of cell volume or AVD is a defining characteristic of apoptosis and suggests it plays a critical role during cell death. Whether it's simply to decrease the size of cells to be easily phagocytized by resident macrophage and/or neighboring cells, or as an essential component of the apoptotic machinery, AVD is unique to this mode of programmed cell death. While many questions remain as to the role AVD plays during apoptosis, what has become apparently clear from studies on AVD and water movement in less common model systems is the existence of cell-type specific mechanisms for apoptosis. The extension of apoptotic studies in models such as neuronal cells, cardiomyocytes, and corneal epithelial cells has illustrated how individual cells employ their endogenous channels, transporters, and/or exchangers to carry out the AVD process and aids our understanding of how ionic and water flux relate to the execution of apoptosis. Finally, additional information gained from studying AVD in diverse model systems extends our knowledge of cell death and the role it plays in human disease.
# Author contributions
CB and JC wrote the review. Both authors contributed to the article and approved the submitted version.
# Funding
This research was supported by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences IZ1AE3090079. |
Strain-Promoted Azide–Alkyne Cycloaddition-Based PSMA-Targeting Ligands for Multimodal Intraoperative Tumor Detection of Prostate Cancer
Strain-promoted azide−alkyne cycloaddition (SPAAC) is a straightforward and multipurpose conjugation strategy. The use of SPAAC to link different functional elements to prostate-specific membrane antigen (PSMA) ligands would facilitate the development of a modular platform for PSMA-targeted imaging and therapy of prostate cancer (PCa). As a first proof of concept for the SPAAC chemistry platform, we synthesized and characterized four dual-labeled PSMA ligands for intraoperative radiodetection and fluorescence imaging of PCa. Ligands were synthesized using solid-phase chemistry and contained a chelator for 111 In or 99m Tc labeling. The fluorophore IRDye800CW was conjugated using SPAAC chemistry or conventional N-hydroxysuccinimide (NHS)− ester coupling. Log D values were measured and PSMA specificity of these ligands was determined in LS174T-PSMA cells. Tumor targeting was evaluated in BALB/c nude mice with subcutaneous LS174T-PSMA and LS174T wild-type tumors using μSPECT/CT imaging, fluorescence imaging, and biodistribution studies. SPAAC chemistry increased the lipophilicity of the ligands (log D range: −2.4 to −4.4). In vivo, SPAAC chemistry ligands showed high and specific accumulation in s.c. LS174T-PSMA tumors up to 24 h after injection, enabling clear visualization using μSPECT/CT and fluorescence imaging. Overall, no significant differences between the SPAAC chemistry ligands and their NHS-based counterparts were found (2 h p.i., p > 0.05), while 111 In-labeled ligands outperformed the 99m Tc ligands. Here, we demonstrate that our newly developed SPAAC-based PSMA ligands show high PSMAspecific tumor targeting. The use of click chemistry in PSMA ligand development opens up the opportunity for fast, efficient, and versatile conjugations of multiple imaging moieties and/or drugs.
# ■ introduction
Prostate cancer (PCa) is the second most common cancer in men worldwide, leading to substantial morbidity and mortality. 1 About 90% of PCa patients have a localized tumor at initial screening and are candidates for surgery.The prostate is located between critical structures, and as a consequence, the surgeon has to perform very narrow tumor resections.In approximately 15−65% of PCa patients, dependent on the disease stage, tumor resection is incomplete, caused by positive surgical tumor margins upon removal of the primary tumor and/or incomplete removal of tumor positive lymph nodes in the pelvis.Innovative approaches to improve intraoperative tumor detection can increase the chance of complete surgical resection of all tumor tissue.
To specifically detect prostate cancer cells during surgery, prostate-specific membrane antigen (PSMA) targeting ligands conjugated to one or multiple imaging moieties can be used.One of these imaging moieties is a fluorophore. Fluorescence imaging allows direct visualization of tumor tissue during surgery.This enables a more precise removal of the primary tumor with less positive surgical margins as a result. Another approach is radioguided surgery using a γ-emitting radio- nuclide, which allows for an intraoperative detection of deeperseated tumor lesions and metastatic lymph nodes.Moreover, dual-labeling of PSMA-targeting ligands can provide a powerful combination of the two complementary modalities mentioned above.Therefore, we focused on fluorescence imaging combined with the radionuclide detection of PCa using duallabeled PSMA-targeting ligands.
To develop a versatile platform that enables easy synthesis of PSMA ligands, strain-promoted azide−alkyne cycloaddition (SPAAC, N 3 -DBCO) can be used, a well-known form of click chemistry. Click chemistry reactions proceed with ease under mild nontoxic conditions (i.e., at room temperature in water) tolerating the presence of a wide range of functional groups. Hence, strain-promoted cycloadditions provide fast and highly efficient chemistry to link PSMA ligands to different functional elements.Nonetheless, there is still uncertainty about the influence of the hydrophobic functional group dibenzocyclooctene that is introduced with SPAAC on the pharmacokinetics and nonspecific uptake of the ligands.Therefore, the aim of the current study is to develop duallabeled small-molecule PSMA ligands that can be conjugated to different functional components using SPAAC chemistry (N 3 -DBCO) and to evaluate the effects of the hydrophobic dibenzocyclooctene group on the binding affinity, pharmacokinetics, and biodistribution of the ligand.
As a first proof of concept for the SPAAC chemistry-based PSMA-targeting platform, we synthesized PSMA ligands that were conjugated with the near-infrared (NIR) fluorophore IRDye800CW using SPAAC chemistry. Moreover, a chelator for either technetium-99m ( 99m Tc) or indium-111 ( 111 In) radiolabeling was added. We compared the SPAAC-based PSMA-targeting ligands with similar ligands where conventional N-hydroxysuccinimide (NHS)−ester coupling of IRDye800CW was used. With these in hand, we could determine the effect of conjugation strategy (SPAAC vs NHS) on the PSMA-binding affinity, biodistribution, and pharmacokinetics of these novel dual-labeled ligands. As a secondary aim, we evaluated the differences in affinity and biodistribution of the 111 In-labeled ligands and their 99m Tc-labeled equivalents.
# ■ results
Design and Synthesis. PSMA-1007 was previously reported to perfectly fit the active site as well as the entrance funnel of PSMA. Therefore, the design of our ligands is based on this high-affinity ligand, meaning that it consists of a naphthylalanine, aminomethyl benzoic acid, a glutamic acid, and a nicotinic acid (nonfluorinated). Importantly, the backbone of PSMA-1007 also contains two glutamic acid residues, of which the most C-terminal glutamic acid is oriented toward the exterior of PSMA, red circle). To synthesize dual-labeled ligands, we replaced the most C-terminal glutamic acid residue in the backbone of PSMA-1007 with a lysine. This lysine is then oriented toward the exterior of PSMA in the same manner as the side chain of the original glutamic acid is, providing space for conjugation of multiple functional moieties. To the side chain of this lysine, we added another lysine or azidolysine to introduce two groups that could be further functionalized with a metal chelator as well as a fluorophore. Most of the synthesis was performed on the solid phase, including the incorporation of a 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) or mercaptoacetylglycylglycylglycine (MAG3) che-lator. After cleavage from the resin, we conjugated all ligands in solution with IRDye800CW via either SPAAC (N 3 -DBCO) or NHS-ester-based chemistry. The following ligands were synthesized: PSMA-N048 (KuE-linker-DOTA-SPAAC-IR-Dye800CW), PSMA-N049 (KuE-linker-MAG3-SPAAC-IR-Dye800CW), PSMA-N050 (KuE-linker-DOTA-NHS-IR-Dye800CW), and PSMA-N051 (KuE-linker-MAG3-NHS-IRDye800CW). The ligands are further referred to as N48, N49, N50, and N51.
Radiolabeling of the Ligands. RCY of 111 In-labeled ligands (N48 and N50) exceeded 93% in all experiments (molar activity 30−33 MBq/nmol). RCY of 99m Tc-labeled ligands ranged between 15 and 69% (molar activity 22−100 MBq/nmol). All ligands were purified, resulting in a radiochemical purity >95% as determined by instant thin-layer chromatography (ITLC) and high-performance liquid chromatography (HPLC). The stability of the radiolabeled ligands in human serum was determined by RP-HPLC. After incubation for 2 h at 37°C, there were no changes in the pattern of the HPLC peaks, indicating that the radiolabel was stably coupled and there was no decomposition of the ligand.
In Vitro Characterization of the Ligands. To characterize the four ligands, we first measured the distribution coefficient (log D at pH 7.4), to determine the effect of the dibenzocyclooctene group on the lipophilicity of the PSMA ligand. The log D of 99m Tc-N49 was −2.4 compared with −3.5 for its NHS-based equivalent 99m Tc-N51 (p < 0.01). The log D of 111 In-N48 (SPAAC) was −3.2 compared with −4.4 for 111 In-PSMA-N50 (NHS) (p < 0.001). Hence, as expected the ligands with the dibenzocyclooctene group for SPAAC chemistry were more lipophilic. The use of MAG3 compared with DOTA as a chelator increased lipophilicity as well (p < 0.01). In addition to the lipophilicity, we determined the IC 50 values of the nonradiolabeled ligands, Next, we verified the PSMA-binding potential of our ligands in an in vitro binding and internalization assay using PSMApositive and -negative cells, in which all four ligands showed PSMA-specific binding. The use of SPAAC chemistry led to a significantly lower total binding of the ligand; 3.9 ± 0.5% for 99m Tc-N49 and 8.9 ± 0.6% for 111 In-N48, compared with NHS-conjugated variants; 22.9 ± 1.6% for 99m Tc-N51 and 17.2 ± 3.5% for 111 In-N50, p < 0.001). As a reference, we determined the total binding of In-PSMA-617 in our assay, which was 13.4 ± 3.5%. For all ligands, the internalized fraction ranged between 61.0 and 80.6% of the total cell-associated activity and did not differ significantly. Minimal binding and internalization were observed in PSMA-negative cells or PSMA-positive cells incubated with an excess of 2-PMPA, demonstrating that, despite their higher lipophilicity, the SPAAC chemistry-based ligands do not suffer from nonspecific binding and uptake.
Pharmacokinetics of SPAAC-Based Ligands. To establish the pharmacokinetics of the SPAAC-based PSMA ligands, the ex vivo biodistribution of 99m Tc-N49 and 111 In-N48 was determined 2, 4, and 24 h p.i. in mice bearing LS174T-PSMA-positive and -negative tumors. 99m Tc-N49 showed PSMA-specific tumor uptake >8%ID/g, which was stable up to 24 h after injectionand . Tumor uptake of 111 In-N48 was 23.8 ± 3.9%ID/g and 22.7 ± 1.5%ID/g 2 and 4 h p.i., respectively. Yet, after 24 h p.i. tumor uptake of 111 In-N48 significantly decreased to 17.6 ± 1.8%ID/ g (p < 0.05). Tumor uptake of the 111 In-labeled ligand N48 was significantly higher compared with the 99m Tc-labeled ligand N49 at all three timepoints measured (p < 0.05 for all timepoints).
For both ligands, the amount of tracer present in the blood over time decreased from 2−3%ID/g (2 h p.i.) to <0.3%ID/g (24 h p.i.). Furthermore, accumulation of the two ligands in other organs including the PSMA-negative tumor, spleen, liver, prostate, and salivary glands was low (<5%ID/g) at the 2 h p.i. timepoint and further decreased over the course of 24 h, leading to high tumor-to-background ratios at all timepoints measured (Supplementary . μSPECT/CT and fluorescence imaging of the mice revealed that both ligands clearly visualized the subcutaneous LS174T-PSMA tumors (left flank) up to 4 h p.i..
Biodistribution Comparison of SPAAC-and NHS-Based Ligands. To determine the effect of conjugation chemistry (SPAAC vs NHS) on the in vivo performance of the PSMA ligands, their tumor and normal tissue uptake was determined in mice bearing both a PSMA-positive and -negative LS174T tumor 2 h after injectionand Supplementary . Despite the high tumor uptake of all Biodistribution was determined in mice bearing subcutaneous LS174T-PSMA and LS174T wild-type xenografts. Data are expressed as %ID/g ± SD; * indicates p < 0.05, ** indicates p < 0.01, and *** indicates p < 0.001. Bioconjugate Chemistry pubs.acs.org/bc Article four ligands tested, the NHS-based ligands showed higher tumor accumulation compared to their SPAAC-based counterparts (not significant, p > 0.05). Tumor uptake of N48 (SPAAC) and N50 (NHS) labeled with 111 In were 21.2 ± 1.2%ID/g and 25.3 ± 2.0%ID/g, respectively. The lowest uptake in the PSMA-expressing tumors was measured for the 99m Tc-labeled ligands; 12.0 ± 1.4%ID/g and 17.7 ±3.7%ID/g for N49 (SPAAC) and N51 (NHS), respectively. In our LS174T s.c. tumor model, uptake of 111 In-N50 was significantly higher than both 99m Tc-labeled ligands (p < 0.01) and control ligand 111 In-PSMA-617 (18.5 ± 1.9%ID/g, p < 0.05). All four ligands showed a significantly higher uptake in the PSMA-positive tumor compared to the PSMA-negative tumor (p < 0.001)and Supplementary .
All ligands showed rapid blood clearance and minimal uptake in muscle, bone, salivary glands, and prostate at 2 h after injection, leading to high tumor-to-organ ratiosand Supplementary . The SPAAC chemistry-based ligands showed a higher liver uptake compared to their NHSbased equivalents; 4.3 ± 0.3%ID/g (SPAAC) vs 1.7 ± 0.3% ID/g (NHS) for 99m Tc and 3.4 ± 0.3%ID/g (SPAAC) vs 1.2 ± 0.1%ID/g (NHS) for 111 In (p < 0.05). Ligand accumulation in the excretory organ, the kidneys, was similar for both conjugation strategies but lower for the 99m Tc-labeled ligands (range 73.9−81.6%ID/g), compared with the 111 In-labeled ligands (>120%ID/g, p < 0.001).
Multimodal Imaging of Subcutaneous PSMA-Positive Tumors. To compare the multimodal imaging potential of our SPAAC chemistry-based ligands to that of their NHS-based equivalents, we scanned mice using both a μSPECT/CT and a fluorescence scanner.shows representative back-and side-view μSPECT/CT imagesand the corresponding fluorescence imagesof all four ligands 2 h after injection. No large differences in imaging potential were observed between the SPAAC and NHS-based ligands. All ligands could clearly visualize the subcutaneous LS174T-PSMA tumors (right flank), without showing clearly visible uptake in the PSMA-negative LS174T tumors (left flank). In addition, multimodal images visualized high renal tracer accumulation in all mice.
Multimodal Imaging of Intraperitoneal Tumors. To evaluate the applicability of the dual-labeled PSMA ligands for image-guided resections in a more clinically relevant setting, we injected six mice with intraperitoneal LS174T-PSMApositive tumors with our ligand 111 In-N50. Two hours after injection of the ligand, a preoperative μSPECT/CT scan was acquired, on which multiple i.p. tumors could be visualized. Next, we carried out image-guided resection of the i.p. tumors, of whichshows one illustrative example. NIR fluorescence imaging immediately revealed a tumor at the injection site, also visible on μSPECT/CTFinally, small nodules beneath the liver were found,6C, yellow arrow). All resected tumors showed fluorescent uptake. The kidneys were highly fluorescent as well, which hampered the detection of tumor lesions in close proximity to these execratory organs in mice.
# ■ discussion
In this study, we used strain-promoted azide−alkyne cycloaddition (SPAAC) chemistry to synthesize dual-labeled PSMA-targeting ligands that could aid in the intraoperative detection and resection of PCa. The literature has indicated that the dibenzocyclooctene group, present in the ligand after SPAAC chemistry, is rather hydrophobic and could negatively impact the affinity and nonspecific binding of small PSMAtargeting ligands. Therefore, we compared the performance of our SPAAC-based ligands with conventional NHS-ester-based PSMA ligands. Withal, the SPAAC-based ligands showed similar high affinity toward PSMA in vitro, and specific uptake and retention in PSMA-positive tumors in vivo, compared to NHS-based ligands. This indicates the feasibility of SPAAC chemistry as a versatile conjugation strategy in high-affinity PSMA ligand design.
Click chemistry has been used in many research fields because of its beneficial characteristics, including high yield, high specificity, and simplicity. 20 More specifically, it enables reasonably fast kinetics under aqueous conditions in the presence of a wide range of functional groups and no high temperatures are required.However, the dibenzocyclooctene group used in SPAAC chemistry is rather lipophilic and could alter the affinity, internalization, and nonspecific binding of small PSMA-targeting ligands in vitro.On the contrary, Wirtz et al. found that higher lipophilicity of their PSMA-I&Tbased ligands is beneficial in terms of affinity and internalization, possibly because of ligand interaction with a lipophilic binding pocket of PSMA.log D determination in our study showed that SPAAC chemistry conjugation of IRDye800CW indeed leads to more lipophilic ligands. We determined the IC 50 values of all four ligands to be in the same order of magnitude (184−475 nM). Importantly, IC 50 values were determined using the nonradiolabeled ligand and without a nonradioactive metal (Re/In) present in the chelator. Complexation of a metal in the chelator could alter ligand properties (e.g., charge, hydrophilicity, etc.) and with it ligand affinity.Overall, the affinity of the ligands was lower than that of the high-affinity PSMA-617 ligand (IC 50 : 8.5 nM). Yet, other IRDye800CW conjugated PSMA ligands reported in the literature also showed higher affinities (25 nM−1.7 μM) compared to the nonfluorescent ones like PSMA-617.Despite their lower affinity, ligand uptake in the PSMApositive tumor of our four ligands was comparable with PSMA-617. Possible explanations for this discrepancy might be the increased internalization of the dual-labeled ligands compared with PSMA-617, which could lead to increased ligand retention in the tumor. Moreover, the dual-labeled ligands Bioconjugate Chemistry pubs.acs.org/bc Article show a longer blood circulatory half-life (possibly due to larger size/charge/lipophilicity) compared to PSMA-617, which could also result in increased tumor uptake.In line with this, the study of showed that the addition of a fluorophore (including IRDye800CW) also decreased PSMA ligand affinity, while it increased ligand uptake in PSMA-positive LNCaP tumors.Next to the addition of the dye, higher ligand lipophilicity (due to the DBCO group) was reported to increase, bus also decrease tumor uptake of various click-chemistry-based tracers. Moreover, a lipophilic character of the tracer can increase nonspecific uptake in other organs. 20,In the case of PSMA ligands, higher ligand lipophilicity of PSMA-I&T-based ligands led to an increase in tumor uptake in LNCaP xenografts.In addition, Boḧmer et al. developed and characterized a copper(I)-catalyzed azide−alkyne cycloaddition (CuAAC) based PSMA ligand for PET imaging called [ 18 F]PSMA-MIC01.This CuAAC-based ligand showed specific tumor uptake in LNCaP xenografts (11.7 ± 4.2%ID/ g, 1 h p.i.) with only minor nonspecific uptake in other organs. In our study, the SPAAC chemistry PSMA ligands similarly showed high PSMA-specific tumor uptake with low nonspecific accumulation in other organs (e.g., PSMA-negative tumor, prostate, salivary glands). Yet, as might be expected, a more than 2-fold higher liver uptake was measured for the SPAAC variants. These data are in line with the statement of Notni et al. that a lipophilic character of the ligand, induced by large hydrophobic groups (i.e., aromatics such as in dibenzocyclooctene), is prone to increase the fraction of slow hepatobiliary clearance.Tumor-specific uptake values of clinically used ligands such as PSMA-617, PSMA-I&T, and PSMA-1007 reported in the literature range from 5 to 13%ID/g (LNCaP, 1/2 h p.i). 17,30−33 Tumor uptake of N49 (9%ID/g) was in a similar range and uptake of N48 (21%ID/g) was perhaps even higher (LS174T-PSMA, 2 h p.i). Nonetheless, no direct comparison could be made due to the use of the LS174T-PSMA xenografts in this study compared with LNCaP xenografts used in the literature. However, a direct comparison of the LNCaP and LS174T-PSMA xenograft models did not show major differences in PSMA-I&T tracer uptake between these models, 32 strongly indicating that the performance of our SPAAC ligands was in a similar range to those of the clinically available ligands. For radioguided surgery, 99m Tc and 111 In are used because they emit γ-photons detectable with a handheld γ probe. In the present study, as a secondary aim, we evaluated differences between DOTA-based ligands suitable for 111 In-labeling (N48, N50) and MAG3-based ligands for labeling with 99m Tc (N49, N51). Chelation of 111 In in DOTA leads to a neutral charge, whereas chelation of 99m Tc in MAG3 leads to a net charge of -1, which might be advantageous since the introduction of negative charges to increase PSMA affinity and ligand uptake in PSMA-positive tumors was reported in multiple studies. However, the use of MAG3 compared with DOTA as a chelator increased lipophilicity, which could also influence tumor uptake and affinity of the PSMA ligand. Our results show that the 111 In-DOTA ligands have a significantly higher uptake in the s.c. LS174T-PSMA-positive tumors compared with the 99m Tc-MAG3 ligands. Besides favorable in vitro and in vivo properties of the 111 In-labeled ligands, labeling of 99m Tc in the MAG3 chelator resulted in low RCY (15−69%), which was highly variable between each labeling. In comparison, labeling with 111 In always resulted in high RCY (≥93%). Furthermore, the use of MAG3 as a chelator led to more difficulties with the synthesis and overall stability of the N51 ligand (data not shown). Consequently, DOTA-based ligands might be preferred over MAG3-based variants.
As shown in the i.p. model presented in this study, tumors located deeper inside the surgical cavity may not be visualized with NIR fluorescence imaging alone due to the absorption of the emitted fluorescent light in overlying tissues (2−3 mm).This further emphasizes the importance of dual-labeled NIR fluorescence and radionuclide targeting agents that, in addition to preoperative tumor localization, allow intraoperative mapping of more deeply situated tumor lesions with a γ probe. The findings of the current study encourage clinical studies with PSMA-targeted dual-labeled ligands to enable image-guided complete resection of all PCa lesions during radical prostatectomy, preventing cancer recurrences and improving the chances for curative PCa surgeries.
The SPAAC-based conjugation strategy presented in this study provides a versatile platform in which PSMA ligands can easily be coupled to different chelators, fluorophores or anticancer drugs. In the future, it offers the opportunity to click various imaging moieties (e.g., fluorophores, radionuclides or MRI contrast agents) to the ligand for preclinical microscopy, as well as clinical diagnostic, pre-and intraoperative imaging of PCa. In addition, theranostic tracers could be synthesized that include therapeutic elements such as αor β − -emitting radionuclides for radioligand therapy, photosensitizers for PSMA-targeted photodynamic therapy, or anticancer drugs including immunomodulatory agents and chemotherapeutics. Finally, moieties that improve the pharmacokinetics of the ligands could be easily incorporated. For example, albumin binders such as Evans blue could be added. In the study of Wang et al. addition of Evans blue to PSMA-617 already led to major improvements in the pharmacokinetics of the ligand (e.g., significantly higher tumor accumulation and highly radiotherapeutic efficacy).In addition, Kuo et al. recently showed that albumin binder optimization and use of these optimal albumin binders could lead to enhanced tumor uptake and tumor-to-kidney absorbed dose ratios in 177 Lu-labeled PSMA ligands.
# ■ conclusions
We developed four dual-labeled ligands which all showed high PSMA affinity and excellent PSMA-specific tumor uptake. We compared an NHS and SPAAC chemistry-based approach to attach the fluorophore IRDye800CW. Overall, no significant differences between the SPAAC chemistry ligands and their NHS-based counterparts were found, while 111 In-labeled ligands outperformed the 99m Tc ligands. These results inspire the use of click chemistry conjugations in PSMA ligand development to enable fast, efficient, and easy coupling of various chelators, dyes, or even anticancer drugs.. Cell Culture. The LS174T colon carcinoma cell line was purchased from the American Type Culture Collection. LS174T-PSMA cells were created by stable transfection with human PSMA using the plasmid pcDNA3.1-hPSMA.Wildtype LS174T colon carcinoma cells were used as a control. All cells were cultured in RPMI 1640 medium supplemented with 10% FCS and 2 mM glutamine (5% CO 2 , 37°C). Additionally, LS174T-PSMA cells were cultured in the presence of 0.3 mg/ mL G418 geneticin.
Radiolabeling. Indium-111: Ligands (1−20 μg) were radiolabeled under metal-free conditions with 111 InCl 3 (Curium) in 0.5 M 2-(N-morpholino)ethanesulfonic acid (MES) buffer (pH 5.5, twice volume of 111 InCl 3 ). Labeling was performed at 90°C for 30 min.Ethylenediaminetetraacetic acid (EDTA, 50 mM) was added to a final concentration of 5 mM after the incubation. Molar activity after labeling ranged from 30 to 33 MBq/nmol. Ligands were purified by a Sep-Pak C18 light cartridge (Waters) and eluted from the cartridge with 50% ethanol in water.
Technetium-99m: Ligands (1−20 μg,) were radiolabeled in 45 μL of ammonium acetate (NH 4 Ac, 0.25 M, pH 8) and 15 μL of freshly prepared disodium tartrate buffer (50 mg/mL in 0.25 M NH 4 Ac), under metal-free conditions. Ascorbic acid buffer was prepared just before labeling (3 mg/mL in 10 mM HCl). Next, 5 μL of freshly prepared stannous chloride dihydrate (SnCl 2 ) buffer (4 μg/mL in ascorbic acid buffer) was added simultaneously with 99m TcO 4 − in saline, followed by incubation for 30 min at 90°C. Molar activity after labeling ranged from 16 to 100 MBq/nmol. Ligands were purified by a Sep-Pak C18 light cartridge (Waters) and eluted from the cartridge with 50% ethanol in water.
ITLC/HPLC: Radiochemical yield (RCY) was determined by instant thin-layer chromatography (ITLC) using silica gelcoated paper (Agilent Technologies) and 0.1 M ammonium acetate containing 0.1 M EDTA pH 5.5 ( 111 In) or 0.1 M Sodium Citrate pH 6.0 ( 99m Tc), as the mobile phase. In addition, RCY was measured using reverse-phase highperformance liquid chromatography (RP-HPLC) on an Agilent 1200 system (Agilent Technologies) with an in-line radiodetector (Elysia-Raytest). A C18 column (5 μm, 4.6 × 250 mm 2 ; HiChrom) was used at a flow rate of 1 mL/min. We used the following buffer system: buffer A, triethylammonium acetate (TEAA, 10 mM, pH 7); buffer B, 100% methanol; and a gradient of 97−0% buffer A (35 min).
Binding, Internalization, and IC50 Assay. Internalization assay: Binding and internalization characteristics of all ligands were compared using LS174T-PSMA and wild-type LS174T cells, cultured to confluency in six-well plates. The cells were incubated with 50 000 counts per minute (cpm) In-or 99m Tc-labeled PSMA ligand (0.1−0.25 pmol/well) in 1 mL of binding buffer (RPMI/0.5% BSA) for 2 h at 37°C. Nonspecific binding was determined by co-incubation with 2-(phosphonomethyl)pentane-1,5-dioic acid (2-PMPA, 21.57 μM). The cells were washed with PBS and incubated with acid buffer (0.1 M acetic acid, 154 mM NaCl, pH 2.6) for 10 min at 0°C to retrieve the membrane-bound fraction. After this, the membrane-bound fraction was collected, the cells were washed and lysed with 1.5 mL of 0.1 M NaOH, and cell lysis (intracellular activity) was collected. Membrane-bound activity and intracellular activity fractions were measured in a γ-counter (2480 WIZARD 2 Automatic γ Counter, PerkinElmer). 5,45 IC 50 : The 50% inhibitory concentration (IC 50 ) of the ligands was determined using LS174T-PSMA cells in a competitive binding assay. The LS174T-PSMA cells were cultured to confluency in six-well plates, followed by incubation on ice for 2 h in 1 mL of binding buffer (RPMI/ 0.5% BSA) with 50 000 cpm of 111 In-labeled PSMA-617 and a series of increasing concentrations (0.01−300 nM) of unlabeled PSMA ligands. After incubation, the cells were washed with 2 mL of PBS and lysed with 1.5 mL of 0.1 M NaOH. Cell lysis was collected from the plate and the cellassociated activity was measured in a γ-counter and IC 50 values were calculated using GraphPad Prism software version 5.03.
Lipophilicity. Log D values of all radiolabeled ligands were determined by adding 300 000 cpm (0.6−1.5 pmol) to a mixture of 3 mL of PBS (pH 7.4) and 3 mL of n-octanol. Tubes were vortexed vigorously for 1 min and centrifuged for 5 min at 201g. The concentration of radioactivity in a defined volume of each layer was measured in a well-type γ-counter.
Subcutaneous Tumor Model. Animal experiments were performed in 8-to 10-week-old male BALB/c nude mice (Janvier). The animals were housed under nonsterile conditions in individually ventilated cages (Blue line IVC, 4−5 mice per cage) with cage enrichment present and free access to water and chlorophyll-free animal chow (Sniff GmbH). The mice were subcutaneously inoculated with 3.0 × 10 6 LS174T-PSMA cells in the right flank and 1.5 × 10 6 LS174T wild-type cells in the left flank, diluted in 200 μL of complete RPMI 1640 medium. When xenografts were approximately 0.1 cm 3 (10−14 days after tumor inoculation), tracers were injected intravenously in the tail vein. The biotechnicians performing the s.c. and i.v. injections were blinded for the experimental groups and tumor-bearing mice were block-randomized into groups based on tumor size. All experiments were conducted in accordance with the guidelines of the Revised Dutch Act on Animal Experimentation and approved by the institutional Animal Welfare Committee of the Radboud university medical center.
Biodistribution, Fluorescence Imaging, and μSPECT/ CT Imaging. N48, N49, N50, and N51 were radiolabeled with 111 In or 99m Tc. Radio-HPLC chromatograms before and after Sep-Pak C18 purification show the radiochemical purity of the product before injection in mice . The mice were injected intravenously with 10 MBq 111 In-labeled N48, N50, or PSMA-617 as control (0.3 nmol, molar activity 33.3 MBq/ nmol) or 15 MBq 99m Tc-labeled N49 or N50 (0.3 nmol, molar activity 50 MBq/nmol) in PBS/0.5% BSA. Two hours post injection (p.i.), the mice were euthanized by CO 2 /O 2 asphyxiation, and images were acquired with the IVIS fluorescence imaging system (Xenogen VivoVision IVIS Lumina II, PerkinElmer), using an acquisition time of 30 s. Subsequently, μSPECT/CT images were acquired (U-SPECT II, MILabs) with a 1.0 mm diameter pinhole mouse collimator tube.The mice were scanned for 30 min followed by a CT scan (spatial resolution 160 μm, 65 kV, 615 μA) for anatomical reference. μSPECT/CT scans were reconstructed with MILabs reconstruction software, using an ordered-subset expectation maximization algorithm, energy windows 154−188 keV and 220−270 keV for 111 In, and 126−154 keV for 99m Tc, 1 Bioconjugate Chemistry pubs.acs.org/bc Article iteration, 16 subsets, voxel size of 0.4 mm. μSPECT/CT scans were analyzed and maximum intensity projections (MIPs) were created using the Inveon Research Workplace software version 4.1 (Siemens Preclinical Solutions). NIRF images were analyzed using Living Image software version 4.2 (Caliper Life Sciences). Tumors, blood, and relevant organs and tissues were dissected, weighed, and radioactivity in each sample was quantified using a well-type γ-counter. The results were expressed as the percentage of injected dose per gram of tissue (%ID/g). Pharmacokinetics. To determine the pharmacokinetics of the ligands, nine groups of five mice received an intravenous injection of 0.3 nmol 111 In-labeled N48 or N50 (10 MBq/ mouse, molar activity 33.3 MBq/nmol) or 99m Tc-labeled N49 (5 MBq/mouse, molar activity 16.7 MBq/nmol) in PBS/0.5% BSA. At 2, 4, and 24 h p.i., the mice were euthanized followed by dissection. Tissues of interest were dissected, weighed, and measured for radioactivity in a γ-counter as described above. For each ligand, two mice from the 24 h groups underwent repeated μSPECT/CT and NIRF imaging (2, 4 and 24 h p.i.). During imaging, the mice were anesthetized with 2.5% isoflurane inhalation anesthesia and kept warm with a heating pad. Images were acquired and analyzed as described above.
Intraperitoneal Tumor Model. LS174T-PSMA cells (1.0 × 10 6 ) in 200 μL of complete RPMI 1640 medium were injected intraperitoneally and grew for 28 days after inoculation. Six male BALB/c nude mice with intraperitoneally growing LS174T-PSMA tumors were intravenously injected with 111 In-labeled N50 (10 MBq and 0.3 nmol/mouse). Two hours p.i., μSPECT/CT imaging was performed preoperatively (30 min), followed by NIRF imaging of the mice in the supine position after surgical removal of skin, abdominal muscle layers, and peritoneum. After in vivo image acquisition, the visualized tumors were resected, followed by NIRF imaging to control whether residual tumor tissue was in situ.
Statistics. Statistical analyses were performed with Graph-Pad Prism, version 5.03. Results are presented as mean ± standard deviation (SD). Differences in in vitro affinity and tumor/organ uptake in vivo were tested for significance using a one-way ANOVA with a Tukey multiple comparison post-test. Differences were considered significant at p < 0.05, two-sided.
## ■ associated content
## * sı supporting information
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.bioconjchem.1c00537.
Supplemental methods on the synthesis of the ligands, HPLC analysis, and conjugation of the IRDye800CW; supplemental results on the crystal structure of PSMA-1007, serum stability, and IC50 of the ligands; radio-HPLCs before and after C18 purification; biodistribution and pharmacokinetics of the ligands; and the MALDI-ToF/HPLC spectra of all ligands (PDF) |
Protocol for developing quality assurance measures to use in surgical trials: an example from the ROMIO study
# Abstract
Introduction Randomised controlled trials (RCTs) in surgery are frequently criticised because surgeon expertise and standards of surgery are not considered or accounted for during study design. This is particularly true in pragmatic trials (which typically involve multiple centres and surgeons and are based in 'real world' settings), compared with explanatory trials (which are smaller and more tightly controlled). Objective This protocol describes a process to develop and test quality assurance (QA) measures for use within a predominantly pragmatic surgical RCT comparing minimally invasive and open techniques for oesophageal cancer (the NIHR ROMIO study). It builds on methods initiated in the ROMIO pilot RCT. Methods and analysis We have identified three distinct types of QA measure: (i) entry criteria for surgeons, through assessment of operative videos, (ii) standardisation of operative techniques (by establishing minimum key procedural phases) and (iii) monitoring of surgeons during the trial, using intraoperative photography to document key procedural phases and standardising the pathological assessment of specimens. The QA measures will be adapted from the pilot study and tested iteratively, and the video and photo assessment tools will be tested for reliability and validity. Ethics and dissemination Ethics approval was obtained (NRES Committee South West-Frenchay, 25 April 2016, ref: 16/SW/0098). Results of the QA development study will be submitted for publication in a peer-reviewed journal. trial registration number ISRCTN59036820, ISRCTN10386621.
# Introduction and rationale
Randomised controlled trials (RCTs) in surgery are notoriously difficult to design and conduct, due to numerous methodological and cultural challenges. Many of these challenges relate to the fact that surgical procedures are complex healthcare interventions, meaning that 'unlike 20 milligram tablets, no two procedures are the same' and achieving standardisation of surgical techniques and processes is difficult. 1 This is partly because surgeons naturally undertake procedures in (slightly) different ways and have differing skill levels, which may influence rates of postoperative complications and reoperation. 2 A lack of consideration for intervention standardisation and surgeon expertise in the context of RCTs may introduce bias, compromising internal validity. This is acknowledged in guidance such as Consolidated Standards of Reporting Trials of Non-Pharmacologic Treatment (CONSORT-NPT) and Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT), which provides a checklist of 33 items to be reported in trial strengths and limitations of this study ► This protocol describes the process of developing quality assurance (QA) measures to use within pragmatic surgical randomised controlled trials, which is an area lacking in methodological guidance. ► Further work will establish if these QA approaches can be used more widely in different contexts. ► This study will explore the feasibility of obtaining digital videos of open surgery during this study; however, it may not be possible to achieve this with good enough quality or without interfering with the process of surgery. The increasing availability and use of digital imaging and information technologies should facilitate the use of these methods. ► QA processes require investment of resource and expertise: future work to streamline them is therefore needed. ► While the addition of QA processes can reduce bias and improve internal validity, they may initially appear to compromise a trial's generalisability. Provided pragmatic standards are set and monitored, this should be avoided.
## Open access
protocols. They recommend reporting precise details of the intervention and its components to enable replication in routine practice, information about standardisation, procedures for monitoring adherence to intervention protocols and consideration for the expertise of care providers. Neither CONSORT-NPT nor SPIRIT differentiate between the information required in pragmatic or explanatory settings. This may be because it is rare for trials to be purely pragmatic or explanatory: trial design is not dichotomous and there is a continuum between the two extremes. In explanatory trials, which determine the efficacy of interventions, great detail may be necessary because the interventions are often novel and their safety needs to be assessed within carefully controlled settings. Pragmatic trials, which determine whether interventions are effective in the real world, are often multicentre studies with large numbers of surgeons. [bib_ref] Improving the reporting of pragmatic trials: an extension of the CONSORT statement, Zwarenstein [/bib_ref] Under such circumstances, specifying each operative step is likely to create difficulties, and ensuring that each step was delivered as planned may be unrealistic. A balance between adequate standardisation and practicality is therefore necessary and appropriate. One way of achieving this is to determine the minimum active ingredients of the intervention-those that are thought to optimise outcomes or those that are different between the interventions in each trial group-and the degree to which they need to be standardised. In this way, monitoring only the key components may be sufficient, rather than monitoring all components and steps, in order to ensure the intervention is actually delivered as planned. [bib_ref] Novel ways to explore surgical interventions in randomised controlled trials: applying case..., Blencowe [/bib_ref] This approach would also account for the fact that most trials sit within the pragmatic-explanatory continuum rather than being one or the other.
It is, therefore, important to provide reassurance about the standards of surgery in all RCTs, while recognising that this may vary according to trial design. One way of achieving this is to undertake quality assurance (QA), defined as the process(es) of 'directing the performance and behaviours of practitioners and institutions toward more appropriate and acceptable health outcomes'.Undertaking QA in surgery has been summarised in a systematic review of laparoscopic colorectal surgical studies. [bib_ref] Methods of quality assurance in multicenter trials in laparoscopic colorectal surgery: a..., Foster [/bib_ref] The review identified three distinct categories of QA measures: (i) trial entry criteria for surgeons and centres, (ii) standardisation of surgical techniques and (iii) monitoring of surgeons and/or units. Despite this, the use of such QA measures is rarely reported. In addition, it did not consider how QA processes may differ between pragmatic and explanatory trials. A recent systematic review of 80 RCTs found that 18% used entry criteria for surgeons or centres, 29% attempted to standardise the surgical procedures under evaluation (although most did not describe what the standards were), and 28% undertook some form of monitoring during the trial. [bib_ref] Systematic review of intervention design and delivery in pragmatic and explanatory surgical..., Blencowe [/bib_ref] An additional problem is that the QA processes were often selected arbitrarily and relied on surgeons' self-reported data, rather than objective measurements, which leaves them open to criticism. Practical, robust approaches to QA in pragmatic surgical studies are lacking. The aim of this study, therefore, was to develop and test QA processes for pragmatic surgical trials, in the context of a predominantly pragmatic RCT evaluating surgical techniques in upper gastrointestinal cancer surgery (Randomised Oesophagectomy-Minimally Invasive or Open (ROMIO), HTA 14/140/78). the rOMIO study The purpose of the ROMIO study is to compare, in patients with cancer of the oesophagus and oesophagogastric junction, the clinical and cost effectiveness of laparoscopically assisted (LAO) and open (OO) surgical procedures in terms of recovery, health-related quality of life, cost and survival. The RCT will be conducted in at least eight UK centres. The ROMIO study is predominantly pragmatic, as demonstrated by the Pragmatic-Explanatory Continuum Indicator Summary II wheel provided in figure 1. [bib_ref] The PRECIS-2 tool: designing trials that are fit for purpose, Loudon [/bib_ref] For example, it is multicentre and involves more than 40 surgeons. It has broad inclusion criteria and it is expected that at least 60% of patients undergoing oesophagectomy will be eligible to participate. The primary outcome is patient centred and secondary outcomes include resource use and other clinical and patient-reported outcomes. Despite this, however, assessing QA is crucial, to ensure that the LAO is performed to a similar standard as the OO, enabling a fair comparison to be made between the two techniques. Within the ROMIO pilot study, work was undertaken to begin the process of establishing QA methods. This protocol outlines plans to test and assess the feasibility of implementing these QA methods for the purposes of a multicentre RCT in surgery.
ObjECtIvEs ► To examine the variability of performance of oesophagectomy and agree on standardisation of surgery that are acceptable for a predominantly pragmatic multicentre trial. ► To pilot a tool to assess the quality of oesophagectomy undertaken within the ROMIO study. ► To pilot a tool to enable ongoing monitoring of surgeons' technical performance throughout the RCT. ► To explore the feasibility of using intraoperative digital photography and videos as methods for assessing QA in an RCT. ► To develop a feedback system for surgeons participating in the trial.
# Methods and analysis
Methods to assess QA will be developed from work initiated in the pilot RCT. There are three categories: (i) entry criteria for surgeons and centres, (ii) standardisation of surgical techniques and (iii) monitoring of surgeons and centres during the trial. Methods to assess QA in each
Open access of these categories will carefully consider the balance between extensive selection (of centres and surgeons) and standardisation (of technique) with the predominantly pragmatic nature of the study.
## Entry criteria for centres
The criteria used to select centres to participate in the ROMIO study will be based on discussions within the study management group, informed by: recommendations for cancer centres (>50 cases per year), experience of team working in trials (agreement of at least two surgeons to enter patients into the trial) and commitment (shown by the provision of centre-level data for submission to the National Oesophago-Gastric Audit).
## Entry criteria for surgeons
Rather than prevent surgeons from participating in the study, the purpose of this aspect of QA was predominantly to enable (i) variations in surgical technique and skill to be described, facilitating contextualisation of the results, (ii) provision of feedback and (iii) to establish whether LAO was broadly being performed to the same standard as OO. The feasibility of collecting videos of the abdominal phase of OO will be established. If found to be possible, participating surgeons will be required to submit one unedited video of the abdominal phase of OO (the 'standard' technique), in line with existing literature. [bib_ref] Surgical skill and complication rates after bariatric surgery, Birkmeyer [/bib_ref] Because LAO represents a 'new' technique, and to prevent surgeons from selecting only their 'best' example, two unedited videos of the abdominal phase of LAO will be required. All videos will be pseudonymised. A schema outlining the proposed development and validation of the video assessment tool is provided in [fig_ref] Figure 2: Schema for the development of a pretrial QA tool to assess surgeons'... [/fig_ref]. Current available methods for assessing the quality of surgeons' technical skills from videos include hierarchical task analysis 9 and the Objective Structured Assessment of Technical Skills (OSATS) tool. [bib_ref] Objective structured assessment of technical skill (OSATS) for surgical residents, Martin [/bib_ref] OSATS is suitable for use with any type of surgical procedure (although has not been formally tested in the context of oesophagectomy) whereas hierarchical task analyses are developed individually for specific procedures. A hierarchical task analysis for oesophagectomy (HTA-O) was developed during the pilot phase of the ROMIO study 9 ; however, it has not yet been formally tested. Both OSATS and HTA-O measures will be piloted simultaneously using 'think aloud' techniques, whereby a researcher will observe a surgeon (from the review team, see the 'Data analysis' section below) while reviewing an operative video. [bib_ref] The Use of Think-aloud methods in qualitative research an introduction to think-aloud..., Charters [/bib_ref] The surgeon will be asked to complete each measure and express their Pragmatic-Explanatory Continuum Indicator Summary II wheel: a visual representation of the predominantly pragmatic nature of the Randomised Oesophagectomy-Minimally Invasive or Open study. [bib_ref] The PRECIS-2 tool: designing trials that are fit for purpose, Loudon [/bib_ref] Trials that take an explanatory approach produce wheels nearer the hub; those with a pragmatic approach are closer to the rim.
Open access thoughts while doing so, and to vocalise any general feelings about the technical skills displayed on the video. The 'think aloud' sessions will be audio recorded, transcribed verbatim and analysed thematically. Based on the findings from this process, the existing measures (OSATS and HTA-O) may be amended or combined, and new domains added. The 'think aloud' process will be iterative-that is, it will be repeated with different videos until the surgeon has no further comments-and stop once the study team are satisfied that no new amendments are necessary. It is anticipated that two surgeons will be involved in this piloting phase.
stAndArdIsAtIOn Of surgICAl tEChnIquEs This phase of QA will be undertaken with careful consideration of the need to balance extensive standardisation with the pragmatic nature of the study, and the practical challenges associated with monitoring adherence to the standards. LAO and OO will be deconstructed into their component parts, the 'key operative components' identified, and the expected similarities and differences for each component of both procedures will be mapped and documented using a typology of surgical interventions. [bib_ref] Standardizing and monitoring the delivery of surgical interventions in randomized clinical trials, Blencowe [/bib_ref] Details of how each component is recommended to be performed (and the degree of flexibility permitted) will be agreed based on evidence from existing literature and consensus among the study team.
## Monitoring of surgery during the trial
A schema outlining the piloting of a 'photo metric' of intraoperative QA for oesophagectomy is provided in [fig_ref] Figure 3: Schema for the development of a QA tool [/fig_ref]. The photo metric is based on the premise that identification of specific anatomical structures on a photograph can be used as a measure of the quality of operative technique (because if the quality is poor, the structures would not be clearly visible). 14 A list of anatomical Open access structures that would be expected to be visible during each key component (as identified above) of LAO and OO was compiled during the pilot phase of the ROMIO study. [bib_ref] Comparing open and minimally invasive surgical procedures for oesophagectomy in the treatment..., Metcalfe [/bib_ref] This provisional list will be refined by pilot testing in the operating theatre (out with the ROMIO study), to establish whether it is technically and logistically possible to collect photos of sufficient quality to demonstrate the required anatomical structures. The refined list will be discussed and agreed by surgeons participating in the ROMIO study. Subsequently, a rating scale will be developed (to include a category to represent that the anatomy cannot be seen or assessed) and the 'photo metric' will undergo reliability and validity testing. Surgeons will be encouraged to submit photos for each patient recruited to the ROMIO study, to demonstrate that the key operative components have been performed.
All trial pathology specimens will be prepared and macroscopically and microscopically assessed in a uniform manner. The pathology data for the trial will be collected using a standardised form and represent data points included within the Royal College of Pathologists Dataset. Data points that will serve as surgical QA indicators include the length of the oesophagus and the number of harvested lymph nodes. The slides of 10% of all cases from each centre will be reviewed by the lead pathologist. Pathologists will be blinded to the randomised allocation for each sample.
dAtA COllECtIOn Patients will be asked to give written informed consent for video and photographic recordings of the procedure (separate to consent for recruitment into the ROMIO study itself), and transfer of the data to the trial coordinating centre (using local and study consent forms). LAO videos will be collected directly from the laparoscopy 'stack' already in routine use for the operations. Video recording will start from when the surgeon inserts the camera port and will end when the camera is removed after the procedure. Processes for obtaining videos of OO will be established as part of this study. Photographs will be taken using the laparoscopy 'stack' (for LAO and OO) and collected for each of the key phases identified during the 'standardisation of surgical procedures' process (described above). Open access feasibility data An additional objective of this study is to collect information about the feasibility of developing QA processes involving the collection of digital images. Technical issues (and their solutions, if appropriate) will be documented as they arise and 'standard operating procedure' documents will be produced for both the video and photography QA aspects of this study. We envisage that it may be particularly problematic to collect digital videos of open surgery, that sufficiently match the quality of laparoscopic procedures. Specific attention will be paid to developing solutions to overcome the barriers to collecting such data, including the need to minimise any interference with the usual process of surgery.
## Contextual information
We will record the total number of procedures performed by each participating surgeon, annual procedural volume and the total volume of oesophagectomies performed in each centre. The number of trainees undertaking procedures under supervision, and the number of procedures involving trainees, will be documented.
# Data analysis
Rating of videos and photos will be undertaken by a team of oesophagogastric surgeons: two participating in the ROMIO study and three routinely undertaking oesophagectomy outside of the trial. Each video and photo will be rated by at least two surgeons. As with similar previous studies, 2 no formal training or guidance will be given regarding the assessment tools. The 'rater' surgeons will not be aware of the surgeon or centre from which the videos and photos were obtained.
video assessments (surgeon entry criteria) Where videos are incomplete (eg, a component of the operation has not been captured), the surgeon will be asked to provide a further video. Videos of surgeons receiving a summary judgement of 'poor' skill (or worse) will automatically be discussed and reviewed by the study team.
Photo assessments (monitoring)
The scoring system will be used to establish a threshold at which the standard of surgery is considered 'sufficient'. This will be determined by iterative review of all photos in discussion with members of the study management group. All photos considered not to meet this standard will be reviewed by the study team. Photo assessments will be correlated with the corresponding lymph node yield for each patient.
fEEdbACk Individual feedback In cases where videos or photos do not meet the expected quality (in terms of the quality of the images as well as the standard of surgery), individualised and private feedback will be provided. In such instances, further operative videos may be requested to clarify that the feedback points have been addressed. However, given that the participants are all consultant oesophagogastric surgeons, we anticipate that this will rarely occur.
## Group feedback
To improve the overall quality of operations, which is another important part of QA, we will develop generic feedback materials for participating surgeons and centres. The generic feedback materials will be developed in two phases: (i) understanding operative techniques, variations and difficulties (by watching the videos and documenting emerging patterns and themes) and (ii) developing an action plan to address the identified difficulties and optimise operative techniques. We will achieve the second phase by developing short videos demonstrating operative techniques, based on the themes detected across the surgeons' submitted videos. These videos will be sent to all participating surgeons at regular intervals during the study. The key issues will also be discussed at ROMIO study investigators' meetings where exemplar videos of 'excellence' and 'room for improvement' will be displayed to allow self-reflection and learning.
After participating in the ROMIO study for 12 months, we will ask surgeons to submit further operative videos, to review progress and standards.
## Future work
This study aims to develop methods for measuring the QA of surgical interventions. Once developed, they will be implemented in the context of an RCT comparing open and laparoscopic surgery for patients with oesophageal cancer. There are numerous analyses that may be undertaken, which will depend on the exact nature of the QA measures that are developed. First, the quality of intervention delivery will be assessed, and to compare what surgeons reported in the case report forms (ie, what they said they did) with the intraoperative photographs (ie, what actually happened). Second, we will explore trends relating to surgeon skill (from the operative videos) and patient outcomes, though numbers are relatively small and this may not, therefore, be possible. Third, the process of developing QA measures for this study will influence future work in this area; specifically, the generation of guidance that can be extrapolated to other RCTs in surgery. This is the focus of a funded fellowship award, which will also examine the process of obtaining consensus about exactly what these QA measures should comprise, accounting for trial design and the nature of the interventions under investigation. Finally, it is important to recognise that improving the QA of surgical interventions may improve standards of surgery within an RCT. While we will document the points at which feedback is given to surgeons and centres, it may not be realistically possible to correlate this (and assign causation) with outcomes. It is well recognised that patients within RCTs generally have more favourable outcomes than those that do not. Improvements to the standards of intervention delivery Open access may, therefore, form a part of the benefits associated with trial participation.
## Public and patient involvement
Patients and the public were extensively involved in the design of the ROMIO study. During these meetings, we asked for their views about the QA aspect. They felt it was an important aspect of the RCT and did not have any issues relating to the acquisition of video and photo data relating to their operation.
EthICs And dIssEMInAtIOn Digital videos will be transferred using Open Document Information Exchange (ODIE) to the National Health Service (NHS) network for analysis by the ROMIO study team and pseudonymised with a unique identifier. The ODIE file hosting system is securely protected through use of an encrypted HTTPS link, meaning third parties cannot read exchanged data. Digital photographs will be uploaded directly by the local site staff into the purpose-designed server ROMIO database hosted on the NHS network. Information capable of identifying individuals will be held in the database with passwords restricted to ROMIO study staff.
Results of the study will be submitted for publication in a peer-reviewed journal and presented at national and international gastrointestinal conferences. Guidance documents relating to: (i) practical considerations for the development of QA procedures in surgical RCTs and (ii) generation of feedback materials to improve QA, will also be published. Feedback materials (in the form of 'gold standard' intraoperative photos of each key phase and operative technique videos) will also be made available.
[fig] Figure 2: Schema for the development of a pretrial QA tool to assess surgeons' overall operative technique in relation to oesophagectomy. HTA-O, Hierarchical Task Analysis for Oesophagectomy 8 ; OSATS, Objective Structured Assessment of Technical Skills; QA, quality assurance; ROMIO, Randomised Oesophagectomy-Minimally Invasive or Open.9 [/fig]
[fig] Figure 3: Schema for the development of a QA tool (photo metric) to document the adequacy of crucial intraoperative steps building on work done in the pilot randomised controlled trial. QA, quality assurance. [/fig]
|
Systematic review of adherence rates by medication class in type 2 diabetes: a study protocol
Introduction: Treatment options for type 2 diabetes are becoming increasingly complex with people often prescribed multiple medications, and may include both oral and injectable therapies. There is ongoing debate about which drug classes provide the optimum second-line and third-line treatment options. In the real world, patient adherence and persistence determines medication effectiveness. A better understanding of adherence may help inform the choice of second-line and third-line drug classes.
Methods and analysis: This systematic review will compare adherence and persistence rates across the different classes of medication available to people with type 2 diabetes. It will include all identified studies comparing medication adherence or persistence between two or more glucose-lowering medications in people with type 2 diabetes. Research databases (MEDLINE, EMBASE, The Cochrane Library, The Register of Controlled Trials, PsychINFO and CINAHL) will be searched for relevant articles, using a comprehensive search strategy. All identified medication trials and observational studies will be included which compare adherence or persistence across classes of diabetes medication. The characteristics and outcomes of all the included studies will be reported along with a study quality grade, assessed using the Cochrane Risk Assessment Tool. The quality of adjustment for confounders of adherence or persistence will be reported for each study. Where multiple (n ≥3) studies provide compare adherence or persistence across the same 2 medication classes, a meta-analysis will be performed.
## Ethics and dissemination: no ethics approval is
required. This review and meta-analysis (where possible) will provide important information on the relative patient adherence and persistence, with the different classes of diabetes therapies. Once complete, the results will be made available by peer-reviewed publication.
Trial registration number: CRD42015027865.
# Introduction
Type 2 diabetes is a growing problem worldwide. [bib_ref] The worldwide epidemiology of type 2 diabetes mellitus-present and future perspectives, Chen [/bib_ref] Alongside this increased prevalence comes an increased burden of disease from complications, which are potentially preventable. [bib_ref] Globalization of diabetes: the role of diet, lifestyle, and genes, Hu [/bib_ref] The importance of patient education and promoting self-care has long been recognised as a key component in chronic disease management and improving patient outcomes. [bib_ref] Effectiveness of self-management training in type 2 diabetes: a systematic review of..., Norris [/bib_ref] One component of self-care is adherence to often complicated medication regimes. Good adherence is associated with reduced risk of diabetes complications, reduced mortality and reduced economic burden. [bib_ref] Medication adherence: its importance in cardiovascular outcomes, Ho [/bib_ref] [bib_ref] The impact of treatment noncompliance on mortality in people with type 2..., Currie [/bib_ref] [bib_ref] The economic consequences of noncompliance in cardiovascular disease and related conditions: a..., Muszbek [/bib_ref] However, a substantial proportion of people with type 2 diabetes do not take medication as prescribed, with only 67-85% of oral medication doses taken, and approximately 60% of insulin doses. [bib_ref] A systematic review of adherence with medications for diabetes, Cramer [/bib_ref] In recent years, new medication classes have been introduced which have increased the complexity of diabetes treatment algorithms leading to multiple second-line and third-line options.Selection from these options usually follows a discussion with the patient and involves assessing potential efficacy, contraindications, potential side effects (especially hypoglycaemia risk and weight change), medication cost, frequency of dosing, whether the medication is required to be administered as an injection and whether monitoring of blood indices is required. However, a key component of medication effectiveness, rarely discussed at the outset, is patient adherence, and this is especially important in diabetes as overall adherence to therapies is poor.
Adherence is determined by many factors: patient-related, condition-related, socioeconomic, health system-related and therapy-related (table 1) factors.While interventions aimed at improving medication adherence should target all these main domains, it is the therapy-related component that will predominantly influence which therapy is chosen. Identification of any substantial differences in adherence rates between medication classes is vitally important. While there have been previous reviews of medication adherence in type 2 diabetes, these have not specifically interrogated differences in adherence rates between medications. The side-effect profile of each class of diabetes medication is different, and is likely to have a differential impact on adherence. Similarly, different dosing regimens and administration routes between classes will influence adherence. Knowledge of any differential in adherence rates between classes will help to inform prescribing practice, and is currently lacking.
A wide range of terms have been used to describe patient use (or non-use) of prescribed medications including: adherence, compliance, concordance and persistence. Two distinct patterns of medication use are described by these terms; missed medication doses (generally described by the terms adherence, compliance and concordance) and duration of use before termination ( persistence). The term concordance is preferred by some to emphasise the joint agreement by the prescribing physician and the patient to use the prescribed medication in a certain way. [bib_ref] Compliance, concordance, adherence, Aronson [/bib_ref] 'Adherence' is the term recommended for use by WHO,however, the term most commonly used in the literature is compliance. [bib_ref] Medication compliance and persistence: terminology and definitions, Cramer [/bib_ref] For this study, we will use the term adherence. Persistence with medication is the preferred term to refer to the duration of medication use. [bib_ref] Medication compliance and persistence: terminology and definitions, Cramer [/bib_ref] This is another key component of effectiveness of clinical medication, and is closely linked with medication adherence.
Knowing the relative adherence and persistence rates of medications used for the treatment of type 2 diabetes is important when selecting from the increasingly complex array of therapies.
# Methods
## Research objective
This systematic review will compare adherence and persistence rates between different classes of therapies for the treatment of hyperglycaemia in people with type 2 diabetes. It will attempt to answer the question: Are there substantially different adherence and persistence rates between the different classes of therapies available for type 2 diabetes? The key components of the study are outlined in table 2.
Literature that compares adherence rates and persistence between different medication classes will be identified and, where appropriate, synthesised and summarised. This review forms an early component of a wider research theme: the University of Surrey-Lilly Real World Evidence projects. These projects aim to use routine primary care records to identify areas for improvement of diabetes management. This review will inform further research as part of these projects, and by the wider research community into medication adherence.
## Interventions
The adherence and persistence with all available glucose-lowering therapies will be explored. These will be grouped by drug class for comparison [fig_ref] Table 3: Classes of medications used in type 2 diabetes [/fig_ref]. We will include medications which are in current use, those which have been previously withdrawn or discontinued, and those which have not yet been licensed, if sufficient data on adherence is available. Data on medications which are not currently in general use may provide important information on the adherence to the medication class, for example, insulin inhalation (trade name, Exubera)-an inhaled insulin discontinued in 2007may provide useful information on adherence to inhaled insulin use.
We will perform secondary analyses to identify differences in adherence and persistence within classes where daily, weekly and depot preparations exist (such as glucagone-like peptide 1 agonists). We will also perform a secondary analysis of the difference in adherence and persistence with the biguanide metformin standard The key components of the systematic review described using the standard PICOS (Participants, Interventions, Comparisons, Outcomes and Study design) formula
## Participants
People with type 2 diabetes who required treatment with one or more medications Interventions Glucose-lowering medications Comparisons Other classes of glucose-lowering medications (metformin to be used as the default comparator where available) Outcomes
Any reported measure of patient adherence or persistence with treatment Study design Both interventional and observational studies Lower socioeconomic status, lack of financial resources/increased medication costs, lower education level, lack of family support 36 Condition-related Presence of depression, 37 or other chronic diseases, [bib_ref] Predictors of nonpersistence with thiazolidinediones in patients with type 2 diabetes, Leblond [/bib_ref] [bib_ref] Influence of previous medication compliance on future compliance in patients with type..., Gwadry-Sridhar [/bib_ref] [bib_ref] Predictors of response to liraglutide in Japanese type 2 diabetes, Toyoda [/bib_ref] shorter duration of diabetes, [bib_ref] Determinants of adherence to diabetes medications: findings from a large pharmacy claims..., Kirkman [/bib_ref] fewer diabetes complications [bib_ref] Therapeutic compliance: a prospective analysis of various factors involved in the adherence..., Bezie [/bib_ref]
## Health system
Healthcare centre/clinic, [bib_ref] Wide clinic-level variation in adherence to oral diabetes medications in the VA, Bryson [/bib_ref] increased distance to nearest pharmacy, [bib_ref] Understanding Predictors of Compliance in Fixed-Dose Combination vs Loose-Dose Combination Therapy for..., Gwadry-Sridhar [/bib_ref] lower continuity of care 43 Therapy-related Concurrent medication use, adverse effects, poor previous experience with medication [bib_ref] Understanding Predictors of Compliance in Fixed-Dose Combination vs Loose-Dose Combination Therapy for..., Gwadry-Sridhar [/bib_ref] formulation and metformin-modified release. In our experience, there is a substantial difference in tolerability of these preparations.
## Comparisons
Medication classes will be compared (within individual studies) with other available medication classes. Given the previously identified 7 heterogeneity in studies designs, and both measures and reporting methods of adherence, it is likely that direct comparison of classes will have to be intrastudy only. For this reason, studies reporting adherence or persistence rates in a single drug class will be excluded.
## Outcome measures
Adherence and persistence will be defined as: ▸ Adherence: The proportion (or estimated proportion) of doses taken over a defined time, or the proportion of people attaining a defined level of adherence. Where possible, we will use the recommended definition of adequate adherence to be a medication possession ratio (MPR) ≥80% or proportion of days covered (PDC) ≥80% 20 or proportion of doses taken (when measured directly) to be ≥80%. MPR is defined as the number of days' supply of medication divided by the medication refill interval (with some variation in the formula used). PDC is defined as proportion of days the patient has medication available in a defined period. PDC allows adjustment for overlaps caused by early collection of prescriptions. ▸ Persistence: The proportion of people who remain on a therapy after a defined period of time, or the mean duration of therapy. Persistence may be defined as medication refills consistent with ongoing use of the medication. [bib_ref] Standardizing terminology and definitions of medication adherence and persistence in research employing..., Raebel [/bib_ref] Failure to refill a prescription after a predefined gap indicates non-persistence. Differing definitions of adherence affect outcomes, with the impact of this effect apparent when stratifying studies by the definition used. [bib_ref] A meta-analysis of the association between adherence to drug therapy and mortality, Simpson [/bib_ref] We will duplicate this method; stratifying studies by the threshold used to define adequate adherence, and use a similar method to stratify studies by the definition of persistence (see Data synthesis section). Primary outcomes ▸ Comparison of adherence or persistence between metformin (standard formulation) and other hypoglycaemic agents. Secondary outcomes ▸ Comparison of adherence or persistence between two or more hypoglycaemic agents, where metformin is not the comparator. ▸ Any differences between interventional and observational reports of studies. ▸ Analysis of the impact of the definition of adherence/persistence on the relative adherence rates to diabetes therapies. ▸ Any reported reasons for differences in adherence or persistence between medication classes. ▸ Any reported differences in disease outcomes (glycaemic control or diabetes complications) in patients due to differing adherence rates between medication classes. ▸ Assessment of the quality of adjustment for confounding in existing studies.
## Included study types
All interventional and observational studies comparing adherence rates and/or persistence with two or more diabetes medication classes will be included.
Interventional studies types will include randomised, non-randomised, and cross-over trials. Interventions can include interventions aimed at improving adherence or the use of medication as the intervention (clinical trials). Observational studies will include cohort and cross-sectional studies (eg, those with surveys of patient reported compliance).
## Study inclusion criteria
Studies will be included in the review if all the participants taking the medications of interest have type 2 diabetes. Studies from the community setting will be included; hospital outpatient studies and community studies, including primary care. Studies measuring adherence using the following methods will be included; self-reported measures, clinician estimated adherence, calculated adherence rates from prescribing or dispensing data, or electronic monitoring of medication use. Electronic monitoring technology broadly involves the use of a medication container which collects data on bottle opening. Studies will be included if they report a sufficient measure of adherence and/or persistence. For this purpose, adherence will be defined as the proportion of doses taken over a given time period ( period defined in the study); reported as a fraction; or reported as a proportion of patients achieving a predefined level of adherence. Persistence will be defined as the proportion of people who remain taking a therapy after a specified treatment period or the mean number of days until discontinuation. These definitions are consistent with previous work. [bib_ref] A systematic review of adherence with medications for diabetes, Cramer [/bib_ref] Exclusion criteria Studies will be excluded if ▸ They include populations with only type 1 diabetes, people without diabetes taking the medication of interest, or mixed populations where adherence rates in those with type 2 diabetes were not reported separately. ▸ The medications studied are not used to treat diabetes. ▸ No clear measure of adherence or persistence is reported. ▸ Adherence or persistence rates in diabetes during pregnancy. ▸ Adherence or persistence rates with medications from a single class only are reported. ▸ Adherence or persistence rates are not reported separately by individual drug or by drug class. ▸ The study participants are hospital inpatients or inpatients at another institution. ▸ The study participants have medication administered by a caregiver or professional.
Search method Studies will be obtained by searching the following databases: ▸ MEDLINE ▸ EMBASE ▸ The Cochrane Library including the Cochrane Controlled Trials Register (CENTRAL) and the Database of reviews of effectiveness (DARE) and the NHS Health Economics Database ▸ The meta Register of Controlled Trials (http://www. controlled-trials.com) ▸ PsychINFO ▸ CINAHL Additionally, bibliographic searches of the included studies will also be performed. Only studies published from 2006 onwards will be included to provide a summary of the most recent 10 years of publications.
## Search terms
The search strategy detailed in online supplementary appendix 1 will be used to search the MEDLINE database. This strategy will be adapted to apply to the other included databases. Only English language studies will be included.
## Study selection
Two clinicians will independently review the titles and abstracts of all the articles identified using the search terms described. Studies will be included which match the above inclusion criteria and do not have any exclusion characteristics. Studies will be eliminated if both reviewers agreed that the study did not meet the required criteria. Studies which either or both reviewers feel merit further analysis will be included for full paper review. Full articles will be collected and collated into unique studies. Full papers for each study will be reviewed separately by each reviewer against the inclusion and exclusion criteria. In the case of disagreement between the reviewers, a third reviewer will be included to resolve the discrepancy. If amendments are needed to the inclusion or exclusion criteria to clarify whether a particular study is included or excluded these will be reported in the final review manuscript alongside details of the paper(s) in question.
Data extraction Data will be extracted using a standardised method into a preformatted database. Data will be extracted on study type and intervention (if relevant), study location, sample size, study population characteristics (age, gender, comorbidities, concurrent medication use-including pill burden, and duration of diabetes), duration of follow-up, inclusion and exclusion criteria, and funding sources. Outcome data extracted will comprise adherence or persistence measurement method, adherence values, persistence values, suggested or measured reasons for any reported interclass adherence or persistence differences, and impact on glucose control and diabetes complications. All extracted data will be verified by a second reviewer.
## Assessment of evidence quality
The internal validity of all included studies will be assessed using the Cochrane risk assessment tools. This enables evaluation of clinical trials based on the following criteria; sequence generation, allocation concealment, blinding of participants, personnel and outcome assessors, incomplete outcome data, selective outcome reporting, and other sources of bias. All publications will be categorised as either low risk, unclear risk or high risk or bias. For cohort studies, the risk assessment tool enables assessment based on the following domains; selection of population, assessment of exposure, assessment of the present of outcomes at baseline (not relevant here), matching or adjustment for confounders, identification of prognostic variables, assessment of the outcome measure, adequacy of follow-up, and similarity of co-interventions between groups. We will report the quality of all the included studies, and will comment on the overall impact of data quality on the conclusions drawn by the review. We will rank the overall quality of evidence on this topic using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) scoring system.Data analysis and presentation We will provide a description of the number of studies included and excluded at each stage of the search process. A study flow diagram will be incorporated in the standard form suggested by the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement. [bib_ref] Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, Moher [/bib_ref] The characteristics of all the studies included will be presented with the study outcome findings. We will provide a text summary of the main results, and a discussion including details of the quality of existing evidence and recommendations for future research.
## Data synthesis
Factors such as age, gender, comorbidities and concurrent medication use influence both medication choice and adherence. Untangling the contribution to non-adherence caused by these confounders from that caused by the medication class will be complex. We will address this issue in several ways: 1. We will perform a comparison of adherence between classes in randomised trials. Where randomisation is sufficient, any differences in adherence are likely to be related to medication factors only. 2. In observational studies, we will assess and report the methods used to identify and adjust for potential confounding factors. We will report which confounders have been adjusted for in each included study, and report those which have not been measured or reported. We will comment on the limitation of the methods used to control for confounders where this has been performed.
3. We compare relative adherence rates for each drug comparison in all studies (via meta-analysis as described below) with those with a method of adjustment or controlling for confounders. These methods may include adjustment via statistical methods, such as regression analysis, or by controlling for confounding via case selection (such as case-control studies). 4. We will perform meta-analyses for each pair of medication classes compared where there are a sufficient number of studies enabling comparison (n ≥3). The possible pairwise comparisons are shown in [fig_ref] Figure 1: A schematic representation of the possible pairwise comparisons of adherence and persistence... [/fig_ref].
For each pairwise comparison, we will undertake the following analyses: studies, and only observational studies which adjust for confounders of persistence.
For all meta-analyses, we will report tests for heterogeneity (Q statistics and I 2 ).
## Review registration
This review has been registered with PROSPERO (registration: CRD42015027865).
## Updates to study protocol
If any updates to the study protocol are required, these will be listed and included as supplementary information alongside a final manuscript, and updated on the PROSPERO register.
## Review time frames
We will aim to complete the review within a period of 1 year from the start of the literature search. This time period excludes time spent following submission of a final manuscript and article peer review.
# Conclusions
This review and, if possible, meta-analysis, will provide important information on patient adherence and persistence with the different classes of diabetes therapies. This is one of the key components of medication effectiveness, and will help to guide clinical decision-making about which is the optimum second-line or third-line diabetes therapy to initiate. The review will also enable key recommendation for further research in this area to be made.
Contributors AM wrote the manuscript and designed the review. ZT, WH, NM, MW, and SdL critically appraised the review design and systematic review proposal manuscript. All authors have approved the final manuscript.
Funding Eli Lilly and Company.
Competing interests AMG, WH, and SdL are funded as part of the University of Surrey-Lilly Real World Evidence (RWE) projects. This centre is funded by Eli-Lilly Pharmaceuticals. Lilly are updated periodically with the research performed as part of the RWE projects although these funders have not had a role in development of this protocol. NM has received fees for serving as a speaker, a consultant or an advisory board member for Allergan, Bristol-Myers Squibb-Astra-Zeneca, GlaxoSmithKline, Eli Lilly, Lifescan, MSD, Metronic, Novartis, Novo Nordisk, Pfizer, Sankio, Sanofi, Roche, Servier, Takeda. MW has received speaker fees from Astra-Zeneca.
Provenance and peer review Not commissioned; externally peer reviewed.
Open Access This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work noncommercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http:// creativecommons.org/licenses/by-nc/4.0/
[fig] Figure 1: A schematic representation of the possible pairwise comparisons of adherence and persistence between different medication classes. Comparison with metformin is the primary outcome (black lines), but the additional pairwise comparisons possible are secondary outcomes of interest (grey lines). [/fig]
[table] Table 3: Classes of medications used in type 2 diabetes [/table]
|
Structure of the human NK cell NKR-P1:LLT1 receptor:ligand complex reveals clustering in the immune synapse
Structure of the human NK cell NKR-P1:LLT1 receptor:ligand complex reveals clustering in the immune synapse REVIEWER COMMENTS Reviewer #1 (Remarks to the Author):In the revision of the manuscript initially submitted about 4 years ago, the authors have constructively addressed the main concerns of the reviewer. Indeed, the authors are to be congratulated in substantiating the main conclusion of the paper by conducting a series of complementary solution based and single molecule imaging experiments while weaving in some analyses of the mouse NKR-P1b complex structures. In doing so the authors present a compelling revision that provides new mechanistic insight into human NKR-P1-LLT1 structure and oligomerisation and how this relates to an increased understanding of NK cell function.Reviewer #2 (Remarks to the Author):My review relates to the SEC-SAXS analysis that was carried on after the first round of reviews. As stated by Reviewer 1, "in the NKR-P1/LLT1 crystal, each monomer of the LLT1 homodimer is observed to contact a different monomer of the NKR-P1 homodimer. The primary interaction mode corresponds closely to the interface found in the NKp65/KACL complex,..., but evidence for the biological relevance of the secondary interaction mode is much less convincing." SEC-SAXS was then a sound approach to "confirm" the existence of such an interaction in solution. Unfortunately, the way the SEC-SAXS experiments were carried on and analysed could not answer the question. Since the complex NKR-P1/LLT1 is in equilibirum with the individual partners, it necessarily dissociates during its elution through the SEC column, even if the concentration of the injected solution is high. It is therefore no surprise that the main elution peak corresponds to several species eluting closely to each other. This reflects well in the Rg profile ofFigure 4, for which no clear plateau is seen. The authors then choose a strange strategy, consisting in merging the curves from several peaks (even those from higher oligomeric species) and analyse them as a large ensemble. It seems to me that this strategy, by blurring back the SEC-SAXS information, contradicts the very interest of SEC-SAXS, which is to isolate the individual curves, and analyse them on their own. The conclusions drawn by the analysis based on OLIGOMER, are necessarily more speculative than what the analysis of a curve from a monodisperse species would be. I would rather suggest to use a deconvolution strategie of the main elution peak, based on EFA or on US-Somo approaches. There will then be a chance to get the curves of the individual species and analyse them accordingly. Alternatively, a more experimental strategy would be to saturate the elution buffer with one of the partners during the SEC-SAXS elution, in order to shift the equiibrium towards the formation of the complex. This strategy is of course much more expensive, but it could then unambiguously provide the scattering curve of the complex alone: see for instance https://doi.org/10.
antibody, it would be useful to determine how the author's determined that the NKR-P1 homodimers were doubly labeled with Alexa-647. And additionally, the determination of the 15 nm localization precision, as this can be greatly affected by dyes per secondary antibody, packing of multiple dyes in a small area leading to either quenching or fret effects on the emission, under labeling of the target, and labeling of the target with secondary antibodies with no dye or bleached dyes conjugated. Since the change in size and event number was modest, it is relevant to report how the authors determined that a majority of NKR-P1 molecules were being labeled with single dyes. This is especially the case as most commercial secondary dyes are labeled with 4-7 dyes per antibody. Because the changes in size and events observed fall within the StDev and the localization precision, it would be relevant to provide additional controls to support the significance of this change. And to ensure that the statistical tests between the measurements account for multiple measurements. Several distributions appear to have more than one population present, possibly single homodimers and multimers, and in that case reporting the mean value may not be as relevant as changes between the overall histograms of the distributions, which do not preassume a single population.
Further support of multimerization at this level could be provided by observing single or up to 3step permanent photobleaching of antibodies in the absence of the MEA buffers to promote blinking, or with another secondary fluorophore that is more stable.
Also-it may be relevant to discuss whether a single homodimer is considered a cluster? or if another terminology may be more relevant to distinguish between lightly cross-linked homodimers, versus singly-distributed homodimers.
## Reviewer #1 (remarks to the author):
In the revision of the manuscript initially submitted about 4 years ago, the authors have constructively addressed the main concerns of the reviewer. Indeed, the authors are to be congratulated in substantiating the main conclusion of the paper by conducting a series of complementary solution-based and single-molecule imaging experiments while weaving in some analyses of the mouse NKR-P1B complex structures. In doing so, the authors present a compelling revision that provides new mechanistic insight into human NKR-P1-LLT1 structure and oligomerisation and how this relates to an increased understanding of NK cell function.
We are grateful to the reviewer for the positive feedback and hope we will continue contributing to understanding NK cell function in the future.
## Reviewer #2 (remarks to the author):
My review relates to the SEC-SAXS analysis that was carried on after the first round of reviews.
As stated by Reviewer 1, "in the NKR-P1/LLT1 crystal, each monomer of the LLT1 homodimer is observed to contact a different monomer of the NKR-P1 homodimer. The primary interaction mode corresponds closely to the interface found in the NKp65/KACL complex, but evidence for the biological relevance of the secondary interaction mode is much less convincing." SEC-SAXS was then a sound approach to "confirm" the existence of such an interaction in solution.
We thank the reviewer for commenting on the SEC-SAXS part of our research. Indeed, at the time of first submitting this manuscript, the presence of the secondary binding mode in solution was assessed by the SAXS experiment performed in batch mode, and we were not able to deconvolute the data corresponding to a mixture of monomers, dimers, and higher interacting species. This is indeed why we have chosen to proceed with the SEC-SAXS measurement, thus separating the higher oligomeric species from dissociating monomers/dimers. Nevertheless, we would like to point out that since the first submission, we have complemented the evidence for the presented secondary binding mode also with a nanoscopy experiment assessing the formation of clusters on cellular surface and by observing the direct effect on NK cell inhibition. Thus, the SEC-SAXS data presented here are not the only experimental evidence for the biological relevance of the secondary binding mode observed in the crystal structure of the NKR-P1/LLT1 complex.
Unfortunately, the way the SEC-SAXS experiments were carried on and analysed could not answer the question. Since the complex NKR-P1/LLT1 is in equilibrium with the individual partners, it necessarily dissociates during its elution through the SEC column, even if the concentration of the injected solution is high. It is therefore no surprise that the main elution peak corresponds to several species eluting closely to each other. This reflects well in the Rg profile of , for which no clear plateau is seen. The authors then choose a strange strategy, consisting in merging the curves from several peaks (even those from higher oligomeric species) and analyse them as a large ensemble. It seems to me that this strategy, by blurring back the SEC-SAXS information, contradicts the very interest of SEC-SAXS, which is to isolate the individual curves, and analyse them on their own. The conclusions drawn by the analysis based on OLIGOMER, are necessarily more speculative than what the analysis of a curve from a monodisperse species would be.
Here we have to agree with the reviewer that the analysis performed by OLIGOMER would be indeed more speculative if it were indeed based on merging curves from several resolved peaks while analyzing them as a large ensemble. However, that is not what we present in the manuscript.
As seen in , we obtained two main resolved peaks in size-exclusion chromatography, one corresponding to the higher oligomeric species and the other containing dissociating monomeric/dimeric species. Although the Rg profile does not have an apparent plateau, we have selected four separate intervals from the first peak and two intervals from the second peak where the Rg delta environment has only a low difference. These intervals are denoted in as columns with diagonal hatching with corresponding frame numbers in the figure legend. Data from these intervals were then merged separatelywe have not mixed or merged data between these intervals. We have performed OLIGOMER analysis on single curves throughout the data as well (this is now shown in the Supplementary Data in the revised manuscript); however, due to the low signal of the first peak that contains the more interesting oligomeric species, it proved necessary to boost the signal-to-noise ratio by merging small intervals with low local Rg difference in order to fit the imperfectly resolved mixture of oligomeric species within the data. We agree this approach does not entirely deconvolute the data; however, it proves the point that we are observing larger oligomeric species in solution which are in equilibrium with monomer/dimeric species, and it also shows that the oligomeric models based on our crystal structure fit the obtained SAXS curves quite well (as shown in .
We have changed the text in paragraph "SEC-SAXS analysis confirms NKR-P1:LLT1 higherorder complex formation" as well as in figure legends and methods to make it more transparent that we have not mixed and merged data from two distinct elution peaks.
I would rather suggest to use a deconvolution strategy of the main elution peak, based on EFA or on US-SOMO approaches. There will then be a chance to get the curves of the individual species and analyse them accordingly. Alternatively, a more experimental strategy would be to saturate the elution buffer with one of the partners during the SEC-SAXS elution, in order to shift the equilibrium towards the formation of the complex. This strategy is of course much more expensive, but it could then unambiguously provide the scattering curve of the complex alone: see for instance https://doi.org/10.1038/emboj.2011.461.
As per the reviewer's suggestion, we have tried to analyze the data with US-SOMO; however, this approach proved unsuccessful. Given the complexity of the mixture, i.e., monomeric and dimeric forms of both proteins interacting in the primary or secondary or both interaction modes, it is impossible to specify how exactly the peak should be deconvoluted, and we are afraid that attempting this could provide biased results. On the contrary, the analysis performed by the OLIGOMER where all possible combinations of all species were subjected as a library to the algorithm, and then the algorithm used it to best-fit the data, thus selecting the most probable combinations of species present in the samples, is completely unbiased. As mentioned above, we sampled through the SEC-SAXS data by analyzing several discrete, distinct data intervals, which at least partially deconvolutes the data, just in a different way. Even if the buffer were saturated with one of the binding partners, still, there would be an equilibrium of different oligomeric species utilizing varying ratios of the primary and secondary interaction modes, thus necessitating the same type of analysis as already applied. We believe the main confusion here was that the reviewer misunderstood how we sampled and merged the data, and we apologize for not being clearer. The corresponding parts of the text were edited to provide the reader with a more descriptive explanation.
## Reviewer #3 (remarks to the author):
This review is only focused on the super-resolution imaging portion of the paper.
The authors present data demonstrating that NKR-P1 increases in average cluster size and events per cluster in the presence of soluble LLT1, and that this change in distribution requires the presence of the three predicted strongest contacts in the LLT1 secondary interaction interface. While the data is consistent with the other findings with complementary methods, the manuscript would be strengthened by further justification of several experimental and analysis design choices and the inclusion of other relevant controls to ensure consistent labeling of the proteins imaged.
One comment is regarding the experimental design, in which LLT1 and LLT1-SIM are introduced to the cells in a soluble form, which differs dramatically from the presentation of LLT1 on the surface of an activated monocyte, B cell, NK or T cell. It would be relevant for the authors to address how this change in presentation of LLT1, and thus the biophysical constraints that would be placed upon LLT1 if it were attached to a cellular membrane, may affect its binding to NKR-P1.
We thank the reviewer for raising this point. Indeed, the experimental setup with one binding partner embedded in the membrane and the other being presented as soluble does not describe the biological reality of intercellular contact. However, this dSTORM experiment aimed to test the hypothesis of the NKR-P1 receptor's ability to form clusters (cross-linked oligomers) in the context of the cell membrane upon engaging the dimeric species of LLT1. This experimental setup allowed us to normalize the concentration of LLT1 throughout the measurement while ensuring the presence of primarily dimeric LLT1 in the solution. Compared to utilizing LLT1 cell transfectants, we have simplified the experimental setup to achieve higher reproducibility of the measurements. Furthermore, when we tested presenting LLT1 on supported lipid bilayers (not shown in the current manuscript), thus getting closer to the biological reality, sadly, mainly monomeric species of LLT1 were present on the lipidic bilayers at the low concentrations of LLT1 necessary for single molecule tracking and bleaching. Thus, this approach proved not suitable at the moment; however, we agree with the reviewer and have updated the manuscript's text to address the difference in LLT1 presentation.
Also, because dSTORM imaging can be very affected by the number of labels on the secondary antibody, it would be useful to determine how the authors determined that the NKR-P1 homodimers were doubly labeled with Alexa-647. And additionally, the determination of the 15 nm localization precision, as this can be greatly affected by dyes per secondary antibody, packing of multiple dyes in a small area leading to either quenching or fret effects on the emission, under labeling of the target, and labeling of the target with secondary antibodies with no dye or bleached dyes conjugated. Since the change in size and event number was modest, it is relevant to report how the authors determined that a majority of NKR-P1 molecules were being labeled with single dyes. This is especially the case as most commercial secondary dyes are labeled with 4-7 dyes per antibody. Because the changes in size and events observed fall within the StDev and the localization precision, it would be relevant to provide additional controls to support the significance of this change. And to ensure that the statistical tests between the measurements account for multiple measurements. Several distributions appear to have more than one population present, possibly single homodimers and multimers, and in that case reporting the mean value may not be as relevant as changes between the overall histograms of the distributions, which do not presume a single population.
Further support of multimerization at this level could be provided by observing single or up to 3-step permanent photobleaching of antibodies in the absence of the MEA buffers to promote blinking, or with another secondary fluorophore that is more stable.
We thank the reviewer for this technical note. However, we used only primary antibodies in our study, directly labeled by the manufacturer with AF647. We used the same batch and concentration of the antibody to ensure consistent labeling within the single measurements for all experiments. Moreover, we always measured conditions with and without the soluble ligand in each series of individual measurements. The negative controls were also evaluated (the condition with NKR-P1 and PBS added instead of protein ligands) to ensure consistent overall measurement quality and avoid possible errors resulting from different microscope settings or sample handling.
Being aware that this experimental approach does not directly describe the size of the receptor clusters (if any) due to potential overcounting artifacts, we instead focused only on the comparison of three distinct states of the receptor: free receptor on the membrane, receptor interacting with ligand potentially inducing clustering/oligomerization by its two interaction modes, and receptor interacting with ligand bearing only primary interaction interface, and assessing the relative change of the observed cluster of events sizes between these three states.
As the cluster of events parameters are averaged over the whole cell (i.e., single plotted point corresponds to cluster parameters observed in a single cell), the statistical tests between the plotted data indeed account for multiple measurementsi.e., cluster data collected on various cells in multiple experiments.
The cluster of events size histogram, which may point at the combination of different oligomer contributions, was assessed for every single cell, but as we are not sure about the value (size) corresponding to the NKR-P1 homodimer alone (random or some pre-existing oligomers may form on the cell surface even without ligand), we were not able to deconvolute the curve.
Also -it may be relevant to discuss whether a single homodimer is considered a cluster? Or if another terminology may be more relevant to distinguish between lightly cross-linked homodimers, versus singly-distributed homodimers.
Thank you for this excellent point. Yes, we are talking about "cluster of events" parameters in the evaluation, which means a cluster of acquired fluorescent events. The size of the "cluster of events" is not the same as the "receptor cluster size", which we are not defining (due to method limitations, as discussed above). Most of these "clusters of events" (CoE) correspond to NKR-P1 homodimers or the homodimers cross-linked by the soluble dimeric ligand. Both the text in the Results and Figure legends sections has been modified to clarify this misunderstanding. Moreover, we have edited and substituted "cluster" for "cluster of events" (CoE) and added an explanation to its legend to avoid any confusion.
## 1
## Reply to reviewers' comments
Reviewer #2 (Remarks to the Author):
I thank the authors for clarifying the way the SEC-SAXS data were pooled together before using the OLIGOMER algorithm. I also understand that deconvolution of the data would reveal impossible in their case. I would then support publication.
We thank the reviewer for supporting our publication.
## Reviewer #3 (remarks to the author):
The reasoning behind the response to review and the edited text help to clarify the conclusions and methodology and resolves the majority of my requests for clarification and any concerns regarding interpretation of results or experimental design.
The only additional detail that I would request be considered for additional prior to final publication are for more description (e.g., catalog number, name) regarding the type and concentration of the fiducials used to be added to the Methods section. And for a note in the methods section and/or document text to indicate whether edges were included in the 'whole cell' analysis, as inclusion of edge regions can cause there to be areas that are unable to measure clustering (no cell present) and apparent large clusters at cell edges where membrane is concentrated in the TIRF field. A comment regarding how edges were excluded from analysis would satisfy this concern.
We are glad that our responses met the reviewer´s remarks.
Concerning the fiducial markers, we used commercial FluoSpheres™ (product number F8801 from Thermo Scientific) markers. These 100 nm large spheres have an excitation maximum of 580 nm. Fiducial markers were diluted 1000× into the water and then diluted 100× into PBS buffer (the final dilution of bead stock concentration was 100000×) and topped on the sample. Beads were let to settle for at least one hour. Just before imaging, the buffer was exchanged for the oxygen-scavenging buffer, and the sample was sealed. The final bead dilution was assessed by optimization to have a reasonable number of fiducial markers surrounding the imaged cells within the field of view. The text in the Methods section was updated accordingly.
The reviewer raises an important point regarding the inclusion/exclusion of the apparent cell edges in the analysis that would result in obvious artificial data bias. Indeed, the cell's edges were excluded from the analysis performed using the ClusterViSu. ClusterViSu offers a freehand selection drawing tool to define the region of interest included for the clustering analysis. This tool was used to define the region of interest that included the whole inner area of the cell -i.e., the whole cell, excluding the concentrated membrane regions prominent in the TIRF field, the cell edges. Occasionally, apparent areas of the cell that did not adhere well to the surface of the supporting glass (non-planar regions of the cell membrane, resulting in the "edging" effect in the TIRF field) were excluded from the analyzed region of interest as well. However, it is important to state that no "picking and choosing" of the region of interest area has been done; the majority of the inner surface of the cells was included in the analysis. The text in the Methods section was updated accordingly.
One comment is regarding the experimental design, in which LLT1 and LLT1-SIM are introduced to the cells in a soluble form, which differs dramatically from the presentation of LLT1 on the surface of an activated monocyte, B cell, Nk or T cell. It would be relevant for the authors to address how this change in presentation of LLT1, and thus the biophysical constraints that would be placed upon LLT1 if it were attached to a cellular membrane, may affect it's binding to NKR-P1.Also, because dSTORM imaging can be very affected by the number of labels on the secondary
The authors present data demonstrating that NKR-P1 increases in average cluster size and events per cluster in the presence of soluble LLT1, and that this change in distribution requires the presence of the three predicted strongest contacts in the LLT1 secondary interaction interface. While the data is consistent with the other findings with complementary methods, the manuscript would be strengthened by further justification of several experimental and analysis design choices and the inclusion of other relevant controls to ensure consistent labeling of the proteins imaged.No further commentsReviewer #2 (Remarks to the Author):I thank the authors for clarifying the way the SEC-SAXS data were pooled together before using the OLIGOMER algorithm. I also understand that deconvolution of the data would reveal impossible in their case. I would then support publication.Reviewer #3 (Remarks to the Author):The reasoning behind the response to review and the edited text, help to clarify the conclusions and methodology and resolves the majority of my requests for clarification and any concerns regarding interpretation of results or experimental design.The only additional detail that I would request be considered for additional prior to final publication are for more description (e.g. catalog number, name) regarding the type and concentration of the fiducials used to be added to the Methods section. And for a note in the methods section and/or document text to indicate whether edges were included in the 'whole cell' analysis, as inclusion of edge regions can cause there to be areas that are unable to measure clustering (no cell present) and apparent large clusters at cell edges where membrane is concentrated in the TIRF field. A comment regarding how edges were exclude from analysis would satisfy this concern. |
Comparison between remifentanil and other opioids in adult critically ill patients
Background and aims: To identify the efficacy and safety of remifentanil when compared with other opioids in adult critically ill patients.Methods:We searched for studies in the Cochrane Library, MEDLINE, and EMBASE that had been published up to May 31st, 2019. Randomized clinical trials using remifentanil comparing with other opioids for analgesia were included. Two reviewers independently applied eligibility criteria, assessed quality, and extracted data. Duration of mechanical ventilation was the primary outcome, and secondary outcomes included weaning time, intensive care unit (ICU), length of stay (LOS), hospital LOS, mortality, side effects, and costs.Results: Fifteen studies with 1233 patients were included. Remifentanil was associated with a significant reduction in the duration of mechanical ventilation in the adult ICU patients when compared with other opioids (P = .01). Remifentanil also reduced the weaning time (P = .02) and the ICU LOS when compared with other opioids (P = .01). There was no difference in the hospital LOS (P = .15), side effects (P = .39), and mortality (P = .79) between remifentanil and other opioids, what's more, remifentanil increased the costs of anesthesia (P < .001) but did not increase cost of hospitalization (P = .30) when comparing with other opioids. Conclusions: Remifentanil reduced the duration of mechanical ventilation, weaning time, and ICU LOS when compared with other opioids in adult critically ill patients. Higher quality RCTs are necessary to prove our findings. PROSPERO registration number: CRD42016041438. Abbreviations: CI = confidence interval, GRADE = Grading of Recommendations, Assessment, Development, and Evaluation, ICU = intensive care unit, IV = inverse variance, LOS = length of stay, M-H = Mantel-Haenszel, RR = risk ratio, SMD = standard mean differences.
# Introduction
Pain is very common among patients admitted to an intensive care unit (ICU) whether at rest or during standard care procedures.Analgesia is often required for ICU patients to relieve pain, improve comfort, reduce stress, and facilitate procedures. Currently, opioids, such as morphine, fentanyl, sufentanil, and remifentanil, are commonly used for pain management in the ICU.However, the accumulation of opioids leads to numerous side effects, such as nausea, vomiting, ileus, hemodynamic instability, respiratory depression, and prolongs the duration of mechanical ventilation and ICU length of stay (LOS).Remifentanil, a 4-anilidopiperidine derivative of fentanyl, is an ultra-short-acting m-opioid receptor agonist. It has a rapid onset of action (1 minute) and a rapid offset of action following discontinuation (3-10 minutes) irrespective of the duration of infusion. Its property of organ-independent metabolism makes the pharmacokinetics of remifentanil unaffected by the renal and live dysfunction which is very common in critically ill patients.Therefore, we hypothesized that remifentanil could be a better analgesic than other opioids in critically ill patients.
The results of an early meta-analysis showed that remifentanil was associated with reduced weaning time, but not associated with a reduction in mechanical ventilation duration, ICU LOS, or mortality. Furthermore, in this meta-analysis, remifentanil was not only compared with other opioid, but also with sedative The author(s) received no financial support for the research, authorship, and/or publication of this article.
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. agent.Another recent meta-analysis found that remifentanil as compared with other opioids was associated with decreased duration of mechanical ventilation, time to extubation, and the length of ICU stay. However, the subjects of this meta-analysis were patients under the treatment of mechanical ventilation and included many post-surgery patients. In 5 of their included studies, remifentanil and other opioids were only used during the operation (not used in the ICU), and some of them were carried out in the anesthesia recovery room, not the ICU.Therefore, this meta-analysis may not represent the value of remifentanil in critically ill ICU patients. On the other hand, recent studies indicated that remifentanil as compared with other opioids was associated with a higher incidence of side effects, such as opioidinduced hyperalgesia.Therefore, we conducted this systematic review and metaanalysis to identify the efficacy and safety of remifentanil when compared with other opioids in adult critically ill patients.
## Study aim
We aim to assess the effects of remifentanil on the duration of mechanical ventilation, weaning time, ICU LOS, hospital LOS, cost, and mortality in critically ill patients; and aim to assess the side effects of remifentanil as well.
# Method
This study is a review and meta-analysis. So ethical approval is not necessary.
This systematic review is performed in accordance with the methods recommended in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guideline.We registered this systematic review and meta-analysis on PROSPERO international prospective register of systematic reviews (https:// www.crd.york.ac.uk/prospero/) on February 11, 2018, the registration number CRD42016041438.
## Search strategy
We searched 3 electronic databases including Cochrane Library, MEDLINE, and EMBASE databases for reports of studies that included databases from their inception to May 31th, 2019. Our search strategy used appropriate medical subject headings and keywords. The search strategy is in the appendix (see Supplement Content, which illustrates the search strategy, http://links.lww. com/MD2/A477). We also manually checked clinical trials.gov and the references of the relevant studies to identify other potentially trials or unpublished reports. Two reviewers (SY and HZ) completed the research of this systematic review independently. A consensus of all the authors was made to resolve the inconsistency of the literature review.
## Study selection
Eligible studies were those that matched the following criteria:
1. Type of study: randomized controlled trial; 2. Human study; 3. Population: adult patients (age ≥ 18 years) admitted to the ICU; 4. Intervention: remifentanil used for analgesia management; 5. Predefined outcomes: duration of mechanical ventilation, weaning time (from the beginning to the end of the mechanical ventilation weaning procedure), ICU LOS, hospital LOS, side effects (nausea/vomiting, hemodynamic instability, and delirium), mortality, and costs.
The study with the most recent publication date was included in the review if there were more than 1 eligible trial from 1 team with the same subjects.
Studies that met any of the following criteria were excluded from the analysis: a study not set in an ICU (remifentanil was not administrated in ICU); a study that remifentanil was compared with non-opioid drugs; a study was in neither English nor Chinese language.
The primary outcome was the duration of mechanical ventilation. The second outcomes were weaning time, ICU LOS, hospital LOS, side effects (nausea/vomiting, hemodynamic instability, and delirium), mortality, and costs.
## Data extraction
Two reviewers (SY and HZ) reviewed the titles, abstracts, and all full-text articles according to a standard data extraction form independently. We resolved disagreements through discussion with a third author (HW). The data extracted in the analysis were as follows: the study ID (combined the first author's name with publication year), country, selected population, size, site, intervention, and outcome. We also checked the additional files and contacted the authors for more details if necessary.
## Quality assessment and publication bias
Two reviewers independently explored the quality of selected RCTs using the Cochrane Collaboration Risk of Bias tool by RevMan 5.3 software.This tool considers 7 different domains: adequacy of sequence generation, allocation sequence concealment, blinding of participants and caregivers, blinding for outcome assessment, incomplete outcome data, selective outcome reporting, and the presence of other potential sources of bias not accounted for in the other 6 domains. Each domain was categorized as low, unclear, or high risk of bias. Two reviewers (SY and HZ) made judgments independently. In cases of disagreement, a resolution was first attempted by discussion and then by consulting a third author (HW) for arbitration.
We assessed the possibility of publication bias by using funnel plots which were implemented in RevMan 5.3 software (Cochrane Library, London, UK). We also used egger's regression test to measure funnel plot asymmetry using Stata 12.0 software (StataCorp LP, Texas).3.5. Grading the quality of the evidence We used the methodology of the Grades of Recommendation, Assessment, Development, and Evaluation (GRADE) Working Group to assess the quality of the evidence for all the outcomes by Stata 12.0 software. This tool included 4 domains: risk of bias, inconsistency, indirection imprecision, and publication bias. The quality of evidence was classified as very low, low, medium, and high.
# Statistical analysis
We used RevMan 5.3 software (Cochrane Library, London, UK) to perform the statistical analysis. We analyzed pooled effects by calculating risk ratio (RR) with 95% confidence intervals (95% CI) for dichotomous variables and standard mean differences (SMD) with 95% CI for continuous variables. Mantel-Haenszel (M-H) and inverse variance (IV) methods were applied for dichotomous variables and continuous variables, respectively. We assess the heterogeneity of the trials using the I 2 statistic as implemented in RevMan5.3 software. I 2 values of 25% to 50% indicated low, 50% to 75% indicated moderate, and >75% indicated high heterogeneity. Heterogeneity was significant when I 2 > 50% and P < .1. The analyses were performed using randomeffects models.The results were expressed using P values. A P values less than .05 was considered statistically significant.
# Results
## Study identification and selection
Our search strategy identified 13,897 relevant citations, while 13,892 from electronic selection and 5 from other references. We assessed 11,176 articles after the removal of duplicates. We screened titles and abstracts to identify potentially eligible studies and retrieved 41 manuscripts for full-text review. Twenty-six of those studies were excluded: 4 studies were not in English or Chinese,12 did not include the relevant outcomes for systematic review,were not set in the ICU,3 studies using remifentanil versus non-opioid for sedation,and 2 studies are not RCTs.In total, 15 studieswere eligible and included in this systemic review, which ultimately included 1233 subjects. A Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram of the selection of studies is shown in. We included 15 RCTs in this review. The other opioids compared with remifentanil included: 3 studies with sufentanil,10 studies with fentanyl,and 5 studies with morphine.The population of included studies consisted of surgical patients (including postcardiac surgeryand other surgical patientsand medical critically ill patients admitted to the ICU.Four studies were multi-center,and the other studies were single-center. Two studies were performed in multi-countries,1 in America,1 in Austria,2 in Denmark,2 in China,1 in Korea,3 in Germany,1 in Greece,1 in Italy,and 1 in Oman.
## Risk of bias assessment
For all RCTs included in this meta-analysis, most of the domains were evaluated as having low risk of bias (allocation sequence concealment, blinding for outcome assessment, incomplete outcome data, and selective outcome reporting of domains). The results of the study quality assessment were summarized in. Funnel plots were visually inspected and did not demonstrate evidence of publication bias in the duration of mechanical ventilation (see Supplement Content, which illustrates funnel plots of the duration of mechanical ventilation, http://links.lww. com/MD2/A474) and the ICU LOS (see Supplement Content, which illustrates funnel plots of the ICU LOS, http://links.lww. com/MD2/A473). The Egger regression test showed that the tests of asymmetry were not significant for main endpoints, including: the duration of mechanical ventilation (t = À1.56; 95% CI À8.81 to 1.41; P = .14); weaning time (t = À0.35; 95% CI À9.01 to 7.00; P = .745); ICU LOS (t = À1.75; 95% CI À9.71 to 1.06; P = .11); side effects: nausea/vomiting (t = 0.49; 95% CI À2.29 to 3.12; P = .657), hypotension (t = 1.69; 95% CI À2.41 to 5.54; P = .233), and dysrhythmia (t = À0.Sixteen studies focus on mechanical ventilation time; they all showed that remifentanil significantly reduced the duration of mechanical ventilation. Fourteen of them were included in the metaanalysis.One study was excluded because of the impossibility of data extraction.Pooled analysis of 13 studiesshowed that remifentanil was associated with a significant reduction in the duration of mechanical ventilation when compared with other opioids (SMD À0.23; 95% CI À0.41 to À0.06; P = .01; IV random; heterogeneity I 2 = 50%, P = .01).
## Second outcomes 4.2.2.1. weaning time.
Five studiesthat recruited 487 subjects observed weaning time. Remifentanil was associated with a significantly shorter time when compared with the other opioids (SMD À0.21; 95% CI À0.40 to À0.03; P = .02; IV random; heterogeneity I 2 = 11%, P = .34) (see Supplement Content, which illustrates remifentanil decrease the weaning time, http://links.lww.com/MD2/A479).
## Icu length of stay.
Fourteen studiesobserved the ICU LOS. Thirteen studiesthat recruited 1034 patients were included in the meta-analysis. Breen's study was excluded because of the impossibility of data extraction.Remifentanil was associated with a reduction in the ICU LOS when compared with the other opioids (SMD À0.33; 95% CI À0.60 to À0.07; P = .01; IV random; heterogeneity I 2 = 77%, P < .01).observed hemodynamic instability: hypotension (RR 1.85; 95% CI 0.87-3.92; P = .11; M-H random; heterogeneity I 2 = 0%, P = .41), and dysrhythmia (RR 1.43; 95% CI 0.55-3.73; P = .47; M-H random; heterogeneity I 2 = 0%, P = .71); 2 studies (132 patients)observed delirium (RR
## Mortality.
Three studiesobserved mortality. Pool analysis showed that there was no difference between remifentanil and other opioids (RR 0.92; 95% CI 0.51-1.66; P = .79; M-H random; heterogeneity I 2 = 0%, P = .82) (see Supplement Content, which illustrates no difference in the mortality, http://links.lww.com/MD2/A476).
## Costs.
Two studies observed the costs.Liu et al observed 60 patients after tumor operation showed that remifentanil increased the cost of ICU than fentanyl.Engoren et alcompared remifentanil with fentanyl and sufentanil for 90 patients undergoing cardiac surgery. They found that remifentanil increased the costs of anesthesia (P < .001) but did not increase costs of hospital (P = .
3) when comparing with other opioids.
## Grade quality evidence
The GRADE quality evidence was assessed with GRAD Epro software, and the results were as follows: for comparison of remifentanil with other opioids, the quality of evidence on the duration of mechanical ventilation was thought to be moderate; the quality of evidence on the duration of mechanical ventilation and weaning time were thought to be moderate; the ICU LOS, the incidence of side effects (nausea/vomiting, hypotension, dysrhythmia, and delirium), and mortality were thought to be low; the hospital LOS was thought to be very low. The main reason for these results was that the heterogeneity was high (I 2 > 50%). Other reasons for the demotion of the studies included: other types of bias (referred to as commercial interference), the lack of blinding, inconsistency, and imprecision.
# Discussion
The main finding of our systematic review and meta-analysis was that remifentanil significantly reduced the duration of mechanical ventilation when compared with other opioids in adult critically ill patients. We also found the following: remifentanil significantly reduced the weaning time and ICU LOS; there was no significant difference in the hospital LOS, side effects, and mortality between remifentanil and other opioids.
Remifentanil is a fentanyl relative u-receptor agonist, and it is mainly combined with a-1-acid glycoprotein, which reaches rapidly blood-brain balance in 1 minute resulting in rapid onset and offset. In addition, remifentanil is different from other fentanyl analogues because it is hydrolyzed by nonspecific esterase in plasma and tissues, which is independent of liver and kidney.Therefore, it is easy to explain why remifentanil could reduce the duration of mechanical ventilation, the weaning time, and the ICU LOS in comparison with other opioids in critically ill patients which usually have a prolonged use of opioid and with organ dysfunction. Our meta-analysis included both surgical patients and medical patients, which may well represent the adult critically ill patients. For another, remifentanil was all used in the ICU in our included RCTs. These were partially different from another meta-analysis: they mainly included short-term postsurgery patients; some of their included studies were carried out during the operation and in the anesthesia recovery room, not the ICU.Our meta-analysis found that there were no significant differences in hospital LOS and mortality between remifentanil and other opioids. However, we only included 3 to 4 studies for hospital LOS and mortality. These results were similar to the other 2 meta-analyses.Hospital LOS and mortality are affected by many reasons which mainly may be the severity of diseases rather than the selection of analgesic therapy.
We also found that there was no significant difference in side effects between remifentanil and other opioids. The results were similar in the subgroup analyses of nauseous/vomiting, hypotension, dysrhythmia, and delirium. But these results were all assessed in small samples. And all the included studies did not observe the opioid-induced hyperalgesia which is the most unique side effect of remifentanil when compared with other opioids. Opioid-induced hyperalgesia has been well illustrated in the postoperative patientsand also should be paid more attention in the future studies of remifentanil in critically ill patients.
In our review, only 2 studies observed the costs. One found remifentanil increase cost of anesthesia while did not increase cost Yang et al.100:Medicine of hospital,another found that remifentanil increase cost of ICU.Remifentanil has unique pharmacokinetics and pharmacodynamics profiles but is expensive when compared to other traditional opioids such as morphine and fentanyl. When we apply this analgesic, we must consider its cost, especially in the anesthesia procedure and ICU. Many studiesshowed that remifentanil decrease mechanical ventilation time and hospital LOS. Further studies are needed to explore the costeffectiveness of different analgesics. There were several limitations in our meta-analysis:
1) the studies of this meta-analysis were small size, but the quality of each study was high; 2) the population of included studies was heterogeneous, and consisted of surgical and medical patients; 3) the type of the other opioids for the included studies were heterogeneous, which included morphine, fentanyl, and sufentanil; 4) the heterogeneities were high in some of the analyses, such as ICU LOS, and hospital LOS.
However, we analyzed the outcomes in subgroups classified by surgical and medical patients to reduce clinical heterogeneity. We also selected a random-effects model rather than a fixed-effects model to address the observed heterogeneity. Therefore, large, well-designed randomized controlled trials are necessary for the future.
# Conclusions
In conclusion, remifentanil reduced the duration of mechanical ventilation, weaning time, and ICU LOS when compared with other opioids in adult critically ill patients. Large-scale random-ized controlled trials are necessary to confirm our findings and to further evaluate the safety and cost of remifentanil in critically ill patients.
# Author contributions
Shuguang Yang helped conceive the idea for the meta-analysis, search the literature, collect the data, perform the statistical analysis, create the figures/tables, and draft the manuscript. Huiying Zhao helped conceive the idea for the meta-analysis, design the study, participate in the literature search, collect the data, interpret the results, and draft the manuscript. Huixia Wang helped search the literature, analyze the data, create the figures, and revise the manuscript. Hua Zhang helped perform the statistical analysis and revise the manuscript. Youzhong An helped design the study and revise the statistical analysis and the manuscript. All authors read and approved the final manuscript for publication. Quality of evidence of the studies that compared remifentanil with other opioids that were included in the meta-analysis, according to Grades of Recommendation, Assessment, Development, and Evaluation (GRADE). |
Association of Optimal Blood Pressure With Critical Cardiorenal Events and Mortality in High-Risk and Low-Risk Patients Treated With Antihypertension Medications
## Emethods 1. national health information database (nhid)
The NHID, a public database for the whole population of South Korea, was established and is being maintained by the National Health Insurance Service (NHIS). Details on the NHID are published elsewhere.The NHID covers data from 2002 onwards, and comprises the following. (a) eligibility data: demographics, income based insurance contributions, and date of death (b) health screening records: medical history, health behavior, physical exam, and laboratory exam (c) reimbursement records of the NHIS: prescribed drugs, medical procedures, outpatient visits, hospitalizations, and lists of medical diagnosis health care provider data: medical institutions, equipment, and human resources
The data in each resource was assembled using de-identified join keys, which replace personal identification numbers assigned to citizens of Korea.
Nationwide health screenings are performed for citizens aged ≥40 years, generally at 2-year intervals, in hospitals or medical centers. During the health screening, information on medical history and health behaviors are obtained using standardized questionnaires. Trained medical staff perform physical examinations including blood pressure (BP) measurement. Blood and urine samples are obtained after at least an 8-h fast.
Serum creatinine and high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterols have been measured from 2009 health screening.
# Emethods 3. covariates
Outlier data were excluded from the health screening records. (a) systolic BP <90 mm Hg or >200 mm Hg (b) diastolic BP <30 mm Hg or >140 mmHg (c) blood glucose <30 mg/dl or >900 mg/dl total cholesterol <130 mg/dl or >320 mg/dl (e) HDL cholesterol <20 mg/dl or >100 mg/dl (f) serum creatinine <0.3 mg/dl or >15.0 mg/dl (g) body mass index <10 or >50 Treated and untreated BP records were collected separately. (a) if antihypertensive prescription ≥90 days in the year of BP measurement, the BP → treated BP (b) if not, the BP → untreated BP Income levels were determined by income based insurance contributions. The amounts of alcohol consumption were calculated as the number of drinks averaged per day.
In each year of follow-up, a new average was calculated for each time period as follows. (a) primary cohort: secondary cohort: Using the yearly updated values, the variables were categorized as follows. In each year of follow-up, the status of health conditions were determined. (a) diabetes mellitus: yes or no (b) hyperlipidemia: yes or no (c) proteinuria: yes or no (d) smoking: never, former, or active smoker (e) antihypertensive compliance: regular, irregular, or never use Antihypertensive compliance was determined as follows. Initiation of antihypertensive treatment was defined as the first year when antihypertensive drugs were prescribed for ≥90 days per year. Among cases that initiated treatment, regular use was considered when antihypertensives were prescribed for >half of each follow-up period (from initiation of treatment to each year of follow-up), irregular use was when the prescription was ≤half of each follow-up period, and never use when the prescription was never for ≥90 days per year.
Using baseline data, the variables were categorized as follows.
## Emethods 4. risk categories
Both cohort participants were grouped into three risk categories by the number of risk factors present at baseline (≥3, 2, or ≤1 risk factors). The primary cohort participants were additionally categorized by using the SCORE system (≥7.5%, 2.5-7.4%, or <2.5%) and the WHO/ISH score (≥20%, 10-19%, or <10%). The secondary cohort participants were also categorized by using a Korean prediction model (≥15%, 7.5-14%, or <7.5%) and the Framingham score (≥30%, 15-29%, or <15%).
## A. determination of risk factors in both cohorts
Five risk factors (hypertension, diabetes mellitus, hyperlipidemia, proteinuria, and active smoking) were identified, using the results of health screenings and information on the prescription of drugs. Information on the prescription of drugs in the reimbursement records were captured using NHIS billing codes (eTable 2).
B. In the primary cohort, the 10-year risk of cardiovascular disease was calculated with following variables. CHD_S0_age = exp(-(exp(α_CHD)) * ((Age -20)**p_CHD)) CHD_S0_age_10 = exp(-(exp(α_CHD)) * ((Age -10)**p_CHD)) CHD_ω = (βchol_CHD * (Total-C * 0.0259 -6)) + (βsbp_CHD * (SBP -120)) + (βcig_CHD * Cig) CHD_S_age = CHD_S0_age**exp(CHD_ω) CHD_S_age_10 = CHD_S0_age_10**exp(CHD_ω) CHD_S10_age = (CHD_S_age_10 / CHD_S_age) Risk10_CHD = (1 -CHD_S10_age) nonCHD_S0_age = exp(-(exp(α_nonCHD)) * ((Age-20)**p_nonCHD)) nonCHD_S0_age_10 = exp(-(exp(α_nonCHD)) * ((Age-10)**p_nonCHD)) nonCHD_ω = (βchol_nonCHD * (Total-C * 0.0259 -6)) + (βsbp_nonCHD * (SBP -120)) + (βcig_nonCHD * Cig) nonCHD_S_age = nonCHD_S0_age**exp(nonCHD_ω) nonCHD_S_age_10 = nonCHD_S0_age_10**exp(nonCHD_ω) nonCHD_S10_age = (nonCHD_S_age_10 / nonCHD_S_age) Risk10_nonCHD = (1 -nonCHD_S10_age)
In participants without diabetes, SCORE risk score = 100 * (Risk10_CHD + Risk10_nonCHD) In males with diabetes, SCORE risk score = 200 * (Risk10_CHD + Risk10_nonCHD) In females with diabetes, SCORE risk score = 400 * (Risk10_CHD + Risk10_nonCHD)
## B-2. calculation of who/ish risk scores in the primary cohort
The scores were calculated using a single comma delimited file extracted from WHO/ISH cardiovascular risk assessment charts. 5 C. In the secondary cohort, the 10-year risk of cardiovascular disease was calculated with following variables.
## C-1. calculation of local risk scores in the secondary cohort
The scores were calculated according to a Korean prediction model,which was developed on the basis of 2013 ACC/AHA risk score.
In males, KRiskFactor = (ln(Age) * 9.362) + (((ln(Age))**2) * 2.425) + (ln(Total-C) * 6.409) -(ln(Age) * ln(Total-C) * 1. In females, Framingham risk score = 100 * (1 -0.95012 ** exp(RiskFactor))
# Emethods 5. outcomes
Previous studies using the disease codes listed in NHIS reimbursement records have reported that 73%~93% of the disease codes for myocardial infarction or stroke are valid.In addition to the disease codes, I used information on revascularization procedures, prescribed peritoneal dialysates, hemodialysis, and kidney transplantation, which were captured by NHIS billing codes, to identify outcomes more reliably.
## A. critical cardiorenal event
Critical cardiorenal event was identified through December 31, 2015 in the primary cohort and through December 31, 2107 in the secondary cohort. (a) Information on critical care unit admission were captured using NHIS code for admission (NHIS clause code, 02) to critical care unit (NHIS item code, 03). Critical care unit admission from cardiorenal diseases (ICD-10, I00-I99 and N18) was verified with the primary medical diagnosis listed in reimbursement records. (b) Information on revascularization procedures of coronary, cerebral, and carotid arteries were captured using NHIS billing codes (eTable 3). Revascularization for myocardial infarction (ICD-10, I21-I22) or stroke (ICD-10, I63-I64) was verified with the primary diagnosis listed in reimbursement records. (c) Information on hemodialysis (NHIS billing code, O7020 and O9991), prescribed peritoneal dialysates (eTable 3), and kidney transplantation (NHIS billing code, R3280; and ICD-10, Z94.0) were captured from NHIS reimbursement records. Using the information, end-stage kidney disease with dialysis for ≥90 days per year or kidney transplantation was verified.
## B. all-cause death
All-cause death was confirmed through December 31, 2015 in the primary cohort and through December 31, 2017 in the secondary cohort, using information on date of death, which was included in eligibility database of the NHID.
## Efigure 1. flow charts of participant selection in the primary (a) and secondary (b) cohorts efigure 5. yearly event rates in prevalent (a and c) or recent (b and d) antihypertensive users
The 1-year rates were estimated by multiplying the combined hazard ratios by the combined mean of the age specific rates in the reference group (systolic BP, 120-129 mm Hg). The summary effects and 95% CIs of the primary and secondary cohorts were calculated by using the DerSimonian-Laird random-effects model. All analyses were adjusted for age, sex, family history of cardiovascular disease, income level, smoking, alcohol consumption, exercise frequency, body mass index, diabetes, hyperlipidemia, and proteinuria. The critical cardiorenal event was a composite of admission to critical care unit with cardiovascular or chronic kidney disease, revascularization for myocardial infarction or stroke, and new onset end-stage kidney disease. Error bars indicate 95% CIs. BP, blood pressure.
## Efigure 7. yearly event rates in risk categories grouped by risk factors after exclusion of proteinuria
The risk categories were grouped by the number of risk factors present at baseline after exclusion of proteinuria: i.e., ≥3, 2, ≤1 of the four risk factors (hypertension, diabetes, hyperlipidemia, and smoking). Among a total of 487,412 primary cohort participants, 34,050 (7.0%), 110,023 (22.6%), and 343,339 (70.4%) had ≥3, 2, and ≤1 risk factors, respectively. Among a total of 915,563 secondary cohort participants, 65,631 (7.2%), 188,669 (20.6%), and 661,263 (72.2%) had ≥3, 2, and ≤1 risk factors, respectively. The 1-year rates were estimated in the primary (A and C) and secondary (B and D) cohorts, by multiplying the hazard ratios by the mean of the age specific rates in the reference group (systolic BP, 120-129). The analyses were adjusted for age, sex, family history of cardiovascular disease, income level, smoking, alcohol consumption, exercise frequency, body mass index, diabetes, hyperlipidemia, and proteinuria, and further adjusted for antihypertensive compliance (regular use, irregular use, and nonuse). The critical cardiorenal event was a composite of admission to critical care unit with cardiovascular or chronic kidney disease, revascularization for myocardial infarction or stroke, and new onset end-stage kidney disease. Error bars indicate 95% CIs. BP, blood pressure.
## Etable 1. time-lagged covariates for subsequent years
|
Registered health problems and demographic profile of integrated textile factory workers in Ethiopia: a cross-sectional study
Background: Textile and garment factories are growing in low and middle-income countries as worldwide demand for inexpensive clothing increases each year. These integrated textile and garment production factories are often built-in areas with few workplaces and environmental regulations, and employees can be regularly exposed to workplace hazards with little regulatory oversight. Consequently, workers' health may be significantly affected due to long term exposure to hazards. This study describes registered health problems and their association to work-related and personal factors among workers in integrated textile factories in Ethiopia.Methods: Institution-based cross-sectional study design was employed for this analysis. A one-year recording of worker's clinical diagnoses (between March 2016 and February 2017) was gathered from the factory clinics of three integrated textile factories. Clinical diagnosis data was obtained as factory workers visited the clinics if feeling unwell. Sociodemographic characteristics and work-related information were obtained from the factory's human resource departments. The sociodemographic and clinical diagnosis statuses of 7992 workers were analyzed. The association between the registered diagnoses and workplace factors (work in textile production, garment production and support process) and personal factors (age, sex and educational status) were studied using logistic regression analysis.Results: The average employee age and years of service were 40 years and 11 years respectively. 60% of workers were females, comprising of 4778 women. 66% of all workers (5276) had 27,320 clinical diagnoses. In total, this caused 16,993 absent working days due to sick leave. Respiratory diseases (34%) and musculoskeletal disorders (29%) were the most prevalent diagnoses, while bodily injuries were the cause of most work absences. Work department, sex and educational status are variables that were most significantly associated with higher prevalence of disease groups.
Conclusions: About two-thirds of the integrated textile factory workers were diagnosed with different types of disease. The textile and garment production department workers were affected at a greater rate than the support process workers, indicating that some diseases may be related to workplace exposure. Further study should investigate rare chronic diseases such as cancer, heart diseases, renal diseases and diabetes.
Keywords: Integrated textile, Respiratory disease, MSD, Injury, Clinical diagnosis, Work-related diseases, Ethiopia
# Background
The textile and garment sector is growing in low and middle-income countries (LMICs); currently, Sub-Saharan African countries seeking to industrialize are expanding textile production capabilities. In Ethiopia, the integrated textile factories quoted as "Farm to Fashion" production, which comprise both textile production and garment processing have a higher priority in the textile and clothing value chain than stand-alone textile or garment factories. Integrated factories have comparative advantages of creating extensive employment opportunities, with some factories recruiting around 6000 workers per factory. Furthermore, the departments are arranged in a chain to produce clothes from raw cotton; the fabrics produced at the weaving/ knitting department use the yarn produced from the spinning department that processes raw cotton; then the final clothes made in the garment department use the fabrics produced from the finishing department. The health risks present in the integrated processes may be different from separate stand-alone factories.
Traditionally, the textile industry has been known to cause respiratory diseases. Since the early 1900s, the main research regarding work and health in the textile and garment industry focused on the respiratory disease, byssinosis. Several studies have indicated that processing raw cotton for clothing generates inhalable dust and endotoxins that might cause respiratory problems. Respiratory issues are not the only hazard that textile and garment production employees are exposed to. Heavy machines and mechanical contact, manual labour, repetitive work, awkward working postures and increased pressure to produce can put workers at risk for work-related diseases, such as musculoskeletal disorders (MSDs) and traumatic injuries. Some studies reported various health problems among workers in the textile and garment sector. Nevertheless, several of these studies are reviews, and few of them are original researches that study health problems often assessed by the self-report.
As aforementioned, work in integrated textile factories may be associated with different health risks than those seen in factories with the sole purpose of fabric production. The textile department workers directly interact with raw cotton and dangerous machines to produce yarn and fabrics, implying an increased risk for respiratory diseases, injuries, and hearing impairment. Workers in the garment department process the fabrics; change the sizes, colours, and textures in various working conditions using different machines. This process increases the risk of MSDs attributive to repetitive work, unhealthy postureand cancer risk from the dyes. The establishment of integrated textile factories has been rising in Ethiopia in the past years; yet there are no national statistics concerning occupational diseases and injuries to show the country's workers' health status.
Despite the prominent presence of health hazards in the textile industry, Occupational Health Services (OHS) is lacking in Ethiopia. According to ILO, only 5-10% of workers in developing countries have access to OHS. Nevertheless, several of the integrated textile factories in Ethiopia do have health clinics. These clinics are likely to diagnose work-related diseases but we have very little knowledge about their activities. Some factories have health clinics with registration books where information about workers' health is available. The occupational disease study in the integrated textile factory helps to see the overall picture of health risks and associated impacts.
This study aims to describe the magnitude of registered health problems and the demographic profile of workers in the integrated textile factories from factory clinics during 1 year. Personal and workplace factors associated with the diseases were also identified and examined. Since workers visit the health clinics if they feel unwell or are injured, factory clinics can be a valuable source of information and knowledge about workrelated diseases and injuries in the integrated textile factories of Low and Middle-Income Countries (LMICs). Using the ILO lists of occupational diseases, respiratory diseases, MSD, injuries and ear diseases/impaired hearing was the main focus of discussion in this paper.
# Methods
## Study design
A 1-year institution-based cross-sectional study was employed that collected workers' health data from clinics in three integrated textile factories. In Ethiopia, integrated textile and garment factories aim to organize production by including the whole production line from raw cotton to clothing. The detailed production process in the integrated textile factories has been described in a previous study.
## Study settings
Three integrated textile factories participated in this study which fulfilled the inclusion criteria. One hundred thirty registered enterprises are active in the Ethiopian textile value chain; 20 grouped under integrated textile factories. The selection criteria for the integrated textile factories for this study include the presence of four functional integrated production departments (spinning, weaving/knitting, finishing and garment), working in three shifts, the presence of health clinics for both emergency and non-emergency consultations and the availability of a health recording system for workers. Three integrated textile factories fulfilled the above inclusion criteria; Factory 1 and 2 are located in Amhara Regional State, both within 600 km from the capital Addis Ababa. Factory 3 is in Tigray Regional State 1300 km North of Addis Ababa. These three factories 1, 2 and 3, were established in 1961, 1986 and 1992, respectively and had clinics from inception.
Each clinic, located within the factory compound provides similar services, including an outpatient department, emergency admission, laboratory facility and drug dispensary. Factory 3 clinic has diagnosed more workers than factory 1 and factory 2 clinics due to the high number of workers employed in factory 3. Each clinic has a general physician, 2-3 health officers, 5-7 nurses, 2-4 health assistants, 2-3 laboratory technicians, and 1-3 pharmacists. The clinics are primary healthcare centres; hence they have a referral system to hospitals and specialized clinics for advanced diagnosis and treatment. Since the factories function in three shifts, the clinics are open 24 h for consultations, but emergency service is available only for night shift workers. Overall, nonemergency cases can visit the clinic in the daytime outside of normal working hours; however, workers can visit the clinic for emergency health conditions while at work with approval from a supervisor.
## Study population
The population study comprised all workers in the three integrated textile factories; these are workers in factories 1, 2 and 3 with 1545, 1380, and 5067 employees respectively. Workers were broadly categorized into three groups: textile production department, garment production department and support department. The textile department included workers in spinning, weaving/knitting and finishing; the garment department included workers in garment production, and the support department included workers from administrative and maintenance services. The main work activities in the textile, garment and support are yarn/fabric making, cloth production and technical/administrative roles, respectively. Therefore, workers in the various department may have different exposure risks. Workers in the textile department are exposed to dust, noise, and dangerous machines, those in the garment department workers are predominately exposed to ergonomic hazards and chemicals, whereas the support department workers are exposed to mechanical and office related hazards. Furthermore, support department workers occasionally enter the production department for machine or production-related support issues that may expose them to hazards in the production sections; otherwise, they work regularly in the office or their workshops.
## Data collection
Data was collected from the clinics and human resource offices of the factories. Each worker's personal and working department profile is available from the human resource registration in a Microsoft Excel database that contains a list with a unique identification number, date of birth, sex, education, working department, and number of years employed. Workers' Medical information was obtained from the factory clinics. Each employee has a patient card labelled with their name and a unique identification number, which is the same as the one used in the human resource database. The registration on the patient card includes information about the date of consultation, type of diagnoses and number of sick leave days absent for each worker. During sick leave, a worker may be admitted to the clinic or given medication to administer at home with paid regular wage. Depending on the medical procedure, the worker may also share the medical cost. All clinical consultations of workers from 1st March 2016 to 28th February 2017 were extracted from the health archives of the factory clinics and registered manually in a logbook prepared for this research purpose.
In this study, no specific diagnostic code system was used. The clinic physicians used many diagnosis types, so the diagnoses were grouped into the following comprehensive categories: respiratory diseases, injuries, musculoskeletal disorders, allergy, ear diseases, eye diseases, gastrointestinal infections, mouth diseases, peptic ulcer diseases, reproductive health problems, skin diseases, neurologic and psychiatric diseases, and other health conditions. These categories have various groups of health conditions and were described as follows. Respiratory diseases included bronchitis, asthma, pneumonia, pulmonary tuberculosis, and upper respiratory tract infections. Injuries included fractures, cuts, dislocations, burns, swellings, soft tissue injuries, and lacerations. Musculoskeletal disorders included back pain, neck stiffness, disc prolapse, joint pain, leg pain, myalgia and arthritis. Allergies included allergic rhinitis, allergic conjunctivitis, allergic sinusitis, allergic reaction, skin allergy and food allergy. Ear diseases included otitis media, ear infection, vertigo, ear pain and hearing problems. Eye diseases included conjunctivitis, trachoma, chalazion, presbyopia, pterygium, blepharitis, blurred vision, short sight, glaucoma. Gastrointestinal infections included intestinal parasites and dysentery. Mouth diseases included dental caries, tonsillitis, oral candidiasis, tooth bleeding, glossitis and gum infection. Peptic ulcer diseases included gastritis, epigastric burn, hernia, dyspepsia and gastrointestinal disorder. Reproductive health problems included mastitis, pelvic inflammatory disease, dysfunctional uterine bleeding, abortion, genital ulcer, cyst, breast tumour, fistula, scrotal swelling, dysmenorrhea, cervical cancer, sexually transmitted infection and vaginal bleeding. Skin diseases included dermatitis, herpes zoster, herpes simplex, wart, skin rash, scabies, cellulitis, tinea corporis, boils, melasma, contact dermatitis and tinea capitis. Neurologic and psychiatric diseases included migraine, neuralgia, nerve problems, peripheral neuropathy, sciatica, anxiety, depression, mental disturbance, and psychosis. Other diseases included cancer, cardiac, kidney, goitre, chronic liver disease, chronic osteomyelitis, rectal prolapse, appendicitis, tumour, bone problem and insomnia-some disease-specific diagnoses used for diabetics, hypertension, anemia, hemorrhoids, and urinary tract infection.
Clinic consultation for antenatal services, chronic disease follow-ups and visits to change the treatment regime were excluded from the study. From 31,512 clinic consultations, 4192 were excluded due to incomplete information. A worker may visit the clinic for a new diagnosis or a previously known health problem requiring medical treatment. A worker may visit the clinic for consultations more than once for the same diagnosis category at a different time or for a different diagnosis. Thus, a worker's clinical investigation of a new disease or further diagnosis for medical treatment to a previously known health condition within 1 year is counted. Four nurses based in each clinic (12 nurses) participated in the data collection with 2 days of training for this work by the principal investigator. The data extraction is checked for consistency and completeness by the first author, and two clerks entered the clinical diagnosis data from the logbook into the Microsoft Excel spreadsheet.
## Measurement of variables
During analysis, the diagnostic categories are the outcome/dependent variables, while workplace and personal factors are the independent variables. The variable work department is one of the workplace factors categorized into three groups; textile, garment and support. The work department (textile production, garment production and support process) that represents and describes the peculiar nature of work-related circumstances found in each department, the machines, raw materials, work process, product, and the physical and psychosocial work environment. The textile and garment department workers are directly engaged in the production, and are more exposed to work environment hazards than the support department; hence the support department is a reference group during analysis. Similarly, the personal factors, education is grouped into three categories: able to read and write, completed grade 1-10th, and those at the college level and above. Further, age is treated as a continuous variable, while male, higher education group and factory 3 are reference categories.
# Statistical analysis
Clinical and demographic data sources were merged in an Excel database, then transferred to SPSS version 22 (SPSS, Chicago, IL, USA) for cleaning and analysis.
Two unique identification variables were created in the database; person specific unique identification and disease specific unique identification. The person specific unique identification was given for each study participant, and can be repeated in the database if the person has been diagnosed more than one time. Accordingly, 7992-person specific unique identifications with 27,320 records were produced. This unique identification helped to analyze and describe the independent variables of personal and workplace factors. Disease specific unique identification variables were also produced by combining the person specific unique identification variable with the specific disease code. This helped to calculate the specific disease prevalence and the number of repeated diagnosis for a specific disease.
Descriptive statistics were used, with an arithmetic mean utilized for the continuous variables such as age, years of service at work and number of days absent. Frequency and percentage were also used to describe categorical variables such as sex, education, health status and work force in each factory. Each of the independent variable mentioned above was stratified and presented by the three working departments (Textile, Garment and Support). Also, the percentage of each dependent variable, the 18 disease groups, was stratified by working department and presented in bar graph. Furthermore, the prevalence, number and of repeated diagnosis, and number of sickness absence days for each disease outcome were calculated and presented in.
A disease diagnosis at least one time per worker was used in the logistic regression analysis. Univariate logistic regression analysis was performed for each disease outcome with an independent variable to determine whether to include the variable in the model. The selection of the variable is based on a P-value < 0.2. A multicollinearity diagnostic test was performed among the independent variables; worker's age and service years were significantly correlated (r = 0.8, p < 0.001); consequently, we decided not to add the variable service years in the logistic regression analyses. Finally, the multivariable binary logistic regressions analysis was employed to identify any of the workplace and personal factors which had significant association with each disease outcome based on AOR with 95% CI and P-value < 0.05. This logistic regression analysis procedure was repeated for each disease outcome in.
# Results
## Demographic and work characteristics
A total of 7992 workers were included in the analysis. The average and (standard deviation) age of workers was 40years with an age range of 18-69 years. The proportion of workers in the textile, garment and support departments was 40, 44 and 16%, respectively. 60% of the workers were females; the proportions of female workers being 87, 40, and 33% among the garment, textile, and support departments respectively. 52% of workers had completed education level grade 1-10, while 7 % of workers had no formal education but could read and write. The highest proportion of uneducated workers was among the support department (13%) and lowest among the garment department (5%). About (66%) of workers had at least one diagnosis during the 1 year of the observation period.
## Prevalence of registered diseases
A total of 27,320 consultations took place with a total of 5276 diagnosed workers, equivalent to five consultations per worker during the study period. The highest proportion of total consultations was due to respiratory diseases (17%). Further, the highest prevalence of diseases diagnosed at least one time per individual worker were respiratory diseases (34%), followed by MSD (29%), gastrointestinal infection (21%), peptic ulcer disease (19%) and injury (17%). A total of 16,993 workdays were registered as sickness absence in 1 year period due to workers' health problems. Injury was the highest cause of sick leave days, (2951) (17%), followed by respiratory diseases (2327) (14%). The number of workdays absence by departments was highest among textile department workers, (9027) (53%) followed by garment workers (6415) (38%) and support workers (1481) (9%). In line with this, the proportion of workdays' absence per number of workers was 2.8, 1.8 and 1.4 in the textile, garment and support respectively.
Disease prevalence across the departments was varied. The prevalence of diagnosis at least for one disease was 69, 65 and 60% among the textile, garment and support department workers respectively. The proportion of textile department workers accounted for 44% of the total workforce; however, the textile department workers' overall proportion of disease diagnoses is about 49%.
In terms of disease type, a higher percent of respiratory, MSDs, injuries, peptic ulcers, AFI, mouth diseases, skin diseases, eye diseases, allergy, hemorrhoids, and hypertension were identified among the textile department workers. Gastrointestinal infections, neurological and psychiatric, urinary tract infection, reproductive illnesses and anemia were the highest reported by garment department workers. Respiratory disease was the most prevalent across the three working departments, with 37, 32 and 31% among textile, support and garment departments, respectively. From the total workers who had at least one respiratory disease, 626 (23%), 135(5%), 84 (3%) and 27 (1%) were diagnosed with bronchitis, asthma, pneumonia and tuberculosis, respectively. MSD was the second most prevalent disorder in all three departments, 31, 28, and 25% among textile, garment, and support respectively. Injury was the third most prevalent issue among the textile department workers, but is the eighth among the garment department workers.
## Factors associated with the diagnostic category
The statistical analysis uses diseases that were diagnosed at least once for individual workers as a unit of analysis. In the analysis support department, male, education status higher than "read and write" and factory 3 was the reference category. The multivariate logistic regression analysis found that working department (Textile and Garment), sex (female worker), educational status (read and write), age (older worker) and factory type (Factory 1&2) are associated with a higher prevalence of diseases compared to the reference category. Textile department workers have significantly higher odds for eight disease groups with adjusted odds ratio ranges (AOR: 1.22-1.79) compared to the support department workers. The garment department workers had significantly higher odds for five disease groups (AOR: 1.30-1.67) compared to the support department workers. Female workers had significantly higher odds for seven disease groups (AOR: 1.11-14.76) compared to male workers.
Furthermore, workers with low educational status had significantly higher odds for four disease groups (AOR: 1.36-1.52) than workers of higher education level. Age of workers was significantly associated with 14 of the.
# Discussion
This study shows that workers in the integrated textile factories were diagnosed with a wide range of diseases in 1 year. Respiratory disease was the leading cause of morbidity followed by MSDs, whereas workplace injuries caused the most days away from work. Working in the textile departments, being female, older age and low educational status are associated with higher risk for most disease groups.
## The size of the problem
The overall disease prevalence in 1 year time is 66%. The majority of the workers were diagnosed at least for one disease in the study period but, some workers had more than one disease diagnosis, which made the total number of consultations for diagnosis 27,320. These figures are higher than reports from other countries. For instance, a cross-sectional study that evaluates the health conditions using clinical examinations of 514 male Indian textile workers found 754 disease conditions, or 1.5 per worker. Another retrospective study from medical records of 1906 workers from mobile clinics in Bangladesh textile and garment reported that 25% of the workers were diagnosed with at least one disease condition. A short survey that examined the occupational health conditions of 845 Indian textile workers found that 447 workers suffered from different diseases. There are several limitations in this comparison with the above studies: difference in the observation period, the difference in diagnosis standard and difference in the study population. Moreover, in the present study, the proportion of total diagnosed cases from the number of all workers in the factories is 3.4, higher than the proportion of the total number of cases diagnosed from the general population 0.50 in Ethiopia. The total number of cases diagnosed in the general population excluding children less than 5 years of age was 48.8 million, given that the general population's count of Ethiopia 98.6 million. According to the Ministry of Health annual morbidity statistics report, the annual rate of outpatient visits for a new and repeated health condition is 0.9, which is about four times less than our study population. The result may indicate that workers from the integrated textile factories were diagnosed with more diseases than the general population; however, the high prevalence rate of diagnosis might also be related to free access to health services in factories and other demographic-related differences.
## Workplace factors
Workers in the textile department had a higher prevalence than other employees for many diseases, primarily respiratory diseases. Several other studies from low and middle-income countries have also shown a high prevalence of respiratory problems among textile department workers. An exposure assessment study by Yifokire and colleagueshas measured dust exposure levels among the textile production workers and the garment department workers in Ethiopia to be higher than the recommended threshold limit value in the ACGIH. This association might be due to the relationship between respiratory diagnoses and high dust levels in the integrated factories. However, the present study cannot answer this question due to the mixture of diagnoses in the categories used and the lack of exposure measurements in these particular factories. In this study, multiple respiratory diseases were described as bronchitis and asthma; these diagnoses might link with dust exposure.
MSD diagnoses are the second prevalent disease group in the present study and are significantly associated with the manufacturing departments. Both textile and garment department workers have higher odds of MSDs compared to the support department workers. Other studies have highlighted ergonomic hazards in the textile and garment department that could increase the risk of MSDs. Additionally, ergonomic hazard exposure assessment studies in Bangladesh and Cambodia found that the tasks in garment production gave a high risk of MSDs. The MSDs may be linked to exposure to ergonomic hazards in the textile and garment departments.
Furthermore, workplace injuries are among the most reported health problems and the primary cause of absence in this study. Both textile and garment department workers handle heavy machinery and have a higher risk of injuries than workers in the support department. This implies that the injuries might be related to the working conditions in the textile and garment departments. However, the prevalence of injuries in this study is lower than in a study of self-reported injuries in another Ethiopian textile factory. The difference may be associated with several factors; one potential reason being that minor injuries managed by first aid may not be included in the diagnostic reports of the factories' health services.
Moreover, some literature indicates that textile factories have high noise levels in their production departments; consequently, one can expect workers in this department to have a high prevalence of ear problems, as the prolonged noise may cause reduced hearing.
However, our study did not show any difference among the textile and support workers regarding ear diseases. However, the diagnosis might not detect reduced hearing among workers in this study, as health offices did not have the equipment to measure hearing ability in the targeted factories. Workers who develop hearing problems may move from the textile department to the support department to reduce their noise exposure or leave the job. Several studies indicated the possibility of self-selection of workers for the job or migration to different departments to mitigate disease. Future studies should consider exposure intensity and interruption by tracing the detail of worker's exposure profiles.
## Personal factors
Sex is significantly associated with most diseases registered in this study; female workers were diagnosed with disease at a higher rate than males. A qualitative indepth interviewing and focus group discussion with 24 female workers from Bangladesh indicated that female workers suffered from several types of diseases in garment factories. The morbidity assessment study by Singh and colleaguesalso revealed that female workers in the textile section had more severe anemia than males, similar to the finding of the present study. This might be related to the monthly menstrual cycle among females. Furthermore, a study in Bangladesh has reported a higher prevalence of different diseases among female workers than male workers, but with lower prevalence of injuries. Similarly, the current study shows a lower prevalence of injuries among females; possibly due to differences in tasks, with men often delegated to working with machinery, expose them to a higher risk of injury. Increased morbidity due to MSDs and respiratory diseases were also reported among female textile workers in India. Likewise, a result from the current study found that females are at higher risk of MSDs than males, but found no difference in the likelihood of respiratory diseases.
Previous studies indicated that high disease prevalence among female textile and garment workers could be linked to poor living conditions and engagement in an unhygienic work environment. These factors need further study to explore the contexts of this working population.
This study shows that the low educational status of workers in the textile and garment factories is associated with several disease groups, including injuries, MSD, peptic ulcer, UTI, AFI and hemorrhoids. Several studies have also revealed that textile and garment production workers with low educational status are at increased risk of different diseases. Another study in India also indicates high overall morbidity among textile and garment workers, which was significantly associated with low educational status. Also, a systematic review indicated that lower educational status could increase the health vulnerabilities of workers. A large study from WHO (n = 30,146) also showed that adults' low educational status was significantly associated with MSD in the LMICs. The increased risk of disease might be associated with the fact that most workers who are low educated are engaged in blue-collar jobs and may not be aware of the presence of different workplace hazards and may have poor access to the safety information at work.
Workers age is also associated with diseases in this study; increased age posing a significantly higher risk for several diseases, including respiratory, MSD, injury, ear diseases and gastrointestinal diseases in comparison to younger workers. Similarly, a study of general health problems among female garment workers in India showed that older age workers have a significantly higher risk for various diseases such as respiratory diseases, gastrointestinal diseases, MSD and eye diseases compared to younger workers. Older workers might be exposed to workplace hazards for many years and have high cumulative exposures. Moreover, workers with work services greater than 5 years had a significantly higher risk for 13 disease groups in the textile department than workers with service less than 5 years. Older factories (1 and 2) had also significantly higher AOR for several disease groups than the recently established factory (3). Therefore, factory clinics can be a good source of information for work-related disease research and action in LMICs setting.
## Different diagnoses than in the general population
The most prevalent cause of morbidity in this study is respiratory health problem (34%), followed by MSD (29%), GI (21%), peptic ulcer (19%) and injuries (17%). The magnitude and type of morbidities are higher and contrast from the general population in Ethiopia. The prevalence of the top leading diseases in the general population of Ethiopia are pneumonia (2.6%), acute upper respiratory infection (2.4%), typhoid (1.7%), dyspepsia (1.6%) and functional intestinal disorder (1.4%). Unlike the general population, most diseases from the textile and garment department in this study are non-communicable diseases related to dust exposure, ergonomic hazards, contact with chemicals and dangerous machines.
According to the ILO report, some of the diseases diagnosed among the textile and garment workers in the integrated factories could be work-related. The diseases are higher in magnitude and different from the diseases found in the general population, especially respiratory diseases, MSD, injury and ear diseases. These diseases might be due to the presence of hazards at the workplaces known to cause those health problems.
According to the "healthy workers effect" concept, a lower morbidity rate is expected among workers compared to the general population. However, the comparison of the morbidity rate of the current study with the general population should be taken cautiously. Generally, workers in integrated factories were distinct from the general population in many ways; they have free access to health service and information, and have a higher average age of 40 years, whereas the average age in the general population is 20 years. In addition, 92% of the study population attended formal education, with only 67% of the general population. This could make a difference in health-seeking behaviour.
In the early nineteenth century, textile workers in the US and Europe suffered from multiple diseases, including a high prevalence of respiratory health problems related to textile cotton exposure. The occupational health and safety standard improvements in developed countries have reduced workers' health problems. Economic globalization pushed textile factories to developing countries, specifically Asia and Africa. It seems that work-related health problems were exported together with the factories but the improved occupational health and safety standards have been left behind.
One of the strengths of this study was the inclusion of workers in the factories from all three work departments of textile, garment, and support. However, we do not know how representative the disease figures are regarding actual prevalence since the workers may also visit other health institutions. Conversely, the factory clinics serve workers free of charge and have a referral to hospitals for advanced diagnosis and treatment; thus, it is very likely that workers can consult the factory clinics to a large extent.
On the other hand, the main weakness of this study was the lack of standard diagnostic codes in the archives from the health clinics that forced us to use large categories for diagnoses. We also had limited information about the worker's exposure profile and could not collect potential confounder variables such as previous health condition, current work exposure at the different departments, housing, living environment, family, behaviour and lifestyle-related information. It is subsequently difficult to know the root causes of various health problems. Using a control group from another industry might have improved the study. Comparing groups inside the factory have advantages to link health problems to work conditions as the workers in the three departments had the same organizational experiences and the same factory culture. There still however could be selfselection of workers for the job, and movement of workers within the departments attributed to health conditions could be an inherent problem.
# Conclusion
About two-thirds of the workers in the integrated textile factories are diagnosed with different types of diseases, with a high prevalence of morbidity. The textile and garment department workers suffered a higher prevalence of diseases than support department workers, indicating that some diseases might be related to work in these departments. Work department, sex, age and educational status were significantly associated with several registered work-related diseases. Besides, respiratory, MSD, injuries and ear diseases were higher in quantity than the general population. Factory clinics seem to be an essential source of evidence to understand the occupational disease burden. Comparison of the risk level among the working departments and the general population needs careful attention while interpreting the result due to the lack of control for the potential confounders. Further study is needed to investigate the reason for repeated clinic consultation and rare chronic diseases such as cancer, heart diseases, renal diseases, and diabetics in relation to worker's exposure profile. |
Perirectal epidermoid cyst in a patient with sacrococcygeal scoliosis and anal sinus: A case report
BACKGROUNDPerirectal epidermoid cysts are rare masses arising from the ectodermal germ cell layer of the hindgut and are predominantly found in middle-aged women. It is often difficult to make an accurate diagnosis of these cysts and it is equally challenging to distinguish it from other developmental cysts.CASE SUMMARYWe report the case of an 18-year-old female patient with a perirectal mass who presented to the hospital with constipation. The patient experienced sacrococcygeal falls and burns on the left buttocks during growth. Three-dimensional computed tomography scans indicated abnormal sacral vertebrae with the sacral canal partially enlarged and opened. Pelvic magnetic resonance imaging showed a 55 mm × 40 mm × 35 mm unilocular cystic mass in the perirectal space and a solitary sinus in the left ischiorectal fossa. The cyst was completely resected posteriorly using the sacrococcygeal approach. The pathology was verified to be an epidermoid cyst. The patient remained recurrence-free after 6 mo of follow-up.CONCLUSIONSuccessful treatment of perirectal epidermoid cysts depends on comprehensive evaluation. This is significant for the surgical approach and prognosis.
# Introduction
Epidermoid cysts are commonly found in different parts of the body; however, perirectal epidermoid cysts are extremely rare. Perirectal cysts are congenital abnormalities considered to originate from caudal embryonic vestiges . Perirectal epidermoid cysts occur mostly in middle-aged women; however, they are rare in younger women [bib_ref] Presacral epidermoid cyst: imaging findings with histopathologic correlation, Yang [/bib_ref]. An abnormal mass in the pelvic floor space is often incidentally discovered during routine examinations. Most patients with perirectal cysts do not exhibit clinical symptoms. However, some patients may present with non-specific symptoms resulting from the compression of adjacent tissues, including urinary retention, constipation and a palpable mass near the anus . To help improve clinical diagnostic strategies and prevent misdiagnoses of the condition, we report a rare case of a perirectal epidermoid cyst occurring in a younger female. year-old female patient presented to the clinic with a complaint of constipation.
## Case presentation
## Chief complaints
## History of present illness
The patient had difficulty in evacuating her bowels for 2 mo.
## History of past illness
Three months after birth, the patient suffered from sacrococcygeal deformity due to an accidental fall and underwent an imaging examination at the local hospital without any therapy. At age five, her left buttock was scalded with boiling water.
## Personal and family history
The patient had no family history of inflammatory bowel disease or hereditary tumors.
## Physical examination
No abnormalities were found upon abdominal examination. Her buttocks were asymmetrical with irregular scar hyperplasia and shrinkage observed at the four o'clock position, which was approximately 3 cm away from the anal opening. Physical examination revealed a soft, mobile and poorly circumscribed mass without tenderness in the right perirectal region. The sacrum and coccyx were displaced to the right. Digital rectal examination revealed a bulge in the retrorectal area resulting in mild stenosis of the lumen [fig_ref] Figure 1: Preoperative image of the mass [/fig_ref].
## Laboratory examinations
Laboratory test results were normal.
## Imaging examinations
Colonoscopy revealed a slight stenosis of the area between the rectum and anal canal without erosion, ulceration or tumor formation. Three-dimensional (3D) computed tomography (CT) revealed abnormal sacral vertebrae with the sacral canal partially enlarged and opened. No destruction of the local sacral
## Final diagnosis
The cystic mass was identified to be an epidermoid cyst.
## Treatment
The cystic mass did not affect the patient's spine and the sinus in her left buttock showed no signs of infection. Complete surgical excision was performed using the posterior transsacrococcygeal approach. The patient was placed in the jackknife position after spinal anesthesia. A longitudinal median incision was made over the mass of the body surface on the left side of the coccyx. The skin, subcutaneous tissue, fat layer and levator ani muscle were cut layer by layer to reveal the space of pelvic floor. The cystic mass was firmly attached to the puborectalis and left posterior wall of the lower rectum. Using both blunt and sharp dissection to carefully dissect the surrounding anatomical structures, the cyst was completely excised. The cyst measured 55 mm × 40 mm × 35 mm and was filled with a soybean curd residue-like material. The rectum was confirmed to be intact by using an intraoperative anoscope. The space of the pelvic floor was stitched into layers with a drainage tube left in situ. Histologically, the cystic cavity was covered with squamous epithelium and composed of gray and white cheese-like layered keratinocytes mixed with exfoliated broken epidermal cells, keratin and cholesterol.
After surgery, the patient was hospitalized for purgative and preventive antibacterial treatment. We changed the dressing after defecation twice daily. The patient was discharged 2 wk after the surgery with a drain and the stitches were removed.
## Outcome and follow-up
Six months after surgery, the patient remained recurrence-free.
# Discussion
The presacral space is a triangular space between the posterior sacrum and anterior rectum that is bounded by the peritoneal reflection superiorly and levator ani muscles inferiorly . Developmental cystic masses arising from the presacral space include several kinds of tumors including dermoid cysts, epidermoid cysts, chordomas, adrenal rest tumors, anterior sacral meningoceles, cystic hamartomas, tailgut and rectal duplication cysts . Epidermoid cysts in the presacral space are uncommon with an incidence of 1 in 40000-63000 hospital admissions and 60% of perirectal epidermoid cysts are congenital October 26, 2022
Volume 10 Issue 30 developmental cysts [bib_ref] Clinical Study and Review of Articles (Korean) about Retrorectal Developmental Cysts in..., Baek [/bib_ref]. Perirectal epidermoid cysts are typically slow-growing; approximately 26%-50% of patients are asymptomatic and they are usually incidentally discovered during imaging [bib_ref] Carcinoma arising in epidermoid cyst: a case series and aetiological investigation of..., Morgan [/bib_ref]. Compression of pelvic structures by an enlarging cyst may present with urinary complaints, constipation, perianal pain or a palpable mass in the precoccygeal region. When there is an infection, these masses may result in perianal discharge, fistulous opening and bleeding in the rectum. Although epidermoid cysts are common October 26, 2022
Volume 10 Issue 30 skin lesions, they rarely develop into squamous cell carcinomas. Malignant tumors arising from epidermoid cysts are reported to appear at a rate of 0.011%-2.2% [bib_ref] Multiple complications from an intracranial epidermoid cyst: case report and literature review, Abramson [/bib_ref]. The exact pathogenesis of epidermoid cysts becoming malignant tumors remains unclear. However, the disease progression may be attributed to chronic inflammatory responses to repeated cyst ruptures and a subtotal resection of the cyst wall[10]. Elevated levels of AFP or HCG may be indicative of germ cell tumors. Colonoscopy may reveal extrinsic rectal compression and exclude intestinal space-occupying lesions. Transrectal ultrasonography may be useful for assessing the location and extent of small cystic masses and their connection with the anal sphincter [bib_ref] Sacrococcygeal neurofibroma: rare cause for chronic pelvic pain, Paul [/bib_ref]. CT examination clearly showed bone destruction by malignant presacral masses. MRI is superior to CT in differentiating between any bone, spinal canal or meningeal involvement and its relationship with surrounding soft tissues and organs to determine the appropriate surgical plan [bib_ref] Management of presacral tumors: Our experience with posterior approach, Saxena [/bib_ref]. Although infections, hemorrhages or calcifications in these lesions may alter the signal intensity, subtle changes in the signal intensity favor epidermoid cysts. However, these findings are not specific. Except for suspicious cancerous lesions, preoperative biopsy should not be performed to prevent tumor dissemination, abscess, fecal fistula or meningitis[1].
Choosing a surgical plan for presacral tumors largely depends on the tumor's location, size and relationship with surrounding tissues and organs. Common surgical approaches are transsacrococcygeal, transabdominal, transsphincter and combined transabdominal and transsacral approaches [bib_ref] Unusual cause of defecation disturbance: a presacral tailgut cyst, Akbulut [/bib_ref]. If the mass is small (≤ 10 cm), located at the caudal level (below S4) and does not invade surrounding structures, the sacral approach is usually adopted [bib_ref] The trans-sphincteric and trans-sacral approaches for the surgical excision of rectal and..., Kanemitsu [/bib_ref]. When the mass is located at a high spinal level (above S3), the transabdominal approach is a better option. When the mass is large, its location near surrounding organs such as the ureter and iliac vessels is unclear and it is difficult to employ a single approach; thus, the combined abdominal sacrococcygeal approach can be selected[A4]. The sphincter approach is an option for patients with small, low-lying lesions . Gynecologists choose the transvaginal approach because it provides a sufficient field of vision, shorter operative time and lower blood loss in low-lying retrorectal lesions [bib_ref] Perianal Tailgut Cyst, Grossi [/bib_ref].
The differential diagnoses of perirectal cystic lesions include tailgut cysts, cystic teratomas, chordomas and anterior sacral meningoceles [bib_ref] Multiple Presacral Teratomas in an 18-year-old Girl: A Case Report, Park [/bib_ref] [bib_ref] Presacral chordoma diagnosed by transrectal fine-needle aspiration cytology, Permi [/bib_ref] [bib_ref] An Unusual Presentation as Recurrent Abortions in a Case of Giant Presacral..., Rege [/bib_ref] [bib_ref] Presacral epidermoid cyst in an elderly female, mistaken for ovarian cystadenoma, Jha [/bib_ref]. In female patients, a high (in terms of location) perirectal mass is misdiagnosed as ovarian cystadenoma . The unique feature of this case is that a perirectal mass with perianal sinus and sacrococcygeal malformation was found in a young female patient. To accurately determine the extent of the mass lesion and rule out other pathologies, the patient underwent several imaging examinations. MRI revealed a solitary abscess in the left ischiorectal fossa which had no sinus interacting with the presacral mass. Given the patient's history of scalding, sterile necrotic tissue may have accumulated in the perianal sinus. Three-dimensional CT scans revealed an abnormal sacral vertebra with the sacral canal partially enlarged and opened; however, there was no damage to the sacral surface bone. The wall of the cyst was remote from the sacral canal, the opened sacral canal was at the S3 Level and the cyst was below S4. Evidence of an anterior sacral meningocele is insufficient. Although extremely rare, benign cysts can also progress into malignant tumors . In our case, a well-defined smooth margin with an absence of invasion may have helped differentiate it from malignant masses. This case highlights the importance of pre-operative imaging and evaluation to identify the nature of the presacral mass which is crucial for surgery and prognosis.
Laparoscopic surgery is an option because of its minimal invasiveness, low risk of complications and complete tumor removal. Considering that the patient was a young female who had not given birth, surgery was performed using the transsacrococcygeal approach to avoid damaging the pelvic organ. During surgery, we cut the cystic mass to reduce its volume for a larger operative space. In cases of suspected malignant tumors, it should be carefully stripped along the capsule to prevent rupture and metastasis. In this regard, the importance of preoperative differential diagnosis is emphasized.
# Conclusion
Distinguishing an epidermoid cyst from other perirectal cystic masses is a significant diagnostic challenge. The clinical manifestations of presacral masses vary and once found, colonoscopy, sacrococcygeal CT and pelvic MRI findings should be further evaluated. Using the tumor's location, size and relationship with the surrounding organs, an appropriate surgical plan should be selected.
[fig] Figure 1: Preoperative image of the mass. bone was observed (Figure 2). Magnetic resonance imaging (MRI) revealed a 55 mm × 40 mm × 35 mm well-circumscribed unicameral cystic mass in the pelvis that adhered to the left rectal wall posteriorly to the sacrum resulting in a right anterior displacement of the rectum. The vaginal wall was compressed and wrinkled (Figure 3). Enhanced strips in the scan extended from the levator ani muscle to the skin of the left buttock (Figure 4). Based on the MRI findings, preoperative diagnoses included a presacral epidermoid cyst, anal sinus and sacrococcygeal scoliosis. [/fig]
[fig] Figure 2: Computed tomography scan examination of the sacral vertebra. A: Frontal view; B: Dorsal view: Three-dimensional computed tomography scan showed a sacrococcygeal scoliosis below the S2 level (white arrows); the sacral canal is partially enlarged and opened. [/fig]
[fig] Figure 3: Magnetic resonance imaging. A: T2-weighted imaging: a well-circumscribed mass (asterisk) compressing the rectum and displacing it right-anteriorly; B: T2-weighted imaging showed a well-defined mass anterior to the sacrum. [/fig]
[fig] Figure 4 Axial: T1-weighted imaging. A: Circular signal (white arrows) on the outside of the levator ani muscle with strips signs connecting to the skin of the left buttock; B: Contrast-enhanced T1-weighted imaging showed enhancement of the strips signs. [/fig]
|
Cumulative Risks of Foster Care Placement by Age 18 for U.S. Children, 2000–2011
Foster care placement is among the most tragic events a child can experience because it more often than not implies that a child has experienced or is at very high risk of experiencing abuse or neglect serious enough to warrant state intervention. Yet it is unclear how many children will experience foster care placement at some point between birth and age 18. Using synthetic cohort life tables and data from the Adoption and Foster Care Analysis and Reporting System (AFCARS), we estimated how many U.S. children were placed in foster care between birth and age 18, finding support for three conclusions. First, up to 5.91% of all U.S. children were ever placed in foster care between their birth and age 18. Second, Native American (up to 15.44%) and Black (up to 11.53%) children were at far higher risk of placement. Foster care is thus quite common in the U.S., especially for historically disadvantaged racial/ethnic groups. Third, differences in foster care placement were minimal between the sexes, indicating that the high risks of foster care placement are shared almost equally by boys and girls.
# Introduction
Of all childhood experiences, foster care placement is among the most tragic. In most instances, children are placed in foster care because the state rules that the child has experienced or is at high risk of experiencing maltreatment (abuse or neglect). Though clearly helpful to some children, foster care placement frequently introduces additional instability to their already-chaotic lives, potentially further harming them. This combination of maltreatment and instability means that children who have experienced foster care suffer not only from elevated rates of mortality in childhood [1], but also from a host of other problems ranging from asthma to behavioral problems to suicidal ideation [bib_ref] Assessing the health status of children entering foster care, Chernoff [/bib_ref] [bib_ref] Mental health problems of children in foster care, Clausen [/bib_ref] [bib_ref] Health status of children in foster care: The experience of the center..., Halfon [/bib_ref] [bib_ref] Timely identification of mental health problems in two foster care medical homes, Jee [/bib_ref] [bib_ref] Child maltreatment 1: Burden and consequences of child maltreatment in high-income countries, Gilbert [/bib_ref] [bib_ref] Child protection and child outcomes: measuring the effects of foster care, Doyle [/bib_ref] [bib_ref] Child protection and adult crime: Using investigator assignment to estimate causal effects..., Doyle [/bib_ref] [bib_ref] Mental health care utilization and expenditures by children in foster care, Harman [/bib_ref]. Children in foster care are five times more likely to be diagnosed with depression, four times more likely to be diagnosed with ADHD, and ten times more likely to be diagnosed with bipolar disorder than other children, for instance [bib_ref] Mental health care utilization and expenditures by children in foster care, Harman [/bib_ref].
Annual foster care entry rates and point-in-time estimates of the number of children in foster care suggest that foster care placement is rare. In 2011, the most recent year for which data are available, only 0.34% of all American children entered foster care within the year, and only 0.54% of children were in foster care on any given day. These estimates imply that despite foster care placement's implications for individual children, its societal importance may be minimal because it affects few children. However, these data may create an inaccurate portrayal of how common foster care placement is because annual and daily estimates of children in foster care do not convey how many children ever experience placement during their entire childhood.
Is foster care placement an event so uncommon it requires minimal consideration? Or, is it an event common enough that it merits serious attention from researchers and policymakers? We provide insight into these questions by using synthetic cohort life tables to estimate the cumulative probability of foster care placement for children from birth to age 18 with data from the Adoption and Foster Care Analysis and Reporting System (AFCARS) in the years 2000-2011. We also stratify the risk of being placed in care by age 18 by race/ethnicity, sex, and year.
# Materials and methods
The Adoption and Foster Care Analysis and Reporting System (AFCARS)
We used data from the Adoption and Foster Care Analysis and Reporting System (AFCARS), which includes all children in foster care from 2000-2011, for the numerator , in concert with race-, age-, sex-, and year-specific estimates of the U.S. population for the denominator. The number of children at risk of experiencing their first foster care placement by age and year is recorded in [fig_ref] Table 1: Number of Children at Risk of First Foster Care Placement by Age [/fig_ref]. The number of children experiencing their first foster care placement by age and year is recorded in .
The AFCARS data are publicly available through the National Data Archive on Child Abuse and Neglect, which is housed at Cornell University, and have been de-identified prior to being made available to researchers in the publicly available version of the data, which are the data we used for all analyses. These data can be accessed after signing a terms of use agreement form here:
http://www.ndacan.cornell.edu/datasets/datasets-listafcars.cfm. Because the Adoption and Foster Care Analysis and Reporting System (AFCARS) data are publicly available and deidentified, the Yale University Institutional Review Board deemed this research exempt. [fig_ref] Table 1: Number of Children at Risk of First Foster Care Placement by Age [/fig_ref] The AFCARS data contain ''case-level information on all children in foster care for whom State and Tribal title IV-E agencies have responsibility for placement, care or supervision and on children who are adopted under the auspices of the State and Tribal title IV-E agency. Title IV-E agencies are required to submit AFCARS data semi-annually to the Children's Bureau. The AFCARS report periods are October 1 through March 31 and April 1 through September 30''. We used the combined reporting files, so what we refer to as 2010 spans October 1, 2009 to September 30, 2010. Because reporting to the AFCARS is mandatory, all states contributed data for all 12 years.
The AFCARS data include information pertinent to the child welfare system, such as when the child was most recently placed in foster care and whether that was the first placement or a higher order placement, as well as basic demographic information such as age, sex, and race. Additionally, the dataset includes information on how long the child has been in care-a number that ranges from a few days to many years-and what type of care arrangement they are in, specifying, for instance, whether the child is living in a pre-adoptive home, kinship foster care, non-kin foster care, a group home, an institution including a juvenile detention center (although children in a juvenile detention center would only be considered to be in foster care if removed from their homes for one of the reasons listed below), or independently.
Roughly two-thirds of the children in the data enter foster care because they experienced maltreatment, with the other one-third of children entering foster care for a variety of reasons tied to their parents. Such reasons include parental drug or alcohol abuse, parental death, parental inability to cope (encompassing many things), parental abandonment, and inadequate housing. Some children were admitted for reasons tied to the child, including having disabilities better provided for in foster care or exhibiting serious behavioral problems.
## Measures
We rely on four measures: (1) age, (2) sex, (3) race/ethnicity, and (4) first admission to foster care. Age at first admission to foster care is based on the difference between the admission date and his/her birth date. Sex is based on caseworker reports of sex. There is little missing data on age or sex; in 2005, 915 of about 800,000 cases, or 0.1%, were missing data on sex or age. We treated these cases as missing completely at random. Because of the small amount of missing data (,0.1%), the choice of method for dealing with it minimally affects the results.
Race/ethnicity is based on caseworker reports, with Asian, Black, Hispanic, Native American, Pacific Islander, and White as the options. Some children were reported as having more than once race, but since we wanted to provide estimates for specific racial/ethnic groups, we were forced to assign them one racial/ ethnic identity. In all cases in which the caseworker reported the child was Native American, we considered the child to be Native American (regardless of additional identities). For those children who were not Native American but were considered to be multiracial, we considered (1) children reported to have Hispanic ethnicity as Hispanic, (2) children reported to be Black (but not Hispanic) as Black, (3) children reported to be Asian or Pacific Islander (but not Hispanic or Black) as Asian, and (4) all remaining children as White. To account for the approximately 2% of cases who had missing racial/ethnic identity information we distributed these cases in the following way. Children whose Hispanic ethnicity was marked ''unable to determine'' and who had missing racial data were distributed amongst Hispanic cases; children whose race was marked ''unable to determine'' and whose ethnicity was unmarked were distributed amongst non-White cases; and children for whom both racial and ethnic data was missing were distributed amongst all cases equally. Because of the small amount of missing data (,2%), the choice of method for dealing with it minimally affects the results. Our measure of first time admissions to foster care in the last year is based on whether the admission was the first placement to foster care and occurred in the reporting year and is based on a constructed variable in the dataset that allowed us to differentiate first and subsequent admissions by matching children based on unique (de-identified) IDs. This measure is crucial for our analysis because if we also included subsequent admissions to foster care, we would count children as experiencing the event more than once, thereby incorrectly inflating the estimates.
## Analytic strategy
To construct the estimates of the cumulative risk of foster care placement between birth and age 18, we use synthetic cohort (or period) life tables. Synthetic cohort life tables, which were originally designed to study mortality, have long been used to study what proportion of a hypothetical cohort of 100,000 individuals would survive to any given age if they were exposed to any year's age-specific mortality rates at each age. (The Census Bureau still uses this method to estimate life expectancy at birth in the United States.) In this adaptation, the synthetic cohort life table provides an estimate of the proportion of a hypothetical cohort of children exposed to the age-specific first time foster care admission rates at each age who could expect to experience foster care placement at some point between birth and age 18.
The key benefit of synthetic cohort life tables for our analysis is that they allow us to estimate the cumulative risk of foster care placement to age 18 using only 1 year of data, a calculation that would be impossible using birth cohort life tables, which require 18 years of data to provide that estimate because they follow a birth cohort through time.
Synthetic cohort life tables also have significant limitations. Most importantly, the estimates they produce may be unreliable when the age-specific rates are changing quickly. This limitation is especially problematic if the number of synthetic cohorts estimated is small. Although this can be a key limitation of this method, it is important to note two things before reviewing the results. First, with one exception (Native Americans), the age-specific risks of foster care placement are changing gradually, not rapidly, in our data, meaning that the bias in our estimates should be minimal. Second, because we have 12 years of data, it would be easy to identify years in which the estimates were unreliable. Indeed, parallel analyses comparing synthetic and birth cohort life tables estimating the cumulative risk of foster care placement in California showed that even with only two synthetic cohorts (based on 2000 age-specific first-time entry rates and 2006 agespecific first-time entry rates), estimates using synthetic cohort and birth cohort life tables were very similar [bib_ref] Children ever in care: an examination of cumulative disproportionality, Magruder [/bib_ref].
Because children who have already been placed in the foster care system are still included in the total population counts provided by CDC Wonder (even though they are no longer at risk of first placement), we adjust the denominator down accordingly. [fig_ref] Table 1: Number of Children at Risk of First Foster Care Placement by Age [/fig_ref] presents the adjusted denominators by age.) For example, we multiply the number of children five years of age in the Properly counting children who enter the population of children at risk of first foster care placement (through immigration) and who leave the population of children at risk of first foster care placement (through emigration and death) is also essential. Because the number of children at risk of first foster care placement is updated annually (based on the number of children in the population according to CDC Wonder), children entering and leaving the population at risk only minimally affect our results.
Because synthetic cohort life tables rely on only one year of data, we provided annual estimates of the cumulative risk of foster care placement for each year from 2000 to 2011. Though our data include the entire population of interest, we present confidence intervals because even in the most complete dataset, there is always some disparity from the population. Our confidence intervals are based on Greenwood's formula for the asymptotic standard error [bib_ref] The Natural Duration of Cancer, Greenwood [/bib_ref]. We used Stata/SE 12 for all analyses.
# Results
The cumulative risk of ever being placed in foster care between birth and age 18 for all American children was 5.91% in 2005 [fig_ref] Table 3: Cumulative Risks of Foster Care Placement from Birth to Age 18 for... [/fig_ref]. By age six, the cumulative risk of foster care placement was 3.11% [fig_ref] Table 3: Cumulative Risks of Foster Care Placement from Birth to Age 18 for... [/fig_ref]. This risk increased to 4.42% by age 12 and 5.91% by age 18 [fig_ref] Table 3: Cumulative Risks of Foster Care Placement from Birth to Age 18 for... [/fig_ref].
Cumulative risks of foster care placement differed dramatically by race/ethnicity. White and Hispanic children had cumulative risks of foster care placement relatively close to those for the population, at 4.86% and 5.35%, respectively [fig_ref] Table 3: Cumulative Risks of Foster Care Placement from Birth to Age 18 for... [/fig_ref]. In contrast, Asian children had the lowest risk at 2.14%, while Black children and Native American children had the highest cumulative risks of placement, at 10.99% and 15.44%, respectively [fig_ref] Table 3: Cumulative Risks of Foster Care Placement from Birth to Age 18 for... [/fig_ref]. Compared to White children, Hispanic children were at 1.10 relative risk of foster care placement, Asian children were at 0.44 relative risk of foster care placement, Black children were at 2.26 relative risk of foster care placement, and Native American children were at 3.18 relative risk of foster care placement (p,.001 for all comparisons to the White population; see [fig_ref] Table 3: Cumulative Risks of Foster Care Placement from Birth to Age 18 for... [/fig_ref] for 95% Confidence Intervals).
Children had the highest risk of first foster care placement during infancy, with 1.09% of all U.S. children first entering care before their first birthday [fig_ref] Figure 1: Age-Specific Risks of First-Time Foster Care Placement [/fig_ref]. The risk of first placement then trailed off until age 13, at which point it increased throughout adolescence. The age-patterning of first placement was similar for all groups [fig_ref] Figure 1: Age-Specific Risks of First-Time Foster Care Placement [/fig_ref].
Between 2000 and 2011, cumulative risks of foster care placement declined slowly, but substantial racial/ethnic disparities persisted [fig_ref] Figure 2: Cumulative Risk of Foster Care Placement by Age 18 [/fig_ref]. The percentage of all U.S. children estimated to ever be in foster care between birth and age 18 ranged from 4.76% in 2009 to 5.91% in 2005 [fig_ref] Figure 2: Cumulative Risk of Foster Care Placement by Age 18 [/fig_ref]. Shifts were largest for Native American children, whose risk ranged from 10.54% in 2011 to 15.44% in 2005, and Black children, whose risk ranged from 8.84% in 2010 to 11.53% in 2001.
Females were more likely to ever be placed in foster care than were males for all years, but these differences were often small [fig_ref] Figure 3: Cumulative Risk of Foster Care Placement by Demographic Group [/fig_ref]. Thus, racial/ethnic stratification in the cumulative risk of placement was more substantial than was sex stratification.
# Discussion
The results from our analyses demonstrate that foster care placement is far more common than often thought. Up to 5.91% of U.S. children (1 in 17) will experience foster care placement at some point between birth and age 18. The risk, however, is not evenly distributed. A shocking 15.44% (1 in 7) of Native American children and 11.53% (1 in 9) of Black children will enter foster care at some point before they turn 18. This risk of being in foster care is shared almost equally by boys and girls, further suggesting the global nature of the problem.
Although this is not the first study to use demographic methods to estimate the cumulative risk of foster care placement [bib_ref] Children ever in care: an examination of cumulative disproportionality, Magruder [/bib_ref] , it nonetheless extends research in this area in three key ways. First, it considers the entire country instead of just one state (California), giving broad insight into how common foster care placement is across the entire United States. Second, it provides cumulative risk estimates throughout the entirety of childhood (to age 18) rather than to age 7. Finally, it provides estimates for the entire period during which U.S. foster care caseloads have declined so sharply (2000-2011) rather than just during the early part of this period . Thus, the current study greatly advances research beyond the important early applications of life tables to study the cumulative risk of foster care placement to age seven in California [bib_ref] Children ever in care: an examination of cumulative disproportionality, Magruder [/bib_ref].
These findings document a pressing need for further research and policy measures on the topic. The prevalence of foster care in the lives of American children, for instance, suggests that further investigation into the consequences of placement in foster care-as distinct from the circumstances that lead to foster care placement-and how foster care can be more beneficial to children is necessary. Moreover, since foster care placement is indicative of poor life circumstances-whether because of the conditions that caused it or because of the instability resulting from removal from one's birth family-it is concerning that Black and Native American children have far greater risks of experiencing such circumstances. Such findings call for additional research on the interaction between social inequality and foster care placement. Finally, since the risk of placement is highest in the first year of life, additional support to pregnant women and first time mothers may be good policy to reduce foster placements.
In light of these high cumulative foster care placement risks and associated outcomes, researchers and policymakers must give far greater attention to this vulnerable group of children.
[fig] 009: (0.009, 0.009) 0.023 (0.023, 0.024) 0.009 (0.009, 0.009) 0.004 (0.003, 0.004) 012 (0.012, 0.013) 0.032 (0.031, 0.032) 0.013 (0.012, 0.013) 0.005 (0.005, 0.005) 0.050, 0.051) 0.041 (0.041, 0.041) 0.094 (0.094, 0.095) 0.045 (0.044, 0.046) 0.017 (0.017, 0.018) 0.140 (0.136, 0.144) 15 0.054 (0.054, 0.054) 0.044 (0.044, 0.044) 0.101 (0.100, 0.102) 0.048 (0.048, 0.049) 0.019 (0.018, 0.020) 0.147 (0.143, 0.150) 16 0.057 (0.057, 0.058) 0.047 (0.047, 0.047) 0.107 (0.106, 0.108) 0.052 (0.051, 0.052) 0.020 (0.020, 0.021) 0.152 (0.148, 0.156) 17 0.059 (0.059, 0.059) 0.049 (0.048, 0.049) 0.110 (0.109, 0.111) 0.054 (0.053, 0.054) 0.021 (0.021, 0.022) 0.154 (0.151, 0.158) doi:10.1371/journal.pone.0092785.t003 [/fig]
[fig] Figure 1: Age-Specific Risks of First-Time Foster Care Placement. doi:10.1371/journal.pone.0092785.g001 Cumulative Risks of Foster Care Placement PLOS ONE | www.plosone.org population in 2005 by the probability of having never been placed in foster care by age five based on 2005 rates for children five years of age. This adjustment, although important for the precision of the results, does not greatly alter the findings. [/fig]
[fig] Figure 2: Cumulative Risk of Foster Care Placement by Age 18. doi:10.1371/journal.pone.0092785.g002 [/fig]
[fig] Figure 3: Cumulative Risk of Foster Care Placement by Demographic Group. doi:10.1371/journal.pone.0092785.g003 [/fig]
[table] Table 1: Number of Children at Risk of First Foster Care Placement by Age (0-17) and Year (2000-2011). [/table]
[table] Table 3: Cumulative Risks of Foster Care Placement from Birth to Age 18 for All U.S. Children and White, Black, Hispanic, Asian, and Native American Children, 2005. [/table]
|
Iron-Modified Blood Culture Media Allow for the Rapid Diagnosis and Isolation of the Slow-Growing Pathogen Francisella tularensis
The life-threatening disease tularemia is caused by Francisella tularensis, an intracellular Gram-negative bacterial pathogen. Due to the high mortality rates of the disease, as well as the low respiratory infectious dose, F. tularensis is categorized as a Tier 1 bioterror agent. The identification and isolation from clinical blood cultures of F. tularensis are complicated by its slow growth. Iron was shown to be one of the limiting nutrients required for F. tularensis metabolism and growth. Bacterial growth was shown to be restricted or enhanced in the absence or addition of iron. In this study, we tested the beneficial effect of enhanced iron concentrations on expediting F. tularensis blood culture diagnostics. Accordingly, bacterial growth rates in blood cultures with or without Fe 21 supplementation were evaluated. Growth quantification by direct CFU counts demonstrated significant improvement of growth rates of up to 6 orders of magnitude in Fe 21 -supplemented media compared to the corresponding nonmodified cultures. Fe 21 supplementation significantly shortened incubation periods for successful diagnosis and isolation of F. tularensis by up to 92 h. This was achieved in a variety of blood culture types in spite of a low initial bacterial inoculum representative of low levels of bacteremia. These improvements were demonstrated with culture of either Francisella tularensis subsp. tularensis or subsp. holarctica in all examined commercial blood culture types routinely used in a clinical setup. Finally, essential downstream identification assays, such as matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF-MS), immunofluorescence, or antibiotic susceptibility tests, were not affected in the presence of Fe 21 . To conclude, supplementing blood cultures with Fe 21 enables a significant shortening of incubation times for F. tularensis diagnosis, without affecting subsequent identification or isolation assays. IMPORTANCE In this study, we evaluated bacterial growth rates of Francisella tularensis strains in iron (Fe)-enriched blood cultures as a means of improving and accelerating bacterial growth. The shortening of the culturing time should facilitate rapid pathogen detection and isolation, positively impacting clinical diagnosis and enabling prompt onset of efficient therapy.
the disease may indicate the intentional aerosolized dissemination of F. tularensis as a bioterror agent. Early administration of effective antimicrobial therapy is essential for favorable patient outcomes [bib_ref] New therapeutic approaches for treatment of tularaemia: a review, Boisset [/bib_ref]. Thus, early detection and isolation of the pathogen are instrumental for effective treatment.
Typically, two subspecies of F. tularensis are found throughout the Northern Hemisphere, which are responsible for the majority of human infections. F. tularensis subsp. tularensis (type A) strains are mainly found in Canada and the United States (5), while F. tularensis subsp. holarctica (type B) strains are mainly found in Europe and Asia [bib_ref] Tularaemia: clinical aspects in Europe, Maurin [/bib_ref]. F. tularensis subsp. tularensis can cause severe invasive diseases, such as pneumonia and bacteremia, whereas subsp. holarctica usually causes mild symptoms and has a low mortality rate [bib_ref] First case of severe pneumonic tularemia in an immunocompetent patient in the..., Sigaloff [/bib_ref]. The clinical significance of F. tularensis bacteremia is unknown, but it appears to occur in conjunction with more severe forms of the disease and with compromising underlying illnesses.
Diagnosing tularemia is not straightforward both due to the nonspecific nature of the initial symptoms and to the fact that F. tularensis is difficult to culture [bib_ref] Detection of selected fastidious bacteria, Doern [/bib_ref]. Furthermore, the handling of this bacterium poses a significant risk of infection to laboratory personnel [bib_ref] Detection of selected fastidious bacteria, Doern [/bib_ref].
In most clinical laboratories, diagnosis of tularemia relies on serological tests (which cannot be implemented earlier than 10 days from illness onset) [bib_ref] The development of tools for diagnosis of tularemia and typing of Francisella..., Johansson [/bib_ref] , except for patients exhibiting bacteremia, in which case, tularemia may be diagnosed incidentally by the detection of F. tularensis in blood cultures (BCs). The culture of the organism is instrumental for diagnosis, antibiotic susceptibility testing, biovar characterization [bib_ref] Identification of Francisella species and discrimination of type A and type B..., Forsman [/bib_ref] [bib_ref] Evaluation of PCR-based methods for discrimination of Francisella species and subspecies and..., Johansson [/bib_ref] , and molecular epidemiological typing [bib_ref] Francisella tularensis strain typing using multiple-locus, variable-number tandem repeat analysis, Farlow [/bib_ref] [bib_ref] Extensive allelic variation among Francisella tularensis strains in a short-sequence tandem repeat..., Johansson [/bib_ref]. However, direct identification of F. tularensis is limited, due to its low growth rates. Usually, the appearance of individual colonies on nonselective optimized agar plates requires 2 to 4 days of incubation, while in liquid media visible growth occurs within 3 to 7 days of incubation [bib_ref] Isolation of Francisella tularensis and Yersinia pestis from blood cultures by plasma..., Aloni-Grinstein [/bib_ref] , depending the inoculum dose.
The importance of early diagnosis incentivized the development of rapid and sensitive approaches for the detection of F. tularensis directly from clinical samples. These are mainly based on immune-labeling [bib_ref] Simultaneous immunodetection of anthrax, plague, and tularemia from blood cultures by use..., Mechaly [/bib_ref] [bib_ref] Detection of Francisella tularensis-specific antibodies in patients with tularemia by a novel..., Sharma [/bib_ref] [bib_ref] Development of an immunoassay test system based on monoclonal antibodies and immunomagnetic..., Vetchinin [/bib_ref] and genetic [bib_ref] A rapid high-throughput sequencing-based approach for the identification of unknown bacterial pathogens..., Israeli [/bib_ref] [bib_ref] Development of a multitarget real-time TaqMan PCR assay for enhanced detection of..., Versage [/bib_ref] assays.
Ever since blood cultures have been routinely performed for clinical diagnostics, isolation of F. tularensis from blood of infected humans has rarely been reported [bib_ref] Francisella tularensis bacteremia: report of two cases and review of the literature, Karagoz [/bib_ref]. This is possibly due to the low sensitivity of "classic" blood culturing systems and also the short duration of the bacteremia phase, early in the infection [bib_ref] Isolation of Francisella tularensis from blood, Provenza [/bib_ref] [bib_ref] Enhancing recovery of Francisella tularensis from blood, Reary [/bib_ref] [bib_ref] Tularemia: history, epidemiology, pathogen physiology, and clinical manifestations, Sjostedt [/bib_ref] , which implies that the timing of the blood collection is critical. In recent years, with the development of more nutritious culture media, continuous-monitoring devices, and improved incubation protocols, the frequency of tularemia cases diagnosed by blood culture has increased [bib_ref] Francisella tularensis bacteremia: report of two cases and review of the literature, Karagoz [/bib_ref] [bib_ref] Enhancing recovery of Francisella tularensis from blood, Reary [/bib_ref] [bib_ref] Septic pneumonic tularaemia caused by Francisella tularensis subsp. holarctica biovar II, Fritzsch [/bib_ref] [bib_ref] Francisella tularensis bacteremia, Haristoy [/bib_ref] [bib_ref] Tularemia in a kidney transplant recipient: an unsuspected case and literature review, Khoury [/bib_ref] [bib_ref] Fatal infection caused by Francisella tularensis in a neutropenic bone marrow transplant..., Sarria [/bib_ref] [bib_ref] An original case of Francisella tularensis subsp. holarctica bacteremia after a near-drowning..., Ughetto [/bib_ref]. Nevertheless, prolonged incubation times are often necessary for detection, requiring in some cases as much as 12 or even 21 days for positive blood cultures to be unambiguously determined [bib_ref] Typhoidal tularemia in a human immunodeficiency virus-infected adolescent, Gries [/bib_ref] [bib_ref] First isolation of Francisella tularensis subsp. tularensis in Europe, Gurycova [/bib_ref].
Generally, the addition of nutritional supplements to accelerate bacterial growth is an effective approach. Iron was shown to be one of the limiting nutrients required for F. tularensis metabolism and growth, as established by previous studies that demonstrated restriction or enhancement of bacterial growth with the absence or addition of iron, respectively [bib_ref] Identification of Francisella tularensis genes affected by iron limitation, Deng [/bib_ref] [bib_ref] Growth of Francisella tularensis LVS in macrophages: the acidic intracellular compartment provides..., Fortier [/bib_ref] [bib_ref] Characterization of the siderophore of Francisella tularensis and role of fslA in..., Sullivan [/bib_ref].
Commercially available ready-to-use clinical blood culturing diagnostic systems include manufacturer-provided media. However, their iron content is not standardized among manufacturers and may vary as a function of the hemin and or extracts present in the culture. In addition, an important source of iron is the analyzed peripheral blood from the sample itself; yet, this source may exhibit significant individual variations as well. Generally, the level of iron required for optimal bacterial growth is ;10 26 M; however, the level of free iron in mammalian tissues is typically ;10 218 M [bib_ref] Inhibition of Francisella tularensis phagocytosis using a novel anti-LPS scFv antibody fragment, Mechaly [/bib_ref]. Early studies during F. tularensis infection have suggested that iron withholding represents one of the typical host responses belonging to the innate nutritional immunity defense mechanism [bib_ref] Optimal testing parameters for blood cultures, Cockerill [/bib_ref]. This reduction of available iron in bacteremic blood samples combined with nonoptimal growth media can potentially limit bacterial growth in blood cultures and may result in false-negative diagnosis.
In this study, F. tularensis growth rates in iron (Fe 21 )-enriched blood cultures were evaluated as a means of improving and accelerating bacterial growth for early diagnosis and isolation of the pathogen. To the best of our knowledge, implementation of the beneficial effect of enhanced iron concentration to expedite F. tularensis blood culture diagnostics has not been attempted previously.
# Results and discussion
Accelerated growth rates of F. tularensis subsp. holarctica in Fe 2+ -supplemented blood cultures. In this study, the feasibility of improving bacterial growth rates in iron-supplemented blood cultures was first evaluated for the attenuated type-B F. tularensis live vaccine strain (LVS). Bacteria were spiked at two concentrations (300 and 3,000 CFU/mL) in Bactec Plus Aerobic/F culture vials supplemented with 100 mM Fe 21 containing 10 mL of naive fresh human blood. This Fe 21 concentration was previously selected after considering iron's stability in medium versus its potential toxicity at higher concentrations [bib_ref] An improvement in diagnostic blood culture conditions allows for the rapid detection..., Makdasi [/bib_ref]. Vials were incubated at 37°C at 150 rpm for 120 h, and bacterial growth was determined at different time points by CFU counts. As shown in , increased F. tularensis growth rates were observed in blood cultures supplemented with Fe 21 compared to nonsupplemented media. This beneficial effect promoted a significant increase of ;6 orders of magnitude in bacterial concentrations following 48 h of incubation in the low-inoculum culture (300-CFU/mL initial concentration) . For the high-inoculum culture (3,000 CFU/mL) , an increase of ;4 orders of magnitude was observed after 32 h. These results strengthen the highly beneficial nature of Fe 21 addition to blood culture vials, potentially shortening F. tularensis detection time in clinical samples.
The current clinical practice of blood drawing from symptomatic adult patients for diagnostic culturing consists of collecting from two distinct sites at least two pairs of samples, each set requiring 20 mL of blood, subsequently divided equally between aerobic and anaerobic cultures [bib_ref] Optimal testing parameters for blood cultures, Cockerill [/bib_ref] [bib_ref] Current blood culture methods and systems: clinical concepts, technology, and interpretation of..., Weinstein [/bib_ref]. The cultures are then incubated in a continuous monitoring system based on fluorescent detection of CO 2 production, which serves as a reporter for potential active microbial metabolism.
Accordingly, we evaluated F. tularensis growth in various routinely employed Bactec vials in the presence or absence of additional iron. Increasing inoculums of F. tularensis (LVS, 30 to 3,000 CFU/mL) were spiked into four different Bactec blood cultures: two aerobic vials (Plus Aerobic and Standard Aerobic) and two anaerobic vials (Standard Anaerobic and Lytic Anaerobic) containing 10 mL of naive fresh human blood. The effect of Fe 21 supplementation in shortening the time necessary for detection was evaluated using Bactec FX40, an automated alert incubator, routinely used for clinical blood culture diagnostics. All Fe 21 -supplemented aerobic cultures were distinguished as positive considerably earlier than nonsupplemented vials [fig_ref] TABLE 1: Effect of Fe 21 supplementation on F [/fig_ref] , as expected on Effect of iron supplementation on the live vaccine strain (LVS) growth rate. Bactec Plus Aerobic/F culture vials containing 10 mL of naive fresh human blood were spiked with the F. tularensis subsp. holarctica LVS at a final concentration of (A) 300 CFU/mL or (B) 3,000 CFU/mL. Fe 21 (100 mM) was added to vials, and then the vials were incubated at 37°C. Bacterial growth was determined by CFU counts 0, 24, 36, 48, 60, 72, 96, and 120 h following incubation. Nonsupplemented blood cultures were used as controls. Results are averages of triplicate counts 6 standard error of the mean (SEM) of results from two blood cultures for each group containing two individual blood donations. ***, P , 0.0001 of Fe 21 -supplemented cultures versus nonsupplemented controls according to two-way ANOVA using Sidak's multiple-comparison test.
the basis of the above-described elevation in bacterial growth rate promoted by Fe 21 addition . This observation was particularly significant considering that the growth rates observed at the low inoculum dose (30 CFU/mL) of F. tularensis in nonsupplemented Plus Aerobic vials were very low, virtually undetectable during the manufacturer-specified 5-day incubation protocol [fig_ref] TABLE 1: Effect of Fe 21 supplementation on F [/fig_ref] , while iron-supplemented vials were detected after 76 6 6 h. Prolonging the incubation of the Bactec FX40 for 2 additional days did not result in positive identification of the nonsupplemented cultures. Thus, Fe 21 supplementation dramatically shortened incubation periods for diagnosis and potential isolation of F. tularensis. Similarly, adding iron to Standard Aerobic and Plus Aerobic/F vials, inoculated with a bacterial dose of 300 CFU/mL, also resulted in significantly reduced detection times (36.7 6 0.3 and 60.4 6 9.6 h, respectively), compared to nonsupplemented cultures. The same effect was observed for the higher inoculum (3,000 CFU/mL), achieving a 39.7 6 6.9-h reduction for detection in Plus Aerobic vials and 22.8 6 5.9 h for Standard Aerobic vials. The difference between bacterial growth rates in the two Aerobic vial culture types probably stems from their formulation: Standard Aerobic vials contain hemin (see [fig_ref] TABLE 1: Effect of Fe 21 supplementation on F [/fig_ref] in the supplemental material). Of note, no growth was detected in Anaerobic blood culture vials, irrespective of iron addition even at the high inoculum of F. tularensis.
Positive cultures were confirmed by direct viable counting by plating at the time of detection. Bacterial counts were between 5 Â 10 7 and 2 Â 10 8 CFU/mL, irrespective of the original inoculum or blood culture type as previously reported.
Accelerated growth rates of the virulent F. tularensis subsp. tularensis in Fe 2+supplemented blood cultures. The highly beneficial effect of Fe 21 supplementation on bacterial growth was further demonstrated for the virulent F. tularensis subsp. tularensis SchuS4 strain. As depicted in [fig_ref] FIG 2: Superior bacterial growth in Fe 21 -supplemented aerobic vials [/fig_ref] , a significant increase of up to 6 orders of magnitude in bacterial counts was achieved in the Plus Aerobic iron-supplemented samples within 48 h of incubation for the low inoculum tested (300 CFU/mL). For the vials containing a high initial inoculum (3,000 CFU/mL), an increase of ;5 orders of magnitude was observed [fig_ref] FIG 2: Superior bacterial growth in Fe 21 -supplemented aerobic vials [/fig_ref]. These results are in agreement with the results obtained with F. tularensis LVS . Improvement of growth rate upon Fe 21 supplementation was also observed when SchuS4 bacteria were cultured in Standard Aerobic Bactec vials [fig_ref] FIG 2: Superior bacterial growth in Fe 21 -supplemented aerobic vials [/fig_ref]. In these cultures, an increase of 4 orders of magnitude in bacterial counts within 24 h of incubation was observed, irrespective of the initial inoculation concentration. This elevation reached approximately 6 orders of magnitude in bacterial counts after 32 h of incubation for the low inoculum tested [fig_ref] FIG 2: Superior bacterial growth in Fe 21 -supplemented aerobic vials [/fig_ref].
In the absence of Fe 21 , lower initial inoculums resulted in a delay in bacterial growth, mostly related to an elongated lag phase. Therefore, and according to the results presented in [fig_ref] TABLE 1: Effect of Fe 21 supplementation on F [/fig_ref] , it is conceivable to expect a more pronounced effect of Fe 21 supplementation in clinical, bacteremic cultures, where initial inocula are expected to be low. Generally, the bacteremia exhibited in more than 50% of the documented cases consists of less than 1 CFU/mL [bib_ref] Guidelines on blood cultures, Towns [/bib_ref]. We note that significant differences in the bacterial growth rates were observed in blood samples obtained from different donors (a total of six-two for cultures of the SchuS4 strain and four for those of LVS). These differences were more pronounced at low inocula, as represented by the high standard deviation values observed between blood samples [fig_ref] FIG 2: Superior bacterial growth in Fe 21 -supplemented aerobic vials [/fig_ref]. Since bacteremic blood will be probably sampled after initiation of the typical nutritional immunity defense response [bib_ref] The effects of Francisella tularensis infection on iron metabolism in man, Pekarek [/bib_ref] , characterized by free iron reduction, it is expected that these cultures will exhibit a low growth rate. Thus, real-life clinical diagnosis should highly benefit from Fe 21 supplementation. In the case of Standard Anaerobic cultures, only a marginal benefit could be attributed to Fe 21 supplementation and no effect at all could be detected when the SchuS4 bacteria were grown in the Lytic Anaerobic cultures, which do not support the F. tularensis growth [fig_ref] FIG 2: Superior bacterial growth in Fe 21 -supplemented aerobic vials [/fig_ref]. While elevation in bacterial growth in anaerobic condition was observed after 48 h, moderate and complete decline was observed following 96 h in Standard and Lytic vials, respectively [fig_ref] FIG 2: Superior bacterial growth in Fe 21 -supplemented aerobic vials [/fig_ref]. These results may explain the inability to achieve positive signals after 7 days in the LVS experiments [fig_ref] TABLE 1: Effect of Fe 21 supplementation on F [/fig_ref].
Finally, we verified the reliability of additional diagnostic tests, which are carried out with bacteria grown in blood cultures, for bacterial typing or for antibiotic susceptibility testing (AST) after supplementation with Fe 21 .
Rapid identification of bacteria by the matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) methodology has been extensively investigated and is frequently implemented in clinical setups as a reliable approach for identifying the bacteria present in positive blood cultures [bib_ref] The impact of blood culture identification by MALDI-TOF MS on the antimicrobial..., Bhavsar [/bib_ref] [bib_ref] Impact of MALDI-TOF-MS-based identification directly from positive blood cultures on patient management:..., Osthoff [/bib_ref]. Accordingly, we addressed the question whether modification of the standard blood cultures by Fe 21 supplementation may compromise the accuracy of MALDI-TOF identification of the F. tularensis pathogen. As depicted in [fig_ref] FIG 3: Direct detection of F [/fig_ref] , positive F. tularensis identification (average values of 1.9) was achieved by MALDI-TOF analysis of bacterial protein extracts derived from blood cultures, regardless of Fe 21 content. The MS spectra obtained from both cultures completely overlapped. Thus, modification of the media allows for significantly earlier detection of growth without affecting F. tularensis MALDI-TOF identification.
Since immunodiagnostics are considered reliable and specific methods for detection of tularemia [bib_ref] Simultaneous immunodetection of anthrax, plague, and tularemia from blood cultures by use..., Mechaly [/bib_ref] [bib_ref] Detection of Francisella tularensis-specific antibodies in patients with tularemia by a novel..., Sharma [/bib_ref] [bib_ref] Development of an immunoassay test system based on monoclonal antibodies and immunomagnetic..., Vetchinin [/bib_ref] , a routinely implemented immunoassay was evaluated in the modified cultures. Accordingly, supplemented and nonsupplemented Bactec Plus Aerobic blood cultures were spiked with bacteria of the SchuS4 virulent strain. Culture samples were collected in the course of 96 h of incubation. The levels of bacterial soluble antigens in the treated samples were determined by time-resolved fluorescence enzyme-linked immunosorbent assay (TRF ELISA) using polyclonal antibodies derived from rabbits immunized with the LVS strain [bib_ref] Simultaneous immunodetection of anthrax, plague, and tularemia from blood cultures by use..., Mechaly [/bib_ref]. Early detection of soluble F. tularensis antigens in supplemented media was observed within 36 h, with signal/noise (S/N) ratio values of 3 or 12 for low-inoculum (300-CFU/mL) or high-inoculum (3,000-CFU/mL) cultures, respectively [fig_ref] FIG 4: Direct immunodetection of F [/fig_ref]. Shown are results from MALDI-TOF MS analysis of F. tularensis bacterial protein extraction derived directly for supplemented and nonsupplemented cultures following the Bactec FX40 incubator's alert. The spectrum obtained from protein extracts derived from supplemented culture completely overlaps the spectrum obtained from nonsupplemented control cultures (red and blue spectra, respectively).
These results are in correlation with the accelerated growth rates promoted by Fe 21 supplementation [fig_ref] FIG 2: Superior bacterial growth in Fe 21 -supplemented aerobic vials [/fig_ref]. Maximal signals (assay saturation) were achieved 48 to 60 h postinoculation of supplemented cultures compared to undetectable levels in nonsupplemented cultures. We may envisage that in the future, using supplemented cultures may represent a means by which soluble antigen detection is implemented as a stand-alone assay for rapid tularemia identification in blood culture, potentially circumventing the need of plating for CFU quantification.
Direct antilipopolysaccharide (anti-LPS) immunofluorescent staining in formaldehydeinactivated blood culture samples was performed for F. tularensis diagnosis. The results described in [fig_ref] FIG 4: Direct immunodetection of F [/fig_ref] , establish that the sensitivity of the assay was actually improved, owing to the higher bacterial density afforded by the supplemented cultures. Finally, the supplemented blood cultures were used as a bacterial source for AST. Several antibiotics are recommended by the Centers for Disease Control and Prevention (CDC) for prophylaxis and treatment of tularemia, including bacteriostatic (e.g., doxycycline) and bactericidal (e.g., ciprofloxacin and gentamicin) drugs. Antibiogram Etest assays were performed in order to examine whether the Fe 21 supplementation affects the susceptibility of the bacteria to antibiotics. Therefore, the MIC of F. tularensis derived from supplemented compared to nonsupplemented blood cultures was determined. The lower density of bacteria in nonsupplemented Plus Aerobic vials following 72 h of incubation impaired the reliability of the test, while in supplemented media MIC values were accurately determined to be similar to the standard values [fig_ref] FIG 4: Direct immunodetection of F [/fig_ref]. In addition, no significant differences in MIC values were observed in Standard Aerobic vials, irrespective of Fe 21 supplementation. It is conceivable that due to the higher density of the bacteria, the Etest could potentially be performed earlier. Thus, the results clearly indicate the highly beneficial value of supplementing cultures for rapid diagnosis and isolation of F. tularensis.
Conclusions. Tularemia, a severe life-threatening disease, is caused by F. tularensis. The low infectious doses, coupled with environmental stability, qualified this pathogen as a Tier 1 biothreat agent. Prompt administration of effective antimicrobial therapy, essential for favorable patient prognosis, requires early pathogen detection, identification, and isolation. Clinical blood samples are relevant specimens for the isolation and diagnosis of F. tularensis. Specifically, blood cultures are considered to be the "gold standard" of sepsis diagnostics. Rapid and accurate identification of the slow-growth pathogen F. tularensis using blood culture methods requires improved nutritious culture media, continuous-monitoring devices, and prolonged incubation protocols.
In order to expedite diagnostic times, supplementation of commercial, routinely used blood culture vials with external Fe 21 was examined for the ability to enhance F. tularensis growth rates. This study consisted of comparison of a variety of blood culturing conditions generated by supplemented or nonmodified aerobic and anaerobic protocols. The data documented in this report show a significant improvement in bacterial growth rate in Fe 21 -supplemented media. These results were demonstrated in different blood donors, a variety of Bactec vial types, and bacterial subspecies (F. tularensis subsp. tularensis or subsp. holarctica). The study suggests that supplemented aerobic cultures represent the conditions of choice for rapid detection of the pathogen, on the basis of both bacterial growth as well as methods involving mass spectrometry, immunodetection, and identification of secreted bacterium-borne biomarkers.
In suspected tularemia cases (either upon arrival from areas of endemicity, contact with wild animals, tick bites, exclusion of more common etiologies of presenting signs, or in a suspected malicious use in bioterror scenarios) we recommend that additional supplemented aerobic vials should be included in the sampling protocol. In such cases, positive Bactec signals may prompt early downstream diagnostic assays for F. tularensis identification. In our previous study [bib_ref] An improvement in diagnostic blood culture conditions allows for the rapid detection..., Makdasi [/bib_ref] , we reported the accelerated bacterial growth of Yersinia pestis in blood cultures by the addition of nutritional supplements. This enabled a shortening of the doubling time, resulting in an increase of 5 orders of magnitude in bacterial loads within 24 h of incubation, allowing rapid detection and isolation of Y. pestis bacteria that grow slowly in vitro. Further research will establish the applicability of this metal ion supplementation for the improvement of the growth of other pathogens, possibly leading to a universal method for accelerating the pathogen's growth and identification.
# Materials and methods
F. tularensis strains. The F. tularensis subsp. holarctica live vaccine strain (LVS) (ATCC 29684) and F. tularensis subsp. tularensis strain SchuS4 were used in this study. The study was conducted in a biosafety level 3 (BSL3) facility in accordance with the biosafety guidelines of the Israel Institute for Biological Research (IIBR).
Culture media. FeSO 4 powder (Merck) was dissolved in double distilled water (DDW) and filtered (0.2-mm-pore filter), forming a 100 mM stock solution. Human blood samples were obtained from the National Blood Services, MDA, Israel, under MDA research permit 08-0290.
F. tularensis strains were grown on cysteine heart agar (CHA) plates (Becton Dickinson, France) enriched with 1% (wt/vol) hemoglobin (Becton, Dickinson, France) at 37°C for 72 h. Colonies were suspended in sterile phosphate-buffered saline (PBS) Biological Industries, Beth Haemek, Israel) and added at a defined concentration [bib_ref] An improvement in diagnostic blood culture conditions allows for the rapid detection..., Makdasi [/bib_ref] into naive fresh human blood. Inoculated blood samples (10 mL/vial) were inserted into four different types of BACTEC blood culture vials (Plus Aerobic/F, Standard Aerobic/ F, Standard Anaerobic/F, and Lytic Anaerobic/F (BD, Sparks, MD, USA). Blood cultures were supplemented with FeSO 4 at a final concentration of 100 mM. Nonsupplemented blood culture vials were used as a control. The inoculated blood culture vials were then shaken at 150 rpm at 37°C in a New Brunswick Scientific C76 water bath for the indicated time periods or at 35°C in Bactec FX40 (Becton Dickinson United Kingdom) until a blood culture alert level was reached. Initial CFU counts (time 0) were determined by plating 0.1-mL blood culture samples of serial 10-fold dilutions in duplicate, while for additional time points drop plating was performed by plating 10 mL of serial 10-fold dilutions in triplicate on CHA plates. The number of CFU was determined following 48 to 72 h of incubation at 37°C.
MIC determination. Etest strips (bioMérieux, France) of selected antimicrobial agents (ciprofloxacin, doxycycline, and gentamicin) were applied to an inoculated CHA surface derived from spiked blood cultures with or without supplements. Each of the strips contains dried antibiotic concentration gradients that are marked with a concentration scale. The plates were incubated for 48 to 72 h at 37°C, and MIC values (micrograms per milliliter) were read directly from the strips according to the manufacturers' instructions.
Immunofluorescence assay for F. tularensis detection. The immunfluorescence assay (IFA) allows visualization of the bacteria in different matrices. A 2-mL aliquot of inoculated BC with or without supplements was applied onto slides, air dried, and fixed in 100% acetone for 30 min. Alexa 488-conjugated anti-F. tularensis LPS (TL-1 monoclonal antibody) [bib_ref] Inhibition of Francisella tularensis phagocytosis using a novel anti-LPS scFv antibody fragment, Mechaly [/bib_ref] was applied on the spots and incubated at 37°C for 30 min. Visualization was carried out using an Axioscop 2 fluorescence microscope (Zeiss) equipped with a Â40 magnification objective and a top-mounted camera (DS-Fi3, Nikon). Images were recorded using NIS-Elements F4.6 software.
Detection of F. tularensis soluble antigens from inoculated blood culture by time-resolved fluorescence. Detection of F. tularensis soluble antigens was conducted using a 3-step sandwich ELISA, based on time-resolved fluorescence (TRF) using europium lanthanide. Microtiter plates were coated with 5 mg/mL polyclonal antibodies against F. tularensis (15) overnight at 4°C. Coated wells were blocked with 2% bovine serum albumin (BSA) for 2 h at 37°C. In order to detect soluble antigens, 0.3 mL of the inoculated BC (with or without supplements) was spun at 14,000 rpm for 5 min. The supernatant was then filtered (0.2-mm pore), and a 50-mL aliquot from the sterile supernatant was applied to the microtiter plates for 30 min at 37°C. Following washing steps (with PBS containing 0.05% Tween 20), biotinylated reporter antibodies (polyclonal anti-F. tularensis) were loaded for an additional 30 min at 37°C. After an additional wash step, streptavidin-europium ab275850 (Abcam) diluted 1:1,000 was loaded for 20 min. After a final wash, enhancement solution (1244-105, DELFIA [dissociation-enhanced lanthanide fluorescence immunoassay]; PerkinElmer, Waltham, MA, USA) was added, and the resulting signal was measured using a microplate reader (excitation of 340 nm and emission of 612 nm). The results were calculated as the ratios between the signal (S) measured for each sample compared to the signal measured with the assay run against antigen-free PBS (noise [N]). S/N ratios above 2 are considered positive.
Bacterial identification by MALDI-TOF MS. Direct bacterial identification by matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) was carried out using a MBT Sepsityper IVD kit (Bruker Daltonics GmbH, Bremen, Germany), according to the manufacturer's instructions. Briefly, 1 mL of blood culture was collected and transferred into a 1.5-mL tube and a 200-mL aliquot lysis buffer was added. The mix was centrifuged for 2 min at 14,000 rpm, and the pellet was washed with 1 mL washing buffer. The mixture was recentrifuged for 1 min at 14,000 rpm. The supernatant was discarded, and the pellet was dried. The sample's proteins were extracted using the standard EX method, according to the manufacturer's instructions. The samples were then identified by the Bruker MALDI-TOF MS instrument (Bruker Datonics, GmbH, Bremen, Germany).
Statistical analysis. Data were analyzed using GraphPad Prism5 software. Results are expressed as means 6 standard error. Statistical significance was determined by two-way analysis of variance (ANOVA) using Sidak's multiple-comparison test. A P value of #0.05 was considered to be significant.
# Supplemental material
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 0.1 MB.
[fig] FIG 2: Superior bacterial growth in Fe 21 -supplemented aerobic vials. Bactec Aerobic and Anaerobic vials (Standard Aerobic, Plus Aerobic, Standard Anaerobic, and Lytic Anaerobic) containing 10 mL of naive fresh human blood were inoculated with the virulent F. tularensis subsp. tularensis strain SchuS4 at (A and C) 300 and (B, D, and E) 3,000 CFU/mL. Vials were incubated at 37°C, and bacterial growth was determined by CFU counts over the course of incubation(24,32, 48, 60, 72, and 96 h). The role of supplemented 100 mM Fe 21 in Plus Aerobic (A and B), Standard Aerobic (C and D), and Standard Anaerobic and Lytic Anaerobic (E) cultures was evaluated. Nonsupplemented vials were used as a control. Results are averages of triplicate counts 6 SEM of results from duplicate blood cultures containing two individual blood donations in each group. ***, P , 0.0001, for Fe 21 -supplemented culture versus nonsupplemented control according to two-way ANOVA using Sidak's multiple-comparison test. [/fig]
[fig] FIG 3: Direct detection of F. tularensis from positive blood culture by MALDI-TOF-MS. Bactec Plus Aerobic/F culture vials containing 10 mL of naive fresh human blood were spiked with LVS. Vials were supplemented with Fe 21 (100 mM), while nonsupplemented blood cultures (BCs) were used as a control. [/fig]
[fig] FIG 4: Direct immunodetection of F. tularensis from inoculated blood culture. Blood cultures (Bactec Plus Aerobic/F culture vials) supplemented with Fe 21 and containing 10 mL of naive fresh human blood were spiked with the SchuS4 strain at (A) 300 CFU/mL and (B) 3,000 CFU/mL. Nonsupplemented blood cultures were used as a control. Vials were incubated at 37°C, and blood cultures were sampled over the course of incubation(24,32, 48, 60, 72, and 96 h) for the detection of soluble F. tularensis antigens by TRF. Results are presented as average 6 SEM of the signal/noise (S/N) ratio from duplicate blood cultures containing two individual blood donations in each group. Positive detection was determined as an S/N ratio of .2 (red dashed line). (C) Immunofluorescence assay (IFA) using Alexa Fluor 488-conjugated anti LPS MAbs for the detection of the F. tularensis directly from blood culture following 48 h of incubation. The scale bar represents 100 mm. (D) Etest assays to determine the MIC (micrograms per milliliter) of ciprofloxacin, doxycycline, and gentamicin were performed directly from supplemented blood cultures (following 72 h of incubation) compared to nonsupplemented blood cultures and standard test values documented by Johansson et al.(45). [/fig]
[table] TABLE 1: Effect of Fe 21 supplementation on F. tularensis early detection in spiked bacteremic blood culture using the Bactec FX40 diagnostic system Initial inoculum Avg ± SD time to detection (h) a [/table]
|
A PLSPM-Based Test Statistic for Detecting Gene-Gene Co-Association in Genome-Wide Association Study with Case-Control Design
For genome-wide association data analysis, two genes in any pathway, two SNPs in the two linked gene regions respectively or in the two linked exons respectively within one gene are often correlated with each other. We therefore proposed the concept of gene-gene co-association, which refers to the effects not only due to the traditional interaction under nearly independent condition but the correlation between two genes. Furthermore, we constructed a novel statistic for detecting gene-gene co-association based on Partial Least Squares Path Modeling (PLSPM). Through simulation, the relationship between traditional interaction and co-association was highlighted under three different types of co-association. Both simulation and real data analysis demonstrated that the proposed PLSPM-based statistic has better performance than single SNP-based logistic model, PCA-based logistic model, and other gene-based methods.
# Introduction
A Genome-wide Association Study (GWAS) typically tests whether certain SNPs have strong associations with predefined trait or disease by applying statistical methods. Hundreds of GWAS's for complex human diseases or traits were completed over the last decade. Nonetheless, the genetic variants discovered in GWAS's account for only a small proportion of the heritability of complex diseases [bib_ref] Sizing up human height variation, Visscher [/bib_ref] [bib_ref] Progress and promise of genome-wide association studies for human complex trait genetics, Stranger [/bib_ref]. One possible reason is that most GWAS analysis methods test the SNP-phenotype association individually, which has relatively low power in detecting multiple SNPs with small causal effects [bib_ref] Finding the missing heritability of complex diseases, Manolio [/bib_ref]. Additionally, in human body, genes tend to work collaboratively, especially within specific pathways or modules that are associated with certain diseases [bib_ref] Network biology: understanding the cell's functional organization, Barabási [/bib_ref] [bib_ref] The modular nature of genetic diseases, Oti [/bib_ref] [bib_ref] A pathway-based view of human diseases and disease relationships, Li [/bib_ref]. Therefore, we suspect that the missing proportion of heritability could be partly due to the ignorance of the joint effect of genes contributing to the disease or trait [bib_ref] Finding the missing heritability of complex diseases, Manolio [/bib_ref] [bib_ref] Human genetic variation and its contribution to complex traits, Frazer [/bib_ref]. Complex diseases often result from multiple genes' interplays within genetic networks, a general term that we used here to represent all kinds of networks defined on gene level, e.g., biological pathways, gene regulatory networks, and gene modules. The idea of multi-gene effect led to the development of genetic network-based analysis for GWAS [bib_ref] Pathway and network analysis with high-density allelic association data, Torkamani [/bib_ref] [bib_ref] Pathway and network-based analysis of genome-wide association studies in multiple sclerosis, Baranzini [/bib_ref] [bib_ref] Common variants conferring risk of schizophrenia: a pathway analysis of GWAS data, Jia [/bib_ref].
Network inference is a challenging task and proper methods should be proposed in constructing a priori topological structures for establishing genetic networks that contribute to diseases or traits of interest. A knowledge-based approach is commonly adopted for genetic network construction and inference [bib_ref] Gene prioritization through genomic data fusion, Aerts [/bib_ref] [bib_ref] CANDID: a flexible method for prioritizing candidate genes for complex human traits, Hutz [/bib_ref] [bib_ref] A knowledge-driven interaction analysis reveals potential neurodegenerative mechanism of multiple sclerosis susceptibility, Bush [/bib_ref] [bib_ref] Knowledge-Driven Analysis Identifies a Gene-Gene Interaction Affecting High-Density Lipoprotein Cholesterol Levels in..., Ma [/bib_ref] , but it is still underdeveloped in testing whether significant relationships between any two nodes in such networks exist. Theoretically, this can be solved by testing the joint effect of two genes. Traditionally, to detect gene-gene interaction, a product term is usually added to the logistic regression model Logit(P)~b 0 zb 1 Azb 2 Bzb 3 A|B, which implies a nearly independence assumption, at least not much correlation, between gene A and gene B for inferring the interaction measurement (b 3 ) [bib_ref] Confounding and effect-modification, Miettinen [/bib_ref] [bib_ref] Interaction: A word with two meanings creates confusion, Ahlbom [/bib_ref]. Nevertheless, one common sense is that the development of most diseases is attributed to the correlated genes in pathways. Another situation is that two SNPs usually locate in the two linked gene regions respectively, or in the two linked exons respectively within one gene. All these situations indicate that the two SNPs may have high correlation rather than independence (or low correlation). Therefore, the assumption of the above logistic model is rarely satisfied, and it will be inevitable to lose efficiency when high correlation existed between the two SNPs. In this paper, we proposed the concept of gene-gene co-association, which refers to the extent to which the joint effects of two genes differs from the main effects, not only due to the traditional interaction under the nearly independent condition but the correlation between two genes, while the part attributed to the correlation has usually been neglected in traditional interaction model using regression method. The proposed gene-gene co-association can be measured by the difference of the correlation between two genes within case and control groups without the independent assumption. This measurement refers to the co-association of two genes contributing to the disease or trait.
For genetic networks derived from GWAS, there are multiple variants (i.e. SNPs) within a gene region, where one single SNP in this region is inadequate to represent the overall effect of the whole gene on a disease. Previous studies suggested that gene-based analysis would allow the formation of pathways to interpret complex diseases and provide the functional bases of an association finding [bib_ref] Genome-wide gene and pathway analysis, Luo [/bib_ref]. Therefore, summarizing SNP effects at gene level to estimate gene-gene co-association appears to be an appealing strategy for constructing genetic networks. In our previous study [bib_ref] A gene-based method for detecting gene-gene coassociation in a case-control association study, Peng [/bib_ref] , a statistic called CCU for detecting gene-gene co-associations was proposed, which was constructed by the difference between the canonical correlation within case and control respectively. Since CCU statistic only uses the first canonical correlation coefficient, it may not be an inefficient estimator of gene-gene co-associations and may have very low power. Recently, another gene-based statistic was proposed to detect gene-gene interaction [bib_ref] Multivariate Detection of Gene-Gene Interactions, Rajapakse [/bib_ref] , which was built based on the difference of the covariance matrix within case and control respectively. Although both the two methods were severely affected by the high multicollinearity problem commonly encountered in GWASs, they motivated us to develop a new gene-based method to detect gene-gene co-association.
In this paper, we proposed a novel statistic to test the coassociation between two genes under a case-control design. The statistic was defined as the standardized difference of path coefficient for the gene pair between cases and controls based on Partial Least Squares Path Modeling (PLSPM) [bib_ref] PLS path modeling, Tenenhaus [/bib_ref] , which has been successfully used to detect associations in GWAS [bib_ref] Gene-based partial least-squares approaches for detecting rare variant associations with complex traits, Turkmen [/bib_ref] [bib_ref] A Latent Variable Partial Least Squares Path Modeling Approach to Regional Association..., Xue [/bib_ref]. To assess the performance of the proposed PLSPM-based statistic, simulation studies were conducted to evaluate its type I error rate and power. Its performance was also compared with single SNPbased logistic regression model [bib_ref] Genome-wide strategies for detecting multiple loci that influence complex diseases, Marchini [/bib_ref] , Principle Component Analysis(PCA)-based logistic regression model [bib_ref] A principal components regression approach to multilocus genetic association studies, Wang [/bib_ref] [bib_ref] Testing association between disease and multiple SNPs in a candidate gene, Gauderman [/bib_ref] , the CCU statistic [bib_ref] A gene-based method for detecting gene-gene coassociation in a case-control association study, Peng [/bib_ref] and the covariance-based statistic [bib_ref] Multivariate Detection of Gene-Gene Interactions, Rajapakse [/bib_ref]. Our method was then applied to real data analysis of Coronary atherosclerotic disease (CAD) association study. Both simulation and real data analysis suggested that the proposed PLSPM-based statistic has advantageous performances compared to other methods.
# Materials and methods
The Modeling Framework [fig_ref] Figure 1: PLSPM-based co-association model [/fig_ref] illustrates the framework for the PLSPM-based statistic between gene A and gene B. We denote the genotype data for gene A and gene B as
[formula] X D~( X D 1 ,X D 2 , Á Á Á ,X D p ) and Y D~( Y D 1 ,Y D 2 , Á Á Á ,Y D q ) respectively among cases, with X C~( X C 1 ,X C 2 , Á Á Á ,X C p ) [/formula]
and Y C~( Y C 1 ,Y C 2 , Á Á Á ,Y C q ) respectively among controls. Then, the path coefficient b D betweenX D and Y D obtained by PLSLM could be viewed as a measure of the correlation between genes A and B among cases. Similarlyb C measures the correlation between A and B among controls. In the algorithm of PLSPM, the path coefficient is calculated as the standardized regression coefficient of the two latent variables. This standardized path coefficient is equal to their correlation coefficient between the two latent variables. Therefore the arrow is merely used to reflect the structure and has no direction effect. No matter whether the path coefficients of the two genes are calculated from A to B or from B to A, technically the result remains the same under PLSPM.
We introduce D~b D {b C as an estimate of co-association between the two genes contributing to the disease, hence the proposed novel PLSPM-based test statistic can be defined as
[formula] U Ã~b D {b C ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Var(b D {b C ) p~D ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Var(D) pð1Þ [/formula]
where Var(b D ),Var(b C ),Var(D)denote the variance of b D ,b C , and D respectively. The framework of the PLSPM for gene-gene co-association resembles structural equation modeling (SEM) with three types of parameters defined: (1) latent variable scores (i.e., j 1 and j 2 ) defined as combinations of their manifest variables (SNPs within the gene); (2) path coefficients (b D and b C ) between the two latent variables in the case and control groups, which are counterparts of correlation coefficients in the SEM framework; (3) loadings (l 0 s) for each block that defines the relationship between the SNPs and their latent variables. In this paper, reflective measurement model was assumed in PLSPM to describe the relationship between SNPs and the latent variables. For estimation of the above parameters, the Lohmöller's PLSPM algorithmwas used. After centering and standardizing the manifest variables (i.e., variables in coding the genotype data such as the additive model) and giving initial values on weights w ij s, the algorithm is essentially an iterative procedure that works by alternating the outer and inner estimation steps. First, in the outer estimation step, we estimate the values of the latent variables j 1 and
[formula] j 2 by v 1~P p j~1 v 1j x j and v 2~P q j~1 v 2j y j , [/formula]
respectively. Then, in the inner estimation step, the endogenous latent variable j 2 is updated with value v 2~e12 v 1 , where e 12 is obtained via the centroid scheme by setting as '+1' or '21', i.e., the sign of the correlation between the outer estimates n 1 and n 2 . After the inner estimation step, weights are updated before moving to the next step: w 1j~c ov(x j ,v 1 ) and w 2j~c ov(y j ,v 2 ). Details of the algorithm and proof of its convergence is similar to the case of the two latent variables as provided in Chapter 2 of the book by Esposito. In GWAS data with case-control design, we separately applied the above algorithm for estimating the path coefficients for cases and for controls.
## Permutation test for the plspm-based statistic
To test whether genes A and B has co-association effect on a disease of interest, we conduct hypothesis testing with null hypothesis
[formula] H 0 : b D {b C~0 : [/formula]
Since PLSPM adopts nonparametric paradigm for estimating b D and b C and does not assume parametric distributional forms for the observed and latent variables, the asymptotic distribution of the path coefficients b D and b C is not available, hence we do not have a distribution available for UÃeither. To solve this problem, we adopted the strategy of a permutation test [bib_ref] Permutation tests for multiple loci affecting a quantitative character, Doerge [/bib_ref] , a common approach for nonparametric statistical inferences. To alleviate the high computation burden, a random permutation test for D~b D {b C was used to obtain p-value in testing the above H 0 . Rejection of the H 0 provides evidence in suggesting a significant co-association between the two genes contributing to the disease.
Significance test of path coefficients and loadings were furnished by bootstrap procedures conducted in the case and control groups, respectively [bib_ref] PLS path modeling, Tenenhaus [/bib_ref] [bib_ref] Bootstrap methods: another look at the jackknife, Efron [/bib_ref]. A large, pre-specified number of bootstrap samples (e.g., 1,000), each with the same number of subjects as the original sample, were generated via re-sampling with replacement. Parameter estimation was done for each bootstrap sample, whose path coefficients or loadings can be viewed as drawings from their sampling distributions. All bootstrap samples together provided empirical estimators for the standard error of each parameter.
## Simulation studies
Simulation studies were conducted to evaluate the performance of the proposed statistic for testing co-association between two genes. We simulated three scenarios by considering different types of co-association: Type I (co-association under nearly independent condition between gene A and gene B), Type II (co-association only caused by correlation between gene A and gene B), and Type III (co-association caused by both correlation and independent term A6B between gene A and gene B).
For scenario 1 (Type I co-association), we simulated two causal SNPs with interactions using software gs2.0 [bib_ref] Generating samples for association studies based on HapMap data, Li [/bib_ref]. The phased haplotype data of two gene regions TNRC9 and NEGR1 of CEU population were downloaded from the Hapmap website (http:// hapmap.ncbi.nlm.nih.gov/) and used to generate the simulation datasets. TNRC9 locates at Chr16:51074034…51089856, including [bib_ref] Pathway and network analysis with high-density allelic association data, Torkamani [/bib_ref] SNPs, and NEGR1 locates at Chr1:71705132…71712343, including 10 SNPs. The pair-wise linkage disequilibrium LD pattern of the two gene regions are shown in . For two causal SNPs, SNP1 from gene A and SNP2 from gene B, gs2.0 [bib_ref] Generating samples for association studies based on HapMap data, Li [/bib_ref] simulated genotypes and the binary phenotype according to logistic interaction model Logit(P)~b 0 zb 1 6(SNP1)zb 2 |(SNP2)zb 3 6 (SNP1|SNP2), where b 3 denoted the interaction effect of two SNPs. Furthermore, we specified different interaction odds ratios (ORs, exp (b 3 )) from 1.0 to 1.5 stepped by 0.1.
For scenario 2 (Type II co-association), to create the coassociation between linked genes under the condition of none interaction, we simulated two linked (correlated) causal SNPs only with marginal effects using software Hapgen2 [bib_ref] HAPGEN2: simulation of multiple disease SNPs, Su [/bib_ref] , and further specified co-association levels by the difference of the marginal effects of two causal SNPs. The phased haplotype data of two For scenario 3 (Type III co-association), again the same C6orf10 and BTNL2 genes was used in this scenario. Gs2.0 [bib_ref] Generating samples for association studies based on HapMap data, Li [/bib_ref] was first used to generate the dataset of Type I co-association, and Hapgen2 [bib_ref] HAPGEN2: simulation of multiple disease SNPs, Su [/bib_ref] for the dataset of Type II co-association. Finally, we mixed the above simulation data with the proportion 1:1 to create the scenario of Type III co-association. The model can be also expressed by Logit(P)~b 0 zb 1 |(SNP1)zb 2 6 (SNP2)zb 3 |(SNP1|SNP2), but the two genes are actually correlated rather than independent as defined in the model of scenario 1.
Current GWAS is still map-based rather than sequence-based, so association might predominantly be indirect. We therefore mainly deal with the indirect association. All the datasets were analyzed with the causal SNPs removed, permitting the effect of the causal SNPs to be detected indirectly. The genotype data were coded according to the additive genetic model [bib_ref] Genome-wide strategies for detecting multiple loci that influence complex diseases, Marchini [/bib_ref] [bib_ref] A complete enumeration and classification of two-locus disease models, Li [/bib_ref].
Under the null hypothesis H 0 (with exp (b 3 ) specified as 1.0 in scenario 1 and (exp (a 1 ),exp (a 2 )) specified as (1.0, 1.0) in scenario 2), 100,000 cases and 100,000 controls were generated and combined to form a hypothetical population from which case and control samples were randomly selected with different sample sizes [fig_ref] Figure 1: PLSPM-based co-association model [/fig_ref]. To examine the stability of the PLSPM-based statistic, we randomly sampled N individuals from the cases and controls for the calculation of the type I error rates under different nominal levels of 0.01, 0.05 and 0.1. A total of 1000 simulations were repeated for each sample size.
To highlight the advantages of our proposed PLSPM-based statistic, four existed methods were used to compare with our method. The first was traditional single SNP-based logistic model. For each simulation, all pair-wise SNPs from genes A and B and their product terms were defined as the independent variables in the single SNP-based logistic regression model [bib_ref] Genome-wide strategies for detecting multiple loci that influence complex diseases, Marchini [/bib_ref]. We considered each of the pair-wise interactions separately, selecting the most significant one (smallest p-value). Significance levels are determined using permutations to adjust the multiple testing. The second was PCA-based logistic model, which was constructed by Logit(P)~b 0 zb 1 6Z 1 zb 2 |Z 2 zb 3 |(Z 1 |Z 2 ), whereZ 1 and Z 2 denoted the first principle component score of gene A and gene B respectively, and b 3 denoted the interaction effect of two genes. The third was the CCU statistic proposed in our previous study, and the last was the recently proposed covariance-based statistic [bib_ref] Multivariate Detection of Gene-Gene Interactions, Rajapakse [/bib_ref].
For scenarios 1 and 2, under the alternative hypothesisH 1, the performance of four different methods (PLSPM-based statistic, CCU statistic [bib_ref] A gene-based method for detecting gene-gene coassociation in a case-control association study, Peng [/bib_ref] , single SNP-based [bib_ref] Genome-wide strategies for detecting multiple loci that influence complex diseases, Marchini [/bib_ref] and PCA-based [bib_ref] A principal components regression approach to multilocus genetic association studies, Wang [/bib_ref] [bib_ref] Testing association between disease and multiple SNPs in a candidate gene, Gauderman [/bib_ref] logistic model) were assessed 1) at different sample sizes under fixed OR; 2) at different co-association levels under fixed sample sizes; and 3) at different minor allele frequency (MAF) of causal SNPs from two genes under fixed OR and fixed sample size to evaluate the performance with various linkage disequilibrium (LD) patterns. For scenario 3, under the alternative hypothesisH 1 , the performance of the four methods were assessed at different sample sizes with fixed co-association level and assessed at different co-association levels with fixed sample sizes. In addition, we compared our PLSPM-based statistic with the covariance-based statistic [bib_ref] Multivariate Detection of Gene-Gene Interactions, Rajapakse [/bib_ref] by repeating 1) and 2) under scenario 1 and 2.
## Application
The proposed PLSPM-based statistic was also applied to a real dataset. The data consisted of genotypes data from three candidate susceptibility genes (LRP5, LRP6, PCSK9), all belonging to the lipid metabolism pathway associated with Coronary atherosclerotic disease (CAD). The dataset contained samples from 498 CAD cases and 509 controls, and the genotyping was conducted by Qilu Hospital of Shandong University in China. The three genes (LRP5, LRP6, PCSK9) were typed with two, nine, three SNPs respectively. All the four methods were conducted in detecting gene-gene co-association contributing to CAD.
# Results
# Simulation results
Type I error rate. [fig_ref] Table 1: Type I error rates of the PLSPM-based statistic in different scenarios [/fig_ref] shows the estimated type I error rates of the PLSPM-based statistic under different nominal levels in both scenario1 and 2. It reveals that the type I error rates of the proposed statistics are close to nominal levels (0.01, 0.05, 0.1) as a function of sample sizes.
Power. [fig_ref] Figure 3: The power of the four methods under different sample sizes [/fig_ref] shows the performances of the four methods under different sample sizes given fixed co-association level for scenarios 1, 2 and 3. It indicates that the powers of the four methods all increase monotonically with sample size in scenarios 1 and 3 [fig_ref] Figure 3: The power of the four methods under different sample sizes [/fig_ref] , while the single SNP-based [bib_ref] Genome-wide strategies for detecting multiple loci that influence complex diseases, Marchini [/bib_ref] and PCA-based [bib_ref] A principal components regression approach to multilocus genetic association studies, Wang [/bib_ref] [bib_ref] Testing association between disease and multiple SNPs in a candidate gene, Gauderman [/bib_ref] logistic model lost their power in detecting gene-gene Type II co-association [fig_ref] Figure 3: The power of the four methods under different sample sizes [/fig_ref]. Obviously, the power of the PLSPM-based statistic is higher than that of the CCU statistic [bib_ref] A gene-based method for detecting gene-gene coassociation in a case-control association study, Peng [/bib_ref]. Only in scenario 1, the single SNP-based logistic model has slight higher power when sample size is larger than 3000, and PCA-based logistic regression model [bib_ref] A principal components regression approach to multilocus genetic association studies, Wang [/bib_ref] [bib_ref] Testing association between disease and multiple SNPs in a candidate gene, Gauderman [/bib_ref] has comparable power with PLSPM-based statistic [fig_ref] Figure 3: The power of the four methods under different sample sizes [/fig_ref] , while they has less power for the other two scenarios. [fig_ref] Figure 4: The power of four methods under different co-association levels [/fig_ref] depicts the power under different co-association levels in the three scenarios. For the case of Type I co-association in scenario 1, the power increases monotonically with the interaction ORs for all the four methods [fig_ref] Figure 4: The power of four methods under different co-association levels [/fig_ref]. In scenario 2, the power of the PLSPM-based statistic and that of the CCU statistic [bib_ref] A gene-based method for detecting gene-gene coassociation in a case-control association study, Peng [/bib_ref] both increases monotonically along with the difference between marginal ORs of the two causal SNPs [fig_ref] Figure 4: The power of four methods under different co-association levels [/fig_ref]. As for scenario 3, the PLSPM-based statistic has the highest power, followed by the two logistic regression models, and then by the CCU statistic [bib_ref] A gene-based method for detecting gene-gene coassociation in a case-control association study, Peng [/bib_ref]. [fig_ref] Figure 5: The power of the four methods under different causal SNPs [/fig_ref] illustrates the power of the four methods under different MAF or LD patterns. For both type I and type II coassociation, PLSPM-based statistic outperforms all other methods with the highest testing power, although the powers of the four methods vary heavily under different MAF or LD patterns. It is notable that the logistic regression models do not work for scenario 2. Specifically, the power for detecting co-association between the 8th SNP within gene A and 8th SNP gene B is quite low for all the four methods due to the low MAF (0.08) of 8th SNP within gene B [fig_ref] Figure 5: The power of the four methods under different causal SNPs [/fig_ref]. This indicates that the proposed PLSPM-based statistic lose its power in detecting rare variation.
One reviewer suggested us compare our proposed PLSPMbased statistic with the recently proposed covariance-based statistic [bib_ref] Multivariate Detection of Gene-Gene Interactions, Rajapakse [/bib_ref]. As the covariance-based statistic [bib_ref] Multivariate Detection of Gene-Gene Interactions, Rajapakse [/bib_ref] didn't work in our simulated data due to that the matrix W defined in their method was not invertible resulted from high collinearity between SNPs, we just attempted to do the calculations using the Moore-Penrose generalized inverse. The results are shown in the Tables S1-S4 in Supplementary Materials S1. In scenario 1, it indicates that the powers of the two methods are comparable in detecting Type I coassociation, and the PLSPM-based method has slight advantage with a lower odds ratio which is more common for SNP data. While in scenario 2, the covariance-based statistic [bib_ref] Multivariate Detection of Gene-Gene Interactions, Rajapakse [/bib_ref] has a higher power in detecting the gene-gene Type II co-association. [fig_ref] Table 2: The results of gene-gene co-association contributing to CAD within the lipid metabolism... [/fig_ref] shows the results of a gene-gene co-association test between three genes that are potentially contributing to CAD within the lipid metabolism pathway using the PLSPM-based statistic, CCU statistic [bib_ref] A gene-based method for detecting gene-gene coassociation in a case-control association study, Peng [/bib_ref] , single SNP-based logistic model [bib_ref] Genome-wide strategies for detecting multiple loci that influence complex diseases, Marchini [/bib_ref] and PCA-based logistic model [bib_ref] A principal components regression approach to multilocus genetic association studies, Wang [/bib_ref] [bib_ref] Testing association between disease and multiple SNPs in a candidate gene, Gauderman [/bib_ref]. The co-association between LRP5 and LRP6 is statistically significant (a~0:05) detected only by PLSPM-based statistic and not by the other three methods.
# Application result
# Discussion
Many methods have been developed for constructing the genetic network, such as Bayesian network [bib_ref] bNEAT: a Bayesian network method for detecting epistatic interactions in genome-wide association..., Han [/bib_ref] , Gaussian network [bib_ref] A graphical model approach for inferring large-scale networks integrating gene expression and..., Scott [/bib_ref] , and Boolean network [bib_ref] Analysis and Practical Guideline of Constraint-Based Boolean Method in Genetic Network Inference, Saithong [/bib_ref]. In these genetic networks for GWAS with case-control design, an 'edge' between any two nodes indicates that the joint effects of the two genes on target trait or phenotype would be different between controls and cases, which implies the co-association (or interaction) between the two genes. Various algorithms have been developed to learn the topological structure (i.e., links between the nodes) from GWAS data. In this paper, we proposed a novel statistic within the framework of PLSPM, which can be used to test on the existence of gene-gene co-association, i.e., whether an edge between any two genes would exist. It provides a preliminary or prior tool as a first step in constructing or learning genetic network structures given a GWAS dataset with case-control design.
The concept of gene-gene co-association was proposed in our previous paper [bib_ref] A gene-based method for detecting gene-gene coassociation in a case-control association study, Peng [/bib_ref]. It can be measured by the difference of the gene-gene correlation between the case and control groups without employing the nearly independence (at least not much correlation) assumption. Several strategies could be used to detect the gene-gene co-association, though some of these methods still didn't jump out of the traditional concept of gene-gene interaction [bib_ref] Confounding and effect-modification, Miettinen [/bib_ref] [bib_ref] Interaction: A word with two meanings creates confusion, Ahlbom [/bib_ref]. In this paper, the proposed PLSPM-based statistic clarified the concept and the measurement of gene-gene coassociation, which refers to the effects not only due to the traditional interaction under nearly independent condition but the correlation between two genes.
Through simulation, the relationship between traditional interaction and co-association was highlighted. The scope of co- association includes the following three scenarios: co-association under nearly independent condition between gene A and gene B [fig_ref] Figure 3: The power of the four methods under different sample sizes [/fig_ref] , co-association only caused by correlation between gene A and gene B [fig_ref] Figure 3: The power of the four methods under different sample sizes [/fig_ref] and co-association caused by both correlation and independent term A6B between gene A and gene B [fig_ref] Figure 3: The power of the four methods under different sample sizes [/fig_ref]. Currently, simulation and real data analysis demonstrated that the proposed PLSPM-based statistic is stable and has higher power than CCU statistic [bib_ref] A gene-based method for detecting gene-gene coassociation in a case-control association study, Peng [/bib_ref] , single SNP-based logistic model [bib_ref] Genome-wide strategies for detecting multiple loci that influence complex diseases, Marchini [/bib_ref] and PCA-based logistic model [bib_ref] A principal components regression approach to multilocus genetic association studies, Wang [/bib_ref] [bib_ref] Testing association between disease and multiple SNPs in a candidate gene, Gauderman [/bib_ref] (see results in [fig_ref] Table 1: Type I error rates of the PLSPM-based statistic in different scenarios [/fig_ref] , [fig_ref] Figure 3: The power of the four methods under different sample sizes [/fig_ref] to [fig_ref] Figure 5: The power of the four methods under different causal SNPs [/fig_ref] and [fig_ref] Table 2: The results of gene-gene co-association contributing to CAD within the lipid metabolism... [/fig_ref]. In addition, the performance of PLSPM-based statistic compared with recently proposed covariance-based statistic [bib_ref] Multivariate Detection of Gene-Gene Interactions, Rajapakse [/bib_ref] indicated that the powers of the two methods are comparable in detecting gene-gene co-association, while the former can deal with the high multicollinearity problem between SNPs (see Supplementary Materials S1).
Observing that two genes in any pathway, two SNPs usually locate in the two linked gene regions respectively or in the two linked exons respectively within one gene are often correlated with each other, we think it is meaningful to fabricate the term, genegene co-association. In Peng et al [bib_ref] A gene-based method for detecting gene-gene coassociation in a case-control association study, Peng [/bib_ref] , CCU statistic was developed for estimating and testing such a gene-gene coassociation within the framework of canonical correlation analysis. Nonetheless, since the CCU statistic [bib_ref] A gene-based method for detecting gene-gene coassociation in a case-control association study, Peng [/bib_ref] was calculated only by the first canonical correlation coefficient, it may lose power in the testing. Our simulation studies confirmed that the novel PLSPMbased statistic had more power than the CCU statistic [bib_ref] A gene-based method for detecting gene-gene coassociation in a case-control association study, Peng [/bib_ref] (see evidence from . Although the power of PLSPMbased statistic is similar as PCA-based logistic model [bib_ref] A principal components regression approach to multilocus genetic association studies, Wang [/bib_ref] [bib_ref] Testing association between disease and multiple SNPs in a candidate gene, Gauderman [/bib_ref] for the case of Type I co-association [fig_ref] Figure 3: The power of the four methods under different sample sizes [/fig_ref] , the former still has a superior performance when the logistic model lose its power for the case of Type II co-association [fig_ref] Figure 3: The power of the four methods under different sample sizes [/fig_ref]. The logistic regression model methods do not work at all because it cannot theoretically handle the scenario of Type II co-association; PLSPM-based statistic outperforms PCA-based logistic regression model [bib_ref] A principal components regression approach to multilocus genetic association studies, Wang [/bib_ref] [bib_ref] Testing association between disease and multiple SNPs in a candidate gene, Gauderman [/bib_ref] because of the advantage of PLSPM method [bib_ref] PLS path modeling, Tenenhaus [/bib_ref] ; PLSPM-based statistic outperforms single SNP-based logistic model [bib_ref] A principal components regression approach to multilocus genetic association studies, Wang [/bib_ref] [bib_ref] Testing association between disease and multiple SNPs in a candidate gene, Gauderman [/bib_ref] since the causal SNPs were excluded and the PLSPM-based statistic reflects the joint effects of multiple SNPs in the genes or regions. Also, the performance of PLSPMbased statistic are comparable with the recently proposed covariance-based statistic [bib_ref] Multivariate Detection of Gene-Gene Interactions, Rajapakse [/bib_ref] , while it is not affected by high multicollinearity between SNPs (see Supplementary Materials S1). The proposed method for detecting gene-gene co-association was developed based on PLSPM. An advantage of the algorithms is that they are robust to the multicollinearity problem, which is commonly encountered in GWAS data because of strong linkage disequilibrium between SNPs [bib_ref] Gene set analysis of SNP data: benefits, challenges, and future directions, Fridley [/bib_ref] [bib_ref] Linkage disequilibrium in the human genome, Reich [/bib_ref] [bib_ref] Linkage disequilibrium and the mapping of complex human traits, Weiss [/bib_ref]. Compared to covariancebased Structural Equation Model (SEM) and other parametric modeling methods, PLSPM is a ''soft modeling'' approach, requiring fewer distributional assumptions, and the variables studied can be numerical, ordinal, or nominal, hence no normality assumptions are needed. This is a very appealing feature for SNP data in genetic analysis and PLSPM has been successfully applied in genome wide association studies. We want to admit that although the proposed PLSPM-based approach has indicated numerous benefits, it has some limitations. Firstly, the current PLSPM-based statistic is based on a random permutation test due to the lack of its asymptotic distribution. Parametric test will be in great demand in future studies. Secondly, the PLSPM-based statistic still lacks efficiency when dealing with rare variation situation (see evidence in [fig_ref] Figure 5: The power of the four methods under different causal SNPs [/fig_ref].
## Supporting information
Supplementary Materials S1 [fig_ref] Table 1: Type I error rates of the PLSPM-based statistic in different scenarios [/fig_ref]. The power of the two methods for detecting Type I co-association under different sample sizes. [fig_ref] Table 2: The results of gene-gene co-association contributing to CAD within the lipid metabolism... [/fig_ref]. The power of the two methods for detecting Type I co-association under different interaction odds ratios. . The power of the two methods for detecting Type II co-association under different sample sizes. . The power of the two methods for detecting Type II co-association under different pairs of marginal odds ratios. (DOC)
[fig] Figure 1: PLSPM-based co-association model. doi:10.1371/journal.pone.0062129.g001 [/fig]
[fig] Figure 3: The power of the four methods under different sample sizes. Note: In Figure 3a, rs189851 (MAF = 0.43) in gene TNRC9 and rs12125823(MAF = 0.44) in gene NEGR1 were defined as causal SNPs with their interaction odds ratio fixed at 1.3. In Figure 3b, rs926594 (MAF = 0.46) in gene C6orf10 and rs2294880 (MAF = 0.45) in gene BTNL2 were defined as causal SNPs with their marginal odds ratio fixed at 1.3 and 1.7 respectively. In Figure 3c, mixed dataset with proportion 1:1 were generated by the same causal SNPs in Figure 3b, with interaction odds ratio 1.3 for Type I co-association and marginal effect odds ratio 1.3 and 1.7 for Type II co-association. doi:10.1371/journal.pone.0062129.g003 [/fig]
[fig] Figure 4: The power of four methods under different co-association levels. Note: In Figure 4a, rs189851 (MAF = 0.43) in gene TNRC9 and rs12125823(MAF = 0.44) in gene NEGR1 were defined as causal SNPs with sample size fixed at 2000. In Figure 4b, rs926594 (MAF = 0.46) in gene C6orf10 and rs2294880 (MAF = 0.45) in gene BTNL2 were defined as causal SNPs with sample size fixed at 4000. In Figure 4c, mixed datasets with proportion 1:1 were generated by the same causal SNPs inFigure 4bwith sample size fixed at 2000, and the horizontal axis denotes different interaction odds ratios for Type I co-association and marginal effect odds ratios for Type II co-association. doi:10.1371/journal.pone.0062129.g004 [/fig]
[fig] Figure 5: The power of the four methods under different causal SNPs. Note: The horizontal axis denotes the positions of the causal SNPs in the corresponding genes (Ai:Bi denotes the causal SNPs are ith SNP in gene A and ith SNP in gene B). InFigure 5a, A,B denotes gene TNRC9 and NEGR1 with causal SNPs' interaction odds ratio fixed at 1.3. InFigure 5b, A,B denotes gene C6orf10 and BTNL2 with causal SNPs' marginal effect odds ratios fixed at 1.3 and 1.7. Results for other pair-wise SNPs are qualitatively similar, hence not shown in theFigure. [/fig]
[table] Table 1: Type I error rates of the PLSPM-based statistic in different scenarios. [/table]
[table] Table 2: The results of gene-gene co-association contributing to CAD within the lipid metabolism pathway using four different methods.PLSPM-based statistic CCU PCA-based logistic model SNP-based logistic model *Only the SNP pairs with the smallest P-value were presented. doi:10.1371/journal.pone.0062129.t002 [/table]
|
State-of-the-Science Review of Non-Chemical Stressors Found in a Child’s Social Environment
Background: Children are exposed to chemical and non-chemical stressors from their built, natural, and social environments. Research is needed to advance our scientific understanding of non-chemical stressors, evaluate how they alter the biological response to a chemical stressor, and determine how they impact children's health and well-being. To do this, we conducted a state-of-the-science review of non-chemical stressors found in a child's social environment. Methods: Studies eligible for inclusion in this review were identified through a search of the peer-reviewed literature using PubMed and PsycINFO. Combinations of words associated with non-chemical stressors and children were used to form search strings. Filters were used to limit the search to studies published in peer-reviewed journals from 2000-2016 and written in English. Publications found using the search strings and filters went through two rounds of screening. Results: A total of 146 studies met the inclusion criteria. From these studies, 245 non-chemical stressors were evaluated. The non-chemical stressors were then organized into 13 general topic areas: acculturation, adverse childhood experiences, economic, education, family dynamics, food, greenspace, neighborhood, social, stress, urbanicity, violence, and other. Additional information on health outcomes, studies evaluating both chemical and non-chemical stressors, and animal studies are provided. This review provides evidence that non-chemical stressors found in a child's social environment do influence their health and well-being in both beneficial (e.g., salutatory effects of greenspace and social support) and adverse (e.g., poor relationships between health and selected non-chemical stressors such as economics, educational attainment, exposure to violence, stress) ways. Conclusions: This literature review identified a paucity of studies addressing the combined effects of chemical and non-chemical stressors and children's health and well-being. This literature review was further complicated by inconsistencies in terminology, methodologies, and the value of non-chemical stressor research in different scientific disciplines. Despite these limitations, this review showed the importance of considering non-chemical stressors from a child's social environment when addressing children's environmental health considerations. children's health and well-being. Children may be more vulnerable to the combined effects of chemical and non-chemical stressors due to their physiology (e.g., metabolic rate, surface-area-to-body-weight) and lifestage-specific activities and behaviors (such as mouthing, crawling or playing close to the ground, playing outdoors) when compared to adults. The combination of their continual physiological development, the nature of their motor, cognitive, and life course developments, and reliance on a caretaker (lack of independence) are all reasons for a child's increased vulnerability to the combined effects of chemical and non-chemical stressors[1][2][3][4][5].As the exposure assessment paradigm has shifted from single chemicals to multiple chemicals to mixtures, the scientific community has realized the need to include non-chemical stressors in studies evaluating chemical stressors[6][7][8][9][10][11][12]. In 2011, a workshop at the Society for Toxicology annual meeting resulted in work addressing the inclusion of non-chemical stressors into cumulative risk assessment [13] followed by additional workshops on cumulative risk and multiple exposures[11]. Previous work on non-chemical stressors and public health has typically excluded considerations of chemical exposures[14][15][16][17][18][19].Research on non-chemical stressors is needed to advance our scientific understanding of non-chemical stressors, and how they alter the biological response to a chemical stressor, in regard to their impact on children's health and well-being. Studies have evaluated how non-chemical stressors directly affect animal health and well-being; for instance, a study that looked at the health effects of cows from crowding and sleep deprivation[20]. Other studies were designed to look at indirect effects on health and well-being, such as a study that determined if exposure to greenspace influenced physical activity[21], thus, influencing health. There are very few studies that address outcomes from the combination of non-chemical and chemical stressors, or the interaction effect(s) of non-chemical and chemical stressors.The primary objective of this research is to conduct a state-of-the-science review of non-chemical stressors found in a child's social environment that might impact individual health. This review synthesizes many studies, included through predetermined criteria, into general topic areas. The secondary objective of this review is to identify which topic areas have begun to research the relationships between chemical and non-chemical stressors, and how the relationships may impact health outcomes. This review will help to identify possible gaps in the science of non-chemical stressors and inform future research design.Materials and MethodsLiterature SearchStudies eligible for inclusion in this review were identified through a search of the peer-reviewed literature using PubMed and PsycINFO. Combinations of words associated with non-chemical stressors (e.g., non chemical AND (stressor OR factor)) AND children were used to form search strings. Initial search strings included the general terms non-chemical stressor OR non-chemical factor AND health. Following the initial search, additional terms (i.e., psychosocial or social determinants or social environment) AND child* AND health were added to the search strings. Finally, search strings included specific non-chemical stressors (e.g., socioeconomic or exposure to violence or social support or acculturation or food access or overcrowd or urbanization or greenspace) AND child* AND health. Filters were used to limit the search to studies published in peer-reviewed journals from 2000-2016 and written in English.Study SelectionThe Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA)[22]method was used for the literature search phase(Figure 1). Publications found using the search strings and filters went through two rounds of screening. During the first round, titles and abstracts were screened and selected based on their relevance and whether they met the inclusion criteria listed below. In the
# Introduction
Children are exposed to both chemical and non-chemical stressors from their built, natural, and social environments. Non-chemical stressors are factors found in these environments, including physical factors (e.g., noise, temperature, and humidity) and psychosocial factors (e.g., poor diet and illicit drug use). The social environment can include elements such as exposure to violence, greenspace access and use, social support, access to education, psychological stress, family income, socioeconomic status, neighborhood quality, acculturation, and food sources. Current research suggests that non-chemical stressors modify the biological response to chemical stressors, thereby impacting Int. J. Environ. Res. Public Health 2019,, 4417 3 of 27 second round, full text articles of selected abstracts were retrieved and reviewed. For articles that met the inclusion criteria, the references were reviewed for additional relevant citations.
## Inclusion criteria
To be included in this review, studies needed to meet the following criteria. (1) experimental or observational studies;study was designed to evaluate a non-chemical stressor found in the social environment of the participants; (3) study evaluated a non-chemical stressor as a variable that impacts health; (4) study cohort was located in the United States, Canada, or Europe;non-chemical stressor classification used the Tulve et al.conceptual framework, which does not include intrinsic biological factors, biological pathogens, or activities/behaviors.
## Figure 1. diagram of record selection, eligibility, and inclusion; adapted from preferred reporting
Items for Systematic Reviews and Meta-Analysis (PRISMA).
## Data extraction and synthesis
Information extracted from the eligible studies included: article identifiers (author(s), year), study characteristics, subject demographics, non-chemical stressors, chemical stressors, outcome measures, and health impact results. After this information was extracted, non-chemical stressor
## Inclusion criteria
To be included in this review, studies needed to meet the following criteria.
(1) experimental or observational studies;
(2) study was designed to evaluate a non-chemical stressor found in the social environment of the participants;
(3) study evaluated a non-chemical stressor as a variable that impacts health; (4) study cohort was located in the United States, Canada, or Europe;non-chemical stressor classification used the Tulve et al.conceptual framework, which does not include intrinsic biological factors, biological pathogens, or activities/behaviors.
## Data extraction and synthesis
Information extracted from the eligible studies included: article identifiers (author(s), year), study characteristics, subject demographics, non-chemical stressors, chemical stressors, outcome measures, and health impact results. After this information was extracted, non-chemical stressor variables from articles were identified and categorized into general topic areas for further analysis. Mutual variables from articles were then quantified, and descriptive analyses were completed on the extracted information.summarizes the characteristics of the studies included in this review.
# Results
## Results of general topics
A total of 146 studies met the inclusion criteria. From these studies, a total of 245 non-chemical stressors were evaluated. The non-chemical stressors were organized into 13 general topic areas: acculturation, adverse childhood experiences (ACE), economic, education, family dynamics, food, greenspace, neighborhood, social, stress, urbanicity, violence, and other (other included several variables that were seen once or twice that were considered non-chemical stressors but did not fit into another general topic area) (Supplemental Materials;.displays the general topic categories resulting from the synthesis of the studies included in this review.
## Acculturation
The results describing the relationships between acculturation (n = 31) and health are diverse. Roughly 94% of the studies reported that acculturation impacted health. Of those, almost 38% found that higher acculturation was related to beneficial health impacts, whereas 48% found that higher acculturation was linked to adverse health impacts. Results also indicated that 14% of the studies had mixed resultsand 6% had either non-significant resultsor suggestive results from qualitative analyses of focus groups. Interestingly, culture of origin did not offer an explanation for the variation on whether the statistical relationship was positive or negative. Of the 31 studies that considered the health impact from acculturation measures, none jointly investigated a chemical exposure.
## Adverse childhood experiences
Eighteen articles studied adverse childhood experiences and all reported negative impacts on health and well-being. These studies included violence, as well as other types of adverse childhood experiences (e.g., divorce, parental mental illness). In 1995, Kaiser Permanente and the Centers for Disease Control and Prevention (CDC) jointly conducted one of the largest investigations of childhood abuse and neglect and later life health and well-being, known as the CDC-Kaiser ACE Study. The CDC-Kaiser ACE study included a retrospective survey and longitudinal tracking of study participants that evaluated health and well-being outcomes against multiple childhood experience variables, such as physical abuse, neglect, substance abuse, or living with a person with mental illness in the home. The CDC-Kaiser ACE study provided strong evidence that a significant
## Acculturation
The results describing the relationships between acculturation (n = 31) and health are diverse. Roughly 94% of the studies reported that acculturation impacted health. Of those, almost 38% found that higher acculturation was related to beneficial health impacts, whereas 48% found that higher acculturation was linked to adverse health impacts. Results also indicated that 14% of the studies had mixed resultsand 6% had either non-significant resultsor suggestive results from qualitative analyses of focus groups. Interestingly, culture of origin did not offer an explanation for the variation on whether the statistical relationship was positive or negative. Of the 31 studies that considered the health impact from acculturation measures, none jointly investigated a chemical exposure.
## Adverse childhood experiences
Eighteen articles studied adverse childhood experiences and all reported negative impacts on health and well-being. These studies included violence, as well as other types of adverse childhood experiences (e.g., divorce, parental mental illness). In 1995, Kaiser Permanente and the Centers for Disease Control and Prevention (CDC) jointly conducted one of the largest investigations of childhood abuse and neglect and later life health and well-being, known as the CDC-Kaiser ACE Study. The CDC-Kaiser ACE study included a retrospective survey and longitudinal tracking of study participants that evaluated health and well-being outcomes against multiple childhood experience variables, such as physical abuse, neglect, substance abuse, or living with a person with mental illness in the home. The CDC-Kaiser ACE study provided strong evidence that a significant positive relationship existed between childhood exposure to multiple adverse childhood experiences (e.g., non-chemical stressors) and adult health and well-being.
Certain studies identified a relationship between exposure to violence or adverse experiences and health impacts that mimicked a dose-response relationship. Those studies offered evidence that the more adverse experiences a participant had, or the more types of violence experienced (physical, sexual, or physical and sexual) or more time exposed to adverse experiences or violence, increased the negative health impact resulting in risky behavior, general adverse health, telomere erosion, comorbid depression, chronic pain, hypertension, neurological responses, and heart disease. Of the eighteen studies on adverse experiences by children included in this review, only one (<6%) investigated the relationship between a chemical stressor and a non-chemical stressor and their combined impact on a child's health. Stein et al.identified not only the negative associations between total adversities and child cognition, but also that associations were stronger when higher levels of organophosphate metabolites were higher, with gender variations.
## Economic
Previous research offers evidence that socioeconomic disadvantages (e.g., lower income, less educational attainment) are linked to poorer general health, higher morbidity and mortality rates, and more susceptibility to chemical exposures. Additionally, research has shown that socioeconomic and sociodemographic characteristics affect choices that result in increased chemical exposures.
For this review, economic measures included wealth, income, disposable income, or an index such as socioeconomic status (SES) or position (SEP), or poverty. Economic influences on health showed consistent associations across all studies (n = 20) and several were significant (n = 17). The lower the income, wealth, or resources, the higher the likelihood of negative health impacts or prevalence of illness studied. These measures were also studied at different scales. Of the studies, 70% reported measures at the individual/household level, while 25% studied variables at the community/neighborhood level. Only one study reported economic measures at a country level. Several studies evaluated whether effects from income were seen as gradients (e.g., a correlational relationship between the level of income and the severity of the health impact). Descriptive analyses suggested that 85% of the articles reported a negative relationship between an economic influence and a health outcome. The remaining studies had mixed results depending on the geographic scaleand the other variables included in the model. Unlike most of the social stressors in this review, three studies (15%) involved chemical stressors (cigarette smoke, chlorpyrifos, and industrial pollution) in their research design.
## Education
Evidence in the literature shows an association between caregiver educational attainment and child health. Additional evidence shows a positive correlation between educational attainment, mortality, and life expectancy. There is also research that offers evidence that education level can be a predictor of chemical usage in a household environment, as well as a component of SES.
Fourteen articles evaluated the relationship between educational attainment and health and well-being, with seven articles linking higher educational attainment to positive health impactsand six articles linking lower educational attainment to negative health impacts. Ultimately, both reporting scenarios conveyed a positive relationship between education level and quality of health. Although 13 articles offered evidence that lower educational attainment for the participant or caregivers had negative impacts on health and well-being, one article found that depending on race and disease, some instances were found where higher parental educational attainment increased the likelihood of disease for the child. None of the educational attainment research included in this review looked at effects on health and well-being from a combination of chemical exposures and educational attainment.
## Family dynamics
Nineteen articles evaluated family dynamics, such as, family communication and relationships, familial support and involvement, and family structure or environment, such as, stability, siblings, parental structure (single vs. couple), or parenting style. Some of the research looked at familial physicaland well-being characteristics] that might influence a child's social environment. Stronger or more positive familial support and better communication and relationships resulted in a positive impact on health and well-being, while unhealthy relationships and lack of family support and time together showed significant negative impacts on health and well-being. Family structure (such as fewer siblingsor a single parent-householdalso showed a negative impact on child health and well-being. Living in a social environment with a depressed mother (or parents), parents using a democratic parenting style, or overweight parentshad a negative impact on a child's health and well-being. None of the research included in the family dynamic category studied effects on health and well-being from a combination of chemical exposures and non-chemical stressors.
## Food
Previous research has been conducted involving relationships between food and health, including topics on food insecurity, food acculturation, and income. Some research focuses on food choices, such as fast food habits, organic vs. non-organic, or farmers market habits, and food choices which emulate the habits of one's caretaker; while other research has focused on such things as community exposure to agricultureor acculturation of caretakersto explain variations in fruit and vegetable intake for children. Food sources (e.g., fast food, convenience store, supermarket, farmers marker), access to food, and food preparation are derived from the social environment of a child and all contribute to children's health. Research has identified trends in food location, food source, and purchasing habits leading to sources of energy, sugar, sodium, and nutrients. There is also a large body of research that identified food as the source for exposures to toxicants (i.e., flame retardants, pesticides, insecticides).
Several studies included in this review (n = 17) evaluated food in a social context. Some of the research studied food (in)security, food access, and skipping/eating patterns. Additional research studied health impacts resulting from food choices. For example, Lumia et al.showed the advantageous health impacts from early life exposure to fish. Chen et al.and Liu et al.reported on the undesirable health impacts from a poor diet. Drewnowski et al.studied supermarket access. One study compared lunches purchased at school versus lunches brought from home and found that both included negative choices that could adversely impact health. Most of these studies offered evidence that poor diet, lack of food, and meal skipping resulted in both short-and long-term negative health effects, such as obesity, hyperactivity, depression, hypertension, and improper nutrient intake. Only one of the seventeen studies (<6%) considered both a chemical and non-chemical stressor by looking at air pollution from oil refineries and industrial settings, and access to food.
## Greenspace
For this review, greenspace, waterways/rivers/lakes/oceans (anything considered bluespace), and natural or restorative environments were classified as greenspace. Greenspace is identified as "urban nature" or "residential greenspace", and includes parks, fields, forests, gardens, and yards. This review identified studies (n = 8) that researched greenspace locations and uses through a social lens. From those studies, various greenspace-related non-chemical stressors were identified (residential greenspace, urban greenspace, greenspace use, proximity to park, perceived lack of greenspace, coastal proximity, neighborhood greenspace, time spent in greenspace) as well as additional non-chemical stressors that were considered within the research (family relations, neighborhood quality, social capital, rural/urban, SES, number of siblings). Five of the greenspace non-chemical stressors investigated (62%) offered evidence of salutary health effects, which included increased social capital, improved asthma outcomes, higher esteem, and better emotional well-being. Results associated with the remaining three non-chemical stressors were either inconclusiveor not significant. Greenspace can be described as a component of both the natural and social environments. For example, the Montreal Study followed immigrant children and their families, and found that social support was significantly correlated with "urban nature" and negatively correlated with emotional stress. Additional greenspace research has shown that higher greenspace use results in positive impacts on health-related quality of life and friendships. Greenspace use has also been correlated with reducing aggressive behavior and increasing emotional well-being among children between the ages of 7 and 18 years old.
Chen et al.reported a significant interaction effect between the quality of the parent-child relationship and residential greenspace. They showed that as the relationship improved, residential greenspace was more strongly associated with better asthma control. Other studies found that urban greenspace usage was related to quality of life for childrenand lowered aggressive behaviors in adolescents, while perceived lack of greenspace was associated with increased body mass index (BMI). Hordyk et al.used a hermeneutic phenomenological approach to observe that urban nature strengthened social cohesion for immigrants and minimized emotional stress. McCracken et al.found that the use of greenspace influenced participants' self-esteem. Another study found that being close to a preferred park was influential for both usage and health, while being close to a park was not. Studies did not consider a combination of non-chemical and chemical exposures when analyzing greenspace and health effects.
## Neighborhood
Eight studies evaluated whether neighborhood characteristics influenced health outcomes. Some evidence exists in the literature to suggest that neighborhood disorder, neighborhood disadvantage, low quality of neighborhood, and neighborhood problemsall adversely impact health, such as via substance abuse, asthma, pulmonary function, or general well-being, to name a few. However, two studies that looked at neighborhood disadvantagedid not report adverse health effects. Another study on neighborhood quality (criminal activity, substance usage, and vandalism)did not report any adverse health impacts. Two studies included both chemical and non-chemical stressors in their analysis. These studies both included considerations and analysis for the exposures from air pollution (NO 2 and air toxic emissions); however, neither study identified interaction effects between the chemical and non-chemical stressors on health and well-being.
## Social support
Social support is described as a social network of peers and/or family that a person may rely on for emotional support throughout the lifecourse. Social support has been shown to improve physical functioning of the ill and reduce depressive symptoms, as well as improve treatment compliance. Social support is generally analyzed by measuring attributes, such as visiting neighbors, neighborhood involvement, or number of friends. Social support is also seen as a potential by-product of living near "like" people, such as immigrant enclaves or participation in outdoor activities.
Social support (n = 29) as a non-chemical stressor often presented as a salutary measure/variable if the social support existed and was positive. For example, even in an economically burdened neighborhood, such as Baltimore, MD, USA, a belief of having a caring adult in the home was correlated with a child's hope. Other research that reported findings of salutary impacts on health and well-being confirmed that both perceptions of social support and actual social support had beneficial health impacts. Similarly, research offered evidence that low quality social relations, lacking social support, or high social disadvantage have adverse health impacts including, but not limited to, general mental/physical health problems, attention deficit hyperactivity disorder (ADHD), obesity, and cardiometabolic disease. Five studies (17%) had either inconclusive or non-significant results for social support impacts on health and well-being. Caldwell et al.was the only study that looked at a non-chemical stressor (social support) and a chemical exposure (prenatal exposure to ethanol). The impairment seen in the animal test groups exposed to alcohol was ameliorated with communal living (vs. isolation).
## Stress
Studies on stress (n = 14) as a non-chemical stressor consistently (86%) had findings of a negative relationship between stress and health and well-being, such that high levels of stress were inversely associated with health outcomes. Only two studies indicated either not significant or mixed results. Five of the studies (28%) included a chemical exposure in addition to the non-chemical stressor in the analysis. Clougherty et al., Clougherty et al., and Cowell et al.found that adverse health outcomes were heightened or exacerbated with the presence of the non-chemical stressor. In an animal study that analyzed footshocks (stress) and chlorfenvinphos, the results found that, independently, the insecticide or the footshocks (stress) had adverse health outcomes, but stress appeared to have a protective effect that diminished the adverse health outcome from the insecticide when the exposure to stress (footshocks) preceded the insecticide exposure. In addition to stress as a non-chemical stressor, stress was identified in a few of the studies as a health outcome or measure which included measures of general stress, cortisol levels, cortisol reactivity, and persistence of cortisol. Stress as a health outcome is briefly acknowledged in the health outcomes section and; however, this section is addressing stress as a non-chemical factor.
## Urbanicity
This review identified five articles that assessed geographic environments (urbanicity: e.g., rural, suburban, urban) on health and well-being. Rural urbanization encompasses stressors, such as a shift from agrarian to industrial society factors, caregiver education, sanitation practices, food options (availability and behaviors), and health care access and practices. Many stressors associated with urban development may have both individual and community level effects. Previous studies have shown that children in urban areas grow faster than those in rural areas, offering evidence that an urban setting can influence physiological changes occurring throughout a child's lifecourse. Other studies show significant differences in food quality consumption, asthma prevalence, and general health conditionsbetween urban and rural regions.
The studies included in this review highlighted research that showed that both higher urbanicityand lower urbanicitywere associated with negative health impacts depending on the health outcome measured. For example, Protano et al.and Wood et al.saw associations between lower urbanicity and increased body weight/obesity. Erinosho et al.identified a link between rural environments and consuming less vegetables. Breslau et al.and Chai et al.both identified stronger correlations between high urbanicity and exposure to violence, which led to indirect impacts, such as, but not limited to, higher incidences of post-traumatic stress disorder (PTSD) in females. In addition, recent research exists on non-chemical stressors (acculturation, meal patterns, food insecurity, exposure to violence, and social determinants) in predominantly urban or rural populations, but does not test for urbanicity as a variable of influence, emphasizing the need for more studies that consider urbanicity as a non-chemical stressor in future studies. None of the studies included in this review researched urbanicity and chemical exposures as combined exposures impacting health and well-being.
## Violence
When children are exposed to violence, it can lead to health problems throughout the lifecourse. Exposure to violence (ETV) can be direct (being the victim of the violence) or indirect (e.g., witnessing the violence). Types of violence to which a child may be exposed include physical, sexual, verbal/threat, crime, or bullying. ETV can happen on an individual scale with someone familiar or a stranger, a social scale (such as schoolyard bullying or athletic team loyalties), a neighborhood or community scale, or a national scale, which could include political strife, terrorism, and war. Several studies outside of the inclusion criteria for this review indicated that drug and alcohol misuse, mental health status, and behavior of adolescents and adults are influenced by early life exposures to violence.
Exposure to violence during childhood resulted in an inverse relationship to a person's health and well-being. Of the 43 studies included in this review, 93% reported an adverse impact on health and well-being irrespective of the type or scale of the violence, whereas 7% of the studies reported either no, mixed, or inconclusive impacts on health and well-being. Of the studies that evaluated violence as a variable affecting health and well-being, only 7% included a chemical exposure.
## Other
Nineteen of the studies did not fit well into the categories and were grouped as "other". Although race and ethnicity can each be considered non-chemical stressors found in a child's social environment, this review did not seek to find research on race or ethnicity because this research used the construct published by Tulve et al.as a foundation for study and search design. There were, however, four studies in this review that identified race, ethnicity, or ancestry as an interactive variable which had significant impacts on the health outcomes for children. There were three studies that identified geographic-type influence (e.g., area of city, crowding, country of originwhich had significant negative impacts on children's health. On the contrary, Waters et al.did not find overcrowding to have a negative effect on health and when Drewnowski et al.measured the effect that distance to (food) market had, they found it to be not significant.
Studies that scrutinized behaviors, such as poor sleepor lower physical activity, showed significant negative health impacts. Likewise, studies that examined living in household chaos, in high turbulence, with high child misfortune, or with high discrimination, all showed significant negative health outcomes. Beaver et al.found that duration of breastfeeding had a significant impact on children's health.
# Additional results
## Health outcomes
This review did not set inclusion criteria based on health variables measured, so we are able to see what the scientific community is studying regarding health outcomes, indicators, or behaviors, and their linkages to non-chemical stressors. Health outcomes, indicators, and behaviors extracted from the studies in this review were organized into 13 categories: asthma, cardiovascular health, chronic conditions (general), diabetes/allostatic, general physical health, life expectancy, mental health, neurological, risk behavior, stress (In addition, note that in the literature stress is studied as both an independent variable that may impact a health outcome, and as a health outcome or measure which results from non-chemical and/or chemical exposures. Health outcomes for stress might include stress, cortisol levels, or the persistence of cortisol.), weight, and other (e.g., cancer, telomere health, dental, sleep health, memory). Mental health, weight, and general physical health were the most studied, followed by asthma and dietary habits.
## Studies evaluating both chemical and non-chemical stressors
Fourteen studies looked at both chemical and non-chemical stressors found in a child's social environment. Seven of these studies evaluated responses to stress and either air pollution (NO2, black carbon, or concentrated ambient fine particles, pesticides, nicotine, or alkylphenols. Three were animal studies. Two of the studies explored linkages between air pollution and community level social stressors. The remaining five studies investigated effects from alcohol and early-life rearing conditions (isolated or communal nest of dams), pesticide exposures and neighborhood poverty, pesticide exposures and ACEs, tobacco smoke and material hardship, and oil refineries and neighborhood poverty. Two animal studies suggested salutary effects from the relationship between a non-chemical and chemical stressor. Gralewicz et al.reported that exposure to stress induced a cortisol response that had a protective effect against exposure to an organophosphate pesticide (chlorfenvinphos). Another study showed that social support (communal nests) had an ameliorating influence on the negative effect of neurochemical changes as a result of alcohol exposure. Five of the studies that analyzed the effect of non-chemical stressors on a biological response to a chemical exposure stated that the non-chemical stressor either had an exacerbating, heightened, or synergisticinfluence on the health outcome. McCormick et al.also offered evidence that adolescents were especially vulnerable to specific stressors because of their developmental stage.
## Studies evaluating both chemical and non-chemical stressors
Fourteen studies looked at both chemical and non-chemical stressors found in a child's social environment. Seven of these studies evaluated responses to stress and either air pollution (NO 2 , black carbon, or concentrated ambient fine particles, pesticides, nicotine, or alkylphenols. Three were animal studies. Two of the studies explored linkages between air pollution and community level social stressors. The remaining five studies investigated effects from alcohol and early-life rearing conditions (isolated or communal nest of dams), pesticide exposures and neighborhood poverty, pesticide exposures and ACEs, tobacco smoke and material hardship, and oil refineries and neighborhood poverty. Two animal studies suggested salutary effects from the relationship between a non-chemical and chemical stressor. Gralewicz et al.reported that exposure to stress induced a cortisol response that had a protective effect against exposure to an organophosphate pesticide (chlorfenvinphos). Another study showed that social support (communal nests) had an ameliorating influence on the negative effect of neurochemical changes as a result of alcohol exposure. Five of the studies that analyzed the effect of non-chemical stressors on a biological response to a chemical exposure stated that the non-chemical stressor either had an exacerbating, heightened, or synergisticinfluence on the health outcome. McCormick et al.also offered evidence that adolescents were especially vulnerable to specific stressors because of their developmental stage. When studying the links between prenatal exposure to black carbon and community violence, Chiu et al.found an interaction effect on childhood asthma. Shmool et al.found that community level social stressors, such as crime and physical disorder, negatively impacted children's health independently, but overcrowding and lack of resources modified the NO 2 levels in the neighborhood.
## Animal studies
Six studies in this review were animal studies. Four of the studies analyzed a response to higher stress brought on by footshocks, social defeat and overcrowding, strangers in their den, or predators within observation range. In addition, studies looked at responses to social situations, such as higher socialized living (communal nests)or social isolation. McCormick et al.found negative gender-specific health outcomes during adolescent lifestage exposures to stress, isolation, and nicotine. Clougherty et al., Gergs et al., and Reber et al.had similar findings of significant negative health effects resulting from higher exposures to stress, while Gralewicz et al.found that stress had created a protective effect on the response of hyposensitivity to pesticide exposure. Caldwell et al.found that increased social environments offered significant positive health effects. Overall, it is understood that animal studies can be used to evaluate cause and effect when human studies are not available.
# Discussion
## Summary of findings
Non-chemical stressors have been studied in both the physical and social sciences as both direct and confounding variables. With little exception, the interrelationships between chemical and non-chemical stressors from a child's social environment are not studied in regard to health and well-being. Reducing this information gap would make relevant scientific contributions to understanding cumulative exposures and risks in a child's social environment.
Through organizing and synthesizing the current relevant literature this review provides evidence that non-chemical stressors found in a child's social environment can influence their health and well-being. This review confirmed that adverse relationships exist between health and selected non-chemical stressors including, but not limited to, economic disadvantage, lower educational attainment, exposure to violence, adverse childhood experiences, stress, and urbanicity. On the other hand, this review also identified the salutary effects of some non-chemical stressors, such as exposure to or experience from greenspace and social supporton health and well-being.
Also identified through the synthesis of extant literature was the paucity of studies evaluating links between chemical and non-chemical stressors and their combined and/or interactive impacts on children's health and well-being. The existing literature on studies testing non-chemical stressors varies greatly in study design, non-chemical stressor considered, and chemical stressor considered, if one was incorporated. When considering the conceptual framework published by Tulve et al., understanding the interrelationships between chemical and non-chemical stressors is paramount, since people are exposed to both chemical and non-chemical stressors at each lifestage throughout their lifecourse, impacting their health and well-being in countless ways.
# Limitations
Limitations with this review include the identification of numerous inconsistencies in terminology (e.g., non-chemical stressors, psychosocial stressors, social determinants of health, environmental stressors), methodologies, and the value of non-chemical stressor research in different scientific disciplines. In order to overcome this, the search terms were increased to attempt to capture more studies for the review. Additionally, often the non-chemical stressor being studied is an index for other interchangeable variables (e.g., SES) resulting in more inconsistencies, or can be measured in many non-standardized ways (e.g., stress). This heterogeneity can be problematic for research in public health. An additional limitation includes the lack of interdisciplinary research that bridges the gap between the physical and social sciences, which could strengthen study designs and methodologies. Another limitation would be the possibility there may be additional studies not captured in this review.
# Conclusions
This review presents evidence of the importance of considering non-chemical stressors found in a child's social environment when addressing children's health, exposure assessment, or risk assessment. This review also highlights a noticeable dearth in research addressing both non-chemical and chemical stressors, supported by the few studies that showed significant health effects from combined exposures to both chemical and non-chemical stressors. In addition, there is a need for future studies to address the interaction effects of chemical and non-chemical stressors. This review uses existing literature to offer evidence that: (1) non-chemical stressors in a child's social environment can impact health; (2) the health impact can appear after various latency periods after exposure; (3) the non-chemical stressor can influence the response to a chemical exposure; and (4) very few studies look at the resulting health effects from exposure to both chemical and non-chemical stressors combined. Therefore, this review highlights the importance of including non-chemical stressors found in a child's social environment when understanding health impacts at each lifestage throughout the lifecourse. It is also important to recognize that not all non-chemical stressors are mutually exclusive. It is important when designing research to not overestimate the burden of the non-chemical exposure by capturing the effect more than once. In this review there were instances of ACE studies that included exposure to violence, there were neighborhood studies that included income measures, and acculturation studies that included social measures. Future research on cumulative exposures including non-chemical and chemical factors needs to acknowledge this to eliminate overestimation of non-chemical stressor effects. Acknowledgments: The authors would like to acknowledge Kiran Alapaty, CarolAnn Gross-Davis, Lisa Melnyk, and Lindsay Stanek for their reviews and editorial contributions.
## Conflicts of interest:
The authors declare no conflict of interest.
## Disclaimer:
The views expressed in this presentation are those of the authors and do not necessarily represent the views or policies of the U.S. Environmental Protection Agency.
## Consent of publication:
It has been subjected to EPA administrative review and approved for publication. Year
## Acronyms and abbreviations
|
Elucidating the Molecular Determinants of Aβ Aggregation with Deep Mutational Scanning
Despite the importance of Ab aggregation in Alzheimer's disease etiology, our understanding of the sequence determinants of aggregation is sparse and largely derived from in vitro studies. For example, in vitro proline and alanine scanning mutagenesis of Ab 40 proposed core regions important for aggregation. However, we lack even this limited mutagenesis data for the more diseaserelevant Ab 42 . Thus, to better understand the molecular determinants of Ab 42 aggregation in a cellbased system, we combined a yeast DHFR aggregation assay with deep mutational scanning. We measured the effect of 791 of the 798 possible single amino acid substitutions on the aggregation propensity of Ab 42 . We found that 75% of substitutions, largely to hydrophobic residues, maintained or increased aggregation. We identified 11 positions at which substitutions, particularly to hydrophilic and charged amino acids, disrupted Ab aggregation. These critical positions were similar but not identical to critical positions identified in previous Ab mutagenesis studies. Finally, we analyzed our large-scale mutagenesis data in the context of different Ab aggregate structural models, finding that the mutagenesis data agreed best with models derived from fibrils seeded using brain-derived Ab aggregates.
treatment with the competitive DHFR inhibitor methotrexate, yeast expressing soluble Ab variants grow rapidly, whereas yeast expressing aggregating Ab variants grow slowly.
Mutagenesis can elucidate the role of individual residues in protein aggregation. For example, in vitro proline [bib_ref] Mapping abeta amyloid fibril secondary structure using scanning proline mutagenesis, Williams [/bib_ref] and alanine [bib_ref] Alanine scanning mutagenesis of Abeta(1-40) amyloid fibril stability, Williams [/bib_ref] scanning mutagenesis of Ab 40 revealed core regions important for aggregation. However, we lack even this limited mutagenesis data for the more disease-relevant Ab 42 and, so far, the majority of mutagenesis studies have been performed in vitro.
Thus, to fully understand the molecular determinants of Ab 42 aggregation in a cell-based system, we combined the yeast growth-based aggregation assay with deep mutational scanning [bib_ref] Deep mutational scanning: assessing protein function on a massive scale, Araya [/bib_ref] to measure the effect of 791 of the possible 798 single amino acid substitution on the aggregation propensity of Ab 42 . We used high-throughput DNA sequencing to track the frequency of each Ab 42 variant during the selection, enabling us to assign a solubility score to every variant. We present the first large-scale, cell-based mutational analysis of Ab, illuminating the physicochemical properties of amino acids that abrogate, promote or do not affect Ab aggregation. Of 791 single amino acid Ab variants we evaluated, 75% maintained or increased aggregation. In addition, we identified 11 positions at which substitutions, particularly to hydrophilic and charged amino acids, disrupted Ab aggregation. These critical positions were similar but not identical to critical positions identified in previous Ab mutagenesis studies. Finally, we analyzed our large-scale mutagenesis data in the context of different Ab aggregate structural models, finding that some structures were plausible whereas others were not.
# Methods
## Library construction
The library was cloned using in vivo assembly [bib_ref] IVA cloning: A single-tube universal cloning system exploiting bacterial In Vivo Assembly, García-Nafría [/bib_ref]. First, a forward primer containing a 59 homology region, an NNK codon, and a 39 extension region was designed for each codon in Ab 42 . The homology and extension regions were at least 15 nucleotides in length and had melting temperatures greater than 55C. Reverse primers were the reverse complement of the 59 homology region.
A separate PCR reaction was performed for each codon. These reactions contained 40 ng template (p416GAL1-Ab-DHFR) and 10 mM forward and reverse primers (IDT, custom oligos) in a total reaction volume of 30 mL. The following cycling conditions were used: 95C 3min, 8x [ 98C 20 sec, 60C 15 sec, 72C 9 min], 72C 9 min. After PCR, 7.5 mL of each product was run on a 1.5% agarose gel for 30 min at 100V to check for a single product. The remaining 22.5 mL aliquots of product were each digested for an hour at 37C with 0.6 mL of DpnI (NEB, R0176S). After digestion, 4 mL of each linear product was transformed into a 50 mL of TOP10F Chemically Competent E. coli (ThermoFisher, C303003) according to manufacturer's instructions, with the following modifications: the protocol was done in a 96 well plate, and cells were recovered in a total volume of 200 mL SOC. After recovery, cells were transferred to a deep well plate with 1.6-1.8 mL of ampicillin LB and shaken overnight. To estimate colony count, 50 mL of culture was plated on an LB + ampicillin agar plate. Deep well plates and agar plates were incubated at 37C overnight. After incubation, all 42 deep well plate cultures were combined and subject to Midiprep (Sigma, NA0200).
## Plasmids, yeast strains and growth conditions
To create a galactose-regulated Ab-DHFR expression system, we directionally cloned DHFR into p416 (URA3, GAL1 promoter, CEN) using Blp and SpeI and then cloned the human Ab 42 coding sequence into the SpeI and HindIII of the same vector. Ab-GFP variants were cloned using the same scheme. All Ab variants were cloned into p416 and transformed in W303 strain (MATa/MATa {leu2-3,112 trp1-1 can1-100 ura3-1 ade2-1 his3-11,15} [phi+]). Cells were grown at 30C in synthetic complete (SC) media lacking uracil and supplemented with 2% glucose.
## Methotrexate selection assay
Transformed yeast were inoculated into 5 mL (low-throughput) or 300 mL (co-culture and deep mutational scan) of C-Ura, 2% glucose media, grown in a rotating/shaking, 30C incubator overnight and then transferred to 5 mL or 300 mL 2% raffinose media to remove the glucose repression acting on the gal1 promoter. After two hours in 2% raffinose, yeast were back-diluted to an OD of 0.01 into 5 mL or 300 mL 2% galactose to induce Ab 42 -DHFR expression in the presence or absence of 80 mΜ methotrexate (TCI America, M-1664) and 1 mM sulfanilamide (Sigma, S-9251). In 5 mL experiments, yeast growth was measured over 48h using a spectrophotometer that detects 660 nm wavelengths. The following equation was used to calculate doubling times from two time points: (Log 10 (OD T2 /OD T1 )/ Log 10 (2))/ DΤ, where OD represents the optical density at 600nm at a time point (T). For co-culture experiments, yeast with aggregating and nonaggregating variants were inoculated at equal densities in 300 mL. Ten OD units of yeast were collected from 300 mL cultures every 12h, spun down, concentrated and stored in -80C. At the end of the experiment, frozen yeast were thawed and then their plasmids were extracted using a DNA Clean and Concentrator kit (Zymo Research, D4013). Extracted plasmids were prepped and sequenced using Sanger sequencing. For the deep mutational scan, 300 mL cultures were sampled at the following timepoints: input, 28h (OD 1.0), 31.5h (OD 2.0), 35h (OD 3.0), 38h (OD 4.5), and 40h (OD 6.0). Cultures were spun down, concentrated and stored in -80C. Plasmids were extracted from yeast with Yeast Plasmid Miniprep 1 kit . Library fragments were amplified in 17 PCR cycles using primers specific to DNA sequences that flank Ab-DHFR in p416, and sequenced by an Illumina NextSeq sequencer using paired-end reads .
# Variant effect analysis
Enrich2 was used to calculate solubility scores for each Ab variant from sequencing fastq files. The Enrich2 pipeline calculates a variant's score in three steps. First, a variant's normalized frequency ratios are tabulated for each timepoint by dividing the frequency of a variant's sequencing reads by all mapped reads and normalizing by the wild-type frequency ratio. Sequencing reads were stringently filtered for quality; we require each base have a Phred score greater than 20 and no uncalled bases. Second, a weighted linear least squares regression line is fit to the normalized frequency ratios across time points. Third, the slope of the regression line is calculated, averaged across the three replicates and log 2 transformed. This averaged log 2 slope reflects a variant's aggregation propensity. Solubility scores below 0 denote variants that are more aggregation-prone than wildtype, whereas scores above 0 indicate that a variant has increased solubility compared to wild-type.
Classifying Ab variants using synonymous mutations Variant classifications (i.e., WT-like, more aggregation-prone, more soluble) were assigned using the distribution of 39 synonymous mutations from the deep mutational scan. We define WT-like as any variant with a score within 6 2 SD of the synonymous variant mean . A variant is more-aggregation prone than wildtype if its score is greater than 0.39 or more soluble if its score is lower than -0.26.
## Data and code availability
Raw sequencing data is available in the NCBI GEO database (accession number GSE139122). Code and variant scores are available at https:// github.com/FowlerLab/amyloidBeta2019. Supplemental material available at FigShare: https://doi.org/10.6084/m9.figshare.8330297.
# Results
First, we verified that the DHFR-based yeast aggregation assay could differentiate between aggregating wild type Ab (Ab WT ) and a nonaggregating (Ab 19 FD ) variant [bib_ref] Linking amyloid protein aggregation and yeast survival, Morell [/bib_ref]. As expected, in a mixed culture treated with methotrexate, Ab 19 FD outcompeted Ab WT [fig_ref] Figure 1 A: -D [/fig_ref]. We used fluorescence microscopy of Ab-GFP fusions to confirm that 30-70% of yeast expressing Ab WT -GFP had cytoplasmic punctae compared to 0-20% of cells expressing Ab 19 FD -GFP across five fields of view [fig_ref] Figure 1 A: -D [/fig_ref]. Thus, we concluded the assay could be used in a deep mutational scan to measure the aggregation propensity of variants of Ab.
Using this assay, we conducted a deep mutational scan of Ab that yielded solubility scores for 791 single amino acid variants, representing 99.1% of the possible single variants. Solubility scores were calculated by taking the weighted least squares slope of each variant's frequency ratios across six time points. (see Methods). The slopes from each replicate were well correlated (Pearson's R 0.78 to 0.92; , [fig_ref] Figure 1 A: -D [/fig_ref]. Replicate slopes were averaged and log 2 transformed to produce final solubility scores such that wild-type had a solubility score of zero . Positive solubility scores indicated less aggregation and negative scores indicated increased aggregation.
Solubility scores ranged from -2.38 (most aggregating) to 1.45 (most soluble). The mean (median) solubility score for all variants was 0.09 (0.08), which was similar to the solubility scores of the 39 synonymous variants in our library (mean: 0.06; median: 0.08). Because we did not expect synonymous variants to affect aggregation propensity, we used their distribution of scores to identify WT-like variants . In total, we found that 344 (43.4%) of Ab variant scores were within two standard deviations of the synonymous score mean and thus had WT-like effects (WT-like range: . Additionally, we found 246 (31.1%) variants to be more aggregation-prone than Ab WT and 201 (25.4%) variants to be more soluble. Therefore, 75% of Ab variants maintained or increased the peptide's propensity to aggregate in yeast cells.
To verify that our deep mutational scan accurately measured variant effects on aggregation, we tested six Ab variants, G38F, K16V, A42V, 19 A-F. Solubility scores for 791 Ab variants. Solubility scores reliably measure the effects of Ab sequence on aggregation propensity. A scatter plot shows the correlation between two of three biological replicates that were averaged to yield final solubility scores (A; [fig_ref] Figure 1 A: -D [/fig_ref]. The distribution of solubility scores (x-axis) of synonymous variants was used to determine cutoffs that define variants that are wild-type-like or more/less aggregation-prone than wild-type. The density plot shows distributions of nonsynonymous (light gray) and synonymous (dark gray) variants and the white lines show the lower (-0.26) and upper (0.39) bounds for wild-type-like variants (B). The scatterplot shows the correlation between our solubility scores (y-axis) and a low-throughput yeast growth assay that measured yeast growth rate as a proxy for Ab solubility (C; [fig_ref] Figure 1 A: -D [/fig_ref]. The heatmap shows the effect of 791 Ab variants on solubility with Ab positions on the x-axis and mutant amino acids on the y-axis. A variant's color denotes its solubility: red is most soluble, white is wild-type-like and, dark blue is most aggregated, whereas yellow variants are missing from our variant library and dots denote the wild-type amino acid at a given position. The annotation tracks on the x-and y-axes display the hydrophobicity of each wild-type and mutant amino acid, respectively. The heatmap's y-axis has been re-ordered using hierarchical clustering on the solubility score vectors (D). For each position, the mean solubility score at each position is depicted using the same color scheme as the main heatmap. Additionally, the mean solubility scores for all hydrophobic and polar substitutions are shown (E; . Heirarchical clustering on the x-axis yielded 6 distinct clusters: 1 (red), 2 (orange), 3 (yellow), 4 (green), 5 (light blue), and 6 (dark blue; F; .
FY, L17S and L34R, that spanned the solubility score range in a lowthroughput validation assay. The growth rate of methotrexate-treated yeast expressing each Ab variant was measured and compared to the aggregation propensity scores , S1B). We found that low-throughput assay results strongly correlated with the solubility scores derived from deep mutational scanning (R 2 = 0.98). Thus, our deep mutational scan reliably measured Ab variant aggregation propensity in the yeast assay.
To explore the effects of each amino acid substitution on Ab aggregation, we created an Ab sequence-aggregation map . Substitutions to aspartic acid and proline were most associated with Ab solubility, as evinced by their median scores of 0.64 and 0.56, respectively . Conversely, the most aggregation-associated substitutions were hydrophobic tryptophan and phenylalanine, with scores of -0.60 and -0.51, respectively. Moreover, hierarchical clustering of all 791 solubility scores by amino acid revealed that hydrophobic substitutions, except alanine, clustered together and were associated with greater aggregation than other classes of substitutions.
Next, we characterized each position in Ab based on its mutational profile. Hierarchical clustering of variant solubility scores by position identified six distinct clusters ; S2B-C). In cluster 1, comprising positions and 41, substitutions tended to decrease Ab aggregation compared to substitutions in other clusters (cluster 1 mean solubility scores = 0.64, all other clusters = -0.28; . In cluster 1, even substitutions to hydrophobic amino acids slightly decreased aggregation (mean solubility score = 0.17). The effects of substitutions in cluster 2 were similar to but less extreme than in cluster 1. Both clusters 1 and 2 are largely comprised of hydrophobic positions in the wild type Ab sequence. Indeed, 80% of Ab positions with hydrophobic wild type residues are in clusters 1 and 2. In stark contrast, within clusters 4, 5 and 6, hydrophobic substitutions generally increase protein aggregation (all mean: -0.15, -0.12 and -0.45; hydrophobic means: -0.29, -0.65, and -1.04). Cluster 3 contains only two positions, 37 and 38. Here, every substitution except proline increased aggregation (all mean: -0.99, hydrophobic mean: -1.56). Given that cluster 1 is characterized by hydrophobic positions where hydrophilic substitutions profoundly decreased aggregation, we suggest that this cluster defined buried b-strands in the Ab sequence.
Next, we compared our solubility scores to previous alanine and proline scans which reported Ab 40 fibril thermodynamic stability in vitro (DDG). DDG values were determined by measuring variant Ab monomer concentration remaining in solution after fibril formation reached equilibrium [bib_ref] Mapping abeta amyloid fibril secondary structure using scanning proline mutagenesis, Williams [/bib_ref] [bib_ref] Alanine scanning mutagenesis of Abeta(1-40) amyloid fibril stability, Williams [/bib_ref]. We found that the effects of proline substitution in our assay were correlated with proline DDG values (R 2 = 0.40), while the effects of alanine substitutions in our assay were less correlated with alanine DDG values (R 2 = 0.17; . In our alanine and proline comparisons, we found the greatest correlation at positions 17-20 and 31-32, where substitutions decreased aggregation in all studies . The most notable disagreement between studies was for alanine substitutions at positions 37 and 38. In our assay, alanine substitutions caused a profound increase in aggregation, whereas the in vitro alanine scan showed the opposite effect.
We also compared our buried b-stand positions from cluster 1 to b-stands proposed based on the in vitro alanine and proline scans, finding some concordance . The single amino acid scans identify three regions that disrupt fibril elongation thermodynamics when mutated. The regions include positions 15-21, 24-28, and Comparison of yeast cell-based solubility scores to in vitro aggregation measurements and Ab structural models. The scatterplot shows the correlation between our solubility scores (y-axis) and two single amino acid scans that measured the effect of proline (orange) or alanine (teal) variants on the thermodynamic stability of aggregates, relative to wild type (DDG) (A; . The first two tracks show unmeasured mutations (dashed gray) and the Ab buried b-strand positions (black) suggested by proline scanning alone, or by proline and alanine scanning together [bib_ref] Mapping abeta amyloid fibril secondary structure using scanning proline mutagenesis, Williams [/bib_ref] [bib_ref] Alanine scanning mutagenesis of Abeta(1-40) amyloid fibril stability, Williams [/bib_ref]. The third track shows positions with the greatest increase in solubility when mutated in our large-scale mutagenesis study, found in cluster 1 (B). The next nine tracks show the secondary structure of nine models of Ab aggregate structure for each Ab position (x-axis; C). The Ab wild-type sequence is shown at the top. 31-36 for the proline scan and positions 18-21, 25-26, and 32-33 for the combined alanine and proline scans [bib_ref] Mapping abeta amyloid fibril secondary structure using scanning proline mutagenesis, Williams [/bib_ref] [bib_ref] Alanine scanning mutagenesis of Abeta(1-40) amyloid fibril stability, Williams [/bib_ref]. Given the generally highly disruptive nature of proline substitutions [bib_ref] Analysis of Large-Scale Mutagenesis Data To Assess the Impact of Single Amino..., Gray [/bib_ref] , it is not surprising that the proline scan would nominate many positions. Our deep mutational scan, on the other hand, does not reveal a central b-strand or strong decrease in aggregation with alanine or proline substitution from positions 24-28. We speculate that this difference is due either to the distinct experimental approaches used or to the different Ab species (Ab 40 vs. Ab 42 ).
# Discussion
We used deep mutational scanning to characterize 791 Ab variants in a yeast-based aggregation assay. Proline and aspartic acid substitutions were most disruptive of Ab aggregation, while tryptophan and phenylalanine increased aggregation most. Additionally, we used unsupervised clustering to determine the regions of Ab most important for aggregation. We conclude that these regions are most likely to form buried b-stands, which are necessary for aggregation and sensitive to amino acid substitutions [bib_ref] The common architecture of cross-beta amyloid, Jahn [/bib_ref] [bib_ref] Alpha Helices Are More Robust to Mutations than Beta Strands, Abrusán [/bib_ref]. These include positions . While other positions could also form b-stands, the positions in cluster 1 are most likely to form the buried cores of Ab aggregates in our cell-based assay.
Due to the noncrystalline nature of Ab fibrils, traditional techniques such as X-ray crystallography and solution-state NMR cannot be used to solve Ab's aggregate structure. Instead, structural models have been developed by amassing constraints, such as the direction and register of b-sheets. For example, solid-state nuclear magnetic resonance studies suggest that Ab fibrils are parallel, in register b-sheets [bib_ref] Dipolar recoupling NMR of biomolecular self-assemblies: determining inter-and intrastrand distances in fibrilized..., Gregory [/bib_ref] [bib_ref] Supramolecular structural constraints on Alzheimer's beta-amyloid fibrils from electron microscopy and solid-state..., Antzutkin [/bib_ref] [bib_ref] Solid-state NMR studies of amyloid fibril structure, Tycko [/bib_ref]. Many of these structural models are problematic because they are generated from constraints derived from in vitro experimental data, which may not be representative of in vivo conditions.
Given that we collected large-scale mutagenesis data in a cell-based system, we examined how our results compared to structural models of Ab fibrils. Some models such as 1IYTand 2NAO [bib_ref] Atomicresolution structure of a disease-relevant Ab(1-42) amyloid fibril, Wälti [/bib_ref] , showed very little to no overlap with either our proposed buried b-strands or those proposed by [bib_ref] Mapping abeta amyloid fibril secondary structure using scanning proline mutagenesis, Williams [/bib_ref] [bib_ref] Alanine scanning mutagenesis of Abeta(1-40) amyloid fibril stability, Williams [/bib_ref]. Other models contained three b-strand regions reminiscent of those suggested by [bib_ref] Mapping abeta amyloid fibril secondary structure using scanning proline mutagenesis, Williams [/bib_ref] [bib_ref] Alanine scanning mutagenesis of Abeta(1-40) amyloid fibril stability, Williams [/bib_ref] : 2MXU [bib_ref] Ab(1-42) fibril structure illuminates self-recognition and replication of amyloid in Alzheimer's disease, Xiao [/bib_ref] , 5KK3 [bib_ref] Atomic Resolution Structure of Monomorphic Ab42 Amyloid Fibrils, Colvin [/bib_ref] , and 5OQV [bib_ref] Fibril structure of amyloid-b(1-42) by cryo-electron microscopy, Gremer [/bib_ref]. Yet other models propose b-strand patterns more similar to ours. These include 2BEG, 2LNQ [bib_ref] Fibril structure of amyloid-b(1-42) by cryo-electron microscopy, Gremer [/bib_ref] , 2LMP and 2LMN [bib_ref] Molecular structure of b-amyloid fibrils in Alzheimer's disease brain tissue, Lu [/bib_ref]. Since our b-strand patterns were derived from data gathered in a cell-based assay, we hypothesized that they would be most consistent with structural models based on in vivo-derived fibrils. Indeed, the 2LMP and 2LMN models were based on fibrils seeded from plaques isolated from the brains of individuals afflicted by Alzheimer's disease. Moreover, every model besides 2LMP and 2LMN was constructed using NMR or cryo-EM data from laboratory grown fibrils. These models are less concordant with our cell-based mutational data, which suggests that there are important structural differences between in vitro and in vivo derived fibrils.
Two major differences exist between the experimental conditions used by [bib_ref] Mapping abeta amyloid fibril secondary structure using scanning proline mutagenesis, Williams [/bib_ref] [bib_ref] Alanine scanning mutagenesis of Abeta(1-40) amyloid fibril stability, Williams [/bib_ref] and in our work, and may explain the difference in b-strands proposed in our respective in vitroand in vivo-derived models. First, [bib_ref] Mapping abeta amyloid fibril secondary structure using scanning proline mutagenesis, Williams [/bib_ref] [bib_ref] Alanine scanning mutagenesis of Abeta(1-40) amyloid fibril stability, Williams [/bib_ref] incubate Ab in the absence of any other proteins, while our yeast-based system provides key players that affect protein aggregation, such as chaperone proteins and molecular crowding. Second, [bib_ref] Mapping abeta amyloid fibril secondary structure using scanning proline mutagenesis, Williams [/bib_ref] [bib_ref] Alanine scanning mutagenesis of Abeta(1-40) amyloid fibril stability, Williams [/bib_ref] incubate Ab peptides at 37C, whereas our yeast-based experiments required a lower temperature of 30C. This temperature difference may yield differences in folding kinetics. Further experiments are required to determine the contribution of these experimental differences to b-strand formation in Ab.
Deep mutational scanning data could contribute to the investigation of Ab fibril structure beyond the analysis of existing models we present. For example, others have used site-saturation mutagenesis and deep mutational scanning data to evaluate proposed structural models [bib_ref] Structural correlates of the temperature sensitive phenotype derived from saturation mutagenesis studies..., Bajaj [/bib_ref] [bib_ref] Protein model discrimination attempts using mutational sensitivity, predicted secondary structure, and model..., Khare [/bib_ref]. Additionally, deep mutational scanning data have now been used to generate distance constraints for the prediction of tertiary protein structure .
In summary, we used deep mutational scanning to elucidate the effects of amino acid substitutions on Ab aggregation in a cell-based model. We used these large-scale mutagenesis data to propose positions critical for Ab aggregation. Our results conflict with some previous in vitro reports of the effects of substitutions on Ab aggregation and with some models of Ab fibril structure. This outcome highlights the difficulties of studying protein aggregation and emphasizes the potential utility of in vivo or cell-based models. We suggest that deep mutational scanning of other aggregation-prone proteins such as a-synuclein or transthyretin could help reveal the relationship between sequence, structure and aggregation.
# Acknowledgments
## Literature cited
[fig] Figure 1 A: -D. A yeast-based aggregation assay distinguishes between soluble and aggregation-prone variants of Ab. A schematic of the assay shows plasmid-based expression of Ab-DHFR and a nonaggregating variant of Ab fused to DHFR, which lead to slow and fast yeast growth in the presence of methotrexate, respectively (A). A stacked bar graph shows the percentage of Ab-DHFR and Ab 19 FD -DHFR in co-culture (y-axis) every 12 hr for 48 hr (x-axis; B). Fluorescence light microscopy shows the aggregation patterns of Ab-GFP (WT) and Ab 19 FD -GFP ( 19 FD) 16h after induction of expression (C). A bar graph shows the percentage of yeast cells with punctae (y-axis) in five fluorescence microscopy images of Ab-DHFR (WT) or Ab 19 FD -DHFR ( 19 FD; x-axis; D). [/fig]
|
Pseudoaneurisma da artéria temporal superficial: relato de três casos
ResumoO pseudoaneurisma da artéria temporal superficial é um evento raro. Ocorre principalmente em homens jovens em decorrência de traumatismo craniano fechado. A maioria dos casos é assintomática, podendo eventualmente haver associação com sintomas vagos. A rotura do pseudoaneurisma e o desenvolvimento de déficits neurológicos são complicações esporádicas. Os autores relatam três casos de pseudoaneurisma de artéria temporal superficial pós-traumatismo craniano fechado. Os pacientes foram manejados com sucesso por ligadura e ressecção dos pseudoaneurismas.Palavras-chave: artérias temporais; falso aneurisma; traumatismos craniocerebrais.AbstractPseudoaneurysms of the superficial temporal artery are rare events. They primarily occur in young men as a consequence of blunt head traumas. The majority of cases are asymptomatic, although there are sometimes vague symptoms. Ruptured pseudoaneurysms and development of neurological deficits are occasional complications. The authors describe three cases of pseudoaneurysm of the superficial temporal artery after blunt head trauma. Patients were successfully managed with ligature and resection of the pseudoaneurysms.
## Introdução
O pseudoaneurisma traumático da artéria temporal superficial (ATS) é raro e representa menos de 1% dos aneurismas descritos . O primeiro caso foi relatado por Thomas Bartholin em 1740 2,3 e, desde então, aproximadamente 200 casos foram descritos na literatura . Embora associado à baixa morbidade, o pseudoaneurisma da ATS pode causar uma série de sintomas locais e até mesmo rotura. Os autores relatam três casos de pseudoaneurisma da ATS secundários a traumatismo craniano fechado.
## Descrição dos casos
## Caso 1
Paciente do sexo masculino, 22 anos, com queixa de desconforto na região frontoparietal direita associado a um aumento de volume local. Relatava história de acidente automobilístico havia três meses com trauma contuso na região temporal direita. Referiu sangramento no momento do trauma e desenvolvimento de hematoma periorbitário. Foi avaliado e manejado com sutura da ferida e observação; porém, após 20 dias, notou um aumento progressivo de volume na região do trauma. Ao exame físico, o paciente apresentava nódulo pulsátil de aproximadamente 1,5 cm de diâmetro, sem sopro ou frêmito, com diminuição da pulsatilidade à manobra de compressão proximal. Com o diagnóstico de pseudoaneurisma da ATS, foi realizada uma arteriografia para afastar lesões associadas, que confirmou o diagnóstico e eliminou a presença de lesões intracranianas (Figura 1a). O paciente foi submetido a tratamento cirúrgico por ligadura e ressecção do aneurisma (Figura 1b). A evolução foi favorável e não houve recorrência do aneurisma.
## Caso 2
Paciente do sexo masculino de 27 anos, com história de agressão direta (soco) em região frontoparietal 15 dias antes. Desenvolveu uma massa pulsátil indolor na região temporal diagnosticada como pseudoaneurisma de ATS, manejada com ressecção cirúrgica e ligadura arterial. O paciente apresentou boa evolução, sem recorrência do aneurisma (Figura 2).
## Caso 3
Paciente do sexo masculino de 29 anos, vítima de acidente automobilístico de alta energia. Ainda em cuidados intensivos devido ao coma, foi evidenciado aumento de volume na região temporal direita diagnosticado como pseudoaneurisma. Foi submetido ao tratamento cirúrgico com boa evolução (Figura 3).
## Discussão
Após a primeira descrição, várias séries de pseudoaneurismas da ATS têm sido relatadas na literatura. Embora tenha sido primariamente descrito após trauma penetrante 5 , a grande maioria dos casos de pseudoaneurisma da ATS ocorre em decorrência de traumatismo fechado 6 . Está associado à prática de vários esportes como hóquei, rúgbi, squash e beisebol. Também é descrito associado a lesões penetrantes por artroplastia temporomandibular, transplante de cabelo, remoção de cistos, lacerações e ferida por arma de fogo 1 . Pelo mecanismo de trauma, é mais prevalente em homens jovens (mais de 80% são do sexo masculino, com média de idade de 33 anos) . Apenas 5% dos aneurismas de ATS não são pseudoaneurismas de origem traumática, sendo classificados como aneurismas ateroscleróticos ou congênitos 9 . A localização mais comum é no ramo anterior da ATS. Raramente está localizado proximalmente ou no ramo posterior. A região do ramo anterior mais comumente lesada é aquela que cruza a proeminência relacionada à junção da fáscia temporal superficial e à linha temporal superior, na porção anterior do crânio. O ramo anterior é literalmente esmagado contra essa proeminência óssea durante o trauma, levando à formação do pseudoaneurisma .
A apresentação clínica é normalmente benigna, já que a ruptura é rara. Existem apenas três casos de ruptura relatados na literatura . A maioria dos pseudoaneurismas (90%) apresenta-se como massa pulsátil assintomática, que pode variar entre 0,5 e 5,7 cm (maioria entre 1 e 1,5 cm de diâmetro). Após o episódio do trauma, o pseudoaneurisma torna-se evidente após duas a seis semanas 1 . Eventualmente, apresenta-se doloroso, causando desconforto local ou uma série de sintomas vagos. Pode estar acompanhado de tontura, alterações visuais ou alterações neurológicas, porém relacionadas ao trauma prévio 1 . Nesses casos, a tomografia computadorizada ou a arteriografia devem ser realizadas para excluir possíveis lesões associadas . Na grande maioria das vezes, o exame clínico e a anamnese detalhada dirigida principalmente para a existência de história de trauma recente são suficientes para o diagnóstico. Deve-se realizar sempre a palpação cuidadosa da massa e a oclusão da ATS proximalmente para verificar a diminuição do pulso local, o que ajuda no diagnóstico diferencial. Exames complementares são reservados aos casos em que não existe uma história típica de traumatismo recente .
Historicamente, vários tratamentos foram utilizados para o tratamento do pseudoaneurisma de ATS, desde a simples compressão 16 até a ligadura proximal da artéria carótida comum 1 . O tratamento deve ser indicado para o alívio dos sintomas locais, por razões estéticas ou para prevenir a ruptura. O tratamento atual é a ressecção do aneurisma após ligadura proximal e distal do ramo anterior sob anestesia local ou geral 5 . O tratamento é seguro e evita a recorrência. A embolização é reservada aos pseudoaneurismas localizados na porção proximal da ATS, onde o acesso cirúrgico é complexo e está acompanhado de risco de lesão da glândula parótida e do nervo facial [bib_ref] Aneurysm of the superficial temporal artery presenting as a parotid mass, Buckspan [/bib_ref]. O índice de complicação da embolização varia de 1 a 3% e inclui inflamação local acompanhada de dor, trombose induzida pelo cateter, ruptura do aneurisma e embolização acidental da artéria carótida interna [bib_ref] Embolization of traumatic aneurysm of the maxillary artery, Field [/bib_ref].
O pseudoaneurisma da ATS, embora raro e de baixa morbidade, deve ser levado em consideração em todo paciente que apresenta história de traumatismo craniano fechado. Todo profissional que atende urgência/emergência deve estar alerta a essa entidade clínica.
# Introduction
Traumatic pseudoaneurysms of the superficial temporal artery (STA) are rare and account for less than 1% of aneurysms reported. The first case described was reported by Thomas Bartholin in 1740 2,3 and since then approximately 200 cases have been described in the literature. Although associated with low morbidity, a pseudoaneurysm of the STA can cause a series of local symptoms and can even rupture. The authors describe three cases of pseudoaneurysm of the STA secondary to blunt head trauma.
## Descriptions of the cases
## Case 1
The patient was a 22-year-old male complaining of discomfort in the anterior, right, parietal region associated with an increase in volume at the site.
He reported an automobile accident 3 months previously in which he had received a blunt trauma to the right temporal region. He described bleeding at the time of the trauma and subsequent periorbital hematoma. He was assessed and treated with suture of the wound and observation; but after 20 days, he noted a progressive increase in volume at the trauma site. On physical examination, the patient exhibited a pulsating nodule of approximately 1.5 cm in diameter, with no murmur or thrill, and the pulsation reduced in response to proximal manual compression. He was diagnosed with pseudoaneurysm of the STA and arteriography was conducted to investigate the possibility of additional lesions, which confirmed the diagnosis and ruled out intracranial injuries [fig_ref] Figura 1: Caso 1 [/fig_ref]. The patient underwent surgical treatment by ligature and resection of the aneurysm [fig_ref] Figura 1: Caso 1 [/fig_ref]. He recovered well and there was no recurrence of the aneurysm.
## Case 2
A 27-year-old male patient had suffered direct aggression (a punch) to the anterior parietal area 15 days previously. He developed a painless pulsating mass in the temporal area that was diagnosed as a pseudoaneurysm of the STA and managed with surgical resection and arterial ligature. He recovered well and there was no recurrence of the aneurysm [fig_ref] Figura 2: Tratamento cirúrgico do segundo caso por ligadura arterial e aneurismectomia [/fig_ref].
## Case 3
This patient was a 29-year-old male who had suffered a high energy automobile accident. While in intensive care because of coma, an increase in volume in the right temporal region was observed and diagnosed as a pseudoaneurysm. He was treated surgically with good results [fig_ref] Figura 3: Tratamento cirúrgico do terceiro caso por ligadura arterial e ressecção do pseudoaneurisma [/fig_ref].
# Discussion
Since the first description, several series of pseudoaneurysms of the STA have been reported in the literature. Although it was first described as secondary to penetrating trauma, 5 the great majority of cases of STA pseudoaneurysm are the result of blunt traumas. It is associated with participation in several sports, such as hockey, rugby, squash, and baseball. It has also been described in relation to penetrating injuries caused by temporomandibular joint replacement, hair transplants, removal of cysts, lacerations, and gunshot wounds. The trauma mechanism and therefore the condition itself is most prevalent among young men (more than 80% of reports were in males, with a mean age of 33 years). Just 5% of aneurysms of the STA are not pseudoaneurysms of traumatic origin and are classified as atherosclerotic or congenital aneurysms. 9 The most common site is the anterior branch of the STA. They are rarely located proximally or along the posterior branch. The part of the anterior branch that is most often injured is where it crosses the prominence at the junction between the superficial temporal fascia and the superior temporal line, on the anterior part of the skull. The anterior branch is literally crushed against this bony prominence during the trauma, leading to formation of the pseudoaneurysm. Clinical presentation is normally benign, since rupture is rare. There are only three cases of rupture reported in the literature. The majority of pseudoaneurysms (90%) present as an asymptomatic pulsating mass, which can range in size from 0.5 to 5.7 cm (the majority are 1 to 1.5 cm in diameter). After the trauma episode, the pseudoaneurysm becomes evident in 2 to 6 weeks. 1 They can occasionally be painful, causing localized discomfort or a series of vague symptoms. There may also be dizziness, visual disturbances, or neurological changes, but these are related to the prior trauma. In these cases, computed tomography or arteriography should be used to rule out possible associated injuries. In the great majority of cases, detailed clinical examination and patient history, primarily focused on the possibility of recent trauma, are sufficient to make a diagnosis. Careful palpation of the mass and proximal occlusion of the STA should always be performed to test for a reduction in the pulse, which is useful for differential diagnosis. Additional examinations are only needed if the typical history of recent trauma is absent. Historically, many treatments have been used to mange pseudoaneurysms of the STA, ranging from simple compression 16 to proximal ligature of the common carotid artery. 1 Treatment should be directed at relief of local symptoms, esthetic objectives, or prevention of rupture. Currently, the preferred treatment is resection of the aneurysm after proximal and distal ligature of the anterior branch under local or general anesthesia. This treatment is safe and prevents recurrence. Embolization is reserved for pseudoaneurysms located in the proximal part of the STA, where surgical access is complex and there is risk of injury to the parotid gland and facial nerve. [bib_ref] Aneurysm of the superficial temporal artery presenting as a parotid mass, Buckspan [/bib_ref] The complication rate of embolization ranges from 1 to 3% and conditions include local inflammation with pain, thrombosis caused by the catheter, rupture of the aneurysm, and accidental embolization of the internal carotid artery. [bib_ref] Embolization of traumatic aneurysm of the maxillary artery, Field [/bib_ref] Although a pseudoaneurysm of the STA is rare and has a low rate of morbidity, it should be considered in all patients with a history of blunt head trauma. All professionals who work in emergency should be aware of this clinical entity.
[fig] Figura 1: Caso 1. (a) Arteriografia evidenciando pseudoaneurisma do ramo anterior da artéria temporal superficial; (b) Tratamento cirúrgico através de ligadura arterial e ressecção do pseudoaneurisma. J Vasc Bras. 2018 Jan.-Mar.; 17(1):76-80 [/fig]
[fig] Figura 2: Tratamento cirúrgico do segundo caso por ligadura arterial e aneurismectomia. [/fig]
[fig] Figura 3: Tratamento cirúrgico do terceiro caso por ligadura arterial e ressecção do pseudoaneurisma. J Vasc Bras. 2018 Jan.-Mar.; 17(1):76-80 [/fig]
[fig] Figure 1: Case 1. (a) Arteriography showing pseudoaneurysm of the anterior branch of the superficial temporal artery; (b) Surgical treatment by arterial ligature and resection of the pseudoaneurysm. J Vasc Bras. 2018 Jan.-Mar.; 17(1):76-80 [/fig]
[fig] Figure 2: Surgical treatment of the second case by arterial ligature and aneurysmectomy. [/fig]
[fig] Figure 3: Surgical treatment of the third case by arterial ligature and resection of the pseudoaneurysm. J Vasc Bras. 2018 Jan.-Mar.; 17(1):76-80 [/fig]
|
A simple, clinically applicable motor learning protocol to increase push-off during gait: A proof-of-concept
ObjectiveTask-specific training is often used in functional rehabilitation for its potential to improve performance at locomotor tasks in neurological populations. As push-off impairment are often seen with these patients, this functional approach shows potential to retrain gait overground to normalize the gait pattern and retrain the ability to improve gait speed. The main objective of this project was to validate, in healthy participants, a simple, low-cost push-off retraining protocol based on task-specific training that could be implemented during overground walking in the clinic.Methods30 healthy participants walked in an 80-meter long corridor before, during, and after the application of an elastic resistance to the right ankle. Elastic tubing attached to the front of a modified ankle-foot orthosis delivered the resistance during push-off. Relative ankle joint angular displacements were recorded bilaterally and continuously during each walking condition.ResultsOn the resisted side, participants presented aftereffects (increased peak plantarflexion angle from 13.4±4.2˚to 20.0±6.4˚, p<0.0001 and increased peak plantarflexion angular velocity from 145.8±22.7˚/s to 174.4±37.4˚/s, p<0.0001). On the non-resisted side, aftereffects were much smaller than on the resisted side suggesting that the motor learning process was mainly specific to the trained leg.ConclusionThis study shows the feasibility of modifying push-off kinematics using an elastic resistance applied at the ankle while walking overground. This approach represents an interesting venue for future gait rehabilitation.
[formula] a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 [/formula]
# Introduction
After neural injury, gait control is often compromised. Reduced push-off force output is one of the most prevalent gait impairments, and correlates to a reduced gait speed [bib_ref] Walking ability is related to muscle strength in children with cerebral palsy, Eek [/bib_ref] [bib_ref] Plantarflexor weakness as a limiting factor of gait speed in stroke subjects..., Nadeau [/bib_ref] [bib_ref] The relationship of lower-extremity muscle torque to locomotor performance in people with..., Kim [/bib_ref] [bib_ref] Strength of lower limb related to gait velocity and cadence in stroke..., Bohannon [/bib_ref] [bib_ref] Mechanisms of disturbed motor control in ankle weakness during gait after stroke, Lamontagne [/bib_ref] , limiting patients in their mobility and activities of daily living. After conventional rehabilitation, gait speed only partially recovers [bib_ref] Deficit and change in gait velocity during rehabilitation after stroke, Goldie [/bib_ref] [bib_ref] Rehabilitation of Gait Speed After Stroke: A Critical Review of Intervention Approaches, Dickstein [/bib_ref] [bib_ref] Motor recovery after stroke: a systematic review of the literature, Hendricks [/bib_ref] , and push-off force output remains sub-optimal. It is therefore warranted to improve the retraining of push-off force output in these populations in order to reduce maladaptive compensations and retrain the ability to modulate/increase gait speed.
Experimentally controlled error-based motor learning, where an externally applied perturbation is used to train the emergence of new motor programs, has been proposed as a potential protocol for training motor recovery after injury [bib_ref] Robot-assisted adaptive training: custom force fields for teaching movement patterns, Patton [/bib_ref]. Considering that the neural control of gait results from the interaction between voluntary commands, involuntary rhythmic movement generators, and complex phase-dependent sensorimotor integration (see for reviews, it is impractical to use purely explicit learning protocols during rehabilitation. Error-based motor learning therefore represents an approach of particular interest for gait retraining, both from its implicit nature [bib_ref] Robot-assisted adaptive training: custom force fields for teaching movement patterns, Patton [/bib_ref] and its taskspecificity [bib_ref] Effects of task-oriented circuit class training on walking competency after stroke: a..., Wevers [/bib_ref] [bib_ref] Gait training facilitates push-off and improves gait symmetry in children with cerebral..., Lorentzen [/bib_ref] [bib_ref] Task-related circuit training improves performance of locomotor tasks in chronic stroke: a..., Dean [/bib_ref] [bib_ref] The effect of progressive task-oriented training on a supplementary tilt table on..., Kim [/bib_ref] [bib_ref] Effects of task-oriented training on mobility function in children with cerebral palsy, Salem [/bib_ref].
In neurologically impaired populations, error-based motor learning protocols have shown the capacity to improve overall walking ability [bib_ref] Modifications in ankle dorsiflexor activation by applying a torque perturbation during walking..., Blanchette [/bib_ref] [bib_ref] Robotic resistance/assistance training improves locomotor function in individuals poststroke: a randomized controlled..., Wu [/bib_ref] [bib_ref] Robotic resistance treadmill training improves locomotor function in human spinal cord injury:..., Wu [/bib_ref] [bib_ref] Size of kinematic error affects retention of locomotor adaptation in human spinal..., Yen [/bib_ref] [bib_ref] Using swing resistance and assistance to improve gait symmetry in individuals post-stroke, Yen [/bib_ref] [bib_ref] Locomotor adaptation to resistance during treadmill training transfers to overground walking in..., Yen [/bib_ref]. At the ankle joint specifically, results are more mitigated. For foot dorsiflexors during the swing phase, error-based motor learning was effective. This is an important finding, as improper activation of these muscles has been associated with increased risk of fall [bib_ref] Foot drop: where, why and what to do?, Stewart [/bib_ref]. The generalizability of these studies may be limited however, as they were performed on motorized treadmills, a situation that does not represent the real-life situations of patients. When assessed, the transfer of treadmill-induced aftereffects to overground walking is unfortunately incomplete [bib_ref] Locomotor adaptation to resistance during treadmill training transfers to overground walking in..., Yen [/bib_ref] [bib_ref] Split-belt treadmill adaptation transfers to overground walking in persons poststroke, Reisman [/bib_ref].
For ankle plantarflexors, the muscles responsible for modulating gait velocity in healthy and impaired populations [bib_ref] Strength of lower limb related to gait velocity and cadence in stroke..., Bohannon [/bib_ref] [bib_ref] Walking speed: reference values and correlates for older adults, Bohannon [/bib_ref] [bib_ref] Evidence for a non-linear relationship between leg strength and gait speed, Buchner [/bib_ref] , error-based motor learning has been less successful. Indeed, in a study using a robotized ankle-foot orthosis, Noel et al. [bib_ref] Using an electrohydraulic ankle foot orthosis to study modifications in feedforward control..., Noel [/bib_ref] were not able to produce aftereffects (a manifestation of motor learning) during the push-off phase of gait, the moment where ankle plantarflexors are at their maximal activation. Part of this inability to modify the locomotor pattern may be related to the fact that the study was here again performed on a treadmill.
While recent advances in robotic orthoses and exoskeletons have made it possible to apply controlled perturbations during gait in the laboratory setting to induce error-based motor learning [bib_ref] Contribution of feedback and feedforward strategies to locomotor adaptations, Lam [/bib_ref] [bib_ref] Effects of walking in a force field for varying durations on aftereffects..., Fortin [/bib_ref] [bib_ref] Chapter 8-challenging the adaptive capacity of rhythmic movement control: from denervation to..., Bouyer [/bib_ref] [bib_ref] The effects of powered ankle-foot orthoses on joint kinematics and muscle activation..., Sawicki [/bib_ref] , the majority of them are cumbersome, expensive and complex to use, making them impractical for standard clinical use [bib_ref] Recent developments and challenges of lower extremity exoskeletons. Spec Issue, Chen [/bib_ref] [bib_ref] Lower-limb robotic rehabilitation: literature review and challenges, Díaz [/bib_ref]. This limits large-scale implementation, and hence reduces substantially the possibility of having an impact on patients' quality of life.
Simpler potential solutions that may be more suitable for clinical use should be studied. For example, elastic-based perturbations that can induce error-based motor learning during gait [bib_ref] Timing-specific transfer of adapted muscle activity after walking in an elastic force..., Blanchette [/bib_ref] [bib_ref] Effects of repeated walking in a perturbing environment: a 4-day locomotor learning..., Blanchette [/bib_ref] would be an interesting low-cost, low maintenance alternative to robotic devices for the clinical setting. Furthermore, considering the limited transfer of learning from treadmill training to overground walking, overground training should also be prioritized.
As retraining of ankle plantarflexors after CNS injury is important in order to regain functional gait speed and return to community ambulation [bib_ref] Functional ambulation velocity and distance requirements in rural and urban communities. A..., Robinett [/bib_ref] , the primary objective of the present study was therefore to validate, in healthy participants, a simple, error-based gait retraining protocol that could be implemented during overground walking in the clinic. To do so, the protocol tested the effect of an elastic tubing resistance applied around the ankle using a modified ankle-foot orthosis. Considering the complex nature of the neural control of walking during push-off (involving interactions between positive sensory feedback and descending drive), a secondary objective of this study was to take advantage of differences in individual adaptation strategies to identify motor strategies that might be more efficient for transferring the adapted motor pattern to regular walking (measured as increases in aftereffect duration) and optimize retention over time. Based on error-based motor learning principles, we hypothesized that this simple, low-cost push-off retraining device would result in increased peak plantarflexion angle once removed.
# Materials and methods
## Participants
A convenience sample of 30 non-disabled participants (see [fig_ref] Table 1: Group demographic data [/fig_ref] for demographic data) was recruited from Université Laval's student population. They had to be naive to the task and aged between 18 and 65 years. The exclusion criteria were known history of neurological or musculoskeletal disorders that could interfere with task execution. All participants read and signed a consent form describing the experimental procedure and their involvement in the study. This protocol was approved by the local ethics committee (CIUSSS-CN #2016-578) and the experimental procedures were in accordance with the Declaration of Helsinki.
## Elastic resistance
Elastic tubing (10.5 cm-long; Thera-Band1 Silver; The Hygenic Corporation, Akron, Ohio) was used to create a 'force field' that resisted plantarflexion. It was attached to the front of the modified AFO and to a strap at the level of the fifth metatarsal head [fig_ref] Fig 1: Picture of the modified AFO with the elastic force field on [/fig_ref]. The elastic force perturbation pulled the foot upwards during swing and resisted push-off (max resistance reached at the end of push-off) but had little effect during the rest of the gait cycle (see Barthélemy et al., 2012 for more details [bib_ref] Rapid changes in corticospinal excitability during force field adaptation of human walking, Barthé Lemy [/bib_ref]. The stiff lateral stems of the AFO ensured that the elastic tubing did not induce compression of the ankle joint.
## Protocol
A simple, low-cost push-off retraining protocol was developed using an elastic resistance used to resist plantarflexion, thus retraining push-off. Participants had to walk at self-selected speed overground, while voluntarily increasing their descending output to overcome the resistance applied on a modified AFO. Participants walked overground in an 80-meter-long corridor. During each of 3 walking conditions, they wore a modified Klenzac ankle foot orthosis (AFO) on their right leg, on which elastic tubing was attached only during the second condition [fig_ref] Fig 1: Picture of the modified AFO with the elastic force field on [/fig_ref]. The first condition (BASELINE) consisted of walking 4x80 meters in the corridor and served to characterize baseline gait parameters. During the second condition (EXPOSURE; 4x80 meters), Theraband Silver elastic tubing [fig_ref] Fig 1: Picture of the modified AFO with the elastic force field on [/fig_ref] was attached to the modified AFO to create a resistance against ankle plantarflexion. The lateral stems of the AFO absorbed elastic tubing tension, and participants therefore only felt a resistance to angular movement, with no joint compression. They were instructed to "push against the elastic to overcome its resistance". The last condition (POST-EX-POSURE; 8x80 meters) was used to measure the persistence of the gait pattern modifications induced by training, and their duration following elastic removal. In order to capture the majority of possible aftereffects and to serve as a wash-out period, the POST-EXPOSURE recording session lasted twice as long as the other two conditions. Each participant chose their preferred gait speed during BASELINE. An experimenter specialized in clinical gait analysis walked slightly behind them during the 3 conditions and, during EXPOSURE and POST-EX-POSURE, systematically provided verbal feedback every 40 meters to minimize major gait deviations and reminded participants to push against the elastic tubing.
## Recordings
Relative ankle joint angles were recorded bilaterally and continuously during each walking condition using electrogoniometers (Biometrics Ltd., Ladysmith, Virginia). Data was transmitted to a desktop computer using wireless communication (Norangle; Noraxon USA inc., Scottsdale, Arizona) and saved at 1000 Hz/channel.
## Data processing
To minimize the amount of equipment worn by participants, individual gait cycles were identified from the continuous recordings using ankle angular velocity rather than foot switches. Gait cycles were aligned on peak plantarflexion angular velocity (an event occurring near toeoff) and a 500 ms window centered around this time was used for data extraction using a custom-made MatLab program (non-published method validated by visual inspection of the results). For each gait cycle, peak ankle plantarflexion angle and angular velocity (referred to as velocity throughout the article) were quantified. A step-by-step time course for these variables was then plotted together with a 95% confidence interval (95%CI) calculated from the mean of the last 50 baseline gait cycles.
## Statistics
To quantify the presence of motor adaptation during elastic exposure, and aftereffects after elastic removal (main study objective), the main variables were evaluated using 2 complementary methods: a time course analysis and a time point analysis.
Time course analysis was performed graphically using the confidence intervals [bib_ref] Inference by eye: confidence intervals and how to read pictures of data, Cumming [/bib_ref] , as previously described in Fortin et al. [bib_ref] Effects of walking in a force field for varying durations on aftereffects..., Fortin [/bib_ref]. Briefly, an 11-points moving average was calculated for the Exposure and Post-Exposure data. The number of consecutive strides outside of the 95%CI was counted and used to quantify the presence and duration of changes.
For time point analysis, the following 5 epochs were defined:
- "Baseline late": mean of the last 50 strides of the BASELINE period;
- "Exposure early": mean of the first 5 strides of the EXPOSURE period;
- "Exposure late": mean of the last 50 strides of the EXPOSURE period;
- "Post-exposure early": mean of the first 5 strides of the POST-EXPOSURE period;
- "Post-exposure late": mean of the last 50 strides of the POST-EXPOSURE period.
As data followed a normal distribution, a repeated measure ANOVA was applied separately to the difference in peak plantarflexion position and velocity, and for aftereffect durations. Level of significance was set at 0.05.
In the presence of the elastic, several motor strategies can be used to overcome the resistance: 1) returning to baseline ankle angle at push-off; 2) returning to baseline ankle push-off velocity; 3) a combination of 1) and 2). To identify which of these motor strategies might optimize retention over time (secondary study objective), a graph of individual relationship between changes in ankle position and changes in ankle velocity was produced using the following formula: 100 � (Exposure late-Baseline late)/Baseline late. A Pearson correlation was used to determine if the slope was significantly different from zero and three zones were arbitrarily added to the graph (±20% of angle change; see Results).
# Results
## Ankle kinematics
Fig 2A shows the peak plantarflexion time course of a representative participant. During the BASELINE condition, baseline late peak plantarflexion angle was 14.6±2.3˚. With the elastic, peak plantarflexion angle was initially reduced to 9.2±1.6˚(exposure early; p<0.05). The participant then gradually adapted to the force field by increasing peak plantarflexion to 17.7 ±3.1˚. Upon elastic removal, there was a significant overshoot in peak plantarflexion to 25.9 ±2.5˚(p<0.05) that gradually returned to baseline values in post-exposure late (13.9±1.9˚; p>0.05). Using a graphical time course analysis, these aftereffects lasted 148 strides for this participant.
Group effects on peak plantarflexion angle. As a group, ANOVA results were statistically significant (p<0.0001, F (4, 29) = 20.91, R 2 = 0.42). Participants showed a statistically significant increase in peak plantarflexion, from 13.4± 4.2˚in baseline late to 20.0±6.4˚in post-exposure early (p<0.0001) for the right lower limb [fig_ref] Fig 2: Effect of the force field on the peak plantarflexion angle [/fig_ref]. Their mean aftereffects duration was 150±156.7 strides (range: 0-462). They returned to 14.5±4.7˚in post-exposure late, a value not statistically different from baseline. Moreover, when looking at the exposure early vs. late, the peak plantarflexion varies from 14.0±4.9˚to 13.1±5.4˚, which is not statistically different (p>0.05).
Regarding the contralateral leg, while small differences in peak plantarflexion were only found during exposure (20.8±4.2˚in baseline late and 22.6±4.5˚in exposure late; p = 0.002), no significant aftereffects were measured. This suggests that no adaptation occurred on the contralateral side.
Group effects on peak plantarflexion angular velocity. In addition to changes in peak plantarflexion angle, there was a significant change in plantarflexion velocity (p<0.0001, F (4, 29) = 15.30, R 2 = 0.35), for the right (experimental) lower limb [fig_ref] Table 2: Group means for the kinematic values [/fig_ref]. When baseline late was compared to post-exposure early (p<0.0001), going from 145.8±22.7˚/s to 174.4±31.0˚/s. On the other hand, post-exposure late velocity was not statistically different from baseline values (p>0.05).
The table shows the peak plantarflexion (Peak PF) mean values for the group, with its standard deviations (SD). The peak plantarflexion velocity with its standard deviations is also shown. ANOVA results for the peak plantarflexion angle and angular velocity are presented below each variable. The aftereffect range (AE range) is presented for each way of calculating it: with the 11-points moving average calculated from the 50 last strides of the baseline condition and from the post-exposure condition. For the left lower limb, no significant changes were observed regarding peak plantarflexion angular velocity.
## Aftereffects duration variability across participants
As mentioned above, the duration of aftereffects ranged from 0 to 462 in the group, with a mean duration of 150±156.7 strides, showing a large variability across participants. This is, however, when the aftereffects duration is calculated based on baseline late data. As presented in [fig_ref] Table 2: Group means for the kinematic values [/fig_ref] , this variability is reduced when aftereffects duration is measured relative to the 50 last strides of the post-exposure condition. In this case, mean aftereffects duration is 80.6±91.2 strides, ranging from 0 to 301 strides. However, even with this "correction", a large variability remains. Detailed analysis of individual adaptation time courses revealed different strategies across participants. Three representative examples are presented in [fig_ref] Fig 3: Fig 3 [/fig_ref] Participants represented in the left column (strategy A) showed the longest aftereffect durations. Looking at their differences with baseline in peak ankle plantarflexion angle and peak plantarflexion velocity, these participants actually pushed more (increased peak plantarflexion angle compared to baseline; arrow A) and faster (increased peak plantarflexion velocity compared to baseline; arrow B) in the presence of elastic resistance than during baseline walking. On the contrary, participants represented in the right column (strategy C) had the least aftereffects. These participants decreased their peak push-off angle (arrow C) and velocity (arrow D) relative to baseline. Finally, the participants represented in the centre column (strategy B) either increased their push-off velocity OR increased plantarflexion angle while maintaining the other variable around the baseline value. They had intermediate aftereffect durations.
The relationship between individual strategies and aftereffect duration is presented for the whole group in [fig_ref] Fig 4: Continuum of all participants in regard of their strategies [/fig_ref] The latter plots the difference in velocity as a function of the difference in ankle position with associated aftereffect duration in the label below each strategy. Each participant is represented by a point on the graph. On the X axis, the difference in peak plantarflexion angle is represented by the exposure late values minus the baseline late values, while on the Y axis, the same difference is shown for plantarflexion (push-off) velocity. As shown on the graph, the R 2 value is 0.497. On this figure, it is possible to see that individual strategies are found as a continuum between the three examples presented in [fig_ref] Fig 3: Fig 3 [/fig_ref] one in the upper-right Examples of adaptive strategies. This figure represents the three adaptive strategies found during the exposure condition in our participants. For each column, the peak plantarflexion angle and the peak plantarflexion velocity is shown for a representative participant. Strategy A represents a participant with an increased peak plantarflexion angle and velocity maintained through the whole condition. Strategy B represents a participant that increased its plantarflexion velocity at the very beginning of the condition without maintaining this increase. Strategy C represents a participant that decreased its plantarflexion angle and velocity during the exposure condition.
## Relation between kinematic parameters
The curve fitting for the linear regression presented on
# Discussion
The main goal of this study was to determine if a motor adaptation protocol similar to those used for dorsiflexor muscles could be adapted to retrain ankle plantarflexors during push-off. By comparing ankle kinematics activity before, during and after exposure to an elastic force field applied during walking overground, our results suggest that the neural control of soleus can indeed be modified if training is performed under specific conditions.
## Aftereffects following the elastic removal
Upon removing the elastic resistance, participants showed increased peak plantarflexion angle and velocity that were maintained over several gait cycles. These results are equivalent to those presented after TA adaptation [bib_ref] Modifications in ankle dorsiflexor activation by applying a torque perturbation during walking..., Blanchette [/bib_ref] [bib_ref] Walking while resisting a perturbation: effects on ankle dorsiflexor activation during swing..., Blanchette [/bib_ref]. Therefore, the presence of aftereffects after the elastic removal supports the hypothesis that it is possible to modify the feedforward (central) command controlling ankle plantarflexors during push-off. Similar modifications in feedforward command were demonstrated at other joints previously [bib_ref] Contribution of feedback and feedforward strategies to locomotor adaptations, Lam [/bib_ref] [bib_ref] Timing-specific transfer of adapted muscle activity after walking in an elastic force..., Blanchette [/bib_ref] [bib_ref] Characteristics of the gait adaptation process due to split-belt treadmill walking under..., Yokoyama [/bib_ref] [bib_ref] Locomotor adaptation to a soleus EMG-controlled antagonistic exoskeleton, Gordon [/bib_ref] [bib_ref] Adaptive representation of dynamics during learning of a motor task, Shadmehr [/bib_ref] , but initial work at the ankle during push-off had previously been unsuccessful [bib_ref] Using an electrohydraulic ankle foot orthosis to study modifications in feedforward control..., Noel [/bib_ref]. The mean duration of this aftereffect was quite long on average (180 cycles) when compared to TA adaptation (20 cycles; [bib_ref] Walking while resisting a perturbation: effects on ankle dorsiflexor activation during swing..., Blanchette [/bib_ref].
This increased plantarflexion during post-exposure is providing further evidence that the aftereffects resulted from a continued use of the adapted motor pattern, and not just from a change in the way participants walked.
## Adaptive strategies
Figs 3 and 4 together suggest that for push-off adaptation, only looking at peak plantarflexion angular position changes might not be sensitive enough to capture the adapted motor strategy used by our participants. Adding information about push-off velocity helps predicting if a participant is going to show large aftereffects or not. Indeed, when participants increased both peak plantarflexion angle AND peak plantarflexion velocity during push-off, they showed longer aftereffects than if either one of these strategies was used alone. This finding suggests that the explosive contraction aspect measured by push-off velocity could be very important for a more complete adaptation during push-off. This explosive aspect has been described previously as the ability to use the muscles' torque-producing capacity explosively and tend to be more conductive to explosive performance with concentric contractions [bib_ref] Contraction type influences the human ability to use the available torque capacity..., Tillin [/bib_ref]. As walking while resisting an elastic force field uses concentric contraction of ankle plantarflexors, this might explain the importance of the explosive aspect of the muscle contraction to predict responders. When designing force field adaptation protocols for the rehabilitation setting, it might therefore be very important to insist, by giving clear instructions to participants, on increasing push-off velocity.
## Effects on the contralateral leg
Regarding the contralateral leg, only a small and transient change in peak plantarflexion angle was observed during the exposure condition, with no carry-over to post-exposure. Push-off velocity, however, was not affected by the elastic force field training. This suggests that the motor learning process is mainly specific to the trained leg. Previous studies using split-belt training with stroke also suggest that training will mainly improve the trained (affected) side [bib_ref] Repeated split-belt treadmill training improves poststroke step length asymmetry, Reisman [/bib_ref].
## Ecological validity of the training environment and clinical implications
To our knowledge, this study is the first to report the ability to use force field adaptation to train plantarflexor muscles to produce more activity during push-off.
The fact that aftereffects consisting of increased plantarflexion were measured shows that even muscles with a strong positive feedback component such as ankle plantarflexors to their motor output are amendable to environmentally-driven plastic modifications in central drive. Such capacity therefore has the potential of being tapped into for the design of future neurorehabilitation protocols using phase and task-specific resistive training, such as can be delivered using robotized gait orthoses.
The increase in muscle activation will not only increase the number of motor units recruited, but in the long run, if this protocol is repeated over several sessions, it might, in addition, lead to structural modifications in muscle fibers, similar to strength training, thereby increasing muscle strength and muscle mass, two additional beneficial factors for rehabilitation. By increasing the strength of the muscles responsible for push-off, it might be possible to increase gait speed of patients. It is well known that lower limb strength is correlated to gait speed [bib_ref] Strength of lower limb related to gait velocity and cadence in stroke..., Bohannon [/bib_ref] [bib_ref] Walking speed: reference values and correlates for older adults, Bohannon [/bib_ref] [bib_ref] Evidence for a non-linear relationship between leg strength and gait speed, Buchner [/bib_ref]. Thus, by increasing their gait speed, it would be possible for patients to meet the minimal requirement to be functional in society. Moreover, in chronic stroke patients, it has been shown that task-related training circuits focusing on functional activities improve performance at locomotor tasks [bib_ref] Task-related circuit training improves performance of locomotor tasks in chronic stroke: a..., Dean [/bib_ref]. As this protocol involves walking overground while resisting the elastic, it could be easily integrated in activities of daily living, or into circuit training during rehabilitation. Finally, as discussed in Blanchette et al. [bib_ref] Walking while resisting a perturbation: effects on ankle dorsiflexor activation during swing..., Blanchette [/bib_ref] , an increase of 5˚in dorsiflexion range in a population of persons with a neuromuscular disease is considered to be clinically significant. In this study, we have shown an increase of almost 7˚towards plantarflexion in the experimental leg. This could mean that this protocol could possibly significantly improve the gait pattern of people living with a neurological impairment, while being specific to the leg you want to train. Further studies in clinical populations are required to better understand the functional implications of this modification in the gait cycle with regard to the overall gait pattern. Based on our findings, we would recommend that such resistance training in further studies or in clinical settings should focus on the impaired leg to maximise improvement.
## Strengths and limitations of the study
This study has some limitations. First, kinematic recordings were made only at the ankle. It is therefore possible that additional modifications in the gait pattern occurring at other joints were not quantified. A second limitation is the absence of kinetic variables. It would be interesting to record the anteroposterior ground reaction forces in future studies in order to validate the efficiency of the modification in the gait pattern. It would also have been interesting to compare gait speed before and after the exposure phase or during the exposure phase to ensure constant gait speed. As push-off training should improve gait speed, objectifying the magnitude of this effect would be clinically useful.
This study also has several strengths. this approach, using an AFO orthosis and elastic tubing, requires little materials and is cheap to implement, thereby facilitating its potential use in the clinic.
# Conclusion
These results show that it is possible to retrain push-off while walking overground with a protocol similar to that previously used for dorsiflexor training. This training results in increased peak plantarflexion angle and velocity in the majority of participants in a time period too short for changes in muscle structure to occur. The next step will be to test this protocol in neurological populations that have impaired push-off control, to objectify the presence of aftereffects and measure the generalizability of this approach as a potential intervention for gait neuro-rehabilitation.
Supporting information S1 File. Dataset.
[fig] Fig 1: Picture of the modified AFO with the elastic force field on. A picture of the modified Klenzac ankle foot orthosis representing the subject's movement (A) recorded by the electrogoniometer placed on the lateral stem of the modified AFO and the elastic resistance that participants had to resist (B). https://doi.org/10.1371/journal.pone.0245523.g001 [/fig]
[fig] Fig 2: Effect of the force field on the peak plantarflexion angle. (A) Time course for one representative participant of the peak plantarflexion angle (˚) during the three walking conditions. Each dot represents a single gait cycle. The 95% lower and upper confidence interval (dashed lines) is based on the mean of the moving average of the last 50 gait cycles peak plantarflexion from the post-exposure late. The black arrow indicates when the PF angle has returned to normal values (95% lower CI). (B) Means of the peak plantar flexion angle (˚) for the group. Asterisk, P < 0.0001. https://doi.org/10.1371/journal.pone.0245523.g002 [/fig]
[fig] Fig 3: Fig 3. Examples of adaptive strategies. This figure represents the three adaptive strategies found during the exposure condition in our participants. For each column, the peak plantarflexion angle and the peak plantarflexion velocity is shown for a representative participant. Strategy A represents a participant with an increased peak plantarflexion angle and velocity maintained through the whole condition. Strategy B represents a participant that increased its plantarflexion velocity at the very beginning of the condition without maintaining this increase. Strategy C represents a participant that decreased its plantarflexion angle and velocity during the exposure condition. [/fig]
[fig] Fig 4: Continuum of all participants in regard of their strategies. The X axis shows the difference in position, while the Y axis shows the difference for the push-off velocity, represented as percentage of change. Zone A represents a change of more than 20% in plantar flexion, zone B a change of ±20% in plantarflexion and zone C represents a change of more than -20% in plantar flexion. In each zone, the mean aftereffect duration is presented for aftereffects calculated from the baseline (AE1) and from the post-exposure late (AE2).https://doi.org/10.1371/journal.pone.0245523.g004 [/fig]
[table] Table 1: Group demographic data. [/table]
[table] Table 2: Group means for the kinematic values. [/table]
|
Coarse Particles and Heart Rate Variability among Older Adults with Coronary Artery Disease in the Coachella Valley, California
[bib_ref] Time-series studies of particulate matter, Bell [/bib_ref] [bib_ref] Reduced heart rate variability and mortality risk in an elderly cohort, Tsuji [/bib_ref] [bib_ref] Heart rate variability from 24-hour electrocardiography and the 2-year risk for sudden..., Algra [/bib_ref] [bib_ref] Prognostic value of heart rate variability in chronic congestive heart failure secondary..., Szabo [/bib_ref] [bib_ref] RR variability in healthy, middle-aged persons compared with patients with chronic coronary..., Bigger [/bib_ref] [bib_ref] Short-and long-term effects of cigarette smoking on heart rate variability, Hayano [/bib_ref] [bib_ref] Effects of aging and of chronic obstructive pulmonary disease on RR interval..., Pagani [/bib_ref] [bib_ref] Quantitative evaluation of cardiac parasympathetic activity in normal and diabetic man, Pfeifer [/bib_ref] [bib_ref] Postural hypotension and low R-R interval variability in Parkinsonism, spinocerebellar degeneration, and..., Kuroiwa [/bib_ref] [bib_ref] Particulate matter and heart rate variability among elderly retirees: the Baltimore 1998..., Creason [/bib_ref] [bib_ref] Ambient pollution and heart rate variability, Gold [/bib_ref] [bib_ref] Air pollution and heart rate variability among the elderly in Mexico City, Holguin [/bib_ref] [bib_ref] Daily variation of particulate air pollution and poor cardiac autonomic control in..., Liao [/bib_ref] [bib_ref] Effects of air pollution on heart rate variability: the VA normative aging..., Park [/bib_ref] [bib_ref] Heart rate variability associated with particulate air pollution, Pope [/bib_ref] [bib_ref] Ambient particulate air pollution, heart rate variability, and blood markers of inflammation..., Pope [/bib_ref] [bib_ref] Traffic-related pollution and heart rate variability in a panel of elderly subjects, Schwartz [/bib_ref] [bib_ref] Daily variation of particulate air pollution and poor cardiac autonomic control in..., Liao [/bib_ref] [bib_ref] Air pollution and daily mortality in the Coachella Valley, California: a study..., Ostro [/bib_ref] [bib_ref] Coarse and fine particles and daily mortality in the Coachella Valley, California:..., Ostro [/bib_ref] [bib_ref] Coarse and fine particles and daily mortality in the Coachella Valley, California:..., Ostro [/bib_ref]
# Materials and methods
Subject recruitment. The study protocol was approved by the Institutional Review Board of the Public Health Institute (Oakland, California). Study participants were recruited from a large cardiology practice and through newspaper advertisements from December 1999 through February 2000. Subjects were eligible if they were ambulatory adults ≥ 60 years of age; were not current smokers; had coronary artery disease manifested by at least one of the following: a) a history of angina Alterations in cardiac autonomic control, assessed by changes in heart rate variability (HRV), provide one plausible mechanistic explanation for consistent associations between exposure to airborne particulate matter (PM) and increased risks of cardiovascular mortality. Decreased HRV has been linked with exposures to PM 10 (PM with aerodynamic diameter ≤ 10 µm) and with fine particles (PM with aerodynamic diameter ≤ 2.5 µm) originating primarily from combustion sources. However, little is known about the relationship between HRV and coarse particles [PM with aerodynamic diameter 10-2.5 µm (PM 10-2.5 )], which typically result from entrainment of dust and soil or from mechanical abrasive processes in industry and transportation. We measured several HRV variables in 19 nonsmoking older adults with coronary artery disease residing in the Coachella Valley, California, a desert resort and retirement area in which ambient PM 10 consists predominantly of PM 10-2.5 . Study subjects wore Holter monitors for 24 hr once per week for up to 12 weeks during spring 2000. Pollutant concentrations were assessed at nearby fixed-site monitors. We used mixed models that controlled for individual-specific effects to examine relationships between air pollutants and several HRV metrics. Decrements in several measures of HRV were consistently associated with both PM 10 and PM 10-2.5 ; however, there was little relationship of HRV variables with PM 2.5 concentrations. The magnitude of the associations (~ 1-4% decrease in HRV per 10-µg/m 3 increase in PM 10 or PM 10-2.5 ) was comparable with those observed in several other studies of PM. Elevated levels of ambient PM 10-2.5 may adversely affect HRV in older subjects with coronary artery disease. and a positive ECG, echocardiographic or nuclear stress test, or angiography (n = 12), b) prior percutaneous coronary intervention (n = 1), c) prior coronary artery bypass surgery (n = 12), or d) a history of myocardial infarction at least 6 months before recruitment (n = 16) (of the 19 subjects in the study, almost all met at least two criteria for eligibility); and residence within 5 miles of either of the two fixed-site air quality monitoring stations in Coachella Valley (located in Palm Springs and Indio).
Exclusion criteria included conditions associated with autonomic dysfunction (e.g., diabetes, chronic renal failure, Parkinsonism, and chronic alcohol abuse), cardiac transplant, cardiac pacemaker, implantable defibrillator, atrial fibrillation, or significant cognitive impairment.
Data collection. During the initial inperson appointment, staff obtained written informed consent and administered a baseline questionnaire, which included questions on subject demographics, medical history, current medications, usual daily activities, and any limitations on activity. The information obtained in the baseline questionnaire was supplemented by abstracting photocopies of the individuals' medical records on standardized forms. Data abstracted from the medical records included, where available, left ventricular ejection fraction (LVEF), history of myocardial infarction, and medications prescribed. During this study, the participants remained under the medical supervision of their regular personal physicians.
Staff also measured the subjects' lung function at baseline using a portable Simplicity spirometer (Mallinckrodt, Inc., St. Louis, MO). Spirometry was conducted following the guidelines of the American Thoracic Society (1995), with reproducibility criteria modified slightly to accommodate the subjects' ages (i.e., results of at least three of the forced expiratory maneuvers were required to be within 15% of one another). Briefly, the subjects were seated and wore a nose clip for the spirometric maneuvers. Each subject performed at least four expiratory maneuvers. Spirometry was rescheduled for subjects who reported a respiratory infection in the preceding 3 weeks.
Twenty-four-hour ambulatory ECGs were digitally recorded for each subject at weekly intervals from 14 February through 31 May 2000, using lightweight Trillium 3000 Holter monitors with disposable electrodes (Forest Medical, Syracuse, NY). During the Holter monitoring, subjects performed their normal daily activities, except those that would interfere with the ECG recording, such as showering. In general, Holter monitoring began at the same time and day every week for each subject. In cases of missed appointments, subjects were rescheduled for monitoring within the next 2 days, if possible.
Staff followed a standardized protocol for subject preparation and placed five electrodes (two channels) in a modified V5 and aVF configuration similar to that used by [bib_ref] Heart rate variability associated with particulate air pollution, Pope [/bib_ref] [bib_ref] Ambient particulate air pollution, heart rate variability, and blood markers of inflammation..., Pope [/bib_ref]. Each Holter monitoring session began with a 20-min resting ECG with the subject supine, during which staff remained with the subject. At each session, staff gave the subject a simple 24-hr time-activity diary to record times spent indoors or outdoors, air conditioner (AC) use (yes or no), and whether windows were open during each 2-hr period (and one 6-hr block from 2400 hr to 0600 hr).
Staff downloaded each subject's monitoring data from a removable flashcard to a personal computer for storage and subsequent editing by an ECG technician. The subjects' physicians were sent a standard Holter report within 24 hr, which resulted in the identification of three subjects during the initial monitoring sessions who had experienced asymptomatic but potentially life-threatening arrhythmias. These patients underwent procedures to implant defibrillator/pacemaker devices and were dropped from the study; no additional recordings were undertaken for these three individuals. An additional subject was found to have continuous atrial fibrillation. None of these subjects' ECG data were included in the analysis. Thus, of the initial 23 subjects, we had multiple ECG recordings from 19 for the analysis.
Ambient pollutant data consisted of continuous measurements of PM 10 , PM 2.5 , and ozone, which were monitored at fixed-site stations operated by the SCAQMD in Indio and Palm Springs, located at either end of the population corridor in Coachella Valley. PM 10-2.5 data were derived by subtracting PM 2.5 mass concentrations from PM 10 . Although the SCAQMD also monitored for carbon monoxide and nitrogen dioxide during the study period, we did not use these data for our analysis because there were many days with missing values. Sulfur dioxide was not monitored in the valley at that time. We obtained daily meteorologic data collected at two valley airports (minimum, maximum, and mean temperature, as well as dew point, relative humidity, and barometric pressure) from the National Climatic Data Center (Asheville, NC).
Data from baseline questionnaires, medical records abstraction, pulmonary function testing, daily diaries, and extracted HRV variables were entered into a SAS database, with 10% double-data entry to check for accuracy. The database was then merged with air quality and meteorologic data for analysis using SAS (version 8; SAS Institute Inc., Cary, NC).
Data analysis. Only normal sinus R-R intervals were used in the HRV analysis. Artifacts, ectopy (both supraventricular and ventricular), and uninterpretable complexes were excluded. We examined time-domain, frequency-domain, and geometric HRV variables. Time-domain variables included a) the standard deviation of normal sinus rhythm ("normal-to-normal" or N-N) beats (SDNN), representing the average of the standard deviations of normal beats of successive 5-min blocks over the duration of the monitoring period (SDNN estimates overall HRV); b) the standard deviation of the average N-N intervals (SDANN) within successive 5-min blocks (an estimate of long-term components of HRV); and c) the root mean square of successive differences (r-MSSD), which is the square root of the mean of the sum of the squares of differences between adjacent normal R-R intervals, which estimates short-term components of HRV and is a sensitive indicator of vagal tone .
Frequency-domain analysis delineates the heart rate signal into its frequency components and quantifies them in terms of their relative intensity or power. We examined three frequency-domain variables: high frequency (HF), low frequency (LF), and total power. HF components (0.15-0.40 Hz) provide an index of parasympathetic activity, whereas LF components (0.04-0.15 Hz) are considered to encompass both sympathetic and parasympathetic activity [bib_ref] Spectral analysis of the heart rate signal, Cerutti [/bib_ref]. Total power is an indicator of overall HRV.
Geometric methods involve analysis of the sample density histogram of R-R interval durations. A plot of the distribution typically depicts the main peak as a triangular shape. The triangular index (TRII) provides an estimate of overall HRV that is more resistant to beat-labeling errors than are its time-and frequency-domain counterparts .
Because the Holter software did not allow for downloading time-domain HRV variables for monitoring periods other than the full 24 hr, we chose specific time intervals for which the HRV variables would be calculated and then extracted them manually into the database. For time-domain variables and the TRII, we chose 0600-0800 hr, 1800-2000 hr, and 24 hr. The two 2-hr intervals were selected to provide the most marked contrasts in ambient PM levels, which tended to be lowest in the early morning and highest in the early evening, based on examination of 24-hr continuous monitoring data from Palm Springs and Indio. Frequency-domain variables generally are measured in 5-min increments and are sensitive to physical activity patterns. Therefore, to reduce intersubject behavioral variability, we chose to examine two 5-min periods: a) minutes 6-10 of the Holter monitoring session, during which the subjects were supine after the hook-up; and b) 0301-0305 hr, when most individuals would be asleep.
In summary, for each individual monitoring day, we obtained SDNN, SDANN, r-MSSD, and TRII for the full 24-hr period and for two 2-hr periods in the morning and evening, as well as frequency-domain HRV variables for two 5-min intervals.
Statistical methods. Most of the HRV variables were log-normally distributed and were log-transformed for the analyses. We applied mixed linear regression models to the continuous HRV variables and pollution metrics, with random-effects parameters to control for interindividual variation and fixed-effects parameters to estimate relationships between the various pollutant metrics and changes in HRV.
We explored the independent influence of meteorologic factors, examining both concurrent and lagged values (up to 3 days) of temperature, humidity, dew point, and barometric pressure. Because only barometric pressure was associated with HRV metrics, it was retained in subsequent models. Air pollutant variables were entered individually into the models; we examined the impact of both concurrent and lagged pollutant values to allow for the possibility of delayed and cumulative effects. Therefore, for the 24-hr measures of HRV, single-day lags and moving averages of up to 4 previous days for each pollutant were considered. For HRV variables measured on a 2-hr (time domain and TRII) or 5-min (frequency domain) basis, we examined 2-, 4-, 6-, 8-, and 24-hr pollutant moving averages. Because HRV is related inversely to heart rate, the models included the subjects' average heart rate during the monitoring periods.
For some of the associations found, we conducted additional analyses to examine potential impacts of behavioral factors that might influence exposure. For example, for the 2-hr evening period, we examined the effect (based on responses in the daily diary) of subjects' keeping windows open, using AC, or being outdoors for > 1 hr. Each of these factors was included separately as a dichotomous variable in models that also included a PM metric (PM 10 , PM 10-2.5 , or PM 2.5 ). We also added an interaction term between the specific factor and the PM metric to these models (e.g., the use of AC between 1800 and 2000 hr and concurrent PM 10 ). [fig_ref] Table 1: Characteristics of HRV study population [/fig_ref] presents demographic and medical data for the 19 participants. The average number of HRV monitoring sessions per subject was 8.8 (range, 4-12). Descriptive statistics for the pollutant and meteorologic variables during the study period are presented in [fig_ref] Table 2: Coarse particles and heart rate variability Environmental Health Perspectives • VOLUME 114... [/fig_ref] , and the time-and frequency-domain HRV variables used in the analysis are summarized in [fig_ref] Table 3: Summary of HRV variables a and average heart rate [/fig_ref].
# Results
Evaluation of potential time-variant confounders through both simple correlation analysis and univariate regressions indicated that the pollutant variables were not confounded by any meteorologic variables. Although barometric pressure was often associated with the HRV measures, it had little impact on the associations of ambient pollutants with HRV. Therefore, the results presented are from fixed-effects models that included only the pollutant term and average heart rate as predictor variables.
Results of the analysis of time-domain HRV variables measured during the evening period (1800-2000 hr) are displayed in . These results indicated associations between decrements in SDNN, SDANN, and TRII in relation to increases in both PM 10 and PM 10-2.5 . The magnitude of the associations between SDNN and PM 10 or PM 10-2.5 increased as the averaging time increased up to 6 hr but began to decrease at 8 hr and diminished to nonsignificance when the averaging time was extended to the prior 24 hr. A similar pattern was observed for SDANN, whereas for TRII the coefficients for both PM 10 and PM 10-2.5 continued to increase modestly at 8 hr relative to an averaging time of 6 hr. There was no evidence of an association between PM 2.5 or ozone and these HRV variables. There was no association between any pollutant variable and r-MSSD, except for a marginally significant but positive association with PM 10-2.5 averaged over the preceding 24 hr.
In contrast to the regressions for the evening monitoring period, there were few associations during the morning monitoring period between pollutant metrics and timedomain variables (data not shown). PM 10-2.5 was associated with both SDNN and SDANN at lags up to 4 hr but not at 24 hr. PM 10 , PM 2.5 , and ozone were not associated with any HRV metrics in the morning session. In addition, there was again a marginally significant positive association between PM 10-2.5 averaged over 24 hr and r-MSSD.
Analysis of the frequency-domain variables during sleep (0300 hr) also indicated sporadic associations between HRV and PM metrics [fig_ref] Table 5: Regression coefficients for frequency-domain HRV variables in relation to PM metrics and... [/fig_ref]. For this monitoring period, the unlagged pollutant variables were measured over the prior hour (i.e., 0200-0300 hr). Total power was associated with all three particulate metrics. The strongest associations for PM 10 and PM 10-2.5 were averaged over the prior 4 hr, whereas for PM 2.5 , only the measurement in the prior hour was statistically significant. There were also several modest associations with changes in the HF and LF components, with no obvious patterns. Ozone was also associated with decreases in all three frequency-domain measures, although the coefficients were of borderline significance (p = 0.08 to 0.10). The daytime posthookup frequency-domain variables also showed no pattern of association with the pollutant metrics (data not shown).
Adding variables representing exposurerelated behaviors (e.g., use of AC) to the models generally resulted in modest increases in the magnitude and significance of the coefficients for PM 10 and PM 10-2.5 (data not shown). However, neither these behavioral variables nor the interactive term coefficients were statistically associated with the HRV metrics. The use of exposure adjustment factors did not alter the generally null to modest findings for PM 2.5 .
Several constitutional and clinical variables [age, sex, lung function, use of betablockers or angiotensin-converting enzyme (ACE) inhibitors, prior smoking status] did not exhibit an association with SDNN, nor did they have much, if any, effect on the magnitude or significance of PM 10-2.5 coefficients. In contrast, inclusion of LVEF in the model increased the absolute magnitude of the PM 10-2.5 coefficient by 36% [from -0.00072 (p = 0.02) to -0.00098 (p = 0.007)], whereas the LVEF coefficient was of borderline statistical significance (p = 0.09).
# Discussion
We found consistent associations of several PM metrics, notably PM 10 and PM 10-2.5 , with short-term decrements in several measures of HRV in a panel of older adults with coronary artery disease. The strongest associations were detected when PM measurements were taken within a few hours before the HRV measures. These associations, however, were no longer present when the PM averaging time was extended to 24 hr or longer. These observations suggest that if there are causal relationships between PM exposures and decreases in HRV, the effects likely occur in close temporal proximity to the exposures.
These findings accord with some previous epidemiologic studies of HRV [bib_ref] Acute exposure to environmental tobacco smoke and heart rate variability, Pope [/bib_ref] , although others have reported more prolonged effects [bib_ref] Particulate matter and heart rate variability among elderly retirees: the Baltimore 1998..., Creason [/bib_ref] [bib_ref] Ambient particulate air pollution, heart rate variability, and blood markers of inflammation..., Pope [/bib_ref]. [bib_ref] Ambient pollution and heart rate variability, Gold [/bib_ref] conducted 25-min ECG measurements in 21 older Boston residents weekly over a 3-month period. They reported significant associations of r-MSSD and SDNN with PM 2.5 within a few hours of obtaining the ECG data. No associations between PM 2.5 and HRV were seen at lags longer than 24 hr. In a subsequent study, however, the same researchers found somewhat stronger associations with 24-hr PM metrics than with 4-hr averages . [bib_ref] Ambient particulate air pollution, heart rate variability, and blood markers of inflammation..., Pope [/bib_ref] reported decrements in several HRV metrics associated with 24-hr averages of PM 2.5 measured up to 2 days before Holter monitoring, although the strongest associations were with same-day measurements. A recent study of 10 elderly subjects involving 2-hr controlled exposures to either filtered air or concentrated PM 2.5 also reported significant decrements in several HRV measures immediately postexposure, which tended to persist (albeit somewhat attenuated) at 24 hr postexposure [bib_ref] Elderly humans exposed to concentrated air pollution particles have decreased heart rate..., Devlin [/bib_ref]. In contrast, other investigators found that a 48-hr PM averaging time had the strongest associations with decrements in HRV [bib_ref] Effects of air pollution on heart rate variability: the VA normative aging..., Park [/bib_ref].
In our study population of individuals with coronary artery disease, we identified PM-associated decreases in SDNN, SDANN, and TRII, but little relationship with r-MSSD. Others have found decrements in SDNN and SDANN, with mixed results regarding r-MSSD [bib_ref] Heart rate variability associated with particulate air pollution, Pope [/bib_ref] [bib_ref] Ambient particulate air pollution, heart rate variability, and blood markers of inflammation..., Pope [/bib_ref]. It is possible that the variable results with the latter metric are caused partly by the effects of a variety of common cardiovascular medications on r-MSSD. [bib_ref] Daily variation of particulate air pollution and poor cardiac autonomic control in..., Liao [/bib_ref] examined HRV in 26 elderly residents of a Baltimore retirement home, reporting significant decreases in HF, LF, and SDNN in relation to indoor and outdoor PM 2.5 only among subjects with preexisting cardiovascular disease. Recently, [bib_ref] Traffic-related pollution and heart rate variability in a panel of elderly subjects, Schwartz [/bib_ref] reported stronger associations of PM 2.5 (especially black carbon) with HRV decrements in subjects with a prior myocardial infarction (n = 3) relative to the other subjects (n = 25), although this observation must be interpreted cautiously because of small numbers. Other studies have reported that subjects with cardiovascular disease may be at increased risk of PM-associated changes in HRV [bib_ref] Air pollution and heart rate variability among the elderly in Mexico City, Holguin [/bib_ref] [bib_ref] Effects of air pollution on heart rate variability: the VA normative aging..., Park [/bib_ref]. [bib_ref] Air pollution and heart rate variability among the elderly in Mexico City, Holguin [/bib_ref] reported decrements in HF and LF variables among 34 elderly nursing home residents with both PM 2.5 and ozone in Mexico City, especially among individuals with hypertension. However, we found little relationship between frequency-domain variables and either of these pollutants, nor did we observe that a history of hypertension affected the PM-HRV associations. However, the levels of both PM 2.5 and ozone were substantially greater in the Mexican study (means of 37.2 µg/m 3 PM 2.5 and 149 ppb ozone in Mexico City vs. 18.6 µg/m 3 and 37 ppb, respectively, in our study, representing the averages of the values recorded at Indio and Palm Springs). In addition, all of the subjects in our study had documented coronary artery a All coefficients and SE × 1,000. PM 10 indicates PM 10 levels measured in the hour just before the HRV measurement (0300 hr); PM 10 2 hr indicates PM 10 levels measured in the 2 hr before the HRV measurement, and so forth. Similar conventions apply to PM 2.5 , PM 10-2.5 , and ozone, except that ozone 8-hr indicates 8-hr averaged ozone levels (1900 hr-0300 hr).
disease, which could represent a more important determinant of susceptibility compared with hypertension alone. Direct comparisons of our quantitative results with those of prior air pollution-HRV investigations are not entirely appropriate because they involve multiple differences in study design (e.g., various pollutant mixtures, averaging times for both pollution and HRV, exposure scenarios, subjects' health status and medications, and Holter monitoring protocols). Nevertheless, our results suggest associations of HRV metrics with PM exposures of the same order of magnitude as several of those previously reported [fig_ref] Table 6: Comparisons of particle-associated decreases in SDNN per 10 µg/m 3 increase in... [/fig_ref] , although not all studies have identified a relationship between PM and HRV.
In contrast to prior studies, however, the strongest signals that we identified were associated with PM 10-2.5 of primarily geologic origins, which dominate the PM 10 mass in the Coachella Valley, as well as throughout much of the arid American West and Southwest. A chemical mass balance analysis of annual average particle composition in Indio conducted previously by the SCAQMD (1990) indicated that geologic and vehicular sources contributed approximately 59% and 8% of PM 10 mass, respectively, with high particulate metal concentrations of silicon, aluminum, iron, and calcium, markers of crustal sources. That we found associations of PM 10-2.5 with HRV decrements whereas others (e.g., [bib_ref] Daily variation of particulate air pollution and poor cardiac autonomic control in..., Liao [/bib_ref] did not may be a dose-related phenomenon: PM 10-2.5 levels were substantially higher in this study than in any of the other investigations. Even within our study, we identified stronger, more consistent associations with the evening than with the morning PM concentrations (mean PM 10-2.5 levels of 47.1 µg/m 3 and 18.3 µg/m 3 for the 2-hr evening and morning periods, respectively), suggesting a concentration-related effect. In addition, PM 10-2.5 composition in the valley likely differs from those found in urban settings. It is also possible that at least some of the particles of interest lie in the intermodal size range (where the PM 2.5 and PM 10-2.5 distributions overlap).
Several publications link ambient PM 10-2.5 with cardiovascular mortality [bib_ref] Airborne coarse particles and mortality, Castillejos [/bib_ref] [bib_ref] Associations between air pollution and mortality in Phoenix, Mar [/bib_ref] [bib_ref] Coarse and fine particles and daily mortality in the Coachella Valley, California:..., Ostro [/bib_ref] and morbidity [bib_ref] Interim results of the study of particulates and health in Atlanta (SOPHIA), Tolbert [/bib_ref]. A recent review of the limited epidemiologic evidence found support for associations between PM 10-2.5 and cardiorespiratory morbidity and mortality [bib_ref] Epidemiological evidence of effects of coarse airborne particles on health, Brunekreef [/bib_ref]. Toxicologic studies also indicate that some PM 10-2.5 may be at least as capable of eliciting proinflammatory effects and oxidative damage as PM 2.5 [bib_ref] Cytotoxicity and induction of proinflammatory cytokines from human monocytes exposed to fine..., Monn [/bib_ref] [bib_ref] Inflammatory effects of coarse and fine particulate matter in relation to chemical..., Schins [/bib_ref]. Although particles may initiate or enhance inflammatory processes in the airways, the HRV changes indicated by our results and those of others [bib_ref] Acute exposure to environmental tobacco smoke and heart rate variability, Pope [/bib_ref] occurred over a time course shorter than that typically linked with pollutant-induced inflammation, suggesting perhaps direct involvement of neural reflexes. Consistent with this observation are the results of a recent controlled exposure study of asthmatic and healthy young adults [bib_ref] Altered heart-rate variability in asthmatic and healthy volunteers exposed to concentrated ambient..., Gong [/bib_ref] , in which concentrated PM 10-2.5 exposures resulted in modest decrements in HRV with no evidence of airway inflammation in induced sputum or changes in exhaled nitric oxide. Using a PM 10-2.5 concentrator, [bib_ref] Altered heart-rate variability in asthmatic and healthy volunteers exposed to concentrated ambient..., Gong [/bib_ref] found reductions in several HRV measures in four healthy adults exposed for 2 hr to concentrations ranging from 46 to 197 µg/m 3 . Although the sample size was quite small, that investigation provides a modest degree of corroboration of our findings.
That we found sporadic and relatively weaker associations of HRV decrements with PM 2.5 is somewhat puzzling, because the ambient PM 2.5 concentrations in our study overlapped with those in several other investigations [bib_ref] Daily variation of particulate air pollution and poor cardiac autonomic control in..., Liao [/bib_ref]. However, the specific composition and sources of the particles may be important determinants of response. [bib_ref] Traffic-related pollution and heart rate variability in a panel of elderly subjects, Schwartz [/bib_ref] found that the black carbon fraction of PM 2.5 showed a stronger relationship with decrements in frequency-and time-domain HRV variables than did secondary particles (e.g., sulfates), although the latter were not measured directly in that analysis. In contrast, in a study of 34 elderly subjects in Seattle, Washington, where residential wood combustion is an important source of PM 2.5 , [bib_ref] Association between short term exposure to fine particulate matter and heart rate..., Sullivan [/bib_ref] found no association of several frequency-domain HRV variables with 1-hr, 4-hr, or 24-hr measurements of ambient or indoor PM 2.5 . However, in that study, concentrations were low relative to those in other epidemiologic studies, and there was limited variability in exposure (median interquartile range was 6 µg/m 3 ); thus, the extent to which wood combustion or other sources might have influenced these results cannot be determined. In a study of 39 boilermakers exposed to PM 2.5 occupationally, [bib_ref] The association of particulate air metal concentrations with heart rate variability, Magari [/bib_ref] reported significant associations of PM-associated vanadium and lead with increases in the SDNN index after adjusting for mean heart rate, age, and smoking status, suggesting that specific particulate metals may affect cardiac autonomic activity.
Having detected consistent associations between PM metrics and several HRV variables in this population of individuals with preexisting cardiovascular disease, we did not identify any subject characteristics, except LVEF, associated with heightened susceptibility to PM. However, with only 19 subjects, this secondary analysis can only be considered exploratory. Moreover, LVEF was strongly correlated with average heart rate (r = 0.46, p < 0.0001), which is inversely related to HRV.
# Conclusions
In this study of elderly subjects with documented coronary artery disease, we detected decrements in several HRV metrics that were consistently associated with elevated PM 10 and PM 10-2.5 concentrations. Associations of these HRV variables with PM 2.5 levels were generally much weaker. The magnitude of the associations (~1-4% decrease in HRV metrics per 10-µg/m 3 increase in PM 10 or PM 10-2.5 ) was comparable with those observed in other studies of PM 2.5 in urban areas. In arid environments characteristic of much of the American West and Southwest, elevated levels of ambient PM 10-2.5 may adversely affect HRV in older subjects with coronary artery disease.
The clinical significance of PM-associated HRV decrements remains to be established. Most studies linking decreases in HRV with increased mortality have examined 24-hr baseline HRV as a predictor of adverse events over the course of months to years. Although these and others' results may offer a partial mechanistic explanation for the repeated observations in time-series studies of PM-associated cardiovascular mortality, it is also possible that such acute decrements in HRV represent an epiphenomenon of some other unmeasured, underlying pathophysiologic processes. 131.4 (24 hr) PM 2.5 (24 hr) -3.5 (-1.9 to -5.1) -2.7 [bib_ref] Effects of air pollution on heart rate variability: the VA normative aging..., Park [/bib_ref] 31.6 (4 min) PM 2.5 (24 hr) -0.9 (-3.0 to 1.4) -2.8 [bib_ref] Association between short term exposure to fine particulate matter and heart rate..., Sullivan [/bib_ref] 49 (20 min) PM 2.5 (24 hr) 0.5 (-2.4 to 3.9) a 1.0 [bib_ref] Effects of particle size fractions on reducing heart rate variability in cardiac..., Chuang [/bib_ref] 33.9 (16 hr) PM 2.5-1 (4 hr) -1.4 (-3.0 to 0.2) a -4.2 Present study 73.2 (2 hr) PM 10 (2 hr) -1.2 (-0.3 to -2.1) -1.6 Present study 73.2 (2 hr) PM 10 (6 hr) -2.4 (-0.7 to -4.1) -3.3 Present study 73.2 (2 hr) PM 10-2.5 (2 hr) -1.2 (-0.2 to -2.2) -1.6 Present study 73.2 (2 hr) PM 10-2.5 (6 hr) -3.0 (-1.0 to -5.1) -4.1
Abbreviations: CI, confidence interval; RSP, respirable particles (< 3 µm) from environmental tobacco smoke. a Subjects with cardiovascular disease.
[table] Table 1: Characteristics of HRV study population (n = 19). [/table]
[table] Table 2: Coarse particles and heart rate variability Environmental Health Perspectives • VOLUME 114 | NUMBER 8 | August 2006 Descriptive statistics of pollutant and meteorologic variables.Table 4. Regression coefficients a for time-domain HRV variables measured in the evening (1800-2000 hr) in relation to different averaging times for PM 10 , PM 10-2.5 , and PM 2.5 . [/table]
[table] Table 3: Summary of HRV variables a and average heart rate. [/table]
[table] Table 5: Regression coefficients for frequency-domain HRV variables in relation to PM metrics and ozone. [/table]
[table] Table 6: Comparisons of particle-associated decreases in SDNN per 10 µg/m 3 increase in PM. [/table]
|
Statins reduce mortality in patients with COVID-19: an updated meta-analysis of 147 824 patients
a b s t r a c tObjectives: There is conflicting evidence about the efficacy of statin use in regard to clinical outcomes in patients with coronavirus disease 2019 . A systematic review and meta-analysis was performed to examine the effect of statin use on mortality in COVID-19 patients. Methods: The electronic databases were searched, from inception to March 3, 2021. Unadjusted and adjusted effect estimates with their 95% confidence intervals (95% CI) were pooled using random-effects models. Results: Twenty-five cohort studies involving 147 824 patients were included. The mean age of the patients ranged from 44.9 to 70.9 years; 57% of patients were male and 43% were female. The use of statins was not associated with mortality when applying the unadjusted risk ratio (uRR 1.16, 95% CI 0.86-1.57; 19 studies). In contrast, meta-analyses of the adjusted odds ratio (aOR 0.67, 95% CI 0.52-0.86; 11 studies) and adjusted hazard ratio (aHR 0.73, 95% CI 0.58-0.91; 10 studies) showed that statins were independently associated with a significant reduction in mortality. Subgroup analyses showed that only chronic use of statins significantly reduced mortality according to the adjusted models. Conclusions: The use of statins was found to be associated with a lower risk of mortality in COVID-19 patients based on adjusted effects of cohort studies. However, randomized controlled trials are still needed to confirm these findings.
# Introduction
The current coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory coronavirus 2 (SARS-CoV-2), remains a major public health problem across the globe, despite the availability of vaccines [bib_ref] Global dynamics of SARS-CoV-2 clades and their relation to COVID-19 epidemiology, Hamed [/bib_ref]. Consequently, there is a continuing need for effective pharmacological therapies that reduce the morbidity and mortality of patients with Statins are widely used drugs in current medical practice. Given their lipid-lowering effect, these drugs are a mainstay in the treat-ment of patients with dyslipidemia and atherosclerosis-related diseases [bib_ref] Safety and efficacy of statin therapy, Adhyaru [/bib_ref]. Recently, statins have emerged as a potential new therapy for patients with COVID-19 due to their pleiotropic effects [bib_ref] Pleiotropic Effects of Statins on the Cardiovascular System, Oesterle [/bib_ref]. However, there are conflicting data about the utility of statins in COVID-19 patients [bib_ref] Effect of statin therapy on SARS-CoV-2 infection-related, Masana [/bib_ref] [bib_ref] Statin Use and In-Hospital Mortality in Diabetics with COVID-19, Saeed [/bib_ref] [bib_ref] She nL , Zha oYC , Yua nY , et al ...., Zhang [/bib_ref]. Therefore, a systematic review and meta-analysis was performed to examine the effect of statin use on mortality in COVID-19 patients.
# Methods
This review was reported according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) statement [bib_ref] Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, Moher [/bib_ref]
## Search strategy
A search was performed in the following six electronic databases: Embase, PubMed, Web of Science, Scopus, CENTRAL, and Cochrane COVID-19 register. The search was conducted from inception to November 25, 2020, and was updated on March 3, 2021, and included, but was not limited to, the following keywords and their combinations: hydroxymethylglutaryl-CoA reductase inhibitors, statin, COVID-19, and SARS-CoV-2. The complete search strategy is shown in Supplementary Material [fig_ref] Table 1 continued: AF, atrial fibrillation [/fig_ref]. There were no restrictions on language or publication year. In addition, a hand-search of reference lists of included articles was also performed to identify further eligible studies.
## Eligibility criteria
Cohort studies and randomized controlled trials that evaluated the effect of statins on mortality in COVID-19 patients diagnosed by reverse transcription PCR were included. Case reports, case series, reviews, abstracts, and editorials were excluded.
## Study selection
All articles from the electronic search were downloaded into EndNote X8 and duplicates were removed. The title and abstract were assessed independently by three review authors (MSB, PC, and AAY) to identify potentially eligible studies. Two review authors (BMT and AAY) independently screened the full-text and recorded reasons for the exclusion. Any disagreement on title/abstract and full-text selection was resolved by a third review author (CDA).
## Exposure and outcome
The use of statins was defined as chronic (i.e., before hospital admission) or in-hospital administration (i.e., during hospitalization) of any type and dose of a statin. The outcome of interest was mortality. All author-reported definitions of mortality were used.
## Data extraction
Data from each included study were extracted independently by two review authors (BMT and IB) using a standardized data extraction form that had been piloted previously. Any disagreement was resolved by a third review author (CDA). If additional data were required, the corresponding author was contacted by e-mail to request this information. The following information was extracted: name of the first author, publication year, country, sample size, population, age, sex, comorbidities, use of statins, and mortality. If available, unadjusted and adjusted effect estimates were extracted.
## Risk of bias assessment
The risk of bias of cohort studies was evaluated using the Newcastle-Ottawa Scale (NOS). Each article was classified as follows: high risk of bias (0-4 points), moderate risk of bias (5-7 points), and low risk of bias (8-9 points). Three review authors (MSB, PC, and BMT) independently assessed the risk of bias and any disagreement was resolved by a third review author (CDA).
# Statistical analysis
All meta-analyses were performed using a random-effects model with an inverse-variance method. Between-study variance (tau-square, τ 2 ) was estimated using the Paule-Mandel method.
Unadjusted risk ratios (uRR), adjusted odds ratios (aOR), and adjusted hazard ratios (aHR) with their 95% CI were pooled for the assessment of the effect of statins on mortality. Heterogeneity among studies was evaluated using the chi-square test (threshold P < 0.10) and the I 2 statistic. Heterogeneity was defined as low if I 2 < 30%, moderate if I 2 = 30-60%, and high if I 2 > 60%. Subgroup analyses were conducted according to the timing of statin use (chronic versus in-hospital). Funnel plots were used to assess publication bias and the Egger's test was performed to measure asymmetry of funnel plots only if 10 or more studies were included. All meta-analyses were conducted using the 'meta' package from R 3.6.3. A two-tailed P < 0.05 was considered statistically significant.
# Results
## Study selection
The electronic search strategy initially identified 857 articles. After the removal of duplicates, 409 articles remained. Following the screening of studies by title/abstract, 359 articles were excluded. After the full-text assessment of 50 articles, 25 articles were excluded. Thus a total of 25 studies (all cohorts) were finally selected [bib_ref] Relation of Statin Use Prior to Admission to Severity and Recovery Among..., Daniels [/bib_ref] [bib_ref] Association of Statin Use With the In-Hospital Outcomes of 2019-Coronavirus Disease Patients:..., Fan [/bib_ref] [bib_ref] Risk Factors Associated With Mortality Among Patients With COVID-19 in Intensive Care..., Grasselli [/bib_ref] [bib_ref] Association between antecedent statin use and decreased mortality in hospitalized patients with..., Gupta [/bib_ref] [bib_ref] Clinical comorbidities, characteristics, and outcomes of mechanically ventilated patients in the State..., Krishnan [/bib_ref] [bib_ref] Beneficial Effect of Statins in COVID-19-Related Outcomes: A National Population-Based Cohort Study, Lee [/bib_ref] [bib_ref] Outcomes of Hospitalized COVID-19 Patients by Risk Factors: Results from a United..., Mallow [/bib_ref] [bib_ref] Effect of statin therapy on SARS-CoV-2 infection-related, Masana [/bib_ref] [bib_ref] Impact of prior statin use on clinical outcomes in COVID-19 patients: data..., Mitacchione [/bib_ref] [bib_ref] Estimating risk of mechanical ventilation and in-hospital mortality among adult COVID-19 patients..., Nicholson [/bib_ref] [bib_ref] Statin Therapy and the Risk of COVID-19: A Cohort Study of the..., Oh [/bib_ref] [bib_ref] Statins in patients with COVID-19: a retrospective cohort study in Iranian COVID-19..., Peymani [/bib_ref] [bib_ref] Atorvastatin associated with decreased hazard for death in COVID-19 patients admitted to..., Rodriguez-Nava [/bib_ref] [bib_ref] Risk Factors Associated With In--Hospital Mortality in a US National Sample of..., Rosenthal [/bib_ref] [bib_ref] Protective role of statins in COVID 19 patients: importance of pharmacokinetic characteristics..., Rossi [/bib_ref] [bib_ref] Statin Use and In-Hospital Mortality in Diabetics with COVID-19, Saeed [/bib_ref] [bib_ref] Is the use of ACE inb/ARBs associated with higher in-hospital mortality in..., Selçuk [/bib_ref] [bib_ref] Statin Use Is Associated with Decreased Risk of Invasive Mechanical Ventilation in..., Song [/bib_ref] [bib_ref] She nL , Zha oYC , Yua nY , et al ...., Zhang [/bib_ref] [fig_ref] Figure 1: Flow diagram of study selection [/fig_ref].
## Study characteristics
The main characteristics of the 25 included studies (147 824 patients) are summarized in [fig_ref] Table 1 continued: AF, atrial fibrillation [/fig_ref]. The mean age of the patients ranged from 44.9 to 70.9 years; 57% of patients were male and 43% were female. Overall, 32% of patients had received statins. Twenty studies reported chronic use of statins and five studies reported in-hospital use. The type of statin was reported in only eight studies. The most common types were atorvastatin (71%), rosuvastatin (13%), and simvastatin (13%). Most studies were from the United States of America ( n = 9) and Italy ( n = 4). The most frequent comorbidities were hypertension (51%), dyslipidemia (41%), and diabetes (33%). The majority of studies ( n = 17) only included hospitalized patients. However, four studies included outpatients and hospitalized patients and four studies included only patients admitted to the intensive care unit (ICU). The adjusted effect esti-mates and adjusted variables for each individual study are shown in Supplementary Material .
## Risk of bias assessment
The NOS evaluation scored 18 studies as having a low risk of bias, six studies as having a moderate risk of bias, and one study as having a high risk of bias .
## Mortality
In 19 studies (114 881 patients), the use of statins was not significantly associated with mortality in COVID-19 patients (uRR 1.16, 95% CI 0.86-1.57; I 2 = 99%) [fig_ref] Figure 2: Forest plot showing the unadjusted risk ratio between statin use and mortality... [/fig_ref]. The funnel plot did not show asymmetry ( Supplementary Material [fig_ref] Figure 1: Flow diagram of study selection [/fig_ref] and the Egger's test was not significant ( P = 0.66).
In 11 studies (102 996 patients), the use of statins was significantly associated with lower mortality in COVID-19 patients (aOR 0.67, 95% CI 0.52-0.86; I 2 = 79%) [fig_ref] Figure 3: Forest plot showing the adjusted odds ratio between statin use and mortality... [/fig_ref]. The funnel plot did not show asymmetry ( Supplementary Material [fig_ref] Figure 2: Forest plot showing the unadjusted risk ratio between statin use and mortality... [/fig_ref] and the Egger's test was not significant ( P = 0.42).
In 10 studies (44 033 patients), the use of statins was significantly associated with lower mortality in COVID-19 patients (aHR 0.73, 95% CI 0.58-0.91; I 2 = 64%) [fig_ref] Figure 4: Forest plot showing the adjusted hazard ratio between statin use and mortality... [/fig_ref]. The funnel plot showed asymmetry [fig_ref] Figure 3: Forest plot showing the adjusted odds ratio between statin use and mortality... [/fig_ref] , sug- gesting publication bias that was confirmed by the Egger's test ( P = 0.03).
## Subgroup analyses
Subgroup analysis by the timing of statin use (chronic versus in-hospital) revealed that only chronic use of statins significantly reduced mortality in COVID-19 patients according to the adjusted models [fig_ref] Figure 2: Forest plot showing the unadjusted risk ratio between statin use and mortality... [/fig_ref]. Moreover, no significant differences were found between the subgroups.
# Discussion
This study found that the use of statins was significantly associated with a lower risk of mortality compared to non-statin users based on adjusted estimates. In addition, the subgroup analyses showed that only chronic use of statins was associated with a significant risk reduction in mortality. The risk of bias was low to moderate in almost all studies.
Several pathways are involved in the pathogenesis of COVID-19. SARS-CoV-2 enters into cells using angiotensin-converting enzyme 2 (ACE2), which is expressed ubiquitously, with predominance in the lungs, heart, kidneys, and vascular system [bib_ref] COVID-19 Outbreak: Pathogenesis, Current Therapies, and Potentials for Future Management, Hossain [/bib_ref]. ACE2 plays a key role in the renin-angiotensin system (RAS) by negatively regulating RAS activation and attenuating the harmful effects of angiotensin II [bib_ref] ACE2, Much More Than Just a Receptor for SARS-COV-2, Samavati [/bib_ref]. After viral entry, a robust local and systemic inflammatory response is elicited leading, in some cases, to an overproduction of proinflammatory cytokines [bib_ref] COVID-19 Outbreak: Pathogenesis, Current Therapies, and Potentials for Future Management, Hossain [/bib_ref]. Furthermore, there are compelling data showing that COVID-19 patients exhibit a hypercoagulable state, as evidenced by a high incidence of thrombotic complications in these patients [bib_ref] The hypercoagulable state in COVID-19: Incidence, pathophysiology, and management, Abou-Ismail [/bib_ref].
Statins are a class of lipid-lowering agents that act primarily by inhibiting 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase [bib_ref] Safety and efficacy of statin therapy, Adhyaru [/bib_ref]. Statins have also been shown to have other pleiotropic effects, whereby they act through many mechanisms independent of low-density lipoprotein cholesterol reduction, including anti-inflammatory, antioxidative, anti-thrombotic, and immunomodulatory functions [bib_ref] Pleiotropic Effects of Statins on the Cardiovascular System, Oesterle [/bib_ref]. In addition, statins significantly increased the ACE2 expression in the heart and kidney in an animal model of atherosclerosis, suggesting a positive effect on RAS balance [bib_ref] Tissue specific up regulation of ACE2 in rabbit model of atherosclerosis by..., Tikoo [/bib_ref]. Although the clinical significance of the pleiotropic effects of statins remains controversial, there is evidence of clinical benefits in a diversity of diseases such as respiratory viral infections, bacterial pneumonia, and venous thromboembolism, among others [bib_ref] Pleiotropic Effects of Statins on the Cardiovascular System, Oesterle [/bib_ref]. Furthermore, statins, especially pitavastatin, could exert a direct antiviral effect by interacting with the main protease enzyme of SARS-CoV-2 ( Reiner et al., 2020 ). Overall, The distinction between the chronic and in-hospital use of statins is an important issue to highlight. In the subgroup analyses, it was found that only patients with chronic use had independently lower mortality. This suggests that prolonged exposure to statins would be required to manifest their beneficial effects in patients with COVID-19. However, these results should be considered only hypothesis-generating, since a detailed description about the timing of statin use was lacking in almost all studies. It is not known whether patients who reported pre-admission use of statins con-tinued to receive them during their hospitalization. Moreover, in the cases that reported in-hospital use, it is not clear whether it was initiated de novo or was a continuation of previous use. In the only two studies that reported continued use of statins during hospitalization [bib_ref] Association of Statin Use With the In-Hospital Outcomes of 2019-Coronavirus Disease Patients:..., Fan [/bib_ref] [bib_ref] Statin Therapy and the Risk of COVID-19: A Cohort Study of the..., Oh [/bib_ref] , only one reported a significant reduction in mortality in COVID-19 patients [bib_ref] Association of Statin Use With the In-Hospital Outcomes of 2019-Coronavirus Disease Patients:..., Fan [/bib_ref]. Therefore, randomized controlled trials are needed to clarify whether the de novo or continued administration of statins has a favorable impact in these patients.
There are few previously published systematic reviews examining the effect of statin use in COVID-19 patients [bib_ref] Statin therapy did not improve the in-hospital outcome of coronavirus disease 2019..., Hariyanto [/bib_ref]. Two reviews concluded that statin use was associated with a significant reduction in mortality or ICU admission. In contrast, one review concluded that statins did not improve in-hospital outcomes. The largest review was performed by , which only included 13 studies involving 52 122 patients. In addition, it was the only review that evaluated chronic and in-hospital use of statins; however, it only combined unadjusted estimates. Compared to these reviews, the present study included substantially more studies and patients. Furthermore, a significant number of adjusted estimates were pooled.
This study has some limitations. First, given that only observational studies were evaluated, there is still a risk of residual confounding that could alter the results. Second, heterogeneity was high in all estimates. Possible reasons for heterogeneity include sample size, types and timing of statins, and heterogeneous populations, among others. Third, there is a risk of misclassification about the timing of statin use (chronic versus in-hospital) due to a lack of detailed information. Finally, some studies included outpatients and others included only ICU patients. This difference in disease severity could be a potential source of selection bias, affecting the pooled effect estimates.
In conclusion, this review found that statins significantly reduce mortality in COVID-19 patients based on adjusted estimates of cohort studies. The subgroup analysis revealed that only patients who were chronically treated with statins had a significant benefit. However, large randomized controlled trials are still needed to confirm these findings. +
# Declarations
Funding: None. Ethical approval: Not applicable. Informed consent: Not applicable. Data availability: Data are available from the corresponding author upon reasonable request.
Conflict of interest: The authors declare that there is no conflict of interest.
# Author contributions
Carlos Diaz-Arocutipa: participated in database search, study review, data analysis, and manuscript preparation. Beatriz Melgar-Talavera: participated in database search, study review, and manuscript preparation. Ángel Alvarado-Yarasca: participated in database search, study review, and manuscript preparation. María M. Saravia-Bartra: participated in database search, study review, and manuscript preparation. Pedro Cazorla: participated in database search, study review, and manuscript preparation. Iván Belzusarri: participated in database search, study review, and manuscript preparation. Adrian V. Hernandez: participated in study review and manuscript preparation.
# Supplementary materials
Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.ijid.2021.08.004 .
[fig] Figure 1: Flow diagram of study selection. [/fig]
[fig] Figure 2: Forest plot showing the unadjusted risk ratio between statin use and mortality in COVID-19 patients. (uRR, unadjusted risk ratio; CI, confidence interval.) [/fig]
[fig] Figure 3: Forest plot showing the adjusted odds ratio between statin use and mortality in COVID-19 patients. (aOR, adjusted odds ratio; CI, confidence interval.) [/fig]
[fig] Figure 4: Forest plot showing the adjusted hazard ratio between statin use and mortality in COVID-19 patients. (aHR, adjusted hazard ratio; CI, confidence interval.) these mechanisms are the different tar gets where statins could act directly or indirectly during SARS-CoV-2 infection explaining their beneficial effect. [/fig]
|
Congenital Zika Syndrome and Extra-Central Nervous System Detection of Zika Virus in a Pre-term Newborn in Mexico
Background. During pregnancy, the Zika virus (ZIKV) replicates in the placenta and central nervous system (CNS) of infected fetuses; nevertheless, the ability of ZIKV to replicate in other fetal tissues has not been extensively characterized.Methods. We researched whether dissemination of congenitally-acquired ZIKV outside the CNS exists by searching for the accumulation of the viral envelope protein, ZIKV ribonucleic acid (RNA), and infectious viral particles in different organs of a deceased newborn with Congenital Zika Syndrome. A real-time qualitative polymerase chain reaction (qPCR) was used to detect ZIKV RNA in the brain, thymus, lungs, kidneys, adrenal glands, spleen, liver, and small intestine. The same tissues were analyzed by indirect immunofluorescence and immunoperoxidase assays using the monoclonal antibody 4G2 to detect ZIKV envelope antigens. Isolation of infectious ZIKV in a cell culture was carried out using brain and kidney samples.Results.A postmortem, virological analysis of multiple organs, such as the kidneys (epithelial cells in the renal tubules), lungs (bronchial epithelia), thymus (epithelial cells inside the Hassall's corpuscles), and brain (neurons, ependymal cells, and macrophages) revealed the presence of ZIKV RNA and envelope antigens. Other tissues of the deceased newborn tested positive by qPCR for Epstein-Barr virus and human herpesvirus 6, including the brain cortex (Epstein-Barr) and the thymus, kidneys, and adrenal glands (human herpesvirus 6). The kidneys were identified as a significant niche for viral replication, given that infectious particles were successfully isolated from renal tissues.Conclusions. Our findings demonstrate the ability of congenitally-acquired ZIKV to produce disseminated infections and the viral tropism towards epithelial cells.
Recently, ZIKV replication was observed in the epithelial cells of the proximal renal tubules of immunodeficient mice, as well as in primary human renal proximal tubular epithelial cells [bib_ref] Zika virus infects renal proximal tubular epithelial cells with prolonged persistency and..., Chen [/bib_ref]. Moreover, other types of primary human renal cells, such as the renal glomerular endothelial cells and the mesangial cells, are also permissive to ZIKV replication [bib_ref] Zika virus infection of the human glomerular cells: implications for viral reservoirs..., Alcendor [/bib_ref]. ZIKV ribonucleic acid (RNA) has been detected in the renal tissues of human fetuses with CZS, supporting the theory that viral replication in renal tissues might be a common trait of flaviviral infections [bib_ref] The spectrum of neuropathological changes associated with congenital Zika virus infection, Chimelli [/bib_ref] [bib_ref] Human Zika and West Nile virus neurological infections: What is the difference?, Wiley [/bib_ref]. Nevertheless, the sole presence of viral RNA is not enough to demonstrate that ZIKV can replicate in the human kidney.
In this work, we searched for ZIKV RNA and antigens in the kidney and other organs of a deceased newborn with CZS. Viral isolation from renal tissues provided evidence that the kidneys are active sites of ZIKV replication in congenitally-infected fetuses. Arbovirus and herpesvirus co-infections were also determined.
# Methods
## Case history
In August 2016, a 22-year-old woman from Southern Veracruz, Mexico, developed a mild febrile illness at 14 weeks of gestation, accompanied by headaches, a skin rash, and general pruritus that lasted for 3 days. In October 2016, at 24 weeks of gestation, a prenatal evaluation with fetal ultrasonography revealed intrauterine growth restriction, and she was referred to the Instituto Nacional de Perinatología in Mexico City. At 28 weeks of gestation, further fetal ultrasonographic analyses confirmed the diagnosis and showed microcephaly, enlarged lateral ventricles, and cortical calcifications, as well as oligohydramnios. Maternal serum and amniotic fluid samples collected at 28 weeks of gestation tested negative for Zika, Dengue, and Chikungunya viruses RNA by real-time quantitative polymerase chain reaction (RT-qPCR). In late November 2016, at 30 weeks of gestation, the patient was admitted with severe preeclamptic symptoms, which led to a Cesarean section. The newborn died 4 hours after delivery due to Respiratory Distress Syndrome. The necropsy was performed with the written informed consent of the mother.
## Necropsy
The necropsy was performed 2 hours after the newborn's death. Fresh tissue samples from different organs were collected and stored at -80°C for virological studies. Samples from the cerebral cortex, thymus, lungs, kidneys, adrenal glands, spleen, liver, and small intestine were fixed in 10% formalin and embedded in paraffin for further histopathological examination and for in situ detection of viral antigens.
# Virological analysis
RT-qPCR assays were used to detect viral RNA from the cerebral cortex, thymus, lungs, kidneys, adrenal glands, spleen, liver, and small intestine. RT-qPCR amplification and melting curve profiles for arbovirus detection (Zika, Dengue, Chikungunya, West Nile virus; Sevilla-Reyes et al., in preparation) were used as indicators of related infections.
All necropsy samples were also tested by qPCR for deoxyribonucleic acid (DNA) of the [bib_ref] Zika virus pathogenesis and tissue tropism, Miner [/bib_ref]
## Next-generation sequencing
A brain sample was fragmented and passed through an insulin needle. The homogenate was clarified by centrifugation and filtered with a 0.45 µm syringe filter. The filtrate was digested with Turbo DNase (Ambion) and RNase I (Invitrogen) to remove non-protected human nucleic acids. Remaining nucleic acids were extracted using the QIAamp viral RNA minikit (Qiagen). Purified RNA was reverse-transcribed with Superscript III reverse transcriptase (Invitrogen) and second-strand DNA synthesis was performed with Klenow fragment polymerase (New England Biolabs) using barcoded primers consisting of a 20-nucleotide-specific sequence upstream of a random nonamer, as previously described [bib_ref] Rapid identification of known and new RNA viruses from animal tissues, Victoria [/bib_ref]. The resulting DNA products were PCR amplified, and libraries were constructed with the Nextera XT DNA library preparation kit (Illumina) and sequenced on a NextSeq Illumina platform (2 × 150 bp run).
## Immunofluorescence and immunoperoxidase assays
The immunofluorescence and immunoperoxidase assays for the detection of ZIKV envelope antigen were carried out using the mouse anti-flavivirus envelope protein monoclonal antibody 4G2 [bib_ref] Dengue virus-specific and flavivirus group determinants identified with monoclonal antibodies by indirect..., Henchal [/bib_ref] , obtained from mouse immune ascitic fluid and an Alexa-555 goat anti-mouse immunoglobulin G (IgG) as a secondary antibody (Jackson ImmunoResearch) or rabbit anti-mouse antibodies labelled with horseradish peroxidase (Vectastain ABC Sytem, Burlingame, CA). Complete methodologies are provided in the Methods section of the Supplementary Appendix.
## In situ apoptotic cell detection
For in situ apoptotic cell detection, we used the DeadEnd peroxidase colorimetric apoptosis detection system kit (Promega), using brain tissue sections as described in the Methods section of the Supplementary Appendix.
## Ultrastructural studies
For ultrastructural evaluation, small tissue fragments from the kidney cortex were analyzed by immunoelectronmicroscopy with the monoclonal antibody 4G2 and a rabbit anti-mouse IgG conjugated to 5 nm gold particles. Complete methodologies are provided in the Methods section of the Supplementary Appendix. A brain sample was fragmented, suspended, and homogenized in serum-free Dulbecco´s Modified Eagle Medium. Clarified supernatants were passed through 0.22 μm filters and used to infect Vero and C6/36 cell monolayers, which were then incubated at 37 o C or 28°C, respectively, in a 5% CO 2 atmosphere for 7-10 days. The supernatants of the infected Vero and C6/36 cells were recovered, and then 4 additional passages of the possible isolates were carried out in Vero cells. The identity of the ZIKV isolates was verified by RT-PCR and Sanger sequencing, as described in the Methods section of the Supplementary Appendix.
# Results
Clinically-diagnosed CZS was confirmed by the pathological examination of the deceased newborn, both macroscopically and microscopically. External body examination of the 30 week-gestation female newborn revealed microcephaly with a head circumference of 23.5 cm (below the third percentile), micrognathia and retrognathia, low-set ears, and a depressed nasal bridge, as well as arthrogryposis. A macroscopic evaluation of the encephalic region revealed microcephaly, with a brain weight of 46.8 grams (Z-score: -5.35). The malformed brain showed an smooth outer cortical surface, with a total lack of gyri (lissencephaly). The cerebral lobes and the brain stem were hypoplastic. A cross-sectional examination of the brain revealed bilateral ventricular enlargement and a slim ribbon of cortical and subcortical white radial calcifications that were more prominent towards the occipital lobes [fig_ref] Figure 1: Central nervous system findings associated with Congenital Zika Syndrome [/fig_ref].
A microscopic examination of the brain sections revealed extensive granular calcifications in the cortical and subcortical white matter, which were surrounded by activated microglia, as shown by ionized calcium binding adapter molecule 1 immunostaining [fig_ref] Figure 1: Central nervous system findings associated with Congenital Zika Syndrome [/fig_ref]. After extensive revision of the histological sections, we found occasional capillaries with minimal or mild inflammatory cells in the perivascular spaces, many of them positive to F4-80, a marker of peripheral macrophages that indicates monocyte emigration from capillary lumens to perivascular spaces [fig_ref] Figure 1: Central nervous system findings associated with Congenital Zika Syndrome [/fig_ref]. In the same sections, we performed the TUNEL technique to detect apoptotic bodies and found many apoptotic cells, particularly in the cerebral cortex [fig_ref] Figure 1: Central nervous system findings associated with Congenital Zika Syndrome [/fig_ref].
ZIKV RNA was detected in the brain (cerebral cortex), thymus, lungs, kidneys, adrenal glands, and spleen, all of which were negative for other arboviruses [fig_ref] Table 1: Results From the Qualitative Polymerase Chain Reaction Assays for Detection of Arboviruses... [/fig_ref]. The cerebral cortex and liver also tested positive for EBV infection, while the thymus, kidneys, adrenal glands, and liver samples tested positive for HHV6 [fig_ref] Table 1: Results From the Qualitative Polymerase Chain Reaction Assays for Detection of Arboviruses... [/fig_ref]. Both the liver and proximal small intestine samples were negative for ZIKV RNA.
Additionally, the complete genome of ZIKV (ZIKV/H.sapiens/Mexico/INPER38b/2016) was recovered from the brain tissue (accession number MG494697). A phylogenetic analysis of the complete genome sequence revealed that it clustered together with sequences collected from Mexico and Central America since late 2015, suggesting that they were still circulating in the region (Supplementary [fig_ref] Figure 1: Central nervous system findings associated with Congenital Zika Syndrome [/fig_ref] in mid-2016.
To analyze whether the presence of viral RNA in the tissues of the deceased newborn were accompanied by viral antigens, cross sections from organs that tested positive for ZIKV RNA [fig_ref] Table 1: Results From the Qualitative Polymerase Chain Reaction Assays for Detection of Arboviruses... [/fig_ref] were analyzed by immunoperoxidase and immunofluorescence staining for the viral envelope (E) protein using the monoclonal antibody 4G2. Viral antigens were not detectable in the spleen or adrenal glands, even though they had previously tested positive for viral RNA. In contrast, in the brain cortex there was strong immunostaining to the E protein around the calcified areas and in macrophages localized in perivascular infiltrates [fig_ref] Figure 2: Detection of envelope antigens of the Zika virus in brain sections [/fig_ref]. Additionally, mild E protein immunostaining was observed in some neurons and in ependymal cells from the lateral ventricles, which also tested positive for apoptosis with the TUNEL technique [fig_ref] Figure 2: Detection of envelope antigens of the Zika virus in brain sections [/fig_ref]. We also provide examples of negative controls in [fig_ref] Figure 2: Detection of envelope antigens of the Zika virus in brain sections [/fig_ref].
In the kidneys, strong immunostaining was detected in both the medullar and cortical tubules [fig_ref] Figure 3: Detection of Flavivirus envelope protein in different fetal tissues by immunoperoxidase assay [/fig_ref]. Immunofluorescence detection of the E protein was preferentially located in the apical cytoplasm of tubular epithelial cells [fig_ref] Figure 1: Central nervous system findings associated with Congenital Zika Syndrome [/fig_ref].
In the thymus, the E protein was only detected in the epithelial cells located inside the Hassall's corpuscles, while no specific signal was observed in the cortex or in other histological elements of the medulla of the thymus [fig_ref] Figure 3: Detection of Flavivirus envelope protein in different fetal tissues by immunoperoxidase assay [/fig_ref]. Finally, in the lungs, strong immunostaining was observed in the cytoplasm of the bronchial epithelial cells [fig_ref] Figure 3: Detection of Flavivirus envelope protein in different fetal tissues by immunoperoxidase assay [/fig_ref] , preferentially displaying an apical localization.
The high levels of ZIKV antigens and RNA detected in the brain, kidney, and lungs made us hypothesize that the virus was actively replicating in these organs. Virion-like particles were observed in sections of renal tissues by electron microscopy [fig_ref] Figure 5: Virological analysis of kidney tissues [/fig_ref] , suggesting the presence of ZIKV in the kidney of the deceased newborn. To confirm the presence of infectious viral particles in the kidney, as well as in the brain and lungs, virus isolation procedures were performed using fresh-frozen tissues. After 4 passages in Vero cells, the infected monolayers displayed a mild cytopathic effect, characterized by cell rounding and detachment. The RT-qPCR results suggested that the isolation of ZIKV was positive from the brain and kidney, but not from the pulmonary tissues [fig_ref] Table 2: Results From the Real-Time Qualitative Polymerase Chain Reaction Assays for Identification of... [/fig_ref]. To confirm the identity of the isolates, partial sequences of the E gene were recovered from the supernatants of the infected cultures showing a single synonymous substitution, in comparison to the viral sequences obtained by next generation sequencing directly from the brain cortex (Supplementary [fig_ref] Figure 2: Detection of envelope antigens of the Zika virus in brain sections [/fig_ref].
[formula] DENV 1-4 - - - - - - - - CHIKV - - - - - - - - WNV - - - - - - - - HSV-1 - - - - - - - - HSV-2 - - - - - - - - VZV - - - - - - - - EBV POSITIVE b - - - - - POSITIVE - CMV - - - - - - - - HHV6 - - POSITIVE POSITIVE POSITIVE - POSITIVE - HHV7 - - - - - - - - HHV8 - - - - - - - - [/formula]
# Discussion
Several studies have demonstrated the link between intrauterine ZIKV infections and the development of a distinctive pattern of birth defects, grouped as CZS, that includes microcephaly and multiple fetal malformations such as arthogryposis and intrauterine growth restriction [bib_ref] Zika virus associated with microcephaly, Mlakar [/bib_ref] [bib_ref] The spectrum of neuropathological changes associated with congenital Zika virus infection, Chimelli [/bib_ref] , characteristic of the Fetal Akinesia Deformation Sequence. Moreover, common histopathological findings in the CNS tissues of autopsied fetuses and neonates with CZS include microcalcifications in the cortical and subcortical white matter, reactive gliosis, glial degeneration [bib_ref] Zika virus associated with microcephaly, Mlakar [/bib_ref] [bib_ref] Zika virus damages the human placental barrier and presents marked fetal neurotropism, Ld [/bib_ref] and, in some cases, perivascular lymphocytic cuffing with macrophage and mononuclear cell infiltrates [bib_ref] Zika virus damages the human placental barrier and presents marked fetal neurotropism, Ld [/bib_ref] [bib_ref] Zika virus: pathology from the pandemic, Ritter [/bib_ref]. Nevertheless, the dissemination of ZIKV to other human fetal tissues has only been partially characterized. The disseminative potential of ZIKV has already been demonstrated in human patients under corticosteroid therapy or with underlying medical conditions such as autoimmune diseases. In such cases, ZIKV RNA has been detected postmortem in the blood, brain, spleen, liver, kidneys, lungs, and heart [bib_ref] Zika virus associated deaths in Colombia, Sarmiento-Ospina [/bib_ref] [bib_ref] Fatal sickle cell disease and Zika virus infection in girl from Colombia, Arzuza-Ortega [/bib_ref]. Additionally, infectious ZIKV was isolated from a pool of macerated organs (heart, lungs, and kidneys), suggesting that either of them might represent an important niche for ZIKV replication in immunosuppressed adult humans [bib_ref] Zika virus epidemic in Brazil. I. Fatal disease in adults: clinical and..., Azevedo [/bib_ref]. In several CZS cases, postmortem detection of ZIKV RNA in different tissues has been assessed by RT-PCR or in situ hybridization methods; still, some of them have failed to detect viral RNA outside the fetal CNS and the placenta [bib_ref] Zika virus associated with microcephaly, Mlakar [/bib_ref] [bib_ref] Congenital Zika virus infection: beyond neonatal microcephaly, Melo [/bib_ref] [bib_ref] Zika virus infection and stillbirths: a case of hydrops fetalis, hydranencephaly and..., Sarno [/bib_ref]. In an analysis that was among the first to detail a CZS case, the presence of viral RNA was reported in the spleen, lungs, liver, and muscle, suggesting the possible dissemination of ZIKV to multiple fetal tissues [bib_ref] Zika virus infection with prolonged maternal viremia and fetal brain abnormalities, Driggers [/bib_ref]. Later, during the ZIKV outbreak in the Americas, the presence of ZIKV RNA was reported in the spleens, kidneys, and livers of 2 fatal cases of CZS in Colombia [bib_ref] Severe neurologic disorders in 2 fetuses with Zika virus infection, Colombia, Acosta-Reyes [/bib_ref]. Further analyses of the tissues of 7 deceased neonates in Brazil demonstrated that ZIKV RNA was present in the brains, lungs, hearts, livers, spleens, and kidneys in 3 of the cases studied [bib_ref] Postmortem findings for 7 neonates with congenital Zika virus infection, Sousa [/bib_ref] , supporting the theory that disseminated infections might be occurring in some cases of CZS. Moreover, in situ hybridization analyses of fetal human tissues have also demonstrated that the liver, kidneys, and spleen might be active sites of ZIKV replication in fetuses [bib_ref] The spectrum of neuropathological changes associated with congenital Zika virus infection, Chimelli [/bib_ref]. Consistent with previous findings, this report provides evidence of the presence of ZIKV RNA in multiple fetal tissues, suggesting that disseminated infections might be affecting different organs in each case; however, due to technical issues, retrospective study comparisons could be difficult.
The dissemination of ZIKV into multiple fetal tissues has also been evaluated by the immunohistochemical detection of viral antigens: mainly by the detection of the viral E protein [bib_ref] Replication of early and recent Zika virus isolates throughout mouse brain development, Rosenfeld [/bib_ref] with the pan-flaviviral antibody 4G2. To date, the presence of the E protein of ZIKV has been reported in brain tissues, as well as in placental tissues [bib_ref] Zika virus damages the human placental barrier and presents marked fetal neurotropism, Ld [/bib_ref] , chorionic membranes, and umbilical cords [bib_ref] Human Zika and West Nile virus neurological infections: What is the difference?, Wiley [/bib_ref]. The presence of the E protein has been reported in neurons, in areas of microcalcification in the brain parenchyma, in degenerating glial cells, and in microglia and endothelial cells [bib_ref] Zika virus: pathology from the pandemic, Ritter [/bib_ref]. In this report, we were also able to detect ZIKV antigens in cells from the ventricular ependymal epithelium. The implications of ZIKV infection of ependymal cells in humans remain unclear; nevertheless, it seems that the viral infection could induce apoptosis, considering that we found many TUNELpositive ependymal cells [bib_ref] Zika virus disrupts neural progenitor development and leads to microcephaly in mice, Li [/bib_ref] [bib_ref] The Brazilian Zika virus strain causes birth defects in experimental models, Cugola [/bib_ref].
Interestingly, the presence of ZIKV antigens in systemic organs during congenital infections has only been reported in animal models [bib_ref] Zika virus infection of rhesus macaques leads to viral persistence in multiple..., Hirsch [/bib_ref] [bib_ref] Infection via mosquito bite alters Zika virus tissue tropism and replication kinetics..., Dudley [/bib_ref] despite several attempts to detect ZIKV E proteins in human fetal tissues using 4G2 [bib_ref] Zika virus associated with microcephaly, Mlakar [/bib_ref] [bib_ref] Zika virus damages the human placental barrier and presents marked fetal neurotropism, Ld [/bib_ref]. In this report, we could detect ZIKV antigens in multiple organs using the monoclonal antibody 4G2. In the brain, we detected the E protein in cells from the ventricular ependymal epithelium and in macrophages located in perivascular spaces. The strong immunoreactivity displayed by peripheral macrophages towards F4/80 and the ZIKV E protein supports the hypothesis that hematogenous dissemination of the virus might play a role in the development of intrauterine ZIKV infections with multi-organ involvement.
The presence of the viral-like particles in the renal tissues and the isolation of infectious ZIKV from renal tissues demonstrate that the kidneys are an active site of ZIKV replication in the fetus. Moreover, the renal tubular epithelium appears to be at risk for ZIKV infection.
ZIKV replication in the kidneys might explain the continuous viral shedding observed in the urine of some congenitally-infected newborns [bib_ref] Prolonged shedding of Zika virus associated with congenital infection, Oliveira [/bib_ref]. Seemingly, viral excretion in urine could be a common trait of flaviviral infections [bib_ref] Yellow fever virus RNA in urine and semen of convalescent patient, Barbosa [/bib_ref] , although viral isolation from the renal tissues of patients with flaviviral infections has not previously reported [bib_ref] Persistent infection with West Nile virus years after initial infection, Murray [/bib_ref] [bib_ref] West Nile Virus in Transplant Recipients Investigation Team. Transmission of West Nile..., Iwamoto [/bib_ref]. Yet, several studies must be performed to rule out the possibility that ZIKV replication in fetal kidneys might entail a higher risk of renal damage.
We also detected ZIKV antigens in pulmonary tissues and in the Hassall's corpuscles in the thymus, which is suggestive of ZIKV replication in these organs. Pulmonary hypoplasia, interstitial lymphocytic pneumonitis, and expansion of the alveolar septa have all been reported as part of the clinical manifestation of CZS in fetuses [bib_ref] Postmortem findings for 7 neonates with congenital Zika virus infection, Sousa [/bib_ref] , and have been hypothesized to be a consequence of viral replication in pulmonary tissues. Even though were unable to isolate infectious ZIKV from the lungs of the infected fetus, we cannot rule out the fact that the lungs might support viral replication. Further, the clinical implications of ZIKV E protein in the thymus remain unclear and, even though the presence of viral antigens in the epithelial cells of the thymus is not conclusive enough to ensure that they can support viral replication, this finding increases the importance of studying the role of the thymus during intrauterine ZIKV infections, especially because of the important role that it plays during the establishment of self-tolerance in the fetus. The presence of EBV and HHV6 in several tissues of the deceased newborn raises even more questions about the factors that might have an influence over congenital ZIKV infections. Even though both viruses have been previously associated with congenital infections [bib_ref] In utero Epstein-Barr virus (infectious mononucleosis) infection, Goldberg [/bib_ref] [bib_ref] Congenital infections with human herpesvirus 6, Adams [/bib_ref] , further studies are needed to understand whether they have an influence in the extra-CNS dissemination of ZIKV, as well as over the severity of the disease.
Infections with HSV-2 have been shown to enhance ZIKV infection in placental tissues, demonstrating the possible role of members of the Herpesviridae family in the severe cases of CZS [bib_ref] HSV-2 enhances ZIKV infection of the placenta and induces apoptosis in first-trimester..., Aldo [/bib_ref]. Many populations have high prevalences of pathogens like HHV6 [bib_ref] Prevalence of human herpesvirus 6 antibody in the population of Belém, De Freitas [/bib_ref] [bib_ref] Seroprevalence of Epstein-Barr virus infection in U.S. children ages 6-19, Dowd [/bib_ref] ; thus, the coinfection with ZIKV or with other arboviruses is likely. Still, the influence of latent infections with HHV over the course and severity of congenital ZIKV infections remains to be explored.
It is important to highlight that, as in previous reports [bib_ref] Pathology of congenital Zika syndrome in Brazil: a case series, Martines [/bib_ref] [bib_ref] Human Zika and West Nile virus neurological infections: What is the difference?, Wiley [/bib_ref] [bib_ref] Zika virus: pathology from the pandemic, Ritter [/bib_ref] [bib_ref] Postmortem findings for 7 neonates with congenital Zika virus infection, Sousa [/bib_ref] , no dramatic pathological changes were found in most of the tissues that were analyzed, confirming that ZIKV causes the most devastating damage to the fetal brain. However, these data may forecast other, not-so-obvious pathologies like renal damage, autoimmune diseases, or other postnatal complications that may arise in the future for congenitally-infected newborns.
In conclusion, we demonstrated the ability of ZIKV to replicate in the human kidney and disseminate to other organs in congenitally-infected fetuses. Moreover, we demonstrated that other congenital viral infections might co-exist with ZIKV, increasing the need to study them to analyze their role in the pathogenesis of congenital ZIKV infection.
## Supplementary data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
[fig] Figure 1: Central nervous system findings associated with Congenital Zika Syndrome. Cross sections of the brain display symmetric lateral ventricle enlargement and diffuse cortical and subcortical calcifications, as well as cortical mantle thinning, which was more accentuated towards the (A) temporal lobe. (B) The cerebral cortex is shown with extensive calcifications. (C) Numerous activated microglial cells were found near to the calcifications. (D) Occasional capillaries showed some inflammatory cells in the perivascular area. (E) Some of these inflammatory perivascular cells were macrophages that showed F4/80 positive immunostaining. (F) The TUNEL technique showed numerous apoptotic cells in cortical and subcortical areas. [/fig]
[fig] Figure 2: Detection of envelope antigens of the Zika virus in brain sections. Representative micrographs of Zika virus detection by immunoperoxidase with the antibody 4G2 in tissue paraffin-embedded fetal tissues showed (A) strong viral envelope (E) protein immunostaining around the calcified areas. (B) Macrophages around a brain capillary showed strong E protein immunostaining. (C) Occasional neurons showed mild E protein immunostaining (arrow). (D) Ependymal cells showed slight E protein immunostaining; additionally, the TUNEL technique showed numerous apoptotic bodies in the ependymal epithelium (inset). (E and F) Negative controls of the assays were performed in tissues from a non-infected newborn. [/fig]
[fig] Figure 3: Detection of Flavivirus envelope protein in different fetal tissues by immunoperoxidase assay. Immunoperoxidase assay with the antibody 4G2 in tissue paraffin-embedded fetal tissues showed that (A) the epithelium from the cortical proximal and distal kidney tubules had strong envelope protein immunoreactivity. (B) Epithelial cells from the Hassall corpuscles in the thymus medulla showed slight E protein immunostaining (arrows), as did (C) the bronchial epithelial lumen. Negative controls of the assays were performed in tissues from a non-infected newborn in the (D) kidney, (E) thymus, and (F) lung. [/fig]
[fig] Figure 4: Detection of the Flavivirus envelope protein in different fetal tissues by immunofluorescence assay. Immunofluorescence assay with the antibody 4G2 in tissue paraffin-embedded fetal tissues showed (A) immunostaining in the epithelial cells from cortical convoluted tubules in the kidney and Hassall´s corpuscles in (B) the thymus and (C) the bronchial epithelial cells of the lung. Controls of the assays were performed in tissues from a non-infected newborn in the (D) kidney, (E) thymus, and (F) lung. [/fig]
[fig] Figure 5: Virological analysis of kidney tissues. Representative electron microscopy micrographs of infected epithelial cells from the kidney. Panel A shows numerous virions in the cytoplasm of an epithelial cell from the proximal convoluted tubule (arrows); Panel B shows immunogold particles (black dots) on spherical electron dense structures, corresponding to virions in the epithelial cell from a proximal convoluted cell. [/fig]
[fig] Notes: Author contributions. M. Y. V.-V., E. E. S.-R., R. L., M. Y.-M., and R. H.-P. contributed equally as primary investigators in this article. The rest of the authors contributed substantially to the completion of this study and preparation of this manuscript. [/fig]
[table] Table 1: Results From the Qualitative Polymerase Chain Reaction Assays for Detection of Arboviruses and Herpesviruses on Multiple Tissue Samples [/table]
[table] Table 2: Results From the Real-Time Qualitative Polymerase Chain Reaction Assays for Identification of the Isolates of Zika Virus Obtained From Brain and Kidney Samples After 4 Passages in Vero Cells aTable displays Cq values b For the positive control, RNA was extracted from Vero cells infected with the Zika virus reference strain PRVABC59. [/table]
|
SNP@Domain: a web resource of single nucleotide polymorphisms (SNPs) within protein domain structures and sequences
The single nucleotide polymorphisms (SNPs) in conserved protein regions have been thought to be strong candidates that alter protein functions. Thus, we have developed SNP@Domain, a web resource, to identify SNPs within human protein domains. We annotated SNPs from dbSNP with protein structure-based as well as sequence-based domains: (i) structure-based using SCOP and (ii) sequence-based using Pfam to avoid conflicts from two domain assignment methodologies. Users can investigate SNPs within protein domains with 2D and 3D maps. We expect this visual annotation of SNPs within protein domains will help scientists select and interpret SNPs associated with diseases. A web interface for the SNP@Domain is freely available at http://snpnavigator.net/ and from http://bioportal.net/.
# Introduction
To facilitate the identification of disease-associated single nucleotide polymorphisms (SNPs) from a large number of SNPs, it is important to select functionally relevant SNPs [bib_ref] Target SNP selection in complex disease association studies, Wjst [/bib_ref]. There are many SNP annotation servers and databases, such as FESD (http://combio.kribb.re.kr/FESD/), PicSNP (http://plaza.umin.ac.jp/~hchang/picsnp/), SNPper (http:// snpper.chip.org/) and SNPs3D (http://www.snps3d.org). These are useful for selecting SNPs without a priori biological knowledge [bib_ref] PicSNP: a browsable catalog of nonsynonymous single nucleotide polymorphisms in the human..., Chang [/bib_ref] [bib_ref] SNPper: retrieval and analysis of human SNPs, Riva [/bib_ref] [bib_ref] SNPs3D: candidate gene and SNP selection for association studies, Yue [/bib_ref] [bib_ref] SNPeffect: a database mapping molecular phenotypic effects of human non-synonymous coding SNPs, Reumers [/bib_ref] [bib_ref] MutDB services: interactive structural analysis of mutation data, Dantzer [/bib_ref] [bib_ref] SNPNB: analyzing neighboring-nucleotide biases on single nucleotide polymorphisms (SNPs), Zhang [/bib_ref] [bib_ref] topoSNP:a topographic database of non-synonymous single nucleotide polymorphisms with and without known..., Stitziel [/bib_ref] [bib_ref] nsSNPAnalyzer: identifying disease-associated nonsynonymous single nucleotide polymorphisms, Bao [/bib_ref] [bib_ref] SIFT: predicting amino acid changes that affect protein function, Ng [/bib_ref] [bib_ref] Human non-synonymous SNPs: server and survey, Ramensky [/bib_ref] [bib_ref] Selecton: a server for detecting evolutionary forces at a single amino-acid site, Doron-Faigenboim [/bib_ref]. They help biologists focus on specific genomic/proteomic regions or gene sets providing functional annotations and visualization.
The SNPs in conserved protein regions have been thought to be strong candidates that can alter protein functions [bib_ref] SIFT: predicting amino acid changes that affect protein function, Ng [/bib_ref].
However, up to now, there have been no web servers that provide extensive protein domain annotation of SNPs. Currently, Ensemblprovides domain annotation of SNPs assigned by Pfam [bib_ref] The Pfam protein families database, Bateman [/bib_ref] , PROSCAN [bib_ref] The PROSITE database, its status in 1995, Bairoch [/bib_ref] and PFscan [bib_ref] A flexible motif search technique based on generalized profiles, Bucher [/bib_ref]. However, these protein domains are all sequence-based functional domains that are based on protein sequence profiles. Structure-based approaches define domains according to the compactness and conservation of protein structural regions [bib_ref] Toward consistent assignment of structural domains in proteins, Veretnik [/bib_ref] while sequence-based domain databases constructed based on sequence similarity of proteins implied evolutionary relationships [bib_ref] A comparison of sequence and structure Protein domain families as a basis..., Elofsson [/bib_ref] [bib_ref] comparative mapping of sequence-based and structure-based protein domains, Zhang [/bib_ref]. If a structure-based domain family and sequence-based domain family are defined at the same location over the same set of protein chains, they should map exactly to each other in a protein. However, it has been known that they have conflicts [bib_ref] A comparison of sequence and structure Protein domain families as a basis..., Elofsson [/bib_ref] [bib_ref] comparative mapping of sequence-based and structure-based protein domains, Zhang [/bib_ref]. SCOP (21) is a representative structure-based classification database for Protein Data Bank (PDB) [bib_ref] The Protein Data Bank, Berman [/bib_ref]. They list all the proteins with known structures and organize them hierarchically. Pfam (15) is a representative sequence-based domain database that contains hidden Markov model-based profiles of many common protein domains constructed using multiple sequence alignments. Previously, Elofsson's group [bib_ref] A comparison of sequence and structure Protein domain families as a basis..., Elofsson [/bib_ref] reported that 70% of SCOP families exist in Pfam, while 57% of Pfam families exist in SCOP. Recent research conducted by Zhang's group [bib_ref] comparative mapping of sequence-based and structure-based protein domains, Zhang [/bib_ref] shows that 80% of SCOP domains overlap with at least one Pfam family. These SCOP domain families correspond to 99.7% of the Pfam families. Although the overlaps increased (SCOP, from 70 to 80%; and Pfam, from 57 to 99.7%), partial mapping between SCOP and Pfam domain could still occur. Zhang's group reported that only 62% of the cases of one-to-one mapping of a SCOP domain to a Pfam domain agreed by 90% or more of their coverage [bib_ref] comparative mapping of sequence-based and structure-based protein domains, Zhang [/bib_ref].
Since a non-synonymous SNP can correspond to an amino acid change, it is necessary to have a good protein domain annotation and visualization server. Here, we introduce the SNP@Domain server providing information for SNPs found within protein domains. SNP@Domain contains all the human SNPs from dbSNP (23) that match SCOP and Pfam domain sequences that are assigned to Ensembl database proteins. A 2D map of Pfam and SCOP domains with SNPs is provided. Additionally, a 3D map of SNPs within domains is provided if protein structures are available.
# Methods and usage
## Identifying snps within protein domains
We annotated protein domains to human proteins in the Ensembl database (ftp://ftp.ensembl.org/pub/human-25.34e/ data/mysql/homo_sapiens_snp_25_34e) and mapped whole SNPs from dbSNP (http://www.ncbi.nlm.nih.gov/SNP/) [bib_ref] ) dbSNP: the NCBI database of genetic variation, Sherry [/bib_ref]. Since the Ensembl database provides Pfam domain annotation information, we performed a structure-based domain assignment using the PDB-ISL method [bib_ref] Fast assignment of protein structures to sequences using the intermediate sequence library..., Teichmann [/bib_ref] [bib_ref] Gapped BLAST and PSI-BLAST: a new generation of protein database search programs, Altschul [/bib_ref]
## Two-and three-dimensional maps of snps within protein domains
SNP@Domain is a web-based tool that was constructed using Java Server Pages and Perl Common Gateway Interface scripts. SNP@Domain provides three query interfaces as shown in [fig_ref] Figure 1: Search interface of SNP@Domain [/fig_ref] : (i) SNP identifier (rs number), (ii) gene identifier (Ensemble protein ID) and (iii) domain identifier (SCOP concise classification strings ID or Pfam ID). SNP@Domain also supports keyword searches with gene and/or domain names. When the user accesses it with a queried SNP or a gene name, the 2D image map of SNPs within protein domains is displayed as shown in [fig_ref] Figure 2: An example of detail information and image maps of an SNP within... [/fig_ref]. This 2D image map utilizes the Generic Genome Browser (Gbrowse; http://www.gmod.org), originally developed by Stein's group [bib_ref] The Generic Genome Browser: a building block for a Model Organism System..., Stein [/bib_ref]. The 2D map has four kinds of horizontal tracks corresponding to SCOP domains, Pfam domains, synonymous and non-synonymous SNPs within a protein. For convenience, synonymous SNPs and non-synonymous SNPs are displayed separately. The queried SNPs are highlighted in the map so they can be easily distinguished. Each SNP in the 2D map links to detailed information of the SNP such as chromosomal position, class, validation, alleles, effects predicted by SIFT server and relationships with disease(s), if available. If the structure of the protein is available in the PDB, SNP@Domain provides a 3D view of the protein highlighting the amino acids affected by SNPs. To avoid sequence conflicts between an Ensembl protein sequence and a PDB sequence, SNP@Domain carries out a BLAST with a query of Ensemble protein sequence against a protein sequence from PDB and parsed hits. We use MDL Chime plugin (http://www.mdli.com/ downloads/) for visualizing 3D structures of proteins which was developed based on RasMol (http://www.umass.edu/ microbio/rasmol/) [bib_ref] RASMOL: biomolecular graphics for all, Sayle [/bib_ref].
[fig] *: To whom correspondence should be addressed. Tel: +82 42 879 8549; Fax: +82 42 879 8519; Email: [email protected] Ó The Author 2006. Published by Oxford University Press. All rights reserved. The online version of this article has been published under an open access model. Users are entitled to use, reproduce, disseminate, or display the open access version of this article for non-commercial purposes provided that: the original authorship is properly and fully attributed; the Journal and Oxford University Press are attributed as the original place of publication with the correct citation details given;if an article is subsequently reproduced or disseminated not in its entirety but only in part or as a derivative work this must be clearly indicated. For commercial re-use, please contact [email protected] [/fig]
[fig] Figure 1: Search interface of SNP@Domain. The user is able to search SNP domain annotations with three inputs including (i) SNP identifier (rs number), (ii) Gene identifier (Ensembl protein ID) or name/symbol, and (iii) Domain identifier (SCOP concise classification strings ID or Pfam ID) or name. [/fig]
[fig] Figure 2: An example of detail information and image maps of an SNP within protein domains. Following the user's query to the SNP (rs number ¼ 'rs3088308'), the SNP's detail information including chromosomal location, class, validation and alleles were displayed. And a summary of domain mapping results and a corresponding 2D image map were shown up. Four tracks of the 2D image map were displayed including (i) Pfam domain, (ii) SCOP domain, (iii) synonymous SNPs and (iv) non-synonymous SNPs within the protein. The 3D image map of the SNP is also available. [/fig]
|
Foot Position Measurement during Assistive Motion for Sit-to-Stand Using a Single Inertial Sensor and Shoe-Type Force Sensors
Citation: Kitagawa, K.; Gorordo Fernandez, I.; Nagasaki, T.; Nakano, S.; Hida, M.; Okamatsu, S.; Wada, C. Foot Position Measurement during Assistive Motion for Sit-to-Stand Using a Single Inertial Sensor and Shoe-Type Force Sensors. Int. J.
# Introduction
## Lower back pain due to assistive motion for sit-to-stand
Assistive motion for sit-to-stand causes lower back pain (LBP) among caregivers due to heavy manual lifting. reported that 69% of nurses experienced LBP caused by patient handling tasks, including assistive motion for sit-to-stand. Furthermore, our investigation on occupational injury revealed that caregivers experience LBP caused by manual lifting for assistive motion. Although stationary lifting devices can be used to prevent LBP, several devices cause upper limb discomfort during the sitting up process. Moreover, these stationary devices cannot be used in all workspaces
## Related studies for foot position measurement
Generally, foot position is measured using vision-based systems, such as optical motion capture systems and video cameras, which cannot be used in actual workspaces due to the large spaces occupied by these systems and the limited measurement areas that are available. Therefore, the measurement of foot position for the prevention of LBP during assistive motion for sit-to-stand must be achieved using a small number of wearable sensors. During gait analysis, the foot trajectory is often measured using an inertial sensor on the shoe. However, integral errors have been a concern when directly calculating longterm foot trajectories using inertial data obtained from the inertial sensor. To reduce integral errors of the inertial sensor, more than 10 algorithms have been developed. However, these algorithms cannot be applied to assistive motion for sit-to-stand given that they are based on specific and frequent gait events. Although an ultrasonic sensor can also be used to measure foot position, these sensors cannot measure foot position when obstacles are present between both feet.
## Potential of inertial sensor and shoe-type force sensors
To address the aforementioned concerns, we have been developing a method for measuring foot position using a single inertial sensor on trunk and shoe-type force sensors. Previous studies have shown that the use of an inertial sensor on the trunk and shoe-type force sensors can be considered an effective method for measuring foot position given the relationship between trunk movement and foot position during manual handling. found that a wide-footed stance is related to trunk rotation during manual lifting tasks. reported that a wide-footed stance increased trunk bending. Our previous study indicated that foot position with a long anteroposterior distance reduced trunk bending. From these relationships, we hypothesized that foot position can be estimated using inertial data obtained from the inertial sensor on the trunk.
Furthermore, Jeong et al. suggested that the foot position during manual lifting affects the ground reaction forces on the toe and heel and the trajectory of the center of pressure (COP). revealed that the COP shifted to the left side when the right foot was placed forward. Moreover, previous studies have shown that shoe-type force sensors could be used to recognize occupational postures, including different foot positions, and to monitor trunk movement, lumbar load, and arm movement during manual handling. From these findings, we hypothesized that foot position can be estimated on the basis of the distribution of the ground reaction force on the foot that is obtained from shoe-type force sensors. Therefore, a single inertial sensor on the trunk and shoe-type force sensors were selected as suitable wearable sensors to estimate the foot position.
## Potential of machine learning-based regression
A single inertial sensor on the trunk and shoe-type force sensors could not measure the quantitative foot position. Thus, the current study utilized machine learning-based regression algorithms for the indirect wearable measurement of quantitative foot position. Machine learning-based regression algorithms have previously been applied for indirect measurements using a small number of wearable sensors. Furthermore, machine learning can be applied to estimate nonlinear relationships. Moreover, machine learning can be used in applications using wearable sensors that require indirect estimations for joint movement, muscle activity, and position. Recent studies have shown that the combination of wearable sensors and machine learning could be used to estimate kinematic values during lifting and squatting. Antwi-Afari et al. reported that several lifting postures could be identified using machine learning and that foot pressure distributions could be obtained from insole force sensors. Matijevich et al. found that lumbar load during manual lifting could be predicted by a machine learning-based regression model with an inertial sensor and shoe-type force sensors. Choffin et al. estimated foot angle while squatting using machine learning and insole force sensors. These previous studies focused on tasks that are similar to our current research, and thereby provided insight into the potential usefulness of the combination of wearable sensors and machine learning for estimating foot position during manual lifting. From these findings, we hypothesized that machine learning is the most suitable method for the indirect estimation of foot position.
## Objective of this paper
The present study therefore aimed to propose and evaluate a new method for measuring foot position during manual lifting using a machine learning technique and only a single inertial sensor on the trunk and shoe-type force sensors.
## Contribution of this paper
Previous wearable systems to prevent LBP could monitor and provide feedback for injury risks such as lumbar loads and trunk angles during manual lifting; moreover, other previous studies have also shown that providing feedback regarding trunk angle and lower limb posture can also help to reduce the risk of LBP. However, even though feedback about the trunk angle was completed in real time, lower limb posture feedback was performed by a personal trainer giving verbal instructions such as "use legs instead of back". Thus, a system that could monitor the posture of the lower limbs without the need for a personal trainer was considered here. Particularly, we focused on foot position as an effective parameter for posture adjustment that can be easily adjusted. However, despite the benefits of being able to monitor the position of the feet during manual lifting, no such monitoring system has been developed in the past.
Therefore, this study contributes to assessing foot position measurement using wearable sensors for the prevention of lower back pain during assistive motion for sit-to-stand. The highlight of this study is that a combination of few wearable sensors and machine learning can estimate foot position during manual lifting. The proposed foot position estimation method can be easily realized because it can be implemented through the use of only a single inertial sensor and shoe-type force sensors. Therefore, the proposed method could be applied for suitable assistive motion training to prevent LBP in nursing schools.
# Materials and methods
## Proposed method for foot position measurement
[formula] 2.1.1. Architecture [/formula]
The architecture of the proposed method is shown in. Wearable sensors are mounted on the trunk and both feet. These wearable sensors were selected based on aforementioned potentials and hypotheses in previous section. The inertial sensor on the trunk measures 3-axes accelerations and 3-axes angular velocities during assistive motion for sit-to-stand, whereas the shoe-type force sensors measure the front and rear force on both feet during manual handling. These sensor data are used for machine learningbased regression algorithms to estimate quantitative foot position. Anteroposterior and mediolateral distances between both feet are outputted quantitatively as the foot position. Components of this architecture are described in. Previous wearable systems to prevent LBP could monitor and provide feedbac injury risks such as lumbar loads and trunk angles during manual lifting; mo ver, other previous studies have also shown that providing feedback regarding trun gle and lower limb posture can also help to reduce the risk of LBP. However, though feedback about the trunk angle was completed in real time, lower limb po feedback was performed by a personal trainer giving verbal instructions such as "use instead of back". Thus, a system that could monitor the posture of the lower limbs wit the need for a personal trainer was considered here. Particularly, we focused on foo sition as an effective parameter for posture adjustment that can be easily adjusted However, despite the benefits of being able to monitor the position of the feet during m ual lifting, no such monitoring system has been developed in the past.
Therefore, this study contributes to assessing foot position measurement using w able sensors for the prevention of lower back pain during assistive motion for sit-to-st The highlight of this study is that a combination of few wearable sensors and mac learning can estimate foot position during manual lifting. The proposed foot positio timation method can be easily realized because it can be implemented through the u only a single inertial sensor and shoe-type force sensors. Therefore, the proposed me could be applied for suitable assistive motion training to prevent LBP in nursing sch
# Materials and methods
## Proposed method for foot position measurement
[formula] 2.1.1. Architecture [/formula]
The architecture of the proposed method is shown in. Wearable sensor mounted on the trunk and both feet. These wearable sensors were selected based on a mentioned potentials and hypotheses in previous section. The inertial sensor on the t measures 3-axes accelerations and 3-axes angular velocities during assistive motio sit-to-stand, whereas the shoe-type force sensors measure the front and rear force on feet during manual handling. These sensor data are used for machine learning-base gression algorithms to estimate quantitative foot position. Anteroposterior and med teral distances between both feet are outputted quantitatively as the foot position. C ponents of this architecture are described in.
## Wearable sensors
## Wearable sensors
The inertial sensor (LP-WS1104, Logical Product Co., Fukuoka, Japan) attached on the trunk measures the three-axis acceleration and angular velocity of the trunk and is used for the machine learning algorithm. The specifications of the inertial sensor were as follows: full range of sensor output in acceleration, ±5 G; acceleration sensitivity, 191.4 mV/g; full range of sensor output in gyro, 1500 degrees per second (dps); and gyro sensitivity, 0.8 mv/dps. The shoe-type force sensors consisted of eight force sensors on each insole. Four sensors were located on the forefoot, and the other four sensors were located from the midfoot to the hindfoot, as shown in. The sum of the output values for each of the four sensors were calculated as the front and rear force on each foot (four total values). FlexiForce sensors (Tekscan Inc., South Boston, MA, USA) were selected for the proposed method given their thinness and flexibility. The specifications of the force sensor were as follows: full range of sensor output, 445 N for one sensor; linearity, <±3% of full range of sensor output; and hysteresis, <4.5% of full range of sensor output. The noise of the force data was reduced via an amplifier circuit (FlexiForce Adapter 1120, Phidgets Inc. Calgary, AB, Canada).
The inertial sensor and data logger for the force sensors were synchronized using 2.4-GHz wireless communication in an IEEE802.15.4-based protocol and antenna (Logical Product Co., Fukuoka, Japan). The shoe-type force sensors measured front and rear vertical forces on each foot during manual handling. The sampling rates for these wearable sensors were set at 100 Hz. The inertial data were saved in a flash memory drive installed in the inertial sensor. Force data were saved in the flash memory drive installed in the data logger (Logical Product Co., Japan). These saved data were exported to a personal laptop personal computer via a USB cable.
## Machine learning-based regression algorithm
A machine learning regression model estimated the quantitative foot position from the wearable sensor data. For this purpose, several parameters were calculated from the time series data of each sensor signal to be used as input for the machine learning model. The extracted parameters included the mean, standard deviation, skewness, kurtosis, maximum, minimum, and root mean square of the data in each measurement. These parameters were determined based on previous studies using machine learning algorithm and wearable sensors. These seven parameters were calculated for 3-axes accelerations, 3-axes gyro, and four force values (front and rear of each foot); thus, a total of 70 features were calculated for each trial. The proposed method estimates foot position based on an entire motion; thus, these parameters were extracted from an entire motion.
Optimal machine learning algorithms depend on target movement and sensors. Thus, this study compared the outcomes of artificial neural network (ANN), Gaussian process, k-nearest neighbor (kNN), linear regression, and support vector regression (SVR) for the proposed method. These five machine learning algorithms were selected given their frequent use for wearable systems in previous studies. In addition, ANN, Gaussian process, kNN, and SVR were expected to build regression models for nonlinear and complex data distribution. These findings are more informative than the fact that these algorithms were simply tested. Furthermore, linear regression was expected to build a simple model for implementation advantages. The details of these machine learning algorithms are descripted in a later section.
## Foot position
Our previous study found that both the anteroposterior and mediolateral distances are important for foot position adjustment aimed at reducing LBP during assistive motion for sit-to-stand. Therefore, the proposed method outputs the anteroposterior and mediolateral distances between both left and right feet. Actual foot positions for the training and validation of the machine learning algorithm were measured using an optical motion capture system (OptiTrack, Corvallis, OR, USA) and optical makers on the both left and right heels. This study defined heel position as actual foot position without considering foot rotation and the center of the foot because our previous study found that changing heel position could reduce lumbar load during assistive motion for sit-to-stand. The data from the optical motion capture system were sampled at 100 Hz to match the sampling rate of the wearable sensors. In addition, the optical motion capture system measured foot positions by means of a marker on each heel in the global frame coordinate.
## Expected intervention
The proposed method will be applied for assistive motion training in each nursing school. The collection of the data to be used for machine learning can be collected from nursing students because nursing students repeat assistive motion during their lessons and during self-training. Furthermore, previous studies have mentioned that improving assistive motion training for nursing students might be best approach for the long-term prevention for LBP because it is difficult to improve the assistive motion of experienced nurses. The proposed method will be used by nursing students during assistive motion training. While undergoing assistive motion training, immediate feedback regarding foot position will be given immediately after one motion. If the foot position is unsuitable, the nursing students will be able to improve their foot position starting from the next motion. Nursing students will be able to acquire a suitable foot position via training with these processes.
## Experiment for evaluation of the proposed method
This study evaluated whether the proposed method could accurately measure quantitative foot position during actual manual handling. Moreover, we compared five common machine learning algorithms (ANN, Gaussian process, kNN, linear regression, and SVR) to determine the optimal algorithm for the proposed method. Details regarding this experiment are described below.
## Participants
Ten young male students (age 23.2 ± 1.03 years, height 171 ± 6.35 cm, weight 59.8 ± 5.14 kg, mean ± standard deviation) at the Kyushu Institute of Technology were recruited. The experiment was explained to all of the participants before participation, after which all participants signed an informed consent form before experimentation began. This study considered the idea that the proposed system could be implemented in nursing schools. In this intervention, training data would be collected at the same nursing school because different conditions such as the sizes or heights of beds and wheelchairs. Therefore, the training data size could be limited because of the number of students in each school. Thus, the use of only ten participants was allowed because this study did not expect to collect a large training dataset for generalization. Note that the participants had no experience as caregivers. All of the experimental procedures were conducted in accordance with the Declaration of Helsinki and the Ethics Committee for Human Research of the Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology (approval number: 19-05).
## Experimental procedure
After putting on the wearable sensors (inertial sensor and shoe-type force sensors) for the proposed method, the participants were asked to lift the lower back of a doll (height 145 cm, weight 10.0 kg) from a wheelchair (height 45.0 cm). This assistive motion for sit-tostand is shown in. This motion, which entails the patient being picked up from the lower back, is common assistive motion. The participants executed a sit-to-stand movement with the doll (simulated patient). The doll was lighter than an actual human due to concerns about the patients placing substantial stress on their lumbar area. After practicing this manual lifting motion (10 min before measurement), measurements were obtained while the participants performed manual lifting using nine different foot position patterns, which are presented inand. Foot distances were normalized to the body height of each participant (unit: %height) in order to avoid the effect of the different body heights between the participants. The left foot was fixed to the anterior, whereas the right foot moved to adjust the foot distances for each foot position pattern. Previous studies indicate that caregivers should stand close to the patient to reduce the lumbar load. Thus, the left foot was always fixed to the same spot between the footrests of the wheelchair during each trial. As shown in, these foot patterns included different combinations of anteroposterior, and 41-55 %height) and mediolateral (10-20, 21-30, and 31-40 %height) foot distances. These foot positions were defined for collecting various foot position data. The participants repeated the manual lifting technique for over five trials for each foot position (a total of 45 trials for each participant), changing the foot distances for each trial in range of each foot position pattern by themselves. Wearable sensors and optical motion capture systems (eight cameras, OptiTrack, Corvallis, OR, USA) measured the assistive motion for sit-to-stand with a 100-Hz sampling rate. The optical markers for the motion capture system were mounted on the heels of both shoe-type force sensors.
lifting technique for over five trials for each foot position (a total of 45 trials for each ticipant), changing the foot distances for each trial in range of each foot position pa by themselves. Wearable sensors and optical motion capture systems (eight cam OptiTrack, Corvallis, OR, USA) measured the assistive motion for sit-to-stand with a Hz sampling rate. The optical markers for the motion capture system were mounte the heels of both shoe-type force sensors. %height is feet distance normalized to body height of each participant. Nine foot positions different combination of anteroposterior and mediolateral distance. Each anteroposterior an diolateral distance has range of value.
# Data analysis
The proposed method was evaluated after obtaining all of the measurement from 450 trials. The input parameters for the machine learning model (mean, stan deviation, skewness, kurtosis, maximum, minimum, and root mean square) were c lated for all wearable sensor signals (front and rear force of each foot, 3-axes accelera and 3-axes angular velocity) using MATLAB R2020b (MathWorks, Natick, MA, USA actual foot position values (anteroposterior and mediolateral feet distances) were c lated using optical motion capture data and the 3D analysis software VENUS 3D R (N Tech. Ltd., Tokyo, Japan). Foot position data were normalized to the body height of participant.
Five machine learning algorithms (ANN, Gaussian process, kNN, linear regre and SVR) were implemented and were verified using the data mining software WEK (University of Waikato, Hamilton, New Zealand). The specifications and paramet %height is feet distance normalized to body height of each participant. Nine foot positions have different combination of anteroposterior and mediolateral distance. Each anteroposterior and mediolateral distance has range of value..
# Data analysis
The proposed method was evaluated after obtaining all of the measurement data from 450 trials. The input parameters for the machine learning model (mean, standard deviation, skewness, kurtosis, maximum, minimum, and root mean square) were calculated for all wearable sensor signals (front and rear force of each foot, 3-axes acceleration, and 3-axes angular velocity) using MATLAB R2020b (MathWorks, Natick, MA, USA). The actual foot position values (anteroposterior and mediolateral feet distances) were calculated using optical motion capture data and the 3D analysis software VENUS 3D R (Nobby Tech. Ltd., Tokyo, Japan). Foot position data were normalized to the body height of each participant.
Five machine learning algorithms (ANN, Gaussian process, kNN, linear regression, and SVR) were implemented and were verified using the data mining software WEKA 3.6 (University of Waikato, Hamilton, New Zealand). The specifications and parameters of each machine learning algorithm are shown in Tables 2-6. The radial basis function (RBF) kernel was applied for Gaussian processes and SVR given that this kernel was used in previous studies related to wearable sensing. . Specifications and parameters of Gaussian process.
## Specification/parameter status/value
Kernel RBF Kernel Hyperparameter γ 1.0 . Specifications and parameters of k-nearest neighbor (kNN).
## Specification/parameter status/value
Weight Uniform Distance
Euclidean Distance Hyperparameter K 1 . Specifications and parameters of linear regression.
## Specification/parameter status/value
Feature Selection M5 Method Ridge Parameter R 1.0 × e −8. Specifications and parameters of support vector regression (SVR).
## Specification/parameter status/value
Training Sequential Minimal Optimization Kernel RBF Kernel Hyperparameter γ 0.01
The accuracy of the proposed method was evaluated from 450 trial data obtained from 10 participants. The accuracies of each machine learning algorithm for the proposed method were calculated using 10-fold cross validation. In the 10-fold cross validation, the dataset was randomly divided into 10 subgroups. Nine subgroups were used as training data, and one subgroup was used as test data. This process was repeated 10 times, and each subsample was used in turn as the test data. As mentioned previously, this study did not expect to prepare a big training dataset for generalization because we considered the fact that the proposed method would be used in different nursing schools. Because of this, it is important that the training data be collected in school that the method is being used in because each school has different conditions. Thus, we considered that these data size and validation processes were suitable. All of the estimated foot positions were obtained from the 10 calculations based on the 10-fold cross validation. The root mean square error (RMSE) between the estimated and actual foot positions was calculated as the error of the proposed method, selecting the machine learning algorithm with the smallest RMSE value as the optimal algorithm. Furthermore, the RMSE values were compared for three feature patterns (only inertial sensor, only shoe-type force sensors, and all wearable sensors) to select the optimal wearable sensors. Through such processes, the optimal combination of wearable sensors and machine learning algorithms can be determined.
Statistical analyses were performed to determine the optimal combination of wearable sensors and machine learning algorithms using EZR 1.32 (Division of Hematology, Saitama Medical Center, Jichi Medical University, Saitama, Japan). Significant differences between the error of the proposed method and zero were evaluated using the paired t-test (significance level: p < 0.05). In addition, power analyses were performed for the paired t-test (α = 0.05). Pearson's correlation between the proposed method and actual values was calculated as the accuracy of the proposed method (significance level: p < 0.05).
The Bland-Altman plot, which represents the mean (horizontal axis) and difference (vertical axis) between estimated and actual values, was used to evaluate fixed and proportional errors of the proposed method using the optimal algorithm. The limitation of agreement (LOA) of the vertical axis in the Bland-Altman plot was used to evaluate the fixed errors. Accordingly, an LOA that does not include zero indicates a fixed error. Pearson's correlation between the mean (horizontal axis) and difference (vertical axis) of the Bland-Altman plot was calculated to evaluate the proportional error (significance level: p < 0.05). Significant correlations between the mean and difference indicates a proportional error.
# Results
## Data specification
The anteroposterior and mediolateral foot distances measured by the optical motion capture system are shown in. Measured foot distanceswere satisfied for the nine different foot positions defined inand.shows vertical acceleration and sagittal angular velocity obtained from the inertial sensor.shows the front force of the foot obtained from the shoe-type force sensor. The results of the inertial data show that vertical accelerations were increased by a longer foot distance. On the other hand, the sagittal angular velocities were decreased by a shorter foot distance. The results of the force data show that the forces of the left (anterior) foot were increased by a longer foot distance. On the other hand, the forces of the right (posterior) foot were decreased by a shorter foot distance. These results indicate that inertial and force data were affected by the foot position changing.
# Results
## Data specification
The anteroposterior and mediolateral foot distances measured by the optical motion capture system are shown in. Measured foot distanceswere satisfied for the nine different foot positions defined inand.shows vertical acceleration and sagittal angular velocity obtained from the inertial sensor.shows the front force of the foot obtained from the shoe-type force sensor. The results of the inertial data show that vertical accelerations were increased by a longer foot distance. On the other hand, the sagittal angular velocities were decreased by a shorter foot distance. The results of the force data show that the forces of the left (anterior) foot were increased by a longer foot distance. On the other hand, the forces of the right (posterior) foot were decreased by a shorter foot distance. These results indicate that inertial and force data were affected by the foot position changing.
## Rmse values
The RMSE values of the anteroposterior and mediolateral foot distances between the proposed method and actual foot position are shown in. Accordingly, the RMSE values of the proposed method using all wearable sensors (inertial sensor and shoetype wearable sensors) were the smallest for both anteroposterior and mediolateral foot distances in three feature patterns. Moreover, the RMSE values of the proposed method using Gaussian process were the smallest for both anteroposterior and mediolateral foot distances in the five algorithms. Based on the aforementioned results, the combination of all of the wearable sensors and the Gaussian processes was determined to be the optimal combination of features and algorithms for the proposed method. The proposed method using the optimal combination of features and algorithm could measure the anteroposterior and mediolateral foot distances with less than 6.5 %height RMSE (anteroposterior: 6.06 %height; mediolateral: 6.30 %height).
# Statistical results
The results of the statistical analyses are summarized in. The errors of the proposed method (Mean ± S.D.) are shown in. Accordingly, no significant differences in both the anteroposterior and mediolateral foot distances were observed between
## Rmse values
The RMSE values of the anteroposterior and mediolateral foot distances between the proposed method and actual foot position are shown in. Accordingly, the RMSE values of the proposed method using all wearable sensors (inertial sensor and shoetype wearable sensors) were the smallest for both anteroposterior and mediolateral foot distances in three feature patterns. Moreover, the RMSE values of the proposed method using Gaussian process were the smallest for both anteroposterior and mediolateral foot distances in the five algorithms. Based on the aforementioned results, the combination of all of the wearable sensors and the Gaussian processes was determined to be the optimal combination of features and algorithms for the proposed method. The proposed method using the optimal combination of features and algorithm could measure the anteroposterior and mediolateral foot distances with less than 6.5 %height RMSE (anteroposterior: 6.06 %height; mediolateral: 6.30 %height).
# Statistical results
The results of the statistical analyses are summarized in. The errors of the proposed method (Mean ± S.D.) are shown in. Accordingly, no significant differences in both the anteroposterior and mediolateral foot distances were observed between
## Rmse values
The RMSE values of the anteroposterior and mediolateral foot distances between the proposed method and actual foot position are shown in. Accordingly, the RMSE values of the proposed method using all wearable sensors (inertial sensor and shoetype wearable sensors) were the smallest for both anteroposterior and mediolateral foot distances in three feature patterns. Moreover, the RMSE values of the proposed method using Gaussian process were the smallest for both anteroposterior and mediolateral foot distances in the five algorithms. Based on the aforementioned results, the combination of all of the wearable sensors and the Gaussian processes was determined to be the optimal combination of features and algorithms for the proposed method. The proposed method using the optimal combination of features and algorithm could measure the anteroposterior and mediolateral foot distances with less than 6.5 %height RMSE (anteroposterior: 6.06 %height; mediolateral: 6.30 %height).
# Statistical results
The results of the statistical analyses are summarized in. The errors of the proposed method (Mean ± S.D.) are shown in. Accordingly, no significant differences in both the anteroposterior and mediolateral foot distances were observed between the error of the proposed method and zero (p > 0.05). Moreover, significant correlations between the proposed method and actual values were observed for both the anteroposterior and mediolateral foot distances (p < 0.05).. Statistical results.
## Foot position paired t-test (error vs. zero) correlation (estimated vs. actual) significant difference power
Anteroposterior N.S. 0.0284 0.891 * Mediolateral N.S. 0.0289 0.692 * %height is feet distance normalized to the body height of each participant. S.D.: standard deviations. Significant differences were evaluated using paired t-tests. Correlation was calculated using Pearson's correlation. N.S.: not significant (p > 0.05); *: significant (p < 0.05).
## Bland-altman plot
Bland-Altman plots, which are presented in, were used to evaluate the proposed method using all of the wearable sensors and Gaussian process. The statistical results for the Bland-Altman plots are detailed in. The LOA of the Bland-Altman plot included zero for both the anteroposterior and mediolateral foot distances, suggesting the absence of fixed errors in the proposed method. However, significant correlations were noted between the difference and mean of the Bland-Altman plots for both the anteroposterior and mediolateral foot distances (anteroposterior: r = 0.682 and p < 0.05; mediolateral: r = 0.758 and p < 0.05), suggesting proportional errors in the proposed method.
# Discussion
## Validity of the proposed method
Our results showed that the proposed method using the optimal combination of wearable sensors and machine learning algorithm could accurately measure anteroposterior foot distances with an RMSE of 6.06 %height. One previous study showed that the number of steps can be a useful measurement unit to provide feedback regarding foot position during manual handling. In addition, the number of steps has enough resolution because our previous study showed that lumbar load might be affected by changing the foot position with at least 10 to 15 %height. Thus, since the anteroposterior RMSE value is shorter than the foot length (foot or shoe size), the proposed
# Discussion
## Validity of the proposed method
Our results showed that the proposed method using the optimal combination of wearable sensors and machine learning algorithm could accurately measure anteroposterior foot distances with an RMSE of 6.06 %height. One previous study showed that the number of steps can be a useful measurement unit to provide feedback regarding foot position during manual handling. In addition, the number of steps has enough resolution because our previous study showed that lumbar load might be affected by changing the foot position with at least 10 to 15 %height. Thus, since the anteroposterior RMSE value is shorter than the foot length (foot or shoe size), the proposed method could be applied to provide feedback about the anteroposterior foot position using the number of steps (foot or shoe). For example, when the measured anteroposterior foot distance is shorter by approximately 10 %height than the optimal distance, the system can provide feedback that says, "please back your rearfoot up by one step". On the other hand, the RMSE of the proposed method for the mediolateral foot distance was 6.30 %height, which was longer than the foot width. Thus, future studies should attempt to minimize measurement errors in the mediolateral foot distances.
Our statistical analyses showed a significant correlation between the proposed method and the actual foot position in both the anteroposterior and the mediolateral feet distances. In addition, there were no significant differences that were seen for the errors of the proposed method and zero. These results indicate that the proposed method involving the use of wearable sensors and a machine learning technique can effectively measure foot positions. Moreover, a significant correlation between the proposed method and the actual foot position in the mediolateral foot distance suggests the possibility of improving errors in the mediolateral foot distance of the proposed method.
The LOA of the Bland-Altman plots showed no fixed error in both the anteroposterior and mediolateral foot distancesand, suggesting that the proposed method could be used to measure foot position during manual lifting without bias or systematic errors. Nonetheless, correlations between the Bland-Altman plots showed proportional errors in both the anteroposterior and mediolateral foot distancesand, indicating the need to focus on proportional errors for the future improvement of the proposed method.
The aforementioned results therefore suggest that the proposed method can be utilized for the monitoring of and giving instructions on quantitative foot positions during assistive motion for sit-to-stand to prevent LBP among caregivers. Furthermore, these results highlight the direction of future improvements in the proposed method.
## Wearable sensors
The RMSE values showed that both the inertial sensor and the shoe-type force sensor were useful for the proposed method. In addition, inertial and force data were affected by the changing foot position. The inertial sensor on the trunk could perceive differences in the foot position given that trunk movement is affected by foot position during manual handling. Moreover, the shoe-type force sensors could measure the force changes caused by the foot position given that the force distribution on the insole changes based on the foot position during manual handling. Therefore, both an inertial sensor and the shoe-type force sensors were considered for the proposed method. Additionally, these wearable sensors can be combined to measure lumbar load and arm movement during manual handling, thereby making them useful for various measurements aimed at preventing LBP for occupational health.
Nonetheless, using only both inertial sensor and shoe-type sensors for the proposed method allowed us to be able to measure the anteroposterior and mediolateral foot distances with less than 7.5 %height RMSE by means of Gaussian process. Improving on such errors by modifying features or algorithms could perhaps allow the use of only either an inertial sensor or shoe-type force sensors for the proposed method. Indeed, reducing the number of sensors would increase user comfort. Furthermore, using only a single inertial sensor for the proposed method can be viable through the use of smartphones given the reliable and valid inertial sensors in current smartphones. According to the current fixation condition for the inertial sensor, smartphones would need to be attached to the users back with a belt. If smartphones are able to be placed in a chest pocket, then the method would be more user friendly. However, the accuracy of the proposed method when placing a smartphone in a chest pocket should be investigated. Such combinations and applications for wearable sensors would ultimately be based on the usability and accuracy required by each user.
## Machine learning algorithms
The proposed method using Gaussian process was able to measure both the anteroposterior and mediolateral foot distances with the smallest RMSE. The reason for these results is that Gaussian process are suitable for machine learning with a small dataset, such as that used in this study. This is because Gaussian process can change the margin of fitting based on the number of data points. Machine learning algorithms that allow the use of small datasets can be useful for actual workspaces given the ease of data acquisition. Therefore, Gaussian process can be considered the optimal machine learning algorithm for the proposed method. On the other hand, the RMSE values of several algorithms were similar. Therefore, there is a possibility that several algorithms other than Gaussian process could also be applied for the proposed method. Future work should continue to explore the optimal algorithm by considering other factors such as implementation or calculation speed.
## Limitation and future works
The proposed method was evaluated by performing assistive motion for sit-to-stand in a laboratory setup. Future studies should evaluate the proposed method in various motions in actual nursing students. An additional limitation is that the foot positions were measured using the heel marker of a global coordinate frame. A single marker on the heel does not approximate a centroid on the foot, and the foot rotation might change the center position of the foot. Moreover, another limitation of this experiment was that the left foot was fixed between the footrests of a wheelchair. There is the possibility that an actual caregiver cannot fix their footing during an assistive motion. The accuracy of the proposed method might be affected by this unstable foot position. Thus, the proposed method should be tested for unstable foot positions during assistive motion. In this study, was another limitation was that the 10 kg doll used in this experiment is lighter than actual patients. This creates a problem, as the method is less reliable due to the difference between the doll used in the experiment and actual patients. According to previous studies, the COP velocity is affected by changing weight. Thus, patient weight might affect the force data obtained from the shoe-type force sensors. Thus, the effect of patient weight on the proposed method should be investigated in future work. In the experiment, each data included the start and end points. The start and end points of assistive motion should be extracted automatically in order for the system to be useful. Previous studies developed an automatic recognition method for patient handling tasks including assistive motion for sitto-stand using footwear sensors, which were similar to our shoe-type force sensors. In future work, these previous methods might be combined with our proposed method for automatic foot position estimation.
As mentioned previously, errors regarding the mediolateral foot width should be improved. Future studies must examine feature selection methodsand other machine learning algorithmsto improve the accuracy of the proposed method. Moreover, linear-based solutions to reduce the proportional error of the proposed method must be developed. The regression model of this study could only be applied for a specific learned task. In future work, a new generalized regression model that learns various patient handling tasks should be considered. Furthermore, readable models such as a biomechanical link modelshould be combined to improve accuracy. The proposed method estimates foot position from an entire motion. In this case, our future system will give feedback regarding foot position immediately after assistive motion. When the system gives feedback indicating that the foot position is not suitable, the caregiver will be able to modify their foot position for the next motion. However, it is desirable that measurements and feedback be made prior to completing an assistive motion. Thus, future work should focus on foot position measurement in the initial posture.
After improvements and further evaluation, monitoring and instruction systems for foot position during manual lifting must be implemented. However, problems remain for system integration using a wireless sensor network. This study could not consider real-time processing, transmission time, and network lifetime. In future work, the proposed method should be implemented for real-time processing based on recent frameworks. For example, Coviello et al. proposed a framework for wireless synchronization based on sending and transmission times. Moreover, Sakuru et al. suggested that sink node selection algorithms based on network lifetime could be applied for wireless sensor networks. Our proposed system will be implemented using a wireless sensor network with these frameworks. Furthermore, a feedback method for foot position using the proposed method should be developed. Foot position might be corrected by means of audio or using a haptic device as well as previous feedback methods used to correct the trunk angle.
Our previous study found that foot position adjustment could reduce lumbar loads during assistive motion for sit-to-stand by using lower limb muscles. In this previous study, the participants performed lifting with the legs and not the back without other instructions. Thus, we considered that a feedback system for foot position would be useful to realize assistive motion using the lower limbs. However, there is a possibility that the best instruction to generate optimal assistive motion to prevent lower back pain requires feedback for another postural parameter, such as the trunk angle. Thus, future studies should consider a combination of both foot position and trunk angle for more effective intervention.
# Conclusions
This study proposed and evaluated a new method for measuring foot position during assistive motion for sit-to-stand in order to prevent LBP among caregivers. The proposed method can be utilized for occupational health for caregivers given that it only uses a few wearable sensors. The following are the contributions of the current study:
## -
The proposed method can measure quantitative foot positions using a single inertial sensor, shoe-type force sensors, and a machine learning algorithm.
## -
The proposed method was evaluated using RMSE values, statistical analysis, and Bland-Altman plots. - Optimal combinations of wearable sensors and machine learning algorithm were explored for the proposed method.
## -
The experimental results showed that a combination of both inertial and shoe-type force sensors and Gaussian process is the optimal combination for the proposed method.
## -
The RMSE values and statistical results indicated that the proposed method could measure foot position during assistive motion for sit-to-stand.
## -
Bland-Altman plots showed that the proportional error should be improved in the proposed method.
The proposed method will certainly contribute to wearable monitoring and instruction systems for manual lifting to prevent LBP among caregivers. We believe that the proposed method can be applied in nursing schools that require training for assistive motion for sit-to-stand.
## Conflicts of interest:
The authors declare no conflict of interest. |
An alternative framework for fluorescence correlation spectroscopy
SupplementaryFigure 4. Comparison with FCS and PCH. a Targeted synthetic fluorescent intensity trace. The time step is 10 µs and the total duration of the trace is 100 s. b PCH curve and the theoretical fit. c FCS curve and best theoretical fit. d The portion of the trace analyzed by our method rebinned at 100 µs. e The number of molecules in the effective volume with =1, arising from the trace in (d). Exact value of the number of molecules is shown by the green line and the PCH and FCS estimates are shown by the dashed and solid pink lines. g On the posterior probability distribution of the molecular brightness we superpose the PCH estimate of the molecular brightness (pink dashed line) and the true value (green dashed line). h On the posterior probability distribution of the background photon emission rate we superpose the PCH estimate of the background photon emission rate (pink dashed line) and the true value (green dashed line). f The posterior probability distribution of the diffusion coefficient obtained by analyzing the trace in (d). The FCS estimate of the diffusion coefficient obtained by analyzing the total trace, shown in (a), illustrated by a pink dashed line with the exact value (green dashed line). The targeted synthetic trace is generated by freely diffusive molecules with diffusion coefficient, molecular brightness and background photon emission rates of of 10 µm 2 s −1 , 5 × 10 4 photons s −1 and 10 3 photons s −1 , respectively. The 95% confidence intervals of the posteriors over the diffusion coefficient, molecular brightness and background photon emission rates are highlighted in cyan.
Here we provide supplementary materials and technical details that complement the main text. These include: (i) Additional analysis results that demonstrate the estimation of molecular brightness and background photon emission rates, joint posterior probability distributions, molecule locations, and additional results for multiple diffusive species. These results are repeated for simulated and experimental data. The FCS estimate is shown by a magenta dashed line. c The joint probability distribution of the diffusion coefficient and the molecular brightness. d The posterior probability distribution of the molecular brightness. e The joint probability distribution of the diffusion coefficient and molecular brightness. f The joint probability distribution of the molecular brightness and background photon emission rates. g The posterior probability distribution of the background photon emission rate and the 95% confidence intervals of the posteriors are highlighted in cyan. The experimental fluorescent intensity trace was produced with a concentration of 100 pM of Cy3 in a 94% glycerol/water mixture.. Estimated joint posterior probability distribution. a Experimental fluorescent intensity trace used inwith a length of 1000 data points and time step 100 µs. b The posterior of the diffusion coefficient. The FCS estimate is shown by a magenta dashed line. c The joint probability distribution of the diffusion coefficient and the molecular brightness. d The posterior probability distribution of the molecular brightness. e The joint probability distribution of the diffusion coefficient and the molecular brightness. f The joint probability distribution of the molecular brightness and background photon emission rates. g The posterior probability distribution of the background photon emission rate and the 95% confidence intervals of the posteriors are highlighted in cyan. The experimental fluorescent intensity trace produced is with a concentration of 1 nM of Cy3 in a 94% glycerol/water mixture. Supplementary. Comparison with FCS and PCH. a Targeted experimental fluorescent intensity trace. The time step is 10 µs with a total time of 5 min. b PCH curve and the theoretical fit. c FCS curve and the best theoretical fit. d The portion of the trace analyzed by our method rebinned at 100 µs. e The concentration of Cy3 in the effective volume with =1, arising from the trace in (d). The experimental concentration is shown by the green line and the PCH estimated is shown by the pink line. g The posterior probability distribution of the molecular brightness with the PCH estimated of the molecular brightness shown by a solid green line. h The posterior probability distribution of the background photon emission rate with the PCH estimate of the background photon emission rate shown by a solid green line. f The posterior probability distribution of the diffusion coefficient obtained by analyzing the trace in (d). The FCS estimate of the diffusion coefficient obtained by analyzing the total time trace, shown in (a), is denoted by a pink solid line. The targeted experimental trace is generated by free diffusive Cy3 in a mixture of water and glycerol with 75% glycerol, a laser power of 100 µW and a concentration of Cy3 at 1 nM, excitation wavelength, NA and refractive index used are 532 nm, 1.42 and 1.4, respectively.
## Supplementary
Supplementary. Experimental traces of free Cy3B dyes using an elongated confocal volume. a Experimental fluorescent intensity trace used in FCS. The time trace is generated by 2.5 nM Cy3B dyes in glycerol/water mixture with 70% glycerol and laser power of 100 µW. b Auto-correlation curve of the trace in (a) and best theoretical fit. c Portion of the trace in (a) to be used as the input to FCS and our method. d Auto-correlation curve of trace in (c). e Posterior probability distribution over the diffusion coefficient estimated from the trace in (c). Traces shown in (a) and (c) are acquired at 100 µs for a total of 10 s and 0.1 s respectively. The laser power use to generate the signal (a) is 100 µW (measured before the beam enters the objective). The estimation of the diffusion coefficient as the results of autocorrelation fitting in (a) matched with Stokes-Einstein prediction, equal to 18.79 µm 2 s −1 and in (d) is 145.75 µm 2 s −1 .. Testing the experimental setup for effects of saturation. a Fluorescence count rate shown here is plotted with respect to the laser intensity. Laser powers are, conservatively, measured at the output of the laser. As such, we expect these to be lower at the entrance of the objective. b Auto-correlation curves of the time traces used in (a). Only three of the six acquired FCS decays are shown for clarity (they all overlap within experimental error). Fluorescent intensity time traces were obtained using a solution of Cy3B dyes in glycerol/water mixture with 70% glycerol at six different laser powers in the 10-100 µW range. Despite a slight nonlinearity in (a), there is no effect of saturation within the experimental error in (b). Supplementary Note 3: Detailed methods description 1. Representation of molecular diffusive motion Consider a particle moving in 1D diffusion. The probability distribution p(x, t) of the particle's location obeys Fick's second lawand is given by the diffusion equation
## Saturation test
## Analysis of additional data for multiple diffusive species
[formula] ∂p ∂t = D ∂ 2 p ∂x 2 Supplementary Equation 1 [/formula]
where D is the particle's diffusion coefficient. Assuming the particle is located at x k−1 at a time t k−1 , i.e. assuming the initial condition p(x, t k−1 ) = δ(x − x k−1 ), and a free space boundary, i.e. lim x→±∞ p(x, t) = 0, we can solve this equation to obtain p(x, t) for any later time t. The solution is
[formula] p(x, t) = exp − (x−x k−1 ) 2 4(t−t k−1 )D 4π(t − t k−1 )D Supplementary Equation 2 [/formula]
which equals to the probability density of a normal random variable with mean x k−1 and variance 2(t − t k−1 )D, see . At time t = t k , we therefore have
[formula] x k |x k−1 ∼ Normal (x k−1 , 2(t k − t k−1 )D) . [/formula]
## Supplementary equation 3
Similarly, solving the diffusion equation for particles following isotropic 3D diffusion in free space, we have
[formula] x k |x k−1 ∼ Normal (x k−1 , 2(t k − t k−1 )D) y k |y k−1 ∼ Normal (y k−1 , 2(t k − t k−1 )D) z k |z k−1 ∼ Normal (z k−1 , 2(t k − t k−1 )D) Supplementary Equation 4 [/formula]
which constitute the molecular motion model used throughout this study.
## Description of stokes-einstein model
For the experimental data, we benchmark our estimates of the diffusion coefficient against the Stokes-Einstein prediction. Namely, for a spherical particle in a quiescent fluid at uniform temperature
[formula] D = kT 6πrη Supplementary Equation 5 [/formula]
where, D is the diffusion coefficient, k is Boltzmann's constant, T is the solution's absolute temperature, r is the hydrodynamic radius of the particle and η is the solution's dynamic viscosity.
## Fcs formulation
The formulation we used in this study to autocorrelate the synthetic and experimental time traces is
[formula] G ex (τ ) = I(t + τ )I(t) I(t) 2 − 1 Supplementary Equation 6 [/formula]
where the I(t) is the number of detected photons at time t. The computational implementation uses the Wiener-Khinchin theorem.
The theoretical functionused to fit the autocorrelation curves, for a 3DG PSF, is
[formula] G th (τ ) = 1 N 1 − F + F e − τ τ F (1 − F ) 1 1 + 4Dτ ω 2 xy 1 (1 + 4Dτ ω 2 z ) 1 2 [/formula]
## Supplementary equation 7
and, for the 2DGL PSFs, is
[formula] G th (τ ) = 1 N 1 − F + F e − τ τ F (1 − F ) 1 1 + 4Dτ ω 2 xy + 1 Supplementary Equation 8 [/formula]
where N is the average number of molecule in the effective volume, D is the diffusion coefficient, τ F is the triplet state relaxation time and F is the fraction molecules populating the triplet state.
To find the best fit, we use χ 2 minimization [10-12]
[formula] χ 2 = τ (G th (τ ) − G ex (τ )) 2 . [/formula]
Supplementary Equation 9
## Definition of molecular brightness
As the definition of molecular brightness in Eq. 2, we use the emission rate of detected photons of a single fluorophore. For a fluorophore located at (x, y, z) this is formulated as
[formula] µ(x, y, z) = µ 0 ϕ d ϕ qe ϕ f σ EXC(x, y, z) CEF(x, y, z) Supplementary Equation 10 [/formula]
where, µ 0 is the maximum excitation intensity which occurs at the center of the confocal volume, ϕ d is the efficiency of the photon collection at the center of the confocal volume, ϕ qe is the quantum efficiency of the detector, ϕ f is the quantum efficiency of the fluorophore (i.e. quantum yield), σ is the absorption cross-section of the fluorophore, EXC(x, y, z) is the excitation profile and CEF(x, y, z) is the detection profile, i.e. collection efficiency function, which equals the fraction of the photons collected by the detector to the total photons emitted by a point source.
## To obtain eq. 2, we cast supplementary equation supplementary equation 10 in the simplified form
[formula] µ(x, y, z) = µ mol PSF(x, y, z) Supplementary Equation 11 [/formula]
where µ mol = µ 0 ϕ d ϕ qe ϕ f σ, which we term molecular brightness at the center of the confocal volume, and PSF(x, y, z) = EXC(x, y, z) CEF(x, y, z), which we term the PSF.
To relate the parameter µ mol to the average photon count rate, which is commonly estimated in bulk experiments, we consider the spatial average of µ(x, y, z) as follows µ(x, y, z) = µ mol PSF(x, y, z) .
## Supplementary equation 12
For the specific choice of a 3DG PSF (see below), the average is computed as follows
[formula] PSF(x, y, z) = +∞ −∞ +∞ −∞ +∞ −∞ exp −2 x 2 ω 2 xy − 2 y 2 ω 2 xy − 2 z 2 ω 2 z dxdydz V eff = π 2 ω 2 xy π 2 ω 2 xy π 2 ω 2 z 1 V eff Supplementary Equation 13 [/formula]
where V eff denotes the effective volume of 3DG PSFand it is given by
[formula] V eff = π 3 2 ω 2 xy ω z . [/formula]
## Supplementary equation 14
Consequently, Supplementary Equationimplies
[formula] µ mol = √ 8 µ(x, y, z) . [/formula]
## Supplementary equation 15
In other words, the molecular brightness is, by definition, approximately 2.8 times larger than the average photon count rate of a single molecule.
## Definition of point spread function models
In this study we use three different point spread functions as approximations to the more realistic Airy function, namely a 3D-Gaussian (3DG), a 2D-Gaussian-Cylindrical (2DGC)and a 2D-Gaussian-Lorentzian (2DGL).
The definition of the PSF for the 3DG case is
[formula] PSF 3DG (x, y, z) = exp −2 x 2 + y 2 ω 2 xy − 2 z 2 ω 2 z Supplementary Equation 16 [/formula]
while, the definition of the PSF for the 2DGC case is
[formula] PSF 2DGC (x, y, z) = exp −2 x 2 ω 2 xy − 2 y 2 ω 2 xy . [/formula]
## Supplementary equation 17
For both cases, ω xy and ω z are the semi-axes lateral and parallel to the optical axis. These are represented in terms of the excitation wavelength λ exc , solution refraction index n sol , and numerical aperture NA of the microscope as ω xy = 0.61λ exc /NA and ω z = 1.5n sol λ exc /NA 2 ; for example see. For more realistic representations, ω xy and ω z can be estimated directly based on calibration experiments with known diffusion coefficients; for example see. The definition of the PSF for the 2DGL case is
[formula] PSF 2DGL (x, y, z) = 1 1 + z zR 2 exp −2 x 2 +y 2 ω 2 xy 1 + z z R 2 Supplementary Equation 18 [/formula]
where ω xy , λ exc , and n sol are similar to the 3DG or 2DG cases and z R = n sol πω 2 xy /λ exc .
## Description of the data simulation
To generate fluorescence intensity time traces that mimic a realistic confocal setup, we simulate molecules movingthrough an illuminated 3D volume. The number of moving molecules N is prescribed in each simulation. To maintain a relatively stable concentration of molecules near the confocal volume, and so to avoid generating traces where every molecule eventually strays into un-illuminated regions, we impose periodic rectangular boundaries to our volume. The boundaries are placed at ±L xy perpendicular to the focal plane and ±L z perpendicular to the optical axis.
We assess the locations of the molecules x n k , y n k , z n k , where k = 1, . . . , K label time levels and n = 1, . . . , N label molecules, at equidistant time intervals t 1 , t 2 , . . . , t K . The time interval between successive assessments δt = t k − t k−1 , as well as the total trace duration T total = t K − t 0 , are prescribed.
Molecule locations at the first assessment x n 1 , y n 1 , z n 1 are sampled randomly from a uniform distribution with limits equal to the boundaries ±L xy and ±L z of our pre-specified volume. Subsequent locations are generated according to the diffusion model described above under a prescribed diffusion coefficient D.
Finally, we obtain individual photon emissions w k by simulating Bernoulli random variables of success probability q k = 1 − e −µ k δt , where the rate µ k gathers single photon contributions from the background and the entire molecule population according to
[formula] µ k = µ back + µ mol N n=1 PSF(x n k , y, n k z n k ) Supplementary Equation 19 [/formula]
where both background and molecular brightness, µ back and µ mol , are prescribed. The PSF model is also prescribed. To avoid artifacts induced by the periodic boundaries we impose in our volume, we ensure that L xy ω xy , L z ω z , or L z z R , where ω xy , ω z and z R characterize the geometry of the confocal volume, see
## Definition of normalized distance and numbers of molecules
As we need to estimate the positions of the molecules with respect to the center of the confocal volume, which is the point of origin, in order to ultimately estimate the number of molecules as a proxy for molecule concentration, for example Supplementaryand, we must address difficulties associated with symmetries of the confocal PSF with respect to rotations around the optical axis or the focal plane.For this, for a molecule at (x n k , y n k , z n k ), when the 3DG PSF is used,, we rely on where d n k is the normalized distance with respect to the optical axis of molecule n at time k.
These distances are obtained by setting the respective PSFs equal to exp(−(d n k ) 2 ) and are unaffected by the aforementioned symmetries, i.e. x n k → −x n k , y n k → −y n k , and z n k → −z n k .
For a given normalized distance , we define the number of molecules N k as the number of estimated (active) molecules within the corresponding distance. That is
[formula] N k = n b n H 1 − d n k Supplementary Equation 23 [/formula]
where H is the Heaviside step function, b n is the load of molecule n, and V is the volume of a designated effective region chosen to agree with the effective volume V eff used in FCS.
## Description of the time trace preparation
The initial time trace consists of single photon arrival times which are computationally too expensive to analyze. Our method instead operates on photon intensity traces which are either obtained directly during an experiment or obtained from individual photon arrival time traces after binning. To transform single photon arrival time traces into intensity time traces, we use time bins of fixed size (main size) that typically span multiple photon arrival times. To speed up the computations further, as some bins have none of very few photons, over certain portions of the trace we use larger bins (auxiliary size). Briefly, the user specifies a minimum number of photons per bin as a lower threshold. As illustrated in Supplementary, those bins, preselected at the main size, containing fewer photons than the specified threshold are enlarged uniformly in order to achieve an average of at least as many photons as specified by the threshold. This occasional adaptation, from the main to the auxiliary bins, becomes important in the analysis of traces from experiments held near single molecule resolution where molecule concentrations are low so that on average only one molecular passage through the confocal volume happens. Consequently, photon intensities are low, and thus the bulk of computational time otherwise would had been spent processing trace portions of poor quality (i.e. with few or no photons).
To carry our the necessary computations, as we detail shortly, we use the Anscombe transformationto approximate the Poissonian likelihoods of photon intensities (see below). This approximation is robust as long as bins contain on average 4 photons or more. Thus, as a minimum requirement, we also use the aforementioned threshold to ensure the validity of the approximations.
Supplementary Note 4: Detailed description of the inference framework
## . description of prior probability distributions
The model parameters in our framework that require priors are: the diffusion coefficient D; the molecular brightness and background photon emission rates µ mol and µ back ; the initial molecule locations x n 1 ,y n 1 ,z n 1 ; and load prior weights q n . As we already mentioned in the main text, a prior on the population of diffusing molecules is implicitly defined by the prior on both b n and q n . Our choices are described below.
## Prior on the diffusion coefficient
To ensure that the D sampled in our formulation attains only positive values, we place an inverse-Gamma prior
[formula] D ∼ InvGamma (α D , β D ) . [/formula]
## Supplementary equation 24
Besides ensuring a positive D, this prior is also conjugate to the motion model we use which facilitates the computations (see below).
## Priors on molecular brightness and background photon emission rates
To ensure that µ mol and µ back sampled in our formulation attain only positive values, we place Gamma priors on both
[formula] µ mol ∼ Gamma (α mol , β mol ) µ back ∼ Gamma (α back , β back ) . [/formula]
## Supplementary equation 25
Due to the specific dependencies of the likelihood (that we will discuss shortly) on the photon emission rates, conjugate priors cannot be achieved for µ mol and µ back . So, the above choice offers no computational advantage (see below) and could be readily replaced with more physically motivated choices if additional information on molecular brightness becomes available.
## Priors on initial molecule locations
Due to the symmetries in the confocal PSF, i.e. a molecule at a location (x, y, z) emits the same average number of photons as a molecule at locations (±x, ±y, ±z), we are unable to gain insight regarding the octant of the 3D Cartesian space in which each molecule is located. To avoid imposing further assumptions on our framework that may determine each molecule's octant uniquely, but may limit the framework's scope to specific experimental setups, we place priors on the initial locations that respect these symmetries. Accordingly, in our framework, at the onset of the measuring period, molecules are equally likely to be located at any of the positions (±x n 1 , ±y n 1 , ±z n 1 ). To facilitate the computations (see below), we place independent symmetric normal distributions, see , on each Cartesian coordinate of the model molecules
[formula] x n 1 ∼ SymNormal µ xy , σ 2 xy y n 1 ∼ SymNormal µ xy , σ 2 xy z n 1 ∼ SymNormal µ z , σ 2 z . [/formula]
## Supplementary equation 26
We want to emphasize that the symmetric priors above do not affect our estimates. According to the motion model we employ, no matter where molecules are initiated, they may subsequently move freely and eventually switch to a different octant if warranted by the data. Our symmetric priors merely indicate that for each individual molecular trajectory considered, there are another 7 symmetric trajectories that are equally likely to have occurred.
## Priors and hyperpriors for molecule loads
To facilitate the computations (described next), we use a finite, but large, model population consisting of N molecules containing both active and inactive molecules. These model molecules are collectively indexed by n = 1, 2, . . . , N . As explained in the main text, estimating how many molecules are actually warranted by the data under analysis is equivalent to estimating how many of those N molecules are active, i.e. b n = 1, while the remaining inactive ones, i.e. b n = 0, have no impact whatsoever and are instantiated only for computational purposes.
To ensure that each load b n takes only values 0 or 1, we place a Bernoulli prior of weight q n . In turn, on each weight q n , we place a conjugate Beta hyperprior b n |q n ∼ Bernoulli(q n ) Supplementary Equation 27 q n .
## Supplementary equation 28
To ensure that the resulting formulation avoids overfitting, we make the specific choices A q = α q /N and B q = β q (N − 1)/N . Under these choices, and in the limit that N → ∞; that is, when the assumed molecule population is allowed to be large, this prior/hyperprior converge to a Beta-Bernoulli process. Consequently, for N 1, the posterior remains well defined and becomes independent of the chosen value of N . In other words, provided N is large enough, its impact on the results is insignificant; while its precise value has only computational implications (see below). w k |{x n k , y n k , z n k , b n } n , µ mol , µ back ∼ Poisson (µ k ) , k = 1, . . . , K Supplementary Equation 40
## Summary of model equations
[formula] µ k = (t k − t k−1 ) µ back + µ mol n b n PSF(x n k , y n k , z n k ) . Supplementary Equation 41 [/formula]
For molecules diffusing in a confocal volume that is extremely elongated over the optical axis, the PSF approaches a cylindrical one. In this case, it is safe to eliminate the
## Description of the computational scheme
The joint probability distribution of our framework is p(D, µ mol , µ back , {q n , b n , x n , y n , z n } n |w), where molecular trajectories and intensities (measurements) are gathered in Due to the nonlinearities in the PSF and the non-parametric prior on q n and b n , analytic evaluation or direct sampling of this posterior is impossible. For this reason, we develop a specialized Markov chain Monte Carlo (MCMC) scheme that can be used to generate pseudo-random samples. This scheme is explained in detail below.
In order to terminate the MCMC sampler, we need to determine when a representative number of samples has been computed. To do so, we divide the samples already computed into four portions and compare the mean values of the diffusion coefficient of the two last ones
[formula] η 1 = 3I/4 i=2I/4 D i I/4 , η 2 = I i=3I/4 D i I/4 [/formula]
Supplementary Equation 48
where, η 1 and η 2 are the mean values of the two last portion of the sampled diffusion coefficients denoted D i and I is the total number of computed MCMC samples thus far. Following, we terminate the sampler when |η 1 −η 2 | < thr , where thr is a pre-specified threshold. Also, to avoid incorporating burn-in samples in the calculations, we ensure a minimum number of iterations I of no less than 10 4 . A working implementation of the resulting scheme in source code and GUI forms, see Supplementary, are available through the provided source code.
## Overview of the sampling updates
Our MCMC exploits a Gibbs sampling scheme. Accordingly, posterior samples are generated by updating each one of the variables involved sequentially by sampling conditioned on all other variables and measurements w.
Conceptually, the steps involved in the generation of each posterior sample (D, µ mol , µ back , {q n , b n , x n , y n , z n } n ) are:
(1) For each n in the molecule population. A working implementation of the framework described in this study is available through the provided source code. Along with this implementation, we provide a graphical user interface (GUI) that can be used to analyze intensity traces from confocal microscopy.
(5) Update jointly the loads b n for all model moleculesUpdate jointly the molecular brightness and background photon emission rates µ mol and µ back , respectively
These steps are described in detail below.
## Sampling of active molecule trajectories
For a given active molecule n, we update the trajectory x n by sampling from the corresponding conditional p(x n |D, µ mol , µ back , {b n , y n , z n } n , {x n } n =n , w), which we achieved through backward sampling.
To be able to sample a trajectory x n in backward sampling, we factorize the density p(x n |D, µ mol , µ back , {b n , y n , z n } n , {x n } n =n , w) as p(x n |D, µ mol , µ back , {b n , y n , z n } n , {x n } n =n , w) = p(x n 1 |x n 2 , D, µ mol , µ back , {b n , y n 1 , z n 1 } n , {x n 1 } n =n , w) × p(x n 2 |x n 3 , D, µ mol , µ back , {b n , y n 2 , z n 2 } n , {x n 2 } n =n , w) . . . × p(x n K−1 |x n K , D, µ mol , µ back , {b n , y n K−1 , z n K−1 } n , {x n K−1 } n =n , w) × p(x n K |D, µ mol , µ back , {b n , y n K , z n K } n , {x n K } n =n , w).
## Supplementary equation 49
According to this factorization, we sample x n , starting from x n K and move backward towards x n 1 . To start the sampling steps, we need to determine each one of the individual densities p(x n K |D, µ mol , µ back , {b n , y n K , z n K } n , {x n K } n =n , w) and p(x n k |x n k+1 , D, µ mol , µ back , {b n , y n k , z n k } n , {x n k } n =n , w). We do this in a forward filtering approachwhich is described in detail below. where w 1:k is an abbreviation for w 1 , . . . , w k and excess parameters are shown by dots. Since the density p(x n k+1 |x n k , D) is already known, to sample x n k in backward sampling, we only need to determine the filter density p(x n k |D, . . . , w 1:k ). To be able to apply forward filtering and compute p(x n k |D, . . . , w 1:k ) efficiently, we use an approximate model, where Supplementary Equation 40, is replaced with T data (w k )|{x n k , y n k , z n k , b n } n , µ mol ∼ Normal (T mean (µ k ), 1) , k = 1, . . . , K.
## Supplementary equation 51
Here, µ k stems fromfor 3D models andfor 2D models; while, T data (w) and T mean (µ) denote Anscombe transformedvariables defined as follows
[formula] T data (w) = 2 w + 3 8 Supplementary Equation 52 [/formula]
T mean (µ) = 2 µ +
[formula] 3 8 − 1 4 √ µ . Supplementary Equation 53 [/formula]
The Anscombe transform exploited here offers a way of transforming Poisson random variables into (approximately) normal oneswhich facilitates the filtering process described next. The approximation we employ is highly accurate for µ 1, while acceptable accuracy is maintained so long as µ > 4 photons. Increasing the accuracy of the above approximation is achieved with either longer data acquisition times or higher laser powers.
Under the Anscombe transform, the densities of both the dynamics,, and observations,, are normally distributed. So, we can compute the filter distribution p(x n k |D, . . . , w 1:k ) of the approximate model similar to the standard theory underlying nonlinear Kalman filters.
More specifically, because the mean of the transformed observation distribution, T mean (µ k ) is a nonlinear function of the location x n k , to apply the Kalman filters we need to approximate the transformed observation distribution in such a way that its mean becomes a linear function of the location x n k . To do so, we use two common approaches: (i) extended Kalman filter (EKF), which locally approximate the transformed observation distribution around selected values; and (ii) unscented Kalman filter (UKF), which globally approximate the transformed observation distribution.
As explained in detail in, the linearization alone is not sufficient to properly approximate the filter. This is because both EKF and UKF assume that the filter is a normal density. This assumption is problematic for our particular case which is symmetric across the origin, i.e. observations provide equal probabilities for the molecule to be in negative or positive side of the center of the PSF, i.e. ±x n k . Due to this symmetry across the yz-plane, the filtering distribution consists of two modes centered symmetrically across the origin. Therefore, we compute an approximate filter distribution of the form p (x n k |D, . . . , w 1:k ) ≈ SymNormal (x n k ; m n k , c n k ) Supplementary Equation 54
where SymNormal (m n k , c n k ) denotes the symmetric normal distribution (see . The filter's parameters m n k and c n k can be computed recursively according to p (x n k |D, . . . , w 1:k ) ∝ p w k |x n k , y n k , z n k , µ mol , µ back , {b n , x n , y n , z n } n which, for our model, reduces to p (x n k |D, . . . , w 1:k ) ∝ Normal (T data (w k ); T mean (µ k ), 1) SymNormal x n k ; m n k−1 , c n k−1 + 2D (t k − t k−1 ) Supplementary Equation 56 and, in turn, is approximated as p (x n k |D, . . . , w 1:k ) ≈ SymNormal (x n k ; m n k , c n k ) .
## Supplementary equation 57
To summarize, in the forward pass of the FFBS, we compute m n k and c n k of the filter of the molecule n, for all time levels k = 1, . . . , K, by linearizing the approximate model around x n 1 = µ xy for k = 1, and around x n k = m n k−1 for k = 2, . . . , K. Since our observation is nonlinear, to calculate the filter, we opt between two different methods: (i) Extended Kalman filter (EKF) and (ii) Unscented Kalman filter (UKF).
In the EKF, we linearize the observations to obtain a closed form for the filter (local approximation) and in the UKF we approximate the joint probability distribution of observations and locations with a multivariate normal distribution (global approximation). The reason to use either of these filters is that the EKF is computationally cheaper but less accurate. According to our analysis it may fail to provide unbiased estimates of the background photon emission rate. On the other hand, the UKF is more robust and provides background emission rate estimates, but these benefits come at an increased computational cost.
In this study, we provide both filters and allow the user to choose between them.
## Extended kalman filter
Within the EKF approximation, the normal probability distribution preceding the symmetric normal ofis linearized in order for their product to become a symmetric normal one.
In this case, we linearize the mean of the observation density Normal (T data (w k ); T mean (µ k ), 1), around the modes of the filter in the previous time step
[formula] T mean (µ k (x n k )) ≈ T mean µ k −m n k−1 + ∂T mean (µ k (x n k )) ∂x n k x n k =−m n k−1 x n k + m n k−1 T mean (µ k (x n k )) ≈ T mean µ k +m n k−1 + ∂T mean (µ k (x n k )) ∂x n k x n k =+m n k−1 x n k − m n k−1 [/formula]
## Supplementary equation 58
where the first term linearizes around x n k = −m n k−1 and the second term linearizes around x n k = +m n k−1 . Under these approximations,attains an analytical solution. In detail
[formula] Normal (T data (w k ); T mean (µ k ), 1) SymNormal x n k ; m n k−1 , c n k−1 + 2D (t k − t k−1 ) = Normal (T data (w k ); T mean (µ k ), 1) Normal x n k ; −m n k−1 , c n k−1 + 2D (t k − t k−1 ) + Normal (T data (w k ); T mean (µ k ), 1) Normal x n k ; +m n k−1 , c n k−1 + 2D (t k − t k−1 ) = Normal x n k ; −m n k−1 + e n k d n k , 1 (d n k ) 2 Normal x n k ; −m n k−1 , c n k−1 + 2D (t k − t k−1 ) + Normal x n k ; +m n k−1 − e n k d n k , 1 (d n k ) 2 Normal x n k ; +m n k−1 , c n k−1 + 2D (t k − t k−1 ) = 1 2 [/formula]
Normal (x n k ; −m n k , c n k ) + 1 2
Normal (x n k ; +m n k , c n k ) = SymNormal (x n k ; m n k , c n k ) .
## Supplementary equation 59
The same calculations apply also for k = 1, where the starting density is replaced with the prior of Supplementary Equation. In this case
[formula] c n 1 = σ 2 xy S(µ 1 ) 2 + σ 2 xy (d n 1 ) 2 m n 1 = µ xy + c n 1 d n 1 e n 1 d n 1 = ∂T mean (µ 1 (x n 1 )) ∂x n 1 x n 1 =µxy e n 1 = T data (w 1 ) − T mean (µ 1 (x n 1 )) x n 1 =µxy [/formula]
## Supplementary equation 60
while for k = 2, . . . , K are
[formula] c n k = c n k−1 + 2D (t k − t k−1 ) 1 + c n k−1 + 2D (t k − t k−1 ) (d n k ) 2 m n k = m n k−1 + c n k d n k e n k d n k = ∂T mean (µ k (x n k )) ∂x n k x n k =m n k−1 e n k = T(w k ) − T(µ k (x n k )) [/formula]
x n k =m n k−1 .
## Supplementary equation 61
Unscented Kalman filter
The unscented Kalman filtertries to fit the joint probability distribution of the observations and locations globally with a multivariate normal distribution to cope with the nonlinearity in Supplementary Equation. Specifically the product ofis approximated as follows
[formula] Normal (T data (w k ); T mean (µ k ), 1)SymNormal x n k ; m n k−1 , c n k−1 + 2D (t k − t k−1 ) ≈ 1 2 BNormal x n k T data (w k ) ; − X n k W n k , xx Σ n k , − xw Σ n k − wx Σ n k , ww Σ n k + 1 2 BNormal x n k T data (w k ) ; + X n k W n k , xx Σ n k , + xw Σ n k + wx Σ n k , ww Σ n k ∝ 1 2 Normal (x n k ; −m n k , c n k ) + 1 2 [/formula]
Normal (x n k ; +m n k , c n k ) = SymNormal (x n k ; m n k , c n k ) Supplementary Equation 62
Since we are faced with a filter which has two symmetric modes, we calculate the filter's mean m n k and variance c n k for one of the modes only, while we recover the other mode's mean and variance by reflection.
The means, auto-and cross-covariances in one mode of the Supplementary Equation 62 are given by
[formula] X n k = +∞ −∞ xq(x)dx W n k = +∞ −∞ T mean (µ k (x)) q(x)dx xx Σ n k = +∞ −∞ (x − X n k ) T (x − X n k )q(x)dx ww Σ n k = +∞ −∞ (T data (µ k ) − W n k ) T (T data (µ k ) − W n k )q(x)dx + 1 xw Σ n k = +∞ −∞ (x − X n k ) T (T data (µ k ) − W n k ) q(x)dx wx Σ n k = +∞ −∞ (T data (µ k ) − W n k ) T (x − X n k ) q(x)dx [/formula]
## Supplementary equation 63
where q(x) = Normal x; m n k−1 , c n k−1 + 2D (t k − t k−1 ) is the probability density of one mode of the filter. The same formula applies to the other mode too.
To calculate the mean value m n k and variance c n k of each normal contributing to the symmetric normal shown above, we need to specify a set of sample points, termed sigma points in the UKF literature, to estimate the mean values and covariance matrix of the bivariate normal on which m n k and c n k depend. To specify sigma points, we first calculate sigma points x sn i and their weights g * i for a standard normal Normal(x; 0, 1) as following according to. We then transform x sn i that will be used in this Normal(x; m k−1 , c k−1 + 2D(t k − t k−1 )). The transformed sigma points are
[formula] x * i = m k−1 + x sn i c k−1 + 2D (t k − t k−1 ). [/formula]
## Supplementary equation 64
Finally, given g * i , x * i , we calculate the mean and covariance of the bivariate normal previously introduced by
[formula] X n k = ∞ −∞ xq(x)dx ≈ i g * i x * i W n k = ∞ −∞ T x (x) q(x)dx ≈ i g * i T x (x * i ) xx Σ n k = ∞ −∞ (x − x M k ) T (x − x M k )q(x)dx ≈ i g * i ( x M k − x * i ) T ( x M k − x * i ) ww Σ n k = ∞ −∞ (T x (x) − w M k ) T (T x (x) − w M k ) q(x)dx + 1 ≈ i g * i ( w M k − T x (x * i )) T ( w M k − T x (x * i )) + 1 xw Σ n k = ∞ −∞ (x − x M k ) T (T x (x) − w M k )q(x)dx ≈ i g * i ( w M k − T x (x * i )) T ( x M k − x * i ) wx Σ n k = ∞ −∞ (T x (x) − w M k ) T (x − x M k )q(x)dx ≈ i g * i ( x M k − x * i ) T ( w M k − T x (x * i )) . [/formula]
## Supplementary equation 65
After computing the means X n k and W n k and auto-covariances and cross-covariances xx Σ k , ww Σ k , xw Σ k , wx Σ k , the mean and variance of each mode of the filter are given by
[formula] m n k = X n k + K n k (T data (w k ) − W n k ) Supplementary Equation 66 c n k = xx Σ n k − K n k ( ww Σ n k )K n k T Supplementary Equation 67 K n k = xw Σ n k ww Σ n k . [/formula]
## Supplementary equation 68
b. Backward sampling
After we compute the filter densities p (x n k |D, . . . , w 1:k ) in the forward filtering step, through the EKF or UKF, we are able to sample the location x n k by using backward sampling as in. Specifically, given a computed filter, we sample sequentially x n k according to
x n K ∼ p x n K |{x n k } k <K , D, µ mol , µ back , {b n , y n , z n } n , {x n } n =n , w Supplementary Equation 69
x n k ∼ p x n k |x n k+1 , {x n k } k <k , D, µ mol , µ back , {b n , y n , z n } n , {x n } n =n , w , k = 1, . . . , K − 1.
## Supplementary equation 70
Due to the specific choices of our problem these reduce to
x n K ∼ SymNormal (m n K , c n K ) Supplementary Equation 71
x n k ∼ SymNormal (m n k , c n k ) × Normal x n k+1 , 2D(t k+1 − t k ) , k = 1, . . . , K − 1where m n k and c n k are the parameters of the filter which are calculated in the forward filtering step.
## Sampling of inactive molecule trajectories
After updating the trajectories of the active molecules, we update the trajectories of the inactive ones. For this, we sample from the corresponding conditionals p({x n , y n , z n } n:b n =0 |D, µ mol , µ back , {q n , b n } n , w). Since the locations of inactive molecules are not associated with the observations in w and hyper-priors {q n } n , these conditionals simplify to p({x n , y n , z n } n:b n =0 |D, {b n } n ) which can be readily simulated jointly in the same manner as standard 3D Brownian motion.
## Sampling of the diffusion coefficient
Now that we have updated the locations of active and inactive molecules, we update the diffusion coefficient D by sampling from the corresponding conditional p(D|µ mol , µ back , {q n , b n , x n , y n , z n } n , w), which, due to the spe-cific dependencies of the variables in our formulation, e.g.,and Supplementary Equation 39, simplifies to p(D|{x n , y n , z n } n ). Because of conjugacy, the latter reduces to D|{x n , y n , z n } n ∼ InvGamma (α , β )
Supplementary Equation 73
where α and β are given by
[formula] α = α D + 3N (K − 1) 2 , β = β D + 1 4 N n=1 K−1 k=1 x n k+1 − x n k 2 + y n k+1 − y n k 2 + z n k+1 − z n k 2 t k+1 − t k Supplementary Equation 74 [/formula]
3.5. Sampling of the molecule prior weights and loads
We update the load prior weights q n by sampling from the corresponding conditional p({q n } n |D, µ mol , µ back , {b n , x n , y n , z n } n , w), which simplifies to p({q n } n |{b n } n ).
For this, we use Supplementary Equationand Supplementary Equation 32, and because of conjugacy, the latter distribution is sampled by sampling each q n separately according to
[formula] p(q n |b n ) ∝ p(b n |q n )p(q n ) = Beta q n ; α q N + b n , β q (N − 1) N + 1 − b n . [/formula]
## Supplementary equation 75
Once the weights q n are updated, we update the loads b n by sampling from the corresponding conditional p({b n } n |D, µ mol , µ back , {q n , x n , y n , z n } n , w). We perform this sampling using a Metropolis-Hastings update with proposals of the form
[formula] (b n ) prop ∼ Bernoulli(q n ). [/formula]
Supplementary Equation 76
In this case, by choosing the proposal distribution similar to the prior distribution, the acceptance ratio becomes
[formula] r b = K k=1 µ back + µ mol N n=1 (b n ) prop PSF (x n k , y n k , z n k ) µ back + µ mol N n=1 (b n ) old PSF (x n k , y n k , z n k ) w k × exp −(t k − t k−1 )µ mol N n=1 (b n ) old − (b n ) prop PSF (x n k , y n k , z n k ) Supplementary Equation 77 [/formula]
where (b n ) old denotes the existing sample.
## Joint sampling of the molecular brightness and background photon emission rates
Finally, after we updated the locations of molecules, and loads, we update the molecular brightness and background photon emission rates µ mol and µ back by sampling from the corresponding conditional p(µ mol , µ back |D, {q n , b n , x n , y n , z n } n , w), which simplifies to p(µ mol , µ back |{b n , x n , y n , z n } n , w). We carry over this sampling using a Metropolis-Hastings update where proposals for µ mol and µ back are computed according to where µ old mol and µ old back denote the existing samples. The acceptance ratio is
[formula] r µ = K k=1 µ prop back + µ prop mol N n=1 b n PSF (x n k , y n k , z n k ) µ old back + µ old mol N n=1 b n PSF (x n k , y n k , z n k ) w k × exp (t k − t k−1 ) µ old back − µ prop back + µ old mol − µ prop mol N n=1 b n PSF (x n k , y n k , z n k ) × µ old mol µ prop mol 2α prop mol −α mol exp µ old mol − µ prop mol β mol + α prop [/formula]
## Supplementary equation 79
We should emphasize, due to the weakness of the extended Kalman filter as compared to the unscented Kalman filter, we consider the background photon emission rate is fixed for EKF. So, in this case we update the molecular brightness µ mol only by sampling from the corresponding conditional p(µ mol |D, {q n , b n , x n , y n , z n } n , w), which simplifies to p(µ mol |{b n , x n , y n , z n } n , w). Again, we carry over this sampling using a Metropolis-Hastings update where proposals for µ mol are computed according to
[formula] µ prop mol ∼ Gamma α prop mol , µ old mol α prop mol [/formula]
## Supplementary equation 80
and the acceptance ration will reduces to
[formula] r µ = K k=1 µ back + µ prop mol N n=1 b n PSF (x n k , y n k , z n k ) µ back + µ old mol N n=1 b n PSF (x n k , y n k , z n k ) w k × exp (t k − t k−1 ) µ old mol − µ prop mol N n=1 b n PSF (x n k , y n k , z n k )+ 1 2 e − (x−µ) 2 2σ 2 √ 2πσ 2 0 µ 2 + σ 2 Bivariate Normal BNormal(µ, Σ) 1 2π √ |Σ| e − 1 2 (x−µ) T Σ −1 (x−µ) µ Σ Poisson Poisson(λ) λ x e −λ x! λ λ Gamma [/formula]
Gamma(α, β)
## Bernoulli
Bernoulli(q) (q − 1)δ0(x) + qδ1(x) q q(1 − q)
Here, the corresponding random variables are denoted by x.
Supplementary . Parameter values used in the generation of the synthetic traces PSF Lxy Lz ωxy ωz,zR N D µ mol µ back T total δt Units µm µm µm µm µm 2 s −1 photons s −1 photons s −1 s sUnits µm µm --µm 2 s −1 -photons s −1 --photons s −1 ---µm µm µm 2 µm 2 µm 2 s −1 photons
## Supplementary references
|
The Carboxy Terminus of YCF1 Contains a Motif Conserved throughout >500 Myr of Streptophyte Evolution
Plastids evolved from cyanobacteria by endosymbiosis. During the course of evolution, the coding capacity of plastid genomes shrinks due to gene loss or transfer to the nucleus. In the green lineage, however, there were apparent gene gains including that of ycf1. Although its function is still debated, YCF1 has proven to be a useful marker for plastid evolution. YCF1 sequence and predicted structural features unite the plastid genomes of land plants with those of their closest algal relatives, the higher streptophyte algae; YCF1 appears to have undergone pronounced changes during the course of streptophyte algal evolution. Using new data, we show that YCF1 underwent divergent evolution in the common ancestor of higher streptophyte algae and Klebsormidiophycae. This divergence resulted in the origin of an extreme, klebsormidiophycean-specific YCF1 and the higher streptophyte Ste-YCF1. Most importantly, our analysis uncovers a conserved carboxy-terminal sequence stretch within YCF1 that is unique to higher streptophytes and hints at an important, yet unexplored function.
# Introduction
Plastid genomes are demonstrably homologous to those of cyanobacteria, but have experienced drastic reductive evolution [bib_ref] Gene transfer to the nucleus and the evolution of chloroplasts, Martin [/bib_ref] [bib_ref] Chloroplast genomes of photosynthetic eukaryotes, Green [/bib_ref] [bib_ref] Genomic perspectives on the birth and spread of plastids, Archibald [/bib_ref]. Of the more than 4000 genes found in sections IV and V cyanobacteria, which have been suggested to be the closest living relatives to the ancestor of the plastid [bib_ref] Genomes of stigonematalean cyanobacteria (subsection V) and the evolution of oxygenic photosynthesis..., Dagan [/bib_ref] for an alternative view, see [bib_ref] An early-branching freshwater cyanobacterium at the origin of plastids, Ponce-Toledo [/bib_ref] , plastid genomes have retained only about 100-200 due to gene loss and endosymbiotic gene transfer [bib_ref] Endosymbiotic gene transfer: organelle genomes forge eukaryotic chromosomes, Timmis [/bib_ref]. A particularly appealing hypothesis for why some genes have been retained is that they are important for the in situ regulation of the photosynthesis redox reactions [bib_ref] Why chloroplasts and mitochondria retain their own genomes and genetic systems: colocation..., Allen [/bib_ref] , which also applies to the retention of genes by mitochondria for the regulation of their key bioenergetic reactions as part of the electron transport chain. Although the genomes of primary green plastids (Chloroplastida, cf. [bib_ref] The revised classification of eukaryotes, Adl [/bib_ref] tend to have fewer genes than those of red algae and glaucophytes [bib_ref] A structural phylogenetic map for chloroplast photosynthesis, Allen [/bib_ref] , there are some exceptions. One is Ycf1.
A few things are unusual about Ycf1. It has no detectable homolog in cyanobacteria, red algae or glaucophytes [bib_ref] YCF1: a green TIC?, De Vries [/bib_ref]. The origin of Ycf1 traces back to either a gain of the gene by the plastid genome, which appear to be extremely rare events, or a radical divergence upon the duplication of a plastid gene in the common ancestor of all Chloroplastida [bib_ref] The evolution of the plastid chromosome in land plants: gene content, gene..., Wicke [/bib_ref] [bib_ref] YCF1: a green TIC?, De Vries [/bib_ref] [bib_ref] The TIC complex uncovered: the alternative view on the molecular mechanism of..., Nakai [/bib_ref]. Regardless of its origin, ycf1 knockouts were shown to be lethal in tobacco [bib_ref] The two largest chloroplast genome-encoded open reading frames of higher plants are..., Drescher [/bib_ref] and Chlamydomonas [bib_ref] A large open reading frame (orf1995) in the chloroplast DNA of Chlamydomonas..., Boudreau [/bib_ref] , although the reason(s) for this lethality, or the function of the protein, were unknown at the time. Based on data gathered from Arabidopsis, it was later proposed that YCF1 is in fact TIC 214, a critical component of the protein translocation machinery at the inner chloroplast envelope [bib_ref] Translocon at the chloroplast inner envelope membrane, Kikuchi [/bib_ref]. If true, TIC214 would be the only TIC component encoded by the plastid genome. Moreover, Ycf1 has been lost in Poaceae and possibly a few other land plants as well [bib_ref] Phylogenetic affinities of the grasses to other monocots as revealed by molecular..., Katayama [/bib_ref] [bib_ref] YCF1: a green TIC?, De Vries [/bib_ref]. For these reasons, the possibility that YCF1 represents "a general TIC translocon" [bib_ref] Translocon at the chloroplast inner envelope membrane, Kikuchi [/bib_ref] has been challenged [bib_ref] YCF1: a green TIC?, De Vries [/bib_ref] , with additional criticism revolving around the extreme variation in YCF1 sequence length, which remains to be explained.
In the context of streptophyte terrestrialization, we recently discussed the evolution of the algal Cte-YCF1 to the land plant version of the protein, Ste-YCF1 [bib_ref] Streptophyte terrestrialization in light of plastid evolution, De Vries [/bib_ref] , along with the transformation of FTSH into . This analysis highlighted a clear distinction between the plastid genomes of the lower (Klebsormidiophyceae + Chlorokybophyceae + Mesostigmatophyceae = KCM) and higher branching streptophyte algae (Zygnematophyceae + Coleochaetophyceae + Charophyceae = ZCC). However, these inferences were hampered by the availability of only a single plastid genome from the lower branching Klebsormidiophyceae. Here we revisit the evolution of ycf1 and its implications for the origin of Ste-YCF1 in land plants, taking advantage of newly sequenced plastid genomes from streptophyte algae [bib_ref] Comparative chloroplast genome analyses of streptophyte green algae uncover major structural alterations..., Lemieux [/bib_ref]. Specifically, we have discovered a conserved motif, currently of unknown function, in the YCF proteins of some ZCC algae and embryophytes.
Klebsormidium spp. YCF1 Proteins Stand out YCF1 sequences are diverse in sequence and size. We first used a phylogenetic approach to gain a preliminary glimpse at YCF1 evolution across streptophyte algae. Though a divergent protein, our YCF1 phylogeny [fig_ref] FIG. 1: -Phylogeny of klesbormidiophycean YCF1 [/fig_ref] shows remarkable resemblance to the most recent species phylogeny based on the phylogenetic analysis of 88 plastid-encoded proteins [bib_ref] Comparative chloroplast genome analyses of streptophyte green algae uncover major structural alterations..., Lemieux [/bib_ref]. When comparing the YCF1 phylogeny to the species phylogeny [fig_ref] FIG. 1: -Phylogeny of klesbormidiophycean YCF1 [/fig_ref] , two YCF1 proteins do not branch as expected within streptophyte algae, both of which are from the genus Klebsormidium. This likely reflects the pronounced structural changes characteristic of the plastid genomes of Klebsormidiophyceae [bib_ref] Comparative chloroplast genome analyses of streptophyte green algae uncover major structural alterations..., Lemieux [/bib_ref]. This prompted us to take a closer look at the protein sequences of streptophyte algal YCF1s.
## Klebsormidiophyceaen ycf1 has undergone drastic sequence expansion
YCF1homologsexhibitawiderangeofsizes(deVriesetal.2015). Thereareneverthelesssomefeaturestheseproteinsallshare:they are predicted to contain six to eight transmembrane TM domains at their N-terminus and to encode charged repeats that cover much of the remaining sequence (cf. de [bib_ref] YCF1: a green TIC?, De Vries [/bib_ref] [bib_ref] YCF1: a Green TIC: response to the de Vries et al. Commentary, Nakai [/bib_ref]. We surveyed all 20 YCF1 proteins with regard to their length, charge, and number of predicted TM domains.
Streptophyte algal YCF1 proteins range between 410 aa (Mesostigma viride) and 4186 aa (K. flaccidum) [fig_ref] FIG. 1: -Phylogeny of klesbormidiophycean YCF1 [/fig_ref]. The largest portion of each YCF1 protein, except for the small YCF1s of the basal branching KCM streptophyte algae M. viride and Chlorokybus atmophyticus, is made up of the repetitively charged C-terminus. Noteworthy was a long negatively charged region from aa position 1777 to 1883 in the YCF1 of Mesotaenium endlicherianum, which is followed by similarly sized positively charged region spanning amino acid (aa) residues 1885-1943. In all cases, the YCF1 N-terminus was predicted to contain 6-7 TMs, usually located within the first 250 aa [fig_ref] FIG. 1: -Phylogeny of klesbormidiophycean YCF1 [/fig_ref]. The only exceptions are the zygnematalean YCF1s of M. endlicherianum and Cylindrocystis brebissonii, which were sometimes predicted to have a seventh TM starting around position 1250 and a seventh and eighth TM at around positions 650 and 700, respectively. The longest streptophyte YCF1s in our dataset (and to our knowledge the longest of any streptophyte) are those of the Klebsormidiophyceae. This, together with the results of the phylogenetic analysis [fig_ref] FIG. 1: -Phylogeny of klesbormidiophycean YCF1 [/fig_ref] , sets the klebsormidiophycean YCF1 clearly apart from both the Cte-YCF1 of the other KCM streptophyte and chlorophyte algae as well as the ZCC/embryophyte Ste-YCF1.
## A novel motif in ste-ycf1
Land plants and ZCC streptophyte algae are united by having a Ste-YCF1 type protein [bib_ref] YCF1: a green TIC?, De Vries [/bib_ref]. This separation was determined through reciprocal HMM detection, which uncovered Cte-YCF1 only when using Cte-YCF1 alignments as a seed and Ste-YCF1s only when using Ste-YCF1 alignments as a seed. Aligning all ZCC Ste-YCF1s with the KCM Cte-YCF1s reveals two regions with high sequence conservation: the first is the hydrophobic N-terminus with its TM domains [fig_ref] FIG. 1: -Phylogeny of klesbormidiophycean YCF1 [/fig_ref] and the second is at the very C-terminus [fig_ref] FIG. 1: -Phylogeny of klesbormidiophycean YCF1 [/fig_ref] and e). Closer inspection of the C-terminal region uncovered a conserved stretch of sequence that includes a 12 aa long motif, RLEDLACMNRFW; we suspect that the single conserved cysteine residue is critical for the function of the motif. We henceforth consider these 12 aa as the 'core motif'. This core motif is present in Zygnematophyceae (the closest algal relatives of embryophytes [bib_ref] Phylotranscriptomic analysis of the origin and early diversificationoflandplants, Wickett [/bib_ref] , Coleochaetophyceae and embryophytes. Directly upstream of this core motif there are eight more conserved residues; the full consensus of the Ste-YCF1 motif is thus IKRFLWPTxRLEDLACMNRFW [fig_ref] FIG. 1: -Phylogeny of klesbormidiophycean YCF1 [/fig_ref]. The only Charophycaeae sequence in this dataset, that of Chara vulgaris, bears a divergent, possibly rudimentary, form of the core motif including the cysteine and differing in four aa, RIEDLVCMERVW (bold = different aa). The fact that Chara possesses the most rudimentary form of the core motif is consistent with the basal branching of the Charophyceae within ZCC algae (cf. [bib_ref] Phylotranscriptomic analysis of the origin and early diversificationoflandplants, Wickett [/bib_ref] [bib_ref] The evolutionary origin of a terrestrial flora, Delwiche [/bib_ref].
To further assess the degree of conservation of the RLEDLACMNRFW motif among embryophytes, we expanded [bib_ref] Comparative chloroplast genome analyses of streptophyte green algae uncover major structural alterations..., Lemieux [/bib_ref]. Relationships of the chlorophyte N. olivaceae, 5 lower branching KCM and 13 higher branching ZCC streptophyte algae, and the land plant P. patens are shown. Klebsormidiophyceae are highlighted in purple. (c) YCF1 proteins are drawn to scale based on their length in aa. Color gradient from red (negative) to white (none) to green (positive) indicate the predicted charge based on the EMBOSS explorer (sliding window size: 5 aa). Boxes on the proteins indicate predicted TM domains where, applying a majority rule, some consensus was found based on TMHMM, Phobius and SOSUI. The frequency with which the TMs were predicted by the different programs was evaluated using CCTOP and is depicted as black (predicted by all algorithms) to white (predicted by few algorithms) coloration of the boxes. Note the length of klebsormidiophycean YCF1s (purple font) and the position of the 200 bp intron removed prior to in silico translation of M. endlicherianum's YCF1 (arrowhead). (d) Depiction of the YCF1 protein sequence alignment of the 20 YCF1s used in (a). The topmost line depicts the full 5015 positions of the alignment (of which 311 aligned residues were used for the phylogeny shown in [a]), black boxes indicate regions where based on a majority rule some consensus was found for at least half of proteins under consideration. The histograms below depict high (red) to low (blue) hydrophobicity (top) and 30-70% (yellow) to < 30% (red) sequence identity (bottom). A zoom-in into the very C-terminal region of the alignment shows that Ste-YCF1 of higher ZCC streptophyte algae shares the core aa-sequence "RLEDLACMNRFW" with land plants. Note the conserved hydrophobic block of N-terminal aa and the conserved Cterminal end. The removal of the intron in the ycf1 ORF of Mesotaenium restores a YCF1 that carries a RxEDLACMNRFW motif. (e) An alignment of 21 land plant YCF1s (3 bryophytes, 3 lycophytes, 6 monilophytes, 3 gymnosperms, Gingko, and 5 angiosperms) demonstrates that RLEDLACMNRFW is conserved throughout all embryophyte Ste-YCF1s. Note the putative RNA editing sites (marked with dotted boxes filled in the color corresponding to the aa after RNA editing) in the S. moellendorffii motif sequence (based on data from S. unicata [cf. [bib_ref] Chloroplast RNA editing going extreme: more than 3400 events of C-to-U editing..., Oldenkott [/bib_ref]. the alignment to include sequences from across the diversity of land plants. We found that the core motif RLEDLACMNRFW is conserved from moss to Arabidopsis thaliana [fig_ref] FIG. 1: -Phylogeny of klesbormidiophycean YCF1 [/fig_ref]. Moreover, land plants have an extended conserved C-terminal region of 39 aa [fig_ref] FIG. 1: -Phylogeny of klesbormidiophycean YCF1 [/fig_ref]. The only land plant sequence showing severe alterations from this core motif (including the cysteine residue) was that of Selaginella moellendorffii. S. moellendorffii is known for its unusual organellar biology characterized by massive RNA editing [bib_ref] Extreme RNA editing in coding islands and abundant microsatellites in repeat sequences..., Hecht [/bib_ref]. In its close relative S. uncinata, RNA editing of the Ycf1 open reading frame (ORF) converts the codons from encoding RPEDPSRMNRPR into encoding RLEDLSCMNRFW [bib_ref] Chloroplast RNA editing going extreme: more than 3400 events of C-to-U editing..., Oldenkott [/bib_ref] ; [fig_ref] FIG. 1: -Phylogeny of klesbormidiophycean YCF1 [/fig_ref]. Ergo, RLEDLACMNRFW is a motif within YCF1 that Arabidopsisshareswith thecommon ancestorofallland plants.
To determine whether this motif has been detected before, we used the 12, 21, and 39 aa motif versions as queries for InterPro scans. The former two did not return any hits. Using the 39 aa motif as a query, InterPro again predicted no family memberships, domains or functions, but detected an unintegrated (i.e. not yet curated) signature match against the unnamed PANTHER protein family PTHR33163 (cf. [bib_ref] PANTHER version 11: expanded annotation data from Gene Ontology and Reactome pathways,..., Mi [/bib_ref] , that consists of 74 land plant YCF1s and does not list the 12, 21, or 39 aa motifs as a conserved feature. None of the motif versions can be found in PFAM (cf. [bib_ref] The Pfam protein families database: towards a more sustainable future, Finn [/bib_ref] and motif searches of the PDB database (cf. [bib_ref] The Protein Data Bank, Berman [/bib_ref] returned no hits. Using the 12, 21, and 39 aa versions of the motif as queries for a BLASTp against the RefSeq database (applying a very low stringency e value cutoff of <1), we detected only streptophyte YCF1 sequences (see supplementary table S1, Supplementary Material online). Hence, the aa sequence RLEDLACMNRFW (and extended versions) indeed represents a novel motif restricted to higher streptophyte YCF1s.
To confirm these BLAST-based results, we used all motif versions as queries in pHMMER searches [bib_ref] HMMER web server: 2015 update, Finn [/bib_ref] against the UniProt database (The UniProt Consortium 2015). The 12 aa motif did not return any significant hits. The 21 aa and 39 aa motifs returned 1788 and 2139 significant hits, respectively. All of these hits, except one, were streptophyte sequences that encompass land plant and 12 ZCC streptophyte algal Ste-YCF1s. The exception was the nuclear-encoded FAR-RED IMPAIRED RESPONSE 1 (FAR1) protein of Medicago truncatula (protein id: MTR_2g039370), containing a 39 aa long C-terminal stretch (FKLFLWPNYRLEDLACINRYWFNTHNGSHFSILRIHMYP [bold = identical residue]) that resembled the 39 aa version of the Ste-YCF1 motif. FAR1 proteins are transcription factors [bib_ref] The FAR1 locus encodes a novel nuclear protein specific to phytochrome A..., Hudson [/bib_ref] [bib_ref] Transposase-derived transcription factors regulate light signaling in Arabidopsis, Lin [/bib_ref] with the ability to bind DNA [bib_ref] Transposase-derived transcription factors regulate light signaling in Arabidopsis, Lin [/bib_ref]. This warrants further investigation, as nucleic acid binding capacities have been described for Chlamydomonas YCF1 [bib_ref] A large open reading frame (orf1995) in the chloroplast DNA of Chlamydomonas..., Boudreau [/bib_ref]. An involvement of the novel Ste-YCF1 motif in DNA/RNA-binding or even transcriptional regulation is hence conceivable (see below).
## Ycf1 sequence divergence and the first filamentous streptophyte algae
What does this mean for the trajectory of YCF1 evolution and ultimately its function? Cte-YCF1 most likely evolved in the common ancestor of all Chloroplastida [bib_ref] The evolution of the plastid chromosome in land plants: gene content, gene..., Wicke [/bib_ref] [bib_ref] YCF1: a green TIC?, De Vries [/bib_ref]. After the split of streptophytes and chlorophytes (cf. [bib_ref] Green algae and the origin of land plants, Lewis [/bib_ref] [bib_ref] Streptophyte algae and the origin of embryophytes, Becker [/bib_ref] , the basal branching streptophyte algae Mesostigmatophyceae and Chlorokybophyceae retained the ancestral Cte-YCF1, sharing it hence with the chlorophytes . In the common ancestor of Klebsormidiophyceae and higher branching ZCC streptophyte algae, the Cte-YCF1 evolved into the highly expanded version found in Klebsormidiophyceae (>3000 aa length) and the Ste-YCF1 (bearing the RLEDLACMNRFW motif). Ste-YCF1 was -YCF1 sequence divergence during streptophyte algal evolution. A cladogram of the evolution of Chloroplastida (thick lines; chlorophytes = yellow, streptophytes = different shades of green) with the trajectory of YCF1 evolution projected on top. The ancestral Cte-YCF1 (blue) is shared by chlorophytes and the KCM streptophyte algae of the Mesostigmatophyceae and Chlorokybophceae. Along the trajectory of streptophyte evolution that includes the first filamentous streptophyte (one of the common ancestors of recent Klebsormidiophyceae and ZCC streptophyte algae + land plants) the Cte-YCF1 sequence started to change (dark purple). From there, YCF1 diverged into the klebsormidiophycean YCF1 (red) and the Ste-YCF1 (purple). The latter is found among ZCC streptophyte algae and land plants. inherited by land plants. Therefore, the changes in YCF1 sequence emerged along the trajectory that includes the first filamentous streptophyte .
We previously hypothesized that changes in YCF1 occurred concomitantly with changes in . Although FtsH is a plastid-encoded component of the photosystem II maintenance machinery [bib_ref] Is ftsH the key to plastid longevity in sacoglossan slugs?, De Vries [/bib_ref] , YCF2's function remains unknown [bib_ref] The two largest chloroplast genome-encoded open reading frames of higher plants are..., Drescher [/bib_ref]. So, what about klebsormidiophycean FtsH/YCF2? Entransia is the only member of the Klebsormidiophyceae with a sequenced plastid genome that encodes the ftsH gene (cf. [bib_ref] Comparative chloroplast genome analyses of streptophyte green algae uncover major structural alterations..., Lemieux [/bib_ref]. Klebsormidium spp. has probably lost ycf2/ftsH. Civá ň and colleagues used a BLASTp approach to distinguish between higher streptophyte YCF2 and FtsH (among Chloroplastida, the latter can be found in Mesostigma, Chlorokybus and chlorophytes). Querying Entransia "FtsH" via BLASTp against the non-redundant protein database of NCBI (NCBI Resource Coordinators 2016) returned the YCF2 sequence of the conifer Pseudolarix amabilis as the best hit (YP_009268451.1; e value of 4 Â 10 À18 , but with a query coverage of only 22%), followed by many other land plant YCF2s with equally low query coverage. In accordance with the approach taken by Civan and colleagues, Entransia FtsH can hence be classified as YCF2. It is noteworthy that the closest streptophyte algal hit, when using the Entransia FtsH/YCF2 as query was found to be Closterium's YCF2 (YP_009256827.1, e value of 2 Â 10 À14 , 27% query coverage and the 61st hit). We interpret this as klebsormidiophycean YCF2 being distinct from its orthologs in other streptophyte algae and land plants; this is reminiscent of the pattern we observed for klebsormidiophycean YCF1. Therefore, our new data further supports the notion that YCF1 and YCF2 sequence divergence occurred contemporaneously. The underlying reason(s) for this parallel divergence is as elusive as the functions of the proteins themselves.
In 1997, Boudreau and colleagues found that Chlamydomonas Cte-YCF1 was able to bind nucleic acids and suggested that it could bind plastid DNA. If so, YCF1 might be important, for example, during plastid division and in the context of a filamentous and, in the case of many ZCC algae and land plants, polyplastidic species. Alternatively, YCF1 might bind specific RNAs instead of DNA, which would point to a regulatory function at a certain stage of the plastid, perhaps regulating a stage transition associated with protein import. This regulatory role might be linked to the recent finding that exogenous treatment of the ZCC alga Spirogyra pratensis with the phytohormone ethylene resulted in a downregulation of the gene expression of Ste-YCF1, concomitant with elevated stress response signaling as well as decreased gene expression of photosynthesis-associated and chlorophyll biosynthesis-associated proteins [bib_ref] Transcriptome profiling of the green alga Spirogyra pratensis (Charophyta) suggests an ancestral..., Van De Poel [/bib_ref]. In summary, the massive divergence in size, the conserved motif in Ste-YCF1, and the loss of the protein in grasses speaks in favor of a function that reaches beyond that of just a TIC component.
# Conclusion
YCF1 remains an enigmatic protein. However, the broader the sampling of sequence data across the diversity of streptophytes becomes, the less puzzling the sequence diversity of YCF1 appears. In this study, we have shown that 1) klebsormidiophycean YCF1 is unique regarding its size expansion and 2) that Ste-YCF1s are united by a novel motif that is RLEDLACMNRFW. The function(s) of this motif remains to be explored, but it is likely significant considering its strict conservation across more than 500 Myr of evolution, which separates Coleochaete and Arabidopsis (cf. [bib_ref] Estimating the timing of early eukaryotic diversification with multigene molecular clocks, Parfrey [/bib_ref].
# Materials and methods
To generate the YCF1 dataset, we sampled annotated ORFs of ycf1 genes from 18 available streptophyte plastid genomes [bib_ref] Ancestral chloroplast genome in Mesostigma viride reveals an early branch of green..., Lemieux [/bib_ref] [bib_ref] A clade uniting the green algae Mesostigma viride and Chlorokybus atmophyticus represents..., Lemieux [/bib_ref] [bib_ref] Comparative chloroplast genome analyses of streptophyte green algae uncover major structural alterations..., Lemieux [/bib_ref] [bib_ref] The chloroplast and mitochondrial genome sequences of the charophyte Chaetosphaeridium globosum: insights..., Turmel [/bib_ref] [bib_ref] The complete chloroplast DNA sequences of the charophyceaen green algae Staurastrum and..., Turmel [/bib_ref] [bib_ref] The chloroplast genome sequence of Chara vulgaris sheds light into the closest..., Turmel [/bib_ref] [bib_ref] Analyses of charophyte chloroplast genomes help characterize the ancestral chloroplast genome of..., Civá Ň [/bib_ref] as well as from the moss Physcomitrella patens [bib_ref] Complete chloroplast DNA sequence of the moss Physcomitrella patens: evidence for the..., Sugiura [/bib_ref] and the chlorophyte alga Nephroselmis olivacea [bib_ref] The complete chloroplast DNA sequence of the green alga Nephroselmis olivaceae: insights..., Turmel [/bib_ref] , and translated them into protein sequences using Geneious R8.1.8 [bib_ref] Geneious basic: an integrated and extendable desktop software platform for the organization..., Kearse [/bib_ref]. For M. endlicherianum, we removed the intron suggested by Civá ň et al. (2014) prior to translation.
Using MAFFT v.7 (Katoh and Standley 2013; G -INS -I), we generated an alignment of all 20 YCF1 proteins and used MEGA7 [bib_ref] MEGA7: Molecular Evolutionary Genetics Analysis version 7.0 for bigger datasets, Kumar [/bib_ref] to compute a maximum likelihood phylogeny applying the LG + G + I + F model [bib_ref] An improved general amino acid replacement matrix, Le [/bib_ref] ; 500 bootstraps, partial deletion [95%] leaving 311 positions, gamma category 5). The 21 embryophyte Ste-YCFs were aligned using the settings L-INS-I in MAFFT v.7. YCF1 charge was predicted using the EMBOSS explorer [bib_ref] EMBOSS: the European Molecular Biology Open Software Suite, Rice [/bib_ref]. Predictions of TM domains were generated using TMHMM [bib_ref] Predicting transmembrane protein topology with a hidden markov model: application to complete..., Krogh [/bib_ref] , SOSUI [bib_ref] SOSUI: classification and secondary structure prediction system for membrane proteins, Hirokawa [/bib_ref]. To build a consensus between all TM predictions we used CCTOP [bib_ref] CCTOP: a Consensus Constrained TOPology prediction web server, Dobson [/bib_ref].
# Supplementary material
Supplementary data is available at Genome Biology and Evolution online.
[fig] FIG. 1: -Phylogeny of klesbormidiophycean YCF1. (a) Maximum likelihood LG + G + I + F phylogeny (500 bootstraps, values shown at each node, partial deletion [95%], gamma category 5) of YCF1 protein sequences from the same species as in (b). NCBI accession numbers of the YCF1 proteins are shown behind the species name. The tree is drawn to scale based on the substitution rate. Note the phylogenetic position of Klebsormidium spp. YCF1. (b) A reference cladogram of species phylogenetic relationships based on [/fig]
|
Predicting Diabetic Neuropathy Risk Level Using Artificial Neural Network and Clinical Parameters of Subjects With Diabetes
Background: A risk assessment tool has been developed for automated estimation of level of neuropathy based on the clinical characteristics of patients. The smart tool is based on risk factors for diabetic neuropathy, which utilizes vibration perception threshold (VPT) and a set of clinical variables as potential predictors. Methods: Significant risk factors included age, height, weight, urine albumin-to-creatinine ratio, glycated hemoglobin, total cholesterol, and duration of diabetes. The continuous-scale VPT was recorded using a neurothesiometer and classified into three categories based on the clinical thresholds in volts (V): low risk (0-20.99 V), medium risk (21-30.99 V), and high risk (≥31 V). Results: The initial study had shown that by just using patient data (n = 5088) an accuracy of 54% was achievable. Having established the effectiveness of the "classical" method, a special Neural Network based on a Proportional Odds Model was developed, which provided the highest level of prediction accuracy (>70%) using the simulated patient data (n = 4158). Conclusion: In the absence of any assessment devices or trained personnel, it is possible to establish with reasonable accuracy a diagnosis of diabetic neuropathy by means of the clinical parameters of the patient alone.
# Introduction
There are approximately 4.7 million people in the United Kingdom living with diabetes and neuropathy affects up to 50% of these patients. [bib_ref] Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes..., Stratton [/bib_ref] Diabetic neuropathy (DN) may be completely asymptomatic, but can cause numbness, tingling, or even painful burning sensations in the lower limbs, rarely affecting the upper limbs. [bib_ref] Advances in the epidemiology, pathogenesis and management of diabetic peripheral neuropathy, Tesfaye [/bib_ref] As neuropathy is a very insidious complication it may remain undiagnosed for several years, so routine regular screening is vital to identify the "at risk" foot. [bib_ref] Foot ulceration and lower limb amputation in type 2 diabetic patients in..., Muller [/bib_ref] Resultant foot ulceration and limb amputation are very serious complications of neuropathy with serious consequences on patients' quality of life and survival.
Commonly used neuropathy screening tools utilize pressure/touch sensation such as the 10 g Semmes Weinstein monofilament, and vibration perception (eg,125 Hz tuning fork, biothesiometer and neurothesiometer). Devices that use vibration perception are generally considered as the gold standard for neuropathy assessment as they can quantitatively predict the onset and progression of the complication. [bib_ref] The natural history of diabetic peripheral neuropathy determined by a 12 year..., Coppini [/bib_ref] The 128 Hz tuning fork was the first tool to use vibration perception but its application was very limited by both observation technique and patient response. [bib_ref] Preliminary data on VibraTip®, a new source of standardised vibration for bedside..., Levy [/bib_ref] The development of the Biothesiometer was based on mechanical and more standardized vibration perception, and was subsequently replaced by the neurothesiometer, which operates on the same principle but is battery-operated. The neurothesiometer produces mechanical vibration with a fixed frequency of approximately 100 Hz while the vibration amplitude is controlled manually using a rotatory control knob. The knob is used to adjust the voltage applied and ranges from 0 to 50 V (0-250 µm in amplitude). The operator applies the handheld probe to the pulp of the great toe and the vibration stimulus gradually increased, until the subject feels the vibration sensation. The voltage displayed on the neurothesiometer is the measured vibration perception threshold (VPT). The major drawback of such a device is its manual observer-dependent operability and its limited vibration intensity. [bib_ref] Vibration perception threshold testing in patients with diabetic neuropathy: ceiling effects and..., Deursen [/bib_ref] With the worldwide increasing prevalence of diabetes, there is a need for more intuitive, operator-independent and smarter diagnostic devices. In the absence of assessment devices or trained personnel, it may be difficult to establish the accurate level of neuropathy, bearing in mind that the clinical parameters of the patient may also play a key role in diabetes. This motivated us to develop an intelligent risk assessment tool based on patient data that can predict the risk level of DN. The software uses patient data and is shown to provide an acceptable level of accuracy, which will continue to improve in performance as quality data are collected over time. As the software is likely to perform better if trained on a larger dataset, we have interfaced it with a newly developed device (VibraScan), [bib_ref] A Smart Device to Replace Neurothesiometer for Measuring Diabetic Vibration Perception Threshold, Dave [/bib_ref] which measures the subject's VPT. With the larger integrated dataset this device can be used as a comprehensive diagnostic tool for DN, while taking into account changing patient parameters.
# Methods
VPTs and various clinical measurements were used to develop a tool for automated prediction of neuropathy. VPT measurements (neurothesiometer) were used to identify the level of DN and were correlated with potential predictors obtained from a clinical diabetes database (n = 5088). Potential predictors included duration of diabetes mellitus, age, height, weight, body mass index, urinary albumin-to-creatinine ratio (ACR), blood glucose, glycated hemoglobin (HbA1c), total cholesterol, and triglyceride. These variables represented the average of a small number of measurements taken at clinic visits in a hospital setting over a 28-year period, and measured at least once a year, and if measured more frequently, averaged over a year. These data were collected repeatedly over this time period for the same subjects. The data were not specifically collected for this study, so no specific protocol was followed and we did not differentiate on the basis of demographic features. There was the possibility of unintended missing data or loss to follow-up, both of which can commonly occur in longitudinal trials. In order to provide data with properties similar to the patient data, 4158 cases were simulated with patients' mean and covariance from the data to improve precision. A Neural Network based on a Proportional Odds Model (NNPOM) was trained using the simulated patient dataset where VPT (volts) was encoded into three categories-low, medium, and high-risk level of neuropathy.
For data classification many conventional statistical methods can be used, but for developing a risk assessment tool a classifier that can handle the data precisely should be used. In a study on 110 patients with diabetes, the Michigan Neuropathic Diabetic Score was used to differentiate normal and abnormal cases. In this study nerve conduction studies were used to assess DN and showed that age, duration of disease, gender, and quality of diabetes control all have a significant relationship with DN, but no correlation was found with hyperlipidemia, blood pressure, or smoking. [bib_ref] Potential risk factors for diabetic neuropathy: a case control study, Booya [/bib_ref] In another study, assessment of DN was performed using the Michigan Neuropathy Screening Instrument questionnairebased examination. This focused on the relationship between risk factors and the prevalence of DN in youths. [bib_ref] Prevalence of and risk factors for diabetic peripheral neuropathy in youth with..., Jaiswal [/bib_ref] Another study was performed to identify risk factors associated with DN by comparing the prevalence of neuropathy in subjects with known diabetes mellitus and those with newonset diabetes. In this study the 10 g monofilament test, pinprick, and VPT were used to categorize patients into normal, and into mild, moderate, and severe neuropathy. A total of 586 patients with established DN were identified. Regression analysis was used to identify the risk factors associated with DN. It was found that age, dyslipidemia, alcohol status, and macro-and microvascular complications were significant risk factors for DN. [bib_ref] Prevalence and risk factors of development of peripheral diabetic neuropathy in type..., Bansal [/bib_ref] For optimal contribution of each variable the right type of transformation should be considered, and we used Box-Cox family [bib_ref] Box-Cox regression models, Drukker [/bib_ref] [bib_ref] An analysis of transformations, Box [/bib_ref] transformations to identify significant variables. The analysis result [fig_ref] Table 1: Significant Variables for DN Prediction [/fig_ref] shows that out of the 13 commonly used variables for assessment of DN, only 7 are found to be statistically significant predictors (P < .05). These are duration of diabetes, age, height, weight, glycemic control (HbA1c), ACR, and cholesterol. These can be considered as the risk factors for development of the DN tool.
These variables can be used to obtain the point prediction of VPT or the predicted VPT within the confidence interval [fig_ref] Table 1: Significant Variables for DN Prediction [/fig_ref]. In order to simplify the interpretation of the outcome variable (VPT) for both clinicians and patients, it was useful to interpret the VPT prediction in terms of cumulative risk levels. For solving this problem in terms of classification, the first step was to divide the data into specific categories based on VPT thresholds. By considering the clinician's expertise in the field of DN, the dataset was categorized into three classes based on VPT measurement-low risk (0-20.99 V), medium risk (21-30.99 V), and high risk (≥31 V).
The structure of the neural network classifier is shown in and uses a feedforward learning algorithm. It is a three-layer feed-forward neural network, which consists of the input layer, hidden layers, and output layers. These layers have 13 nodes in the input layer (as per input variables), a number of nodes in the hidden layer, and 3 nodes in the output layer (required classification).
In order to train the network, various training algorithms are available, of which resilient backpropagation algorithms and scaled conjugate gradient algorithms are used for fast pattern classification of artificial neural network (ANN). [bib_ref] Comparison of neural network training functions for hematoma classification in brain CT..., Sharma [/bib_ref] At the output node, the softmax activation function is used that yields the probability values of the classification provided by the nodes. The softmax usually takes the un-normalized vector and normalizes in terms of probability distribution over predicted output classes. Here x is the input, which incorporates the value of all the 13 attributes , , . Each y i represents the output values belonging to class i. The softmax function is applied to output probabilities of multiple output classes and can be given as the input x multiplied by weights W and added to bias b to generate the predicted output y i as in Eq. (1).
[formula] y softmax W x b i i j i i j = + ∑ ( ) .(1) [/formula]
The softmax function normalizes the three-dimensional vector into the range of [bib_ref] Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes..., Stratton [/bib_ref] and y i is the predicted probability as the risk factor. In order to evaluate how well the classifier is working, the loss function is used to measure the inconsistency between the predicted output and actual output. The loss function is known as the performance goal. Here the cross-entropy loss function is given by Eq.
(2) to get the classification success.
crossentropy label y label y
[formula] i i i , ( )= − ( ) ∑ log(2) [/formula]
The neural network implementation was achieved using pattern classification feature of neural network toolbox on MATLAB platform 16 with the following steps:
1. Preprocess the medical data to achieve the class balance before using a neural network for classification.
# Results
To the best of our knowledge, there are currently no available tools that can determine DN risk by analyzing patients' clinical data. The focus here was to develop a risk assessment tool to help clinicians analyze a complex relationship between risk factors and the level of DN. After obtaining the results of the classical statistical analysis, which was based on summarized patient data, an accuracy of 54% was achieved. [bib_ref] Comprehensive risk assessment of diabetic neuropathy using patient data, Dave [/bib_ref] However, after looking into the complexity and nonlinearity of the data, the idea of using an ANN as classifier was considered. A simulation was run on the neural network with 13 attributes and some of the basic hyper-parameters were kept fixed. The confusion matrix of the overall data set (consisting of testing, validation, training sets) is shown in [fig_ref] Table 2: Confusion Matrix with a Dataset of 13 Attributes [/fig_ref]. The confusion matrix is the summary of prediction results on classification problems and presents the accuracy of the classifier. It gives correct classification as "true positive" or "true negative" as shown in green-colored boxes (diagonal of the square) and incorrect classification is listed by "false positive" and "false negative" as shown in red-colored boxes (adjacent to the diagonal). The numbers in the boxes are correctly classified predictions out of 5088 instances.
When the neural network was trained using seven attributes (the significant predictors) there was a slight decrease in classification from 67.9% to 67.4%; so the dataset with significant predictors is used for further analysis. Note that all 13 variables in were merely to show that all the commonly used factors for DN were considered but the simulation was done with only 7 variables in order to reduce the computational load while still achieving the equivalent results.
In order to improve accuracy, the data need to be either collected precisely or processed in a way to reduce noise. Instead of choosing the prolonged process of collecting patient data again, the second option of increasing predictor precision was to generate a simulated patient dataset. For generating simulated patient data with similar properties, the real patient data mean and covariance of each predictor were used with reduced standard error in order to increase the precision of data. By this method a larger dataset was obtained for improving prediction. The simulated data were generated using the R-code of simstudy software with genData function.Using this function, it generates multivariate normal data. It requires the mean, standard error, correlation matrix, or correlation coefficient of real patient data. If data are generated by using the same standard error of the real patient data then the spread of the data may remain the same for each predictor; therefore, standard error was reduced for independent variables. By comparing the summary of the simulated data with the real patient data it was possible to reduce the skewness of the simulated data and the accuracy was significantly increased by reducing the standard error of each predictor.
Many available classifiers can predict the numerical values from labeled patterns but less consideration has been given to ordinal classification problems where labels of the dependent variables or targets have a natural ordering. In the current scenario, labels are ordered as low, medium, and high risk. While dealing with the problems of misclassification it would be more erroneous if a patient of low risk was classified as high risk rather than as medium risk. This encouraged us to use an ordinal classification model to carefully handle ordinal labels. [bib_ref] Ordinal regression methods: survey and experimental study, Gutierrez [/bib_ref] However, in order to adopt a probabilistic framework, NNPOM was considered to be more useful. [bib_ref] Ordinal models for neural networks, Mathieson [/bib_ref] [bib_ref] Ordinal regression neural networks based on concentric hyperspheres, Gutiérrez [/bib_ref] The model approaches ordinal classification by estimating the latent variable belonging to ordinal categories and is seen to perform well when classes are defined from a discretized variable. [bib_ref] Ordinal regression methods: survey and experimental study, Gutierrez [/bib_ref] This model is adjusted such that it updates the weights by minimizing the cross-entropy loss in each iteration. NNPOM is a linear combination of nonlinear basis functions, which can be adjusted by three hyper-parameters as hidden neurons (M), number of Iterations (iter), and the value of regularization parameter (λ). [bib_ref] A mixture of experts model for predicting persistent weather patterns, Pérez-Ortiz [/bib_ref] The training of NNPOM was performed on MATLAB framework.The model was trained with prior setting of all the hyper-parameters and the range was explored using different numbers of hidden neurons M ϵ {10, 40, 60, 80, 85, 120, 130}, iter ϵ {1000, 1500} and λ ϵ {0.01, 0.001}. The accuracy of the model is calculated based on the percentage of correctly classified class. By increasing the hidden units, training time increases and by changing these hyper-parameters, the results change from 68.25% to 70.1%. However, by selecting M = 120, iter = 1500, and λ = 0.01, the highest level of correctly classified output obtained was 70.1%. This version of trained model with highest accuracy was selected for the risk assessment tool.
To make it user-friendly, an application was developed as shown in the example illustrated in [fig_ref] Figure 2: Risk assessment tool with clinical variables [/fig_ref] using the trained neural network model. Based on the input values provided to the tool, VPT probabilities were predicted from the learnt model with 2% chance of low risk, 23% chance of medium risk, and 75% chance of high risk. The result shows that this subject is very likely to have DN.
Another example was tested by keeping all the parameters the same and just decreasing the age by 10 years to determine the categories and observe the change in the risk level. As seen in [fig_ref] Figure 3: Diabetic neuropathy prediction with changed variables [/fig_ref] , by keeping all the other inputs (apart from age) the same, the risk level of DN is reduced from high to medium risk. Thus, based on a combination of clinical inputs, the tool can predict the subject's VPT and determine the risk level of DN.
# Discussion
This risk assessment tool is the first of its kind to provide a novel method for the reliable screening, diagnosis, and monitoring of DN. The study was conducted by using both the longitudinal data and continuous VPT measurements. Out of 13 variables, 7 variables were identified as significant predictors. For easy interpretation of the risk level of DN, VPT was divided into three categories based on VPT thresholds. We initially considered all 13 variables for training with these attributes, and the model achieved an accuracy of 67.9%. The ANN was then trained with seven best predictors and achieved an accuracy of 67.4%. This comparison gave us a good indication to use only significant predictors rather than all the attributes.
In order to improve the precision of summarized patient data (which provided only 54% accuracy due to missing data), data with similar properties were simulated with covariance and the mean value of patient data, but with reduced standard errors. The NNPOM was trained on simulated data and achieved an accuracy of 70.1% with the seven predictors. Considering the noise and imbalance in data this method worked significantly well for a risk assessment tool for DN. The limitation of quality of data was handled by using simulated patient data and the quality of data was improved by increasing the precision of variables. As a result the achieved accuracy can be further improved by collecting clinically significant patient data over time. We recognize that performance may not improve by leveraging larger datasets and may hit the ceiling; however, it is imperative that if more quality data are collected over time, the performance may increase for a particular country or population. Since the current level of accuracy is based on the summarized or simulated patient data, we found that the results were good enough when compared with the validation data. We intend to evaluate this further and this paper highlights an important step in that direction.
# Conclusion
In the absence of established assessment devices or trained personnel it may prove difficult to get a diagnosis of DN. The undoubted importance of patient clinical characteristics led us to develop a risk assessment tool based on patient data alone. This intelligent software-based tool is both userfriendly and can provide the risk level of DN with reliable accuracy. As the software would perform better if trained on a larger dataset, its performance will continually improve as quality data are collected over time. We, therefore, interfaced this software with our neuropathy device (VibraScan 23 ) in order to enrich quality patient data. The resultant larger integrated dataset enables the device to be used as a comprehensive tool for the risk assessment of DN.
Consensus for screening for DN has been historically very difficult to standardize and lags behind the much more state-of-the-art digital screening for retinopathy for example. As this intuitive artificial intelligence device aptly offers a low or noncontact patient facility, it provides a platform for the safe, quick, and effective screening of the diabetes population at large. [bib_ref] Diabetic neuropathy: are we still barking up the wrong tree and is..., Coppini [/bib_ref] With a current global prevalence of diabetes of 9.3% (just under half a billion people), and an estimation to rise by 50% in 2045, this innovative technique certainly opens a new horizon in modern diabetes care.Abbreviations ACR, Urine albumin to creatinine ratio; AI, artificial intelligence; ANN, artificial neural network; BMI, body mass index; HbA1C, glycated haemoglobin; HDL, high-density lipoprotein; LB, lower bound; LDL, low-density lipoprotein; RBG, random blood glucose; UB, upper bound.
[fig] Figure 2: Risk assessment tool with clinical variables.HbA1c, glycated hemoglobin. [/fig]
[fig] Figure 3: Diabetic neuropathy prediction with changed variables.HbA1c, glycated hemoglobin. [/fig]
[table] Table 1: Significant Variables for DN Prediction. Repeat steps 2-8 to train the neural network using seven best predictors. 10. Compare the accuracies of the dataset with a different set of attributes. [/table]
[table] Table 2: Confusion Matrix with a Dataset of 13 Attributes. [/table]
|
The HLA-B*35 allele modulates ER stress, inflammation and proliferation in PBMCs from Limited Cutaneous Systemic Sclerosis patients
Introduction: HLA-B*35 is associated with increased risk of developing pulmonary hypertension in SSc patients. We previously reported that HLA-B*35 induces endothelial cell dysfunction via activation of ER stress/UPR and upregulation of the inflammatory response. Because PBMCs from lcSSc-PAH patients are also characterized by activation of ER stress/UPR and inflammation, the goal of this study was to assess whether the presence of HLA-B*35 contributes to those characteristics.Methods: PBMCs were purified from healthy controls (n = 49 HC) and lcSSc patients, (n = 44 with PAH, n = 53 without PAH). PBMCs from each group were stratified for the presence of HLA-B*35. Global changes in gene expression in response to HLA-B*35, HLA-B*8 or empty lentivirus were investigated by microarray analysis in HC PBMCs. Total RNA was extracted and qPCR was performed to measure gene expression.Results: ER stress markers, in particular the chaperones BiP and DNAJB1 were significantly elevated in PBMC samples carrying the HLA-B*35 allele. IL-6 expression was also significantly increased in HLA-B*35 lcSSc PBMCs and positively correlated with ER stress markers. Likewise, HMGB1 was increased in HLA-B*35-positive lcSSc PBMCs. Global gene expression analysis was used to further probe the role of HLA-B*35. Among genes downregulated by HLA-B*35 lentivirus were genes related to complement (C1QB, C1QC), cell cycle (CDNK1A) and apoptosis (Bax, Gadd45). Interestingly, complement genes (C1QC and C1QB) showed elevated expression in lcSSc without PAH, but were expressed at the low levels in lcSSc-PAH. The presence of HLA-B*35 correlated with the decreased expression of the complement genes. Furthermore, HLA-B*35 correlated with decreased expression of cyclin inhibitors (p21, p57) and pro-apoptotic genes (Bax, Gadd45) in lcSSc B35 subjects. FYN, a tyrosine kinase involved in proliferation of immune cells, was among the genes that were positively regulated by HLA-B*35. HLA-B*35 correlated with increased levels of FYN in lcSSc PBMCs.Conclusions: Our study demonstrates that HLA-B*35 contributes to the dysregulated expression of selected ER stress, inflammation and proliferation related genes in lcSSc patient PBMCs, as well as healthy individuals, thus supporting a pathogenic role of HLA-B*35 in the development of PAH in SSc patients.
# Introduction
The contribution of genetic factors to the development of systemic sclerosis (SSc, Scleroderma) is well documented [bib_ref] The role of genetics and epigenetics in the pathogenesis of systemic sclerosis, Broen [/bib_ref]. Genetic studies showed a higher incidence of SSc in families with a history of disease compared to the general population (1.5-1.7 % vs 0.026 %). Also, family studies revealed that the relative risk of developing SSc in firstdegree relatives of affected individuals is higher than in third-degree relatives [bib_ref] Autophagy is a key feature in the pathogenesis of systemic sclerosis, Frech [/bib_ref]. The susceptibility loci within the MHC (major histocompatibility complex) region consistently showed strong association with SSc in different cohorts and were confirmed in a large-scale genome-wide association study (GWAS) [bib_ref] Immunochip analysis identifies multiple susceptibility loci for systemic sclerosis, Mayes [/bib_ref]. Of particular interest is HLA-B*35 (human leukocyte antigen class B), which was shown to be associated with increased risk for developing pulmonary arterial hypertension (PAH) in Italian SSc patients [bib_ref] HLA-B35 upregulates the production of endothelin-1 in HLA-transfected cells: a possible pathogenetic..., Santaniello [/bib_ref] [bib_ref] Postmenopause is the main risk factor for developing isolated pulmonary hypertension in..., Scorza [/bib_ref]. This was confirmed in a study of Brazilian SSc patients. HLAB*35 was also associated with SSc in a Choctaw Indian tribe with increased prevalence of SSc [bib_ref] KIR-HLA class i and pulmonary tuberculosis in the Amerindian population in Chaco, Habegger De Sorrentino [/bib_ref]. Furthermore, the association between HLA-B*35 and various other disorders as well as severe viral infections has been reported [bib_ref] KIR-HLA class i and pulmonary tuberculosis in the Amerindian population in Chaco, Habegger De Sorrentino [/bib_ref] [bib_ref] Associations between polymorphisms of HLA-B gene and postmenopausal osteoporosis in Chinese Han..., Li [/bib_ref] [bib_ref] HLA-B35, a common genetic trait, in a familial case of Henoch-Schoenlein purpura..., Pellegrin [/bib_ref]. In particular, studies in patients with HIV (human immunodeficiency virus) infection from different geographical areas have shown a correlation between HLA-B*35 phenotype and progression of AIDS (Acquired Immune Deficiency Syndrome) [bib_ref] HLA and HIV-1: heterozygote advantage and B*35-Cw*04 disadvantage, Carrington [/bib_ref] [bib_ref] HLA-B35 is associated with accelerated progression to AIDS, Itescu [/bib_ref] [bib_ref] Progression to AIDS in French haemophiliacs: association with HLA-B35, Sahmoud [/bib_ref].
We have previously observed that the presence of HLA-B*35 contributes to endothelial cell dysfunction by significantly increasing production of endothelin-1 (ET-1) and significantly decreasing endothelial nitric oxide synthase (eNOS) in conjunction with the upregulation of endoplasmic reticulum (ER) stress and unfolded protein response (UPR) in cultured endothelial cells (ECs) [bib_ref] HLA-B35 and dsRNA induce endothelin-1 via activation of ATF4 in human microvascular..., Lenna [/bib_ref] [bib_ref] HLA-B35 upregulates endothelin-1 and downregulates endothelial nitric oxide synthase via endoplasmic reticulum..., Lenna [/bib_ref]. Furthermore, HLA-B*35-dependent activation of ER stress/UPR correlated with upregulation of the interferon-regulated genes and other inflammatory genes, including IL-6.
A subsequent study using peripheral blood mononuclear cells (PBMCs) obtained from limited cutaneous systemic sclerosis (lcSSc) patients also demonstrated elevated levels of several ER stress markers, particularly in lcSSc patients with PAH. A positive correlation between selected ER stress/UPR markers (BiP/GRP78, glucose regulated protein, and DNAJB1) and IL-6 was also observed, suggesting that ER stress/UPR may have a role in the altered function of circulating immune cells in patients with lcSSc [bib_ref] Increased expression of endoplasmic reticulum stress and unfolded protein response genes in..., Lenna [/bib_ref].
Given the association of the HLA-B*35 with the ER stress and UPR in endothelial cells, in this study, we examined in greater detail the potential contribution of HLA-B*35 to the dysregulated pathways in lcSSc lymphocytes.
# Materials and methods
## Study participants
The study subjects consisted of the patients described in our previous study [bib_ref] Increased expression of endoplasmic reticulum stress and unfolded protein response genes in..., Lenna [/bib_ref] [bib_ref] Limited systemic sclerosis patients with pulmonary arterial hypertension show biomarkers of inflammation..., Pendergrass [/bib_ref] , as well as additional healthy controls and lcSSc patients (described in Additional file 1: . The Boston University Medical Center Institutional Review Board (Boston, MA, USA) reviewed and approved the conduct of this study. Informed consent was obtained from all patients and healthy subjects. Subjects included 97 patients with lcSSc according to diagnostic [17] and subtype criteria [bib_ref] Scleroderma (systemic sclerosis): classification, subsets and pathogenesis, Leroy [/bib_ref] (44 with PAH and 53 without PAH), as well as 49 normal healthy controls.
Patients with lcSSc were stratified into those with or without PAH based on echocardiography or right heart catheterization (RHC); in all patients designated as PAH (n = 44), the diagnosis was confirmed by RHC (mean pulmonary arterial pressure (mPAP) ≥ 25 mm Hg, pulmonary capillary wedge pressure (PCWP) ≤15 and a pulmonary vascular resistance (PVR) ≥3 Wood units), or with PCWP > 15, but ≤ 18 considered to have PAH if adjudicated by the attending pulmonologist on the basis of PVR, PAd-PCWP gradient and trans-pulmonary gradient (and were enrolled in the RE-VEAL Registry as patients with PAH). Patients were considered not to have PAH if echocardiography demonstrated a pulmonary artery systolic pressure <35 mm Hg and normal right ventricular size and function. The modified Rodnan skin score (mRSS) was determined for each patient on the day of the PBMC collection [bib_ref] The modified Rodnan skin score is an accurate reflection of skin biopsy..., Furst [/bib_ref].
SSc disease duration was measured from the onset of the first non-Raynaud's phenomenon symptom of SSc. The mean ± SD disease duration in lcSSc patients was 10 ± 9 years. The mean ± SD age in lcSSc patients was 58 ± 9 years (80 % were women and 45 % had PAH). The mean age for the healthy controls (HCs) was 44 ± 18 years (21 of the 49 HCs were under 30 years old), and 43 % were women.
## Peripheral blood mononuclear cell isolation
Blood was collected from healthy controls and patients in CPTTM tubes designed for one-step cell separation (Becton Dickinson, Mountain View, CA, USA). The sample was then immediately mixed and centrifuged at 1,800 RCF at ambient temperature for 30 min. The PBMC cell layer was then transferred to a 15 ml tube, and PBMCs were washed twice with PBS and lysed in RNeasy RLT buffer (Qiagen, Valencia, CA, USA).
## Lentiviral infection of pbmcs
A lentiviral vector expressing HLA-B*35 (or HLA-B*8) was generated by Applied Biological Materials Inc (Richmond, BC, Canada). Briefly, the cDNA encoding HLAB*35/B*8 was cloned in the shuttle vector pLenti-II-HA-CMV, which contains a HIS tag driven by a separate CMV promoter, and was used to generate recombinant lentiviruses. Lentivirus pLenti-II-HA-CMV was used as a control vector.
Healthy control PBMCs were plated in six-well plates at a density of 0.8-1x10 6 cells/well in RPMI supplemented with 10 % FCS and 1 % AA overnight prior to transduction.
Transductions were performed using M.O.I.'s ranging from 0.1 to 1 mixing the appropriate volume of virus with 8 mg/ ml Polybrene (Sigma, St. Louis, MO, USA), and adding the mixture to the cells together with RPMI to achieve a total volume of 500 μL per well. After 5-6 h incubation at 37°C an additional 500 μL of complete RPMI was added, cells were centrifuged for 30 min at 1200 rpm and culture medium was aspirated and replaced by fresh RPMI. The transduced cells were collected after 72 h. Total RNA was extracted using Qiagen's RNeasy Mini Kits according to the manufacturer's protocol.
# Microarray data analysis
The RNA quality and yield were assessed with an Agilent 2100 Bioanalyzer (Agilent, Santa Clara, CA, USA) and a NanoDrop Technologies ND-1000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). All microarray experiments were performed in one batch. Two hundred nanograms of total RNA were amplified and purified using a TotalPrep RNA Amplification Kit (Applied Biosystems/Ambion, Foster City, CA, USA). The amplified complementary DNA was hybridized on Illumina HT-12 arrays, and the data were extracted with Illumina Genome studio software. Pathway analysis was performed using BRB-ArrayTools (National Cancer Institute, USA). Overrepresented Biocarta pathways were identified using Efron-Tibshirani's GSA test p <0.005. Efron-Tibshirani's test uses 'maxmean' statistics to identify gene sets differentially expressed. All heatmaps show unsupervised hierarchical analysis results (data have been submitted to Gene Expression Omnibus (GEO) public repository, accession number GSE73355).
## Quantitative real-time pcr
Real-time RT-PCR was performed using IQTM SYBR Green Supermix (BioRad, Waltham, MA, USA and MyiQ™ Single-Color Real-Time PCR Detection System (BioRad, Waltham, MA, USA). The amount of template used in the PCR reactions was cDNA corresponding to 200 ng reversetranscribed total RNA. DNA polymerase was first activated at 95°C for 3 min, denatured at 95°C for 30 s, and annealed/extended at 61°C for 30 s, for 40 cycles according to the manufacturer's protocol. Expression of the housekeeping genes β-actin, GADPH, and 18S served as internal positive controls in each assay performed. After measurement of the relative fluorescence intensity for each sample, the amount of each mRNA transcript was expressed as a threshold cycle (c(t)) value. The primer sequences are available upon request.
# Statistical analysis
Comparisons of RT-PCR expression were analyzed using Mann-Whitney non-parametric analyses. Correlations were calculated using Spearman non-parametric correlations.
# Results
The presence of HLA-B*35 allele exacerbates activation of selected ER stress/UPR genes in lcSSc PBMCs
The study subjects (lcSSc with PAH and without PAH), as well as healthy controls (HC) were stratified for the presence of HLA-B*35 allele (18 % of the HCs and 27 % of the lcSSc were B35-positive, of those 27 % of PAH and 26 % of NoPAH patients were B35-positive). We examined the correlation between the presence of HLA-B*35 and the expression of ER stress markers, focusing on the genes previously shown to be elevated in lcSSc-PAH PBMCs [bib_ref] Increased expression of endoplasmic reticulum stress and unfolded protein response genes in..., Lenna [/bib_ref]. Among the previously tested ER stress markers, the chaperones BiP and DNAJB1 were consistently elevated in PBMC samples carrying the HLA-B*35 allele compared to samples negative for HLA-B*35. BiP was elevated in both B35-positive healthy controls (p < 0.05) and in B35positive lcSSc patients (p < 0.05) [fig_ref] Figure 1: HLA-B*35 correlates with higher expression of selected ER stress/UPR genes [/fig_ref] , upper panel). Likewise, DNAJB1 was expressed at higher levels in B35positive HCs (p < 0.01) and B35-positive lcSSc patients (p < 0.0001). Furthermore, the highest levels of BiP and DNAJB1 were present in B35-positive lcSSc-PAH samples (BiP, p < 0.05 lcSSc-PAH B35+ vs lcSSc-PAH B35-; DNAJB1 p < 0.005 lcSSc-PAH B35+ vs lcSSc-PAH B35-) [fig_ref] Figure 1: HLA-B*35 correlates with higher expression of selected ER stress/UPR genes [/fig_ref] , lower panel). Among the other UPR genes associated with lcSSc, only ATF4 (activating transcription factor 4) was elevated in B35-positive lcSSc vs B35-negative lcSSc (p < 0.0005), but these differences were not seen in HC samples [fig_ref] Figure 1: HLA-B*35 correlates with higher expression of selected ER stress/UPR genes [/fig_ref]. These results suggest that the HLA-B*35 allele may primarily influence the expression of chaperons, such as BiP and DNAJ in PBMCs. Notably, a significantly increased expression of the ER stress genes was also observed in comparisons of HC B35-and lcSSc B35-samples. This suggests that besides HLA-B*35 other stressful conditions such as inflammation, infection or oxidative stress may contribute to elevated ER stress and UPR gene expression.
## Global gene expression analysis after transduction of hla-b*35
Microarray analyses were used to further understand the role of HLA-B*35 allele. Lentivirus vector was used to ectopically express HLA-B*35 or HLA-B*8 (another antigen of class I, not known to be associated with an increased risk for developing PAH in patients with lcSSc). PBMCs were isolated from healthy controls and transduced with lentiviruses (empty lentivirus served as an additional control). The basal expression levels of a number of genes were significantly changed in response to lentivirus carrying B35 compared to B8 (or control).
Sixty-four pathways were over-represented in HLA-B*35 vs. HLA-B*8 comparison (List of pathways in Additional file 2: . Among the upregulated pathways were heat shock proteins (BiP, DNAJB1), eicosanoid metabolism (ALOXA5P, arachidonate 5-lipoxygenaseactivating protein), kinases (FYN, ATM), and inflammation (HMGB1, high-mobility group protein B1). Genes with decreased expression levels were related to the cell cycle pathway (inhibitor CDNK1A, cyclindependent kinase inhibitor 1A), the apoptotic pathway (Bax and Gadd45, growth arrest and DNA-damageinducible 45), and the complement pathway (C1QB and C1QC, complement component 1, q subcomponent, C and B chain) [fig_ref] Figure 2: Heatmap showing the expression of gene clusters [/fig_ref]. Genes that showed the most pronounced changes in the array were further confirmed in PBMC cell lines isolated from four different HCs transduced with lentivirus carrying HLA-B*35 (B8 and control virus) by qPCR (Additional file 3: [fig_ref] Figure 1: HLA-B*35 correlates with higher expression of selected ER stress/UPR genes [/fig_ref]. Interestingly, one of our top hits, ALOX5P, was not consistently changed in the transduced HCs used for verification and was not further investigated.
## The presence of hla-b*35 allele in pbmcs enhances inflammation
We have previously reported that interleukin 6 (IL-6) mRNA levels were significantly elevated in lcSSc vs healthy control PBMCs, with the highest levels in lcSSc-PAH PBMCs [bib_ref] Increased expression of endoplasmic reticulum stress and unfolded protein response genes in..., Lenna [/bib_ref]. When HC and lcSSc PBMCs were stratified based on the presence of the HLA-B*35 allele, IL-6 was expressed at a higher level in HLA-B*35positive PBMCs. The association between HLA-B*35 and higher IL-6 was observed in lcSSc PBMCs obtained from patients with and without PAH, but not in healthy controls [fig_ref] Figure 3: IL-6 and HMGB1 are elevated in B35-positive subjects [/fig_ref]. We have previously noted a positive correlation (r = 0.53, p < 0.0001) between mRNA expression of IL-6 and BiP in PBMC samples from patients with lcSSc [bib_ref] Increased expression of endoplasmic reticulum stress and unfolded protein response genes in..., Lenna [/bib_ref]. Notably, IL-6 expression was also associated with the presence of HLA-B*35. When lcSSc PBMC samples were stratified based on the presence of HLA-B*35 allele, the correlation between IL-6 and BiP The array analysis identified an injury response alarmin family member, HMGB1 upregulated in the presence of HLA-B*35. HMGB1 was elevated in lcSSc PBMCs vs HC PBMCs (p < 0.0001). Furthermore, the expression level of HMGB1 was elevated in B35-positive HC (p < 0.05) and lcSSc (p < 0.0001) samples. However, no differences were observed between lcSSc-NoPAH and lcSSc-PAH PBMCs or in the further stratification for the presence of antigen HLA-B*35 [fig_ref] Figure 3: IL-6 and HMGB1 are elevated in B35-positive subjects [/fig_ref]. These results suggest that HLA-B*35 may influence the expression of selected inflammatory genes. Complement complexes are part of the innate immune system and their activation is known to be involved in the pathogenesis of systemic autoimmune diseases [bib_ref] The complement system in systemic autoimmune disease, Chen [/bib_ref]. Complement genes, C1QC and C1QB, were downregulated in HC PBMCs transduced with lentivirus B35. Interestingly, both genes were elevated in PBMCs from lcSSc patients without PAH, but were expressed at significantly lower levels in lcSSc-PAH samples when compared to NoPAH samples (p < 0.005) [fig_ref] Figure 4: Expression of selected complement genes is decreased in HLA-B*35 positive lcSSc PBMCs [/fig_ref]. Further stratification for the presence of B35 revealed that HLA-B*35 correlated with the low levels of the complement genes, with the lowest levels observed in B35positive lcSSc-PAH samples (lcSSc PAH B35+ vs lcSSc PAH B35-, p < 0.01).
## Hla-b*35 correlates with low expression of cell cycle inhibitors and pro-apoptotic genes
Healthy subject PBMCs transduced with the HLA-B*35 lentivirus showed downregulation of the genes related to growth arrest and apoptosis (p21, p57, BAX, Gadd45). Analysis of patient PBMCs also showed significantly lower levels of the cyclin-dependent kinase (CDK) inhibitors, p21 and p57, in B35-positive lcSSc PBMCs compared to B35-negative lcSSc (p < 0.01 and p < 0.001, respectively) [fig_ref] Figure 5: HLA-B*35 is associated with low levels of selected cyclin inhibitors and pro-apoptotic... [/fig_ref]. Healthy controls showed significantly decreased p21, but not p57, in B35positive samples. Further stratification for the presence of HLA-B*35 in lcSSc revealed no difference in lcSSc-NoPAH B35-vs lcSSc-NoPAH B35+ while moderately lower levels were observed in lcSSc-PAH B35positive compared to lcSSc-PAH B35-negative samples (Additional file 4: [fig_ref] Figure 2: Heatmap showing the expression of gene clusters [/fig_ref]. Pro-apoptotic genes, such as Bax and Gadd45, were also downregulated in HLA-B*35 positive samples obtained from HC and lcSSc subjects [fig_ref] Figure 5: HLA-B*35 is associated with low levels of selected cyclin inhibitors and pro-apoptotic... [/fig_ref]. Low levels were also observed in B35-positive lcSSc-NoPAH and PAH samples (lcSSc PAH B35+ vs lcSSc PAH B35-, p < 0.05) (Additional file 4: [fig_ref] Figure 2: Heatmap showing the expression of gene clusters [/fig_ref].
The above global gene expression analysis indicated that two proliferation-associated genes, FYN tyrosine kinase and ATM serine/threonine kinase, are upregulated in HLA-B*35 transduced PBMCs. Accordingly, the levels of FYN were elevated in B35-positive HCs (p < 0.05) and B35-positive lcSSc samples (p<0.005) [fig_ref] Figure 5: HLA-B*35 is associated with low levels of selected cyclin inhibitors and pro-apoptotic... [/fig_ref]. Interestingly, FYN levels were lower in lcSSc-PAH vs lcSSc-NoPAH but elevated in B35-positive subjects in both subpopulations (Additional file 4: [fig_ref] Figure 2: Heatmap showing the expression of gene clusters [/fig_ref]. In contrast to the microarray results, ATM was expressed at lower levels in B35-positive HCs, but its expression did not differ in B35-positive and B35-negative lcSSc samples [fig_ref] Figure 5: HLA-B*35 is associated with low levels of selected cyclin inhibitors and pro-apoptotic... [/fig_ref]. Also, no differences were observed between lcSScNoPAH and PAH, but further stratification for the presence of HLA-B*35 revealed slightly increased levels in lcSSc-NoPAH B35+ vs lcSSc-NoPAH B35-and slightly decreased levels in lcSSc-PAH B35+ vs lcSSc-NoPAH B35-(Additional file 4: [fig_ref] Figure 2: Heatmap showing the expression of gene clusters [/fig_ref]. These results suggested that the presence of HLA-B*35 may influence apoptotic and proliferative responses in PBMC subpopulations.
# Discussion
This study explored the potential contribution of HLA-B*35 to the immune dysregulation in lcSSc. An unbiased approach based on the microarrays from the human PBMCs transduced with the HLA-B*35 carrying lentivirus revealed a number of genes modulated in response to HLA-B*35. Selected genes were then verified in PBMCs obtained from patients with lcSSc as well as healthy controls. Among the genes that significantly correlated with the presence of HLA-B*35 in PBMCs were the heat shock proteins, inflammatory genes, complement genes, and genes related to cell growth and apoptosis. The upregulation of heat shock proteins typically occurs in response to various stressful conditions, including inflammation, infection, and various environmental toxins. In particular, the HSP group, which includes BiP (HSPA5) and its cochaperone DNAJ (HSP40), is required for protein folding and is highly expressed during ER stress [bib_ref] DNAJs: more than substrate delivery to HSPA, Dekker [/bib_ref]. Notably, higher levels of heat shock proteins were also present in PBMCs obtained from healthy individuals carrying the Inflammation and, in particular, elevated levels of IL-6 have been linked to the development of PAH [bib_ref] Inflammation and immunity in the pathogenesis of pulmonary arterial hypertension, Rabinovitch [/bib_ref]. Recent studies have suggested that blocking IL-6 improves both skin and interstitial lung disease in patients with dSSc (http://acrabstracts.org/abstracts/autotaxin-is-highlyexpressed-in-systemic-sclerosis-ssc-skin-mediates-dermalfibrosis-via-il-6-and-is-a-target-for-ssc-therapy/). In our study, increased levels of IL-6 in HLA-B*35-positive lcSSc PBMCs suggests that this is a genetic risk factor leading to enhanced sensitivity of HLA-B*35 leukocytes to activation. Further, our observation that the highest IL-6 levels and the highest expression of ER stress markers, BiP and HSP40, are found in B35-positive lcSSc-PAH samples, suggests that this relationship between ER stress and IL-6 plays a key role in the development of lcSSc-PAH.
Notably, we also found higher levels of HGMB1 in both HLA-B*35-positive lcSSc subjects and healthy controls. Serum levels of HGMB1 were previously shown to be elevated in SSc [bib_ref] Clinical significance of serum HMGB-1 and sRAGE levels in systemic sclerosis: association..., Yoshizaki [/bib_ref]. HMGB1, as well as HSPs, are part of the alarmin family, the endogenous molecules constitutively available and released after injury. Alarmins can promote activation of innate immune cells, recruitment and activation of antigen-presenting cells for host defense and tissue repair through activation of TLRs (Toll-like receptors) [bib_ref] Alarmins: awaiting a clinical response, Chan [/bib_ref]. Thus, elevated HGMB1 may represent another important mediator of the effect of HLA-B*35 on immune dysregulation in lcSSc patients. Previous studies have identified altered expression levels of several additional inflammatory mediators in lcSSc PBMCs, including MCP1, IL-13, and IL-7R [bib_ref] Interferon and alternative activation of monocyte/macrophages in systemic sclerosis-associated pulmonary arterial hypertension, Christmann [/bib_ref] [bib_ref] Altered immune phenotype in peripheral blood cells of patients with scleroderma-associated pulmonary..., Risbano [/bib_ref] [bib_ref] Identification of candidate genes in scleroderma-related pulmonary arterial hypertension, Grigoryev [/bib_ref]. However, the presence of HLA-B35 had no effect on the expression of those genes .
Among the HLA-B*35-regulated genes related to the immune system were the complement genes, C1QB and C1QC. Both genes were moderately elevated in lcSSc subjects without PAH in comparison to healthy controls; however, their expression was significantly reduced in lcSSc-PAH samples. HLA-B*35-transduced PBMCs had reduced levels of C1Q genes and this finding was verified in PBMCs from healthy controls as well as lcSSc samples with and without PAH. Complement is part of the innate immune system and its major function is to recognize and eliminate pathogens. In particular, formation of immune complexes is one of the principal ways of activating the classical pathway of the complement system. If the complement system fails in this function, waste material can accumulate and evoke an autoimmune response. Genetic deficiency of C1Q is a strong risk factor for development of SLE (systemic lupus erythematosus), triggering proinflammatory mediators, such as C5a and C3, and impaired cytokine production resulting in persistent and recurrent viral infections, known to be an exacerbating factor for SLE [bib_ref] Complement deficiencies in systemic lupus erythematosus, Bryan [/bib_ref] [bib_ref] The complement system in systemic lupus erythematosus: an update, Leffler [/bib_ref] [bib_ref] C1q: its functions within the innate and adaptive immune responses and its..., Sontheimer [/bib_ref] [bib_ref] Systemic lupus erythematosus due to C1q deficiency with progressive encephalopathy, intracranial calcification..., Troedson [/bib_ref] , but much less is known about the role of complement in SSc. The biological significance of the reduced levels of C1Q in carriers of the HLA-B*35 allele remains to be clarified.
Lastly, we found significantly decreased levels of selected cyclin inhibitors and pro-apoptotic genes in HLA-B*35-positive PBMCs obtained from lcSSc patients and healthy controls. On the other hand, expression of a tyrosine-protein kinase FYN was upregulated in HLA-B*35 positive PBMCs. FYN plays a role in many biological processes including regulation of cell growth and survival [bib_ref] Function of the Src-family kinases, Lck and Fyn, in T-cell development and..., Palacios [/bib_ref] [bib_ref] T-cell receptor proximal signaling via the Src-family kinases, Lck and Fyn, influences..., Salmond [/bib_ref]. It participates in the downstream signaling pathways that lead to T-cell differentiation and proliferation following T-cell receptor (TCR) stimulation. These results suggest that the presence of HLA-B*35 may favor proliferation of the immune cells and thus contribute to the increased inflammatory response. However, more studies are needed to fully appreciate the functional significance of the presence of HLA-B*35 allele in patients with SSc.
# Conclusions
In summary, the current study further extends our previous findings on the role of HLA-B*35 in endothelial cells [bib_ref] HLA-B35 upregulates endothelin-1 and downregulates endothelial nitric oxide synthase via endoplasmic reticulum..., Lenna [/bib_ref]. In both cell types HLA-B*35 induced ER stress and inflammation related genes. Importantly, the current study verified these experimental findings in cells obtained from lcSSc patients. Notably, the presence of HLA-B*35 correlated with increased levels of alarmins, including HSPs and HMGB1, in healthy individuals, indicating that the presence of HLA-B*35 induces a stress response and is likely to sensitize endothelial and immune cells to further stressful conditions. While some of the biological consequences of HLA-B*35, including modulation of the complement and apoptotic responses, requires further investigation, this study supports the pathological role of HLA-B*35 in SSc.
## Additional files
Additional file 1: . Clinical and hemodynamic data of study subjects. PAP = pulmonary artery pressure. PCWP = pulmonary capillary wedge pressure. CO/CI = Cardiac output (L/min)/ cardiac index (L/min/m 2 ). PVR = pulmonary vascular resistance. ILD = interstitial lung disease.
[fig] Figure 1: HLA-B*35 correlates with higher expression of selected ER stress/UPR genes. PBMCs were isolated from HC (n = 49), lcSSc (n = 97, NoPAH n = 53 and PAH n = 44), and grouped according to the presence of the HLA-B*35 allele: HC B35+ (n =9), HC B35-(n = 40); lcSSc B35+ (n =26), lcSSc B35-(n =71); lcSSc NoPAH B35+ (n =14), lcSSc NoPAH B35-(n = 39), lcSSc PAH B35+ (n = 12) and lcSSc PAH B35-(n = 32). mRNA levels of BiP (left panel), DNAJB1 (middle panel), and ATF4 (right panel) were measured by qPCR. Expression of the housekeeping genes β-actin, GADPH, and 18S served as internal positive controls. Data are expressed as the fold-change normalized to mRNA expression in a single HC sample. Each data point represents a single subject; horizontal lines show the mean. ER endoplasmic reticulum, PBMCs peripheral blood mononuclear cells, HC healthy controls, lcSSc limited cutaneous systemic sclerosis, PAH pulmonary arterial hypertension, ATF4 activating transcription factor 4 was higher in B35-positive samples compared to B35negative samples (r = 0.36 vs r = 0.26). [/fig]
[fig] Figure 2: Heatmap showing the expression of gene clusters. PBMCs were isolated from healthy control and transduced with 0.1-0.5-1 MOI of lentivirus encoding HLA-B*35, HLA-B*8, or control lentivirus for 72 h. The global changes in gene expression were investigated by Illumina HT-12 arrays (Illumina Inc, San Diego, CA, USA). Among genes downregulated by HLA-B*35 lentivirus compared to HLA-B*8, we observed genes related to complement (C1QB, C1QC), cell cycle (CDNK1A), and apoptotic (Bax, Gadd45) pathways. Genes with increased expression levels were related to proliferation (FYN, ATM), inflammation (HMGB1), and ER stress/UPR (HSPA1A and DNAJB1). Expression values above the mean are indicated in dark blue, those below the mean are indicated in light blue. PBMCs, peripheral blood mononuclear cells, MOI multiplicity of infection, ER endoplasmic reticulum, UPR unfolded protein response [/fig]
[fig] Figure 3: IL-6 and HMGB1 are elevated in B35-positive subjects. PBMCs were isolated from HC (n = 49), lcSSc (n = 97, NoPAH n = 53 and PAH n = 44) and grouped according to the presence of the HLA-B*35 allele: HC B35+ (n = 9), HC B35-(n = 40); lcSSc B35+ (n =26), lcSSc B35-(n =71); lcSSc NoPAH B35+ (n =14), lcSSc NoPAH B35-(n = 39), lcSSc PAH B35+ (n = 12) and lcSSc PAH B35-(n = 32). mRNA levels of IL-6 (top panel) and HMGB1 (bottom panel) were determined by qPCR. Expression of the housekeeping genes β-actin, GADPH, and 18S served as internal positive controls. Data are expressed as the fold-change normalized to mRNA expression in a single HC sample. Each data point represents a single subject; horizontal lines show the mean. Right top panel depicts linear regression analysis of the relationship between expression of BiP and IL-6 in PBMCs from lcSSc B35-negative and B35positive patients. IL-6 interleukin-6, HMGB1 high-mobility group protein B1, PBMCs, peripheral blood mononuclear cells, HC healthy controls, lcSSc limited cutaneous systemic sclerosis, PAH pulmonary arterial hypertension HLA-B*35 allele, supporting the view that genetic factors could contribute to the increased levels of ER stress at least in a restricted population of SSc patients. [/fig]
[fig] Figure 4: Expression of selected complement genes is decreased in HLA-B*35 positive lcSSc PBMCs. PBMCs were isolated from HC (n = 49), lcSSc (n = 82, NoPAH n = 43 and PAH n = 39) and grouped according to the presence of the HLA-B*35 allele: HC B35+ (n =9), HC B35-(n = 40); lcSSc NoPAH B35+ (n =14), lcSSc NoPAH B35-(n = 29), lcSSc PAH B35+ (n = 12) and lcSSc PAH B35-(n = 27). mRNA levels of C1QC and C1QB were measured by qPCR. Expression of the housekeeping genes β-actin, GADPH and 18S served as internal positive controls in each assay performed. lcSSc limited cutaneous systemic sclerosis, PBMCs peripheral blood mononuclear cells, PAH pulmonary arterial hypertension, HC healthy controls, qPCR quantitative polymerase chain reaction [/fig]
[fig] Figure 5: HLA-B*35 is associated with low levels of selected cyclin inhibitors and pro-apoptotic genes in lcSSc PBMCs. PBMCs were isolated from HC (n = 49), lcSSc (n = 81) and grouped according to the presence of the HLA-B*35 allele: HC B35+ (n = 9), HC B35-(n = 40); lcSSc B35+ (n = 25) and lcSSc B35-(n = 56). mRNA levels of p21, p57 a, Bax, Gadd45 b and FYN, ATM c were measured by qPCR. Expression of the housekeeping genes β-actin, GADPH, and 18S served as internal positive control in each assay performed. lcSSc limited cutaneous systemic sclerosis, PBMCs peripheral blood mononuclear cells, HC healthy controls [/fig]
[fig] FVC: (%) = estimated forced vital capacity. DLCO = carbon monoxide diffusing capacity. SPAP = estimated systolic pulmonary artery pressure by echocardiogram. ILD was defined as present (Y = yes) or absent (N = no) based on high-resolution computed tomography assessment of the lungs. (XLS 73.5 kb) Additional file 2:Table S2. List of Gene Sets: 64 pathways sorted by LS permutation p-value. (XLS 152 kb) Additional file 3:Figure S1.Validation of array results in HC PBMCs transduced with lentivirus. Expression levels of selected genes upregulated (HSPA1A, known as BiP, DNAJB1, HMGB1, FYN, and ATM) and downregulated (CDKNA1, known as p21, Bax, Gadd45, C1QC, and C1QB) in the array analysis were verified by qPCR in four PBMC cell lines freshly isolated from healthy controls transduced with lentivirus encoding HLA-B*35 or HLA-B*8. Empty lentivirus served as additional control. Graph represents average of four different HC PBMC cell lines. (TIF 127 kb) Additional file 4: Figure S2. HLA-B*35 is associated with low levels of selected cyclin inhibitors and pro-apoptotic genes in lcSSc PBMCs. PBMCs were isolated from HC (n = 49), lcSSc (n = 81, NoPAH n = 43, and PAH n = 38) and grouped according to the presence of the HLA-B*35 allele: HC B35+ (n = 9), HC B35-(n = 40); lcSSc NoPAH B35+ (n = 14), lcSSc NoPAH B35-(n = 29), lcSSc PAH B35+ (n = 12) and lcSSc PAH B35-(n = 26). mRNA levels of p21, p57 (a), Bax, Gadd45 (b), and FYN, ATM (c) were measured by qPCR. Expression of the housekeeping genes β-actin, GADPH and 18S served as internal controls in each assay performed. (TIF 1.63 mb) Abbreviations AIDS: Acquired immune deficiency syndrome; ALOXa5p: Arachidonate 5lipoxygenase-activating protein; ATF4: activating transcription factor 4; BiP/ GRP78: glucose regulated protein; C1QC/C1QB: complement component 1, q subcomponent, C / B chain; CDKN1A: Cyclin-dependent kinase inhibitor 1; EC: endothelial cells; eNOS: endothelial nitric oxide synthase; ER: endoplasmic reticulum; ET-1: endothelin-1; HC: healthy controls; HIV: human immunodeficiency virus; HLA-B*: human leukocyte antigen class B; HMGB1: high-mobility group protein B1; IL-6: interleukin 6; lcSSc: limited cutaneous systemic sclerosis; MHC: major histocompatibility complex; mPAP: mean pulmonary arterial pressure; PAH: Pulmonary arterial hypertension; PBMC: Peripheral blood mononuclear cell; PCWP: pulmonary capillary wedge pressure; PVR: pulmonary vascular resistance; SLE: systemic lupus erythematosus; SSc: systemic sclerosis, Scleroderma; TLRs: Toll-like receptors; UPR: unfolded protein response. [/fig]
|
Prenatal Diagnosis of Talipes Equinovarus by Ultrasound and Chromosomal Microarray Analysis: A Chinese Single-Center Retrospective Study
Li, R.; Jing, X.; Han, J.; et al. Prenatal Diagnosis of Talipes Equinovarus by Ultrasound and Chromosomal Microarray Analysis: A Chinese Single-Center Retrospective Study.Genes 2022, 13, 1573. https://
# Introduction
Talipes equinovarus (TE) is the most common congenital malformation of the foot [bib_ref] The relationship between isolated pes equinovarus and aneuploidies and perinatal outcomes: Results..., Sucu [/bib_ref]. Its incidence is about 1-3 in 1000 of live births, and the proportion of male to female fetuses affected is about 2:1. TE includes four elements: metatarsus adductus, cavus foot, heel varus, and equinus, which is detected antenatally in over half of these cases by fetal ultrasound. It can be unilateral or bilateral (60-70%) and can be either an isolated deformity (50-70%) or a manifestation of chromosomal abnormalities and other genetic syndromes (30-50%) [bib_ref] Congenital talipes equinovarus (clubfoot), Mckinney [/bib_ref]. Its etiology is not yet clear, but genetic and environmental factors are known to play an important role. Despite a high prevalence of TE, only a few pathogenic genes are known. The PITX1, IGFBP3, TBX4, and RBM10 genes have been found to be associated with TE [bib_ref] Altered transmission of HOX and apoptotic SNPs identify a potential common pathway..., Ester [/bib_ref] [bib_ref] Dual hindlimb control elements in the Tbx4 gene and region-specific control of..., Menke [/bib_ref] [bib_ref] Three novel missense mutations in the filamin B gene are associated with..., Yang [/bib_ref]. Although some TE fetuses, known as positional TE, return to a normal position, this is associated with intrauterine factors that limit fetal movements, such as oligohydramnios, twins, and uterine malformations [bib_ref] The clubfoot over the centuries, Sanzarello [/bib_ref]. Therefore, assessing the fetal genetic and clinical prognosis in fetuses with TE is essential.
In the past two decades, chromosomal microarray analysis (CMA) has been well studied and utilized in exploring genomic changes in fetuses with structural anomalies sonographically identified in prenatal settings. About 6% of fetuses with abnormal ultrasonography and normal karyotype can be detected with clinically significant chromosomal variations through CMA testing [bib_ref] Prenatal diagnosis of clubfoot: Chromosomal abnormalities associated with fetal defects and outcome..., Viaris De Le Segno [/bib_ref]. Amihood et al. [bib_ref] Prenatal clubfoot increases the risk for clinically significant chromosomal microarray results-Analysis of..., Singer [/bib_ref] performed CMA testing in 269 prenatal cases of singleton pregnancies with TE in 2020 and detected 16 (5.9%) clinically significant variants. In contrast, Alvarado et al. [bib_ref] Copy number analysis of 413 isolated talipes equinovarus patients suggests role for..., Alvarado [/bib_ref] used CMA for genetic testing in 413 postnatal cases with isolated TE in 2013, and clinically relevant variants were identified in 2.4% of cases. Compared to postnatal studies, the detection rate of CMA in patients with prenatal ultrasound findings of TE is higher, and as the phenotype of TE may be a diagnostic clue for certain fetal syndromes, we believe that CMA may be necessary. However, there are only a few studies of the detection rate of CMA in the prenatal diagnosis of TE.
In this study, we review the clinical and molecular findings of 164 Chinese patients diagnosed with fetal TE at our center to explore its molecular etiology and examine the detection rate of TE by CMA, which may provide more information for clinical screening and genetic counseling.
# Materials and methods
## Study cohort
This was a retrospective cohort study by reviewing all prenatal cases of fetal talipes equinovarus diagnosed at the Prenatal Diagnosis Center, Guangzhou Women and Children's Medical Center, from July 2013 to January 2022. All cases underwent a routine ultrasound scan for fetal anatomy, and associated abnormalities were recorded. Throughout the examination, if two long bones of the lower leg (tibia and fibula) were seen in the same plane as the sole, the diagnosis of fetal talipes equinovarus was made [bib_ref] Prenatal diagnosis of clubfoot: Chromosomal abnormalities associated with fetal defects and outcome..., Viaris De Le Segno [/bib_ref]. We divided pregnancies into unilateral or bilateral groups and isolated or non-isolated groups according to the type of TE. All parents of fetal TE received genetic counseling, which included the potential risks of invasive surgery and the possible implications of the findings, by the Maternal-Fetal Medicine team at our center.
We reviewed clinical data from all cases in our medical record database, including demographic data for pregnancies, indications for invasive examinations, ultrasound findings, karyotype/CMA results, and outcomes of pregnancy. Pregnancy outcomes are recorded partly autopsy results after the termination of pregnancy and partly by telephone or case review, focusing on clinical outcomes, gestational age at birth or termination of pregnancy, neonatal sex, presence of talipes equinovarus, other abnormalities, et al. This study was approved by the Ethics Committee of Guangzhou Women and Children's Medical Center. Informed consent was obtained from the pregnant women before the invasive procedure.
# Chromosomal microarray analysis
At our prenatal diagnostic center, CMA has replaced karyotyping as a first-line method for detecting fetal structural abnormalities since 2013. Genomic DNA was extracted from chorionic villi, amniocytes, cord blood using the Qiagen DNA Blood Midi/Mini kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer's protocol. Informed consent was taken to obtain a parental blood sample in order to run a trio analysis. We analyzed submicroscopic genomic imbalances using whole-genome high-resolution microarray analysis with CytoScan HD arrays and CytoScan 750 K arrays (Affymetrix, Santa Clara, CA, USA) according to the manufacturer's protocols. The built reference genome was aligned on GRCh37/hg19. CytoScan 750K or CytoScan HD arrays are used to detect whole genome copy number variants (CNVs), as well as loss of heterozygosity (LOH) and isodisomy of uniparental disomy (iso-UPD), and to detect mosaicism at >30%. The mean turnaround time (TAT) for CMA from uncultured specimens was seven days. The process has been described in detail elsewhere [bib_ref] Implementation of high-resolution SNP arrays in the investigation of fetuses with ultrasound..., Liao [/bib_ref].
Data were analyzed following American College of Medical Genetics guidelines, which categorizes all selected variants as pathogenic (P), likely pathogenic (LP), variants of unknown significance (VOUS), likely benign, or benign [bib_ref] Technical standards for the interpretation and reporting of constitutional copy-number variants: A..., Riggs [/bib_ref].
# Statistical analysis
Statistical analyses were performed using SPSS 25.0 (IBM, Armonk, NY, USA). Chisquare test or Fisher's exact test was used to compare the characteristics among these subgroups. A p value of < 0.05 was considered statistically significant.
# Results
Between July 2013 and January 2022, a total of 212 pregnancies were consulted in our center for fetal TE. Forty-eight pregnancies were excluded as further testing was refused. The mean maternal age was 29.8 (range 20.1-46.0) years, and the median gestational age of the fetus was 25.0 (range 12.7-33.3) weeks. The majority of patients were diagnosed in the second trimester (128/164 (78.0%)), 32 (19.5%) pregnancies were diagnosed in the last trimester, and the fewest number of patients, 2.4% (4/164), were diagnosed in the first trimester. Of these fetuses, 100 (61.0%) were male and 64 (39.0%) were female. This ratio was similar to that previously reported in the literature [bib_ref] Congenital talipes equinovarus (clubfoot), Mckinney [/bib_ref]. After diagnosis, 103 (62.8%) women chose to continue the pregnancy, while 55 (33.5%) chose to terminate of pregnancy (TOP), and six (3.7%) patients were lost to follow-up. The flowchart of genetic analysis progression is shown in [fig_ref] Figure 1: Flowchart of genetic analysis progression in cohort of fetuses with TE [/fig_ref]. Among the 164 patients who met the inclusion criteria, CMA detected 17 (10.4%) clinically significant variants, including 16 (9.8%) fetuses with pathogenic copy number variant (pCNV) and 1 (0.6%) with likely pathogenic copy number variant (lpCNV), within which four fetuses were trisomy 18 (3/164) and mosaic trisomy 21 (1/164). Among 17 cases with clinically significant variants, there were 9 cases (52.9%) with CNVs < 10Mb, but another 8 cases (47.1%) were detected with CNVs > 10Mb. The most common CNV was the 22q11.2 microdeletion syndrome (n = 4). Non-isolated TE was combined with the most common abnormalities in the cardiovascular system (n = 4) and neurologic system (n = 4), and the most common was ventricular septal defect (n = 3). [fig_ref] Table 1: Clinical features in fetuses with talipes equinovarus [/fig_ref] shows the clinical and Among the 164 patients who met the inclusion criteria, CMA detected 17 (10.4%) clinically significant variants, including 16 (9.8%) fetuses with pathogenic copy number variant (pCNV) and 1 (0.6%) with likely pathogenic copy number variant (lpCNV), within which four fetuses were trisomy 18 (3/164) and mosaic trisomy 21 (1/164). Among 17 cases with clinically significant variants, there were 9 cases (52.9%) with CNVs < 10Mb, but another 8 cases (47.1%) were detected with CNVs > 10Mb. The most common CNV was the 22q11.2 microdeletion syndrome (n = 4). Non-isolated TE was combined with the most common abnormalities in the cardiovascular system (n = 4) and neurologic system (n = 4), and the most common was ventricular septal defect (n = 3). [fig_ref] Table 1: Clinical features in fetuses with talipes equinovarus [/fig_ref] shows the clinical and chromosomal characteristics of these 17 clinically significant variants. [fig_ref] Table 2: Stratified analysis of CNVs detection and pregnancy outcome in TE [/fig_ref] shows that the detection rate of CNVs in singleton pregnancies is significantly higher in non-isolated TE than in isolated TE (10/37, 27.0% vs. 6/111, 5.4%, p < 0.05). In twin pregnancies, 6.3% (1/16) were pCNV, which was not statistically different from singleton pregnancies (1/16, 6.3% vs. 16/148, 10.8%, p = 0.891). In terms of pregnancy outcomes, the rate of TOP was significantly higher in the non-isolated TE group than in the isolated TE group (26/37, 70.3% vs. 22/111, 19.8%, p < 0.05) and higher in the unilateral group than in the bilateral group (28/69, 40.6% vs. 20/79, 25.3%, p < 0.05), both of which were statistically significant. We compared unilateral and bilateral foot for isolated and non-isolated TE in singleton pregnancies intergroup and intragroup [fig_ref] Figure 2: Comparison of CMA detection rates of fetuses with talipes equinovarus [/fig_ref]. We found statistically significant differences in the detection rates of bilateral TE between isolated and non-isolated TE (2/63, 3.2% vs. 6/16, 37.5%, p < 0.05). CMA detected 25 cases of VOUS, but for financial reasons, some families rejected the suggestion to perform parental CMA verification because the price was close to $1000. Demographic and chromosomal data of VOUS are shown in [fig_ref] Table 1: Clinical features in fetuses with talipes equinovarus [/fig_ref]. We compared unilateral and bilateral foot for isolated and non-isolated TE in singleton pregnancies intergroup and intragroup [fig_ref] Figure 2: Comparison of CMA detection rates of fetuses with talipes equinovarus [/fig_ref]. We found statistically significant differences in the detection rates of bilateral TE between isolated and non-isolated TE (2/63, 3.2% vs. 6/16, 37.5%, p < 0.05). CMA detected 25 cases of VOUS, but for financial reasons, some families rejected the suggestion to perform parental CMA verification because the price was close to $1000. Demographic and chromosomal data of VOUS are shown in Supplementary Table S1.
# Discussion
In this study, we performed CMA on fetuses with an ultrasound diagnosis of TE and performed a follow-up evaluation to illuminate the genetic and clinical value of CNVs in fetal TE. We found that fetal TE with associated structural malformation correlates with a higher probability of clinically significant variants. The overall detection rate of clinically significant variants is similar to previous literature [bib_ref] Genetics analysis of fetal foot varus during the second and third trimester..., Li [/bib_ref] (18/166, 10.8% vs. 17/164, 10.4%). In contrast, the proportion of CNVs that are undetectable by karyotyping (<10 Mb) among all clinically significant variants is nearly twice as high in our study (9/17, 52.9% vs. 5/18,
# Discussion
In this study, we performed CMA on fetuses with an ultrasound diagnosis of TE and performed a follow-up evaluation to illuminate the genetic and clinical value of CNVs in fetal TE. We found that fetal TE with associated structural malformation correlates with a higher probability of clinically significant variants. The overall detection rate of clinically significant variants is similar to previous literature [bib_ref] Genetics analysis of fetal foot varus during the second and third trimester..., Li [/bib_ref] (18/166, 10.8% vs. 17/164, 10.4%). In contrast, the proportion of CNVs that are undetectable by karyotyping (<10 Mb) among all clinically significant variants is nearly twice as high in our study (9/17, 52.9% vs. 5/18, 27.8%). Moreover, our study includes twin pregnancies and performs a comprehensive comparison of isolated and non-isolated TE, as well as unilateral and bilateral TE.
Talipes equinovarus is categorized into isolated and non-isolated TE. The isolated type is regarded as an isolated anomaly of the lower limbs that may be associated with polygenic inheritance; its prognosis is considered benign [bib_ref] Prenatal diagnosis of isolated clubfoot: Diagnostic accuracy and long-term postnatal outcomes, Fantasia [/bib_ref]. Non-isolated TE affects approximately 25% of fetuses and has been associated with deletion syndromes, aneuploidies, sex chromosomal abnormalities, neuromuscular diseases, microdeletions, and duplications [bib_ref] Genetics of clubfoot; recent progress and future perspectives, Basit [/bib_ref]. In our cohort, there was a significant difference in the detection rate of isolated TE and non-isolated TE (6/111, 5.4% vs. 10/37, 27.0%, p < 0.05), a finding which is similar to previous studies [bib_ref] Prenatal clubfoot increases the risk for clinically significant chromosomal microarray results-Analysis of..., Singer [/bib_ref].
We made comparisons of unilateral and bilateral TE in singleton pregnancies and found no significant differences (8/69, 11.6% vs. 8/79,10.1%, p = 0.774). Surprisingly, when comparing unilateral with bilateral foot for isolated and non-isolated TE in singleton pregnancies, we found a statistically significant difference in the detection rate of bilateral TE between isolated TE and non-isolated TE (2/63, 3.2% vs. 6/16, 37.5%, p < 0.05). This result suggested that when fetal TE is detected, in addition to excluding other structural abnormalities, it is crucial to preclude whether the contralateral foot also shows TE specifically. The reason is that chromosomal abnormalities are prevalent in bilateral TE. If other anomalies or bilateral TE are combined, we recommend further genetic testing.
Our data found a high incidence of 22q11.2 microdeletion syndrome (DiGeorge syndrome, DGS) in TE, which accounted for 23.5% (4/17) of all CNVs. Microdeletion of 22q11.2 is the most common microdeletion syndrome [bib_ref] Orthopaedic manifestations within the 22q11.2 Deletion syndrome: A systematic review, Homans [/bib_ref] , and TBX1 correlates with the most prominent phenotypes characteristic of this syndrome. Patients with 22q11.2 microdeletion syndrome display a broad array of phenotypes, and the most common findings include cardiac anomalies, hypocalcemia, and hypoplastic thymus. Case 10 in our non-isolated TE group was combined with ventricular septal defect, one of the common phenotypes above. Although skeletal anomalies are not a defining feature of DGS, studies reported that 1.1-13.3% of fetuses with this syndrome might have TE phenotype [bib_ref] Orthopaedic manifestations within the 22q11.2 Deletion syndrome: A systematic review, Homans [/bib_ref]. Interestingly, three of these four cases of fetuses with DGS were not combined with other structural abnormalities. To our knowledge, in the previous literature, there was only one case of a fetus with DGS with isolated TE detected by CMA, in Amihood et al. [bib_ref] Prenatal clubfoot increases the risk for clinically significant chromosomal microarray results-Analysis of..., Singer [/bib_ref]. Unfortunately, these three cases all chose to termination of pregnancy and refused autopsy, so we cannot be sure if other common phenotypes of this syndrome were combined. No relevant literature has mentioned whether isolated TE is associated with the deletion of this fragment. We believe it is necessary to perform further studies to illustrate their correlation.
We included LOH in our study. LOH, also known as absence of heterozygosity, refers to long contiguous stretches of homozygosity in a chromosome. The pathogenesis of LOH includes homozygous mutation of recessive diseases and increased susceptibility to complex diseases [bib_ref] Effects of genome-wide heterozygosity on a range of biomedically relevant human quantitative..., Campbell [/bib_ref] [bib_ref] Regions of homozygosity and their impact on complex diseases and traits, Ku [/bib_ref] , imprinting effects caused by uniparental disomy (UPD) [bib_ref] Mechanisms leading to uniparental disomy and their clinical consequences, Robinson [/bib_ref] , hidden mosaicism or confined placental mosaicisms [bib_ref] Uniparental disomy and human disease: An overview, Yamazawa [/bib_ref] , and potential association with tumorigenesis [bib_ref] The consequences of uniparental disomy and copy number neutral loss-of-heterozygosity during human..., Lapunzina [/bib_ref]. When LOH in the imprinting regions is confirmed to have been inherited from only one parent, it can cause imprinting disorders, such as Prader-Willi syndrome (maternal UPD15) and Angelman syndrome (paternal UPD15). Moreover, when LOH occurs on non-imprinted chromosomes, it may expose the causative gene of recessive genetic disorders. Liu et al. [bib_ref] Absence of heterozygosity detected by single-nucleotide polymorphism array in prenatal diagnosis, Liu [/bib_ref] reported that approximately 55% of LOH carriers had ultrasound abnormalities, and multiple malformations were the most common findings. In contrast, TE is generally dominantly inherited [bib_ref] Genotype-phenotype correlation in clubfoot (talipes equinovarus), Hordyjewska-Kowalczyk [/bib_ref] , but four of the five cases of LOH we identified occurred in isolated TE. Is this a coincidence, or is there a correlation between the two? More research is needed to answer this question.
In our data, case 3 was identified with a microduplication of 1.42 Mb in the chromosome 17p12 region, which was suggested to be pathogenic according to OMIM and DECIPHER databases. The clinical condition was Charcot-Marie-Tooth type 1A (CMT1A), associated with the PMP22 gene, and its phenotype is mainly characterized by multiple foot abnormalities and sensory abnormalities [bib_ref] Mosaicism for the Charcot-Marie-Tooth disease type 1A duplication suggests somatic reversion, Liehr [/bib_ref]. Notably, this case was a fetus with isolated TE, so it is reasonable to suspect that there is a more significant correlation with the genetic mechanism of TE, but this link remains to be explored. Case 14 was detected 25.38Mb duplication in chromosome 5p15.33p13.2, which contained 25 known OMIM disease genes. According to the DECIPHER and ClinGen databases, several studies have reported that patients carried segments of the presently detected segments and had a clinical phenotype that included TE [bib_ref] Trisomy of the short arm of chromosome 5 due to a de..., Cervera [/bib_ref].
It seems that several CNVs of non-isolated TE are occasional findings unrelated to foot malformations, such as Wolf-Hirschhorn syndrome, 21q22.13q22.3 microdeletion, et al. However, the clinical manifestations of these findings include hypotonia and sensory neurological dysfunction. Interestingly, among other anomalies combined with non-isolated TE, in addition to the cardiovascular system, neurological anomalies are the most common, such as porencephaly, holoprosencephaly, etc. Therefore, does this also support that some of the TE components are neuromuscular in origin? Martin et al.revealed that defects in neuronal development caused by the overexpression of Limk1 might lead to muscle atrophy and talipes equinovarus. Nevertheless, despite TE frequently occurring in neuromuscular abnormalities, no consistent neuromuscular abnormality is found in patients with isolated TE by electrophysiological examination or muscle biopsy [bib_ref] Congenital clubfoot: An electromyographic study, Bill [/bib_ref] [bib_ref] Muscle imbalance in the aetiology of idiopathic club foot. An electromyographic study, Feldbrin [/bib_ref] [bib_ref] Histologic and histochemical analysis of muscle specimens in idiopathic talipes equinovarus, Herceg [/bib_ref].
In most cases, pregnancy outcome largely depended on whether prenatal ultrasound combined severe malformations and the chromosomal results of prenatal diagnosis. Parents always chose to terminate the pregnancy for pCNV with a poor prognosis, such as 22q11.2 microdeletion syndrome. For fetuses with VOUS, parents always chose TOP when subsequent ultrasound findings worsened. A negative result by CMA may help to increase parents' confidence to continue the pregnancy. The pregnancy termination and live birth rates in cases with clinically significant variants detected by gene testing were found to be statistically significantly different from those with negative results (14/17, 82.4% vs. 41/147, 27.9%, p < 0.001). This suggests that the genetic test findings in fetuses of TE can affect parental decisions. In cases with clinically significant variants in genetic diagnosis, a small proportion of patients chose to continue the pregnancy after genetic counseling. In contrast, the majority chose termination of pregnancy due to the possible postnatal phenotype and poor prognosis. Therefore, it is necessary to perform genetic testing in patients with a prenatal ultrasound diagnosis of fetal TE, which may provide physicians and parents with more information about possible phenotypes after birth and to help them make pregnancy decisions and post-birth management.
There are some limitations to this study. First, this study is a retrospective study and the results lack of some crucial parameters, such as parental examination, family history, and results of genetic population screening tests such as spinal muscular atrophy. Second, fewer patients underwent karyotyping, and balanced translocations of chromosomes may have been missed. Third, although our study is the first prenatal study of singleton and twin pregnancies with TE using CMA in prenatal studies, the relatively limited number of twin pregnancies did not allow for definitive conclusions. We would like to collect more cases of twin pregnancies to further elucidate the potential differences between them. The final limitation is that TE may be related to a single gene or methylation. At the same time, we did not perform further tests such as whole exome sequencing (WES) and multiplex ligation-dependent probe amplification (MLPA) in our study.
# Conclusions
This study is the most comprehensive prenatal study using CMA to perform a detailed molecular analysis of cases of fetal TE diagnosed in a Chinese population. Ultrasound detection of fetal TE may be a diagnostic clue for some fetal syndromes, and ultrasound abnormalities were associated with increased risk of CMA findings in both singleton and twin pregnancies and both isolated and non-isolated TE. Therefore, when fetal TE is diagnosed by prenatal ultrasound, attention should be paid to whether it is combined with other structural abnormalities and to further genetic testing that could exclude these genetic disorders, which may help with diagnosis and counseling of prenatal TE.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/genes13091573/s1, [fig_ref] Table 1: Clinical features in fetuses with talipes equinovarus [/fig_ref] : Clinical characteristics of TE fetuses with VOUS. Data Availability Statement: The data that support the findings of this study are not publicly available as the information contained could compromise the privacy of research participants. Further inquiries can be directed to the corresponding author.
[fig] Figure 1: Flowchart of genetic analysis progression in cohort of fetuses with TE. [/fig]
[fig] Figure 2: Comparison of CMA detection rates of fetuses with talipes equinovarus. (A) Comparison of CMA detection rate of isolated and non-isolated TE in fetuses with singleton and twin pregnancies; (B) Comparison of CMA detection rate of unilateral and bilateral TE in fetuses with isolated and non-isolated TE. * p > 0.05; # p < 0.05. [/fig]
[fig] Funding: This research was funded by the sub-project of the National Key R&D Program (2021YFC2701002), the National Natural Science Foundation of China (81801461, 81873836, 81771594), the Natural Science Foundation of Guangdong Province (2019A1515012034), science and technology planning project of Guangdong Province, China (Chinese charity number: 2016A020218003), Science and Technology Program of Guangzhou China (Chinese charity number: 201607010341), and Guangzhou Institute of Pediatrics/Guangzhou Women and Children's Medical Center (NO: IP-2019-008). Institutional Review Board Statement: This study was approved by the Ethics Committee of Guangzhou Women and Children's Medical Center (Approval Code: [2015]278B01; Approval Date:2015-4-14). Informed Consent Statement: Not applicable. [/fig]
[table] Table 1: Clinical features in fetuses with talipes equinovarus. *: one of twin fetuses; #: likely pathogenic CNV; TE: talipes equinovarus; VSD: ventricular septal defect; CPCs: choroid plexus cysts; CLP: cleft lip and palate; HPE: holoprosencephaly; SGA: small for gestational age; iCTR: increased cardiothoracic ratio; FGR: fetal growth restriction; CH: cystic hygroma; TOP: termination of pregnancy. [/table]
[table] Table 2: Stratified analysis of CNVs detection and pregnancy outcome in TE. [/table]
|
The use of the concept of transition in different disciplines within health and social welfare: An integrative literature review
Aims: To continuing the quest of the concept of transition in nursing research and to explore how the concept of transition is used in occupational therapy, oral health and social work as well as in interdisciplinary studies in health and welfare, between 2003-2013. Design: An integrative literature review. Methods: PubMed, CINAHL, PsycINFO, DOSS, SocIndex, Social Science Citation Index and AMED databases from 2003-2013 were used. Identification of 350 articles including the concept of transition in relation to disciplines included. Assessment of articles are in accordance to Meleis' typologies of transition by experts in each discipline.Chosen key factors were entered into Statistical Package for the Social Sciences (SPSS).Results:Meleis' four typologies were found in all studied disciplines, except development in oral health. The health-illness type was the most commonly explored, whereas in social work and in occupation therapy, situational transitions dominated.K E Y W O R D Sliterature review, nursing theory, occupational therapy, oral health, social welfare, social work, theory-practice gap, transition
## | introduc ti on
In nursing research, transition has been described as the "passage from one life phase, condition, or status to another," as "periods in between fairly stable states" and as "processes that occur over time," which can be divided "into stages and phases" . From the work of various researchers, transition has been presented as a central concept in the discipline of nursing for the last four decades, [bib_ref] Role insufficiency and role supplementation: A conceptual framework, Meleis [/bib_ref] [bib_ref] Systematic review: Bridging the gap in RPN-to-RN transitions, Suva [/bib_ref]. During this period of time, the components of transition have been identified and described, expanded by one additional typology [bib_ref] Transitions: A central concept in nursing, Schumacher [/bib_ref] and the concept has been further extended and redefined [bib_ref] Experiencing transitions: An emerging middle-range theory, Meleis [/bib_ref] -all in nursing research. role insufficiency and role supplementation [bib_ref] Role insufficiency and role supplementation: A conceptual framework, Meleis [/bib_ref]. In addition to being educated in nursing, Meleis has a graduate education in sociology, as well as medical and social psychology [bib_ref] Transition theory, Im [/bib_ref] , which probably influenced the development of the transition theory. Meleis drew on theories that are now considered as classical sociological theoretical approaches, such as role theoryand symbolic interactionism. Additionally, [bib_ref] Transition: A literature review, Kralik [/bib_ref] noted anthropology as the discipline where transition historically has been described. This makes the concept of transition adaptable and interesting also to other disciplines in health and welfare such as occupational therapy, oral health and social work [bib_ref] Adulthood transitions in health and welfare: A literature review, Munck [/bib_ref].
Transition, in the mentioned research areas, is not highlighted as the main element of a theoretical framework, as in nursing. However, in sociology and in relation to illness, the phenomenon of transition has been captured by concepts such as career, change, illness trajectoriesand biographical disruption [bib_ref] Chronic illness as biographical disruption, Bury [/bib_ref].
Nevertheless, transition as a concept has been used in health and social welfare since at least the 1930s [bib_ref] The transition from institutional to social adjustment, Proehl [/bib_ref].
# | background
To demonstrate the diversity of the concept in nursing, four typologies of transition are described: developmental, situational, health-illness and organizational [bib_ref] Transitions: A central concept in nursing, Schumacher [/bib_ref].
The developmental category involves responses of stages in the life cycle such as parenthood, mainly focusing on the individual perspective. [bib_ref] Role insufficiency and role supplementation: A conceptual framework, Meleis [/bib_ref] emphasized two kinds of developmental transitions, which in particular are associated with health problems-the transition from childhood to adolescence and the transition from adulthood to old age. Situational transitions consist of various changes in educational and professional roles [bib_ref] Transitions: A central concept in nursing, Schumacher [/bib_ref]. Other studied situational transitions are changes in the family, for example widowhood, or transitions due to migration, homelessness and leaving an abusive relationship. Health-illness transitions focus on how individuals and families experience different illness contexts but also on transitions among levels of care during the course of illness [bib_ref] Transitions: A central concept in nursing, Schumacher [/bib_ref].
In their work, sitions as the fourth type of transition, which represents "changes in the wider social, political, or economic environment or by intraorganizational changes in structure or dynamics" (p. 21).
In an attempt to broaden the knowledge of the concept, [bib_ref] Transition: A literature review, Kralik [/bib_ref] reviewed articles from diverse professional fields published between 1994-2004, where transition was used. They described these as "health literature" but also included a social focus. In their review, only research with qualitative methodologies was included.
The concept of transition, as described in nursing, is complex and multidimensionally identified as awareness, engagement, change and difference, time span, critical points and events [bib_ref] Experiencing transitions: An emerging middle-range theory, Meleis [/bib_ref]. Future research might identify other factors important for understanding the complexity of transition to prevent and minimize risks for unhealthy transitions. This research will probably be found in other disciplines in health and welfare, which provides an opportunity to discover transition processes in other diverse populations and contexts as suggested [bib_ref] Experiencing transitions: An emerging middle-range theory, Meleis [/bib_ref]. The concept of transition embodies an important issue and seems currently to be commonly used in various fields outside nursing. However, to the best of our knowledge, the concept is not explored and reviewed in relation to other disciplines in the field of health and social welfare.
## | aim
To continuing the quest of the concept of transition in nursing research and to explore how the concept of transition is used in occupational therapy, oral health and social work, as well as in interdisciplinary studies in health and welfare, between 2003-2013.
Research questions:
- Are the four typologies (developmental, situational, health-illness and organizational) elaborated by [bib_ref] Experiencing transitions: An emerging middle-range theory, Meleis [/bib_ref] applicable for the studied disciplines and if, which differences and/or similarities are found?
- Where contexts are the concept of transition used in the various disciplines?
- Who underwent the transition and from what perspectives are the transition regarded in different disciplines?
## | me thod
## | design
This integrative literature review [bib_ref] The integrative review: Updated methodology, Whittemore [/bib_ref] of transition as a concept was used to examine the literature in various disciplines in health and welfare.
## | inclusion and exclusion
The primary criteria for inclusion in this review were scientific articles published between 2003-2013, original papers written in English and articles that illuminated the concept of transition as a part of the results in the context of nursing, oral health, occupational therapy, social work, or interdisciplinary studies in health and social welfare. Exclusion criteria were if the participants were predominantly <18 years old, if it was a review article, if there was a meeting abstract or if the abstract was missing, if the article did not focus on transition and if the transition concerned another area than the above-mentioned disciplines or did not focus on the outcome of the person(s) in transition.
## | literature search
This literature review used PubMed, CINAHL, PsycINFO, DOSS, SocIndex, Social Science Citation Index and AMED databases. The following Medical Subject Heading (MeSH) terms were used in the search: Transition AND nurs*, OR transition AND car*, OR transition AND social work*, OR transition AND social car*, OR transition AND occupational therap*, OR transition AND oral health*, OR transition AND dental hygien*. To make the data manageable, each search was performed for both title and abstract. If the number of articles in one search for abstracts (e.g., transition AND nurs*) exceeded 300, the number of hits in the title was selected instead.
## | search outcome
After duplicates (N = 2,365) were removed, 2,523 unique references were found in the databases and were screened (PubMed 580, CINAHL 132, PsycInfo 255, DOSS 119, SocIndex 381, Social Sciences Citation Index 882, AMED 174).
## | article selection
All retrieved titles and abstracts were screened to determine the eligibility by an interdisciplinary review team of 11 researchers, all an "expert" in at least one of the studied disciplines. The interrater reliability was tested in two steps. First, 50 articles were assessed using a template as fulfilling the criteria, doubtful fulfilling the criteria or as not meeting the criteria. Good concurrence was reached in 68%. In step two, after a discussion of the reviewed articles and the differences in grading, 25 articles were assessed in a similar way and concurrence was reached in 84%.
After this, the abstracts (N = 2,523) were distributed to the review team (200-250/reviewer) in accordance with each researcher's expertise to be assessed according to the inclusion and exclusion criteria. Most excluded articles were primarily based on transitions outside the health and welfare sector. After this screening, 475 fulltext articles were assessed for eligibility and 125 of them were excluded, resulting in 350 articles included for analysis .
# | analysis
To make the analytic work more systematic and manageable, chosen key factors were entered into Statistical Package for the Social Sciences (SPSS), according to variables included in the analysis . The results are presented regarding the typology of transition, the context of transition in studied disciplines, who performed the transition and from what perspective the transition was regarded. Presented results are exemplified by one or two articles.
## | ethics
Research Ethics Committee approval.
## | re sults
Of the 350 articles included, most studies were conducted in North America (60%) and Europe (22%). In nursing (N = 106) and occupational therapy (N = 39), the pattern for methods used in research was similar . Almost half of these studies used qualitative methods (mainly interviews), about a quarter used quantitative methods (mostly questionnaires), a fifth part was described as "theoretical F I G U R E 1 Flow diagram of the systematic review process summaries," and approximately 10% combined qualitative and quantitative methods. In social work (N = 42), the qualitative methods dominated and there were less theoretical summaries. In oral health (N = 11), no studies had a qualitative design and more than 80% used quantitative design (mostly questionnaires). In interdisciplinary research, (N = 152) quantitative and qualitative methods were used equally and fewer studies were described as theoretical summaries compared with other studied research areas. describes different aspects of the concept of transition in relation to the five disciplines included in this study.
## | typologies of transition
All articles considered as research on transition could be coded into one of the four types of transition defined by Meleis and her colleagues except five (1.4%), [bib_ref] Transitions: A central concept in nursing, Schumacher [/bib_ref]. In all, health-illness transition was the type of transition (41.7%) most often described, followed by situational (36.9%), developmental transition (10.3%) and organizational (9.7%). In oral health, interdisciplinary studies and nursing, the health-illness type of transition was the most common (54.5%, 51.3% and 41.5%, respectively), whereas in social work and in occupational therapy, situational transitions dominated (64.3% and 46.2%, respectively) .
Health-illness transitions were most often due to removals in the healthcare system [bib_ref] Change in self-reported oral health in relation to use of dental services..., Brennan [/bib_ref] [bib_ref] Communication and coordination during transition of older persons between nursing homes and..., Kirsebom [/bib_ref] , or the transition occurred in specialist care, such as for HIV [bib_ref] Best practices in transitioning youth with HIV: Perspectives of pediatric and adult..., Fair [/bib_ref] and in palliative care, or during a time period between in-/outpatient care and home [bib_ref] Hospital to home health care transition: Patient, caregiver and clinician perspectives, Foust [/bib_ref].
Situational transitions mainly regarded conditional changes focusing on caregivers [bib_ref] Self-care needs of caregivers dealing with stroke, Cook [/bib_ref] [bib_ref] Transitions in men's caring identities: Experiences from home-based care to nursing home..., Eriksson [/bib_ref] , out-of-home care such as foster care [bib_ref] Resilience of youth in transition from out-of-home care to adulthood, Daining [/bib_ref] , becoming a professional [bib_ref] The nature and implications of support in graduate nurse transition programs: An..., Johnstone [/bib_ref] [bib_ref] The dawn is too distant: The experience of 28 social work graduates..., Yan [/bib_ref] , professional practices [bib_ref] Procedures when young people leave care-Views of 111 Swedish social services managers, Höjer [/bib_ref] or retiring [bib_ref] The first steps into the third age: The retirement process from a..., Jonsson [/bib_ref] [bib_ref] The influence of changes in dental care coverage on dental care utilization..., Manski [/bib_ref].
Organizational transitions illuminated, for instance, programs or strategies to facilitate the changeover from newly qualified professionals to practice [bib_ref] Application of guided imagery to facilitate the transition of new graduate registered..., Boehm [/bib_ref] , social changes in the societyand improvements, or changed working roles and routines [bib_ref] Social workers' experiences of virtual psychotherapeutic caregivers groups for Alzheimer's, Parkinson's, Stroke,..., Damianakis [/bib_ref]
## | contexts of transitions
Among the disciplines, the prominent context of transition varied.
"Between different contexts" was the most common category in nursing (39.6%), social work (42.9%) and in interdisciplinary research (31.6%), whereas "in daily life" dominated occupational therapy (35.9%). In oral health, almost two out of three studies concerned transition in "outpatient care" . To further understand how F I G U R E 2 Key factors from the included articles which formed the variables included in the analysis. (a) Other: type of transition other than those described by [bib_ref] Experiencing transitions: An emerging middle-range theory, Meleis [/bib_ref] or included in more than one type. the concept of transition is used, these three most commonly used contexts are elaborated for the studied disciplines.
## | between different contexts
Social work included most articles focusing on transition between contexts. Most of these articles concerned youngsters or young adults and leaving care; "leaving care" could mean the transition between foster care and residential care or between out-of-home care and independent living [bib_ref] Resilience of youth in transition from out-of-home care to adulthood, Daining [/bib_ref] [bib_ref] Young people leaving care in Sweden, Höjer [/bib_ref] , or it could mean young people's transition to homelessness forced by a premature home-leaving [bib_ref] Homeless young people, families and change: Family support as a facilitator to..., Mayock [/bib_ref]. Such situational changes were sometimes explic- [bib_ref] Smoothing things over: The transition from pediatric to adult care for kidney..., Remorino [/bib_ref]. Other transitions described in nursing were moving from institutional care to community care [bib_ref] Ready or not: Transitioning from institutional care to community care, Fields [/bib_ref]. For instance, communication and coordination were studied during the transition of older persons between nursing homes and hospitals [bib_ref] Communication and coordination during transition of older persons between nursing homes and..., Kirsebom [/bib_ref]. Becoming a nurse [bib_ref] Being a nurse after having been a nursing student-worker: An approach of..., Costa [/bib_ref] or advancing in the nursing career [bib_ref] There really is a difference: Nurses' experiences with transitioning from RNs to..., Delaney [/bib_ref] was other transitions between contexts.
In occupational therapy, transition between contexts concerned, for instance, the retirement process, that is from working life to retirement [bib_ref] The first steps into the third age: The retirement process from a..., Jonsson [/bib_ref]. Other examples were the transition from secondary education into further education and/or payed employment for young adults with physical disabilities [bib_ref] Relationship of therapy to postsecondary education and employment in young adults with..., Bjornson [/bib_ref] and the transition from hospital to home.
For oral health, only one article was found highlighting the transition from working life to retirement and its effects on dental care use [bib_ref] The influence of changes in dental care coverage on dental care utilization..., Manski [/bib_ref].
## | in daily life
Daily life included transitions at home and at the workplace. For occupational therapy, the context in daily life was the most common. In nursing, the most common transition in daily life considered working life; this transition could include the move from the status of a recent graduate to that of an experienced nurse [bib_ref] Preparing transition-age students with high-functioning autism spectrum disorders for meaningful work, Lee [/bib_ref] [bib_ref] How first experiences influence newly qualified nurses, Tapping [/bib_ref] and it could include a promotion or a changed role during a transfer from one professional role to another [bib_ref] Role transition from mental health nurse to IAPT high intensity psychological therapist, Robinson [/bib_ref] [bib_ref] Role transition from caregiver to case manager-Part II, Schmitt [/bib_ref].
In interdisciplinary research, transition in daily life mostly concerned the management of long-term diseases such as diabetes [bib_ref] Young women with type 1 diabetes' management of turning points and transitions, Rasmussen [/bib_ref] ,
## | outpatient care
Transition in outpatient care for oral health most often considered changes after different oral health treatments, described by self- [bib_ref] Role transition in primary care settings, Holt [/bib_ref] [bib_ref] Recklessness, rescue and responsibility: Young men tell their stories of the transition..., Reeves [/bib_ref].
For occupational therapy, transition in outpatient care was only found in one article evaluating information of self-care needs for patients at home after stroke [bib_ref] Self-care needs of caregivers dealing with stroke, Cook [/bib_ref].
## | who underwent the transition?
Regarding who went through the transition, the patient-client was the most common category in all disciplines, except for social work . In interdisciplinary studies, this was the case in eight out of ten articles, while about half of the studies for oral health, occupational therapy and nursing belonged to this category.
In interdisciplinary studies, patients-clients underwent the youths' transition to adulthood [bib_ref] Adolescence and young adulthood in Spina Bifida: Self-report on care received and..., Ruck [/bib_ref] , elders transition in changing housing [bib_ref] Stability and changes in living arrangement among unmarried older persons: The influence..., Choi [/bib_ref] [bib_ref] Transition of new residents to long-term care: Basing practice on residents' perspective, Heliker [/bib_ref] , menopause transition [bib_ref] Sexual desire during the menopausal transition and early postmenopause: Observations from the..., Woods [/bib_ref] and palliative care patients' experiences [bib_ref] Transition towards end of life in palliative care: An exploration of its..., Larkin [/bib_ref].
In oral health, patients reporting outcomes using different dental services [bib_ref] The effectiveness of out-of-hours dental services: I. pain relief and oral health..., Anderson [/bib_ref] [bib_ref] Change in self-reported oral health in relation to use of dental services..., Brennan [/bib_ref] , patients' experiences of oral health related to quality of life after dental treatments [bib_ref] The transition from breast cancer 'patient'to 'survivor, Allen [/bib_ref] [bib_ref] Assessing the minimally important difference in the oral impact on daily performances..., Tsakos [/bib_ref] and changes in dental coverage status in relation to oral status [bib_ref] The influence of changes in dental care coverage on dental care utilization..., Manski [/bib_ref] were found.
For occupational therapy, the patient-client experienced, for example, quality in rehabilitation [bib_ref] Quality in rehabilitation after a working age person has sustained a fracture:..., Lindahl [/bib_ref] , accommodation transitions for individuals with acquired brain injury [bib_ref] Accommodation outcomes and transitions following community-based intervention for individuals with acquired brain..., Sloan [/bib_ref] , occupational transitions after injury or disease [bib_ref] Evaluating the support needs of injured workers in managing occupational transitions after..., Shaw [/bib_ref] [bib_ref] Perceptions of possibilities of returning to work with chronic musculoskeletal disorders, Švajger [/bib_ref] , mental illness [bib_ref] Rip that book up, I've changed': Unveiling the experiences of women living..., Mckay [/bib_ref] and transition from the hospital to home care [bib_ref] Stroke patients' experience with the Australian health system: A qualitative study, White [/bib_ref].
In social work, the "person" was the predominating category, followed by "the professional" and "several perspectives." The "person" as a category most often referred to young adults who were, or had been in, out-of-home care, such as foster care [bib_ref] Supporting youth in the transition from foster care: Formal and informal connections, Collins [/bib_ref] , followed by young people in different types of socially difficult situations, such as homelessness [bib_ref] Homeless young people, families and change: Family support as a facilitator to..., Mayock [/bib_ref].
Transitions were also viewed both from clients (mostly youth) and
professionals [bib_ref] The school to work transition for young people in state care: Perspectives..., Tilbury [/bib_ref].
In nursing, almost three of ten studies concerned the transition of professionals, which was rarely or never the case in the other disciplines. Often, this regarded the transition from nursing student to staff nurse [bib_ref] Transition from nursing student to staff nurse: A personal reflection: Practice placements..., Pearson [/bib_ref] [bib_ref] Graduate nurses' transition and integration into the workplace: A qualitative comparison of..., Walker [/bib_ref] or from one nursing role to another [bib_ref] There really is a difference: Nurses' experiences with transitioning from RNs to..., Delaney [/bib_ref] [bib_ref] Role transition from mental health nurse to IAPT high intensity psychological therapist, Robinson [/bib_ref]. Moreover, in nursing, 8.5% of those who performed the transition were relatives. This category was rarely mentioned in other disciplines. In occupational therapy, students and employees/job seekers (12.8% for each group) more often were those who performed the transition compared with those in other disciplines .
## | from which perspective was transition seen?
Depending on the disciplines, the perspectives on transition varied. Still, in total, the viewpoint most often was from the patient's/ client's perspective (28.0%) followed by the person's perspective (13.1%) .
In occupational therapy, the patient/client perspective and the professional perspective were almost equal, while in oral health, more than half of the studies (54.5%) considered the patient/client perspective. This perspective was the most common also in interdisciplinary, although to a less degree (38.2%).
In nursing, however, the professional perspective was twice as common as the second most used perspective, which was from the patient's/client's point of view. In social work, transition from the perspective of the person was used in 38.1% of the studies. In oral health, studies focused on the patients' view of oral health [bib_ref] Oral health-related quality of life of stroke survivors on discharge from hospital..., Mcmillan [/bib_ref] , oral treatments [bib_ref] The effectiveness of out-of-hours dental services: I. pain relief and oral health..., Anderson [/bib_ref] , dental service and coverage [bib_ref] The influence of changes in dental care coverage on dental care utilization..., Manski [/bib_ref].
In occupational therapy, transition viewed from the patient/client's perspective considered, for instance the clients' perspective on the return to work [bib_ref] Moving from full-time healing work to paid employment: Challenges and celebrations, Bergmans [/bib_ref] [bib_ref] Perceptions of possibilities of returning to work with chronic musculoskeletal disorders, Švajger [/bib_ref] , recovery patterns (Prvu [bib_ref] Analyzing change in recovery patterns in the year after acute hospitalization, Bettger [/bib_ref] and the experience of mental illness [bib_ref] Rip that book up, I've changed': Unveiling the experiences of women living..., Mckay [/bib_ref]. Transition viewed from the professional perspective in this discipline focused on home-based rehabilitation [bib_ref] Rehabilitation in home care is associated with functional improvement and preferred discharge, Cook [/bib_ref] and school involvements [bib_ref] The role of occupational therapy in providing person-centred transition services: Implications for..., Michaels [/bib_ref].
In interdisciplinary studies, articles taking the patient's/client's perspective explored, for instance, the menopause transition (Smith-DiJulio,, patients' transition between care settings [bib_ref] Integrated transitional care: Patient, informal caregiver and health care provider perspectives on..., Toscan [/bib_ref] or transitions to a community-based setting [bib_ref] Transitioning residents from nursing facilities to community living: Who wants to leave, Nishita [/bib_ref].
In social work, many studies taking the perspective of the person concerned young people in out-of-home care or the leaving care processes [bib_ref] Everything was strange and different': Young adults' recollections of the transition into..., Reimer [/bib_ref]. In many of these cases, the youngsters went through a double transition as their living conditions changed in parallel with their transition from childhood to "instant adulthood" [bib_ref] I remember thinking, why isn't there someone to help me? Why isn't..., Rogers [/bib_ref].
In nursing, compared with other disciplines, the perspective was more often taken from the professionals and focused mostly on the transition from nursing student to occupational nurse [bib_ref] Outcomes of newly practicing nurses who applied principles of holistic comfort theory..., Goodwin [/bib_ref] , as well as when shifting from one role to another, such as becoming CEO/president of their organization [bib_ref] Making the transition: an interview with nurse chief executive officers at Catholic..., Patton [/bib_ref] or becoming a nurse specialist [bib_ref] Role transition from mental health nurse to IAPT high intensity psychological therapist, Robinson [/bib_ref].
Of those patients/clients who underwent the transition (N = 209), the transition was regarded from the patient's/client's perspective in 45% of the cases (not shown in . This relationship was the case in the disciplines of nursing (43.1%), occupational therapy (47.6%) and interdisciplinary research (44.2%). In oral health, most articles (85.7%) showed this relationship, whereas for social work, the corresponding figures were 20%.
The transition of patients/clients was regarded from a professional perspective in almost one fourth (22%) of the studies (not shown in . In social work, all these studies considered the social worker's role in clients' transition in an institutional framework and focused on interventions [bib_ref] Social work and transitions of care: Observations from an intervention for older..., Fabbre [/bib_ref] , procedures [bib_ref] Procedures when young people leave care-Views of 111 Swedish social services managers, Höjer [/bib_ref] or assessments (B. R. [bib_ref] Transitioning from group care to family care: Child welfare worker assessments, Lee [/bib_ref].
Among all examined articles (N = 350), 37 reported on the professional who underwent the transition (not shown in . In 95% of these, the transition was also described from a professional perspective. All except three were found in nursing.
## | d iscuss i on
This analysis continues the quest of [bib_ref] Transition: A literature review, Kralik [/bib_ref] beyond the mere scoop of nursing research by exploring how the concept of transition is used in occupational therapy, oral health and social work, as well as in nursing and interdisciplinary studies. The main result is that all included articles, except five, could be categorized according to [bib_ref] Experiencing transitions: An emerging middle-range theory, Meleis [/bib_ref] and that all types of transitions were found in all disciplines included, except for development, which was not found for oral health. Those five articles categorized as another typology than those described by [bib_ref] Experiencing transitions: An emerging middle-range theory, Meleis [/bib_ref] were not further analysed in our study. However, a new transition typology has recently been found in a review, named as the lifestyle transition [bib_ref] Adulthood transitions in health and welfare: A literature review, Munck [/bib_ref] , indicating that the concept of transition in adulthood is developing.
In relation to the literature review by [bib_ref] Transition: A literature review, Kralik [/bib_ref] , our study indicates that the concept of transition has been used in a larger extent during 2003-2013 compared with the previous 10 years. This could be explained by differences in study design. Our study shows that the concept of transition is used in both qualitative and quantitative studies, as well as in theoretical summaries.
In this review, nursing research is the discipline presenting most articles on transition, except for interdisciplinary studies, which, however, also include the research area of nursing. This dominance could be explained by the fact that transition has been used in nursing research for a long time and has developed as an important concept in nursing [bib_ref] Experiencing transitions: An emerging middle-range theory, Meleis [/bib_ref].
In nursing, the health-illness type of transition dominates in combination with the transition of patients regarded from the professionals' perspective. This finding is interesting because the notion of healthy transition processes versus unhealthy has been described as the core of the transition and of special concern for nurses. [bib_ref] Experiencing transitions: An emerging middle-range theory, Meleis [/bib_ref] argued that there are two reasons for this: 1) nurses often are the primary caregivers of people undergoing transitions associated with health problems and 2) nurses tend to be involved in preparing clients for impending transitions and to be those "who facilitate the process of learning new skills related to clients' health and illness experiences" . Another reflection is that the most recently added typology-organizational transition-constitutes the third largest transition type ( Although Meleis and her colleges introduced and developed the concept of the transition in relation to nursing research and practice, they have presented transition as "a multiple concept embracing the elements of process, time span and perception" . Such broad aspects are useful in various disciplines and with these disciplines, the current study demonstrated how the concept of transition is used in occupational therapy, oral health and social work, as well as in interdisciplinary research in health and welfare. However, the concept is used less in occupational therapy, oral health and social work compared with nursing. This might be a consequence of these disciplines' lack of a developed theory using transition as a core concept. Moreover, in disciplines other than nursing, such as social work, the influence from sociology might imply that other concepts such as career, changeand
trajectory (e.g.,are used in parallel to transition.
Although the concept of transition is used in all studied disciplines, there are interesting differences, such as that in oral health, organizational transitions are the second most frequently used type of transition. However, in this discipline, the total number of studies is few and the concept seems to be used more often in the last five years of the study period.
Further, the results show considerable differences between disciplines regarding which contexts the examined transition occurred in, even when the same category was used. For instance, in social work and nursing, the category "between different contexts" was used differently. In nursing, the in-between stages often appeared in the hospital context, while in social work, "between contexts" reflected the transferal from one "home" to another, commonly referring to young people in foster care or residential treatment. These differences seemed to reflect the different type of work where social workers and nurses are engaged in (for instance child welfare and nursing, respectively). Similarly, the context of "daily life" was most commonly used in occupational therapy and "outpatient care" was most frequent in oral health, corresponding to the various working conditions of these professionals.
The utility of transition, as a concept in research in health and welfare, is further strengthened by how the diversity in professional assignment is also noticeable by the category "who underwent the transition." For all disciplines in health care (i.e. nursing, occupational therapy and oral health), the "patient" was the one who went through the transition most often. This finding corresponds to these professionals' work in healthcare institutions, such as hospitals or outpatient clinics. In social work, on the other hand, the category of "person" reflects the work with people in the community. Consequently, this study demonstrates the usefulness of the concept of transition for studies in various disciplines which is in line with , who raised the importance of "understanding the transition of the process to assist people to move through it" in a healthy direction. Thus, the use of the concept of transition is important for all professionals to understand various changes in health and well-being, as well as in relation to professional, personal or organizational development.
Nevertheless, this study also shows an interesting disciplinary difference, which could not be explained from variations of working conditions. Regarding which perspective the transition was viewed, the "patient/client" or the "person" was the most prominent in occupation therapy, oral health and social work. However, in nursing research, the transition process was regarded mainly from the point of view of the professional. This self-reflecting focus on the nurse as a professional could not be understood clearly by the specific working conditions for nurses compared with other professionals. However, it might be explained by the long tradition of using transition in research and, in particular, the concept's usefulness for providing "directions for nursing practice," as well as for the "development of nursing therapeutics" . In other studied disciplines, such a theoretical framework is lacking.
# | limitation
Adoption of a systematic review approach enhances the scientific rigour of a review. However, this review was performed by systematic steps guided by a theoretical framework and not a systematic review, which may be a limitation. Several researchers from different disciplines were included in the analysis, which could be seen as a possible limitation concerning inter-reliability, but also a strength related to the study aim. Another possible limitation is that only four disciplines are included in the study. In health and social welfare, also other disciplines such as medicine, psychology and social pedagogy could have been considered. This review included articles published at latest 2013, an extension by including also the more recent years would probably have further enhanced the value of the analysis. At the same time, our study points to the need to continuously study concepts such as transition to investigate how it develops in disciplines where the term is well established as well as its dissemination to and use in other subject fields such as medicine and psychology to name a few important disciplines in health and social welfare not yet included in reviews on the concept of transition.
## | con clus ions
The results from this study show that although there were differences between various disciplines, few articles could not be categorized in the four types of transitions identified and explored in nursing. Further, the most recently added typology-organizational transition-was in our data and was used almost as much as developmental transition. This outcome indicates that the decision to extend the concept with the "new" typology was well founded.
Our main conclusion is that the concept of transition appears useful in health and social welfare research in a broad perspective and in diverse ways. This implies its flexibility but also the usefulness of the concept to explore various changes.
## Co n fli c t o f i nte r e s t
No conflict of interest has been declared by the authors.
## O rci d
Ulrika Lindmark https://orcid.org/0000-0003-2786-707X
Pia H. Bülow https://orcid.org/0000-0001-7341-945X Jan Mårtensson https://orcid.org/0000-0002-7406-8732
[fig] For instance ,: Suto (2009) studied migrant women and resettlement,and Sabata, Bruce, and Sanford (2006) investigated home health discharge planning for transition to work for people with disabilities. [/fig]
|
Immunogenicity after two doses of inactivated virus vaccine in healthcare workers with and without previous COVID‐19 infection: Prospective observational study
Vaccines have been seen as the most important solution for ending the coronavirus disease 2019 (COVID-19) pandemic. The aim of this study is to evaluate the antibody levels after inactivated virus vaccination. We included 148 healthcare workers (74 with prior COVID-19 infection and 74 with not). They received two doses of inactivated virus vaccine (CoronaVac). Serum samples were prospectively collected three times (Days 0, 28, 56). We measured SARS-CoV-2 IgGsp antibodies quantitatively and neutralizing antibodies. After the first dose, antibody responses did not develop in 64.8% of the participants without prior COVID-19 infection. All participants had developed antibody responses after the second dose. We observed that IgGsp antibody titers elicited by a single vaccine dose in participants with prior COVID-19 infection were higher than after two doses of vaccine in participants without prior infection (geometric mean titer: 898 and 607 AU/ml). IgGsp antibodies, participants with prior COVID-19 infection had higher antibody levels as geometric mean titers at all time points (p < 0.001). We also found a positive correlation between IgGsp antibody titers and neutralizing capacity (r s = 0.697, p < 0.001). Although people without prior COVID-19 infection should complete their vaccination protocol, the adequacy of a single dose of vaccine is still in question for individuals with prior COVID-19. New methods are needed to measure the duration of protection of vaccines and their effectiveness against variants as the world is vaccinated. We believe quantitative IgGsp values may reflect the neutralization capacity of some vaccines. K E Y W O R D S coronaVac, immunogenicity, inactivated virus vaccine, neutralizing antibody, SARS-CoV-2 antibody J Med Virol. 2022;94:279-286.wileyonlinelibrary.com/journal/jmv
have previously been infected with COVID-19. [bib_ref] Effect of previous SARS-CoV-2 infection on humoral and T-cell responses to single-dose..., Prendecki [/bib_ref] [bib_ref] Binding and neutralization antibody titers after a single vaccine dose in health..., Saadat [/bib_ref] [bib_ref] Antibody responses to the BNT162b2 mRNA vaccine in individuals previously infected with..., Ebinger [/bib_ref] CoronoVac is an inactivated virus vaccine developed by Sinovac Life Sciences in early 2020. After demonstrating the safety and efficacy of the vaccine in Phase 1/2 studies, [bib_ref] Safety, tolerability, and immunogenicity of an inactivated SARS-CoV-2 vaccine in healthy adults..., Zhang [/bib_ref] [bib_ref] Safety, tolerability, and immunogenicity of an inactivated SARS-CoV-2 vaccine (CoronaVac) in healthy..., Wu [/bib_ref] Phase 3 studies were initiated in Brazil, Indonesia, Chile, and Turkey. CoronaVac efficacy was found to be 83.5% in the Phase 3 study in Turkey. [bib_ref] Efficacy and safety of an inactivated whole-virion SARS-CoV-2 vaccine (CoronaVac): interim results..., Tanriover [/bib_ref] In an April 2021 report, the World Health Organization (WHO) stated that the CoronaVac vaccine is effective in preventing COVID-19, and they approved it for emergency use in June 2021.Individuals and clinicians are trying to understand the effectiveness of vaccines and how long the protection lasts as vaccination rates rise. Assessments of vaccine effectiveness are based on realworld data that takes time. In vaccination research with influenza, smallpox, and polio, it has been stated that the neutralizing antibody response predicts vaccine protection, and it is known to play an important role in severe acute respiratory syndrome coronavirus 2. [bib_ref] Commercial serology assays predict neutralization activity against SARS-CoV-2, Suhandynata [/bib_ref] However, neutralizing antibody tests are not widely used because of technical and financial issues. Vaccines induce an immune response against viral spike protein. Anti-spike antibodies are produced by the immune system and may serve as an indicator of protection.The first goal of this study is to follow up the serological responses to the inactivated virus vaccine, namely CoronaVac, among healthcare workers (HCWs) who had or did not have past COVID-19.
The secondary goal is to investigate the impact of the level of quantitative antibody on neutralization capacity. Inclusion criteria for participants with prior COVID-19 infection were SARS-CoV-2 polymerase chain reaction (PCR) positivity in the preceding months, as well as contact and symptomatic disease in any period before vaccination.
Participants who had not previously been infected with COVID-19 were eligible if they had not SARS-CoV-2 PCR positivity, no history of quarantine, no history of therapy against COVID-19 and SARS-CoV-2 IgGsp levels <50 AU/ml in their serum before vaccination.
Antibody levels were measured just before the first dose (Day 0), just before the second dose (Day 28), and 28 days after the second dose (Day 56).
Serological testing for antibodies to the receptor of the S1 subunit of the viral spike protein (IgGsp) was performed with the Abbott Architect SARS-CoV-2 IgG Quant II (Abbott).
Neutralizing antibodies from 31 participants were tested. This After incubation, supernatants were discarded and cells were supplemented with DMEM and 2% methylcellulose solution. After 4 days at 37°C, supernatants were removed and the six-well plates were fixed and inactivated using 4% paraformaldehyde and stained with crystal violet. Serum dilutions with a PRNT of 50% (PRNT50) are referred to as titers. Unless stated otherwise, cut-off titers were set at <1:20.
A standardized questionnaire asking about the demographic characteristics of the participants-age, gender, socioeconomic status, marital status, comorbidity, smoking, and alcohol use-and the use of paracetamol and/or nonsteroidal anti-inflammatory drugs (NSAIDs) was filled out by all participants with the guidance of an infectious diseases specialist. Socioeconomic status was determined using the income and expenditure self-report. The use of paracetamol and/or NSAIDs was defined for both chronic use and 24 h before and after vaccination.
To evaluate the body mass index (BMI) of participants, their heights and weights were measured on a standard scale using a measuring instrument. The Pittsburgh sleep quality index (PSQI) was applied to assess sleep states, and those with PSQI ≥ 5 were classified as having poor quality sleep. The disease severity of the previously infected participants was categorized into three groups according to WHO classification: mild, moderate, and severe.In any of the participants during follow-up, there was no symptomatic disease, though this is based on the participants' self-reports, as SARS-CoV-2 PCR tests were not administered.
The design of this study was followed using the Strengthening the Reporting of Observational Studies in Epidemiology checklist.
This study was approved by the Baskent University Institutional Review Board and Ethics Committee (Project no: KA21/51) and was supported by the Baskent University Research Fund.
## | statistical analyses
Antibody levels were presented as GMTs and confidence intervals.
Antibody levels were converted to base 10 logarithms and used in statistical analysis for intergroup comparison. By testing the compatibility of numerical data to a normal distribution, parametric data were evaluated using Student's t tests in paired comparison and Mann-Whitney U tests for those who did not have parametric properties. Analysis of the nominal data χ 2 test or Fisher's exact test was used. When changes of numerical variables were obtained with more than two measurements in a single group, the Friedman test was used if the variables were not normally distributed, and analysis of variance was used for repeated measures. Spearman's correlation coefficient was used to measure the strength of association between antibody levels and neutralized antibody results. Statistical analysis was done with IBM ® SPSS© 25 software. Situations below 5% of the type 1 error level were interpreted as statistically significant.
# | results
## | demographic characteristics
Of the 148 HCW participants included in the study, 104 (70.3%) were women, and the median age was 39 years (min 22-max 64).
Forty-five of the participants (30.4%) had comorbidity; the most prevalent being thyroid dysfunction (13/148, 8.8%). This was followed by hypertension (11/148, 7.4%) and rheumatic disease (10/148, 6.8%). The demographic characteristics of the participants are given in .
## | evaluation of iggsp antibody titers
## | evaluation neutralizing antibodies
The serological response to the vaccine, natural immune response, and correlation with IgGsp antibody titers was evaluated. Sera from 31 participants were evaluated by neutralization assays and a response equal to or greater than 1/160 dilution was detected (IgGsp antibody titers ranged from 70 to 1526 AU/ml). Neutralizing antibody responses without vaccine were detected 1/320 dilution for those who had past COVID-19 infection and the lowest antibody titers.
However, the lowest antibody titers in second-dose vaccinated participants who had not been previously infected with COVID-19 had 1/160 dilution of neutralizing capacity . We found a positive correlation between IgGsp antibody titers and neutralizing capacity- CoronaVac has been shown to be effective in preventing symptomatic or severe illness. [bib_ref] Efficacy and safety of an inactivated whole-virion SARS-CoV-2 vaccine (CoronaVac): interim results..., Tanriover [/bib_ref] It was also found to be effective According to current evidence, the natural immune response can last up to 9 months after COVID-19 infection. [bib_ref] SARS-CoV-2 antibody dynamics and B-cell memory response over time in COVID-19 convalescent..., Achiron [/bib_ref] We also demon-
# | conclusion
It is well known that not only spike antibodies but also T-and B-cell timing. This is a pioneering study, and more comprehensive research on the subject is required.
## Conflict of interests
The authors declare that there are no conflict of interests.
# Data availability statement
The data that support the findings of this study are available on request from the corresponding author (Dr. Tuğba Y. Yalçın). The data are not publicly available due to patient privacy.
## Orcid
[fig] F: I G U R E 1 (A) Anti-SARS-CoV-2 IgGsp responses on Days 0, 28, 56 in groups with and without COVID-19 history. (B) Showing the geometric mean titers with line graphs of COVID-19 previously infected and not-infected groups at all measurement times.^*Comparison between groups. *Not previously infected participants intergroup comparison. **Previously infected participants intergroups comparison. COVID-19, coronavirus disease 2019; IgG, immunoglobulin G; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2 I G U R E 2 (A) Anti-SARS-CoV-2 IgGsp responses by time after disease in the group with prior COVID-19. (B) Showing the geometric mean titers with line graphs of the groups according to by the time after disease with prior COVID-19 at all measurement times. COVID-19, coronavirus disease 2019; IgG, immunoglobulin G; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2 against several SARS-CoV-2 variants, albeit some variants drastically reduced neutralizing antibody efficacy. 15,16 Neutralizing antibodies have been shown to be highly predictive of SARS-CoV-2 infection protection. 17 According to our findings, a single dose of CoronaVac did not produce a reliable antibody response in those who had not previously been infected (n = 48/74, 64.8%), whereas two doses of the vaccine produced antibodies 28 days after the second dose. Individuals without prior COVID-19 infection had lower antibody responses during the whole study period than those in the previously COVID-19-infected group. The SARS-CoV-2 Quant Assay user manual estimates a 1050 AU/ml SARS-CoV-2 IgG titer, which corresponds to a 95% probability of being at or above the 1:80 neutralization dilutions. However, we detected higher neutralizing capacity at lower quantitative antibody values. This could be due to nonstandardized neutralizing antibody measurements and working with not the same strain. The strain used in our study is the most widely circulating one in our country. It still remains unclear how long immunity lasts after COVID-19 infection, the protective level of antibody titer protect against reinfection, and when people should be vaccinated. These are all important questions in terms of the future trajectory of the SARS-CoV-2 virus. [/fig]
|
Structural insights into the mechanism and E2 specificity of the RBR E3 ubiquitin ligase HHARI
RING-in-between-RING (RBR) ubiquitin (Ub) E3 ligases function with Ub E2s through a RING/HECT hybrid mechanism to conjugate Ub to target proteins. Here, we report the crystal structure of the RBR E3, HHARI, in complex with a UbcH7~Ub thioester mimetic which reveals the molecular basis for the specificity of this cognate E2/RBR E3 pair. The structure also reveals mechanistically important conformational changes in the RING1 and UBA-like domains of HHARI that accompany UbcH7~Ub binding and provides a molecular basis by which HHARI recruits E2~Ub in an 'open' conformation. In addition to optimally functioning with an E2 that solely performs transthiolation, our data suggests that HHARI prevents spurious discharge of Ub from E2 to lysine residues by: (1) harboring structural elements that block E2~Ub from adopting a 'closed' conformation and (2) participating in contacts to ubiquitin that promote an open E2~Ub conformation.
T he reversible posttranslational modification of proteins by ubiquitin serves as a fundamental regulatory process used by cells to control nearly every aspect of eukaryotic biology [bib_ref] The ubiquitin code, Komander [/bib_ref]. The wide-ranging processes controlled by ubiquitination include cell cycle control, DNA repair, signal transduction, and immunity, among others [bib_ref] Ubiquitin signalling in DNA replication and repair, Ulrich [/bib_ref] [bib_ref] The role of ubiquitylation in receptor endocytosis and endosomal sorting, Haglund [/bib_ref] [bib_ref] Emerging regulatory mechanisms in ubiquitindependent cell cycle control, Mocciaro [/bib_ref] [bib_ref] Proteasome-independent functions of ubiquitin in endocytosis and signaling, Mukhopadhyay [/bib_ref]. Ub conjugation elicits its effects by altering the stability, localization, intermolecular interactions, and activity of target proteins 1 and requires the sequential interactions and activities of three enzymes (E1, E2, and E3) [bib_ref] Structural insights into functional modes of proteins involved in ubiquitin family pathways, Hanzelmann [/bib_ref] [bib_ref] Ubiquitin-like protein activation by E1 enzymes: the apex for downstream signalling pathways, Schulman [/bib_ref] [bib_ref] Structural and functional insights to ubiquitinlike protein conjugation, Streich [/bib_ref].
Following ATP-dependent activation of Ub that results in the formation of an E1~Ub thioester intermediate, E1 transfers Ub onto the catalytic cysteine of E2 in a process termed thioester transfer (or transthiolation). The resulting E2~Ub intermediate interacts with members of three different families of Ub E3 ligases (RING, HECT, and RING-in-between-RING (RBR)) that catalyze Ub conjugation to target proteins by distinct mechanisms [bib_ref] New insights into ubiquitin E3 ligase mechanism, Berndsen [/bib_ref] [bib_ref] Structural insights into the catalysis and regulation of E3 ubiquitin ligases, Buetow [/bib_ref] [bib_ref] E2 enzymes: more than just middle men, Stewart [/bib_ref]. The catalytic mechanism of RING E3s involves interactions between a zinc-binding RING domain and the E2~Ub thioester intermediate that facilitates nucleophilic attack of the E2~Ub thioester by a lysine on the target protein directly [bib_ref] New insights into ubiquitin E3 ligase mechanism, Berndsen [/bib_ref] [bib_ref] Structural insights into the catalysis and regulation of E3 ubiquitin ligases, Buetow [/bib_ref] [bib_ref] E2 enzymes: more than just middle men, Stewart [/bib_ref]. This contrasts with the mechanism of HECT E3s, which are structurally unrelated to RING E3s. HECT E3s participate in a process analogous to E1-E2 thioester transfer in which the conserved N-lobe of the catalytic HECT domain recruits E2~Ub and subsequently transfers Ub to a catalytic cysteine within the conserved C-lobe to form an E3~Ub thioester linkage that then undergoes attack by the target protein lysine to form the isopeptide bond [bib_ref] New insights into ubiquitin E3 ligase mechanism, Berndsen [/bib_ref] [bib_ref] Structural insights into the catalysis and regulation of E3 ubiquitin ligases, Buetow [/bib_ref] [bib_ref] E2 enzymes: more than just middle men, Stewart [/bib_ref]. RBR E3s are a distinct class of Ub E3 ligases that harbor three tandem zinc-binding domains termed RING1, in-between RING (IBR) and RING2 (collectively called an RBR domain) [bib_ref] RBR ubiquitin ligases: diversification and streamlining in animal lineages, Marin [/bib_ref] [bib_ref] RBR E3 ubiquitin ligases: new structures, new insights, new questions, Spratt [/bib_ref]. Surprisingly, RBR E3s were recently discovered to function through a RING/HECT hybrid mechanism [bib_ref] UBCH7 reactivity profile reveals parkin and HHARI to be RING/HECT hybrids, Wenzel [/bib_ref] in which the RING1 domain initially recruits the E2~Ub thioester intermediate, similar to canonical RING E3s. However, rather than facilitating Ub discharge from E2~Ub onto target protein lysine residues directly like a canonical RING E3, E2~Ub binding to the RING1 domain of RBR E2s is followed by thioester transfer of Ub to a catalytic cysteine residue in the RING2 domain similar to HECT E3s [bib_ref] UBCH7 reactivity profile reveals parkin and HHARI to be RING/HECT hybrids, Wenzel [/bib_ref] [bib_ref] Structure of a HOIP/E2~ubiquitin complex reveals RBR E3 ligase mechanism and regulation, Lechtenberg [/bib_ref] [bib_ref] Structure and function of Parkin E3 ubiquitin ligase reveals aspects of RING..., Riley [/bib_ref] [bib_ref] The E3 ligase HOIP specifies linear ubiquitin chain assembly through its RING-IBR-RING..., Smit [/bib_ref] [bib_ref] Structure of the human Parkin ligase domain in an autoinhibited state, Wauer [/bib_ref] [bib_ref] Structure of the HHARI catalytic domain shows glimpses of a HECT E3..., Spratt [/bib_ref] [bib_ref] LUBAC synthesizes linear ubiquitin chains via a thioester intermediate, Stieglitz [/bib_ref] [bib_ref] Structure of parkin reveals mechanisms for ubiquitin ligase activation, Trempe [/bib_ref]. An interesting question that arises is how RBRs prevent Ub discharge to lysine residues upon E2~Ub binding to the RING1 domain (as for canonical RING E3s) and ensure that Ub is instead transferred to the RING2 catalytic cysteine prior to catalysis of ubiquitination (as for HECT E3s). A recent study suggested that one way that HHARI achieves this is by specifically inhibiting E2~Ub conformations that are primed for catalysis of ubiquitin discharge onto lysine residues [bib_ref] Molecular insights into RBR E3 ligase ubiquitin transfer mechanisms, Dove [/bib_ref] ; however, the structural basis for this observation is unknown.
HHARI (or ARIH1) is a member of the Ariadne family of RBR E3 ligases, which also includes TRIAD1 (ARIH2), ANKIB1 (KIAA1386), and Cullin-9 (PARC) [bib_ref] RBR ubiquitin ligases: diversification and streamlining in animal lineages, Marin [/bib_ref]. In addition to the RBR domain, HHARI harbors an unstructured N-terminal domain rich in glycine and acidic residues, a UBA-like domain, and a helical C-terminal Ariadne domain that is characteristic of Ariadne subfamily members [bib_ref] Structure of HHARI, a RING-IBR-RING ubiquitin ligase: autoinhibition of an Ariadne-family E3..., Duda [/bib_ref] Overall structure of HHARI in complex with UbcH7-Ub. a HHARI is shown as a surface representation with individual domains labeled and color coded. UbcH7 (magenta) and Ub (gold) are shown as cartoons and the isopeptide bond linking C86K of UbcH7 to Gly76 of Ub is shown as sticks. A schematic outlining the domain organization of HHARI is shown on the bottom with regions of disorder in the structure indicated by hatched boxes. b Cartoon representation of the HHARI/UbcH7-Ub complex colored as in a. The isopeptide bond linking C86K of UbcH7 to Gly76 of Ub is shown as sticks. Zinc ions coordinated by the RING1, IBR, and RING2 domains of HHARI are shown as green spheres and the sulfur atom of the HHARI catalytic cysteine (Cys357) is shown as a yellow sphere HHARI determined in the absence of E2 revealed an autoinhibitory HHARI conformation in which the RING1 domain is positioned far away from the RING2 catalytic cysteine residue, which is covered by Ariadne domain. The precise mechanisms by which HHARI autoinhibition is relieved and the conformational changes required for E2-RBR E3 Ub thioester transfer are unknown; however, recent studies have shown that Ariadne subfamily members HHARI and TRIAD1 are activated through interaction with Cullin RING ligases (CRLs) in their neddylated state [bib_ref] TRIAD1 and HHARI bind to and are activated by distinct neddylated Cullin-RING..., Kelsall [/bib_ref] [bib_ref] Two Distinct Types of E3 Ligases Work in Unison to Regulate Substrate..., Scott [/bib_ref].
An additional area of interest pertains to the specificity of HHARI interactions and function with ubiquitin E2s. UbcH7 (UBE2L3) was initially identified as RING1-dependent high-affinity binding partner of HHARI [bib_ref] The ubiquitin-conjugating enzymes UbcH7 and UbcH8 interact with RING finger/IBR motif-containing domains..., Moynihan [/bib_ref] , and there is evidence that HHARI and UbcH7 function together in pharyngeal development [bib_ref] ARI-1, an RBR family ubiquitin-ligase, functions with UBC-18 to regulate pharyngeal development..., Qiu [/bib_ref] , suggesting that they constitute a biologically relevant cognate E2/RBR E3 pair. Furthermore, a systematic screen of thirty-four different ubiquitin E2s for activity with the Ariadne subfamily member Triad1 revealed that only UbcH7induced Triad1 autoubiquitination [bib_ref] TRIAD1 and HHARI bind to and are activated by distinct neddylated Cullin-RING..., Kelsall [/bib_ref]. This is consistent with in vitro autoubiquitination assays with the untagged RBR domain of HHARI, which revealed significantly higher levels of activity with UbcH7 compared to UbcH5c 14 . These results suggest that there is a high degree of specificity in the HHARI/UbcH7 interaction, however, the molecular basis governing this specificity is unknown.
Here, we report the crystal structure of HHARI in complex with a UbcH7~Ub thioester mimetic. The structure reveals mechanistically important UbcH7~Ub-induced conformational changes in the RING1 and UBA-like domains of HHARI that play a key role in determining the specificity of this E2/E3 pair, as well as in promoting recruitment of UbcH7~Ub in the 'open' conformation. We identify a residue specific to UbcH7 (Lys96) that plays a major role in determining its activity with HHARI, and we find that introducing this residue into UbcH5b substantially increases its activity, identifying this residue as a key specificity determinant. Overall, our structural and biochemical studies indicate that HHARI ensures transfer of Ub from E2 to the RING2 catalytic cysteine as opposed to discharge from E2 directly to lysine residues upon RING1 binding in at least three ways: (1) by evolving a mechanism to specifically recruit an E2 that solely performs transthiolation (UbcH7), (2) by harboring a loop insertion in the RING1 domain (unique to RBR E3s) that is involved in determining specificity of HHARI for UbcH7 and is incompatible with the UbcH7~Ub binding in the 'closed' conformation primed for Ub discharge to lysine residues, and (3) contacts between Ub and the UBA-like domain that promote recruitment of UbcH7~Ub in the inactive 'open' conformation.
# Results
Overall structure of the HHARI/UbcH7-Ub complex. Recent data are consistent with a mechanism for RBR Ub E3 ligase activity in which the RBR RING1 domain initially recruits the E2~Ub thioester intermediate, followed by transfer of Ub from E2 to a catalytic cysteine residue within the RBR RING2 domain [bib_ref] UBCH7 reactivity profile reveals parkin and HHARI to be RING/HECT hybrids, Wenzel [/bib_ref] [bib_ref] Structure and function of Parkin E3 ubiquitin ligase reveals aspects of RING..., Riley [/bib_ref] [bib_ref] The E3 ligase HOIP specifies linear ubiquitin chain assembly through its RING-IBR-RING..., Smit [/bib_ref] [bib_ref] Structure of the human Parkin ligase domain in an autoinhibited state, Wauer [/bib_ref] [bib_ref] Structure of the HHARI catalytic domain shows glimpses of a HECT E3..., Spratt [/bib_ref] [bib_ref] LUBAC synthesizes linear ubiquitin chains via a thioester intermediate, Stieglitz [/bib_ref] [bib_ref] Structure of parkin reveals mechanisms for ubiquitin ligase activation, Trempe [/bib_ref]. In order to gain insights into the molecular basis by which the RBR Ub E3 ligase HHARI recruits the E2~Ub thioester intermediate, we used an E2~Ub thioester intermediate mimetic that is stabilized by isopeptide bond formation between a lysine residue that substitutes for the E2 catalytic cysteine, and the C-terminal glycine of Ub (denoted as E2-Ub throughout) [bib_ref] Essentiality of a non-RING element in priming donor ubiquitin for catalysis by..., Dou [/bib_ref] [bib_ref] Structure of a RING E3 ligase and ubiquitin-loaded E2 primed for catalysis, Plechanovova [/bib_ref]. We focused our efforts on the ubiquitin E2, UbcH7, because biochemical and biological data show that HHARI and UbcH7 interact with high affinity and function biologically as a cognate E2/E3 pair [bib_ref] Structure of HHARI, a RING-IBR-RING ubiquitin ligase: autoinhibition of an Ariadne-family E3..., Duda [/bib_ref] [bib_ref] The ubiquitin-conjugating enzymes UbcH7 and UbcH8 interact with RING finger/IBR motif-containing domains..., Moynihan [/bib_ref] [bib_ref] ARI-1, an RBR family ubiquitin-ligase, functions with UBC-18 to regulate pharyngeal development..., Qiu [/bib_ref]. Using a slight modification of previously published protocols [bib_ref] Essentiality of a non-RING element in priming donor ubiquitin for catalysis by..., Dou [/bib_ref] [bib_ref] Structure of a RING E3 ligase and ubiquitin-loaded E2 primed for catalysis, Plechanovova [/bib_ref] , we were able to produce large quantities of pure UbcH7~Ub thioester intermediate mimetic, which as expected, forms an apparent 1:1 monomeric complex with full-length recombinant HHARI as assessed by gel filtration .
We determined the crystal structure of HHARI/UbcH7-Ub , b) to a resolution of 3.5 Å by molecular replacement using the structure of apo HHARI 23 (HHARI APO ; PDB: 4KBL) as the initial search model (Methods). Coordinates corresponding to UbcH7 and Ub were subsequently placed into unambiguous electron density evident after one round of refinement (Supplementary . There is one HHARI/UbcH7-Ub complex per asymmetric unit and the final model was refined to R/R free values of 0.226/0.252 [fig_ref] Table 1: Data collection and refinement statisticsFig [/fig_ref]. Electron density corresponding to the isopeptide bond linking UbcH7 and Ub is evident in the structure .
Overall, UbcH7-Ub-bound HHARI (HHARI E2−Ub ) adopts an elongated architecture similar to that observed in HHARI APO structures 23 (PDB ID: 4KBL and 4KC9) with the RING1 and RING2 domains separated by~90 Å at opposite ends of the structure and the catalytic cysteine residue of HHARI (Cys357) buried at the RING2/Ariadne domain interface . Almost all contacts between UbcH7 and HHARI involve the RING1 domain and the Ub molecule projects away from the surface of UbcH7, adopting an 'open' conformation that results in contacts to the UBA-like domain of HHARI and . That UbcH7 predominantly contacts the RING1 domain and that UbcH7-Ub is recruited to HHARI with Ub in the open conformation is fully consistent with previously published results [bib_ref] Molecular insights into RBR E3 ligase ubiquitin transfer mechanisms, Dove [/bib_ref] [bib_ref] Structure of HHARI, a RING-IBR-RING ubiquitin ligase: autoinhibition of an Ariadne-family E3..., Duda [/bib_ref] [bib_ref] The ubiquitin-conjugating enzymes UbcH7 and UbcH8 interact with RING finger/IBR motif-containing domains..., Moynihan [/bib_ref]. The UbcH7 and HHARI catalytic cysteines residues are separated by~54 Å, thus, the structure suggests that the UbcH7~Ub intermediate can be recruited to HHARI in its autoinhibited conformation and that UbcH7-Ub binding does not induce large-scale conformational changes required for thioester transfer of Ub from UbcH7 to HHARI. and the three available apo HHARI structures are colored the indicated shade of gray. The RING1 Loop2 E3 region of HHARI is boxed and a magnified cartoon representation of this region is presented as an inset to the right. The distance between the Cα atoms of His234 from Loop2 E3 of apo and UbcH7-Ub bound HHARI structures is shown. Disordered regions in the UBA-like domains of apo HHARI that become ordered upon UbcH7-Ub binding are shown as semitransparent gray spheres. b The HHARI/ UbcH7-Ub (top) and HHARI APO (bottom) structures are shown as cartoon representations in the same orientation with selected residues shown as sticks. c Comparison of the HHARI RING1 domain to the RING domains of HOIP, c-Cbl, and RNF25. The Cα atoms of the eight residues involved in coordination of two zinc ions (green spheres) were superimposed and the structures are shown as cartoon representations with selected side chains shown as sticks. A sequence alignment of the Loop2 E3 region of the RING domains is shown at the bottom with the loop insertion of HHARI boxed in red and highly conserved residues colored black. Cysteine residues involved in zinc coordination are indicated with a black star and the 'linchpin' arginine residue of canonical RING E3s that is involved in stabilization of the closed E2~Ub conformation is indicated with a black triangle. . c Structure-based sequence alignment of ubiquitin E2-interacting regions of selected RING domains. RING domain residues involved in contacts to the indicated ubiquitin E2 are colored magenta, and the conserved zinc coordinating residues are indicated with black stars. The 'linchpin' arginine residue of canonical RING E3s is indicated with a black triangle. An atypical residue involved in zinc atom coordination (Cys324 of FANCL) is highlighted with a black box in the alignment. Structures used to prepare the alignment are the same as in b. d Structure-function analysis of the HHARI RING1 /UbcH7-Ub interface. WT and mutant proteins were utilized in HHARI autoubiquitination assays for the indicated time points, as described in the Methods section. Full-length HHARI is catalytically inactive due to Ariadne domain-mediated autoinhibition so a catalytically active HHARI construct lacking the Ariadne domain (HHARI ΔAri ) was used in these assays, as previously described [bib_ref] Structure of HHARI, a RING-IBR-RING ubiquitin ligase: autoinhibition of an Ariadne-family E3..., Duda [/bib_ref] closer to the surface of UbcH7 in the HHARI/UbcH7-Ub structure to facilitate additional intermolecular contacts . As noted previously [bib_ref] RBR E3 ubiquitin ligases: new structures, new insights, new questions, Spratt [/bib_ref] [bib_ref] Molecular insights into RBR E3 ligase ubiquitin transfer mechanisms, Dove [/bib_ref] , the RING1 domains of many RBR E3 ligases (including HHARI) are unique compared to canonical RING domains in that they have a one to six residue insertion in the Loop2 E3 region that is positioned between two conserved residues involved in zinc coordination and Supplementary . Interestingly, the aforementioned Loop2 E3 conformational change observed in HHARI upon UbcH7 binding involves precisely the two residues comprising the HHARI Loop2 E3 insertion (His234 and Gly235; . Importantly, the conformational change in Loop2 E3 is unique to the HHARI/ UbcH7-Ub interaction as comparison of the RING domain of other E3s in the apo-vs. E2-bound states reveals Loop2 E3 conformations that are essentially identical . The role that the RING1 Loop2 E3 insertion and conformational change plays in HHARI mechanism will be discussed in further detail below.
Comparison of the HHARI APO and HHARI E2−Ub structures reveals additional HHARI conformational changes within the UBA-like domain. Relative to HHARI APO structures, the UBA-like domain undergoes a rigid body rotation towards the UbcH7~Ub thioester intermediate mimetic that facilitates contacts to Ub . The majority of the ubiquitininteracting region of the UBA-like domain involves helix 4 (H4), which is immediately C-terminal to the three-helices comprising the canonical UBA domain fold, and a loop region immediately after H4. H4 is disordered in two of three HHARI APO structures and the loop following the helix is disordered in all HHARI APO structures and likely becomes ordered in the HHARI/UbcH7-Ub structure due to contacts to Ub (see details below).
The HHARI/UbcH7 interface. A majority of the contacts between HHARI and UbcH7 involve a surface on the RING1 domain formed by two loops termed Loop1 E3 and Loop2 E3 that are positioned on either side of the domain, and an extended α helix (Helix 5, H5; [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref]. Two loops, termed Loop4 E2 and Loop7 E2 , and helix 1 (h1) of UbcH7 engage in the majority of contacts to the HHARI RING1 domain. Specifically, Loop4 E2 contacts H5 and Loop1 E3 of the RING1 domain; Loop7 E2 engages in an extensive network of contacts with Loop2 E3 ; and h1 of UbcH7 is perched above Loop1 E3 . Altogether, a total of~1500 Å 2 of surface area are buried at the HHARI RING1/UbcH7 interface with a mix of hydrophobic and hydrogen bond-mediated interactions.
At the center of the RING1/UbcH7 interface, Phe63, Pro97, and Ala98 of UbcH7 engage a hydrophobic patch on the RING1 domain formed by HHARI residues Ile188, Cys211, Glu214, and Tyr215 [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref]. This constitutes the conserved network of interactions present in most canonical RING E3/E2 structures determined to date [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref] , c and Supplementary [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref]. As expected, mutations designed to disrupt this interaction, including an I188D HHARI mutation and a F63A UbcH7 mutation, significantly diminish HHARI activity as assessed by a previously established HHARI autoubiquitination assay that circumvents HHARI autoinhibition through deletion of the autoinhibitory Ariadne domain [bib_ref] Structure of HHARI, a RING-IBR-RING ubiquitin ligase: autoinhibition of an Ariadne-family E3..., Duda [/bib_ref] [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref] and [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref]. On the Loop1 E3 side of the RING1 domain, Tyr190 of HHARI engages in a series of van der Waals contacts with the aliphatic portions of Arg6 and Lys9 on h1 and Lys100 within Loop7 E2 of UbcH7 [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref]. The significant loss of activity resulting from a Y190A substitution in HHARI demonstrates importance of this set of interactions for normal HHARI function [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref].
Contacts more specific to the HHARI/UbcH7 complex take place on the other side of the RING1 domain, where the E2-induced conformational change in Loop2 E3 facilitates an extensive set of unique intermolecular interactions with UbcH7 [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref]. Most prominently, Lys96 of UbcH7 projects towards an acidic surface on HHARI that is unmasked as a result of Loop2 E3 flipping upward towards the surface of the E2 compared to its conformation in HHARI APO structures (Figs. 2b, 3a and . Lys96 of UbcH7 engages in a network of hydrogen bonds with this unmasked region, including the side chain of Ser230 and the backbone carbonyl oxygens of Cys231 and Cys236 (which are also involved in HHARI RING1 domain zinc ion coordination) [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref]. Lys96 is also positioned to engage in longer-range electrostatic interactions with Asp237 of HHARI Loop2 E3 . The importance of Asp237 to HHARI function is demonstrated by the significant loss of autoubiquitination activity observed with a D237A mutant [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref] , though surprisingly a D237R mutant behaved similar to wild type suggesting that electrostatic interactions specifically between the D237 side chain and UbcH7 K96 do not play an important role in activity [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref]. Notably, this network of interactions is not compatible with the Loop2 E3 conformation adopted in HHARI APO structures, as Lys96 of UbcH7 would severely clash with the loop in its apo conformation [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref] , right and . An additional set of hydrogen bonds enabled by the Loop2 E3 conformational change involves the side chains of His234 of HHARI and Thr99 and Gln103 of UbcH7.
To test the contribution of the length of Loop2 E3 on HHARI functionality with UbcH7, we generated the following variants: (1) HHARIΔ H234 (2) HHARIΔ H234/G235 , and (3) HHARI with a four residue (NTYS) insertion after G235 (HHARI NTYS ). These four residues correspond to the sequence of HOIL, which has the longest Loop2 E3 insertion among human RBR E3s. To test the contribution of Loop2 E3 side chain composition, we generated the following variants: (1) HHARI H234G , (2) HHARI H234A , (3) HHARI H234W , and (4) HHARI H234Q/G235D in which the HHARI 'HG' motif was mutated to the corresponding sequence of the Ariadne family member AriH2 (also known as Triad1). The general conclusion resulting from autoubiquitination assays with these proteins is that the length of the Loop2 E3 insertion is important for HHARI activity, with both insertions and deletions significantly decreasing HHARI activity with UbcH7, whereas the side chain composition of Loop2 E3 residues is less important .
As mentioned previously, the RING1 domain of most RBR E3s differ from canonical RING domains in that they harbor a one to six residue insertion between the two zinc ion coordinating cysteine residues in Loop2 E3 [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref]. Interestingly, the region of HHARI Loop2 E3 that clashes most severely with Lys96 of UbcH7 in the 'apo' state is the two-residue His234-Gly235 insertion [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref] , right and . Together, these structural observations suggest that the HHARI Loop2 E3 conformational change induced upon UbcH7 binding is required for the HHARI/UbcH7 interaction to take place, and by extension, that a similar conformational change might be required in order for UbcH7 to function with other RBR E3s harboring a Loop2 E3 insertion.
Specificity in HHARI/ubiquitin E2 interactions. HHARI and UbcH7 function together biologically [bib_ref] ARI-1, an RBR family ubiquitin-ligase, functions with UBC-18 to regulate pharyngeal development..., Qiu [/bib_ref] and are defined as a cognate E3/E2 pair due to the high affinity with which they interact [bib_ref] Structure of HHARI, a RING-IBR-RING ubiquitin ligase: autoinhibition of an Ariadne-family E3..., Duda [/bib_ref] [bib_ref] The ubiquitin-conjugating enzymes UbcH7 and UbcH8 interact with RING finger/IBR motif-containing domains..., Moynihan [/bib_ref] and the fact that UbcH7 lacks intrinsic reactivity with lysine but can efficiently transfer Ub to cysteine [bib_ref] UBCH7 reactivity profile reveals parkin and HHARI to be RING/HECT hybrids, Wenzel [/bib_ref] , a functionality that is a unique requirement of HECT and RBR E3 ligases. Although the highly promiscuous ubiquitin E2, UbcH5b, exhibits some degree of activity with HHARI in our autoubiquitination assay, UbcH7 exhibits a much higher level of activity . This comparatively high degree of activity does not appear to be a result of UbcH7 autoubiquitination as only low molecular weight UbcH7-Ub species are evident even at the longest time points of the assay . In order to gain insights into the apparent preference of HHARI for UbcH7 compared to UbcH5b, we performed a comparative analysis of residues at the HHARI/E2 interface via structure-based sequence alignment and by creating a HHARI/UbcH5b model and . The results of this analysis reveal that one of the most prominent differences between UbcH7 and UbcH5b is Ser94 of UbcH5b, which corresponds to Lys96 of UbcH7. As noted above, Lys96 of UbcH7 engages in an intricate network of interactions with HHARI that are facilitated by Loop2 E3 conformational change that unmasks the UbcH7 interacting surface, whereas the HHARI/UbcH5b model shows that Ser94 cannot engage in an equivalent set of interactions due to its shorter side chain and lack of basicity .
UbcH7:
## Hhari
ΔAri :
[formula] Time (min) [/formula]
UbcH5b:
HHARI ΔAri :
[formula] Time (min) Uba1 UbcH7 UbcH7-Ub(n) UbcH5b-Ub(x2) UbcH5b-Ub UbcH5b [/formula]
UbcH7-Ub To investigate the potential role that UbcH7 Lys96 and UbcH5b Ser94 play in the differential ability of these ubiquitin E2s to function with HHARI, we generated E2 mutants in which these residues were swapped. K96S UbcH7 activity is diminished compared to WT , underscoring the importance of the network of interactions involving Lys96 for UbcH7 functionality with HHARI. Consistent with our structural analysis, the S94K UbcH5b mutant exhibits a striking gain of function with HHARI compared to WT . The UbcH5b gain of function does not appear to be due to UbcH5b ubiquitination as only low molecular weight UbcH5b-Ub adducts are observed in these assays. Consistent with our biochemical data, a K96S substitution in UbcH7 decreases its affinity for full-length HHARI by~30-fold (WT = 35 nM, K96S = 1020 nM K d ), whereas a S94K substitution in UbcH5b increases its affinity for HHARI by~16-fold (WT = 19.5 μM S94K = 1.2 μM K d ) as assessed using isothermal titration calorimetry .
We subjected UBE2T, which harbors an arginine at the position corresponding to K96 of UbcH7 to the same assay and found that WT UBE2T exhibited a very-low level of activity with HHARI and that an R99K variant of UBE2T had a significantly higher level of activity (though still very low compared to UbcH7 WT and UbcH5b S94K . It is also worth noting that Ube2L6, which was identified along with UbcH7 as a high-affinity binding partner of HHARI via pulldown assays, also harbors a lysine residue at the position corresponding to Lys96 of UbcH7 (also Lys96). As expected we found that UBE2L6 interacts with HHARI as assessed by analytical gel filtration chromatography , however, autoubiquitination assays show that UBE2L6 fails to exhibit robust levels of activity despite the presence of K96 (Supplementary . Furthermore, in contrast to WT UbcH7, the comparatively low level of UBE2L6 activity was not further attenuated with a K96S UBE2L6 substitution . UBE2S is another ubiquitin E2 that we found interacts with HHARI by gel filtration but has very little activity with HHARI in autoubiquitination assays . E1-E2 thioester transfer assays indicate that the ability of the mutant E2 proteins used throughout this study to be charged with Ub is not diminished relative to WT .
Overall, the data suggest that the presence of a lysine residue at the position corresponding to K96 of UbcH7 has a positive effect on E2 functionality with HHARI but that an intricate combination of factors collaborate to mediate high levels of activity. Since a single S94K substitution of UbcH5b unlocks relatively high levels of activity, the framework of residues outside of S94 are close to 'ideal'. On the other hand, the very modest gain of function when a lysine residue is introduced into UBE2T suggests that the framework of residues outside of the position of interest that are required for high activity with HHARI are significantly less than 'ideal'. Another point worth noting is that while a UbcH7 K96R variant retains very high levels of activity, this is not the case for UbcH5b S94R (relative to UbcH5b S94K ; nor is it the case for Ube2T which as noted harbors an arginine at the relevant position . This suggests that there are very subtle differences in how these E2s are interacting with the RING1 domain that have a significant effect on activity. Finally, it is worth noting that there is not a direct correlation between the ability of ubiquitin E2s to interact with HHARI as assessed by gel filtration chromatography and the ability to function with HHARI in autoubiquitination assays. This is not surprising since after binding to the RING1 domain, the ubiquitin E2 must transfer Ub to the RING2 catalytic cysteine, a process that likely involves E2/RING2 contacts (see below). It is possible that the ubiquitin E2s we tested that bind with high affinity to HHARI but do not harbor activity in autoubiquitination assays exhibit this behavior due to a suboptimal ability to interact and function with the HHARI RING2 domain.
## Basis for recruitment of e2~ub in the open conformation. e2
Ub thioester intermediates are known to adopt conformations ranging from 'closed' in which the Ub thioester folds back towards the surface of the E2 and engages in a network of interactions centered around helix 3 of the E2 (also known as the crossover helix), to 'open' and 'backbent' conformations in which Ub thioester extends away from this surface of the E2 30-32 . A series of recent studies has revealed that the mechanism by which canonical RING E3s catalyze isopeptide bond formation is through promotion of closed E2~Ub conformations that optimally position the Ub C-terminus and E2 active site for catalysis [bib_ref] Structure of a RING E3 ligase and ubiquitin-loaded E2 primed for catalysis, Plechanovova [/bib_ref] [bib_ref] BIRC7-E2 ubiquitin conjugate structure reveals the mechanism of ubiquitin transfer by a..., Dou [/bib_ref] [bib_ref] Structure of an E3:E2 approximately Ub Complex Reveals an Allosteric Mechanism Shared..., Pruneda [/bib_ref] [bib_ref] Essential Role for Ubiquitin-Ubiquitin-Conjugating Enzyme Interaction in Ubiquitin Discharge from Cdc34 to..., Saha [/bib_ref] [bib_ref] The Mechanism of Linkage-Specific Ubiquitin Chain Elongation by a Single-Subunit E2, Wickliffe [/bib_ref].
An interesting apparent contradiction of HHARI biochemical behavior is that despite harboring a RING1 domain that recruits E2 in a manner resembling the RING domain of canonical RING E3s, the HHARI RING1/E2 interaction: (1) does not enhance reactivity of E2~Ub toward free lysine, as assessed by lysine discharge assays [bib_ref] UBCH7 reactivity profile reveals parkin and HHARI to be RING/HECT hybrids, Wenzel [/bib_ref] and (2) is incapable of functioning with full-length HHARI in autoubiquitination [bib_ref] Structure of HHARI, a RING-IBR-RING ubiquitin ligase: autoinhibition of an Ariadne-family E3..., Duda [/bib_ref] and CRL monoubiquitination assays 25 , as evidenced by a complete lack of activity when the RING2 domain catalytic cysteine (Cys357) is mutated to serine. A recent study revealed that these observations can at least partially be explained by HHARI failing to induce the E2~Ub closed conformation and instead promoting recruitment of E2~Ub in the catalytically inactive open conformation [bib_ref] Molecular insights into RBR E3 ligase ubiquitin transfer mechanisms, Dove [/bib_ref].
Analysis of the HHARI/UbcH7-Ub structure reveals the molecular basis for HHARI selectivity of UbcH7-Ub in the open conformation. First, as noted above, Ub projects away from the surface of UbcH7 in the complex and engages in a network of contacts with the HHARI UBA-like domain [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref]. Interestingly, the general architecture of the HHARI/UbcH7-Ub complex resembles that of the HECT NEDD4L /UbcH5b-Ub structure [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref] ; PDB: 3JVZ). HECT E3s, like RBR E3s, must have mechanisms in place to prevent spurious discharge of Ub to lysine residues. In the HECT NEDD4L /UbcH5b-Ub structure Ub extends away from the UbcH5b active site and engages in contacts to the NEDD4L C-lobe akin to how Ub extends away from the UbcH7 active site and engages in contacts to the HHARI UBA-like domain in our structure [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref]. It is likely that these interactions serve as analogous mechanisms to promote the open E2-Ub conformation thereby preventing discharge to lysine residues.
The UBA-like/Ub interaction observed in the HHARI/UbcH7-Ub structure is atypical in that it does not involve contacts between the Ile44 hydrophobic patch of Ub and the α1-α2 loop UBA core [fig_ref] Figure 6: Analysis of the HHARI UBA-like /Ub interface [/fig_ref] -c and [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref]. Rather, in the HHARI/UbcH7-Ub structure, the β1-β2 and α1-β3 loops of Ub straddle H4 of the UBA-like domain, which as mentioned above is outside (C-terminal) of the three helix bundle that defines the UBA domain. The β1-β2 loop of Ub is also wedged between H4 and a loop that immediately follows [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref]. Importantly, this loop is disordered in all HHARI APO structures and presumably becomes ordered due to contacts to Ub in the HHARI/UbcH7-Ub structure. We found that introduction of K156A, A159D, I164W, and A159D/I164W mutations into the H4 region of the HHARI UBA-like domain resulted in a slight loss of activity in autoubiquitination assays [fig_ref] Figure 6: Analysis of the HHARI UBA-like /Ub interface [/fig_ref].
Of all available structures of UBA/Ub complexes, it is interesting to note that the HHARI UBA-like /Ub interaction most closely resembles that of the HOIL-1L UBL /HOIP UBA complex 39 (PDB: 4DBG; [fig_ref] Figure 6: Analysis of the HHARI UBA-like /Ub interface [/fig_ref]. In the HOIL-1L UBL /HOIP UBA complex, the β1-β2 and α1-β3 loops of the Ubl domain straddles an α-helix that is C-terminal of the conserved three helix bundle of the UBA domain in a manner closely resembling the HHARI UBA-like /Ub interaction [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref]. Although the positions on H4 involved in contacts to Ub/Ubl are conserved on the equivalent helix of HOIP UBA , the identity/similarity of these residues is poorly conserved [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref]. Also, although the Ub/Ubl residues involved in contacts to the HHARI UBA-like and HOIP UBA domains reside on the same surface, the network of contacts is different, with hydrogen bonds predominating at the HHARI UBA-like /Ub interface and hydrophobic contacts predominating at the HOIP UBA/ HOIL UBL interface [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref]. Overall, the binding modes of the HHARI UBA-like /Ub and HOIP UBA/ HOIL UBL complexes are superficially similar, with similar surfaces of the UBA-like/UBA domains interacting with similar surfaces of the Ub/Ubl, though the details of the interactions differ. Interestingly, the region of the HHARI UBA-like domain involved in contacts to Ub in our structure was recently identified as being important for HHARI +Neddylated-Cullin RING ligase (CRL)-mediated ubiquitination of a phosphopeptide substrate [bib_ref] Two Distinct Types of E3 Ligases Work in Unison to Regulate Substrate..., Scott [/bib_ref]. Surprisingly, we found that the binding affinity of UbcH7-Ub for HHARI (K D = 272 nM) was lower than that of free UbcH7 (K D = 35 nM) , which we speculate might be the result of the tendency of UbcH7~Ub to adopt the closed conformation 22 that is incompatible with HHARI binding.
With regards to the flexible Ub C-terminus there are several notable contacts observed in the structure including His162 of HHARI, which is wedged between the side chains of Leu71 and Arg74 of Ub and engages in a hydrogen bond to the backbone oxygen of Arg72 [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref]. A HHARI H162A mutation significantly diminishes activity relative to WT, highlighting the importance of this residue [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref]. Closer to the covalent linkage between UbcH7 and Ub Asn94 of UbcH7 engages in a pair of hydrogen bonds with the backbone oxygen of Arg74 and Pro88 of UbcH7 engages in van der Waal contacts to Gly76 of Ub, the carbonyl oxygen of which projects towards the side chain of UbcH7 His119 [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref].
While contacts between Ub, UbcH7, and the UBA-like domain are a means of promoting an open UbcH7-Ub conformation, there are also mechanisms in place that hinder UbcH7-Ub binding to HHARI in the closed conformation. We created a model of a HHARI/UbcH7-Ub(closed) using the structure of RNF38/UbcH5b-Ub 40 (PDB: 4V3K) as a guide [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref]. The model reveals that the Loop2 E3 insertion of the HHARI RING1 domain (His234-Gly235) sterically clashes with Ub in the closed conformation, in particular residues Leu8 and Leu71 [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref]. These clashes involve both main chain and side chain atoms of the Loop2 E3 insertion of HHARI and take place whether the Loop2 E3 insertion is in the conformation observed in all HHARI APO structures or that observed in the HHARI/UbcH7-Ub structure. Since Loop2 E3 is very closely positioned to Ub(closed) in all canonical RING E3s, it is likely that other RBR E3s with Loop2 E3 insertions would select against closed E2-Ub conformations through a similar mechanism. It is also noteworthy that the surface of UbcH7 that interacts with Ub in the closed conformation is partially engaged by Loop2 E3 of HHARI RING1, in particular the interaction between His234 of HHARI and Gln103 in helix 3 of UbcH7. Finally, the 'linchpin' arginine residue conserved within canonical RING domains that engages in contacts with Ub to stabilize the E2-Ub closed conformation is not conserved in RBR E3s [fig_ref] Figure 3: The HHARI RING1 /UbcH7-Ub interface [/fig_ref] and [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref]. This lack of sequence conservation appears to be a secondary factor in accounting for the lack of UbcH7's lack of reactivity with HHARI RING1 because a D237R mutant containing the 'linchpin' arginine residue functions similar to wild type and a D237R/C357A double mutant lacking the RING2 domain catalytic cysteine residue harbors no apparent activity in HHARI autoubiquitination assays (Supplementary [fig_ref] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation [/fig_ref].
HHARI activation and comparision to other RBR E3s. Analysis of our HHARI/UbcH7-Ub complex shows that HHARI E2−Ub adopts a similarly elongated structure as HHARI APO which indicates that E2~Ub binding does not trigger conformational changes necessary for thioester transfer. However, the Cα atoms of the UbcH7 and HHARI catalytic cysteines are 8 Å closer compared to the HHARI APO structure and superimposition of the IBR domains reveal conformational changes in two hinge regions that account for this difference and may provide insights to the larger structural changes that occur during thioester transfer [fig_ref] Figure 6: Analysis of the HHARI UBA-like /Ub interface [/fig_ref]. The first hinge region is centered on Helix 8 within the IBR domain which rotates~7 degrees towards E2~Ub, and the second hinge region is centered on the linker that connects the RING1 and IBR domains (R1I linker), which rotates~8 degrees toward the RING2 domain .
In addition to Hinges 1 and 2, an additional region that is likely to be involved in conformational changes required for the E2 and HHARI active sites to come in proximity during thioester transfer is the linker that connects the IBR and RING2 domains (IR2 linker). In all HHARI APO structures, the IR2 linker is disordered and though this region is helical in the E2 bound state , the elevated B-factors of residues in this linker are consistent with a high degree of flexibility. As noted previously, the IR2 forms a short helix in the HHARI RING2 NMR structure [bib_ref] Structure of the HHARI catalytic domain shows glimpses of a HECT E3..., Spratt [/bib_ref] and in the HOIP/UbcH5b-Ub structure [bib_ref] Structure of a HOIP/E2~ubiquitin complex reveals RBR E3 ligase mechanism and regulation, Lechtenberg [/bib_ref]. The IR2 linker harbors many of the residues recently identified as being involved in binding to ubiquitin thioester by HHARI [bib_ref] Molecular insights into RBR E3 ligase ubiquitin transfer mechanisms, Dove [/bib_ref] and HOIP, [bib_ref] Structure of a HOIP/E2~ubiquitin complex reveals RBR E3 ligase mechanism and regulation, Lechtenberg [/bib_ref] which led to the proposal that this region undergoes a coil-to-helix transition that effectively assembles the ubiquitin binding site, thereby facilitating thioester transfer of Ub to the active site cysteine of the RING2 domain. Our structure thus suggests that the coil-to-helix transition can occur upon UbcH7-Ub binding. However, since the Ub thioester is held in the open conformation through contacts with the UBA-like domain in our HHARI/UbcH7-Ub structure, additional conformational changes must occur for ubiquitin and the IR2 linker/RING2 to come into proximity during HHARI activation.
Interestingly, Neddylated-CRLs have recently been shown to interact with HHARI and Triad1 UBA-like domains via the Nedd8 molecule that triggers HHARI/Triad1 activation [bib_ref] TRIAD1 and HHARI bind to and are activated by distinct neddylated Cullin-RING..., Kelsall [/bib_ref] [bib_ref] Two Distinct Types of E3 Ligases Work in Unison to Regulate Substrate..., Scott [/bib_ref]. It is tempting to speculate that Nedd8 binding to the UBA-like domain releases the interaction with ubiquitin which can then 'swing around' via its flexible C-terminus to participate in interactions with the IR2 linker/helix identified by [fig_ref] Figure 6: Analysis of the HHARI UBA-like /Ub interface [/fig_ref]. This implies that if an event analogous to pUb/Ub allo binding occurs during HHARI activation, that the UBA-like domain must be displaced from it position in the HHARI/ UbcH7-Ub structure. Whether Nedd8 binding to the HHARI UBA-like domain triggers conformational changes that free up the pUb/Ub allo binding site from Parkin and HOIP structures that is sterically blocked by the first helix of the UBA-like domain of HHARI is unclear. Related to this, since pUb binding to Parkin has been shown to trigger conformational changes in the pUBH helix that results in IBR conformational changes that facilitate E3 activation [bib_ref] Mechanism of phosphoubiquitin-induced PARKIN activation, Wauer [/bib_ref] , it is also unclear whether a mechanism akin to pUBH/h E2 helix extension is conserved in HHARI. Considering that the first helix of UBA domains is typically involved in Ub/ Ubl interactions, it is also possible that a canonical UBA-like/ Nedd8 interaction mechanistically mimics the pUBH/h E2 interaction as they are located at similar regions of the structure.
The IR2 linker region of a symmetry related HOIP molecule engages in contacts to the Ub thioester in the HOIP/UbcH5b-Ub structure that involves the I44 hydrophobic patch of the Ub thioester [bib_ref] Structure of a HOIP/E2~ubiquitin complex reveals RBR E3 ligase mechanism and regulation, Lechtenberg [/bib_ref]. As noted above, Dove et al. identified a potential interaction between Ub and the IR2 linker [bib_ref] Molecular insights into RBR E3 ligase ubiquitin transfer mechanisms, Dove [/bib_ref] , which is in principle consistent with the observations in the HOIP/UbcH5b-Ub structure. With that said, the set of key ubiquitin residues identified in the Dove et al. study that are involved in contacts to HOIP, Parkin, and HHARI were overlapping but distinct [bib_ref] Molecular insights into RBR E3 ligase ubiquitin transfer mechanisms, Dove [/bib_ref] , suggesting differences in the details of how the ubiquitin donor interacts with different RBR E3s. Furthermore, in order for Ub donor to fit onto a similar binding pocket as that observed in the HOIP/UbcH5b-Ub structure it would require distinct conformational changes in HHARI, as there are steric clashes between the Ub donor and elements of RING1 and the IBR (IBRs superimposed; [fig_ref] Figure 6: Analysis of the HHARI UBA-like /Ub interface [/fig_ref].
Consistent with the idea that flexibility of the IR2 linker is important for HHARI activation, we found that introduction of proline residues and deletions in this region results in a significant reduction in HHARI activity . Furthermore, we found that mixing I188D HHARI, which is deficient in E2 binding with C357A HHARI which cannot be charged with Ub, resulted in no activity in HHARI autoubiquitination assays . Together with the IR2 mutant/truncations described above, the most likely explanation for these results is that recruitment of E2~Ub and thioester transfer of Ub to the RING2 active site cysteine occurs in cis, and that conformational changes in the IR2 linker facilitate the E2 and E3 active sites coming into proximity during this process. Presumably conformational flexibility of the linker between RING2 and the Ariadne domain (residues 395-400), which is completely disordered in the HHARI/UbcH7-Ub and all HHARI APO structures, is also required for liberation of the RING2 domain from the autoinhibitory Ariadne domain, which thereby frees the HHARI catalytic cysteine and results in its placement proximal to the UbcH7~Ub thioester linkage .
# Discussion
The HHARI/UbcH7-Ub structure presented herein reveals how HHARI selects for E2~Ub thioester intermediate in the open conformation and selects against interaction with E2~Ub in the closed conformation, thereby providing a molecular basis for how HHARI restricts discharge of Ub from E2 to the catalytic cysteine on the RING2 domain rather than functioning like a canonical RING E3 that primes E2~Ub for discharge of Ub to lysine residues. The Loop2 E3 insertion of the RING1 domain is one of the major determinants of the steric selection against E2~Ub recruitment in the closed conformation and the presence of this insertion in a majority of RBR E3 ligases suggests that this may be a relatively conserved mechanism for how this family of E3s restricts Ub discharge to the E3 catalytic cysteine. Our structural and biochemical experiments have identified an amino acid present in UbcH7 (Lys96) that is a major determinant of the specificity of the HHARI/UbcH7~Ub interaction and we find that introduction of this amino acid into UbcH5b results in a conversion from comparatively low to high levels of activity with HHARI, potentially providing a platform to study the role of HHARI/E2 specificity in a cellular context. Finally, although HHARI remains in the autoinhibited conformation in the HHARI/UbcH7-Ub structure, a comparison to HHARI APO structures reveals hinge regions that are likely to drive the conformational changes required for thioester transfer of Ub from E2 to the catalytic cysteine in the RING2 domain.
During revision of this manuscript, a related study describing a HHARI/UbcH7-Ub crystal structure was published [bib_ref] Structural Studies of HHARI/UbcH7~Ub Reveal Unique E2~Ub Conformational Restriction by RBR RING1, Dove [/bib_ref]. As it pertains to the HHARI/UbcH7 interface and contacts important for molecular recognition, both studies resulted in the same major conclusions. Both studies also resulted in the same major conclusion with regards to the mechanism by which the HHARI Loop2 E3 insertion prevents recruitment of UbcH7~Ub in the closed conformation. There is one significant difference between the two papers with regards to the location of the Ub thioester mimetic in the complex. In our structure, the single HHARI/UbcH7-Ub complex in our asymmetric unit is a 1:1:1 complex in which the ubiquitin molecule is very well-ordered and interacts in the open conformation with the HHARI UBA-like domain in a non-canonical manner as described above. In the recently published Dove et al. study, only one of two Ub molecules in the asymmetric unit of their crystal is modeled, and rather than E2 and Ub interacting with the same HHARI molecule as a 1:1:1 complex, the Ub molecule is involved in crystal contacts with the UBA-like domain of a symmetry related HHARI molecule in a manner resembling a canonical UBA/Ub interaction [bib_ref] Structural Studies of HHARI/UbcH7~Ub Reveal Unique E2~Ub Conformational Restriction by RBR RING1, Dove [/bib_ref]. The authors biochemically probed the canonical Ub/UBA-like interface observed in their crystal and did not observe an effect on HHARI autoubiquitination 44 , though it is formally possible that the observed interaction may occur with free Ub or Nedd8 from during bona fide activation of HHARI (such as through interactions with Neddylated-CRL). Elucidation of the precise architecture of the active HHARI/UbcH7~Ub complex, including the potential role of UBA-like interactions with Ub/Nedd8 will be an exciting focus of future investigation that will require further characterization of the factors that trigger HHARI activation, such as interactions with neddylated CRLs [bib_ref] TRIAD1 and HHARI bind to and are activated by distinct neddylated Cullin-RING..., Kelsall [/bib_ref] [bib_ref] Two Distinct Types of E3 Ligases Work in Unison to Regulate Substrate..., Scott [/bib_ref].
# Methods
Protein expression and purification. The DNA fragments encoding residues 1-557 (full-length) and 1-400 (ΔAriadne) of HHARI, respectively were cloned into vector pSMT3 45 encoding for an N-terminal 6×His-SMT3 tag with a ULP1 cleavage site. For crystallization, UbcH7 with the mutation C86K (to enable isopeptide linkage to ubiquitin) was cloned into pET28 vector with C-terminal uncleavable 6×His tag. For enzymatic assays, wild type (WT) UbcH7 was cloned into the same vector. WT UbcH5b was cloned into pMTTH vector that codes for a C-terminal 6×His tag with a TEV cleavage site. Primer sequences used to generate all constructs in this study are listed in [fig_ref] Table 1: Data collection and refinement statisticsFig [/fig_ref].
All proteins were expressed in BL21 (DE3) Codon Plus E.coli (Agilent Technologies) after induction with 0.1 mM isopropyl-β-D-1-thiogalactoside (IPTG) overnight at 18°C. For expression of HHARI constructs, 0.1 mM ZnCl 2 was added before induction. Bacterial cultures were collected by centrifugation, and lysed by sonication in lysis buffer (20 mM Tris HCl pH 8.0, 350 mM NaCl, 20 mM Imidazole, 2 mM 2-Mercaptoethanol (βME)), in the presence of DNase. The suspension was centrifuged and the supernatant applied to Ni-NTA agarose (Qiagen). For HHARI constructs, 6×His-SMT3 tags were removed by addition of ULP1 protease at a ratio of 1:1000 (w/w) overnight at 4°C. HHARI constructs were further purified using HiLoad 26/600 Superdex 200 pg size-exclusion chromatography column (GE Healthcare) equilibrated in buffer (20 mM Tris HCl pH 8.0, 350 mM NaCl, 2 mM βME) and MonoQ 10/100 GL (GE Healthcare) in 20 mM Tris HCl pH 8.0, 50 mM NaCl, 2 mM βME and eluted with a gradient from 50-1000 mM NaCl. For UbcH5b constructs, 6×His tags were removed by addition of TEV protease at a ratio of 1:100 (w/w) overnight at 4°C. UbcH7 constructs and UbcH5b constructs were further purified on a HiLoad 26/600 Superdex 75 pg size-exclusion chromatography column (GE Healthcare) equilibrated in buffer (20 mM Tris HCl pH 8.0, 350 mM NaCl). UbcH7 constructs were further purified by using MonoS 10/100 GL (GE Healthcare) in 20 mM Bis-Tris pH 6.5, 50 mM NaCl and eluted with a gradient from 50-1000 mM NaCl.
UbcH7-Ub conjugate. UbcH7-Ub conjugate was generated based on published methods [bib_ref] Essentiality of a non-RING element in priming donor ubiquitin for catalysis by..., Dou [/bib_ref] [bib_ref] Structure of a RING E3 ligase and ubiquitin-loaded E2 primed for catalysis, Plechanovova [/bib_ref] with some modification. Briefly, 0.5 μM Uba1, 5 μM UbcH7 and 25 μM Ub were mixed in a buffer containing 50 mM Tris HCl pH9.5, 50 mM NaCl, 2 mM ATP, 10 mM MgCl 2 , and 1 mM βME. The mixture was incubated at 35°C for 16 h. The UbcH7-Ub conjugate was purified by using MonoS 10/100 GL (GE Healthcare) in 20 mM Bis-Tris pH 6.5, 50 mM NaCl and eluted with a gradient from 50-1000 mM NaCl and HiLoad 26/600 Superdex 75 pg sizeexclusion chromatography column (GE Healthcare) equilibrated in buffer (20 mM Tris HCl pH 8.0, 100 mM NaCl).
HHARI/UbcH7-Ub complex formation. HHARI was mixed with a 2-fold molar excess of UbcH7-Ub and incubated overnight at 4°C. The mixture was then applied to a Superose 12 10/300 GL size-exclusion chromatography column equilibrated in 20 mM Tris HCl pH 8.0, 50 mM NaCl and 2 mM βME. Complex formation and purity was confirmed using SDS-PAGE, and complex containing fractions were pooled and concentrated to 10 mg ml −1 for crystallization.
HHARI/UbcH7-Ub crystallization. Crystals were grown at 18°C using the hanging drop vapor diffusion method by mixing 1 μl of protein complex with 1 μl reservoir solution (100 mM sodium cacodylate pH 6.17, 350 mM sodium acetate, 10% Glycerol). Crystals appeared after 3 days and grew to their final size within 7 days. Crystals were briefly soaked in a cryoprotectant solution (100 mM sodium cacodylate pH 6.17, 2 M sodium acetate, 20% Glycerol) before flash-freezing in liquid nitrogen.
Structure determination and refinement. X-ray diffraction data for the HHARI/ UbcH7-Ub complex was collected on a Pilatus 6MF detector at Advanced Photon Source (Argonne, Illinois, USA), NE-CAT beamline 24-ID-C. All data were indexed, integrated, and scaled using HKL2000 [bib_ref] Processing of X-ray Diffraction Data Collected in Oscillation Mode, Otwinowski [/bib_ref]. The HHARI/UbcH7-Ub crystal belongs to space group P6 5 22 with unit cell dimensions a = 154.2, b = 154.2, c = 285.5. There is one HHARI/UbcH7-Ub complex per asymmetric unit.
A complete data set for the HHARI/UbcH7-Ub crystals was collected to a resolution of 3.5 Å. The program PHASER 47 was used to find an initial molecular replacement solution using a multiple ensemble search comprising the UBA-RING1, IBR, and RING2-Ariadne fragments derived from a structure of HHARI determined in the absence of E2 (PDB: 4KBL). After one round of refinement the resulting maps were inspected and electron density for UbcH7 was evident. Coordinates for UbcH7 (PDB: 4Q5E) were manually placed into the electron density and following an additional round of refinement coordinates for ubiquitin (PDB: 4II2) were manually placed into unambiguous electron density. The model was refined to R/R free values of 0.226/0.252 via iterative rounds of refinement and rebuilding using PHENIX [bib_ref] PHENIX: a comprehensive Python-based system for macromolecular structure solution, Adams [/bib_ref] and COOT 49 .
Autoubiquitination assay. Autoubiquitination assays were performed in 20 mM HEPES pH 7.5, 50 mM NaCl, 2.5 mM MgCl 2 and using 150 nM Uba1, 500 nM E2, 250 nM HHARI and 20 μM Ub. Reactions were initiated by addition of 2.5 mM ATP and were incubated at room temperature. Reaction samples were quenched at the indicated time points in SDS-PAGE loading buffer containing 10% βME, resolved on 4-15% gradient gels (Bio-Rad), and visualized by SYPRO-Ruby stain (BioRad).
E2 Charging assay. E1-E2 thioester transfer assays to control for charging of the various E2s used throughout this study were performed as described for the HHARI autoubiquitination assays except that E3 was not added and the reaction time was 3 min.
Western blot. Samples from autoubiquitination assays were resolved on SDS-PAGE and were transferred to PVDF (ThermoFisher) membranes using wet transfer method. The membranes were blocked with 5% non-fat milk, and incubated with primary and secondary antibodies sequentially with 3-5 washes for 10 min. Antibodies are as follows: UbcH7 (BostonBiochem A-640) used at a 1:1000 dilution, UbcH5b (ThermoFisher PA5-30968) used at a 1:2000 dilution, Ub (ThermoFisher PA 1-187) used at a 1:2000 dilution, and ECL-labeled anti-rabbit antibody (ThermoFisher NA934VS) used at a 1:10,000 dilution. Signal was developed with ECL substrates visualized by BioRad Geldoc.
ITC. ITC experiments were performed on a nanoITC (TA instruments) at 25°C in buffer containing 50 mM NaCl, 10 mM Na 2 HPO 4 , 2 mM KH 2 PO 4 pH 7.0. Aliquots (2 μl each) of 200 μM UbcH7 proteins were injected into cell containing 20 μM HHARI full length. For UbcH5b proteins, 500 μM into 50 μM HHARI full length were used. Twenty measurements were made and the data were analyzed using NanoAnalyze (TA instruments).
Size-exclusion analysis. Full-length HHARI and the indicated ubiquitin E2s were incubated in 20 mM Tris HCl pH 8.0, 50 mM NaCl, 2 mM βME at 4°C overnight before going through Superose 12 10/300 GL size-exclusion column equilibrated in the same buffer. Fractions were analyzed on SDS-PAGE and visualized with Coomassie blue.
Data availability. Atomic coordinates and structure factors are deposited in the RCSB with accession code 5TTE. The data that support the findings of this study are available from the corresponding author upon reasonable request.
Received: 12 January 2017 Accepted: 15 June 2017
[fig] Figure 3: The HHARI RING1 /UbcH7-Ub interface. a (left) Cartoon representation of the HHARI RING1 /UbcH7-Ub interface with residues involved in intermolecular interactions shown as sticks. Zinc ions are shown as green spheres and hydrogen bonds are shown as dashed lines. (right) Model of an apo HHARI/UbcH7 complex created by superimposing RING1/UbcH7 from the HHARI/UbcH7-Ub structure onto RING1 of apo HHARI structures. Disordered regions in the UBA-like domains of apo HHARI that become ordered upon UbcH7-Ub binding are shown as semitransparent gray spheres. A magnified view of The HHARI Loop2 E3 region is boxed and a magnified view is shown in the inset. b Structure-based sequence alignment of RING-interacting regions of selected ubiquitin E2s. E2 residues involved in contacts to the indicated RING domain are colored blue, the catalytic cysteines are indicated with a red star, and dashes indicate gaps in the alignment. Structures used to prepare the figure are c-CBL/UbcH7 (PDB:1FBV), RNF25/UbcH5b (PDB:5D1L), HOIP/ UbcH5c (PDB:5EDV), FANCL/UBE2T (PDB:4CCG), and RNF4/Ube2N (PDB:5AIT) [/fig]
[fig] Figure 5: Molecular basis for HHARI recruitment of E2~Ub in the open conformation. a (top) HHARI and UbcH7 from the HHARI/UbcH7-Ub structure are shown as cartoon representations with semitransparent surfaces. Ub is shown as a cartoon representation. Selected residues are shown as sticks. (bottom) A HHARI/UbcH7-Ub model with UbcH7-Ub in the 'closed' conformation shown in the same orientation as the top panel. The model was prepared by superimposing UbcH5b-Ub in the closed conformation (orange) (PDB: 4V3K) onto UbcH7 from the HHARI/UbcH7-Ub structure. The surface of the HHARI UBA-like domain that interacts with the core of Ub from the HHARI/UbcH7-Ub structure is colored yellow. Loop2 E3 of HHARI is boxed and two magnified views of this region are shown in the insets. The two Ub residues that clash most severely with HHARI Loop2 E3 in the closed conformation are colored red and shown as ball and sticks. b Comparison of the HHARI/UbcH7-Ub (left) and HECT NEDD4L /UbcH5b-Ub (right) complexes. Structures are shown as cartoon representations in the same orientation with proteins and domains labeled and color-coded [/fig]
[fig] Figure 6: Analysis of the HHARI UBA-like /Ub interface. a Structural comparison of the UBA domains of HHARI, HOIP (PDB: 4DBG), and Dsk2p (PDB: 1WR1). The conserved three helix core of the UBA domains are colored gray and N-and C-terminal extensions unique to each protein are colored as indicated. b Comparison of the Ub/Ubl binding mode of the UBA domains of HHARI, HOIP, and Dsk2p. Ub/Ubl (gold) from the HHARI/UbcH7-Ub, HOIP UBA / HOIL1 UBL (PDB: 4DBG), and Ub/Dsk2p UBA (PDB: 1WR1) structures were superimposed and are shown in the same orientation. c (left) The surfaces of Ub/ Ubl that interact with the UBA-like domain of HHARI, HOIL, and Dsk2p are outlined on the surface of the Ubl domain from the HOIP UBA /HOIL UBL structure. (right) The Ub/Ubl from the HHARI UBA-like /Ub, HOIP UBA /HOIL UBL , and Dsk2p UBA /Ub structures were superimposed and are shown as cartoon representations. The fragments of the UBA and UBA-like domains that contact the Ub/Ubl domains are shown as cartoon representations colored as in the left panel. d Structure-function analysis of the HHARI UBA-like/Ub interface. The indicated HHARI ΔAri variants were subjected to autoubiquitination assays with UchH7 for the indicated time points, as described in the Methods section [/fig]
[table] Table 1: Data collection and refinement statisticsFig. 2 HHARI conformational changes accompany UbcH7-Ub binding. a Comparison of HHARI structures in the apo and UbcH7-Ub-bound states. The RING1 domains of the structures were superimposed. The HHARI/UbcH7-Ub structure is colored as in [/table]
|
The genome organization of Neurospora crassa at high resolution uncovers principles of fungal chromosome topology
These authors contributed equally to this work.AbstractThe eukaryotic genome must be precisely organized for its proper function, as genome topology impacts transcriptional regulation, cell division, replication, and repair, among other essential processes. Disruptions to human genome topology can lead to diseases, including cancer. The advent of chromosome conformation capture with high-throughput sequencing (Hi-C) to assess genome organization has revolutionized the study of nuclear genome topology; Hi-C has elucidated numerous genomic structures, including chromosomal territories, active/silent chromatin compartments, Topologically Associated Domains, and chromatin loops. While low-resolution heatmaps can provide important insights into chromosomal level contacts, high-resolution Hi-C datasets are required to reveal folding principles of individual genes. Of particular interest are high-resolution chromosome conformation datasets of organisms modeling the human genome. Here, we report the genome topology of the fungal model organism Neurospora crassa at a high resolution. Our composite Hi-C dataset, which merges 2 independent datasets generated with restriction enzymes that monitor euchromatin (DpnII) and heterochromatin (MseI), along with our DpnII/MseI double digest dataset, provide exquisite detail for both the conformation of entire chromosomes and the folding of chromatin at the resolution of individual genes. Within constitutive heterochromatin, we observe strong yet stochastic internal contacts, while euchromatin enriched with either activating or repressive histone post-translational modifications associates with constitutive heterochromatic regions, suggesting intercompartment contacts form to regulate transcription. Consistent with this, a strain with compromised heterochromatin experiences numerous changes in gene expression. Our high-resolution Neurospora Hi-C datasets are outstanding resources to the fungal community and provide valuable insights into higher organism genome topology.
# Introduction
Eukaryotic genome topology, or the spatial organization of chromosomal DNA within the nucleus, plays a critical role in genome function, as DNA folding has roles in gene expression, epigenetics, the cell cycle, DNA replication and repair, development, and countless other DNA-templated processes [bib_ref] Chromosome territories, nuclear architecture and gene regulation in mammalian cells, Cremer [/bib_ref] [bib_ref] Beyond the sequence: cellular organization of genome function, Misteli [/bib_ref] [bib_ref] Gene regulation in the third dimension, Dekker [/bib_ref] [bib_ref] Enhancer function: new insights into the regulation of tissue-specific gene expression, Ong [/bib_ref] [bib_ref] Organization and function of the 3D genome, Bonev [/bib_ref] [bib_ref] The three-dimensional organization of mammalian genomes, Yu [/bib_ref] [bib_ref] Developmental enhancers and chromosome topology, Furlong [/bib_ref] [bib_ref] Heterochromatin drives compartmentalization of inverted and conventional nuclei, Falk [/bib_ref]. Of these, genome topology arguably impacts transcriptional regulation to the greatest extent, yet research is only beginning to illuminate how long-range contacts impact gene expression. For proper levels of transcription in higher eukaryotes, enhancer and/ or silencer regulatory elements, which can be thousands of basepairs (bp) distant on the linear chromosome, associate in close proximity with their cognate promoter within the 3D spatial organization of the nucleus [bib_ref] Looping and interaction between hypersensitive sites in the active b-globin locus, Tolhuis [/bib_ref] [bib_ref] Remote control of gene transcription, West [/bib_ref] [bib_ref] Gene regulation in the third dimension, Dekker [/bib_ref] [bib_ref] The pluripotent regulatory circuitry connecting promoters to their long-range interacting elements, Schoenfelder [/bib_ref]. Further, genes in the same biological pathway can spatially colocalize in RNA Pol II transcriptional hubs for coordinated, temporal control of gene expression [bib_ref] Chromatin decondensation and nuclear reorganization of the HoxB locus upon induction of..., Chambeyron [/bib_ref] [bib_ref] The locus control region is required for association of the murine b-globin..., Ragoczy [/bib_ref] [bib_ref] Preferential associations between co-regulated genes reveal a transcriptional interactome in erythroid cells, Schoenfelder [/bib_ref] [bib_ref] The pluripotent regulatory circuitry connecting promoters to their long-range interacting elements, Schoenfelder [/bib_ref] [bib_ref] Clustering of mammalian Hox genes with other H3K27me3 targets within an active..., Vieux-Rochas [/bib_ref]. Aberrant DNA organization that disrupts normal genome function can lead to disease in humans. Both abnormal chromosomal numbers resulting from chromosomal segregation defects, such as the trisomy of chromosome 21 causing Down's syndrome [bib_ref] The sequence of human chromosome 21 and implications for research into Down..., Gardiner [/bib_ref] , as well as more-subtle disruptions to genome topology, can drastically impact normal genome function. For example, a single genome rearrangement that disrupts a topological boundary can improperly allow the promoter of an oncogene to spatially associate with an enhancer, thereby stimulating oncogenic growth [bib_ref] Insulator dysfunction and oncogene activation in IDH mutant gliomas, Flavahan [/bib_ref] [bib_ref] The three-dimensional organization of mammalian genomes, Yu [/bib_ref] [bib_ref] Integrative detection and analysis of structural variation in cancer genomes, Dixon [/bib_ref] [bib_ref] Highly rearranged chromosomes reveal uncoupling between genome topology and gene expression, Ghavi-Helm [/bib_ref] [bib_ref] Subtle changes in chromatin loop contact propensity are associated with differential gene..., Greenwald [/bib_ref]. Given the critical link between genome topology and function, in-depth studies of genome organization in both humans and more simplistic model organisms are essential.
Historically, genome topology was examined microscopically using electron or fluorescence microscopy, the latter using Fluorescent In Situ Hybridization (FISH) to label individual DNA sequences or fluorescently tagged proteins binding to specific genomic features (e.g. the centromeres) [bib_ref] Chromosome territories, nuclear architecture and gene regulation in mammalian cells, Cremer [/bib_ref]. These techniques contributed to numerous scientific advancements regarding the organization of chromatin-the complex of DNA and proteins critical for genome structure and function-including the condensation of mitotic chromosomes, the formation of individual chromosome territories, the association of compacted, silent heterochromatin with the nuclear periphery while the decompressed, transcriptionally active euchromatin typically localizes to the center of the nucleus, and the colocalization of certain chromosomal features in some organisms, such as telomeres, which independently cluster yet are segregated from centromere clusters [bib_ref] Structure and biochemistry of the nuclear envelope, Franke [/bib_ref] [bib_ref] The new cytogenetics: blurring the boundaries with molecular biology, Speicher [/bib_ref] [bib_ref] The nuclear envelope as a chromatin organizer, Zuleger [/bib_ref] [bib_ref] Nucleolus and nuclear periphery: velcro for heterochromatin, Padeken [/bib_ref] [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref] [bib_ref] Heterochromatin drives compartmentalization of inverted and conventional nuclei, Falk [/bib_ref] [bib_ref] 3D genomics across the tree of life reveals condensin II as a..., Hoencamp [/bib_ref]. The advent of chromosome conformation capture (3C) and its subsequent coupling to highthroughput sequencing (Hi-C) revolutionized the study of genome topology: these exquisite procedures capture interacting genomic loci as individual DNA ligation products. Early reports examined genome topology across cell stages, in single cells, for inactivated X-chromosomes, and even between individual genes [bib_ref] Capturing chromosome conformation, Dekker [/bib_ref] [bib_ref] Comprehensive mapping of long-range interactions reveals folding principles of the human genome, Lieberman-Aiden [/bib_ref] [bib_ref] Spatial partitioning of the regulatory landscape of the X-inactivation centre, Nora [/bib_ref] [bib_ref] A high-resolution map of the threedimensional chromatin interactome in human cells, Jin [/bib_ref] [bib_ref] Organization of the mitotic chromosome, Naumova [/bib_ref] [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref] [bib_ref] Cell-cycle dynamics of chromosomal organization at single-cell resolution, Nagano [/bib_ref]. Many discoveries regarding eukaryotic genome organization have been reported, or confirmed, using Hi-C data. These include metazoan chromosomes occupying individual territories in the nucleus [bib_ref] Iterative correction of Hi-C data reveals hallmarks of chromosome organization, Imakaev [/bib_ref] [bib_ref] Three-dimensional folding and functional organization principles of the Drosophila genome, Sexton [/bib_ref] [bib_ref] 3D genomics across the tree of life reveals condensin II as a..., Hoencamp [/bib_ref] , active euchromatin and silent heterochromatin forming segregated compartments [bib_ref] Comprehensive mapping of long-range interactions reveals folding principles of the human genome, Lieberman-Aiden [/bib_ref] [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref] , and Topologically Associated Domains (TADs)-in which internal chromatin is more apt to interact than DNA outside TAD boundaries-structurally organizing the genome over Megabases of DNA [bib_ref] Spatial partitioning of the regulatory landscape of the X-inactivation centre, Nora [/bib_ref] [bib_ref] Three-dimensional folding and functional organization principles of the Drosophila genome, Sexton [/bib_ref] [bib_ref] Regulation of single-cell genome organization into TADs and chromatin nanodomains, Szabo [/bib_ref]. While lower resolution Hi-C datasets monitor chromatin interactions on (sub-)chromosomal scales, only high-resolution datasets with deep sequencing coverage have enough detail to scrutinize individual gene folding. reported the gold-standard high-resolution Hi-C dataset of human cells to capture and elucidate novel genomic structures [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref]. This study was the first to introduce in situ Hi-C ligation, in which Hi-C ligation products are formed in the nucleus, rather than by large-volume proximity ligation, to more accurately capture genomic contacts. Further, presented contact matrices at an astonishing 1 kilobase (kb) resolution: interactions could be observed across the human genome between any two 1,000 basepair segments ("bins"; in Hi-C datasets, chromosomal DNA is divided into nonoverlapping bins of a uniform size, and the array of bins in a contact matrix reports the interaction value between 2 bins for all bins across a genome). This work defined "resolution" as the smallest bin size at which 80% of all genomic loci (not all bins) have !1,000 contacts with any other locus, thereby providing outstanding structural detail of local contacts. Indeed, Rao et al. detailed 6 novel chromatin compartments (2 active "A" euchromatic and 4 silent "B" heterochromatic compartments) defined by specific epigenetic histone marks, nuclear lamina interactions, and cell cycle stages [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref]. Chromatin loops were also observed, where 2 distant loci, bound by a CTCF dimer, strongly interact to form a loop anchor [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref]. Later work highlighted roles of the protein complexes cohesin and condensin for the formation of loops and chromosomal territories; cohesin extrudes DNA to form loops [bib_ref] Cohesin loss eliminates all loop domains, Rao [/bib_ref] [bib_ref] Organization of chromosomal DNA by SMC complexes, Yatskevich [/bib_ref] [bib_ref] 3D genomics across the tree of life reveals condensin II as a..., Hoencamp [/bib_ref]. While these advances have been critical to discern human genome topology, the large size of the diploid human genome ($6.6 Â 10 9 basepairs) and its complex underlying pathways make mechanistic studies of human genome organization difficult. Thus, further study of genome topology in simple model organisms can provide additional insight into eukaryotic chromosome organization; if Hi-C is to be used to dissect genome topology, high-resolution datasets are imperative.
The filamentous fungus Neurospora crassa is an outstanding model system for chromosome conformation studies [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref] [bib_ref] Chromatin structure and function in Neurospora crassa, Courtney [/bib_ref]. Its relatively small, haploid genome (4.1 Â 10 7 basepairs in length, which is $167x smaller than the diploid human genome; note most fungi are predominantly haploid except for a short time during the sexual cycle when their nuclei are diploid, which differs from the predominantly diploid human cells) compacts similarly as DNA in higher organisms: the ratio of genome size to nuclear volume in fungi ($9.8 Â 10 6 bp/lm 3 ) mirrors that of some human cells ($1.3 Â 10 7 bp/lm 3 ), and TADs reportedly form across the Neurospora genome . Further, the Neurospora genome is partitioned into euchromatin and heterochromatin [bib_ref] The genome sequence of the filamentous fungus Neurospora crassa, Galagan [/bib_ref] [bib_ref] Relics of repeat-induced point mutation direct heterochromatin formation in Neurospora crassa, Lewis [/bib_ref] [bib_ref] Neurospora chromosomes are organized by blocks of importin alpha-dependent heterochromatin that are..., Galazka [/bib_ref]. Neurospora heterochromatin is homologous to that in humans: it is divided into constitutive and facultative subtypes. Permanently silent constitutive heterochromatin is found at gene-poor, repetitive, AT-rich DNA sequences and is posttranslationally marked by methylation of cytosines in DNA and tri-methylation of lysine 9 on histone H3 (H3K9me3; catalyzed by the DIM-5/KMT-1 histone methyltransferase [HMTase] and bound by Heterochromatin Protein-1 [HP1]) [bib_ref] Trimethylated lysine 9 of histone H3 is a mark for DNA methylation..., Tamaru [/bib_ref] [bib_ref] HP1 is essential for DNA methylation in Neurospora, Freitag [/bib_ref] [bib_ref] Relics of repeat-induced point mutation direct heterochromatin formation in Neurospora crassa, Lewis [/bib_ref] [bib_ref] DNA methylation and normal chromosome behavior in Neurospora depend on five components..., Lewis [/bib_ref] [bib_ref] Neurospora importin a is required for normal heterochromatic formation and DNA methylation, Klocko [/bib_ref] [bib_ref] Histone methylation by SET domain proteins in fungi, Freitag [/bib_ref] [bib_ref] Chromatin structure and function in Neurospora crassa, Courtney [/bib_ref]. In contrast, temporarily silent facultative heterochromatin covering gene-rich regions is marked by post-translational di-or tri-methylation of lysine 27 on histone H3 (H3K27me2/3; catalyzed by the SET-7/KMT-6 HMTase) and/or di-methylation of lysine 36 on H3 (H3K36me2; catalyzed by the ASH1L HMTase) [bib_ref] Regional control of histone H3 lysine 27 methylation in Neurospora, Jamieson [/bib_ref] [bib_ref] Loss of HP1 causes depletion of H3K27me3 from facultative heterochromatin and gain..., Jamieson [/bib_ref] [bib_ref] Genome-wide redistribution of H3K27me3 is linked to genotoxic stress and defective growth, Basenko [/bib_ref] [bib_ref] Histone methylation by SET domain proteins in fungi, Freitag [/bib_ref] [bib_ref] Chromatin structure and function in Neurospora crassa, Courtney [/bib_ref]. Euchromatin in Neurospora is demarcated by post-translational di-or tri-methylation of lysine 4 on H3, tri-methylation of lysine 36 on H3 (catalyzed by the SET-2 HMTase), and acetylation of Nterminal tails on histone H3 and H4, among others [bib_ref] Relics of repeat-induced point mutation direct heterochromatin formation in Neurospora crassa, Lewis [/bib_ref] [bib_ref] H2B-and H3-specific histone deacetylases are required for DNA methylation in Neurospora crassa, Smith [/bib_ref] [bib_ref] Heterochromatin protein 1 forms distinct complexes to direct histone deacetylation and DNA..., Honda [/bib_ref] [bib_ref] The coding and noncoding transcriptome of Neurospora crassa, Cemel [/bib_ref] [bib_ref] Histone H3 lysine 4 methyltransferase is required for facultative heterochromatin at specific..., Zhu [/bib_ref]. All told, the genetically tractable N. crassa is an excellent and cost-efficient model of the human genome: high-resolution fungal in situ Hi-C datasets, using fewer Illumina sequencing reads for deep coverage of chromatin contacts across the genome, can elucidate fundamental principles of chromosome conformation inherent to eukaryotes.
Here, we present high-resolution in situ Hi-C datasets of the wild-type N. crassa genome that effectively monitor both euchromatin and heterochromatin, thereby illuminating chromosomal topology across the entire Neurospora genome. We independently generated datasets that monitor contacts in active and silent chromatin using common restriction enzymes with 4-base recognition sequences: DpnII (^GATC) for euchromatin and MseI (T^TAA) for heterochromatin; the latter facilitated the creation of a new Neurospora reference genome version, where we placed unassigned Supercontigs 8-20 in heterochromatic regions on chromosomes [Linkage Groups (LG)] I and V. By either merging the individual DpnII and MseI fastq files and building a single Hi-C contact matrix or generating a DpnII/MseI double digest in situ Hi-C library, we can assess the conformation of individual genomic loci at 500 bp bin resolution. Conservatively, more distant yet strong intrachromosomal contacts can be observed at 1 kb or 2.5 kb bin resolution. We show that chromatin internal to silent genomic regions has extensive, dense, and random internal contacts, while gene-rich, active chromatin forms "globules" $20-40 kb in length that are hierarchically packaged into "Regional Globule Clusters"-analogous to TADs. Further, we observe small euchromatic segments enriched with activating (H3K4me3 or H3K27ac) or repressive (H3K36me2) histone post-translational modifications that contact H3K9me3-marked constitutive heterochromatic regions, and many genes associating with silent chromatin are misregulated when heterochromatin is compromised (e.g. in a Ddim-5 strain), possibly reflecting a novel fungal gene regulation mechanism. All told, our high-resolution in situ Hi-C datasets of wild type N. crassa are valuable resources for the study of eukaryotic genome topology and should facilitate future studies in other conditions or genotypes to characterize fungal chromosome conformation.
# Materials and methods
Strains, culture conditions, crosslinking, and isolation of Neurospora spheroplasts Wild type (WT) N. crassa strains N150 and N3752 were used for all experiments; both strains are called "74-OR23-1VA" and share the same Fungal Genetics Stock Center number (FGSC#2489) but can be considered independent strains due to differing acquisition times and asexual laboratory propagation. Neurospora culture growth, formaldehyde crosslinking, and spheroplasting were performed essentially as described [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref]. The detailed Materials and Methods text describing the in situ Hi-C method as well as a comprehensive, step-by-step protocol for in situ Hi-C adapted to N. crassa are both provided in the Supplementary material.
## Hi-c library construction
Hi-C libraries were generally constructed as previously described [bib_ref] Comprehensive mapping of long-range interactions reveals folding principles of the human genome, Lieberman-Aiden [/bib_ref] [bib_ref] Neurospora chromosomes are organized by blocks of importin alpha-dependent heterochromatin that are..., Galazka [/bib_ref] [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref] , but the protocol was adjusted to generate ligation products in the nucleus (in situ Hi-C) [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref] [bib_ref] Architectural alterations of the fission yeast genome during the cell cycle, Tanizawa [/bib_ref] , which more accurately reflects in vivo genomic contacts. The protocol was also refined for efficient use of reagents [bib_ref] Hi-C 2.0: an optimized Hi-C procedure for high-resolution genome-wide mapping of chromosome..., Belaghzal [/bib_ref]. Specific changes include: spheroplast lysing by glass bead vortexing combined with permeabilizing nuclear membranes with SDS [bib_ref] Architectural alterations of the fission yeast genome during the cell cycle, Tanizawa [/bib_ref] ; Klenow blunting of 5 0 DNA overhangs in a smaller volume to maintain the required 30 lM nucleotide concentration but using less Biotin-14-dATP (Invitrogen cat# 19524-016); in situ ligating DNA strands in a smaller volume; and streamlining organic extraction of DNA ligation products.
## Preparation of hi-c libraries for illumina sequencing
Hi-C libraries for Illumina sequencing were prepared using a NEBNext Ultra II kit (New England Biolabs cat# E7645) with corresponding Multiplex Barcode sets (NEB cat# E7335 and E7500) according to the manufacturer's protocol, with the following exceptions, as all steps were performed using the Hi-C library attached to the magnetic streptavidin beads: following the adapter ligation and USER enzyme digestion, magnetic beads were washed 5 times with 1x BW buffer [5 mM Tris-HCl (pH 8 at 25 C), 0.5 mM Na-EDTA, 1 M NaCl, 0.05% (v/v) Tween-20] and once with TE/10, and were resuspended in 15 ll TE/10 [10 mM Tris-HCl (pH 8 at 25 C), 0.1 mM sodium EDTA]; PCR enrichment of the barcoded Hi-C library off streptavidin beads used either 8 or 15 cycles (or an initial 15 PCR cycles, whereupon the beads were washed 3 times with TE/10, and were re-amplified by an additional 8 cycles, for a total of 23 cycles); and, following the separation of the aqueous PCR product from the magnetic beads, the PCR product was cleaned with a 1:1 ratio of Ampure XP beads (Agencourt, Beckman-Coulter) per the manufacturer's protocol, resuspended in 25 ll TE/10, and quantified by a Qubit HS reaction. Prior to sequencing, all libraries were assessed for quality via Fragment Analyzer and quantity of barcoded DNA by qPCR [Genomics and Cell Characterization Core Facility (GC3F), University of Oregon]. Indexed in situ Hi-C libraries were pooled and sequenced on either an Illumina HiSeq 4,000 as 100 nucleotide (nt) paired-end sequencing runs or an Illumina NovaSeq 6000 as 59 nt paired-end sequencing runs at the Genomics and Cell Characterization Core Facility [GC3F] at the University of Oregon. Hi-C dataset fastq files are provided in .
Bioinformatic data analysis, including mapping of Hi-C libraries and merging of DpnII and MseI datasets Paired-end reads were initially mapped to the corrected Neurospora genome assembly version 12 (nc12) [bib_ref] The genome sequence of the filamentous fungus Neurospora crassa, Galagan [/bib_ref] [bib_ref] Neurospora chromosomes are organized by blocks of importin alpha-dependent heterochromatin that are..., Galazka [/bib_ref]. The MseI dataset analysis allowed further correction of nc12 by providing locations for the unassigned Supercontigs 8-20; the new N. crassa genome assembly, here termed "version 14" (nc14), was used to construct all additional genome contact maps.
All in situ Hi-C contact maps and other analyses presented in this manuscript were generated using the program suite HiCExplorer (https://hicexplorer.readthedocs.io/en/latest/index. html; accessed 2022 September 3) (Ram ırez et al. 2018), using version 3.5 for all analyses, except the program hicFindEnrichedContacts (version 2.2.3) for observed vs. expected matrix generation. Briefly, raw fastq files were mapped to the Neurospora genome (either nc12 or nc14) using bowtie2 (version 2.3), with the --local and --reorder flags, and the initial, high resolution (0.5 kb bin) Hi-C contact matrix was built with hicBuildMatrix using the default settings (double digest matrices were built using the restriction and dangling sequences of both DpnII and MseI); lower resolution matrices were created by the program hicMergeMatrixBins. In situ Hi-C matrix files are listed in Supplementary Table 2. Pearson correlation between replicates was determined and plotted by the command hicCorrelate, while the plots of Hi-C counts in relation to genomic distance for replicate matrices, or different merged dataset resolutions, were generated by the program hicPlotDistVisCounts. Comparison of matrices was performed with the program hicCompareMatrices. Images displaying bin contacts in the in situ Hi-C matrix were made with hicPlotMatrix, and Neurospora "TADs" were plotted with the commands hicFindTads and hicPlotTads. Contact maps are either presented as raw or Knight-Ruiz corrected [bib_ref] A fast algorithm for matrix balancing, Knight [/bib_ref] images, the latter accounting for differences in restriction enzyme site distribution or sequencing bias of underlying DNA, while observed vs expected datasets are presented to show contact strength independent of genomic distance. Contact quantification was performed with the shell script process.sh, which uses the hicConvertFormat subprogram in hicExplorer to convert the Hierarchical Data Format version 5 [.h5] matrix produced from hicExplorer into a homer format matrix; the python script dataconvert.py converts the homer matrix into an NxN array, which is used by the python script epigenetic-mark-Quant-v2.py, adapted from [bib_ref] Neurospora chromosomes are organized by blocks of importin alpha-dependent heterochromatin that are..., Galazka [/bib_ref] , to calculate bins that are enriched for specific epigenetic marks and counts the number of enriched bins within and between chromosomes. All scripts are in the publicly available Klocko-Lab GitHub (https://github.com/ Klocko-Lab/Chip_Quantification; accessed 2022 September 3).
As replicate Hi-C datasets are locally reproducible, DpnII or MseI replicate fastq files were independently merged, resulting in a combined DpnII dataset with 269.9 million (M) reads (89.0M reads deemed "valid" by the stringent quality control standards used by HiCExplorer) and a merged MseI dataset with 93.8M reads . To generate a comprehensive dataset representative of the chromatin composition of the WT Neurospora genome from individual DpnII (euchromatic-specific; GATC recognition sequence) and MseI (heterochromatic-specific; TTAA recognition sequence) datasets, we calculated the percent of the Neurospora genome covered by H3K9me3 demarcating constitutive heterochromatin, which have an increased number of MseI sites due to the AT-rich nature of the underlying DNA. To this end, we generated bed files of a previously published H3K9me3 merged dataset [bib_ref] Loss of HP1 causes depletion of H3K27me3 from facultative heterochromatin and gain..., Jamieson [/bib_ref] [bib_ref] Nucleosome positioning by an evolutionarily conserved chromatin remodeler prevents aberrant DNA methylation..., Klocko [/bib_ref] , subtracted the end point from the starting point to calculate the number of basepairs covered by H3K9me3 at each region, summed the total number of bases, and divided the H3K9me3covered bases by the genome size to get the percentage (15.76%) of the Neurospora genome covered by H3K9me3. The comprehensive in situ Hi-C dataset was built by combining raw DpnII and MseI fastq files that contain the appropriate numbers of valid MseI or DpnII PE100 reads that, when combined, would provide 15.76% MseI-derived contacts and 84.24% DpnII-derived contacts. Here, 21.4M reads (3.8M valid reads) were removed from MseI replicate #4, meaning a fastq file containing 72.4M total MseI reads (16.7M valid MseI reads) was merged with a fastq file containing 270M total DpnII reads (89.2M valid DpnII reads). This single fastq dataset containing 342.2M DpnII and MseI reads (105.9M valid DpnII and MseI reads) was then used to build a high-resolution (0.5 kb bin) Hi-C contact matrix, and that comprehensive matrix (the "merged, single fastq" matrix) was used for subsequent contact analysis and generation of heatmaps for figures. A similar method was used to examine the robustness of merging DpnII and MseI datasets [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , where 23,340,020 valid reads from each of the 3 DpnII replicate fastq files and 4,359,980 valid reads from each of the 3 MseI replicate fastq files (giving 27.7M total valid reads at the percentages of the chromatin composition of the Neurospora genome) were selected and merged in every possible combination.
Neurospora crassa Chromatin Immunoprecipitation sequencing (ChIP-seq) datasets were reported previously (WT H3K9me3merged from GSE68897 and GSE98911; WT H3K27me2/3-merged from GSE82222 and GSE100770; WT H3K27ac-GSE118495; WT CenH3-GSE71024; WT H3K4me3-GSE121356; WT/Ash1L (Dset-2) H3K36me2-GSE118495) [bib_ref] Loss of HP1 causes depletion of H3K27me3 from facultative heterochromatin and gain..., Jamieson [/bib_ref] [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref] [bib_ref] Nucleosome positioning by an evolutionarily conserved chromatin remodeler prevents aberrant DNA methylation..., Klocko [/bib_ref] [bib_ref] Selection and Characterization of Mutants Defective in DNA Methylation in Neurospora crassa, Klocko [/bib_ref] [bib_ref] Histone H3 lysine 4 methyltransferase is required for facultative heterochromatin at specific..., Zhu [/bib_ref]. Datasets were downloaded and remapped to either the nc12 genome or our corrected nc14 genome. SAM file outputs were converted to sorted BAM files with SAMtools ), which were used to create bedgraph or bigwig files at 50 bp bin resolution with DeepTools (Ram ırez et al. 2016) for display on Integrative Genomics Viewer (IGV) [bib_ref] Integrative genomics viewer, Robinson [/bib_ref] ; the count feature of the IGVtools program within IGV was also used to create TDF files at 50 bp bin resolution. IGV images of ChIP-seq enrichment tracks were used for figure construction.
Neurospora WT and Ddim-5 polyadenine messenger RNA sequencing (polyA mRNA-seq) datasets were previously reported (GSE82222) . Here, processed HTseq count files [bib_ref] HTSeq-a Python framework to work with high-throughput sequencing data, Anders [/bib_ref] were downloaded and BED files of differentially expressed genes (here, gene expression in the mutant was up-or downregulated four times the WT expression level: increased genes have a log 2 value of 2 or more and decreased genes are log 2 ¼ -2 or less) were created with DESeq2 [bib_ref] Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2, Love [/bib_ref]. IGV images of RNA-seq bed files were used for figure construction.
# Results
In situ Hi-C of a wild type Neurospora strain using the restriction enzyme DpnII primarily captures the chromosome conformation of euchromatic regions of the genome Previous work elucidated the N. crassa genome organization at a lower resolution using proximity ligation Hi-C; the resulting contact matrices were generated and analyzed with a software pipeline that is no longer supported [bib_ref] Comprehensive mapping of long-range interactions reveals folding principles of the human genome, Lieberman-Aiden [/bib_ref] [bib_ref] Neurospora chromosomes are organized by blocks of importin alpha-dependent heterochromatin that are..., Galazka [/bib_ref] [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref]. Prior to new Hi-C library construction, we chose to switch to HiCExplorer, a comprehensive and easy-to-use data analysis package [bib_ref] High-resolution TADs reveal DNA sequences underlying genome organization in flies, Ram Irez [/bib_ref]. Re-analysis of the previously published proximity ligation Hi-C dataset of the wild type (WT) strain NMF39 (strain 74-OR23-1VA [FGSC #2489], asexually propagated in the Freitag lab, Oregon State University) confirmed the reproducibility of the HiCExplorer software package: a previously generated, yet unpublished, iteratively corrected [bib_ref] Iterative correction of Hi-C data reveals hallmarks of chromosome organization, Imakaev [/bib_ref] contact heatmap of the Neurospora LG II mirrored a Knight-Ruiz (Knight and Ruiz 2013) corrected contact heatmap generated by HiCExplorer [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , demonstrating the HiCExplorer package accurately generates and displays contact strength matrices for N. crassa Hi-C datasets.
The published proximity ligation Hi-C dataset captured chromatin interactions using the restriction enzyme HindIII , which cleaves a less common restriction site sequence (A^AGCTT; sites every 4,096 basepairs [bp] on average) [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. However, recent work has cast doubt on the accuracy of proximity ligation Hi-C [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref] [bib_ref] Comparison of Hi-C results using in-solution versus innucleus ligation, Nagano [/bib_ref] and we were concerned that the paucity of HindIII restriction sites would fail to capture crucial long-range contacts at individual genes. Therefore, we adapted and refined our Hi-C protocol to capture chromosome conformation in fungal nuclei (in situ) [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref] [bib_ref] Architectural alterations of the fission yeast genome during the cell cycle, Tanizawa [/bib_ref] using DpnII (^GATC; sites every 256 bp on average; enzymatic activity not inhibited by cytosine methylation, allowing ligation of methylated genomic regions), as DpnII restriction sites are more abundant in gene-rich euchromatin [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Using a single WT strain (N150; 74-OR23-1VA), we initially generated 3 replicate DpnII in situ Hi-C libraries amplified with 15 PCR cycles for the final library barcoding step [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref] , akin to previous Hi-C library generation [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref] ; these replicate datasets are highly similar when compared [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref]. However, Neurospora heterochromatic regions, rich in adenine and thymine DNA basepairs, are known to be depleted by increased numbers of PCR cycles during library barcoding [bib_ref] Methylated DNA is over-represented in whole-genome bisulfite sequencing data, Ji [/bib_ref] , suggesting 15 PCR cycle libraries may not accurately report heterochromatic contacts. Therefore, we repeated the Hi-C library construction using the same frozen spheroplast samples of WT N150, amplifying the final Hi-C library with only 8 PCR cycles [bib_ref] Methylated DNA is over-represented in whole-genome bisulfite sequencing data, Ji [/bib_ref] ; we also reamplified our Hi-C ligation products bound to streptavidin beads (previously with 15 PCR cycles) with another 8 cycles (a "23 PCR cycle library") to assess if library reamplification also underrepresented AT-rich regions. Comparison of the 8 and 15 PCR replicates show heterochromatin depletion in the latter [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref] , while the "23 PCR cycle" replicates [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref] qualitatively exhibit even greater AT-rich depletion than 15 PCR cycle libraries [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref]. Unsurprisingly, independently merging the 8, 15, and 23 PCR cycle replicates into separate datasets for comparison shows similar results: the 8 PCR cycle merged matrix has more AT-rich heterochromatin signal as compared to the merged 15 and 23 PCR cycle datasets . , showing tracks of H3K9me3 ChIP-seq (delineating heterochromatic regions), genes (track with rectangles; several NCU gene numbers are shown), and the restriction enzyme sites and recognition sequences of HindIII, DpnII, and MseI. Genomic distance indicated above the tracks, and ChIP-seq enrichment scale to the right. b) Heatmap of the chromosome conformation of LG II, generated from the merged DpnII in situ Hi-C dataset, combining 4 replicates (this work; above diagonal), and the previously published proximal ligation Hi-C (HindIII) dataset (below diagonal) . Each heatmap is displayed as the raw read count per 20 kilobase (kb) bins; here, and in images throughout this work, only one half of an individual square plot is included in figures to allow comparisons, since each contact heatmap generated is a mirror image reflected about the diagonal. All images in this manuscript are similarly displayed, with the number of bins (vertical markings) and genomic distance, in megabases (Mb) (horizontal markings) shown on the plot axes. Scale bars, provided on all images or groups of images, indicate the number of mapped reads per bin, either on a log 10 scale (log transformed) or as absolute values. CenH3 and H3K9me3 ChIP-seq tracks presented above and to the left to indicate the centromeric and heterochromatic regions of LG II, respectively; ChIP-seq enrichment scales shown; similar scales are used for ChIP-seq enrichment tracks throughout the manuscript. c) Heatmap displaying the merged DpnII in situ Hi-C dataset plotting raw read counts per 10 kb bin (above diagonal) or 5 kb bin (below diagonal). d) Heatmap displaying the merged DpnII in situ Hi-C dataset of raw read counts (above diagonal) and the calculated observed vs. expected plot (below diagonal) to highlight genomic interactions different from those expected, based on the inverse relationship of contact strength vs. linear distance. Both plots are at 10 kb resolution; scale bar of log 10 observed vs. expected contact strength shown below. e) Heatmap displaying the merged DpnII in situ Hi-C dataset of raw read count per 20 kb bin (above diagonal) and the Knight-Ruiz (KR) corrected (below diagonal) read count per 20 kb bin.
[formula] (a) (b) ( c) (e) (d) [/formula]
Given the depletion of AT-rich DNA, we abandoned the 15 and 23 PCR cycle in situ Hi-C datasets, and from this point forward, exclusively analyzed the 8 PCR cycle DpnII in situ Hi-C libraries.
To confirm our 3 N150 replicates with a biological replicate of an independently propagated strain, we generated another DpnII in situ Hi-C library of the WT strain N3752 (74-OR23-1VA) for 4 WT DpnII chromosome conformation replicates. Heatmaps of contact matrices indicate the 4 replicates are highly reproducible and include strong contacts off-diagonal at the highest resolutions . When the replicate contact matrices were compared, 1 replicate of N150 and the N3752 replicate show fewer euchromatic-heterochromatic contacts , possibly reflecting subtle differences in the distribution of nuclei in cell cycle stages [bib_ref] Organization of the mitotic chromosome, Naumova [/bib_ref] [bib_ref] Architectural alterations of the fission yeast genome during the cell cycle, Tanizawa [/bib_ref]. Raw and corrected replicates correlate well: each display the inverse relationship between contact strength and genomic distance typical of Hi-C datasets , c and d), and Pearson scores of replicate Hi-C contact matrices range between 0.86 and 0.92 , verifying replicate robustness. We note that all in situ Hi-C contact matrices in this manuscript are derived from the average genome topology across an entire population of unsynchronized nuclei, which limits the detail of chromatin structure obtained from individual nuclei.
As local euchromatic contacts were highly reproducible, we merged all 4 replicates into a single DpnII in situ Hi-C dataset, which when combined, had 269.9M total reads, 89.0M of which were deemed valid Hi-C ligation products using the stringent quality control standards in HiCExplorer; considering the 136,425 DpnII sites in the Neurospora genome, this reflects $654 contacts per site on average. To assess the accuracy of the in situ protocol at capturing chromatin contacts exclusively in the nucleus, we examined the number of ligation products containing reads that map to the mitochondrial genome. Our merged DpnII in situ Hi-C data have substantially reduced numbers of mitochondrialspecific reads as Hi-C ligation products relative to the published proximity ligation dataset . These data suggest our merged DpnII in situ Hi-C dataset accurately captured genomic contacts in the nucleus.
To assess if this new dataset reflects Neurospora chromosome conformation at a high resolution, we compared the merged in situ DpnII dataset to the published proximity ligation HindIII data containing $11.2M valid contacts ); the latter is typically displayed with chromosome-level contact matrices using 40 kb bins. Our merged in situ raw contact count matrix displays a greater saturation of data points and density of information across a single, representative chromosome (LG II) at a higher (20 kb) resolution [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , reflecting the capture of a high number of short-and long-range contacts in euchromatic regions. Heterochromatin, marked by H3K9me3 and/or the centromeric histone variant CenH3 [bib_ref] Trimethylated lysine 9 of histone H3 is a mark for DNA methylation..., Tamaru [/bib_ref] [bib_ref] Heterochromatin is required for normal distribution of Neurospora crassa CenH3, Smith [/bib_ref] , has few contacts with euchromatin [fig_ref] Figure 7: Model of genome topology of N [/fig_ref] show similar results; intrachromosomal contacts are slightly enriched relative to contacts between chromosomes across the whole genome. At higher resolutions, including 10 and 5 kb bins, numerous local intrachromosomal contacts are observed [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , demonstrating the depth of our merged DpnII dataset. To examine chromosomal interactions independent of genomic distance, we plotted the log 10 change in observed contacts within the merged DpnII dataset relative to the expected inverse relationship between contact strength and genomic distance (e.g. the contact strength between 2 loci should decrease as the distance separating the loci increase); any deviations between observed contact numbers and the number of expected interactions highlight strong interactions, or a dearth of contacts, between genome features not occurring by chance. Observed contacts are massively increased above expected at centromere flanks and subtelomeres, while few contacts between heterochromatin and euchromatin are observed [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , consistent with heterochromatic regions strongly associating and thereby segregating euchromatin across the Neurospora genome [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref]. strong
To balance our merged matrix, we employed Knight-Ruiz (KR) correction to account for differences in restriction site number or potential ligation/sequencing bias in our Hi-C libraries [bib_ref] A fast algorithm for matrix balancing, Knight [/bib_ref] [bib_ref] High-resolution TADs reveal DNA sequences underlying genome organization in flies, Ram Irez [/bib_ref] , possibly at the expense of local 3-dimensional chromatin structures, as correcting bin signals has the potential to nullify chromatin contacts; to allow for objective data interpretation, many figures in this manuscript present both raw and KR corrected matrix heatmaps. Compared to the raw contact matrix, the KR-corrected DpnII dataset displays uniform intrachromosomal euchromatic contacts, while more distant heterochromatic region interaction signals became stronger [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] and 7; below diagonal). We note some smaller heterochromatic regions have increased euchromatic contacts upon KR-correction (evident by the loss of lowsignal horizontal/vertical lines emanating from heterochromatic regions), suggesting matrix correction may somehow bias chromatin contacts between compartments. All told, our merged in situ DpnII Hi-C dataset captures the Neurospora chromosome conformation at a high resolution.
Heterochromatin specific chromosome conformation is revealed by in situ Hi-C with MseI While our DpnII in situ Hi-C dataset efficiently monitors euchromatic contacts, we were concerned by the paucity of silent chromatin interactions in these data. Previous reports demonstrated strong interactions in heterochromatic regions in Neurospora [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref] , but the reduced number of DpnII sites in AT-rich DNA may compromise any assessment of long-range contacts for heterochromatic DNA [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Indeed, heatmaps of raw contact numbers display few interheterochromatic contacts, but KR-corrected heatmaps, which account for restriction site differences in interaction bins, show strong, longrange association of heterochromatic regions [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Therefore, we adapted our Hi-C protocol for the restriction enzyme MseI, as its T^TAA recognition sequence is highly prevalent in AT-rich, heterochromatic DNA [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] ; [bib_ref] Nucleosome positioning by an evolutionarily conserved chromatin remodeler prevents aberrant DNA methylation..., Klocko [/bib_ref] , allowing efficient capture of chromosome conformation in gene-poor, H3K9me3-marked silent regions of the Neurospora genome. We performed in situ Hi-C using MseI, generating 2 replicates of the WT N150 strain and a third biological replicate of the independently propagated WT strain N3752. Heatmaps of MseI replicates show reproducibility in heterochromatic contact capture across a single chromosome, with strong Hi-C signal presenting at heterochromatic regions in high-resolution matrices . The 3 replicates show similar heterochromatic interactions with variability in medium range off-diagonal euchromatic contacts (within $100 kb), possibly due to variable capture across the reduced number of MseI sites found in euchromatin [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Each MseI replicate dataset is well correlated, with similar contacts and strong Pearson correlation scores . All raw and KR-corrected replicate matrices display the typical, inverse relationship between contact strength and genomic distance in plots of genomic distance versus Hi-C contact numbers .
Since the 3 MseI replicates are reproducible, we merged the replicates into a single MseI in situ Hi-C dataset, containing 72.4 million reads, 16.7 million of which were deemed valid by HiCExplorer quality control parameters. Similar to DpnII Hi-C datasets, MseI Hi-C datasets show few mitochondrial ligation products, confirming the accuracy of the in situ Hi-C protocol . Considering the 195,726 MseI sites across the Neurospora genome, the number of MseI reads represents $85 contacts per site-fewer than our DpnII data-but given the density of MseI sites in heterochromatic AT-rich DNA [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , these ligation products are more likely to capture internal folding of, and contacts between, silent chromatin. In support, plots of genomic distance vs. Hi-C read counts of merged matrices (both raw and KR-corrected) show increased short-range and reduced long-range contacts, presumably within and between heterochromatic regions (below) in MseI contact matrices relative to DpnII datasets , a and b).
Raw contact count heatmaps of the merged MseI in situ Hi-C matrix across LG II mainly display increased Hi-C interactions within and between constitutive heterochromatic regions: internally, strong contacts occur across the entire length of silent regions, while strong, long-range contacts involve distant heterochromatic regions across Megabases (Mb) of linear chromosomal distance; strong euchromatic contacts are primarily restricted to neighboring bins on-diagonal [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref] ; above diagonal). Similar results were observed for the other 6 chromosomes [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , above diagonal). KR correction of MseI matrices eliminates most long-range heterochromatic contacts, potentially due to disparate MseI site positions and resultant discrepancies in silent and active chromatin contact strength [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] ; below diagonal), again highlighting our caution in exclusively relying on KR-corrected matrices to elucidate valid chromatin structures. Our MseI data effectively captures intra-and interchromosomal heterochromatic interactions, as strong contacts between centromeres are observed, as are interactions between subtelomeres; interspersed heterochromatic regions also associate, but these silent regions are segregated from neighboring euchromatin, as shown in single chromosome [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] and whole genome contact heatmaps [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , further supporting the bundling of heterochromatin .
To directly compare the heterochromatic contacts captured in MseI in situ Hi-C matrices to our euchromatic-specific DpnII in situ Hi-C data, we examined the DpnII and MseI datasets on a magnified, $1 Mb region of the right arm of LG II that extends from $3.5 to 4.5 Mb and encompasses several heterochromatic regions. The MseI Hi-C data effectively captured heterochromatin contacts in this region, which are displayed as strong "triangles" of Hi-C interactions [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref]. Adjacent H3K9me3-marked regions are more apt to associate, but strong contacts between heterochromatic regions separated by distances of $1 Mb were still observed [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref]. In contrast, the strongest contacts in gene-rich euchromatin occur locally, and despite the proximity of silent chromatin on the linear chromosome, few euchromatinheterochromatin interactions are seen at this resolution [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref] ; bottom). In fact, neighboring euchromatic regions are more inclined to associate [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref]. Taken together, DpnII and MseI in situ Hi-C have the capacity to independently capture and assess euchromatic and heterochromatic contacts, respectively, across the Neurospora genome.
MseI in situ Hi-C facilitates the correction of the N. crassa reference genome Initially, we used version 12 of the Neurospora genome (nc12) for our Hi-C data processing; this genome version contains 7 chromosomes (Linkage Groups and 13 other Supercontigs (SCs), numbered 8-20, whose location in the genome could not be determined, possibly due to repetitive sequences at SC boundaries [bib_ref] The genome sequence of the filamentous fungus Neurospora crassa, Galagan [/bib_ref] [bib_ref] Neurospora chromosomes are organized by blocks of importin alpha-dependent heterochromatin that are..., Galazka [/bib_ref]. Our MseI data show strong interactions between the unplaced SCs and LGs I and V, the latter 2 exhibiting gaps in local contacts [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , suggesting SCs 8-20 are found in these chromosomes. Therefore, we corrected the Neurospora reference genome, creating version 14 (nc14; the number was chosen to account for a previous correction of LG VI) ) by placing the DNA sequences of SCs 8-20 in the reference genome fasta file using MseI Hi-C data to guide sequence placement. Despite these improvements, nc14 still contains numerous gaps that could not be corrected with Hi-C data.
First, we noticed that SCs 10, 11, 13, and 14 strongly associate with a small heterochromatic region on LG I, centered $6.40 Mb in the nc12 genome [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Comparison of these small SCs showed their approximate order: the strongest contacts between these small SCs highlight their order in the LG [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , while binary comparisons between the small SCs and LG I (from 6.15 to 6.55 Mb) confirmed insertion endpoints [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Together, these data suggest an order of SC 14, 10, 13, with SC 11 slightly downstream, all of which are internal to the LG I heterochromatic region [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Further, we noticed SC 17 is located at $3.7 Mb in LG I, within the centromere, as shown by the binary comparison of SC 17 and the LG I centromere [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. We corrected LG I in the nc14 genome by placing SCs 14, 10, 13, 11, and 17 into locations with unknown "N" bases, which creates a $200 kb larger heterochromatic region from $6.4 to 6.65 Mb and extends the centromere [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Remapping either our MseI Hi-C reads or published H3K9me3 Chromatin Immunoprecipitation sequencing (ChIP-seq) data [bib_ref] Loss of HP1 causes depletion of H3K27me3 from facultative heterochromatin and gain..., Jamieson [/bib_ref] [bib_ref] Nucleosome positioning by an evolutionarily conserved chromatin remodeler prevents aberrant DNA methylation..., Klocko [/bib_ref] to the amended nc14 genome accurately reflects internal and long-range heterochromatic contacts across LG I and H3K9me3 enrichment of the corrected silent regions [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , b and f).
The left arm of Neurospora LG V was also corrected: this is the location of the repeated ribosomal DNA gene copies comprising the Nucleolus Organizing Region (NOR). Previous work suggested localize to LG V, but the order was mostly unknown. Our MseI Hi-C data independently confirmed that these smaller SCs are present on LG V [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , and examination of the SC 9 sequence showed the presence of telomeric repeats, supporting the hypothesis that SC 9 terminates the LG V left arm [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Interchromosomal comparisons of smaller SCs confirmed that the LG V left arm begins with SC 9, followed by SCs 8, 15, 19, 18, 12, and 16 [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , c and d). We inserted these SCs into the nc14 genome for improving the LG V left arm sequence and remapped our MseI data. The resulting Hi-C heatmap clarifies the organization of LG V, showing that the NOR on the left arm of LG V is sequestered from most other genomic loci, although some heterochromatic contacts exist [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Closer examination of the MseI Hi-C data of the LG V left arm shows a cross-shaped pattern centered on the rDNA gene NCU15719 [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , which was also observed on SC 8 in nc12 [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , suggesting the sequence of this rDNA gene is repeated multiple times, preventing discrimination between individual rDNA gene copies in genomic experiments (note: SC 20 is identical to NCU15719 and was excluded). While Hi-C contact data implies promiscuity in NCU15719 genomic contacts, the increased copy number of this gene may bias the Hi-C data, making it incorrectly appear that the NCU15719 rDNA gene has increased Hi-C interactions. Previous estimates place the total number of rDNA copies between 130 and 172 [bib_ref] Premeiotic change of nucleolus organizer size in Neurospora, Butler [/bib_ref] , and since older versions of the Neurospora genome had 55 rDNA gene copies [bib_ref] The genome sequence of the filamentous fungus Neurospora crassa, Galagan [/bib_ref] , NCU15719 could be repeated $75 to 117 times, a copy number supported by the enrichment of CenH3 ChIP-seq data relative to background [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. All told, our MseI Hi-C data assisted in creating a more accurate reference genome for N. crassa.
The complete genome topology of N. crassa is revealed upon merging DpnII and MseI datasets at proportions consistent with the chromatin composition of the Neurospora genome Using DpnII and MseI restriction enzymes, we independently captured the chromosome conformation of GC-rich euchromatic and AT-rich heterochromatic regions, respectively (above). To more comprehensively explore Neurospora's genome topology, we envisioned merging fastq files of our DpnII and MseI datasets and building a single contact matrix. First, to test if merging DpnII and MseI fastq files into a single matrix would reproducibly and accurately present Hi-C contact data, we built 9 different fastq file datasets to examine if unique combinations of Hi-C reads produce similar contact matrices. To this end, we selected valid reads from 3 DpnII replicates and 3 MseI replicates at a ratio representative of the chromatin composition of the Neurospora genome: each DpnII replicate fastq file had 23.34M (84.24%) valid reads and each MseI replicate fastq file had 4.36M (15.76%) valid reads; the latter mirrors the percentage of genomic H3K9me3 (see Materials and Methods). Upon merging all combinations of DpnII and MseI replicate fastq files, contact matrices were built from the 9 fastq datasets, each with 27.7M valid reads, and Hi-C contacts were displayed in heatmaps. The resulting 9 heatmaps are essentially identical [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , and binary comparisons of Hi-C contacts in the 9 matrices are highly correlated, with Pearson correlation values between 0.83 and 0.99 [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. These data, along with the similarity to a double digest dataset (below), demonstrates that merging fastq files to produce a single in situ Hi-C contact matrix is a valid approach. Next, we generated a comprehensive DpnII and MseI Hi-C dataset of the Neurospora genome by merging fastq files with percentages of valid Hi-C reads that reflect the Neurospora genome chromatin, mapping those reads to the nc14 reference genome, and subsequently building the contact matrix; we note that merging preprocessed matrices introduces mathematical bias [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. This comprehensive dataset has 342M reads, 106M of which are considered valid Hi-C contacts . Plots of genomic distance vs. raw and KR-corrected contact numbers across multiple resolutions, from 50 to 1 kb bins, showed the typical inverse relationship in the number of Hi-C contacts relative to genomic distance for this single fastq contact matrix. Even at the highest resolutions, bins separated by $200 kb to 1 Mb of linear genomic distance average at least 1 contact, indicating high contact data depth , c and d). Raw and KR-corrected heatmaps of this comprehensive DpnII/MseI merged, single fastq dataset at 20 and 10 kb bin resolution across LG II show similar interactions to the DpnII-only heatmap, but heterochromatic contacts are more prevalent (compare Figs. 3a and 1e). In particular, the KR-corrected matrix displays higher saturation of, and more detail for, silent chromatin contacts, including intracentromeric and long-range heterochromatic region interactions ; the other 6 Neurospora chromosomes have similarly improved heterochromatin contacts [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Long-range interactions between chromosomes are readily observed in raw and KR-corrected contact matrices, including strong bundling of heterochromatin and contacts between euchromatic chromosome arms . Across the entire genome, the 7 centromeres associate but are isolated from other chromosomal features, including the 14 independently clustered telomeres. Subtle contacts between the euchromatic arms of the 7 Neurospora chromosomes are also prevalent [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. The right arm of the KRcorrected LG II at a high resolution (5 kb bins) elucidates topological differences between silent and active chromatin , as heterochromatin strongly associates across hundreds of kilobases of linear distance, while regions of gene-rich active DNA are more apt to contact nearby euchromatin and remain segregated from heterochromatin.
To examine chromatin interactions independent of genomic distance, we plotted the difference in the observed Hi-C contact signal relative to the expected signal-derived from the inverse relationship of contact strength to genomic distance-across and internal to a single chromosome (LG II) (Figs. 3c, bottom, d and e), all 7 Neurospora chromosomes [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , and across the entire Neurospora genome [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. We observe centromeres and interspersed heterochromatic regions, as well as the flanking euchromatic DNA surrounding these silent regions, contact more than expected, indicating that internal compaction of heterochromatin causes neighboring euchromatin to strongly interact [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] ; centromere clustering across the genome is also observed [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Similarly, intra-and interchromosomal subtelomeres show strong observed contacts [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , indicating chromosome folding facilitates the association of subtelomeres. Closer examination of Hi-C contacts independent of genomic distance on the right arm of LG II shows strong interactions between heterochromatic regions, increased associations between euchromatic regions, and few interactions between silent and active chromatin , although the latter intercompartment contacts are possibly depleted in this dataset, given that merging datasets with different underlying biases impacts matrix balancing (see Discussion). Despite this caveat, our comprehensive in situ Hi-C dataset, which merges independent DpnII and MseI datasets as a single fastq for generating the contact matrix, collectively illuminates the interactions of individual genome features at a high resolution.
Double digest in situ Hi-C of N. crassa elucidates fungal genome topology and supports merging DpnII and MseI fastq files for a comprehensive Hi-C dataset
To expand our analysis of Neurospora genome architecture and confirm the robustness of our merged DpnII and MseI dataset, we generated DpnII/MseI double digest in situ Hi-C datasets, as ligation of DNA between euchromatic and heterochromatic regions may be impacted in single enzyme Hi-C. Since blunting of sticky ends occurs before ligation in Hi-C, differentially digested DNA molecules can still be ligated [bib_ref] Hi-C 3.0: improved protocol for genome-wide chromosome conformation capture, Lafontaine [/bib_ref]. We generated 5 double digest replicates of the N150 strain and a biological replicate using the N3752 strain. While our double digest replicates have varying valid read numbers (Supplementary , all display similar Hi-C contacts at a low resolution across a single chromosome (LG II) and at a high resolution across a small portion of the LG II right arm [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref] ; 2 replicates have mostly on-diagonal contacts, possibly due to reduced valid read numbers (Supplementary . All replicates, both as raw and KR corrected 50 kb bins, display similar inverse relationships between genomic distance and Hi-C contact numbers [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref] , and have high Pearson correlation scores when double digest replicates were compared in pairs [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref].
Given the similarity of double digest replicates, we merged all replicates into a single DpnII/MseI double digest dataset that comprises 456M reads, 76.5M of which are valid Hi-C contacts (Supplementary -less than our merged, single fastq dataset (above) but vastly more than the published proximity ligation dataset ). Graphs of raw or KR corrected Hi-C contacts in the double digest vs genomic distance at multiple resolutions show the typical inverse relationship between Hi-C contacts and genomic distance , e and f). Heatmaps of raw and KR-corrected interactions of the merged double digest dataset across LG II at 10 kb bin resolution are highly comparable to heatmaps of the merged, single fastq dataset, with the exception that long-range heterochromatic interactions are less prevalent (compare . Similar contacts are observed for the other 6 Neurospora chromosomes and genome wide [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref]. Plots of the observed vs. expected double digest dataset interactions, which highlight nonrandom contacts independent of genomic distance, show a similar chromatin topology as the merged, single fastq dataset: reduced heterochromatin-euchromatin interactions, but increased telomere bundling and a greater propensity of centromere flanks to contact , although the double digest dataset displays slightly less-but still overall increased-compaction within centromeres . Closer examination of the LG II right arm shows strong contacts between euchromatic regions despite the presence of intervening heterochromatin ; increased association of euchromatic regions is highlighted in plots of contacts independent of genomic distance . We observe reduced longrange heterochromatic contacts across the LG II right arm , possibly due to the increased likelihood of forming ligation products between silent and active DNA in the double digest reaction.
To compare the double digest dataset to the single fastq merged dataset, we merged 64.5M valid DpnII reads with 12.0M valid MseI reads (76.5M valid reads overall) in a single fastq file, mapped the reads to nc14, built the contact matrix, and compared this merged dataset to the double digest dataset across and within LG II to highlight any discrepancies between these 2 Hi-C datasets. While euchromatic contacts are virtually identical in raw and KR-corrected double digest datasets relative to the merged, single fastq Hi-C dataset , e and f), the raw Hi-C contact matrix of the double digest dataset has reduced contacts within and between heterochromatic regions . KR-correction of the double digest dataset partially alleviates the paucity of heterochromatic contacts but increases heterochromatic-euchromatic interactions Comprehensive genome organization of Neurospora crassa created from merging DpnII (euchromatic) and MseI (heterochromatic) fastq files to build a single contact matrix. a). Heatmaps displaying the raw (above diagonal) and corrected read counts (below diagonal) of genomic interactions at 20 kb (top heatmap) and 10 kb (bottom heatmap) bin resolution of the complete in situ Hi-C dataset, created from merging DpnII and MseI fastq files into a single fastq file, which was processed into a single contact matrix to avoid mathematical bias; this single matrix contains valid DpnII (89.24M) and MseI (16.69M) Hi-C reads at a ratio that accurately depicts the percentages of euchromatin (84.24%) and heterochromatin (15.76%), respectively, in the Neurospora genome (see Materials and Methods). CenH3, H3K9me3, and H3K27me2/3 ChIP-seq enrichment tracks, as well as a gene track, presented above and to the left of each plot. b). Heatmaps plotting interchromosomal contacts at 20 kb bin resolution between LG II and LG III created using raw read count (left) or the KR-corrected read count (right) matrices. CenH3 and H3K9me3 ChIP-seq enrichment tracks presented above and to the left of each plot. c). Triangle heatmaps of KR-corrected read count (top) and the plot of the calculated observed vs expected of KR-corrected read count of Hi-C contacts at 5 kb bin resolution of the right arm of LG II (identical to the region shown in [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref]. Both triangle heatmaps were cropped and rotated 45 from square (reflective) heatmaps. Central IGV image displays tracks of MseI (TTAA) and DpnII (GATC) restriction sites, H3K9me3 ChIPseq enrichment and genes. d). Heatmap of the calculated observed vs. expected signal of KR-corrected read count (above diagonal) and the KR-corrected read counts of the complete DpnII-MseI in situ Hi-C dataset (below diagonal) at 10 kb bin resolution. Centromeric, intratelomeric, and interspersed heterochromatic regions highlighted in E are marked by colored lines. e) Heatmaps of the calculated observed vs expected KR-corrected contacts for the centromere, an interspersed heterochromatic region , and intrachromosomal subtelomere interactions of LG II highlighted in panel (d). CenH3, H3K9me3, and H3K27me2/3 ChIP-seq enrichment, as well as gene tracks presented below each plot.
## And 24a). quantification of strong interactions between
H3K9me3-marked regions independent of genomic distance (e.g. in observed vs expected matrices) shows reduced numbers of heterochromatic contacts in the double digest dataset, particularly for contacts between interchromosomal heterochromatic regions [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref]. Despite this subtle reduction in heterochromatic contacts, we conclude that our DpnII/ MseI double digest dataset effectively captures Neurospora genome topology in a manner similar to the merged, single fastq DpnII and MseI contact matrix.
## Examination of chromatin folding at a high resolution
Using both the merged, single fastq and double digest datasets, we examined major intra-and interchromosomal contacts of silent regions, including centromeres, subtelomeres, and interspersed heterochromatic regions, to characterize their topology at a high resolution. Centromeric chromatin ("centrochromatin"), enriched with the centromere-specific histone variant CenH3 [bib_ref] Heterochromatin is required for normal distribution of Neurospora crassa CenH3, Smith [/bib_ref] , strongly interacts, and is random and self-contained , [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref]. All centrochromatin contacts present equally, especially in KR-corrected heatmaps, with a hierarchical compaction emanating from CenH3-containing nucleosomes. Most centromeric DNA is occluded from other DNA, apart from nearby euchromatic loops that contact presumably surfaceexposed centromeric chromatin. Flanking pericentromeric regions, which each have robust interactions, strongly associate-implying a folded structure [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref]. While centromeric nucleosomes containing CenH3 are still marked with H3K9me3, centromeres minimally associate with neighboring constitutive heterochromatic regions, and no interactions with H3K27me2/3-marked facultative heterochromatin occur. Apart from strong interchromosomal centromeric contacts [fig_ref] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII [/fig_ref] , the (a) ( d) . Comprehensive genome organization of Neurospora crassa derived from the DpnII and MseI double digest contact matrices. a) Heatmap displaying the raw (above diagonal) and corrected read counts (below diagonal) of genomic interactions 10 kb bin resolution of the DpnII and MseI double digest in situ Hi-C dataset, which contained 76.5M valid reads. CenH3, H3K9me3, and H3K27me2/3 ChIP-seq enrichment tracks, as well as a gene track, presented above and to the left of the heatmap. b) Heatmap of the calculated observed vs expected signal of KR-corrected read count (above diagonal) and the KR-corrected read counts of the DpnII and MseI double digest in situ Hi-C dataset (below diagonal) at 10 kb resolution. Centromeric region highlighted in C is marked by a diagonal line. c) Heatmap of the calculated observed vs expected signal of the KR-corrected read counts across the centromere at 5 kb resolution. d) Triangle heatmaps of KR-corrected read count (top) and the plot of the calculated observed vs expected signal of KR-corrected read count at 5 kb bin resolution of the right arm of LG II, as in , e, f) Heatmaps displaying the log 2 ratio of in situ Hi-C contacts at 10 kb resolution comparing the 76.5M valid reads of the double digest to a merged fastq file that combines fastq files containing 64.5M valid DpnII reads with 12.0M valid MseI reads (76.5M valid reads overall), processed as a single matrix. The change in raw read counts (above diagonal) or KR-corrected counts (below diagonal), across (e) LG II or (f), the centromere of LG II is displayed. CenH3, H3K9me3, and H3K27me2/3 ChIP-seq enrichment, as well as gene tracks presented with each heatmap. . Characterization of individual chromatin region topology within the Neurospora genome. a) Heatmaps displaying the raw (above diagonal) and corrected read count (below diagonal) of interactions at and surrounding the LG IV centromere at 2.5 kb bin resolution. The heatmap to the left uses the single in situ Hi-C matrix merged from MseI and DpnII fastq files (with 106M valid reads), while the right heatmap uses the DpnII and MseI double digest contact matrix. Arrowheads highlight pericentromeric interactions. CenH3 and H3K9me3 ChIP-seq enrichment tracks presented above and to the left of the image, respectively. b) Heatmaps displaying the raw read count of contacts between subtelomere IIL, and subtelomeres IIR, IVL, and IVR at 2.5 kb bin resolution in the DpnII and MseI double digest contact matrix. Subtelomere schematics (below and left) and lines indicate positions of apparent subtelomeres marked by H3K9me3; telomeres consist of $20 repeats of 5 0 -TTAGGG-3 0 , averaging $120 bp, at chromosome ends [bib_ref] Characterization of chromosome ends in the filamentous fungus Neurospora crassa, Wu [/bib_ref]. H3K9me3 and H3K27me2/3 ChIP-seq enrichment tracks presented below and to the left at same y-axis scales as in . KR-(Continued) H3K9me3/CenH3-marked centromeres are essentially segregated from other silent loci, supporting previous observations [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref].
[formula] (c) (b) (f) (e)(a) (b) (c) (d) (e) [/formula]
At chromosome termini, the subtelomeres-defined as $100-300 kb regions that extend beyond the repetitive telomeric DNA that are enriched with both H3K9me3 and H3K27me2/3 [bib_ref] Regional control of histone H3 lysine 27 methylation in Neurospora, Jamieson [/bib_ref] [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref] -strongly associate but segregate from centromeres (e.g. [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref]. The strongest contacts between the left and right subtelomeres of individual chromosomes, as well as interchromosomal subtelomere contacts, overlap H3K27me2/3 enriched regions, as evident in raw and KR-corrected contact heatmaps of telomere pairs [fig_ref] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI [/fig_ref] ; the strength of the individual binary pairings varies, implying an unknown organizational system , compare contacts of right (R) subtelomere of LG II and left (L) and right subtelomeres of LG IV with the left subtelomere of LG II; Supplementary Figs. S26-S28]. In contrast, individual subtelomeres have extensive contacts at, and extending from, the telomere across H3K9me3-enriched chromatin: several subtelomeres, including LG IIL, strongly contact neighboring constitutive heterochromatic regions LG IIL). Other subtelomeres contact nearby euchromatin (e.g. LG IIR, gray arrowheads) or reside in larger heterochromatic domains (e.g. LG IVL/R; .
Smaller, interspersed H3K9me3-rich constitutive heterochromatic regions have increased contacts that are occluded from facultative heterochromatin or euchromatin . The majority of these regions are flanked by extensive contacts across $two to five 500 bp bins at 1 or both ends of the H3K9me3 enriched region , suggesting the heterochromatic boundary associates with internal heterochromatin. A set of 100 randomly selected interspersed heterochromatic regions from all 7 chromosomes were inspected for increased heterochromatin/euchromatin boundary associations: 73 silent chromatin regions show increased interactions on both sides, and 94 silent regions have at least 1 side with increased contacts, possibly to prevent heterochromatin spread. All told, gene-poor, constitutive heterochromatic regions stochastically fold and strongly associate in the silent (B) compartment; inside this compartment, centromeres are segregated from other silent heterochromatic regions.
To examine euchromatin folding, we plotted nonlog transformed KR-corrected heatmaps in both our merged, single fastq and double digest in situ Hi-C datasets to highlight off-diagonal structures. We observe that euchromatin forms "globules" that are approximately 20-40 kb in size blue line highlights globule boundaries), suggesting the chromatin fibers in active genic regions locally compact into small loops, highlighting a hierarchical organization to fold gene rich Neurospora DNA into the active "A" compartment in the center of the nucleus.
## Acetylated or methylated euchromatic genes associate with constitutive heterochromatin for gene regulation
To examine the topology of euchromatin for TAD-like structural elements, we used the hicPlotTads program to examine Neurospora euchromatin in our single fastq, merged in situ Hi-C dataset at increasingly higher resolutions. This program defined several TADs internal to euchromatic domains , thin line overlay), although visual inspection showed that the predicted TAD boundaries are most likely incorrect, as euchromatin outside of predicted TAD boundaries still strongly associates with TAD-internal euchromatin . As computational TAD prediction can vary [bib_ref] Comparison of computational methods for the identification of topologically associating domains, Zufferey [/bib_ref] , it is likely that these TAD boundaries were improperly called. We hesitate to argue that the Neurospora genome is organized by true TAD-like structures, considering the definition and establishment of TADs (see Discussion). However, this failed TAD prediction did have an overall positive outcome.
As we generated hicPlotTads output files, we displayed several published ChIP-seq tracks beneath high-resolution Hi-C contact heatmaps to glean information about the underlying euchromatin composition. We were struck by the presence of thin interactions between H3K9me3-marked constitutive heterochromatin and nearby euchromatin . These euchromatic-heterochromatic interactions are observed at all resolutions examined but are best displayed at higher resolutions ( 2.5 kb bins), suggesting only small segments of euchromatic DNA associate with heterochromatin . Closer examination showed that interacting euchromatin is enriched for histone post-translational modifications, including the active marks H3K27ac and promoter-specific H3K4me3, as well as the repressive H3K36me2 catalyzed by ASH1 here, Dset-2 H3K36me2 ChIP-seq elucidates the ASH1-specific H3K36me2, since two H3K36 methyltransferases are present in Neurospora, . WT H3K27me2/3 minimally interacts with constitutive heterochromatin, and the density of WT H3K36me2 prevents correlational analysis . Quantification of the 2.5 kb bins in both the distance-normalized, single fastq and double digest datasets showed that $15% of all H3K9me3-enriched bins form intra-and interchromosomal contacts with euchromatin enriched for H3K27ac, H3K4me3, and ASH1-specific H3K36me2, although H3K9me3 to H3K9me3 contacts were the most prevalent .
To assess whether these topological contacts may influence transcription, we reanalyzed published poly-adenine RNA-sequencing datasets from WT and Ddim-5 strains ). Here, if euchromatic genes require contact with H3K9me3-marked silent chromatin for proper expression, loss of H3K9me3 would alter gene expression patterns. We observe multiple genes that form contacts . Continued corrected heatmaps displaying the contacts extending 25 kb (or 60 kb for IVR) of 4 individual chromosome ends at 500 bp bin resolution using the single in situ Hi-C matrix merged from MseI and DpnII fastq files (giving 106M valid reads; above diagonal) or the DpnII/MseI double digest contact matrix (below diagonal). Lines and ovals (above) denote strong subtelomeric contacts as possible structures, which include the absolute telomeric sequences. H3K9me3 and H3K27me2/3 ChIP-seq enrichment tracks presented below at same y-axis scales as in . Arrowheads show regions on heterochromatic LG IIR subtelomeric regions that interact with euchromatin in the single DpnII and MseI merged matrix that gives 106M valid reads. d). corrected heatmaps displaying contacts of 3 interspersed heterochromatic regions at 500 bp bin resolution using the single in situ Hi-C matrix merged from MseI and DpnII fastq files (with 106M valid reads; above diagonal) or the DpnII/MseI double digest contact matrix (below diagonal). Position of each region indicated above, and H3K9me3 ChIP-seq enrichment track presented below. Arrowheads highlight dense globules either internal or immediately proximal to heterochromatic regions that may limit heterochromatin spread. e). KR-corrected in situ Hi-C heatmaps of contacts of a $500 kb euchromatic region of LG IV at 2.5 kb bin resolution using the single in situ Hi-C matrix merged from MseI and DpnII fastq files (above) or the DpnII/MseI double digest contact matrix (below). Triangles outline possible local euchromatin clusters observed in the KR-corrected matrix data. Gene and H3K9me3 ChIP-seq enrichment tracks shown below heatmap.
with permanently silent chromatin and are enriched with histone post-translational modifications that have drastically altered (4-times) gene expression . Specifically, many genes enriched with H3K27ac or H3K4me3 that associate with silent chromatin are downregulated while several other heterochromatin-contacting genes with ASH1-catalyzed H3K36me2 are upregulated . We conclude that small euchromatic regions enriched for active or repressive histone modifications (a) (b) (c) . Long-range interactions between constitutive heterochromatin and individual euchromatic regions enriched with specific histone marks. a). KRcorrected Hi-C heatmaps of genomic interactions within the $700 kb region of the terminal right arm of LG II at progressively higher bin resolutions (5, 2.5, and 1 kb) aligned with the indicated ChIP-seq enrichment tracks and plotted on a log 10 scale using the single in situ Hi-C matrix merged from MseI and DpnII fastq files (giving 106M valid reads). Triangle lines within the Hi-C heatmaps are the program-calculated Topologically Associated Domains (TADs), while the vertical lines originating from the 1 kb heatmap highlight the borders of the H3K9me3-marked constitutive heterochromatin domains.
Arrowheads mark examples of euchromatic genes interacting with H3K9me3-marked heterochromatin. b). Identical 2.5 kb bin resolution KR-corrected Hi-C heatmap and 3 ChIP-seq enrichment tracks presented in A, and IGV images of bed files displaying genes that are 4-times increased or decreased in their expression in Ddim-5 strains. Dashed lines highlighting individual constitutive heterochromatin-euchromatic region interactions, colored as in panel (a): the 45 interactions originating from H3K9me3 regions are colored identically to the H3K9me3 ChIP-seq enrichment track, while the corresponding vertical lines ending in ChIP-seq enrichment tracks colored per tracks; multiple lines for each interaction between heterochromatin and euchromatin and different peaks of enrichment are shown. c). Graph of in situ Hi-C contact quantification across an observed vs expected 2.5 kb matrix (to normalize for genomic distance), scaled as the % of total contacts, with contacts originating at WT H3K9me3 enriched peaks and ending at one of the 4 possible enriched marks, colored as in A, or a region that is not enriched (see Materials and Methods). Only interchromosomal contacts log 2 ! 3.5 and intrachromosomal contacts log 2 ! 2.5 were counted. Numerical values of the interactions between WT H3K9me3 and either ASH1-catalyzed H3K26me2, WT H3K4m3, or WT H3K27ac are provided to the right of each bar; 664,816 total contacts originate in H3K9me3 in the single fastq matrix while 946,655 total contacts originate in H3K9me3 in the double digest. Bar graphs of the terminal enriched region are colored as the ChIP-seq tracks in (a). associate with constitutive heterochromatin for proper gene expression, possibly revealing a novel fungal mechanism for regulation of gene expression through chromatin topology.
# Discussion
Here, we characterize the genome topology of the filamentous fungus N. crassa at a high-resolution with 2 comprehensive Hi-C datasets: one that merges fastq files of Hi-C datasets independently generated by restriction enzymes monitoring euchromatin (DpnII) and heterochromatin (MseI), and a double digest dataset using both DpnII and MseI, each essentially generated with most current Hi-C protocol (e.g. Hi-C 3.0) [bib_ref] Hi-C 3.0: improved protocol for genome-wide chromosome conformation capture, Lafontaine [/bib_ref]. These datasets show the Neurospora chromosome conformation in exquisite detail, unveiling principles of fungal genome topology [fig_ref] Figure 7: Model of genome topology of N [/fig_ref]. Both datasets equally represent the conformation of euchromatin, suggesting gene contacts are accurate, which will support future gene regulation studies. However, heterochromatin-specific interactions appear more pronounced in the merged, single fastq Hi-C contact matrix-possibly, the DpnII/MseI double digest subtly over-represents rarer euchromatin-heterochromatin contacts and slightly depletes biologically relevant long-range contacts between heterochromatic regions; the latter is readily apparent in KR-corrected DpnII or raw MseI contact matrices. We slightly favor the merged, single fastq dataset, as it has more valid Hi-C interactions and therefore a greater depth of genomic contacts; we believe this dataset accurately displays genome topology within individual compartments (e.g., within heterochromatic or euchromatic regions), given the similarities of this dataset to the double digest contact matrix, as well as our computational experiment that shows excellent correlation between merged combinations of sampled DpnII and MseI replicates. However, our merged, single fastq dataset most likely underestimates contacts between euchromatic and heterochromatic compartments: upon KR correction, the underlying biases produced by the restriction enzyme used to generate that dataset are multiplicative [bib_ref] Probabilistic modeling of Hi-C contact maps eliminates systematic biases to characterize global..., Yaffe [/bib_ref] [bib_ref] Iterative correction of Hi-C data reveals hallmarks of chromosome organization, Imakaev [/bib_ref] , but when these datasets are combined (e.g., fastq files are merged), the inherent biases of each dataset, within the merged dataset, become additive upon KR correction. This means crosscompartment contacts appear weaker in the single fastq, merged Hi-C matrices. While the overall fungal topology can be presented by merging fastq files of independently generated DpnII and MseI datasets, use of either DpnII or MseI independently in a single Hi-C experiment can monitor the contacts within a specific chromatin compartment in the Neurospora nucleus, an excellent advantage for the use of Neurospora in genome topology studies. Specifically, MseI highlights contacts between heterochromatic regions >1 Mb apart while DpnII captures more intra-compartmental euchromatic contacts, despite these restriction enzymes having identical digestion frequencies (e.g. DpnII and MseI both have recognition sites every 4 4 , or 256, bases). This is in contrast to higher organisms, such as humans, where completely different chromosome capture protocols, using different digestion enzymes (e.g. micrococcal nuclease), are needed to monitor short range (loops) or long range (compartments) structural features [bib_ref] Systematic evaluation of chromosome conformation capture assays, Oksuz [/bib_ref]. Thus, our independent DpnII and MseI datasets are superb resources for future work to assess the contribution of individual genes for roles in forming the normal conformation of active or silent chromatin. In fact, our MseI dataset allowed for the correction of the Neurospora reference genome, although numerous gaps are still present in our reported genome version (nc14), which would require long-read (e.g. PacBio) sequencing to fill. Considering that all our Hi-C libraries were generated with the latest in situ protocol to ligate contacting DNA in the nucleus [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref] [bib_ref] Architectural alterations of the fission yeast genome during the cell cycle, Tanizawa [/bib_ref] , we believe that our data are of high quality and effectively capture valid DNA contacts across the Neurospora genome. Previous work reports high-quality metazoan datasets show chromosomal territories with few interchromosomal contacts . However, chromosome territories in fungi are controversial [bib_ref] Lack of chromosome territoriality in yeast: promiscuous rejoining of broken chromosome ends, Haber [/bib_ref] [bib_ref] Spatial organisation and behaviour of the parental chromosome sets in the nuclei..., Lorenz [/bib_ref] [bib_ref] Population genomics of domestic and wild yeasts, Liti [/bib_ref] [bib_ref] Principles of chromosomal organization: lessons from yeast, Zimmer [/bib_ref] , as fungal chromosomes minimally isolate, and substantial interchromosomal contacts are observed despite application of the in situ Hi-C protocol [bib_ref] Architectural alterations of the fission yeast genome during the cell cycle, Tanizawa [/bib_ref]. Indeed, a recent publication on the genome architecture across the Tree of Life highlights distinctions in topology between fungi and metazoans, with fungi exhibiting independent bundling of centromeres and telomeres [bib_ref] 3D genomics across the tree of life reveals condensin II as a..., Hoencamp [/bib_ref]. Altogether, we suggest 1 measure of fungal Hi-C dataset quality is reduced mitochondria contacts, due to the isolation of nuclear and mitochondrial genomes in different organelles [bib_ref] Systematic evaluation of chromosome conformation capture assays, Oksuz [/bib_ref]. By these standards, our presented fungal Hi-C data are of exceptional quality, given the few [bib_ref] Neurospora chromosomes are organized by blocks of importin alpha-dependent heterochromatin that are..., Galazka [/bib_ref]. Boxes: novel genomic interactions identified in this work. Clockwise from the top left: euchromatic contacts on the bundled heterochromatic centromere; specific heterochromatic interactions of intra-and intertelomeric contacts; euchromatic genes enriched for acetylation or ASH1L-catalyzed H3K36me2 contact H3K9me3-marked constitutive heterochromatic regions for the regulation of gene expression; euchromatin is initially packaged into globules of $20-40 kb in size that are stacked into "Regional Globule Clusters": euchromatic loops organized by association of heterochromatic regions. mitochondrial ligation products observed, and are consistent with the formation of weak territories with substantial interchromosomal, euchromatic contacts for the 7 Neurospora chromosomes. An exciting, unexplored possibility is that interchromosomal contacts regulate gene expression.
Our in situ Hi-C data allows us to make several general conclusions about fungal genome organization: heterochromatin is segregated from euchromatin, yet associations between heterochromatic regions mediate euchromatin looping; heterochromatic regions have stochastic internal contacts; heterochromatic histone marks correlate with the topology of some genomic features (e.g., H3K9me3-marked silent regions are most often stochastically organized); and euchromatin enriched for certain histone marks associates with constitutive heterochromatin to impact gene expression. Our data are consistent with the compartmentalization of eukaryotic genomes into euchromatic (A) and heterochromatic (B) compartments [bib_ref] Comprehensive mapping of long-range interactions reveals folding principles of the human genome, Lieberman-Aiden [/bib_ref] [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref] , which supports microscopic observations of the segregation of heterochromatin at the nuclear periphery while euchromatin has a more central localization in the nucleus [bib_ref] Structure and biochemistry of the nuclear envelope, Franke [/bib_ref] [bib_ref] The new cytogenetics: blurring the boundaries with molecular biology, Speicher [/bib_ref] [bib_ref] The nuclear envelope as a chromatin organizer, Zuleger [/bib_ref] [bib_ref] Nucleolus and nuclear periphery: velcro for heterochromatin, Padeken [/bib_ref] [bib_ref] Heterochromatin drives compartmentalization of inverted and conventional nuclei, Falk [/bib_ref] ; this subnuclear architecture supports our hypothesis that KR-correction of MseI-derived matrices excludes valid bundles of smaller, interspersed heterochromatic regions. Further, our in situ Hi-C data confirm microscopic observations in Neurospora showing the 7 centromeres, labeled with infrared fluorescent protein-tagged CenH3, cluster into a single focus localized to the nuclear membrane, yet these centromeres are segregated from telomeres, which present as 2-4 telomere clusters at the nuclear periphery when labeled with green fluorescent protein-tagged TRF1 [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref]. This Rabl-like chromosome organization observed in Neurospora is seen in other fungi, including Epichloë festucae, species of the Verticillium clade, Agaricus bisporus (Common mushroom), Saccharomyces cerevisiae, and Schizosaccharomyces pombe [bib_ref] A threedimensional model of the yeast genome, Duan [/bib_ref] [bib_ref] Cohesin-dependent globules and heterochromatin shape 3D genome architecture in S. pombe, Mizuguchi [/bib_ref] [bib_ref] Architectural alterations of the fission yeast genome during the cell cycle, Tanizawa [/bib_ref] [bib_ref] Repeat elements organise 3D genome structure and mediate transcription in the filamentous..., Winter [/bib_ref] [bib_ref] Repetitive elements contribute to the diversity and evolution of centromeres in the..., Seidl [/bib_ref] [bib_ref] 3D genomics across the tree of life reveals condensin II as a..., Hoencamp [/bib_ref] , highlighting the importance of heterochromatic centromeres and telomeres for fungal genome topology. In contrast, the active, gene-rich euchromatic DNA in Neurospora is packaged into globules that are $20-40 kb in size-smaller than human globules [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref] [bib_ref] Cohesin loss eliminates all loop domains, Rao [/bib_ref] but consistent with globules in yeast [bib_ref] Cohesin-dependent globules and heterochromatin shape 3D genome architecture in S. pombe, Mizuguchi [/bib_ref]. Cohesin forms these globules in other species, as enrichment of the cohesin subunit Rad-21 occurs at globule boundaries, and loss of Rad-21 depletes loops [bib_ref] Cohesin-dependent globules and heterochromatin shape 3D genome architecture in S. pombe, Mizuguchi [/bib_ref] [bib_ref] Cohesin loss eliminates all loop domains, Rao [/bib_ref]. Given the conservation of cohesin from fission yeast to humans, cohesin likely forms globules in Neurospora, and mutations in the Neurospora rad-21 gene may elucidate local globule folding.
To structurally organize active DNA in the Neurospora genome, euchromatic globules are hierarchically packaged into layered "Regional Globular Clusters"; we suggest this term to distinguish the euchromatic structural elements in fungi from TADs in higher eukaryotes, and to clarify a previous report that Neurospora has TADs . In a Regional Globule Cluster, euchromatin encompassing several hundred thousand basepairs of DNA and bordered by constitutive heterochromatin systematically loops upon the bundling of this flanking silent chromatin. Regional Globule Clusters are also observed in lower resolution Hi-C datasets [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref]. One might consider Regional Globular Clusters as primitive fungal TADs, but they certainly are not analogous to higher eukaryotic TADs. Considering the formal definition of Megabase-sized TADs [bib_ref] Topological domains in mammalian genomes identified by analysis of chromatin interactions, Dixon [/bib_ref] [bib_ref] Spatial partitioning of the regulatory landscape of the X-inactivation centre, Nora [/bib_ref] [bib_ref] Three-dimensional folding and functional organization principles of the Drosophila genome, Sexton [/bib_ref] [bib_ref] Regulation of single-cell genome organization into TADs and chromatin nanodomains, Szabo [/bib_ref] , true TADs have few long-range contacts beyond their boundaries. Our data shows that the smaller Regional Globular Clusters are layered and have extensive interlayer contacts, suggesting that higher eukaryotic proteins isolating external contacts from TADs, such as condensin II [bib_ref] 3D genomics across the tree of life reveals condensin II as a..., Hoencamp [/bib_ref] , are not present in Neurospora.
Eukaryotic genomes are also organized by chromatin loops that average $200 kb in size [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref]. Higher eukaryotes encode the protein CTCF for loop formation [bib_ref] CTCF tethers an insulator to subnuclear sites, suggesting shared insulator mechanisms across..., Yusufzai [/bib_ref] : each CTCF monomer binds a convergently oriented 17 basepair asymmetric DNA sequence, and CTCF dimerization forms the loop base [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref] ; binding site choice for CTCF dimerization allows dynamic loop formation [bib_ref] CRISPR inversion of CTCF sites alters genome topology and enhancer/promoter function, Guo [/bib_ref]. Loops in humans are evident in Hi-C heatmaps as an enriched long-range contact at the intercept of 2 CTCF binding sites; enrichment is lost upon targeted CTCF degradation or binding sequence alteration [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref] [bib_ref] CRISPR inversion of CTCF sites alters genome topology and enhancer/promoter function, Guo [/bib_ref] [bib_ref] Targeted degradation of CTCF decouples local insulation of chromosome domains from genomic..., Nora [/bib_ref] , and CTCF is necessary to insulate promoters from aberrant enhancer/silencer contacts [bib_ref] Insulator dysfunction and oncogene activation in IDH mutant gliomas, Flavahan [/bib_ref]. The Neurospora genome does not contain the asymmetric 17 basepair CTCF binding sites, nor does it encode a CTCF homolog, which explains the lack of loop anchor enrichment foci here and in previous Neurospora Hi-C datasets [bib_ref] Normal chromosome conformation depends on subtelomeric facultative heterochromatin in Neurospora crassa, Klocko [/bib_ref] However, it is conceivable that constitutive heterochromatin takes the place of CTCF in fungi. In Neurospora, the heterochromatic regions interspersed throughout the genome would act as an "anchor" at the base of a euchromatic Regional Globular Cluster in a manner analogous to CTCF [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref] , although the dynamics of Regional Globular Clusters would be restricted to the subtle differences in the bundling between heterochromatic regions. Perhaps the seemingly unnecessary interspersed heterochromatic regions in the Neurospora genome are retained during evolution to structurally organize fungal chromosomes through formation of Regional Globular Clusters. One open question is if Regional Globular Clusters (or TADs) in any species have "epigenetic memory" for propagation across generations. In mammals, TADs are lost during mitosis and reform early in G1, implying TADs do not have any inherent epigenetic memory; it has been proposed that other factors, including histone post-translational modifications, reassemble TADs [bib_ref] Organization of the mitotic chromosome, Naumova [/bib_ref]. Perhaps in Neurospora, constitutive heterochromatic regions, which display epigenetic inheritance and spatially interact in the nucleus [bib_ref] Relics of repeat-induced point mutation direct heterochromatin formation in Neurospora crassa, Lewis [/bib_ref] [bib_ref] Neurospora chromosomes are organized by blocks of importin alpha-dependent heterochromatin that are..., Galazka [/bib_ref] , are a simple mechanism to rapidly reform Regional Globular Clusters-and genome topology-during cell cycle progression. It is possible that liquidphase separation of heterochromatin, as reported for human HP1a [bib_ref] Liquid droplet formation by HP1a suggests a role for phase separation in..., Larson [/bib_ref] [bib_ref] The role of phase separation in heterochromatin formation, function, and regulation, Larson [/bib_ref] and proposed for genome compartmentalization [bib_ref] Heterochromatin drives compartmentalization of inverted and conventional nuclei, Falk [/bib_ref] facilitates constitutive heterochromatin interactions.
However, in Neurospora, loss of H3K9me3 (in a Ddim-5 strain lacking the single H3K9 methyltransferase) or HP1 minimally impacts heterochromatin contacts during fungal growth [bib_ref] A histone H3 methyltransferase controls DNA methylation in Neurospora crassa, Tamaru [/bib_ref] [bib_ref] Trimethylated lysine 9 of histone H3 is a mark for DNA methylation..., Tamaru [/bib_ref] [bib_ref] Relics of repeat-induced point mutation direct heterochromatin formation in Neurospora crassa, Lewis [/bib_ref] [bib_ref] Neurospora chromosomes are organized by blocks of importin alpha-dependent heterochromatin that are..., Galazka [/bib_ref] , even though HP1 is important for genome topology in higher eukaryotes, as this protein is necessary to reorganize silent chromatin in the Drosophila genome during early embryogenesis [bib_ref] HP1 drives de novo 3D genome reorganization in early Drosophila embryos, Zenk [/bib_ref]. Thus, in fungi, another inherent property of heterochromatin must drive liquid-phase separation to establish genome topology.
Liquid-phase separation would explain our high-resolution observations of internal heterochromatin organization, as the stochastic, pattern-less nature of Hi-C contacts internal to centromeres, telomeres, and interspersed heterochromatic regions is consistent with both random and variable internal packaging. At chromosome ends, heterochromatin facilitates intra-and interchromosomal subtelomeric contacts: individual subtelomeres are compacted at H3K9me3-marked constitutive heterochromatin, but H3K27me2/3-marked facultative heterochromatin facilitates contacts between subtelomeres. Some subtelomeres are contained within larger H3K9me3 domains, suggesting subtelomeres are uniquely recognized, perhaps by dual repressive histone marks. In Neurospora, loss of H3K27me2/ 3 (in a Dset-7 strain lacking the single H3K27 methyltransferase) compromises genome topology, as subtelomere clusters have reduced association with the nuclear membrane ), but H3K9me3 loss does not reduce subtelomeric interactions . Thus, organization of chromosome ends must occur through distinct mechanisms mediated by both constitutive and facultative heterochromatin. One possibility is that constitutive heterochromatin condenses individual subtelomeres through liquid phase separation [bib_ref] Liquid droplet formation by HP1a suggests a role for phase separation in..., Larson [/bib_ref] while other proteins-such as EPR-1 or PAS [bib_ref] Evolutionarily ancient BAH-PHD protein mediates Polycomb silencing, Wiles [/bib_ref] and/or Shelterin [bib_ref] Shelterin components mediate genome reorganization in response to replication stress, Mizuguchi [/bib_ref] , which recognize/modulate facultative heterochromatin and bind telomere repeats, respectively-manage interchromosomal subtelomere contacts. Future work examining the role of these proteins/complexes in genome topology could clarify this hypothesis.
Small euchromatic segments enriched with active or repressive histone post-translational modifications associate with H3K9me3-marked heterochromatin in our comprehensive Hi-C datasets. Specifically, the small euchromatic regions that topologically associate with constitutive heterochromatin are enriched with activating marks such as acetylation of H3K27 and the trimethylation of H3K4 [bib_ref] Histone H3 lysine 4 methyltransferase is required for facultative heterochromatin at specific..., Zhu [/bib_ref] , as well as a repressive mark, H3K36me2, catalyzed by ASH1 ; these cross-compartment interactions may be more prevalent in the merged, single fastq dataset images due to the disparate frequencies of DpnII and MseI sites (in euchromatin and heterochromatin, respectively), focusing the interactions between genes and heterochromatic regions when dataset biases become additive upon fastq merging (above). While many of these contacts occur between heterochromatin and euchromatin separated by a few thousand basepairs, these intercompartment interactions may represent another layer of gene regulation: for activation, peaks of H3K27ac or H3K4me3 may indicate active genes or promoters require neighboring heterochromatin for maximal transcription, as has been observed for the Neurospora methionine synthase gene met-8 [bib_ref] The highly expressed methionine synthase gene of Neurospora crassa is positively regulated..., Yang [/bib_ref]. Conversely, silent genes may associate with constitutive heterochromatin at the nuclear periphery to prevent aberrant transcription initiation in the active nucleus center. Many of the genes marked by ASH1L-specific H3K36me2 are minimally transcribed and are also tri-methylated at H3K27 , suggesting Neurospora utilizes several mechanisms for repression. However, we observe minimal H3K27me2/ 3 enrichment across H3K36me2-marked genes that associate with silent chromatin-perhaps the association of ASH1-targeted genes with H3K9me-marked heterochromatin allows rapid transcriptional activation, analogous to poised developmental genes marked by H3K4me3 and H3K27me3 in higher eukaryotes [bib_ref] A double take on bivalent promoters, Voigt [/bib_ref]. Currently, the underlying mechanism for euchromatin looping to silent chromatin is unknown; an active recruitment mechanism by unknown protein(s) and a passive association mechanism, similar to liquid droplet formation [bib_ref] The role of phase separation in heterochromatin formation, function, and regulation, Larson [/bib_ref] , are both feasible. We cannot rule out that these euchromatic regions are also marked with a low level of H3K9me3, which would be bound by HP1 to facilitate oligomerization with HP1-enriched constitutive heterochromatin, to colocalize these regions for regulation. Consistent with this hypothesis, the promoter of the frq gene encoding a master regulator of the Neurospora circadian rhythm has low levels of H3K9me3 that cycle over circadian time. Still, proper gene transcription requires uncompromised heterochromatin: RNA-sequencing of a Ddim-5 (aka: Dkmt1) strain, which lacks tri-methylation of H3K9, has numerous previously unexplained gene expression changes, including aberrant transcriptional activation and repression relative to a WT strain . We show several genes that contact constitutive heterochromatin in a WT strain have altered expression upon H3K9me3 loss: some acetylated genes become repressed while H3K36me2-marked genes are activated in a Ddim-5 (Dkmt1) strain, although the correlation is inexact; of course, a specific combination of histone marks would be required for proper gene regulation in WT cells. It is unclear if altered gene expression occurs because topological changes no longer form, or if some component of H3K9me3-marked chromatin is necessary for gene regulation despite these long-range contacts still forming. Highresolution Hi-C of Ddim-5 (Dkmt1) or Dhpo strains may distinguish between these hypotheses. In retrospect, the re-localization of H3K27me2/3 that occurs in a Ddim-5 (Dkmt1) strain [bib_ref] Genome-wide redistribution of H3K27me3 is linked to genotoxic stress and defective growth, Basenko [/bib_ref] [bib_ref] Loss of HP1 causes depletion of H3K27me3 from facultative heterochromatin and gain..., Jamieson [/bib_ref] presumably would not reestablish constitutive heterochromatin to restore proper gene expression, but our data partially explains the fact that few transcriptional changes are observed in a Dset-7 (Dkmt6) strain lacking H3K27me2/3 , as constitutive heterochromatin-and any long-range association between genes and silent chromatin-still forms.
In summary, our high-resolution Neurospora Hi-C datasets identify a novel mechanism involving genome topology for controlling gene expression; further experiments should clarify how euchromatin-heterochromatin interactions impact transcriptional regulation. We maintain that high-resolution Hi-C datasets are valuable tools for fungal researchers to elucidate the role of chromatin topology on genome function in fungal systems modeling human genome architecture or within fungal pathogens.
## Data availability
All WT N. crassa strains are available upon request. All in situ Hi-C high-throughput sequencing data of the WT Neurospora genome have been deposited to the National Center of Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) public repository under the accession number GSE173593. Supplementary figures and tables are available at figshare: https:// doi.org/10.25387/g3.17138534 and discussions; and G3 editor Job Dekker and the 2 anonymous reviewers for comments that improved this manuscript.
# Funding
Funding for this project was provided by start-up funds from the University of Colorado Colorado Springs (UCCS) College of Letters, Arts, and Sciences (ADK) and a UCCS Committee on Research and Creative Works internal seed grant (ADK). Students and ADK were partly supported by an Academic Research Enhancement Award (AREA) grant from the National Institutes of Health (1R15GM140396-01; to ADK) during the final stages of manuscript preparation.
## Conflicts of interest
None declared.
## Literature cited
[fig] Figure 1: High-resolution chromosome conformation of Neurospora crassa using the restriction enzyme DpnII. a) Integrative genomics viewer (IGV)(Robinson et al. 2011) image of a portion of Linkage [/fig]
[fig] Figure 2: High-resolution Hi-C of Neurospora crassa using the restriction enzyme MseI. a) Heatmap displaying the merged MseI in situ Hi-C dataset of LG II, either plotting raw read counts (above diagonal) or KR-corrected counts (below diagonal), at 20 kb bins. The magnified heatmap shows the intrachromosomal interaction between the left and right telomeres of LG II at 5 kb resolution. CenH3 (centromeric) and H3K9me3 (heterochromatic) ChIP-seq enrichment tracks presented above and to the left of each plot. Horizontal line at bottom of heatmap highlights region displayed in panel (c). b). Heatmap displaying the interactions between LG II and LG III in the merged MseI in situ Hi-C dataset at 20 kb. Magnified plots at 5 kb resolution show the interactions between the LG II/LG III centromeres (dashed lines) or the interactions between the left subtelomeres of LG II and LG III (box). c). Heatmaps of MseI (top) and DpnII (bottom) contacts at 5 kb bin resolution of the right arm of LG II (highlighted in A by horizontal line). Both triangle heatmaps were cropped and rotated 45 from square (reflective) heatmaps. Central IGV image displays tracks of MseI and DpnII restriction sites, H3K9me3 ChIP-seq enrichment and genes. [/fig]
[fig] Figure 7: Model of genome topology of N. crassa. Center: general model of interactions of silent genome features within the nucleus of a wild type strain of N. crassa; adapted from [/fig]
|
Histoplasma capsulatum: An Unusual Case of Pericardial Effusion and Coarctation of the Aorta
Histoplasma capsulatum is a fungus that is endemic in many parts of the world and can present with a wide variety of symptoms. Here we present a case of a previously healthy 19-year-old female who presented with shortness of breath. She was found to have a right lung mass and coarctation of the aorta on computed tomography imaging. Pathology revealed granuloma caused by Histoplasma capsulatum. She later developed massive pericardial effusion, requiring emergent pericardiocentesis. She was treated with anti-fungal therapy and recovered well. This case illustrates an unusual presentation of newly diagnosed coarctation of the aorta complicated by Histoplasma pericardial effusion. Imaging and pathology slides are reviewed.
# Introduction
Histoplasma capsulatum is a dimorphic fungus that is endemic in many parts of the world including North, Central and South America. In the United States, it is most prevalent in the Ohio and Mississippi River Valleys. It grows well with humid conditions in acidic soil, with growth accelerated in sites contaminated with bird and bat excrement. Disturbance of the soil results in aerosolized spores that can travel over miles. Histoplasmosis can present with a wide range of symptoms from asymptomatic to life-threatening. Severity of illness after inhalation depends on the intensity of exposure and the immune status of the host. The majority of cases will resolve without therapy. Pulmonary manifestations are the most common symptomatic presentation [bib_ref] Clinical practice guidelines for the management of patients with histoplasmosis: 2007 update..., Wheat [/bib_ref].
## Case report
A 19-year-old previously healthy Hispanic female presented with chief complaint of progressively worsening shortness of breath and heart palpitations for 1 week. She had previously recently been evaluated at an outside hospital without imaging and was diagnosed with gastritis, then discharged home. Prior medical history was positive for childhood asthma and removal of a lipoma from her right flank. There was no family history of heart disease, but she did note a paternal aunt with a history of unspecified mediastinal cancer. The patient emigrated to the United States from Mexico 2 years ago and had visited Mexico 2 months prior to presentation. She had never been pregnant and was not sexually active. She denied smoking, alcohol or drug use. On presentation, her vital signs were blood pressure 128/84, pulse 111 beats per minute, respiration rate 12 Chest X-ray showed a right peri-hilar mass with some tracheal compression [fig_ref] Figure 1: Chest X-ray showing right peri-hilar mass with tracheal compression [/fig_ref]. Computed tomography (CT) scan revealed a 1.2 × 1.1 cm mass in the right hilar region with associated diffuse mediastinal lymphadenopathy [fig_ref] Figure 2: Computed tomography [/fig_ref]. A previously undiagnosed severe isthmic coarctation of the aorta was noted immediately distal to the origin of the left subclavian artery with surrounding collateral arteries. Coarctation was confirmed with echocardiogram that showed peak gradient around 40 mm Hg. Systolic function was normal. Initial imaging was suspicious for malignancy or infectious process. There was concern that the coarctation may be related to the mediastinal process. Initial bronchoscopy with bronchoalveolar lavage (BAL) and fine-needle aspiration (FNA) were non-diagnostic. Subsequent endobronchial ultrasound (EBUS) with trans-bronchial needle biopsy of the mediastinal lymph nodes showed coalescing fibrocaseous granulomas with necrosis [fig_ref] Figure 3: Wedge resection of right middle lobe of lung showing coalescing fibrocaseous granulomas [/fig_ref]. Gram stain, culture and acid-fast bacilli (AFB) stain were negative. QuantiFERON gold test was also negative. Serology was positive for Histoplasma (titer 1:256). Pathologic analysis showed necrotizing granulomas with small fungal yeast that was morphologically most consistent with Histoplasma [fig_ref] Figure 5: Wedge resection of right middle lobe of lung, magnification of fibrocaseous necrosis... [/fig_ref].
Repeat CT chest revealed pericardial thickening with massive pericardial effusion [fig_ref] Figure 6: Computed tomography [/fig_ref]. There was 2 cm pericardial thickening that was confirmed with echocardiogram. Emergent pericardiocentesis was required and 260 cc of yellow blood-tinged fluid was removed, then a pericardial drain was placed. The patient was treated initially with amphotericin B. Due to allergic reaction, she was transitioned to voriconazole. Surgical repair of the coarctation of the aorta was delayed until resolution of the Histoplasma infection. She eventually underwent surgical coarctectomy with end-to-end anastomosis with good results.
# Discussion
Pulmonary manifestations are the most common symptomatic presentation of histoplasmosis [bib_ref] Clinical practice guidelines for the management of patients with histoplasmosis: 2007 update..., Wheat [/bib_ref]. Several other authors have presented cases of histoplasmosis pericarditis [bib_ref] A large urban outbreak of histoplasmosis: clinical features, Wheat [/bib_ref] [bib_ref] Chest pain resulting from histoplasmosis pericarditis: a brief report and review of..., Wang [/bib_ref] [bib_ref] Recurrent massive pleural effusion due to pleural, pericardial, and epicardial fibrosis in..., Kilburn [/bib_ref] [bib_ref] Unusual manifestations of histoplasmosis, Thompson [/bib_ref] [bib_ref] Pericarditis caused by Histoplasma capsulatum, Picardi [/bib_ref] [bib_ref] Pericarditis due to histoplasmosis, Young [/bib_ref] [bib_ref] Acute histoplasma pericarditis, Saslaw [/bib_ref]. Pericarditis is a known complication of pulmonary histoplasmosis, and has been reported in up to 5-10% of symptomatic patients [bib_ref] A large urban outbreak of histoplasmosis: clinical features, Wheat [/bib_ref] [bib_ref] Chest pain resulting from histoplasmosis pericarditis: a brief report and review of..., Wang [/bib_ref]. Patients may present with a range of symptoms including fever, headache, myalgia, cough, chest pain, respiratory failure or death. The vast majority of symptomatic cases show pulmonary infiltrates with mediastinal lymphadenopathy. Pericarditis from histoplasmosis is generally attributed to an inflammatory response rather than an infectious process within the pericardium [bib_ref] Chest pain resulting from histoplasmosis pericarditis: a brief report and review of..., Wang [/bib_ref]. Treatment is indicated for moderate to severe acute or chronic pulmonary, disseminated and central nervous system histoplasmosis. Treatment is generally initiated with amphotericin B intravenously, followed by oral itraconazole. In patients with mild manifestations, treatment is usually unnecessary. Histoplasma pericarditis is recommended to be treated with non-steroidal anti-inflammatory (NSAID) therapy in mild cases [bib_ref] Clinical practice guidelines for the management of patients with histoplasmosis: 2007 update..., Wheat [/bib_ref]. It is unclear if antifungal therapy alters the course of histoplasmosis pericarditis due to lack of clinical trials. Review of literature regarding histoplasmosis pericarditis did not reveal any previously published reports in conjunction with aortic coarctation. The aortic coarctation in this case is believed to be congenital rather than related to the histoplasma infection. However, the coarctation likely contributed to the severity of presentation. Here we provide a review of imaging showing an active pulmonary Histoplasma infection complicated by pericardial effusion in conjunction with aortic coarctation for the first time.
[fig] Figure 1: Chest X-ray showing right peri-hilar mass with tracheal compression. [/fig]
[fig] Figure 2: Computed tomography (CT) of chest showing mediastinal mass and mediastinal lymphadenopathy. [/fig]
[fig] Figure 3: Wedge resection of right middle lobe of lung showing coalescing fibrocaseous granulomas (hematoxylin-eosin stain, magnification, × 100). [/fig]
[fig] Figure 4: Wedge resection of right middle lobe of lung depicting giant cell histiocytes, epithelioid histiocytes, lymphocytes and fibrocaseous necrosis (hematoxylin-eosin stain, magnification, × 100). [/fig]
[fig] Figure 5: Wedge resection of right middle lobe of lung, magnification of fibrocaseous necrosis center of granuloma with budding yeast consistent with histoplasmosis species (Grocott methenamine silver stain, magnification, × 400). [/fig]
[fig] Figure 6: Computed tomography (CT) of chest showing pericardial effusion with pericardial thickening. [/fig]
|
Context and Time in Causal Learning: Contingency and Mood Dependent Effects
Defining cues for instrumental causality are the temporal, spatial and contingency relationships between actions and their effects. In this study, we carried out a series of causal learning experiments that systematically manipulated time and context in positive and negative contingency conditions. In addition, we tested participants categorized as non-dysphoric and mildly dysphoric because depressed mood has been shown to affect the processing of all these causal cues. Findings showed that causal judgements made by non-dysphoric participants were contextualized at baseline and were affected by the temporal spacing of actions and effects only with generative, but not preventative, contingency relationships. Participants categorized as dysphoric made less contextualized causal ratings at baseline but were more sensitive than others to temporal manipulations across the contingencies. These effects were consistent with depression affecting causal learning through the effects of slowed time experience on accrued exposure to the context in which causal events took place. Taken together, these findings are consistent with associative approaches to causal judgement.
# Introduction
The ability to learn about causal relationships between events is adaptive and enables people to learn to control their environment or, at least, to interact with it effectively [bib_ref] Associative Accounts of Causality Judgment, Shanks [/bib_ref]. It isn't surprising then that psychological disturbance can affect people's judgments about causal relationships [bib_ref] Depression and the causal inference process, Abramson [/bib_ref] [bib_ref] Response latencies for the causal explanations of depressed, paranoid, and Normal individuals:..., Bentall [/bib_ref]. Moreover, the underlying mechanisms through which such effects occur are of considerable interest, not least because they can inform about the psychological disturbance itself but also because they can inform us about the nature of causal learning processes. For example, the link between cause and effect is not directly observable and so causal learning involves a psychological process that extracts cues to causality, the temporal and contingency relationships between events [bib_ref] On the origin of personal causal theories, Young [/bib_ref]. For example, cause and effect occur successively, often in close spatial proximity, and the effect should be contiguous and contingent upon the occurrence of the cause [bib_ref] Temporal Contiguity and the Judgment of Causality by Human Subjects, Shanks [/bib_ref] [bib_ref] Contiguity and the outcome density bias in action-outcome contingency judgements, Vallee Tourangeau [/bib_ref]. However, it is the combination of these cues that is critical in terms of defining causality, as any of them taken in isolation could be misleading [bib_ref] Time as guide to cause, Lagnado [/bib_ref] [bib_ref] Judging probable cause, Einhorn [/bib_ref].
Time, in terms of succession and contiguity, is often considered to be the essential causal cue, while patterns of cause-effect cooccurrence provide additional corroborative information [bib_ref] Time as guide to cause, Lagnado [/bib_ref] [bib_ref] Structure and strength in causal induction, Griffiths [/bib_ref] [bib_ref] Theory-based causal induction, Griffiths [/bib_ref]. In other words, if information about the contingency between cause and effect is consistent with a causal relationship, but time information is not, then people are less likely to judge that a causal relationship exists [bib_ref] Temporal Contiguity and the Judgment of Causality by Human Subjects, Shanks [/bib_ref]. However, determining whether or not this is the case is complicated because there is considerable inter-connectedness between these variables. Time, for example, not only defines succession and contiguitybut also the density or rate of cause-effect experiences which is relevant to contingency [bib_ref] Causal inferences as perceptual judgments, Anderson [/bib_ref]. Spatial proximity is defined by the context in which events occur, and the context itself can define action-effect contingency [bib_ref] Covariation in natural causal induction, Cheng [/bib_ref]. Moreover, the passage of time can constitute a change of context [bib_ref] Context, time, and memory retrieval in the interference paradigms of Pavlovian learning, Bouton [/bib_ref] and context is not a discretely occurring countable event, but a continuous and sometimes temporally defined aspect of a causal learning task [bib_ref] Depressive realism and outcome density bias in contingency judgements: The effect of..., Msetfi [/bib_ref]. Thus, it can be argued that time and contingency, and space or context and contingency, are interdependent and in studies so far, context was not a variable that was not manipulated and measured explicitly.
Our strategy in the current paper is to consider context explicitly, alongside the other causal cues, and to manipulate time and context systematically across different contingency conditions. Furthermore, we will also test distinct groups of participants categorized by levels of depressed mood. This is useful because depression is not only associated with changes in causal sensitivity [bib_ref] Depressive realism and outcome density bias in contingency judgements: The effect of..., Msetfi [/bib_ref] [bib_ref] Depressive realism: Wiser or quieter?, Blanco [/bib_ref] [bib_ref] Judgement of contingency in depressed and nondepressed students: Sadder but wiser, Alloy [/bib_ref] [bib_ref] Contingency judgements by depressed college students: Sadder but not always wiser, Benassi [/bib_ref] but also with disturbed time perception [bib_ref] The Effect of Mild Depression on Time Discrimination, Msetfi [/bib_ref] [bib_ref] Time perception, depression and sadness, Gill [/bib_ref] [bib_ref] Time experience and time judgment in major depression, mania and healthy subjects...., Bschor [/bib_ref] and impaired context processing [bib_ref] Depressive realism and outcome density bias in contingency judgements: The effect of..., Msetfi [/bib_ref] [bib_ref] Impaired context maintenance in mild to moderately depressed students, Msetfi [/bib_ref]. Therefore, levels of depression will have specific effects on causal judgments, and moderate the effects of time and context manipulations.
Our starting point is a brief discussion on the status of the various causal cues in relation to current theoretical perspectives on causal judgment, before we then discuss how studying the effects of depressed mood on causal judgment might be a useful method of informing this debate.
## Contingency and time in causal learning
Causal judgments often relate to contingencies with which people are actively involved rather than being passive observers and a considerable body of research has focused on whether judgments of such action-outcome contingencies bear any relation to objective mathematical quantifiers of the same relationships. Given the importance of contingency in causality, we will frame our arguments around this literature.
The contingency between an action and outcome can be quantified by a normative metric, known as delta P or DP [bib_ref] A note on measurement of contingency between two binary variables in judgment..., Allan [/bib_ref]. DP is a value, similar to a correlation coefficient, which describes both the direction (generative, preventative) and strength (strong, weak) of a contingency relationship. It differs from the correlation coefficient in two ways in that it concerns the frequency of binary events (on/off) and is a measure of a one-way, rather than bidirectional, relationship. DP is calculated as the difference between the conditional probabilities of an outcome given the presence of an action, p(Outcome|Action), and the absence of an action, p(Outcome|No Action). Therefore, the calculated value can vary between +1, indicating a perfect positive contingency and a generative causal relationship between action and effect, through the continuum to -1, where the outcome is less likely to happen in the presence of the action than in its absence, a preventative negative contingency. In a situation where the conditional probabilities are equal, the effect is no more likely to occur in the presence of the outcome than in its absence, and the DP is zero. In other words, there is no contingency relationship between action and outcome.
This definition of contingency includes the assumption that four possible action-effect conjunctions are relevant to the DP calculations. [fig_ref] Figure 1 262: contingency tables showing the four possible combinations of action -effect information [/fig_ref] (upper panel) shows a standardized contingency matrix in which the frequencies of each conjunction are given in each cell and denoted by the letters A, B, C and D. In many experimental designs, each conjunction would take place during a discrete experimental trial and the experimenter could manipulate the cell frequencies and thus the contingency experienced by the participant. [fig_ref] Figure 1 262: contingency tables showing the four possible combinations of action -effect information [/fig_ref] shows examples of such manipulations, a negative contingency (left), a zero contingency (middle) and a positive contingency (right) condition. After being exposed to such conditions, participants would be asked to assess the contingency, perhaps by rating their degree of control over the effect on a numeric scale, which could then be compared to systematic manipulations of DP, similar to those shown in [fig_ref] Figure 1 262: contingency tables showing the four possible combinations of action -effect information [/fig_ref].
Studies have shown that people are very sensitive to changes in manipulated contingencies. Their ratings of the strength of causal relationships distinguish between conditions, like those shown in [fig_ref] Figure 1 262: contingency tables showing the four possible combinations of action -effect information [/fig_ref] , in which effects are and are not contingent upon their actions [bib_ref] Rating causal relations: Role of probability in judgments of response-outcome contingency, Wasserman [/bib_ref] [bib_ref] The judgment of contingency and the nature of the response alternatives, Allan [/bib_ref] [bib_ref] Judging the importance of constant and variable candidate causes: A test of..., Vallee-Tourangeau [/bib_ref]. In addition, ratings are highly correlated with much more subtle variations in DP. For example, Wasserman et al., [bib_ref] Rating causal relations: Role of probability in judgments of response-outcome contingency, Wasserman [/bib_ref] exposed their participants to 25 different conditions that involved subtle variations in contingency. Ratings were almost perfectly positively correlated with DP (r = .97), demonstrating remarkable isomorphism between causal ratings and variations in contingency. Notably, causal ratings were weakest in conditions in which contingency was zero (DP = 0), in spite of the fact that these conditions did involve contiguous pairings between actions and outcomes. Contingency, it seems, was a crucial cue to causality.
In spite of the importance of contingency for causal learning, temporal contiguity does have a profound effect on people's ability to detect causal relationships. For example, Shanks, Pearson and Dickinson [bib_ref] Temporal Contiguity and the Judgment of Causality by Human Subjects, Shanks [/bib_ref] exposed their participants to positive contingencies in which the DP was .75 and the temporal delay between actions and outcomes varied from 0s to 16 s. A 2 s delay significantly reduced causal ratings and delays of greater than 4s completely eliminated any perception of causality. Thus, even in conditions with a strong contingency between action and outcome, degrading temporal contiguity attenuated and even eliminated the perception of causality. These and other findings suggested that time was a more critical cue to causality than contingency because, even in the case of a strong contingency, changing the temporal parameters of the task eliminated the perception of cause [bib_ref] Temporal Contiguity and the Judgment of Causality by Human Subjects, Shanks [/bib_ref] [bib_ref] Contiguity and the outcome density bias in action-outcome contingency judgements, Vallee Tourangeau [/bib_ref] [bib_ref] Instrumental judgements and performance under variations in action-outcome contingency and contiguity, Shanks [/bib_ref]. That being said, two assumptions underlie that conclusion, that the experienced temporal sequence of eventoutcome conjunctions is consistent with that programmed by the experimenter, and thus, the contingency experienced by the participant is also isomorphic with that programmed by the experimenter. As we discuss later on in this paper, and has been discussed elsewhere [bib_ref] Contiguity and covariation in human causal inference, Buehner [/bib_ref] , these are two assumptions which can be questioned. Theoretical Approaches to Causal Learning
The debate on whether time or contingency is the more critical or important causal cue relates to the two broad theoretical frameworks that each use different explanatory mechanisms to account for the effects of time and contingency on causal learning. One approach explains causal learning through a simple timesensitive error correction learning algorithm and the development of simple associations between actions and outcomes, such that the strength of the association is correlated with the strength of the causal relationship [bib_ref] Human instrumental learning: A critical review of data and theory, Shanks [/bib_ref]. Another theoretical approach holds that people use causal knowledge, including knowledge of the role of time in causality, to generate propositions or inferences about whether causality is present, before using contingency information to assess the strength of that relationship [bib_ref] Structure and strength in causal induction, Griffiths [/bib_ref] [bib_ref] Theory-based causal induction, Griffiths [/bib_ref] [bib_ref] Two Mechanisms of Human Contingency Learning, Sternberg [/bib_ref] [bib_ref] The propositional approach to associative learning as an alternative for association formation..., Houwer [/bib_ref]. Consistent with the latter view, situation specific knowledge of temporality in causal relationships mitigates some time effects (e.g., delayed distal effects of cause [bib_ref] Abolishing the effect of reinforcement delay on human causal learning, Buehner [/bib_ref] [bib_ref] Rethinking temporal contiguity and the judgement of causality: Effects of prior knowledge,..., Buehner [/bib_ref] [bib_ref] Knowledge mediates the time frame of covariation assessment in human causal induction, Buehner [/bib_ref] , and temporal information has been shown to override contingency information and mislead, producing erroneous causal judgments [bib_ref] Time as guide to cause, Lagnado [/bib_ref]. Thus, temporal information and knowledge should take precedence over contingency in inferring causality. Although there are other key differences between associative and knowledge based models (such as state of association versus truth of inference [bib_ref] The propositional approach to associative learning as an alternative for association formation..., Houwer [/bib_ref] , of relevance at this point are the different mediating mechanisms for time and contingency effects.
Associative models, such as the Rescorla-Wagner model (RWM: [bib_ref] A theory of Pavlovian conditioning: Variations in the effectiveness of reinforcement and..., Rescorla [/bib_ref] , explain both time and contingency effects through the interference of background context in the development of actioneffect associations. For any one effect, there is a finite amount of associative strength available to the relevant action. Thus, although all stimuli present at the same time as the putative cause will develop associative links, they must compete with one another for their share of the overall associative strength available. Causal judgments change with variations in contingency because every time an effect occurs in the absence of the action (see [fig_ref] Figure 1 262: contingency tables showing the four possible combinations of action -effect information [/fig_ref] , the context-effect association becomes stronger and this, in turn, interferes with the development of the action-effect association by reducing the amount of associative strength available to it. Similarly, delays between action and outcome allow the experimental context to be more temporally contiguous with the outcome than the action, allowing the context rather than the action to develop associative strength [bib_ref] Temporal Contiguity and the Judgment of Causality by Human Subjects, Shanks [/bib_ref]. A reduced rate of actioneffect occurrence over time would produce the opposite effect [bib_ref] Depressive realism and outcome density bias in contingency judgements: The effect of..., Msetfi [/bib_ref]. Thus, while time and contingency are important to the development of associations and determining the strength of causal relationships, according to the associative model, the explanatory mechanism for both effects is through the development of interfering contextual associations.
Inference or knowledge based models can also explain the effects of time and contingency on causal judgments. However, these models hold that knowledge of time, in terms of contiguity and succession, is a starting point where causal knowledge is used to generate initial causal models or hypotheses about a given situation, with contingency information used secondarily to test causal hypotheses [bib_ref] Spontaneous assimilation of continuous values and temporal information in causal induction, Marsh [/bib_ref] [bib_ref] The influence of time delays in real-time causal learning, Lagnado [/bib_ref] or assess the strength of the causal relationship [bib_ref] Structure and strength in causal induction, Griffiths [/bib_ref] [bib_ref] Theory-based causal induction, Griffiths [/bib_ref] [bib_ref] Causal inferences as perceptual judgments, Anderson [/bib_ref]. Importantly, the assumption of the existence of causal knowledge can, quite naturally, explain data which associative models find difficult to incorporate. For example, although delay effects are generally robust, temporal knowledge provided in the form or instruction or cover story has been shown to mitigate their effect. Thus plausible delays consistent with a cover story, such as a grenade being fired towards a target and producing an explosion several miles away, are less deleterious to causal learning than unexplained delays [bib_ref] Abolishing the effect of reinforcement delay on human causal learning, Buehner [/bib_ref] [bib_ref] Rethinking temporal contiguity and the judgement of causality: Effects of prior knowledge,..., Buehner [/bib_ref] [bib_ref] Knowledge mediates the time frame of covariation assessment in human causal induction, Buehner [/bib_ref] [bib_ref] How temporal assumptions influence causal judgments, Hagmayer [/bib_ref]. Although the effect of the plausibility of delay is often taken as supporting knowledge based models of causal judgments, it has also been argued that associative models, which code for time [bib_ref] Information and expression of simultaneous and backward associations: Implications for contiguity theory, Matzel [/bib_ref] , can account for knowledge based time effects [bib_ref] Temporal Contiguity and Contingency Judgments: A Pavlovian Analogue, Allan [/bib_ref].
This short discussion shows that the functions of time, context and contingency distinguish theoretical accounts of causal learning. Associative models are time sensitive but explain contingency and time effects on the causal learning process through the development of contextual associations. Knowledge about temporality in causal relationships, on the other hand, is critical to establishing the existence of cause according to some knowledge-based accounts of causal learning. Contingency information is used subsequently to establish the strength of such causal relationships, though time can impact experienced contingencies also through changes in event-outcome conjunction categorization (i.e. cell A might be experienced as cell B [bib_ref] Contiguity and covariation in human causal inference, Buehner [/bib_ref]. However, in terms of which perspective, if any, is best supported by empirical evidence, and as we have argued above, time and context are inter-twined. Moreover, thus far, the interplay between time and context in causal learning has not explicitly been studied. In the next section, we argue that introducing an individual difference variable into experimental work, namely depression, may prove fruitful in terms of elucidating the mechanisms responsible for time effects on causal learning.
## Depression and causal learning
Depression effects on causal learning are particularly important because existing evidence suggests that they are moderated by experiences of time and context. For example, a growing body of evidence shows that even quite mild levels of depression affect people's ratings of the causal consequences of their actions [bib_ref] Judgement of contingency in depressed and nondepressed students: Sadder but wiser, Alloy [/bib_ref] [bib_ref] Contingency judgements by depressed college students: Sadder but not always wiser, Benassi [/bib_ref] [bib_ref] Illusion of control for self and others in depressed and non-depressed college..., Martin [/bib_ref] [bib_ref] Judgement of contingency: Cognitive biases in depressed and nondepressed subjects, Vasquez [/bib_ref]. In order to explore the mechanisms underlying these effects, several studies have manipulated exposure to context by varying the length of time -the duration of the inter-trial interval (ITI) -during the causal learning procedure when no other events take place [bib_ref] Depressive realism and outcome density bias in contingency judgements: The effect of..., Msetfi [/bib_ref] [bib_ref] Depressive realism and the effect of intertrial interval on judgements of zero,..., Msetfi [/bib_ref]. For participants categorized as non-depressed, the trend was for long ITIs to increase the perception of causality in the presence of a zero or positive contingency, but weaken that impression in the presence of a negative contingency. In other words, these time effects were asymmetrical over contingencies and were consistent with the idea that the temporal manipulations affected the strength of contextoutcome associations.
On the other hand, participants categorized as mildly depressed displayed quite a different pattern of effects. A contrast between medium (3s) and long ITI (15s) exposure did not affect ratings of a zero contingency [bib_ref] Depressive realism and outcome density bias in contingency judgements: The effect of..., Msetfi [/bib_ref] but the difference between very short (0.5s) and long (15s) ITI exposure decreased the perception of causality with zero and positive contingencies [bib_ref] Depressive realism and the effect of intertrial interval on judgements of zero,..., Msetfi [/bib_ref] , and increased the perception of negative cause when the contingency was negative (E3). Although these latter findings were reported as of borderline significance using a conservative rejection criterion, the trend was again asymmetrical but diametrically in opposition to those effects displayed by non-depressed groups. A more conservative position would of course be that the time/context manipulations had no effect on the causal judgments made by mildly depressed participants.
Irrespective of the theoretical interpretation of these particular findings, it is evident that simultaneously manipulating time and exposure to the context affected the causal judgments of mildly depressed and non-depressed people differently. This could equally be due to mood related changes in time perception [bib_ref] The Effect of Mild Depression on Time Discrimination, Msetfi [/bib_ref] [bib_ref] Dysphoric mood states are related to sensitivity to temporal changes in contingency, Msetfi [/bib_ref] or processing of context [bib_ref] Impaired context maintenance in mild to moderately depressed students, Msetfi [/bib_ref]. However, exploring the underlying reason for this difference, whether located at the level of time or context processing, can inform about the relative contributions of time and context as causal cues.
In the series of experiments reported here, then, we planned to test predictions about how a range of temporal manipulations will affect causal judgments, in particular durations of action-outcome delays and inter-trial intervals. Predictions can then be extended to the effects of depressed mood on causal judgments. For example, numerous studies suggest that time perception is slowed in depression [bib_ref] Time experience and time judgment in major depression, mania and healthy subjects...., Bschor [/bib_ref] [bib_ref] Time passes slowly for patients with depressive state, Kitamura [/bib_ref] , even mildly dysphoric states [bib_ref] The Effect of Mild Depression on Time Discrimination, Msetfi [/bib_ref]. If slowed time perception underlies the effect of depression on causal learning, this would suggest that time effects will be magnified in participants who are categorized as depressed. However, it is equally possible that impaired processing and maintenance of context representations [bib_ref] Impaired context maintenance in mild to moderately depressed students, Msetfi [/bib_ref] are responsible for depression effects, in this is the case, then all time/context exposure manipulations will have reduced effects in depressed participants in comparison to controls. Studying how depression levels moderate the effects of the time and context manipulations, we have just described, will inform about the combinatorial process underlying causal judgments as well as elucidate the mechanisms through which depression affects causal learning.
## Experiment 1
Previous studies involving depression, temporal manipulations or context in causal learning [bib_ref] Depressive realism and outcome density bias in contingency judgements: The effect of..., Msetfi [/bib_ref] [bib_ref] Judgement of contingency in depressed and nondepressed students: Sadder but wiser, Alloy [/bib_ref] have used a limited range of conditions, mainly those in which the contingency between cause and effect was zero and outcomes were frequent. This means that extant data on depression effects in causal learning currently provide an insufficient baseline against which to explore the effects of temporal and contextual manipulations in the subsequent experiments we plan to report here. Therefore, in Experiment 1, we used an instrumental causal learning task to test a range of preventative (DP = 2.5), generative (DP = +.5) and non-contingency conditions (DP = 0), with different levels of outcome density (low, high). The cover story and visual stimuli used in the task included an explicit and realistic context. The goal of this was to directly consider the role of context in causal learning alongside the other causal cues. Thus, although indirect evidence for the role of context can be obtained from temporal manipulations as will be tested in subsequent experiments, direct evidence can also be derived from explicit ratings of the causal relationship between the context and the outcome. This approach to direct measurement of context is supported by data from previous studies we have carried out in which causal relationships were embedded in realistic virtual contexts, with participants then required to rate the causal relationships between the context, action and outcome [bib_ref] 5-HT Modulation by Acute Tryptophan Depletion of Human Instrumental Contingency Judgements, Chase [/bib_ref]. As expected, context ratings were higher than action ratings with zero contingencies and this pattern was reversed with positive contingencies [bib_ref] 5-HT Modulation by Acute Tryptophan Depletion of Human Instrumental Contingency Judgements, Chase [/bib_ref] -see supplementary data. Moreover, ratings were sensitive to relatively small elevations in depressed or dysphoric mood. Those previous findings indicate the suitability of a virtual context procedure, like that used in Experiment 1, to provide a more comprehensive data set than those currently available, against which to consider the findings of the subsequent experiments reported here.
# Method
Ethics statement. Ethics approval was obtained from the ethical review committees of the Universities of Limerick and Oxford for all experiments reported here, and written informed consent was obtained from all participants prior to participation.
Participants. University students were recruited via a mass screening method, which required all volunteers to complete the Beck Depression Inventory, hereafter BDI [bib_ref] An inventory for measuring depression, Beck [/bib_ref] , as a measure of their current mood state before being invited to participate. BDI scores were taken again during participation and used to assign 50 participants to the high BDI group (scores of 9 or above: n = 24 with n = 4 males) or the low BDI group (scores of 8 or below: n = 26 with n = 17 males). These criteria are consistent with no dysphoria in the low BDI group and mild dysphoria in the high BDI group and have been used in many previous studies [bib_ref] Depressive realism and outcome density bias in contingency judgements: The effect of..., Msetfi [/bib_ref] [bib_ref] Judgement of contingency in depressed and nondepressed students: Sadder but wiser, Alloy [/bib_ref].
As the procedure was not fully randomised, groups were matched on potential confounds such as age, working memory capacity, and estimated pre-morbid IQ, which could have contributed to any between groups effects (see [fig_ref] Table 1: Demographic characteristics of participants in Experiment 1 compared across low and high... [/fig_ref]. Working memory capacity was measured using the forward version of the digit span test, and premorbid IQ was estimated using demographic data (for method and equations see [bib_ref] A demographically based index of premorbid intelligence for the WAIS-R, Barona [/bib_ref]. Independent groups t-tests showed that there were no between group differences in age, digit span score or estimated IQ. As expected, the high BDI group produced significantly higher scores on the BDI at both at screening and during their visit to the lab (see [fig_ref] Table 1: Demographic characteristics of participants in Experiment 1 compared across low and high... [/fig_ref].
Design and Materials. In this experiment, we used a 26(36262) fully factorial mixed design. The within subjects factors were contingency (negative, zero, positive) outcome density [OD] (low, high) and cue (action, context). The between subjects variable was BDI group (low, high). Thus, each participant was exposed to six different contingency conditions, where the programmed DP values were: 20.5 low OD (0.0|0.5), 20.5 high OD (.5|1.0), 0 low OD (.25|.25), 0 high OD (.75|.75), +0.5 low OD (0.5|0.0), and +0.5 high OD (1.0|0.5), where the first value in each parentheses is p(Outcome|Action) and the second value is p(Outcome|NoAction). The cue variable refers to the two different causal ratings that participants were required to provide for each condition, action and context. Each condition was located within a distinct virtual context represented by pictures on the computer screen. The action was a key press on the computer keyboard and the outcome was a 2s auditory stimulus. Following each condition, participants were asked to rate their own control (action), and that of the context, over an auditory outcome using a judgement scale which varied from 2100 (labelled totally prevent) through 0 (labelled no influence) to +100 (labelled totally control). Order of presentation was counterbalanced using a Latin squares design. Presentation of Procedure. During their visit to the laboratory, participants were briefed verbally about the nature of the experiment and then given a written information sheet to read. After giving written informed consent, participants provided demographic information, completed the digit span task, and questionnaires measuring mood state. Following this, instructions about the causal learning task requirements were displayed on the computer screen. The cover story required participants to imagine that they were in a house in which there was a hidden stereo system. They could control the music switching on in each of the rooms in the house (distinct contexts) using a remote control. However, participants were told that the remote control had been working intermittently, and that sometimes music switches on when no one is touching the remote control. The task was therefore to test the remote control in each of the rooms separately.
For each test, participants were told that they would be taken to the particular room and that they should wait to receive a message on the computer screen saying that they were allowed to test the remote control by pressing the spacebar on the computer keyboard. This would happen on many occasions (experimental trials) while they were in the room, and participants could choose to press the space bar at that point, or simply observe. In order for participants to properly gauge what happens when they did not press the button, they were asked to press on approximately half of the possible occasions.
Each experimental trial was constructed such that the message signalling the possibility of the action would stay on the screen for 3s. If the spacebar were pressed during this period, the button on the remote control shown on the screen would show as depressed. No further responses were possible during that particular response time window. At the end of the time window, the music would either play for 2s at a probability of p(Outcome|Action) or the room would remain silent. If the spacebar were not pressed during the time window, then the music would switch on for 2 s at a probability of p(Outcome|NoAction). This 5s procedure constituted one experimental trial, of which they were 40 in each condition, separated by a 3s inter-trial intervals (ITIs) during which the same visual stimuli (the virtual context) remained the same as during the trial.
At the end of each set of 40 trials, a judgement window was displayed and participants were required to rate the causal relationship between their own action and the outcome, and between the distinct context and the outcome using sliders displayed on the computer screen. The judgement sliders were constructed with increments of +/21, so that the full range of the judgement scale (2100 to +100) could be used. After completing all six conditions, participants were thanked, debriefed and paid a nominal fee for their participation. All participants were also provided support information relevant to mood states.
# Results and discussion
Participants rated the control of their actions over the outcome as well as that of the context in six different conditions, including negative, positive and zero contingencies with a low and high density of outcomes. These data are shown in [fig_ref] Figure 2: Ratings of the causal strength of the action and the context [/fig_ref] and suggest that participants' ratings distinguished between action and context, and between the contingencies and the density of outcome occurrence. However, the experimental manipulations seemed to have a weaker effect on ratings made by the high BDI group in comparison to the low BDI group.
These observations were examined using a mixed (36262) 62 factorial analysis of variance, with contingency, outcome density and cue as within subjects factors and BDI group as the between subjects variable. An alpha level of .05 was used here and throughout unless stated otherwise. As we might have expected, contingency affected all ratings, F(2, 96) = 43.07, p,.001, g 2 = .47, MSE = 1077.63, but the direction of the contingency effect depended on which cue, action or context, was rated, F(2, 96) = 113.32, p,.001, g 2 = .70, MSE = 1084.63, as well as participant group, because the three-way interaction between contingency, cue and BDI group was significant, F(2, 96) = 7.59, p = .001, g 2 = .14, MSE = 1084.63. Before exploring that interaction further, it is important to note that the density of outcomes affected ratings, F(2, 96) = 131.67, p,.001, g 2 = .73, MSE = 2592.03, but that this effect depended on the cue rated, F(1, 48) = 4.40, p = .041, g 2 = .08, MSE = 1774.47, and the contingency, F(2, 96) = 8.80, p,.001, g 2 = .16, MSE = 1051.46, but not BDI group, p = .12. In general, although the ratings of low and high outcome density conditions were located in different regions of the judgment scale the pattern of difference was similar. High outcome density conditions always received action ratings and context ratings that were more towards the positive end of the judgment scale than low outcome density conditions. For negative contingency conditions, this meant that high outcome density action ratings were weak and located nearer to zero on the judgment scale than low outcome density action ratings.
Further analyses of the contingency, cue and BDI group interaction revealed quite straightforward effects. For action ratings, the contingency by BDI group interaction was significant, F(2, 96) = 6.68, p = .002, g 2 = .12, MSE = 1004.28. Both low and high BDI groups were sensitive to the difference between contingency conditions, where negative , zero , positive ratings (p,.001 for both groups, with all pairwise comparisons p,.001). However, as revealed by the significant interaction, the size of the contingency effect was greater for the low BDI group (g 2 = .89) than the high BDI group (g 2 = .72). For the context ratings, again the simple interaction between contingency and BDI group was reliable, F(2, 96) = 3.73, p = .028, g 2 = .07, MSE = 1877.98. For the low BDI group, context ratings were significantly affected by contingency, F(2, 50) = 23.65, p,.001, g 2 = .49, MSE = 1343.89, with the ordering of mean context ratings being in the opposite direction to action ratings, negative . zero . positive, with all pairwise comparisons significant with p,.005. In the high BDI group, however, contingency had no effect on context ratings, F(2, 46) = 1.34, p = .26, thus negative = zero = positive context ratings.
In summary, when participants, categorized as having low or high BDI scores, were exposed to a series of six different contingency conditions, low BDI participants' ratings were consistent with a greater degree of discrimination between the contingencies than high BDI participants, and this was consistent irrespective of the density of outcomes. Previous work has only found differences between mood groups in specific high outcome density conditions when each participant was only exposed to one contingency condition [bib_ref] Depressive realism and outcome density bias in contingency judgements: The effect of..., Msetfi [/bib_ref] [bib_ref] Judgement of contingency in depressed and nondepressed students: Sadder but wiser, Alloy [/bib_ref] ; the findings of this experiment show more widely spread mood effects which are present in conditions more similar to the real world in which there are numerous contingencies to judge and compare. In addition, these results provide direct evidence for the first time of mood effects on people's ratings of the context's causal relationship with the outcome in a wide range of conditions. The high BDI group's context ratings did not vary as a function of contingency as the low BDI group's ratings did.
The findings do show very clearly how, for low BDI groups, their ratings of the causal role of the action varied systematically and with their causal ratings of the context. This is certainly consistent with the view of context as playing an important role in causal learning. The high BDI group, on the other hand, did discriminate between contingencies in terms of their action ratings but not their context ratings. High BDI context ratings did not change significantly across contingency although they did reflect levels of outcome density. This pattern is suggestive of causal judgments that interact less with context, or background, as possible causes of effects, though are highly sensitive to the background rate of outcomes.
The findings of Experiment 1 therefore provide us with baseline measures of causal relationships between action, context and outcomes across multiple contingency conditions. They are informative in their own right, in particular as previous studies have always been suggestive that mood effects on causal judgments only occur in very specific conditions, with zero contingencies [bib_ref] Depressive realism: A meta-analytic review, Moore [/bib_ref] especially with long temporal intervals between trials. In fact, one criticism of this area of research has been that these findings are so specific as be rather meaningless in the real world [bib_ref] Perspectives on depressive realism: Implications for cognitive theory of depression, Haaga [/bib_ref]. However, the results of Experiment 1 do show that pattern of differences in causal learning attributable to depressed mood are more pervasive than previously thought.
## Experiment 2
In the next set of experiments, we varied the durations of the inter-trial interval (ITI) and the action-outcome delay in conditions with a moderately positive contingency and a high density of outcomes (DP = 0.5). Both are manipulations of time but are simultaneously manipulations of exposure to context and both theoretical models made specific predictions about the effects of these manipulations.
From the perspective of associative learning theory, predictions are straightforward. Long action outcome delays mean that the constantly present context enjoys greater contiguity with the outcome that the action and this will strengthen context-outcome associations. Longer durations in between experimental trials (ITIs) would have the opposing effect, creating long periods of context exposure in the absence of the outcome, thus weakening the association between context and outcome. However, these consequences of the strength of the context association will be dependent on the specific contingency condition tested. When contingencies are positive or zero, strong and weak context associations will weaken and strengthen ratings of the action's causal relationship respectively. In the case of a negative contingency, however, the effects would be reversed such that the stronger context association would actually promote a stronger preventative causal relationship between action and outcome. In other words, associative theory would predict asymmetrical time/ context effects that are contingency dependent.
These predictions diverge from those made by knowledge-based models and the following discussion explains the reasons for this and then describes specific predictions. We particularly refer here to causal structure models [bib_ref] Theory-based causal induction, Griffiths [/bib_ref] as examples of what Lagnado and Sloman [bib_ref] Time as guide to cause, Lagnado [/bib_ref] refer to as hypothesis driven accounts of learning. According to this view, knowledge of temporality in causal relationships is key to determining whether or not a causal structure exists. This is only one part of the process. Following that, contingency data is used to determine the strength of the causal relationship. In the experiments reported here, both components of the process are required as participants are asked to rate how much, if any, control they have over the music switching on. Accordingly, in order to make the rating, they must use their knowledge of the plausibility, ontology and form of causal relationships to establish whether it exists or not, before establishing its strength [bib_ref] Theory-based causal induction, Griffiths [/bib_ref].
According to Griffiths and Tenenbaum, it seems likely that people make different assumptions about generative and preventative relationships (with different strength parameterization calculations following on from this). However, given the plausibility of a generative relationship in the experimental scenario used here (remote button usually causes music to switch on), and the relative implausibility of a preventative relationship (remote button doesn't usually prevent music from switching on), it seems likely that people would assume a generative relationship and use their temporal knowledge in this way, such that delay is incompatible with causality. Specifically then, in this generative scenario, delay effects should be contingency independent, should be symmetrical across positive and negative contingencies, and should eliminate the perception of causality. This would be the case unless the notion of prevention is very clearly part of the causal scenario [bib_ref] Causal induction: The Power PC theory versus the Rescorla-Wagner Theory, Buehner [/bib_ref] , which was not the case here (i.e. the rating scale allowed for prevention, but the scenario did not include it). We will return to these points later in the general discussion.
While predictions about delay do seem to distinguish the models, predictions around ITI duration do not. Instead they relate to how rates and probability of event occurrence are linked to knowledge in the first place and then use of contingency information to qualify initial assessments. As with associative theory, asymmetrical causal ratings would be predicted. Increasing the duration of the ITI will reduce the base rate of the effect [p(Outcome|No Action)] because of the conceptual similarity between ITIs and the D cell of the contingency table, and would increase the perception of generative causality and decrease the perception of preventative causality, as predicted by associative theory as well. However, there is some ambiguity to these ITI predictions. If rate and probability are processed online over time, rather than over N trials [bib_ref] On the origin of personal causal theories, Young [/bib_ref] (N trials would likely be controlled by the experimenter), then for a given time window, longer ITIs might also decrease the perceived rate of action-effect cooccurrences as well. In other words, both relevant conditional probabilities [p(Outcome|no Action) and p(Outcome|Action)] would decrease, maintaining the overall contingency, having no effect on causal strength. Thus, as causal strength is based on contingency information experienced over time, knowledge based theory could also predict the ITIs would have no effect on the strength of causal ratings.
One methodological issue, however, is that manipulating durations, which occur within or between trials, also affects the overall duration of exposure to a particular contingency condition [bib_ref] Depressive realism and the effect of intertrial interval on judgements of zero,..., Msetfi [/bib_ref]. Conditions with longer ITIs and delays necessarily involve a longer procedure time than shorter ITIs and delays if numbers of trials are held constant across conditions. We therefore carried out two versions of the experiments reported next; one in which the number of experimental trials was held constant while procedure time was varied (version A), and another in which the number of experimental trials was varied while procedure time was held constant (version B). We only report the details of the individual experiments where relevant, as there were no significant differences between the two.
# Method
Participants. All participants completed the Beck Depression Inventory on two occasions, a maximum of 14 days before participation and then again during the visit to the lab. Participants were assigned to mood groups on the basis of the BDI scores taken in the lab (Experiment 2a: N = 50; Experiment 2b: N = 53). In this experiment, we used median BDI scores to assign participants to the low and high BDI groups. Consequently, those who scored 5 or below on the BDI were assigned to the low BDI group while those who scored 6 or above were assigned to the high BDI group. Mood effects have been observed using the same criteria in previous experiments [bib_ref] 5-HT Modulation by Acute Tryptophan Depletion of Human Instrumental Contingency Judgements, Chase [/bib_ref]. The data for five participants were excluded, one female participant due to computer malfunction, and four other participants due to low response rates p(R) ,.13). The characteristics of the final sample are shown in [fig_ref] Table 2: Experiment 2a and Experiment 2b demographic characteristics, comparisons across experiment and mood... [/fig_ref] and comprised N = 46 participants in Experiment 2a and N = 52 participants in Experiment 2b.
The low and high BDI groups were compared on a range of relevant demographic variables using multivariate analysis of variance, including age, years of education, and scores on the digit span test. There were no significant differences between BDI groups on these variables, or across Experiments 2a and 2b (see [fig_ref] Table 2: Experiment 2a and Experiment 2b demographic characteristics, comparisons across experiment and mood... [/fig_ref]. All participants received a nominal payment for their participation.
Design and Materials. This experiment used a 26 (26262) mixed factorial design, where BDI group (low, high) was the between subjects variable. The within subjects variables were delay (short 0 s, long 4 s), ITI length (short 3 s, long 15 s) and cue (action, context). These durations of delay and ITI have been used in previously published studies [bib_ref] Temporal Contiguity and the Judgment of Causality by Human Subjects, Shanks [/bib_ref] [bib_ref] Depressive realism and outcome density bias in contingency judgements: The effect of..., Msetfi [/bib_ref]. The factorial combination of experimental manipulations resulted in four conditions experienced by all participants, with two ratings -action and contextmade for each condition. Order of presentation was counterbalanced using a Latin squares design.
The procedural details were the same as the previous experiment except that durations were varied. [fig_ref] Table 3: Component durations [/fig_ref] shows the duration of each component of the procedure. In all conditions, there was a 3s period in which participants could choose to make the action. One difference between this experiment and Experiment 1, is that if an action occurred during the 3 s response window, then the outcome followed immediately or after 4s at a probability of p(Outcome|Action) rather than at the end of the response window. If no action was recorded by the end of the response window, then an outcome followed immediately or after 4s at a probability of p(Outcome|No Action). Outcomes lasted for 2 s. In Experiment 2a, each condition included 60 trials with the duration of each condition varying accordingly, whereas in Experiment 2 b, the number of trials was varied in order to hold the duration of each condition constant (see [fig_ref] Table 3: Component durations [/fig_ref]. In all cases, the overall procedure time, including all four conditions, lasted for 64 minutes.
In all conditions, there was a moderately positive contingency (DP = .5) between the action and the outcome, such that actions always resulted in an outcome with a probability of 1.0 and trials with no action resulted in an outcome at a probability of .5.
Procedure. Procedural details are identical to Experiment 1.
# Results and discussion
As with the previous experiment, participants rated the effectiveness of their own actions, and the effectiveness of the context, in controlling the occurrence of the music. These data, combined across Experiments 2a and 2b, are described below and shown in [fig_ref] Figure 3: Effects of ITI length and delay duration on ratings of positive [/fig_ref].
In general, participants judged that the causal role of their own actions was stronger than the causal role of the context. However, the effects of delay and ITI length seemed to depend on both BDI group and the specific causal role rated. The data were analyzed using Analysis of Variance with the following variables, cue, delay and ITI, included as within subjects factors, with BDI group (low, high) and Experiment (2a versus 2b) entered as between subjects variables.
The analyses showed that action ratings were indeed significantly higher than context ratings, cue: F(1, 94) = 50.35, p ,.001, g 2 = .35, MSE = 3040.38, and that delay affected ratings also, F(1, 94) = 19.40, p,.001, g 2 = .17, MSE = 1361.91. However, delay effects depended on cue, cue 6 delay: F(1, 94) = 6.16, p = .015, g 2 = .06, MSE = 1372.49, as well as ITI length, cue 6 delay 6 ITI: F(1, 94) = 6.69, p = .011, g 2 = .07, MSE = 782.27, and BDI group, cue 6 delay 6 ITI 6 BDI: F(1, 94) = 9.92, p = .002, g 2 = .10, MSE = 782.27. None of the effects or interactions involving Experiment (2a versus 2b) were reliable or were explored further.
In order to explore the four-way interaction between cue, delay, ITI, and BDI group in more detail, we split the data by BDI group and carried out further analyses. For the low BDI group [fig_ref] Figure 3: Effects of ITI length and delay duration on ratings of positive [/fig_ref] , left), the cue by delay interaction was reliable, F(1, 43) = 7.97, p = .007, g 2 = .16, MSE = 970.62, although the cue by ITI interaction was not reliable, p = . [bib_ref] Contingency judgements by depressed college students: Sadder but not always wiser, Benassi [/bib_ref]. Although we might have expected an effect of ITI length on action judgments, and there was some suggestion of this in [fig_ref] Figure 3: Effects of ITI length and delay duration on ratings of positive [/fig_ref] , the ITI effect on action judgments was not reliable, p = .18. Simple effects analysis showed that whereas long delays reduced action ratings significantly, F(1, 43) = 16.14, p,.001, g 2 = .27, context ratings remained the same, p = .30.
The pattern was different in the high BDI group. The cue, delay by ITI interaction was reliable, F(1, 51) = 15.38, p,.001, g 2 = .23, MSE = 923.16. When delays were short (squares, [fig_ref] Figure 3: Effects of ITI length and delay duration on ratings of positive [/fig_ref] , ratings were not dissimilar to the low BDI group, with action rated significantly higher than context, p , .003, g 2 ..16, and with no discernible effect of ITI length, p ..12, g 2 ,.05. However, when delays were long, action judgments increased and context judgments decreased significantly with longer ITIs, both ps , .02, both g 2 . .10. In other words, ITI effects were strongly evident but only in long delay conditions. In long ITI conditions, there was no difference between short and long delay action judgments, F(1, 51) = 1.04, p = . [bib_ref] Two Mechanisms of Human Contingency Learning, Sternberg [/bib_ref].
Taken together these findings show that manipulations of the time and context affect causal judgments but that the nature of these effects depends on levels of depressed mood. For low BDI groups, there were trends, some significant, towards effects observed in previous studies. Longer action-outcome delays reliably decreased people's judgments of causal relationships, whereas longer ITIs did not reliably affect causal judgments. These effects were not mirrored in context judgments. Thus, this pattern of findings is most consistent with the notion of a timebased moderation of causal judgments in low BDI participants.
However, it could be argued for several reasons that the data from high BDI groups was consistent with effects related to time and context. This is because slowed time experience in depression should magnify the effects of time manipulations, whereas impaired representation or processing of context would reduce the strength of time effects. Consistent with the slowed time experience view, the delay and ITI effects were stronger in high BDI groups' judgments. However, this magnification of time effects concurrently affected context ratings in a manner consistent with contextual mediation of time effects.
The implications of this are as follows. In Experiment 1, we observed strong mirroring of action-context ratings in low BDI as a function of contingency effects. This might suggest then that the time manipulations, tested with identical contingency conditions in Experiment 2, were not strong enough to produce the action context mirroring in low BDIs that we observed in Experiment 1. However, for high BDIs, their slower time experience increased perceived duration to the extent that time effects were stronger and thus mirrored in context ratings. The fact that ITI effects, which involve much longer durations and therefore more accrued time, were stronger in long delay conditions is consistent with this interpretation of the findings.
## Experiment 3
Negative or preventative contingencies (see [fig_ref] Figure 1 262: contingency tables showing the four possible combinations of action -effect information [/fig_ref] may be more informative than positive contingencies. This is because if the plausible causal relation is generative, knowledge based accounts predict that action outcome delays will reduce or eliminate the perception of action-outcome causality. However, associative theory would not make the same prediction. This is because contingency effects on causal learning are based on the strength of context as well as action associations. In the case of negative contingencies, context associations are very strong in a positive direction, whereas action associations are strong in an inhibitory direction. This means that context associations could be asymptotic, such that increasing the action outcome delay would likely not increase the strength of the context association. Moreover, any effect of delay on strengthening the context association would be likely to increase rather then decrease the If perceived rates of action-outcome occurrences over time are also reduced, this could mean that ITI has no effect on ratings.
# Method
Participants. Ninety-nine participants took part either in the fixed trials (Experiment 3a: N = 50) or the fixed time (Experiment 3b: N = 49) experiment. They completed the BDI as in the previous experiment and were assigned to the low and high BDI groups using the same median split criteria (BDI = 5 cut off). However, data from seven participants were excluded for low response rates (n = 2: p(response) ,. [bib_ref] Context, time, and memory retrieval in the interference paradigms of Pavlovian learning, Bouton [/bib_ref] and high response rates (n = 4: p(response) . .85), and one further participant didn't use the keyboard as instructed which resulted in missing data. The characteristics of the final sample are shown in [fig_ref] Table 4: Experiment 3a and Experiment 3b demographic characteristics, comparisons across experiment and mood... [/fig_ref] and comprised N = 44 participants in Experiment 3a and N = 49 in Experiment 3b. There were no significant differences across experiments or mood groups on demographics, although high BDI groups scored significantly higher on the BDI at both time points than low BDI groups.
Design and procedure. The design was identical to Experiment 2, except that a negative contingency was tested (DP = 2.5).
Actions resulted in an outcome at a probability of .5, whereas trials with no actions always ended in an outcome (p = 1). The instructions and judgment scale were identical to the previous experiment, as were the procedural details.
# Results and discussion
Participants were exposed to a moderately negative contingency and rated the extent to which their actions and the context controlled or prevented the occurrence of the music. These data are combined across Experiments 3a and 3b. Action ratings made by low and high BDI groups are shown in [fig_ref] Figure 4: Effects of ITI length and delay duration on ratings of negative [/fig_ref] (filled symbols). For high BDI groups only, longer ITIs and longer delays appeared to be associated with more negative action ratings.
In order to explore these observations, the data were analyzed in the same way as in Experiment 2. Context ratings were significantly different to action ratings, such that context was rated as a facilitator of the outcome and the action was rated as a preventer of the outcome, Cue: F(1, 88) = 189.72, p,.001, g 2 = .68, MSE = 4526.87. Notably, action-outcome delay had no significant effect on ratings, F,1. Membership of the BDI group did significantly influence ratings, BDI: F(1, 88) = 5.62, p = .02, g 2 = .06, MSE = 2721.05, and the difference between action and context ratings, BDI 6 Cue: F(1, 88) = 4.09, p = .046, g 2 = .044, MSE = 4526.87. As there was a three way interaction between cue, delay and BDI group, F(1, 88) = 4.46, p = .038, g 2 = .048, MSE = 1711.78, this was the starting point for further analysis involving the BDI variable. Finally, there was also a significant interaction between cue and ITI length, F(1, 88) = 4.27, p = .042, g 2 = .046, MSE = 903.23, which also required further analysis. Firstly, further examination of the cue by delay by BDI interaction, confirmed that both low and high BDI groups rated the context very differently to the action, all Fs .79, all p,.001, all g 2 ..63, and this difference was consistent across short and long delays, all ps ..13. Further tests showed that the source of the three-way interaction was that high BDI participants rated the context as a stronger cause of the outcome than low BDI participants, specifically in long delay conditions, F(1,88) = 11.44, p = .001, g 2 = .12, MSE = 1448.83. This was not the case with short delays or with action ratings, all Fs,.2.12, all ps..14. We also checked whether long delays produced reliably more negative ratings in long versus short delay conditions, as suggested in [fig_ref] Figure 4: Effects of ITI length and delay duration on ratings of negative [/fig_ref] , however, that effect was not reliable, F(1,43) = 2.10, p = .16, g 2 = .05, MSE = 1636.75.
Finally, we also examined the cue by ITI length interaction. Despite a trend towards ITI effects on causal ratings of the action for all participants, simple effects analysis showed that this was not significant, F(1, 88) = 3.78, p = .055, g 2 = .04, MSE = 753.17. There was no suggestion of any ITI effect on context ratings, p = .25.
Overall, in Experiment 3, the context was rated as a facilitator of the outcome whereas the action was rated as a preventer. Delay and ITI length had no effect on ratings made by the low BDI group. However, the high BDI group rated the context as more strongly facilitative than the low BDI group, specifically in long delay conditions, although this effect was not mirrored reliably in action ratings where the delay effect did not reach criterion. Therefore, in this particular condition, in which the context was a strong cause of the effect by nature of the specific contingency tested, temporal manipulations had no effect on action ratings.
## Experiment 4
The previous experiment was designed to pit the predictions of associative and knowledge based models against each other. Moreover, we also wanted to test whether depression effects on causal learning were consistent with slowed time perception in depression strengthening the effects of temporal manipulations, with subsequent effects on causal learning through contextual associations. The lack of delay effects on action ratings of negative contingencies in the previous experiment is not entirely consistent with either theoretical approach, however it might be that the specific negative contingency tested in Experiment 3 was less likely than any other negative contingency to produce temporal effects.
The specific condition tested in Experiment 3 (see [fig_ref] Figure 1 262: contingency tables showing the four possible combinations of action -effect information [/fig_ref] , lower left) was programmed such that 50% of action trials and 100% of no action trials, during which the context was always present, resulted in an outcome. This configuration, with such a level of high outcome density, would mean that the context and outcome were frequently paired and the association between context and outcome would be very strong and possibly near to the limits of associative or causal strength. In addition, and from a more probabilistic perspective on delay, if a long delay between action and outcome means that the trial is processed as a contextoutcome trial instead, cell C rather than cell A [bib_ref] Contiguity and covariation in human causal inference, Buehner [/bib_ref] , the original 100% likelihood of outcome occurrence after a no action event is at ceiling and cannot therefore be increased. This does not explain why ITI effects were not observed in Experiment 3, however, but we cannot discount the fact that the findings might be related to the specific contingency tested. Therefore, Experiment 4 will repeat the previous experiment (including fixed trials and fixed time versions) with a medium outcome density negative contingency. While the absolute level of contingency (DP = 2.5) will be the same, only 50% of trials will end in an outcome where 25% of action trials and 75% of no action trials will be followed an outcome. This means that the context will be paired with the outcome on fewer occasions than Experiment 3, and the problematic 100% outcome rate after no action trials is no longer the case.
# Method
Participants. Recruitment and BDI completion was carried out on the same basis as previous experiments. One hundred participants completed Experiment 4a (fixed trials: N = 50) or 4b (fixed time: N = 50). However, the data for nine participants were excluded due to response rates ,.15 or ..85. The final sample comprised 44 participants in Experiment 4a and 47 in Experiment 4b (N = 91). The characteristics of the sample are shown in [fig_ref] Table 5: Experiment 4a and Experiment 4b demographic characteristics, comparisons across experiment and mood... [/fig_ref].
Education data was missing in Experiment 4a and therefore no experiment or mood group comparisons are reported for that variable. BDI scores were always significantly different between the low and high BDI groups but not across experiments. Experiment 4b participants were significantly older than Experiment 4a participants. Also high BDI participants were significantly younger than low BDI. Experiment was included as a variable in all analyses as in the previous experiments.
Design and procedure. All details were the same as previous experiments except that a moderate outcome density, negative contingency condition was tested. Outcomes occurred on 25% of action trials and 75% of no action trials.
# Results
Participants rated the action and the context in relation to control over the outcome and the data are shown combined across fixed time and fixed trials versions in [fig_ref] Figure 5: Effects of ITI length and delay duration on ratings of negative [/fig_ref]. The data show a similar pattern to the previous experiment. Action ratings are in the negative portion of the scale and context ratings in the positive end, with little evidence of delay or ITI effects in low BDI ratings. High BDI ratings also looked similar to the previous experiment, except that the effect of longer delays, which pushed their actions ratings in a more negative direction, seemed more pronounced in [fig_ref] Figure 5: Effects of ITI length and delay duration on ratings of negative [/fig_ref] than in the previous experiment.
The data were analysed as in Experiment 3. The difference between context and action ratings was significant for all participants, Cue: F(1, 87) = 130.86, p,.001, g 2 = .60, MSE = 4393.04. In addition, the effect of delay was moderated both by cue and by BDI group, Delay 6 Cue 6 BDI: F(1, 87) = 4.00, p = .049, g 2 = .604, MSE = 4393.04. Examining the low and high BDI groups separately showed that for low BDI groups there was no effect of delay on either cue, Fs ,1, ps ..5. However, in the high BDI group, there was a cue dependent delay effect, F(1, 43) = 4.56, p = .038, g 2 = .10, MSE = 1875.73, such that there was no delay effect on context ratings, F,1, but long delay action ratings were significantly more negative than short delay action ratings, F(1, 43) = 8.73, p = .005, g 2 = .17, MSE = 1431.82.
This pattern of results is very similar to the previous experiment. Again, delay and ITI manipulations did not affect low BDI groups' context and action ratings. However, for the high BDI groups, stronger action ratings of preventative cause in long delay conditions, an effect that was weak and did not reach criterion in the previous experiment, was reliable and based on medium to large effects according to Cohen's [bib_ref] A power primer, Cohen [/bib_ref] criteria. This is consistent with the delay effect increasing in strength in negative contingencies when there is a lower level of outcome density, high levels of which promote very strong context-outcome links. While this may be the case, in a supplementary analysis, we combined the data from both negative contingency experiments, and checked whether the cue 6 delay 6 BDI group interaction depended on the specific contingency (25/75 versus 50/100) condition tested. There was not the case and there were no reliable differences in the interaction between the two experiments, F(1, 179) ,1, p = .99. This finding with high BDI groups only is consistent with slowed time experience magnifying time effects in both positive and negative contingencies in a manner that has a knock on effect on the strength of context associations
# General discussion
In this series of experiments, we set out to explore the processes underlying causal learning, in particular how time and context manipulations affect causal judgments, with depression included as an important moderator variable. Unlike previous studies [bib_ref] Temporal Contiguity and the Judgment of Causality by Human Subjects, Shanks [/bib_ref] [bib_ref] Abolishing the effect of reinforcement delay on human causal learning, Buehner [/bib_ref] [bib_ref] Rethinking temporal contiguity and the judgement of causality: Effects of prior knowledge,..., Buehner [/bib_ref] , and mindful that time manipulations simultaneously affect exposure to the context, we explicitly included measures of context causality. We also tested groups of low and high scorers on a depression scale for whom time and context processing impairments have been documented [bib_ref] The Effect of Mild Depression on Time Discrimination, Msetfi [/bib_ref] [bib_ref] Time perception, depression and sadness, Gill [/bib_ref] [bib_ref] Impaired context maintenance in mild to moderately depressed students, Msetfi [/bib_ref]. We found that for participants with little evidence of depression, effects of time manipulations were only apparent in specific positive contingency conditions. However, for participants scoring higher on the depression scale, time effects were generally stronger and present with positive and negative contingencies. We will discuss these findings in more detail below, also using the contrast between low and high depression scorers, and inconsistencies with previous findings, to inform the theoretical implications of this work. In this series of experiments, we firstly verified that the strength of instrumental causal judgments, made across a range of contingency conditions, varied with similar ratings made about the relation between the context and outcome. This was the case for participants scoring low on the depression scale. However, high BDI groups' context ratings did not vary with action ratings in the same way across contingencies, suggesting a degree of 'decontextualization' in their causal judgments in the absence of any specific temporal or contextual experimental manipulations. This effect was accompanied by action ratings that did not discriminate between contingencies to the same extent that low BDI groups' ratings did. This pattern of findings supports the idea that context typically does have a key function in causal learning, as suggested by associative theories [bib_ref] On the origin of personal causal theories, Young [/bib_ref] , but that this mechanism is vulnerable to low levels of depression and perhaps other psychopathologies.
Further experiments were then designed to test the interrelations between time and context in positive (Experiment 2) and negative (Experiments 3-4) contingencies. Two different time periods were manipulated, the delay between action and outcome and the empty delay between experimental trials, that are thought to weaken and strengthen respectively people's assessments of the causal role of the action. These time effects distinguished low and high BDI scorers. We found that longer time delays did not always reduce the perception of cause for low BDI scorers. While delayed outcomes were perceived as less causal with positive contingencies, as in previous studies, they had no effect on causal judgments when the contingency was negative. It could be argued that this finding suggests that, all other factors being equal including knowledge of temporality and the plausibility of delay in that causal situation, time delays do not always eliminate the perception of a causal relationship as knowledge based theory might suggest [bib_ref] Time as guide to cause, Lagnado [/bib_ref]. We will return to this theoretical point shortly, however, it is the findings from high BDI scorers that are more suggestive of underlying mechanisms.
Firstly, it is useful to reconsider the original predictions that we made about time and depression. Depression has consistently been associated with slowed time perception [bib_ref] The Effect of Mild Depression on Time Discrimination, Msetfi [/bib_ref] [bib_ref] Time experience and time judgment in major depression, mania and healthy subjects...., Bschor [/bib_ref] , which could increase the impact of time manipulations, but also impaired context processing [bib_ref] Depressive realism and outcome density bias in contingency judgements: The effect of..., Msetfi [/bib_ref] [bib_ref] Impaired context maintenance in mild to moderately depressed students, Msetfi [/bib_ref] , which could decrease time effects if they are context based. In the time manipulation experiments reported here, time effects on ratings made by the high BDI groups were stronger than for low BDI participants and it is possible that slowed time perception is the cause of this effect. However, with positive contingencies, both delay and ITI influenced causal ratings, with action and context ratings being influenced in opposition. This might suggest then, that for these participants, slowed awareness of time exerts its effect on causal learning through extended exposure to context and the consequential effect of that on the strength of context associations, rather than the effect of time per se.
Even more informative is the finding that delayed outcomes increased rather than eliminated high BDI participants' perceptions of preventative cause in negative contingency conditions. This means that, for these participants, delay effects were asymmetrical around zero across the judgment scale (ratings of positive short . positive long . negative short . negative long).
While knowledge based theory would predict symmetrical delay effects that eliminate the perception of causality (ratings of positive short . positive long . negative long . negative short), this was not the case for high BDI participants. The asymmetry of delay effects is however consistent with the idea that the stronger relation between context and outcome in long delay conditions would make an inhibitory association between action and outcome even stronger as predicted by associative theory.
These findings are consistent with some, but not all, previous work. For example, Vallee-Tourangeau, Murphy and Baker [bib_ref] Contiguity and the outcome density bias in action-outcome contingency judgements, Vallee Tourangeau [/bib_ref] reported findings consistent with our low BDI groups that variable (degraded) versus constant contiguity deleteriously affected positive action-outcome contingency ratings significantly but not negative contingency ratings. However, in contrast, Mutter, DeCaro and Plumlee [bib_ref] The role of contingency and contiguity in young and older adults' causal..., Mutter [/bib_ref] found symmetrical rather than asymmetrical delay effects with their younger participants. In their study, depression was not a variable of interest and they found that long delays reduced the perception of causality in negative as well as positive contingencies. Their older participants, like our low BDI groups and Vallee-Tourangeau et al. 's participants, displayed no delay effect on negative contingencies. For the most part then, it seems that the effects of delayed outcomes are specific to positive contingency conditions. However, when time effects are enhanced, here due to mild depressed mood, then the full range of delay effects are observable but the nature of the effects are contingency dependent and can enhance and eliminate the perception of causality.
Other inconsistencies between the findings reported here and previous studies relate to the effects of ITI duration on causal ratings. In the present series of experiments, ITI effects on causal ratings were, for the most part, absent or weak and not reliable. However, in our previous work [bib_ref] Depressive realism and the effect of intertrial interval on judgements of zero,..., Msetfi [/bib_ref] , although we reported ITI effects to be weak with positive contingencies, they were strong when the contingency was negative. One reason for this inconsistency in ITI effects, as well as the delay effects mentioned above, might be theoretically important procedural differences. In the present study, and Vallee-Tourangeau et al. [bib_ref] Contiguity and the outcome density bias in action-outcome contingency judgements, Vallee Tourangeau [/bib_ref] , in which patterns of delay effects similar to ours were reported, the time manipulations were tested on a within subjects basis. Mutter's study [bib_ref] The role of contingency and contiguity in young and older adults' causal..., Mutter [/bib_ref] and our own previous work, in which different patterns of delay and ITI effects were reported than those described here, involved between subjects tests of time variables. The within versus between subjects distinction of time effect tests is important as it might imply that time effects are cumulative, such that with multiple conditions time effects are influenced by preceding conditions, thereby explaining the difference in findings.
So far, we have discussed several specific pieces of evidence that inform the theoretical implications of this work and we address this now in detail. Causal structure models, as one example of a knowledge based approach, postulate that time is the primary cue to causality and that contingency is information that is considered subsequently in the process. So, for example, Lagnado [bib_ref] Time as guide to cause, Lagnado [/bib_ref] found that when time information misleads, erroneous causal attributions result. However, it is also clear that knowledge about the plausibility of temporality in a given situation [bib_ref] Abolishing the effect of reinforcement delay on human causal learning, Buehner [/bib_ref] and assumptions about the functional form of generative (positive) and preventative (negative) causes [bib_ref] Theory-based causal induction, Griffiths [/bib_ref] will mean that the effects of time information will be situational. So, for example, Griffiths and Tenenbaum [bib_ref] Theory-based causal induction, Griffiths [/bib_ref] showed how the effects of outcome density manipulations on causal judgments of zero contingencies were reversed by framing the same situations as involving generative or preventative causes. In experiments, but not real life, such information is provided either explicitly or implicitly by the causal scenario. In the present set of experiments, the causal scenario was the same throughout. It could also be argued that a generative causal relationship was most likely assumed because people's causal knowledge of remote controls and music switching on would be consistent with that. If that were the case, outcomes occurring 4s after the action would not be consistent with existing knowledge, and delayed outcomes should therefore eliminate or ameliorate the perception of causality before contingency itself enters the causal process. On the contrary, participants in the present study identified the preventative causal relationship evident in negative contingency condition whatever the delay between action and outcome. Thus, we argue that findings that delay effects (low BDIs) are dependent on the contingency tested and that sometimes delays enhance causal perception (High BDIs, Experiment 4) are not entirely compatible with causal model theory.
There is an alternative argument, of course, which is consistent with other evidence that temporal contiguity between action and outcome is not essential for accurate causal learning. As mentioned above, if the experimental scenario includes a plausible reason for a delay between action and outcome to occur [bib_ref] Abolishing the effect of reinforcement delay on human causal learning, Buehner [/bib_ref] , or if another stimulus is inserted into the delay period [bib_ref] Effect of Local Context of Responding on Human Judgment of Causality, Reed [/bib_ref] then delay effects can be reduced or eliminated in positive contingencies. It may be then that a negative contingency can itself act as a plausible reason for the delay. In other words, people may assume the plausibility of a preventative relationship between action and outcome, and then knowledge of the temporal structure of preventative causality is relevant. Thus action delayed outcome trials might be perceived as consistent with preventative cause. This initial 'modal decision' that a preventative relational structure exists would then allow contingency information to enter the causal process.
However, the experienced contingency would then depend on the duration of the temporal window used to determine whether two events co-occur or not [bib_ref] Contiguity and covariation in human causal inference, Buehner [/bib_ref] [bib_ref] Temporal predictability facilitates causal learning, Greville [/bib_ref]. Thus depending on how event-outcome conjunctions were reclassified in the delayed time frame, a negative contingency could be experienced as more negative and thus this would be consistent with our results; or it might be experienced as random occurrences of outcomes that are simply not linked to any action response window resulting in an experienced zero contingency. The latter outcome would be consistent with Mutter's results showing that delays eliminated the perception of negative cause [bib_ref] The role of contingency and contiguity in young and older adults' causal..., Mutter [/bib_ref]. Notwithstanding, this is currently an area of theoretical imprecision as temporal windows are not only argued to be dynamic and changing in response to incoming information [bib_ref] Temporal predictability facilitates causal learning, Greville [/bib_ref] but likely depend on the continuous or discrete trial nature of the procedure used [bib_ref] Contingency and contiguity trade-offs in causal induction, Buehner [/bib_ref]. Furthermore, it is also unclear how and under what conditions the modal switch involved in preventative and generative cause functions. It therefore could be argued that our findings are consistent with causal structure models.
Another important question, however, is whether such a conclusion would be consistent with the nature of the enhanced time effects we observed on the causal ratings of high BDI participants. Our findings support the hypothesis that slowed time perception in depression would augment the effect of increased delay or ITI. It could be argued then that the high BDI evidence points towards time dominating the causal process in these experiments in which contingency and existing causal knowledge were held constant. Despite this, for several reasons, we would argue for context as the explanatory mechanism for the effects. At baseline, high BDIs produced causal ratings that were less contextualized than other participants. However, delay and ITI effects in positive contingency conditions affected both action and context ratings in opposition, implicating a time effect through context. In these conditions, high BDI judgments were more contextualized than they had been at baseline. Then when negative contingencies were tested, delayed outcomes increased the perception of preventative cause. Taken together, these effects could be parsimoniously linked to time based fluctuations in context associations as predicted by associative theory. We might also speculate that cumulative effects of time over conditions, in relation to the difference in findings from between versus within subjects' designs, fit more readily with a context based associative learning framework than a causal model perspective.
Thus far we have discussed the findings from low and high BDI participants based on the assumption that both sets of people arrive at their causal ratings using the same causal processes but that these same processes are enhanced or impaired due to state changes in basic cognitive processing. However, we must acknowledge an alternative possibility that the two sets of participants used different processes or were at different stages of the same processes when they made their causal judgments. For example, Balleine and Dickinson [bib_ref] Goal-directed instrumental action: contingency and incentive learning and their cortical substrates, Balleine [/bib_ref] argued that instrumental action is underpinned by two different processes, goal directed action-outcome learning and more habit based stimulus response learning. Anatomically distinct from each other, goal directed action is evident early on in the process and this then transfers to more habitual behavior as learning progresses, which is stimulus driven and more independent of the outcome. Evidence from humans and animals also shows that higher levels of stress promote habit based performance over goal directed action [bib_ref] Stress Prompts Habit Behavior in Humans, Schwabe [/bib_ref] , indicating that state changes can influence the function of these processes. Along similar lines, Sternberg and McClelland [bib_ref] Two Mechanisms of Human Contingency Learning, Sternberg [/bib_ref] argue that when there is time, and presumably cognitive resources, available to them, people will make inference based causal judgments. However, with less time and cognitive resources, associative processes would be used. It is possible then that low and high BDI groups' causal ratings represent either different causal learning processes due to the availability of cognitive resources, or different stages of the learning process.
Two process theories do provide an intuitive account of the data reported here. However, one question relevant to these data remains outstanding. Previous research has suggested that under certain conditions, people with higher levels of depression are more accurate or realistic in their causal judgments than others [bib_ref] Judgement of contingency in depressed and nondepressed students: Sadder but wiser, Alloy [/bib_ref]. Realistic causal judgments, observed in some studies, must be reconciled to judgments which are also strongly affected by time and context, possibly due to slowed time perception, as in the present study. One putative reason for this is that slowed time awareness confers a normative advantage in relation to single judgments of a contingency, as these judgments would be less contextualized and more consistent with DP. However, in studies with repeated judgments, as in the current work, the effects of slowed time perception would accumulate over the course of repeated judgments, with appropriate contextualization of the individual judgments being compromised and perhaps unpredictable. Given that multi-judgment experiments bear more resemblance to the myriad of causal judgments made in the real world, although slowed time perception may confer a depressive realism advantage in some experimental settings, this is unlikely to confer similar advantage in the real world.
# Conclusions
We set out to explore the role of time and context in causal learning, with levels of depression included as a moderator variable. Findings are not entirely consistent with either causal structure models or associative theories. Neither of these models can fully explain the absence of time effects on negative contingencies unless they make additional assumptions. For example, if contingency acts as a form of prior knowledge then the question of the psychological precedence of time over contingency becomes irrelevant because both time and contingency would exert their effect through prior knowledge. However, if mildly depressed participants data are considered to be representative of enhanced time effects through slowed time perception, then findings are more consistent with an associative model. These findings are also consistent with the idea that the crucial difference in causal learning, between those scoring low and high on a depression scale, is located in contextual learning.
# Author contributions
Conceived and designed the experiments: RMM. Performed the experiments: RMM CW. Analyzed the data: RMM CW RAM. Contributed reagents/materials/analysis tools: RMM CW. Wrote the paper: RMM CW RAM.
[fig] Figure 1 262: contingency tables showing the four possible combinations of action -effect information. Note: The upper table shows generic information from which P is calculated, where A, B, C and D refer to the frequencies of action -effect conjunctions. DP = A/(A+B) -C/(C+D). The lower tables show examples of three contingency conditions with a DP of 2.5, 0 and +.5 respectively (left to right). P(O|A) refers to the conditional probability of the outcome given the presence of the action and P(O|noA) refers to the conditional probability of the outcome given the absence of the action. doi:10.1371/journal.pone.0064063.g001 [/fig]
[fig] Figure 2: Ratings of the causal strength of the action and the context. Error bars correspond to the standard error of the mean. NB: LD = low outcome density, HD = high outcome density, Neg = negative contingency, Zero = zero contingency, Pos = positive contingency. doi:10.1371/journal.pone.0064063.g002 [/fig]
[fig] Figure 3: Effects of ITI length and delay duration on ratings of positive (100/50) contingencies. Data are combined across Experiments 2a and 2b. Error bars correspond to the standard error of the mean. A = action ratings, C = context ratings. doi:10.1371/journal.pone.0064063.g003 [/fig]
[fig] Figure 4: Effects of ITI length and delay duration on ratings of negative (50/100) contingencies. Data are combined across Experiments 3a and 3b. Error bars correspond to the standard error of the mean. A = action ratings, C = context ratings. doi:10.1371/journal.pone.0064063.g004 [/fig]
[fig] Figure 5: Effects of ITI length and delay duration on ratings of negative (25/75) contingencies. Data are combined across Experiments 4a and 4b. Error bars correspond to the standard error of the mean. A = action ratings, C = context ratings. doi:10.1371/journal.pone.0064063.g005 [/fig]
[table] Table 1: Demographic characteristics of participants in Experiment 1 compared across low and high BDI groups. [/table]
[table] Table 2: Experiment 2a and Experiment 2b demographic characteristics, comparisons across experiment and mood group. [/table]
[table] Table 3: Component durations (s) and numbers of trials in each condition in Experiment 2. [/table]
[table] Table 4: Experiment 3a and Experiment 3b demographic characteristics, comparisons across experiment and mood group. [/table]
[table] Table 5: Experiment 4a and Experiment 4b demographic characteristics, comparisons across experiment and mood group. [/table]
|
Reverse-Bumpy-Ball-Type-Nanoreactor-Loaded Nylon Membranes as Peroxidase-Mimic Membrane Reactors for a Colorimetric Assay for H2O2
Herein we report for the first time fabrication of reverse bumpy ball (RBB)-type-nanoreactor-based flexible peroxidase-mimic membrane reactors (MRs). The RBB-type nanoreactors with gold nanoparticles embedded in the inner walls of carbon shells were loaded on nylon membranes through a facile filtration approach. The as-prepared flexible catalytic membrane was studied as a peroxidase-mimic MR. It was found that the obtained peroxidase-mimic MR could exhibit several advantages over natural enzymes, such as facile and good recyclability, long-term stability and easy storage. Moreover, the RBB NS-modified nylon MRs as a peroxidase mimic provide a useful colorimetric assay for H 2 O 2 .
# Introduction
Reverse bumpy ball (RBB)-type nanoreactors are rapidly attracting increasing interest [bib_ref] Hollow-Shelled Nanoreactors Endowed with High Catalytic Activity, Perez-Lorenzo [/bib_ref] [bib_ref] Enhancing the Exploitation of Functional Nanomaterials through Spatial Confinement: The Case of..., Vaz [/bib_ref]. The term RBB refers to a hollow porous sphere in which numerous nanoscale-sized catalysts remain supported on or partially embedded in the inner walls of the shell, in contrast to the yolk-shell structure [bib_ref] Hollow-Shelled Nanoreactors Endowed with High Catalytic Activity, Perez-Lorenzo [/bib_ref]. RBB-type nanoreactors have been considered to offer additional benefits with respect to the yolk-shell countparts. For instance, the RBB-type nanoreactors can provide a greater quantity of catalytically active sites per nanoreactor [bib_ref] Highly Active Nanoreactors: Nanomaterial Encapsulation Based on Confined Catalysis, Sanles-Sobrido [/bib_ref] [bib_ref] Pd nanoparticles in silica hollow spheres with mesoporous walls: a nanoreactor with..., Chen [/bib_ref]. Moreover, due to the higher contact between the catalysts and the support shell, potential synergistic effects between the catalysts and the supports may be more efficiently exploited [bib_ref] Hollow-Shelled Nanoreactors Endowed with High Catalytic Activity, Perez-Lorenzo [/bib_ref] [bib_ref] Toward Fundamentals of Confined Catalysis in Carbon Nanotubes, Xiao [/bib_ref]. Due to the distinctive features of the RBBs, synthesis of various RBBs and relative applications have been pursued in recent years. Nanoparticle catalysts (i.e., Au [bib_ref] Design of SERS-Encoded, Submicron, Hollow Particles Through Confined Growth of Encapsulated Metal..., Sanles-Sobrido [/bib_ref] [bib_ref] Polymer shell as a protective layer for the sandwiched gold nanoparticles and..., Liu [/bib_ref] , Pd [bib_ref] Pd nanoparticles in silica hollow spheres with mesoporous walls: a nanoreactor with..., Chen [/bib_ref] [bib_ref] Encapsulating Pd nanoparticles in double-shelled graphene@ carbon hollow spheres for excellent chemical..., Zhang [/bib_ref] [bib_ref] Polymer-templated synthesis of hollow Pd-CeO 2 nanocomposite spheres and their catalytic activity..., Du [/bib_ref] , Pt [bib_ref] Highly Active Nanoreactors: Nanomaterial Encapsulation Based on Confined Catalysis, Sanles-Sobrido [/bib_ref] [bib_ref] A general and feasible method for the fabrication of functional nanoparticles in..., Yin [/bib_ref] [bib_ref] Pt@CeO 2 Multicore@Shell Self-Assembled Nanospheres: Clean Synthesis, Structure Optimization, and Catalytic Applications, Wang [/bib_ref] , Mn 3 O 4 [bib_ref] Hollow silica nanosphere having functionalized interior surface with thin manganese oxide layer:..., Anisur [/bib_ref] , and Fe 3 O 4 [bib_ref] Fe/N/C hollow nanospheres by Fe(III)-dopamine complexation-assisted one-pot doping as nonprecious-metal electrocatalysts for..., Zhou [/bib_ref] have been attached to the inner walls of various types of porous spheres including silica [bib_ref] Highly Active Nanoreactors: Nanomaterial Encapsulation Based on Confined Catalysis, Sanles-Sobrido [/bib_ref] [bib_ref] Pd nanoparticles in silica hollow spheres with mesoporous walls: a nanoreactor with..., Chen [/bib_ref] [bib_ref] Hollow silica nanosphere having functionalized interior surface with thin manganese oxide layer:..., Anisur [/bib_ref] CeO 2 [bib_ref] Polymer-templated synthesis of hollow Pd-CeO 2 nanocomposite spheres and their catalytic activity..., Du [/bib_ref] [bib_ref] Pt@CeO 2 Multicore@Shell Self-Assembled Nanospheres: Clean Synthesis, Structure Optimization, and Catalytic Applications, Wang [/bib_ref] carbon [bib_ref] Encapsulating Pd nanoparticles in double-shelled graphene@ carbon hollow spheres for excellent chemical..., Zhang [/bib_ref] and polymers [bib_ref] Polymer shell as a protective layer for the sandwiched gold nanoparticles and..., Liu [/bib_ref] , achieving high catalytic activity.
However, the recyling process of RBB-type nanoreactors in liquid media is tedious and often laborious as a result of the required isolation by centrifugation/sedimentation or filtration [bib_ref] Pd nanoparticles in silica hollow spheres with mesoporous walls: a nanoreactor with..., Chen [/bib_ref] [bib_ref] Rattle-type microspheres as a support of tiny gold nanoparticles for highly efficient..., Liu [/bib_ref] [bib_ref] Synthesis of mesoporous silica hollow nanospheres with multiple gold cores and catalytic..., Chen [/bib_ref] , which has hampered the recovery and reusability of the RBB-type nanoreactors in liquid media. Compared with the dispersion of catalysts in solution, thin film-type catalysts possess more favorable properties from a practical viewpoint. For instance, switching the reaction off or on through thin film catalysts is technically easier to realize, just behaving like a "dip catalyst" [bib_ref] Based on a Silver-Nanoparticle-Embedded Polymer Thin Film, Hariprasad [/bib_ref] [bib_ref] Preparation of catalytic films of the Au nanoparticle-carbon composite tubular arrays, Gong [/bib_ref] [bib_ref] Palladium Nanoparticle-Embedded Polymer Thin Film "Dip Catalyst" for Suzuki-Miyaura Reaction, Hariprasad [/bib_ref] [bib_ref] Palladium nanoparticles supported in a polymeric membrane: an efficient phosphine-free "green" catalyst..., Faria [/bib_ref]. In addition, it is easier to separate thin film catalysts from reaction solutions, offering the feasibility and ease of multiple reuse [bib_ref] Based on a Silver-Nanoparticle-Embedded Polymer Thin Film, Hariprasad [/bib_ref] [bib_ref] Preparation of catalytic films of the Au nanoparticle-carbon composite tubular arrays, Gong [/bib_ref] [bib_ref] Palladium Nanoparticle-Embedded Polymer Thin Film "Dip Catalyst" for Suzuki-Miyaura Reaction, Hariprasad [/bib_ref] [bib_ref] Palladium nanoparticles supported in a polymeric membrane: an efficient phosphine-free "green" catalyst..., Faria [/bib_ref]. On the other hand, recently, fabrication of flexible membrane reactors (MRs) has been pursued partly because the gained flexibility can allow the construction of catalytic reactors with arbitrary geometries [bib_ref] 3D graphene/nylon rope as a skeleton for noble metal nanocatalysts for highly..., Zhang [/bib_ref] [bib_ref] Contra-diffusion synthesis of ZIF-8 films on a polymer substrate, Yao [/bib_ref] , and great attention has been paid to preparing catalytic films on porous flexible substrates. Among common flexible materials, nylon membranes are particularly attractive [bib_ref] 3D graphene/nylon rope as a skeleton for noble metal nanocatalysts for highly..., Zhang [/bib_ref] [bib_ref] Contra-diffusion synthesis of ZIF-8 films on a polymer substrate, Yao [/bib_ref] due to their unique merits such as toughness, high tensile strength, elasticity, and high resistance to acids and alkalis, as exemplified by a recent study by List and co-workers, in which nylon fabric was used as a support for preparing versatile organotextile catalysts.
In the present study, we report fabrication of RBB-type-nanoreactor-based flexible membranes. Template carbonization method is used to synthesize the RBB-type nanoreactors. Then, the RBB-type nanoreactors are loaded on flexible nylon membranes through a facile filtration approach. The as-prepared flexible catalytic membrane as the peroxidase-mimic membrane reactors is studied. Furthermore, it is tested as a H 2 O 2 sensor.illustrates the synthesis process of the RBB-type nanoreactors. In Step 1, SiO 2 nanospheres (NSs) were first modified with 3-aminopropyltrimethoxysilane to introduce amine groups on their surface, serving as the sacrificial core. Then, negatively charged Au NPs were deposited on the amino-functionalized SiO 2 NSs through the electrostatic interactions. In Step 2, the C precursor layers were coated on the surface of the SiO 2 @Au NSs by the self-polymerization of dopamine, forming the SiO 2 @Au@polydopamine sandwich configuration. In Step 3, the as-obtained product was calcined in N 2 atmosphere to carbonize the PDA shell [bib_ref] Preparation of catalytic films of the Au nanoparticle-carbon composite tubular arrays, Gong [/bib_ref] [bib_ref] Dopamine as a Carbon Source: The Controlled Synthesis of Hollow Carbon Spheres..., Liu [/bib_ref] and, finally, the SiO 2 cores were removed by 2 mol/L NaOH etching for 48 h. The morphology of the final products was characterized by transmission electron microscope (TEM). As shown in, the final products exhibited the characteristic morphology of the RBB configuration: hollow NSs with NPs embedded in the inner walls of the shell. reactors (MRs) has been pursued partly because the gained flexibility can allow the construction of catalytic reactors with arbitrary geometries [bib_ref] 3D graphene/nylon rope as a skeleton for noble metal nanocatalysts for highly..., Zhang [/bib_ref] [bib_ref] Contra-diffusion synthesis of ZIF-8 films on a polymer substrate, Yao [/bib_ref] , and great attention has been paid to preparing catalytic films on porous flexible substrates. Among common flexible materials, nylon membranes are particularly attractive [bib_ref] 3D graphene/nylon rope as a skeleton for noble metal nanocatalysts for highly..., Zhang [/bib_ref] [bib_ref] Contra-diffusion synthesis of ZIF-8 films on a polymer substrate, Yao [/bib_ref] due to their unique merits such as toughness, high tensile strength, elasticity, and high resistance to acids and alkalis, as exemplified by a recent study by List and co-workers, in which nylon fabric was used as a support for preparing versatile organotextile catalysts.
# Results and discussion
In the present study, we report fabrication of RBB-type-nanoreactor-based flexible membranes. Template carbonization method is used to synthesize the RBB-type nanoreactors. Then, the RBB-type nanoreactors are loaded on flexible nylon membranes through a facile filtration approach. The as-prepared flexible catalytic membrane as the peroxidase-mimic membrane reactors is studied. Furthermore, it is tested as a H2O2 sensor.illustrates the synthesis process of the RBB-type nanoreactors. In Step 1, SiO2 nanospheres (NSs) were first modified with 3-aminopropyltrimethoxysilane to introduce amine groups on their surface, serving as the sacrificial core. Then, negatively charged Au NPs were deposited on the amino-functionalized SiO2 NSs through the electrostatic interactions. In Step 2, the C precursor layers were coated on the surface of the SiO2@Au NSs by the self-polymerization of dopamine, forming the SiO2@Au@polydopamine sandwich configuration. In Step 3, the as-obtained product was calcined in N2 atmosphere to carbonize the PDA shell [bib_ref] Preparation of catalytic films of the Au nanoparticle-carbon composite tubular arrays, Gong [/bib_ref] [bib_ref] Dopamine as a Carbon Source: The Controlled Synthesis of Hollow Carbon Spheres..., Liu [/bib_ref] and, finally, the SiO2 cores were removed by 2 mol/L NaOH etching for 48 h. The morphology of the final products was characterized by transmission electron microscope (TEM). As shown in, the final products exhibited the characteristic morphology of the RBB configuration: hollow NSs with NPs embedded in the inner walls of the shell. [fig_ref] Figure 2: Schematic illustration of the filtration device [/fig_ref] shows a schematic illustration of the filtration-based fabrication process of the RBB-structured NSs-based catalytic flexible membrane. As illustrated, a piece of 0.20 μm pore nylon membrane was inside in the filter and when the RBB NS-containing solution in the syringe was filtered through the nylon membrane, the RBB NSs were trapped within the nylon membrane, forming the RBB NS-modified nylon membrane. [fig_ref] Figure 2: Schematic illustration of the filtration device [/fig_ref] shows a schematic illustration of the filtration-based fabrication process of the RBB-structured NSs-based catalytic flexible membrane. As illustrated, a piece of 0.20 µm pore nylon membrane was inside in the filter and when the RBB NS-containing solution in the syringe was filtered through the nylon membrane, the RBB NSs were trapped within the nylon membrane, forming the RBB NS-modified nylon membrane. reactors (MRs) has been pursued partly because the gained flexibility can allow the construction of catalytic reactors with arbitrary geometries [bib_ref] 3D graphene/nylon rope as a skeleton for noble metal nanocatalysts for highly..., Zhang [/bib_ref] [bib_ref] Contra-diffusion synthesis of ZIF-8 films on a polymer substrate, Yao [/bib_ref] , and great attention has been paid to preparing catalytic films on porous flexible substrates. Among common flexible materials, nylon membranes are particularly attractive [bib_ref] 3D graphene/nylon rope as a skeleton for noble metal nanocatalysts for highly..., Zhang [/bib_ref] [bib_ref] Contra-diffusion synthesis of ZIF-8 films on a polymer substrate, Yao [/bib_ref] due to their unique merits such as toughness, high tensile strength, elasticity, and high resistance to acids and alkalis, as exemplified by a recent study by List and co-workers, in which nylon fabric was used as a support for preparing versatile organotextile catalysts.
# Results and discussion
In the present study, we report fabrication of RBB-type-nanoreactor-based flexible membranes. Template carbonization method is used to synthesize the RBB-type nanoreactors. Then, the RBB-type nanoreactors are loaded on flexible nylon membranes through a facile filtration approach. The as-prepared flexible catalytic membrane as the peroxidase-mimic membrane reactors is studied. Furthermore, it is tested as a H2O2 sensor.illustrates the synthesis process of the RBB-type nanoreactors. In Step 1, SiO2 nanospheres (NSs) were first modified with 3-aminopropyltrimethoxysilane to introduce amine groups on their surface, serving as the sacrificial core. Then, negatively charged Au NPs were deposited on the amino-functionalized SiO2 NSs through the electrostatic interactions. In Step 2, the C precursor layers were coated on the surface of the SiO2@Au NSs by the self-polymerization of dopamine, forming the SiO2@Au@polydopamine sandwich configuration. In Step 3, the as-obtained product was calcined in N2 atmosphere to carbonize the PDA shell [bib_ref] Preparation of catalytic films of the Au nanoparticle-carbon composite tubular arrays, Gong [/bib_ref] [bib_ref] Dopamine as a Carbon Source: The Controlled Synthesis of Hollow Carbon Spheres..., Liu [/bib_ref] and, finally, the SiO2 cores were removed by 2 mol/L NaOH etching for 48 h. The morphology of the final products was characterized by transmission electron microscope (TEM). As shown in, the final products exhibited the characteristic morphology of the RBB configuration: hollow NSs with NPs embedded in the inner walls of the shell. [fig_ref] Figure 2: Schematic illustration of the filtration device [/fig_ref] shows a schematic illustration of the filtration-based fabrication process of the RBB-structured NSs-based catalytic flexible membrane. As illustrated, a piece of 0.20 μm pore nylon membrane was inside in the filter and when the RBB NS-containing solution in the syringe was filtered through the nylon membrane, the RBB NSs were trapped within the nylon membrane, forming the RBB NS-modified nylon membrane. Success in fabrication of the RBB NS-modified nylon membrane is indicated by the visual color change of the membrane from the white before filtration to the black after filtration and by the colorless filtrate. From the plane-view scanning electron microscope (SEM) images of bare nylon membrane and the RBB NS-modified nylon membrane, shown in [fig_ref] Figure 3: SEM images of the nylon [/fig_ref] ,b, respectively, it can be seen that the surface of the nylon membrane was modified with the RBB NSs. Cross-section SEM images of the RBB NS-modified nylon membrane [fig_ref] Figure 3: SEM images of the nylon [/fig_ref] also indicate that the RBB NSs were introduced into the nylon membrane. A low magnification SEM image of the RBB NS-modified nylon membrane indicating that large-area RBB NS-modified nylon membrane could be obtained is shown in [fig_ref] Figure 3: SEM images of the nylon [/fig_ref]. Furthermore, a high-magnification SEM image of the RBB NS-modified nylon membrane [fig_ref] Figure 3: SEM images of the nylon [/fig_ref] reveals that the RBB NSs should be intercepted by small pores of the nylon filter during filtration. Nylon is naturally hydrophilic and has an open pore structure, facilitating the flow of the influent through the membrane. Meanwhile, the nylon membrane is a depth filter and can retain effectively particles larger than 0.20 µm. As a result, the RBB NSs can be effectively immobilized within the nylon membrane. In addition, the resulting RBB NS-modified nylon membrane is not compact and still maintains the open pore structure of the nylon membrane. These structure features of the RBB NS-modified nylon membrane are favorable for mass transfer in catalytic application. Moreover, the photograph (inset of [fig_ref] Figure 3: SEM images of the nylon [/fig_ref] shows the RBB NS-modified nylon membrane is highly flexible, with no observed change after repeated flexion. Success in fabrication of the RBB NS-modified nylon membrane is indicated by the visual color change of the membrane from the white before filtration to the black after filtration and by the colorless filtrate. From the plane-view scanning electron microscope (SEM) images of bare nylon membrane and the RBB NS-modified nylon membrane, shown in [fig_ref] Figure 3: SEM images of the nylon [/fig_ref] ,b, respectively, it can be seen that the surface of the nylon membrane was modified with the RBB NSs. Cross-section SEM images of the RBB NS-modified nylon membrane [fig_ref] Figure 3: SEM images of the nylon [/fig_ref] also indicate that the RBB NSs were introduced into the nylon membrane. A low magnification SEM image of the RBB NS-modified nylon membrane indicating that large-area RBB NS-modified nylon membrane could be obtained is shown in [fig_ref] Figure 3: SEM images of the nylon [/fig_ref]. Furthermore, a high-magnification SEM image of the RBB NS-modified nylon membrane [fig_ref] Figure 3: SEM images of the nylon [/fig_ref] reveals that the RBB NSs should be intercepted by small pores of the nylon filter during filtration. Nylon is naturally hydrophilic and has an open pore structure, facilitating the flow of the influent through the membrane. Meanwhile, the nylon membrane is a depth filter and can retain effectively particles larger than 0.20 μm. As a result, the RBB NSs can be effectively immobilized within the nylon membrane. In addition, the resulting RBB NS-modified nylon membrane is not compact and still maintains the open pore structure of the nylon membrane. These structure features of the RBB NS-modified nylon membrane are favorable for mass transfer in catalytic application. Moreover, the photograph (inset of [fig_ref] Figure 3: SEM images of the nylon [/fig_ref] shows the RBB NS-modified nylon membrane is highly flexible, with no observed change after repeated flexion. To investigate the peroxidase-like activity of the RBB NS-modified nylon membrane, the catalytic oxidation of 3,3,5,5-tetramethylbenzidine (TMB), a benign and noncarcinogenic color reagent, in the presence of H2O2 was tested. As shown in, the RBB NS-modified nylon membrane could catalyze the oxidation of TMB in the presence of H2O2 and produce a deep blue color, with maximum absorbance at 650 nm [bib_ref] Nanocomposite hydrogels based on liquid crystalline brush-like block copolymer-Au nanorods and their..., Nguyen [/bib_ref]. A kinetic study showed that the RBB NS-modified nylon membrane exhibited its highest catalytic activity at approximately pH 3.5. In addition, also like natural enzymes, the peroxidase-mimic catalytic activity of the RBB NS-modified To investigate the peroxidase-like activity of the RBB NS-modified nylon membrane, the catalytic oxidation of 3,3,5,5-tetramethylbenzidine (TMB), a benign and noncarcinogenic color reagent, in the presence of H 2 O 2 was tested. As shown in, the RBB NS-modified nylon membrane could catalyze the oxidation of TMB in the presence of H 2 O 2 and produce a deep blue color, with maximum absorbance at 650 nm [bib_ref] Nanocomposite hydrogels based on liquid crystalline brush-like block copolymer-Au nanorods and their..., Nguyen [/bib_ref]. A kinetic study showed that the RBB NS-modified nylon membrane exhibited its highest catalytic activity at approximately pH 3.5. In addition, also like natural enzymes, the peroxidase-mimic catalytic activity of the RBB NS-modified nylon membrane was dependent on temperature, showing a maximum at approximately 40˝C. But unlike natural enzymes, the RBB NS-based peroxidase-mimic MR could exhibit facile and good recyclability and long-term stability. The recyclability of the RBB NS-based peroxidase-mimic MR was examined by recycling the same MR. Between each cycle, the membrane was directly withdrawn from the TMB-H 2 O 2 reaction solution and rinsed with deionized water. As revealed from, the MR retained almost unchanged catalytic activity towards TMB oxidation by H 2 O 2 in seven successive cycles, indicating good recyclability of the RBB NS-modified nylon MR. The peroxidase-mimic activity stability test was further investigated by testing the peroxidase-mimic membrane every day. When not in use, it was stored without any other specific care at room temperature. From(blue dotted line), it can be seen that the MR could maintain a stable catalytic activity for at least 25 d.. But unlike natural enzymes, the RBB NS-based peroxidase-mimic MR could exhibit facile and good recyclability and long-term stability. The recyclability of the RBB NS-based peroxidase-mimic MR was examined by recycling the same MR. Between each cycle, the membrane was directly withdrawn from the TMB-H2O2 reaction solution and rinsed with deionized water. As revealed from(bars), the MR retained almost unchanged catalytic activity towards TMB oxidation by H2O2 in seven successive cycles, indicating good recyclability of the RBB NS-modified nylon MR. The peroxidase-mimic activity stability test was further investigated by testing the peroxidase-mimic membrane every day. When not in use, it was stored without any other specific care at room temperature. From(blue dotted line), it can be seen that the MR could maintain a stable catalytic activity for at least 25 d. Furthermore, the catalytic activity of the RBB NS-modified nylon MR is H2O2 concentration dependent. As shown in, the absorbance of this system increased with increasing H2O2 concentration. Therefore, the RBB NS-modified nylon MR can be used as H2O2 sensor, which has potential applications in biomedical fields [bib_ref] Carbon nanotube/polysulfone composite screen-printed electrochemical enzyme biosensors, Sánchez [/bib_ref] [bib_ref] Transition Metal Hexacyanoferrates in Electrocatalysis of H 2 O 2 Reduction: An..., Sitnikova [/bib_ref]. As shown in, under the optimal conditions (i.e., 40 °C, pH 3.5), the absorbance at 652 nm was proportional to H2O2 concentration from 10-80 mmol/L with a detection limit of 0.8 mmol/L. Furthermore, the catalytic activity of the RBB NS-modified nylon MR is H 2 O 2 concentration dependent. As shown in, the absorbance of this system increased with increasing H 2 O 2 concentration. Therefore, the RBB NS-modified nylon MR can be used as H 2 O 2 sensor, which has potential applications in biomedical fields [bib_ref] Carbon nanotube/polysulfone composite screen-printed electrochemical enzyme biosensors, Sánchez [/bib_ref] [bib_ref] Transition Metal Hexacyanoferrates in Electrocatalysis of H 2 O 2 Reduction: An..., Sitnikova [/bib_ref]. As shown in, under the optimal conditions (i.e., 40˝C, pH 3.5), the absorbance at 652 nm was proportional to H 2 O 2 concentration from 10-80 mmol/L with a detection limit of 0.8 mmol/L.
# Results and discussion
## Experimental section
## Chemicals
Tetraethyl orthosilicate (TEOS), (3-aminopropyl)triethoxysilane (APTS) and hydrogen tetrachloroaurate(III) hydrate were purchased from Alfa Aesar (Ward Hill, MA, USA). Dopamine chloride and 3,5,3',5'-tetramethylbenzidine (TMB) were purchased from Sigma (St Louis, MO, USA). All other chemicals of at least analytical reagent grade were obtained from Sinopharm Chemical Reagent Co., Ltd. (Beijing, China). Nylon filters (0.20 μm) were purchased from Lanyi Chemical Reagent Co., Ltd. (Beijing, China). Aqueous solutions were prepared using deionized water produced by a Milli-Q water system (Millipore, Darmstadt, Germany).
## Synthesis of the rbb-structured nss
SiO2 NSs (ca. 200 nm in diameter) were synthesized according to the St ber method. For preparation of amino-functionalized silica NSs, 100 mg silica NSs were dispersed in toluene, and the solution was stirred vigorously after adding 25 μL APTS (0.1 μmol/L). Afterwards, the precipitates were collected by centrifugation, washed three times with ethanol, and then dried overnight under vacuum at 60 °C. Gold NPs (ca. 4.5 nm in diameter) were synthesized according to previous report [bib_ref] Wet Chemical Synthesis of High Aspect Ratio Cylindrical Gold Nanorods, Jana [/bib_ref]. To prepare the SiO2@Au composite NSs, 20 mg amino-functionalized silica NSs were added to the above gold NP solution with stirring for 30 min, followed by centrifugation and drying overnight under vacuum at 60 °C. The obtained SiO2@Au NSs were added into the freshly prepared 1 mg/mL dopamine tris buffer (10 mmol/L, pH 8.5) for polymerization of dopamine [bib_ref] Different synthesis methods allow to tune the permeability and permselectivity of dopamine-melanin..., Bernsmann [/bib_ref] [bib_ref] Electrochemical glucose biosensor with improved performance based on the use of glucose..., Lin [/bib_ref]. The mixture was stirred for 2 h, followed by centrifugation, washing with deionized water and drying in vacuum at 60 °C overnight. The obtained powders were carbonized under N2 atmosphere at 500 °C for 3 h with a heating rate of 5 °C·min −1 . Afterwards, the powders were treated with 2 mol/L NaOH solution for 48 h to remove the SiO2 core, producing the final product, the RBB-structured NSs.
## Preparation of the flexible catalytic membranes
RBB NS-containing solution (1 mL, 0.05 mg/mL) was transferred to a syringe and filtered with a nylon filter membrane. Then, the nylon filter membrane was washed with deionized water and ethanol, followed by drying at 60 °C under vacuum.
## Experimental section
## Chemicals
Tetraethyl orthosilicate (TEOS), (3-aminopropyl)triethoxysilane (APTS) and hydrogen tetrachloroaurate(III) hydrate were purchased from Alfa Aesar (Ward Hill, MA, USA). Dopamine chloride and 3,5,3',5'-tetramethylbenzidine (TMB) were purchased from Sigma (St Louis, MO, USA). All other chemicals of at least analytical reagent grade were obtained from Sinopharm Chemical Reagent Co., Ltd. (Beijing, China). Nylon filters (0.20 µm) were purchased from Lanyi Chemical Reagent Co., Ltd. (Beijing, China). Aqueous solutions were prepared using deionized water produced by a Milli-Q water system (Millipore, Darmstadt, Germany).
## Synthesis of the rbb-structured nss
SiO 2 NSs (ca. 200 nm in diameter) were synthesized according to the Stöber method. For preparation of amino-functionalized silica NSs, 100 mg silica NSs were dispersed in toluene, and the solution was stirred vigorously after adding 25 µL APTS (0.1 µmol/L). Afterwards, the precipitates were collected by centrifugation, washed three times with ethanol, and then dried overnight under vacuum at 60˝C. Gold NPs (ca. 4.5 nm in diameter) were synthesized according to previous report [bib_ref] Wet Chemical Synthesis of High Aspect Ratio Cylindrical Gold Nanorods, Jana [/bib_ref]. To prepare the SiO 2 @Au composite NSs, 20 mg amino-functionalized silica NSs were added to the above gold NP solution with stirring for 30 min, followed by centrifugation and drying overnight under vacuum at 60˝C. The obtained SiO 2 @Au NSs were added into the freshly prepared 1 mg/mL dopamine tris buffer (10 mmol/L, pH 8.5) for polymerization of dopamine [bib_ref] Different synthesis methods allow to tune the permeability and permselectivity of dopamine-melanin..., Bernsmann [/bib_ref] [bib_ref] Electrochemical glucose biosensor with improved performance based on the use of glucose..., Lin [/bib_ref]. The mixture was stirred for 2 h, followed by centrifugation, washing with deionized water and drying in vacuum at 60˝C overnight. The obtained powders were carbonized under N 2 atmosphere at 500˝C for 3 h with a heating rate of 5˝C¨min´1. Afterwards, the powders were treated with 2 mol/L NaOH solution for 48 h to remove the SiO 2 core, producing the final product, the RBB-structured NSs.
## Preparation of the flexible catalytic membranes
RBB NS-containing solution (1 mL, 0.05 mg/mL) was transferred to a syringe and filtered with a nylon filter membrane. Then, the nylon filter membrane was washed with deionized water and ethanol, followed by drying at 60˝C under vacuum.
## Instruments and measurements
The morphologies of the samples were observed using a field emission scanning electron microscopy (SEM, Supra 55, Zeiss, Oberkochen, Germany) and a field emission transmission electron microscopy (TEM, JEM-2100F, JEOL, Tokyo, Japan). For UV-vis absorption measurements, quartz microcuvettes with 10 mm path lengths and 1 mm window widths were used on a UV-vis spectrophotometer (UV-1800, Shimadzu, Tokyo, Japan). For measurement of TMB oxidation by H 2 O 2 catalyzed by the RBB NS-based peroxidase-mimic MR, the RBB NS-modified nylon membranes were immersed in NaOAc buffer (25 mmol/L, N 2 saturation, pH 3.5) containing H 2 O 2 at different concentrations and 800 µmol/L TMB. The reaction was kept at 40˝C for 10 min. UV-vis absorption spectra were recorded to monitor the time-dependent absorbance changes at 652 nm.
# Conclusions
In conclusion, we have demonstrated the fabrication of a RBB NS-based peroxidase-mimic MR. The obtained RBB-structured NSs could be firmly captured by the nylon membrane by filtration, producing flexible membranes. The obtained catalytic membrane could be used as a peroxidase-mimic MR to catalyze the oxidation of TMB by H 2 O 2 and exhibited several advantages over natural enzymes such as facile and good recyclability, long-term stability and easy storage. Moreover, the RBB NS-modified nylon MRs as a peroxidase mimic provides a colorimetric assay for H 2 O 2 with a detection limit of 0.8 mmol/L.
[fig] Figure 1: (a) Synthetic route to the Au NPs-embedded RBB-structured NSs.Step 1: Au loading; Step 2: PDA coating; Step 3: calcination and the core removal by alkaline etching; (b) TEM image of the RBB-structured NSs. Scale bar: 50 nm. [/fig]
[fig] Figure 2: Schematic illustration of the filtration device (i) and the filtration process (ii). [/fig]
[fig] Figure 3: SEM images of the nylon (a) and the RBB-structured NS-modified nylon membranes (b-d); (e) High magnification SEM image of the RBB-structured NS-modified nylon membrane. The black arrows indicate the RBB-structured NS, and the blue arrows indicate the nylon fibre network. Inset of (e): The flexibility of the RBB-structured NS-modified nylon membrane. [/fig]
[fig] Figure 4: (a) UV-vis absorption spectra of the TMB-H2O2 mixture (0.8 mmol/L TMB, 50 mmol/L H2O2) in the absence (blue line) and presence of the RBB NS-based MR (red line) after 10 min incubation; (b,c) Plots of the peroxidase-like activity of the RBB NS-based MR against pH and temperature; (d) The recyclability of the RBB NS-based MR indicated by the normalized absorbance of the TMB oxidized the catalytic oxidation of 0.8 mmol/L TMB by 10 mmol/L H2O2 in seven successive cycles with the same MR (upper) and the long-term stability of the catalytic activity of the RBB NS-based peroxidase-like MR (bottom). [/fig]
[fig] Figure 5: (a) Typical absorption spectrum of the TMB solution in the presence of H2O2 at various concentrations using the RBB NS-modified nylon MRs as a peroxidase mimic; (b) Linear calibration plot between the absorbance at 652 nm and concentration of H2O2. The insert shows the dependence of the absorbance at 652 nm on the concentration of H2O2 in the range 10 mmol/L to 200 mmol/L. [/fig]
|
Fine Mapping of Five Loci Associated with Low-Density Lipoprotein Cholesterol Detects Variants That Double the Explained Heritability
Complex trait genome-wide association studies (GWAS) provide an efficient strategy for evaluating large numbers of common variants in large numbers of individuals and for identifying trait-associated variants. Nevertheless, GWAS often leave much of the trait heritability unexplained. We hypothesized that some of this unexplained heritability might be due to common and rare variants that reside in GWAS identified loci but lack appropriate proxies in modern genotyping arrays. To assess this hypothesis, we re-examined 7 genes (APOE, APOC1, APOC2, SORT1, LDLR, APOB, and PCSK9) in 5 loci associated with low-density lipoprotein cholesterol (LDL-C) in multiple GWAS. For each gene, we first catalogued genetic variation by re-sequencing 256 Sardinian individuals with extreme LDL-C values. Next, we genotyped variants identified by us and by the 1000 Genomes Project (totaling 3,277 SNPs) in 5,524 volunteers. We found that in one locus (PCSK9) the GWAS signal could be explained by a previously described low-frequency variant and that in three loci (PCSK9, APOE, and LDLR) there were additional variants independently associated with LDL-C, including a novel and rare LDLR variant that seems specific to Sardinians. Overall, this more detailed assessment of SNP variation in these loci increased estimates of the heritability of LDL-C accounted for by these genes from 3.1% to 6.5%. All association signals and the heritability estimates were successfully confirmed in a sample of ,10,000 Finnish and Norwegian individuals. Our results thus suggest that focusing on variants accessible via GWAS can lead to clear underestimates of the trait heritability explained by a set of loci. Further, our results suggest that, as prelude to large-scale sequencing efforts, targeted re-sequencing efforts paired with large-scale genotyping will increase estimates of complex trait heritability explained by known loci.
# Introduction
In the past few years, genome-wide association studies (GWAS) have identified hundreds of genetic variants associated with quantitative traits and diseases, providing valuable information about their underlying mechanisms (for a recent example, see [bib_ref] Biological, clinical and population relevance of 95 loci for blood lipids, Teslovich [/bib_ref]. More than 2,000 common variants appear associated with over 200 conditions (as reported by the NHGRI GWAS catalog on 12/ 2010) and for a few, like age-related macular degeneration [bib_ref] Genetic variants near TIMP3 and high-density lipoprotein-associated loci influence susceptibility to age-related..., Chen [/bib_ref] and type 1 diabetes [bib_ref] Genome-wide association study and meta-analysis find that over 40 loci affect risk..., Barrett [/bib_ref] , these common variants already account for a large fraction of trait heritability. In contrast, for most complex traits and diseases, common variants identified by GWAS confer relatively small increments in risk and explain only a small proportion of trait heritability [bib_ref] Finding the missing heritability of complex diseases, Manolio [/bib_ref]. For example, for low-density lipoprotein cholesterol (LDL-C), GWAS based on up to ,100,000 individuals examined at ,2.5 million common variants [bib_ref] Biological, clinical and population relevance of 95 loci for blood lipids, Teslovich [/bib_ref] [bib_ref] Newly identified loci that influence lipid concentrations and risk of coronary artery..., Willer [/bib_ref] [bib_ref] Common variants at 30 loci contribute to polygenic dyslipidemia, Kathiresan [/bib_ref] , have identified 35 loci associated with trait variation, with some also involved in modulating the risk of cardiovascular diseases. Common variants at these loci are estimated to account for 12.2% of the variability in LDL-C levels, about one-fourth of its genetic variance [bib_ref] Biological, clinical and population relevance of 95 loci for blood lipids, Teslovich [/bib_ref]. Several hypotheses have been formulated about the nature of the remaining heritability of lipid levels and other complex traits [bib_ref] Finding the missing heritability of complex diseases, Manolio [/bib_ref] [bib_ref] Uncovering the roles of rare variants in common disease through whole-genome sequencing, Cirulli [/bib_ref] , ranging from the potential role of copy number variants to contributions from a large number of common SNPs each with very small effects. In our view, common and rare variants that are poorly represented in common genotyping arrays might account for an important fraction of trait heritability. Ignoring these variants might not only preclude identification of important trait associated loci but also compromise estimates of heritability. Thus, fine mapping appears the logical next step after GWAS. Here, we have focused on seven genes located in five of the loci associated with LDL-C in our original GWAS for blood lipid levels (APOE, APOC1, APOC2, SORT1, LDLR, APOB and PCSK9) [bib_ref] Newly identified loci that influence lipid concentrations and risk of coronary artery..., Willer [/bib_ref]. A sixth locus (corresponding to SNP rs16996148) that included a large number of genes and no obvious functional candidates was not further examined here. Together, the 5 SNPs identified in the original GWAS analyses of these 5 loci in .8,000 individuals (with follow-up genotyping of .10,000 individuals) explained only 3.1% of LDL-C variability. We set out to re-assess the contribution of these loci to trait heritability by evaluating a broader spectrum of variants. To catalog genetic variation in these regions, we first sequenced the exons and flanking regions of the seven genes in 256 unrelated Sardinians [bib_ref] Heritability of cardiovascular and personality traits in 6,148 Sardinians, Pilia [/bib_ref] , each with extremely low or high LDL-C, and in an additional 120 HapMap samples (parents from the 30 CEU and 30 YRI trios). To assess the effect of identified polymorphisms, we genotyped detected variants and additional variants selected based on an early release of the 1000 Genomes Project in a cohort of 5,524 volunteers from the SardiNIA project [bib_ref] Heritability of cardiovascular and personality traits in 6,148 Sardinians, Pilia [/bib_ref]. Our results show that at these five loci, a combination of rare and common variants, some novel and some previously identified, are associated with LDL-C, and that, taken together they double the variance explained by the common variants detected in GWAS.
# Results
To refine the contribution of five loci implicated by GWAS in the variability of LDL-C, we sequenced the exons and flanking regions of seven genes in 256 unrelated Sardinians [bib_ref] Heritability of cardiovascular and personality traits in 6,148 Sardinians, Pilia [/bib_ref] with LDL-C levels that were either extremely low (116 individuals, mean LDL-C = 70.4616.0 mg/dl) or high (140 individuals, mean LDL-C = 205.9619.6 mg/dl) (Materials and Methods), as well as an additional 120 HapMap samples (parents from the 30 CEU and 30 YRI trios). Observed heterozygosity per base pair per individual was 1.28610 23 in the selected Sardinian individuals, 1.31610 23 in the CEU and 1.99610 23 in the YRI.
Sequencing identified 782 variants, all submitted to dbSNP and now included in dbSNP releases 130 and later (for a complete list see [fig_ref] Table 1: Association Analysis results [/fig_ref]. As expected, more variants were found in the HapMap YRI samples than in the HapMap CEU or in Sardinian individuals with extreme lipid levels [fig_ref] Table 2: Heritability estimates in all study samples [/fig_ref]. Overall, we observed a 2:1 trend for enrichment of rare variants (MAF,1%) in the high LDL-C group compared to the low LDL-C group, similar to the observation by Johansen and colleagues [bib_ref] Excess of rare variants in genes identified by genome-wide association study of..., Johansen [/bib_ref] , but this enrichment was only statistically significant for APOB (P = 0.03 using an exact test). To test for LDL-C association, we used logistic regression to compare individuals in the two categories, yielding 10 variants (in APOE, APOC1, SORT1, APOB, and PCKS9) with P,0.1 [fig_ref] Table S4: Case-control association analysis results [/fig_ref]. Because of the modest number of sequenced individuals and because no signal reached significance after Bonferroni adjustment, we judged these initial association analyses -which focused only on sequenced samples and only at coding regions -inconclusive.
In addition to the loci discussed so far, our re-sequencing and genotyping effort also included B3GALT4 and B4GALT4, two loci that approached genome-wide significance in our initial GWAS analysis (each with 5610 28 ,p,5610 26 ) [bib_ref] Newly identified loci that influence lipid concentrations and risk of coronary artery..., Willer [/bib_ref]. SNPs in these loci did not reach genome-wide significance in two subsequent metaanalyses [bib_ref] Biological, clinical and population relevance of 95 loci for blood lipids, Teslovich [/bib_ref] [bib_ref] Common variants at 30 loci contribute to polygenic dyslipidemia, Kathiresan [/bib_ref] and were not significantly associated with LDL-C in the data generated here [fig_ref] Table 1: Association Analysis results [/fig_ref] and . Because we have no evidence that these two genes are associated with LDL-C, they are not discussed further. Variants identified in the two genes were also deposited in dbSNP.
To increase the power to detect association, we genotyped 5,524 individuals in the SardiNIA cohort [bib_ref] Heritability of cardiovascular and personality traits in 6,148 Sardinians, Pilia [/bib_ref] using the Metabochip (see Materials and Methods). The chip included 285 variants newly discovered by sequencing, together with an additional 2,992 derived from an early analysis of 1000 Genome Project Pilot haplotypes (considering variants 6250 Kb from each gene). Of the 3,277 SNPs that were genotyped, 1,868 passed quality control filters (see Materials and Methods and [fig_ref] Table S5: Statistics of detected genotyped and imputed SNPs for each region [/fig_ref]. To further supplement the number of variants at each locus, we carried out two rounds of genotype imputation. First, we used haplotypes for 256 sequenced SardiNIA samples to impute genotypes for SNPs that failed assay design or genotyping on the Metabochip. Second, using the haplotypes of 60 CEU samples from the 1000 Genomes Pilot, we successfully imputed an additional 5,066 variants [bib_ref] Genotype Imputation, Li [/bib_ref] (Materials and Methods and [fig_ref] Table S5: Statistics of detected genotyped and imputed SNPs for each region [/fig_ref]. After imputation, 7,488 SNPs were available for analysis, with an average minor allele frequency of 18% and an average imputation r 2 of 0.84 for 5,620 imputed SNPs (554 and 5,066 from Sanger and 1000 Genomes imputations, respectively; see [fig_ref] Table S5: Statistics of detected genotyped and imputed SNPs for each region [/fig_ref] for gene specific counts).
At three loci, SORT1, APOB and LDLR, GWAS-identified variants were very strong proxies for the best available association signal, with similar allele frequencies and r 2 .0.88 [fig_ref] Table 1: Association Analysis results [/fig_ref] , and . In those three genes, the variant showing strongest association was non-coding and not in strong linkage disequilibrium (r 2 .0.4) with any tested coding variant.
## Author summary
Despite the striking success of genome-wide association studies in identifying genetic loci associated with common complex traits and diseases, much of the heritable risk for these traits and diseases remains unexplained. A higher resolution investigation of the genome through sequencing studies is expected to clarify the sources of this missing heritability. As a preview of what we might learn in these more detailed assessments of genetic variation, we used sequencing to identify potentially interesting variants in seven genes associated with low-density lipoprotein cholesterol (LDL-C) in 256 Sardinian individuals with extreme LDL-C levels, followed by large scale genotyping in 5,524 individuals, to examine newly discovered and previously described variants. We found that a combination of common and rare variants in these loci contributes to variation in LDL-C levels, and also that the initial estimate of the heritability explained by these loci doubled. Importantly, our results include a Sardinianspecific rare variant, highlighting the need for sequencing studies in isolated populations. Our results provide insights about what extensive whole-genome sequencing efforts are likely to reveal for the understanding of the genetic architecture of complex traits.
## Apoe
The left panel shows the association results at 7 loci. For each gene, the strongest variant is listed first, and any second detected independent signal is listed with results from the conditional analysis (Materials and Methods). The column Type indicates whether the SNP was directly genotyped (Metabochip) or imputed using 1000G reference haplotype (1000G) or the Sardinian reference panel (Affy+Sanger). The right panel shows the association results for the GWAS SNPs previously described [bib_ref] Newly identified loci that influence lipid concentrations and risk of coronary artery..., Willer [/bib_ref] , the correlation with the top SNP listed in the left panel, and its p-value in the conditional analysis (Adjusted P-value).
a Effect sizes are standardized (see Materials and Methods), and represent the change in trait LDL-C values associated with each copy of the reference allele, measured in standard deviation units. b SNP rs583104 is also 1 Kb from PSRC1 transcript. c r 2 = 0.967 with Metabochip second-independent SNP, rs429358. After adjusting for the two independent SNPs, rs7412 and rs429358, the p-value for rs4420638 was 0.5. doi:10.1371/journal.pgen.1002198.t001
The most strongly associated marker at the SORT1 locus, rs583104 (p-value = 1.2610 29 ) was in high LD (r 2 = 0.77) with rs12740374 (p-value = 2.2610 28 ), an intronic SNP in the CELSR2 gene that alters the hepatic expression of the SORT1 gene by creating a C/EBP (CCAAT/enhancer binding protein) transcription factor binding site [bib_ref] From noncoding variant to phenotype via SORT1 at the 1p13 cholesterol locus, Musunuru [/bib_ref]. Both markers were genotyped, so that under the hypothesis that rs12740374 is the causal variant underlying this association signal, the modest difference in pvalues may be attributable to statistical fluctuation. At the remaining two loci, APOE and PCSK9, evidence for association peaked at low frequency (1-5%) variants not in strong linkage disequilibrium with the original GWAS signals. In both cases our analyses pointed to variants that were well studied in other contexts, but which are not included in typical GWAS panels or in the HapMap panel of European haplotypes commonly used to impute missing genotypes. Thus these variants were missed in previous GWAS analyses. In PCSK9, variant rs11591147, which leads to a non-synonymous R46L change in exon 1, was more strongly associated (P = 2.9610 215 , frequency (T) = 0.037, effect = 212.9 mg/dl; [fig_ref] Table 1: Association Analysis results [/fig_ref] than GWAS variant rs11206510, a SNP ,10 Kb upstream of the transcription start site of the gene (P = 5.7610 27 , frequency (C) = 0.24, effect = 23.7 mg/dl) . Furthermore, rs11591147 explained the GWAS association signal (association at GWAS variant rs11206510 became non-significant (P = 0.013) when non-synonymous variant R46L/rs11591147 was included as a covariate, . This coding variant has been previously implicated in the regulation of blood lipid levels, including LDL-C, and in the susceptibility to coronary and ischemic heart disease [bib_ref] Sequence variations in PCSK9, low LDL, and protection against coronary heart disease, Cohen [/bib_ref] [bib_ref] R46L, low-density lipoprotein cholesterol levels, and risk of ischemic heart disease: 3..., Benn [/bib_ref]. At the APOE gene cluster, the strongest evidence of association was observed at the missense variant (R176C, also known as R158C [bib_ref] Genotyping compared with protein phenotyping of the common apolipoprotein E polymorphism, Hansena [/bib_ref] rs7412 (P = 1.8610 231 , frequency (T) = 0.037, effect = 218.8 mg/dl) . This variant did not account for the previously reported GWAS signal; marker rs4420638 indeed remained significantly associated (P = 6.4610 210 ) after adjusting for rs7412. The missense variants at APOE and PCSK9 were not typed in the HapMap II data set, and were only recently added to genotyping arrays (Illumina 1MDuo). Thus they have not been assessed by any GWAS reported to date.
We next conditioned on the top association signal at each of the 5 loci and sought to identify additional independently associated variants. To declare statistical significance at secondary signals, we used a p-value threshold of 1610 24 ; corresponding to an adjustment for 500 independent tests across the five regions examined. At LDLR, we found an independently associated rare missense variant (rs72658864/V578A, P = 2.5610 26 in the basic model, P = 3.9610 26 in the conditional model, frequency (C) = 0.005; effect = 23.7 mg/dl) [fig_ref] Table 1: Association Analysis results [/fig_ref] and . This variant appears to be specific to Sardinia (where we identified it in our SardiNIA cohort [bib_ref] Heritability of cardiovascular and personality traits in 6,148 Sardinians, Pilia [/bib_ref] by Sanger sequencing in 3 out of 256 individuals with extreme LDL-C; by Illumina genotyping in 51 out of 5,800 randomly ascertained individuals; and by Solexa sequencing in 1 out of 505 individuals, unpublished data). It is absent in the HapMap data set, not detected in 280 Northern European individuals sequenced within the 1000 Genomes Project, and monomorphic in .10,000 Finnish [bib_ref] A genome-wide association study of type 2 diabetes in Finns detects multiple..., Scott [/bib_ref] [bib_ref] Association of 18 confirmed susceptibility loci for type 2 diabetes with indices..., Stancáková [/bib_ref] and Norwegian [bib_ref] Rapid changes in the prevalence of obesity and known diabetes in an..., Midthjell [/bib_ref] [bib_ref] The Nord-Trondelag Health Study 1995-97 (HUNT 2): Objectives, contents, methods and participation, Holmen [/bib_ref] [bib_ref] Incidence of and risk factors for type-2 diabetes in a general population, Joseph [/bib_ref] individuals genotyped with the MetaboChip (Materials and Methods, [fig_ref] Table S6: Clinical characteristics of study cohorts [/fig_ref]. Reassuringly, the variant was also observed, albeit with a lower frequency (0.00035), in TaqMan genotyping an independent sample of 5,661 Sardinians from different villages in Sardinia [bib_ref] Analysis of 12,517 inhabitants of a Sardinian geographic isolate reveals that predispositions..., Biino [/bib_ref] (Materials and Methods). The change in lipid levels associated with this rare variant (23.7 mg/dl) is 4 times greater than that observed for the strongest associated common variant at the locus (5.7 mg/dl for rs73015013). At the APOE locus, we found a strong independent signal at non-synonymous variant rs429358 (C130R, also known as C112R [bib_ref] Genotyping compared with protein phenotyping of the common apolipoprotein E polymorphism, Hansena [/bib_ref] [fig_ref] Table 1: Association Analysis results [/fig_ref] and (P = 1.2610 212 in the basic model, P = 5.8610 211 in the conditional analysis, frequency (C) = 0.071, effect = 9.3 mg/dl), which, together with rs7412, defines the three major isoforms of APOE (e2, e3 and e4) [bib_ref] Genotyping compared with protein phenotyping of the common apolipoprotein E polymorphism, Hansena [/bib_ref] [bib_ref] Human E apoprotein heterogeneity. Cysteine-arginine interchanges in the amino acid sequence of..., Weisgraber [/bib_ref]. This variant strongly correlates (r 2 = 0.96) with the originally reported GWAS signal, rs4420638 (P = 4.6610 212 , frequency (G) = 0.097, effect = 7.8 mg/dl). So, at this locus, the initial GWAS analysis picked up one independent signal (a proxy of rs429358/C130R) but missed the strongest associated variant in the region (rs7412/R176C). There was no clear evidence for residual association after accounting for the two missense variants . Interestingly, the frequency of the derived allele C at rs429358 was remarkably lower in Sardinia (freq = 0.07, see [fig_ref] Table 1: Association Analysis results [/fig_ref] than that observed in the Finnish and Norwegian individuals (see and several other European ancestry samples (freq,0.20) [bib_ref] Role of the apolipoprotein E polymorphism in determining normal plasma lipid and..., Sing [/bib_ref] [bib_ref] Sequence diversity and large-scale typing of SNPs in the human apolipoprotein E..., Nickerson [/bib_ref] [bib_ref] Contributions of 18 additional DNA sequence variations in the gene encoding apolipoprotein..., Stengård [/bib_ref] , resulting in a strikingly lower frequency of the e4 haplotype (2.5% vs. 15%) [bib_ref] Role of the apolipoprotein E polymorphism in determining normal plasma lipid and..., Sing [/bib_ref]. Finally, at PCSK9, we observed a possible independent association at SNP rs2479415, in the non-coding region flanking the transcript (P = 1.1610 27 in the basic model, P = 8610 25 in the conditional model, frequency (T) = 0.59, effect = 23.6 mg/dl) [fig_ref] Table 1: Association Analysis results [/fig_ref] and . This variant showed an independent trend also in ,10,000 Finnish and Norwegian individuals (one-sided P = 0.055 after conditioning for rs11591147).
When the 5 GWAS SNPs were replaced by the 8 variants described here (1 each for SORT1 and APOB, 2 for APOE, PCSK9 and LDLR) the variance accounted for by those loci increased from 3.1% to 6.5%. Similar estimates were also obtained with ,10,000 Finnish and Norwegian individuals, where, on average, analysis of these 8 variants increased variance explained from 3.5% to 7.1% [fig_ref] Table 2: Heritability estimates in all study samples [/fig_ref] and Materials and Methods).
# Discussion
We conducted fine mapping of five loci associated with LDL-C at an unprecedented level of resolution. In particular, we sequenced individuals with extreme phenotype levels, and subsequently genotyped variants identified by us and by the 1000 Genomes Project in a larger sample. In a final step we also imputed additional variants in the region to account for limitations of genotyping assay design. At all but one of the loci, APOB, the most strongly associated variant was directly genotyped or sequenced, suggesting that our initial selection included the crucial variants. In three loci, we found strongly associated rare or low frequency variants -which (except for a variant in LDLR, which appears to be specific to Sardinia) had been extensively Figure 1. Regional Association plots. Association results around LDLR, PCSK9 cluster and APOE. In each panel, the box at left (A, C and E) shows the association results in the main analysis; and at right (B, D and F) the results after conditioning for the strongest associated variant, highlighted with a purple dot in both plots, and its name written at the top. Arrows highlight independent signals and the most associated SNP detected in the previous GWAS [bib_ref] Newly identified loci that influence lipid concentrations and risk of coronary artery..., Willer [/bib_ref]. Each SNP is also colored according to its LD (r 2 ) in Sardinians with the top variant, with symbols that reflect genomic annotation as indicated in the legend. The rugs above indicate the position of the SNPs that were analyzed by direct typing (MetaboChip), or imputed by using haplotypes from sequenced samples (Affy+Sanger) or 1000 Genomes haplotypes (1000G). Plots were drawn using the LocusZoom standalone version [bib_ref] LocusZoom: regional visualization of genome-wide association scan results, Pruim [/bib_ref]. Genomic coordinates are given according to build 36 (hg18). doi:10.1371/journal.pgen.1002198.g001 characterized in previous non-GWAS studies. In these cases, although the associated variants had been previously described, they had not been thoroughly examined in together with GWAS associated variants at the same loci -so that the relative contributions of GWAS identified SNPs and previously described variants remained unclear.
In summary, we observed that:
(a) At SORT1 and APOB loci, association peaked at variants with similar effect size and frequency to the variants identified in GWAS; (b) At the LDLR locus, in addition to confirming the GWAS signal, a rare variant with a large effect was found. This variant is currently unique to the island of Sardinia; (c) At the APOE locus, an independently associated low frequency variant was identified. The signal was previously missed in GWAS because the variant was not included in the available genotyping chips or in the HapMap reference panels. An independently associated common variant similar in frequency and effect size to the original GWAS signal was also identified. (d) At the last locus, PCSK9, the GWAS signal could be explained by a low frequency coding variant not included in the available GWAS genotyping chips or in the HapMap reference panels. Furthermore, there was evidence for one other independently associated variant.
The strongest signals identified at APOE (both variants) and PCSK9 (the top hit) are likely to be the causal variants underlying the association signals. For SORT1, the variant exhibiting strongest association appears to be in strong linkage disequilibrium with a recently proposed functional polymorphism. In contrast, biological interpretation remains unclear for the other identified polymorphisms and requires further studies. Our results lead to several important major conclusions. First, it is striking that prior LDL-C GWAS have often missed signals due to low frequency variants (in two of the loci examined here, we identified strongly associated variants with frequency 1-5% that were missed in the original GWAS, because they were untyped or missing on imputation panels and poorly tagged by nearby SNPs). Sequencing in individuals with extreme trait values, along with large-scale imputation and genotyping, provided a better evaluation of the contribution of these loci to variation in LDL-C levels. A similar design was recently used to fine-map loci associated with fetal hemoglobin levels, a trait for which three loci can now account for about half of total variance [bib_ref] Fine-mapping at three loci known to affect fetal hemoglobin levels explains additional..., Galarneau [/bib_ref].
Second, we show that in one of the five loci we fine-mapped, a previously missed low frequency variant can account for the GWAS signal -consistent with the hypothesis that at least some GWAS signals will be due to disequilibrium with nearby low frequency or rare variants [bib_ref] Rare variants create synthetic genome-wide associations, Dickson [/bib_ref]. There is considerable debate on how frequently this scenario will occur [bib_ref] Synthetic associations are unlikely to account for many common disease genome-wide association..., Anderson [/bib_ref]. Our observations are compatible with some of the arguments made on both sides of this debate [bib_ref] Rare variants create synthetic genome-wide associations, Dickson [/bib_ref] [bib_ref] Synthetic associations are unlikely to account for many common disease genome-wide association..., Anderson [/bib_ref]. For example, in the case of PCSK9, a single low frequency variant explains the observed common variant association signal but did not appear to reduce the ability of the genomewide association study to localize the functional element of interest. Furthermore, the effect of this variant was too small to be detectable in most linkage studies (including our own linkage analysis of .35,000 relative pairs in Sardinia). Further, a single low frequency variant (and not a cluster of variants) was sufficient to explain this association signal.
Finally, our results show that if estimates are based only on the common variation assessed through GWAS, heritability at identified loci is likely to be underestimated. A more complete dissection, including common, low frequency and rare variants (some of which will be population specific), dramatically increased the proportion of heritability associated with the 5 loci examined here, from 3.1% to 6.5%. Notably, the variance explained by each locus increased when a rare variant was found as a primary or secondary hit (LDLR, APOE and PCSK9), even when the top GWAS SNP highly correlates with a strong association signal (LDLR and APOE). By contrast, only slight improvements were observed at loci where the most associated marker highly correlated with the GWAS SNPs and there was no evidence for additional independent signals, even when the GWAS variant is unlikely to be functional (SORT1 and APOB).
Genome-wide association studies have proven to be an extremely productive strategy for identifying regions of the genome associated with complex traits, often leading to unexpected insights into complex trait biology. A major efficiency of these studies is that, by focusing on a subset of variants that can be genotyped using array based platforms, they can conveniently and economically survey many common variants in large numbers of individuals. Our results emphasize the utility of these genome-wide studies in identifying trait association regions, but also emphasize that caution is needed when genome-wide study results are used to quantify the overall contribution of a locus to trait heritability. In our opinion, and consistent with our results, accurate estimates of heritability will require more extensive examination of each identified locus.
Broadly, this observation is consistent with recent simulation studies [bib_ref] Quantifying the underestimation of relative risks from genome-wide association studies, Spencer [/bib_ref] which explore, in the context of a dichotomous trait, the relationship between effect sizes observed at GWAS SNPs and at true causal variants for the same locus. These simulation studies suggest that, most of the time, effect sizes estimated from GWAS would be similar to true effect sizes but that, some of the time, effect sizes estimated from GWAS might substantially underestimate the true effect size -especially in a scenario where rare variants are more likely to be causal. In cases where the effect size was underestimated by GWAS variants, a noticeable increase in heritability ensues. It is also interesting to note that the effect sizes estimated here for rare and low frequency variants (all .10 mg/dl) are larger than the effect sizes of any of the common variants identified in GWAS studies. Effect sizes of more rare alleles associated with familial hypercholesterolemia are even larger (see [bib_ref] Mutations in PCSK9 cause autosomal dominant hypercholesterolemia, Abifadel [/bib_ref] for examples of PCSK9 variants with effects .100 mg/dl). This is consistent with the intuition that alleles with a large impact on LDL-C levels will be under strong natural selection and will, thus, be prevented from reaching high frequency in the population. Although rare and low frequency alleles with more modest impacts on LDL-C values are also likely to exist, we cannot detect them using available sample sizes and their detection must await studies of much larger sample sizes.
In conclusion, these results underline that the subsequent sequencing of the coding regions around GWAS associations in individuals with extreme values followed by large scale imputation and genotyping is an important step in assessing the contribution of associated genomic regions to trait heritability. If similar trends to those described here are observed at the remaining LDL-C associated loci, extending our approach described to all known LDL-C susceptibility loci could lead to an increase in the proportion of variance they explain from ,12% to ,24%, exceeding half of the genetic variance for this trait. Due to economic considerations, our sequencing efforts focused on the coding regions of each gene and only on genes that appeared very likely to be involved in lipid metabolism. In each locus, we augmented the set of discovered variants with variants discovered by the 1000 Genomes Project, but that will likely miss very rare as well as population specific variants. We expect that more extensive fine-mapping efforts that more comprehensively examine noncoding regions could identify additional trait associated variants. Ultimately, unbiased whole genome sequencing based association analyses might be required to fully explain the heritability of a trait like LDL-C, facilitating the comprehensive assessment of rare, population specific, and non-SNP variation. In the meantime, directed sequencing and large scale genotyping appears to be a promising approach.
# Materials and methods
# Ethics statement
All individuals studied and all analyses on their samples were done according to the Declaration of Helsinki and were approved by the local medical ethics and institutional review committees.
## Samples description
The SardiNIA project is a population based study of agingrelated traits that includes 6,148 related individuals from the Ogliastra region of Sardinia, Italy [bib_ref] Heritability of cardiovascular and personality traits in 6,148 Sardinians, Pilia [/bib_ref] [bib_ref] Genome-wide association scan shows genetic variants in the FTO gene are associated..., Scuteri [/bib_ref]. During physical examination, a blood sample was collected from each individual and divided into two aliquots, one for DNA extraction and the other to characterize several blood phenotypes, including lipids levels. Specifically, LDL-C values were derived using the [bib_ref] Association of 18 confirmed susceptibility loci for type 2 diabetes with indices..., Stancáková [/bib_ref] ; Norwegian studies are: The Nord-Trøndelag Health Study (HUNT 2) [bib_ref] Rapid changes in the prevalence of obesity and known diabetes in an..., Midthjell [/bib_ref] [bib_ref] The Nord-Trondelag Health Study 1995-97 (HUNT 2): Objectives, contents, methods and participation, Holmen [/bib_ref] and The Tromsø Study (TROMSØ) [bib_ref] Incidence of and risk factors for type-2 diabetes in a general population, Joseph [/bib_ref]. Baseline clinic characteristics of the SardiNIA, Finnish and Norwegian studies are reported in .
The independent Sardinian sample used for assessing the frequency of the rare variant at LDLR consists of 5,661 individuals belonging to 884 families enrolled in the SharDNA study [bib_ref] Analysis of 12,517 inhabitants of a Sardinian geographic isolate reveals that predispositions..., Biino [/bib_ref] , which recruited volunteers from a cluster of villages located in the Ogliastra region: Talana, Urzulei, Baunei, Triei, Seui, Seulo, Ussassai, Perdasdefogu, Escalaplano and Loceri. Observed heterozygotes were unrelated to those observed in the SardiNIA study based on demographic records to track origin of individuals up to 10 generations.
## Sequencing
Sequencing of the 256 Sardinians and the 120 HapMap samples (parents from the 30 CEU and 30 YRI trios) was carried out at the University of Washington Genome Sequencing Center through the NHLBI Resequencing & Genotyping Service (Debbie Nickerson, PI). To select the 256 individuals to be sequenced, we adjusted LDL levels by age and sex and then identified individuals in the top and bottom 5% of the distribution (individuals under lipid-lowering therapy were not considered). Among those, we selected all unrelated individuals who had at least one sibling in the study and were genotyped with 500 K or 10 K arrays [bib_ref] Genome-wide association scan shows genetic variants in the FTO gene are associated..., Scuteri [/bib_ref] , to facilitate downstream follow-up and imputation analyses.
Among the 782 variants detected by sequencing, two loss-offunction variants were observed. However, these were identified only on HapMap samples (see . A common in-frame insertion in APOB was observed in Sardinia and in HapMap CEU samples but was not associated with LDL-C after multiple testing adjustment (rs17240441, P = 3.0610 24 ; see and S1D, . The observed heterozygosity per bp/per individual was 0.00128, 0.00131 and 0.00199 in Sardinia, CEU and YRI samples, respectively. Concordance rate of HapMap II and III phases genotypes with those obtained from Sanger sequencing was 99.63%, while a lower rate (98.1%) was observed with genotypes obtained from the low-pass sequencing 1000 Genomes Project (43 CEU and 42 YRI samples were common between the two datasets), indicating the slightly lower accuracy of next-generation sequencing technologies and in particular of low-pass sequencing approaches [bib_ref] Low-coverage sequencing: Implications for design of complex trait association studies, Li [/bib_ref].
## Genotyping
Genotyping was carried out with Metabochip arrays (Illumina), which were designed in collaboration with several international consortia [bib_ref] Newly identified loci that influence lipid concentrations and risk of coronary artery..., Willer [/bib_ref] [bib_ref] Variants in MTNR1B influence fasting glucose levels, Prokopenko [/bib_ref] [bib_ref] Design of the Coronary ARtery DIsease Genome-Wide Replication And Meta-Analysis (CARDIoGRAM) Study:..., Preuss [/bib_ref] with the aim to fine map association loci detected through GWAS for a variety of traits. Part of the design included a set of wild-card SNPs chosen by individual research groups, and the SardiNIA study promoted the inclusion of all variants detected by sequencing individuals with extreme LDL-C values. In particular, assays were successfully designed for 285 of the 782 variants discovered by sequencing and 178 passed quality controls filters (some of those were polymorphic only in HapMap individuals, but we included all detected variants on the chip to assess heterozygosity on a large sample). Briefly, 3,277 variants were included on MetaboChip, and 1,868 passed quality checks. For a detailed description of markers discarded by each filter see [fig_ref] Table S9: Metabochip Genotype Quality Control Details [/fig_ref]. Concordance rate of Sanger and Metabochip genotypes was 99.47% at QCed markers, evaluated comparing genotypes of the 256 sequenced samples.
Metabochip genotyping was performed using Illumina Infinium HD Assay protocol with Multisample Beadchip format, and GenomeStudio was used for genotype calling. All samples had a call rate.98%, and there was no evidence for mis-specified family relationships (evaluated using Relpair software [bib_ref] Improved inference of relationship for pairs of individuals, Epstein [/bib_ref]. We discarded markers if any of the following was true: a) call rate ,95%, b) MAF = 0, c) Hardy-Weinberg Equilibrium P,10 26 or d) excess of Mendelian Errors [fig_ref] Table S9: Metabochip Genotype Quality Control Details [/fig_ref].
A total of 5,524 Sardinian individuals were genotyped, of which 5,382 had lipid measurements available and were not under lipid lowering therapy. In the Finnish and Norwegian studies, a total of 10,823 samples were genotyped, of which 10,027 had LDL-C measurement available and were not under lipid lowering therapy.
Genotyping of the rare LDLR variant rs72658864 on the SharDNA samples was carried out using TaqMan single SNP genotyping assays (Applied Biosystems). Given the rarity of the variant, DNA of a known heterozygote from the SardiNIA project was included in each well plate to allow detection of intensities of both alleles. The genotype of this sample was called as heterozygote in all plates.
## Imputation and statistical analyses
To better represent genomic variation, we merged genotypes from the 256 sequenced Sardinian samples with genotypes available from Affymetrix 500 K [bib_ref] Genome-wide association scan shows genetic variants in the FTO gene are associated..., Scuteri [/bib_ref] and/or Metabochip for all variants +/22 Mb spanning the gene's transcript. We then phased the haplotypes using MACH [bib_ref] Genotype Imputation, Li [/bib_ref] and used this reference set of haplotypes to impute sequence variants in the rest of the cohort [bib_ref] Family-based association tests for genome-wide association scans, Chen [/bib_ref]. We then focused on variants within +/250 Kb of the gene transcript. To further fine map the region, we used 120 haplotypes from the 60 CEU samples sequenced within the 1000 Genomes Project (June 2010 release of haplotypes based on March 2010 genotypes release) to impute variants outside the coding regions and flanking sequences targeted in our sequencing study. MACH software was used for imputation, with the same sized window used for the Sardinian-based imputation (+/22 Mb). The results obtained with these two rounds of imputation are identified in the text, as well in table and figure legends, as ''Affy+Sanger'' and ''1000G'', respectively.
For association, LDL-C levels were adjusted for age, age squared and sex, and the distribution of residuals was normalized using a quantile transformation. The association test was performed using Merlin (-fastassoc option), which uses a variance component framework to account for genetic correlation across family members [bib_ref] Family-based association tests for genome-wide association scans, Chen [/bib_ref] [bib_ref] Merlin-rapid analysis of dense genetic maps using sparse gene flow trees, Abecasis [/bib_ref].
Comparison of imputed genotypes with experimental genotypes, carried out on a set of 1,097 individuals that were genotyped with the 6.0 Affymetrix Arrays (unpublished data), showed that the average per genotype error rate between imputed and experimental genotypes was 3.7% and 4.1% for imputations based on 1000 Genomes and Sanger haplotypes, respectively.
In the Finnish and Norwegian studies we applied a similar strategy to analyze variants (rs547235 and rs562338 on APOB, rs2479415 on PCSK9 and rs429358 on APOE) that were not included on Metabochip. We defined a set of reference haplotypes of the 60 HapMap CEU founders by merging genotypes from the 1000 Genomes project and those from our Sanger sequencing, using SNPs located +/22 Mb of APOB, PCSK9 and APOE. We then used this reference panel to carry out imputation and successively used imputed dosages for testing association with LDL-C. Association analysis was performed using the same trait transformation and covariates as in the SardiNIA study. Imputation and association tests were performed separately for Finnish diabetics , Finnish non-diabetics (N = 5,678), Norwegian diabetics (N = 1,171) and Norwegian non-diabetics (N = 1,436). Results were then meta-analyzed using an inversevariance method, which combines p-values from each study using weights proportional to the variance of the beta coefficient (effect) . A combined estimate of allele frequencies was obtained using the same weights.
## Variance explained
We evaluated the variance explained by a set of markers by including all of them into the linear model in addition to the clinical covariates (age, age squared, gender), and by subtracting the variance explained by this model versus the basic model (only clinical covariates). Analyses were performed using the lmekin function in R kinship package which uses a variance component framework to account for genetic correlation across family members. In particular, since variance is not purely additive across loci, heritability in [fig_ref] Table 2: Heritability estimates in all study samples [/fig_ref] has been calculated using all 8 SNPs (or 5 SNPs) in the model rather than adding values observed at specific loci [fig_ref] Table 1: Association Analysis results [/fig_ref]. For the Finnish and Norwegian samples, the LDL-C variance explained was calculated in each study group separately, and a combined estimate was calculated by weighting each study according to its sample size [fig_ref] Table 2: Heritability estimates in all study samples [/fig_ref].
## Conditional analyses
We conducted conditional analyses to test for residual associations after accounting for a key SNP. The procedure consists of adding a SNP into the regression model as covariate and testing the effect of another SNP. Specifically, we performed this analysis by adding the strongest associated variant (key SNP) as covariate in order to test 1) whether that variant could explain the GWAS association signal; and 2) if additional independent signals were present. For the latter analysis, a threshold of P,1610 24 was used to declare significance, corresponding to a Bonferroni threshold for 500 independent tests. A graphical representation of association results from the conditional analysis is shown in , 1D, 1F and in and . Association results at B4GALT4(A) and B3GALT4 (B). Similarly to , the strongest associated variant is highlighted with a purple dot, and its name written nearby. Arrows highlight independent signals and the most associated SNP detected in our original GWAS. Each SNP is also colored according with its LD (r 2 ) in Sardinians with the top variants, with symbols that reflect genomic annotation, as in . The rugs on top indicate the position of the SNPs that were analyzed by direct typing (MetaboChip), imputed by using haplotypes from sequenced samples (Affy+Sanger) or imputed by using 1000 Genomes haplotypes (1000G). (TIF) Association results at SORT1 and APOB. Similar to , the box at left (A and C) shows the association results in the main analysis; and at right (B and D) show the results after conditioning for the strongest associated variant, highlighted with a purple dot in both plots, and its name written at the top. Arrows highlight the independent signal (if any) and the most associated SNP detected in our original GWAS. Each SNP is also colored according with its LD (r 2 ) in Sardinians with the top variant (used as covariate), with symbols that reflect genomic annotation, as indicated in the legend in panel A. The rugs on top indicate the position of the SNPs that were analyzed by direct typing (MetaboChip), imputed by using haplotypes from sequenced samples (Affy+Sanger) or imputed by using 1000 Genomes haplotypes (1000G). (TIF) Association results at APOE after adjusting for the two missense variants. The figure shows the association results at APOE after adjusting for the two independent signals rs7412 (indicated with a purple dot) and rs429358 indicated with an arrow. Each SNP is also colored according with its LD (r 2 ) in Sardinians with the top variant (rs7412) in the main analysis, with symbols that reflect genomic annotation, as indicated in the legend. The rugs on top indicate the position of the SNPs that were analyzed by direct typing (MetaboChip), imputed by using haplotypes from sequenced samples (Affy+Sanger) or imputed by using 1000 Genomes haplotypes (1000G). (TIF)
## Urls
## Supporting information
[table] Table 1: Association Analysis results. [/table]
[table] Table 2: Heritability estimates in all study samples. [/table]
[table] Table S1: List of all variants detected by sequencing in the 256 Sardinians and 120 HapMap Samples and relative counts (given on a separate excel file). The table lists the variants detected by Sanger sequencing, their genomic position in build 36, the corresponding alleles, the biological function and the observed frequency in Sardinians, CEU and YRI samples. (XLS) [/table]
[table] Table S2: Summary of variants detected by sequencing in the 256 Sardinians and 120 HapMap Samples. The table summarizes the variants detected by sequencing in different types of biological function. (DOCX)Table S3 Enrichment of rare variants (MAF,0.01). The table lists the number of carriers of coding mutations (MAF,0.01) for each gene in individuals with high or low LDL-C levels. Shaded rows indicate whether a trend for enrichment is observed, although significance was clear only at APOB. (DOCX) [/table]
[table] Table S4: Case-control association analysis results. Association signals showing a p-value,0.1 when comparing individuals with high and low LDL values. (DOCX) [/table]
[table] Table S5: Statistics of detected genotyped and imputed SNPs for each region (+/2250 Kb from gene's transcript). The table summarizes the variants detected and analyzed in each step (sequencing, genotyping, imputation) for each gene. (DOCX) [/table]
[table] Table S6: Clinical characteristics of study cohorts. The table describes the clinical characteristics for the Sardinians, Finnish and Norwegian populations used for association analyses. (DOCX) Table S7 Association results in the Finnish and Norwegian individuals. The table describes the association results for all SNPs in the diabetics and non-diabetics Finnish and Norwegian samples. (DOCX) [/table]
[table] Table S9: Metabochip Genotype Quality Control Details. Statistics of quality controls filters. Note that a marker could have failed more than one check. (DOCX) [/table]
|
Medical Therapy for Heart Failure with Preserved Ejection Fraction
Heart failure with preserved ejection fraction (HFpEF) is a challenging disease state that has long been plagued by heterogeneity in diagnostic criteria and underlying etiologies. Due in part to the complexity of defining this disease and the simplistic approach of only studying medications that have shown significant improvement in heart failure with reduced ejection fraction, there have been a multitude of negative trials in this population. In the past few years, however, there have been medications that have finally shown to benefit patients with HFpEF. In particular, the blockbuster class of medications called SGLT2 inhibitors have provided a treatment option that improves outcomes in this group of patients. There is increasing focus on HFpEF research that aims to improve the phenotyping of these patients to more successfully tailor therapy and improve patient outcomes.
# Introduction
Heart failure with preserved ejection fraction is an elusive disease that has long been challenged by the lack of a unified definition and proven therapies. The original term "congestive heart failure" broadly encompassed clinical symptoms such as dyspnea and pulmonary edema but did not distinguish underlying ejection fraction (EF) and cardiac structural changes. [bib_ref] Heart Failure With Preserved Ejection Fraction In Perspective, Pfeffer [/bib_ref] Approximately 50% of patients with heart failure have a preserved ejection fraction.Diastolic heart failure emerged in the 1980s as a term to describe patients with normal ejection fraction but impaired left ventricular (LV) relaxation. [bib_ref] Left ventricular diastolic dysfunction and diastolic heart failure, Gaasch [/bib_ref] Initially, clinicians struggled with defining and accurately diagnosing diastolic dysfunction due to the lack of standardized echocardiographic assessments. Many argued that imaging could not adequately assess for diastolic dysfunction and believed invasive hemodynamics were needed, but the latter required resources, logistics, and standardization, which were significant barriers. [bib_ref] Left ventricular diastolic dysfunction and diastolic heart failure, Gaasch [/bib_ref] [bib_ref] Defining diastolic heart failure: a call for standardized diagnostic criteria, Vasan [/bib_ref]
## Diagnosis
In recent years, diastolic heart failure has become known as heart failure with preserved ejection fraction (HFpEF). While the name has evolved to include a clear definition and diagnostic criteria, pathophysiologic understanding and treatment strategy has remained stagnant compared to progress in heart failure with reduced ejection fraction (HFrEF). Experts have identified the following as major gaps in the understanding of HFpEF: definition, subtypes, and end points in clinical trials. [bib_ref] Heart Failure With Preserved Ejection Fraction Expert Panel Report: Current Controversies and..., Parikh [/bib_ref] The definition has evolved from the previous definition of diastolic dysfunction in the context of an EF > 50% and is now more broadly defined as structural abnormalities resulting from high filling pressures, diastolic abnormalities, elevated biomarkers, and elevated left heart filling pressures by noninvasive and invasive hemodynamic assessment. [bib_ref] Heart Failure With Preserved Ejection Fraction Expert Panel Report: Current Controversies and..., Parikh [/bib_ref] This diagnosis can be made after excluding noncardiac causes of dyspnea, which can mimic heart failure.
One of the challenging aspects of the treatment of HFpEF is that unlike HFrEF, there is no clearly defined neurohormonal pathway that underlies the pathophysiology. Additionally, studies of HFpEF are challenged by pooling patients into a single "HFpEF" disease, when in reality there are varying etiologies. [bib_ref] Heart Failure With Preserved Ejection Fraction Expert Panel Report: Current Controversies and..., Parikh [/bib_ref] There are also many overlapping disease states that mimic or coexist with HFpEF, including pulmonary hypertension and infiltrative diseases such as sarcoidosis or amyloidosis. Amyloid, for example, is a disease that presents with overlapping symptoms of HFpEF but has vastly different pathophysiology, comorbidities, treatment options, and outcomes. Hahn and colleagues evaluated endomyocardial biopsies of 108 patients with HFpEF and found that 14% of their samples were positive for cardiac amyloidosis. [bib_ref] Endomyocardial Biopsy Characterization of Heart Failure With Preserved Ejection Fraction and Prevalence..., Hahn [/bib_ref] While these patients technically meet the diagnostic criteria of HFpEF, their treatment strategy and outcomes will differ greatly. In fact, it has been speculated that the unintentional inclusion of amyloidosis patients in major HFpEF trials may have contributed to many of these trials failing to meet statistical significance. [bib_ref] The Impact of Patients With Cardiac Amyloidosis in HFpEF Trials, Oghina [/bib_ref] Another disease state with marked overlap is pulmonary hypertension (PH). The most common form is pulmonary hypertension due to left heart disease, which is characterized by elevated pulmonary pressures due to chronic left atrial pressure overload, leading to increased pressure in the pulmonary veins (called "post-capillary" PH). In some cases, chronic pressure overload can lead to irreversible remodeling of the pulmonary arteries and an increase in pulmonary vascular resistance, which leads to combined pre-and post-capillary PH. Invasive hemodynamics should be measured when there is a high likelihood of PH and there are implications on management and prognostication; however, there is debate regarding which variable is associated with the strongest correlation with clinical outcomes. [bib_ref] 2022 ESC/ERS guidelines for the diagnosis and treatment of pulmonary hypertension, Humbert [/bib_ref] The most recent 2022 guidelines are from the European Society of Cardiology, and the only class of medication with potential benefit in HFpEFassociated PH are phosphodiesterase inhibitors (PDE5i). PDE5i have been shown to improve exercise capacity, quality of life, and hemodynamics in small studies of patients with severe precapillary components to their PH and with pulmonary vascular resistance values ≥ 5 Wood units; however, there is insufficient evidence to provide a recommendation on their use. [bib_ref] 2022 ESC/ERS guidelines for the diagnosis and treatment of pulmonary hypertension, Humbert [/bib_ref] The guidelines give a class III recommendation on the use of PDE5i in post-capillary PH.
In addition to distinguishing symptoms among various cardiac etiologies, it is important to correctly differentiate symptoms that might be attributed to HFpEF from noncardiac causes. Two recently developed scoring systems integrate variables to further differentiate HFpEF from noncardiac causes of dyspnea and can aid in determining when further diagnostic testing is needed.The 2022 American College of Cardiology/American Heart Association/Heart Failure Society of America Heart Failure guidelines have incorporated the HF 2 PEF score that includes six variables with a total possible score of nine points .The odds of HFpEF doubled for each 1-unit score increase (OR 1.98 [1.74-2.3]; P < .001) with an area under the curve of 0.841 (P < .001).This tool can be helpful because the overlap in diagnosis is large, but the therapeutic approach may differ significantly. It is critically important to evaluate all the components included in the score to be able to more accurately diagnose HFpEF.
## Treatment
## Renin angiotensin aldosterone system inhibitors
Finding treatment strategies to improve outcomes of HFpEF has been challenging. Initial studies for HF medical therapy did not use ejection fraction (EF) as part of the inclusion criteria. [bib_ref] Heart Failure With Preserved Ejection Fraction In Perspective, Pfeffer [/bib_ref] The first studies that assessed all-cause mortality as the primary outcome in heart failure were the Cooperative North Scandinavian Enalapril Survival Study (CONSENSUS) and Veterans Administration Cooperative Study (V-HeFT). These studies required clinical criteria consistent with heart failure, such as dyspnea and fluid retention, but measurements of myocardial function were not required. 12,13 A post-hoc analysis of the V-HeFT trial demonstrated that patients with a normal EF (> 45%) had a lower rate of death (8% vs 19%), leading subsequent trials to use EF of 35% to 40% as a cutoff to improve sample size and event rates. While this led to significant advancement in HFrEF research, it hindered studies for HFpEF. [bib_ref] Heart Failure With Preserved Ejection Fraction In Perspective, Pfeffer [/bib_ref] [bib_ref] Heart failure with normal ejection fraction. The V-HeFT Study. Veterans Administration Cooperative..., Cohn [/bib_ref] Since this time, various cutoffs (40% or 50%) have been used for HFpEF trials. There is significant variability in reported EFs read by various providers, and the cutoff values used in various studies are somewhat arbitrary, impacting diagnosis and inclusion in studies. Although EF is perhaps the most common criteria for inclusion of HFpEF trials, it is also not specific to the underlying etiology and is associated with large heterogeneity in patient phenotypes. This has likely hindered the progress in identifying beneficial therapies in specific patients.
Most clinical studies evaluating medical treatment for HFpEF have simply followed positive studies with the same drugs studied in HFrEF patient populations. Beta blockers and renin angiotensin aldosterone system (RAAS) inhibitors both reduced mortality in patients with HFrEF, and the Heart Outcomes Prevention Evaluation (HOPE) trial demonstrated that patients at high risk of cardiovascular events but without heart failure have a survival benefit from angiotensin converting enzyme (ACE) inhibitors. [bib_ref] Heart Outcomes Prevention Evaluation Study Investigators. Effects of an angiotensin-converting-enzyme inhibitor, ramipril,..., Yusuf [/bib_ref] The use of these medications also improves outcomes in patients with diabetes, coronary artery disease, and atrial fibrillation, all of which are common comorbidities in the HFpEF population. [bib_ref] Heart Failure With Preserved Ejection Fraction In Perspective, Pfeffer [/bib_ref] [bib_ref] Heart Failure With Preserved Ejection Fraction Expert Panel Report: Current Controversies and..., Parikh [/bib_ref] Due to the broad spectrum of patients who had observed benefits from RAAS inhibition, it was hypothesized that patients with heart failure symptoms but without a reduced left ventricular ejection fraction (LVEF) would also benefit from RAAS inhibitors.
Two large trials assessed angiotensin receptor blockers (ARB) in HFpEF. [bib_ref] Effects of candesartan in patients with chronic heart failure and preserved left-ventricular..., Yusuf [/bib_ref] The effects of candesartan in patients with chronic heart failure and preserved LVEF (CHARM-Preserved) trial compared candesartan to placebo, and the irbesartan in patients with heart failure and preserved ejection fraction (I-Preserve) trial compared irbesartan to placebo. The two studies both included patients with class II-IV symptoms Figure 1 H 2 FPEF score: a simple, evidence-based approach to help in diagnosis of heart failure with preserved ejection fraction.and a recent hospitalization for heart failure, but CHARM-Preserved used an EF cutoff of ≥ 40% whereas I-Preserve only included those with an EF ≥ 45%. In CHARM-Preserved, the primary outcome (cardiovascular death or unplanned hospital readmission due to HF) was met in 22% of the candesartan patients and 24% of placebo patients, with no statistically significant difference between them (HR 0.89 [0.77-1.03]; P = .12). Mortality was not different between patients treated with candesartan compared with placebo (11.2% vs 11.3%), but hospitalizations for heart failure trended toward significance with candesartan (15.9% vs 18.3%, HR 0.85 [0.72-1.01]; P = .07), which was consistent with a prespecified adjusted analysis. In the I-PRESERVE trial, there was no statistically significant difference in the primary outcome (death from cardiovascular causes or hospital readmission for cardiovascular causes) or any secondary outcomes. The reduction in rehospitalizations seen with candesartan was not reproduced with irbesartan (HR 0.95 [0.85-1.08]; P = .44).
A secondary analysis was done assessing the effects on various EF ranges, including data from CHARM-Preserved, CHARM-Added (effects of candesartan in patients with chronic heart failure and reduced left-ventricular systolic function taking angiotensin-converting-enzyme inhibitors) and CHARM-Alternative (effects of candesartan in patients with chronic heart failure and reduced left-ventricular systolic function intolerant to angiotensin-convertingenzyme inhibitors) in EF ≤ 40%. [bib_ref] Heart failure with midrange ejection fraction in CHARM: characteristics, outcomes and effect..., Lund [/bib_ref] Patients with an EF of 40% to 49% showed a significant reduction in the combined outcome of cardiovascular mortality and HF hospitalization (HR 0.72 [0.55-0.95]; P = .02) but there was no benefit seen in EF > 50%, suggesting that the patients with a moderately reduced EF may have driven the improvement in rehospitalizations. If a more conservative EF cutoff of 50% had been used, it is possible that no benefit would have been observed with candesartan. A Cochrane review done for four ARB trials in HFpEF found no significant reduction in mortality or hospitalizations (HR 0.92 [0.83-1.02]). Although the data is conflicting and modest, ARBs have been given a Class 2b recommendation in the 2022 guidelines for reductions in hospitalizations, with the caveat that benefits are largely seen in patients with EF on the lower end.Due to the benefits seen in many other populations, ACE inhibitors were also tested in HFpEF patients. The perindopril in elderly people with chronic heart failure (PEP-CHF) trial assessed perindopril in elderly people > 70 years of age with diastolic heart failure, EF > 40%, and a hospitalization within the previous 6 months. [bib_ref] The perindopril in elderly people with chronic heart failure (PEP-CHF) study, Cleland [/bib_ref] There was no difference in the primary outcome of time to death or unplanned HF hospitalization (HR 0.92 [0.70-1.21]; P = .545) or either of the individual components. Overall, the study had a lower event rate than anticipated and a large discontinuation rate, but ACE inhibitors do not appear to possess the same beneficial effects in HFpEF patients as a whole despite their benefits in patients with overlapping disease states.
One of the most controversial but impactful trials in HFpEF is the TOPCAT (treatment of preserved cardiac function heart failure with an aldosterone antagonist) trial. [bib_ref] Spironolactone for heart failure with preserved ejection fraction, Pitt [/bib_ref] This study compared spironolactone to placebo in patients with signs and symptoms of heart failure, a LVEF ≥ 45%, and either a prior HF hospitalization or elevated natriuretic peptide level. The primary outcome was cardiovascular death, HF hospitalization, or aborted cardiac arrest. It was conducted primarily in North and South America, Russia, and Georgia. The patients included were 68 years old on average, 51% female, 63% with class II symptoms, and 71% had been hospitalized within the previous 6 months. Overall, there was no statistically significant difference in the primary outcome of death from cardiovascular causes, aborted cardiac arrest, or hospitalization for the management of heart failure with 18.6% in the spironolactone group and 20.4% in placebo (HR 0.89 [0.77-1.04]; P = .14). Death from cardiovascular causes (9.3% vs 10.2%) and aborted cardiac arrest (0.2% and 0.3%) were not different between the two groups, but spironolactone reduced hospitalizations (12% vs 14.2%; HR 0.83 [0.69-0.99]; P = .04). The surprising result from the trial was a marked regional difference in event rates.
A subgroup analysis compared the outcomes for patients in the Americas to patients in Russia and Georgia.In the Americas population, the primary outcome occurred in 29.5% of patients versus 8.9% in the Russia/Georgia patients; both rehospitalizations and mortality were significantly improved in the Americas population but not the Russia/ Georgia group (P < .001). In the Americas, spironolactone was associated with a significant benefit (HR 0.83) but this was not the case in the Russia/Georgia arm (HR 1.10). In the overall study, there was a higher incidence of hyperkalemia and elevations in serum creatinine with spironolactone, but this was not observed in the Russia and Georgia population even though compliance was reported to be higher, thus raising concerns for actual compliance. Hence there is concern that the Russian/Georgian arm of the study was conducted poorly and that the data from this region may have negatively impacted the entire study. Due to the improvements seen in mortality and hospitalizations in the Americas, mineralocorticoid receptor antagonists are given a Class 2b recommendation in the 2022 guidelines for the reduction in hospitalizations.The angiotensin receptor and neprilysin inhibitor sacubitril-valsartan, found to have a survival benefit in patients with HFrEF compared to enalapril, was subsequently studied in HFpEF patients. In the phase 2 study, sacubitril-valsartan was shown to significantly reduce N-terminal (NT)-pro hormone brain natriuretic peptide (NT-proBNP) compared with valsartan. [bib_ref] The angiotensin receptor neprilysin inhibitor LCZ696 in heart failure with preserved ejection..., Solomon [/bib_ref] The outcomes trial PARAGON-HF (prospective comparison of ARNI with ARB global outcomes in HF with preserved ejection fraction) enrolled patients with EF ≥ 45% and either elevated natriuretic peptide levels or a recent hospitalization and compared sacubitril-valsartan to valsartan after a run-in period. [bib_ref] Angiotensin-Neprilysin Inhibition in Heart Failure with Preserved Ejection Fraction, Solomon [/bib_ref] The primary outcome of HF hospitalizations and death was not statistically significant (rate ratio 0.87 [0.75-1.01]; P = .06). The rates of hospitalizations were lower with sacubitril-valsartan but not statistically significant (RR 0.85 [0.72-1.00]), but these results are exploratory as the primary outcome was not met. In the prespecified subgroup analysis, patients with EF ≤ 57% significantly benefited from sacubitril-valsartan (RR 0.78 [0.64-0.95]) while those > 57% did not. Overall, sacubitril-valsartan was associated with a higher incidence of hypotension and angioedema but lower rates of hyperkalemia and increases in serum creatinine. Sacubitril-valsartan is also given a Class 2b recommendation in the 2022 guidelines to reduce hospitalizations, again emphasizing the benefits observed in patients with a lower EF. 2
## Loop diuretics
Elevated filling pressures are one of the cornerstone symptoms in HFpEF. Although diuretics are given to reduce congestion, there is limited data to guide their use overall and no specific studies in HFpEF. The DOSE (diuretic optimization strategies evaluation) trial is the largest diuretic trial, and while there was no EF cutoff used in the trial, the average EF of those enrolled was 35%.This trial was a 2 × 2 factorial design assessing high-dose versus low-dose diuretics and continuous versus bolus dosing of furosemide. The primary outcome was patient-perceived dyspnea using a visual analog scale. There was no difference in the primary outcome with bolus versus continuous infusion (P = .47) but high dose was associated with a nonsignificant improvement in symptoms (P = .06) compared with low dose. There were no differences in secondary outcomes with bolus versus continuous infusion, but high dose was associated with an improvement in net fluid loss at 72 hours (P = .001) and change in weight at 72 hours (P = .01). From this trial, we learned that patients admitted with volume overload benefit from at least 2.5-times their home-dose equivalent in intravenous diuretics. Alternative strategies for elevated filling pressures include implantable systems for chronic monitoring of intracardiac and pulmonary artery pressures such as CardioMEMS™ heart sensor, which looked at outcomes in New York Heart Association (NYHA) class III heart failure patients in the CHAMPION trial. [bib_ref] Wireless pulmonary artery haemodynamic monitoring in chronic heart failure: a randomised controlled..., Abraham [/bib_ref] While this is beyond the scope of this paper, this strategy is associated with a significant reduction in HF hospitalizations in patients with HFpEF and HFrEF and also has shown benefits recently in NYHA class II patients. 26
## Sodium-glucose cotransporter-2 inhibitors
Trailing the multiple trials for pharmacotherapies that have failed to demonstrate effectiveness in improving outcomes in HFpEF, sodium-glucose cotransporter-2 (SGLT2) inhibitors have forever changed this paradigm. After the EMPEROR-REDUCED trial (empagliflozin outcome trial in patients with chronic heart failure and a reduced ejection fraction) demonstrated the incremental improvement in outcomes for HFrEF patients when added to guidelinedirected medical therapy, the EMPEROR-PRESERVED trial (empagliflozin outcome trial in patients with chronic heart failure with preserved ejection fraction) proved something similar for HFpEF patients. In this randomized double-blind placebo-controlled trial, patients with EF > 40% with a clinical diagnosis of chronic heart failure based on elevated NT-proBNP (≥ 900 pg/mL or ≥ 300 pg/mL for patients with or without atrial fibrillation, respectively) were enrolled and randomized to receive either 10 mg empagliflozin or placebo. [bib_ref] Empagliflozin in Heart Failure with a Preserved Ejection Fraction, Anker [/bib_ref] This trial succeeded in demonstrating a statistically significant 21% reduction in primary composite outcomes including heart failure outcomes or cardiovascular death (P < .001). In fact, this trial demonstrated renalprotective benefits with significantly lower progression of chronic kidney disease (average change in estimated glomerular filtration rate per year). While this trial failed to demonstrate improvement in functional status, other trials since then have demonstrated improvement in functional status. [bib_ref] The SGLT2 inhibitor dapagliflozin in heart failure with preserved ejection fraction: a..., Nassif [/bib_ref] More recently, the DELIVER-HF trial (dapagliflozin evaluation to improve the lives of patients with preserved ejection fraction heart failure) has demonstrated treatment with dapagliflozin 10 mg daily in primary outcomes (worsening HF event or cardiovascular death) in patients with EF > 40% and with and without recent HF hospitalization. [bib_ref] Dapagliflozin in Patients Recently Hospitalized With Heart Failure and Mildly Reduced or..., Cunningham [/bib_ref] This latter trial, presented at the latest European Society of Cardiology meeting in 2022, further confirmed the survival benefits in patients treated with SGLT2 inhibitors. Proposed mechanisms of action for SGL2 inhibitors are, in fact, not related to the RAAS inhibition as is traditionally thought in HFrEF pharmacotherapies. While many have hypothesized that the glucosuric effects may augment diuresis, thus improving decongestion, animal models have shown direct impact of SGLT2 inhibition in cardiometabolics and cardiac energetics (see [fig_ref] Figure 2: Effects of SGLT2 inhibition at the myocardial level [/fig_ref]. [fig_ref] Figure 3: American College of Cardiology/American Heart Association 2022 Guidelines for the Management of... [/fig_ref] summarizes the 2022 guideline recommendations on treatment of HFpEF. 2 4 outlines some practical tips in initiating and monitoring the effects of SGLT2 inhibitors. Absolute contraindications for initiation of SGLT2 inhibitors include patients with type 1 diabetes and history of diabetic ketoacidosis given their elevated risks of developing ketoacidosis. Relative contraindications include hypotension (up to 6.6% in EMPEROR-PRESERVED) and acute renal failure (up to 12.1% in EMPEROR-PRESERVED). Despite concerns regarding worsening renal function in patients with chronic kidney disease (CKD), a recent study demonstrated delay in progression of CKD in this population. [bib_ref] DAPA-CKD Trial Committees and Investigators. Dapagliflozin in Patients with Chronic Kidney Disease, Heerspink [/bib_ref] Patients with CKD may experience up to a 10% to 15% decline in creatinine clearance; however, more than a 20% decline should raise concerns for acute renal failure, and SGLT-2 inhibitors should be stopped.Lastly, SGLT2 inhibitors may increase the risk of genitourinary infections. If this occurs, the SGLT2 inhibitor should be stopped and the infection treated with an antibiotic or antifungal. Clinicians should consider reinitiating the SGLT2 inhibitor after the infection resolves based on patient-centered decision-making.
# Conclusion
While HFpEF patients have not had the robust armamentarium that HFrEF patients have, evidence-based options are finally available to support the management of these complex patients. The newest blockbuster class of SGLT2 inhibitors are proving to have dramatic effects across a wide array of cardiovascular patients, and for the first time we are seeing signs of a pharmacotherapeutic benefit in HFpEF patients. The use of the newly developed scoring Initiation and up-titration of empagliflozin in patients with heart failure with preserved ejection fraction. EF: ejection fraction; NYHA: New York Heart Association; SGLT2-i: sodium-glucose transporter-2 inhibitor; CrCl: creatinine clearance; BMP: basic metabolic panel.
systems will be valuable in not only diagnosing HFpEF but also improving the homogeneity of populations enrolled in our clinical trials. Using EF cutoffs with evidence of diastolic dysfunction is a nonspecific inclusion criterion and could have led to missed opportunities in accurately testing the effects of medications, such as spironolactone and sacubitril-valsartan. As we continue to improve the phenotyping of HF and our understanding of the diversity of phenotypes within HFpEF, the future holds promise for more individualized and targeted therapies based on the true molecular pathophysiology of diastolic dysfunction in select populations.
## Key points
- The HF 2 PEF score is a validated tool that can be used to evaluate cardiac from noncardiac causes of dyspnea and aid in the diagnosis of heart failure with preserved ejection fraction (HFpEF). - Angiotensin receptor blockers, aldosterone antagonists, and angiotensin receptor-neprilysin inhibitors (sacubitril/valsartan) can be considered in HFpEF, especially in patients with ejection fractions closer to 50%. - The newest class of medications, sodium-glucose cotransporter-2 inhibitors, are the first to show improved outcomes in HFpEF, with a reduction in heart failure hospitalizations and mortality, and should be utilized in all HFpEF patients unless a contraindication exists. - Further effort to improve diagnosis and identify specific phenotypes of HFpEF are needed to tailor trials and therapy since significant heterogeneity exists across the HFpEF spectrum.
[fig] Figure 2: Effects of SGLT2 inhibition at the myocardial level.30 CaMKII: calcium/calmodulin-dependent protein kinase II; FAO: fatty acid oxidation; Na/H: sodium/hydrogen; NLRP3: NLR family pyrin domain-containing 3; OXPHOS: oxidative phosphorylation; ROS: reactive oxygen species; and SGLT2i: sodium-glucose cotransporter-2 inhibition * Signifies myocardial processes. [/fig]
[fig] Figure 3: American College of Cardiology/American Heart Association 2022 Guidelines for the Management of Heart Failure. 2 HFpEF: heart failure with preserved ejection fraction; LVEF: left ventricular ejection fraction; SGLT2: sodium-glucose cotransporter-2; ARNI: angiotensin receptor/neprilysin inhibitor; MRA: magnetic resonance imaging; ARB: angiotensin receptor blocker * greater benefit in patients with LVEF closer to 50%. [/fig]
[fig] Figure 4: outlines some practical tips in initiating and monitoring the effects of SGLT2 inhibitors. Absolute contraindications for initiation of SGLT2 inhibitors include patients with type 1 diabetes and history of diabetic ketoacidosis given their elevated risks of developing ketoacidosis. Relative contraindications include hypotension (up to 6.6% in EMPEROR-PRESERVED) and acute renal failure (up to 12.1% in EMPEROR-PRESERVED). Despite concerns regarding worsening renal function in patients with chronic kidney disease (CKD), a recent study demonstrated delay in progression of CKD in this population. 31 Patients with CKD may experience up to a 10% to 15% decline in creatinine clearance; however, more than a 20% decline should raise concerns for acute renal failure, and SGLT-2 inhibitors should be stopped. 32 Lastly, SGLT2 inhibitors may increase the risk of genitourinary infections. If this occurs, the SGLT2 [/fig]
|
Validation of the Social support and Pain Questionnaire (SPQ) in patients with painful temporomandibular disorders
Background: The present study aimed to validate of Social support and Pain Questionnaire (SPQ) for use in Chinese patients with painful temporomandibular disorders (TMD). Methods: The Chinese version of SPQ was produced by translation and cross-culturally adaptation of the original English version according to international guidelines. The Chinese version of SPQ was then distributed to a total of 118 patients with painful TMD. Reliability of the SPQ was evaluating using internal consistency and test-retest methods and validity of the SPQ was determined by construct validity and convergent validity. The exploratory factor analysis (EFA) was used to assess the construct validity of SPQ. And convergent validity was assessed by correlating the SPQ scores with the score of a global oral health question. Results: The Chinese version of SPQ has a high internal consistency (Cronbach's alpha value, 0.926) and good test-retest reliability ((intraclass correlation coefficient (ICC), 0.784). Construct validity was evaluated by EFA, extracting one factor, accounting for 74.8% of the variance. All factor loadings of the six items had exceeded 0.80. As regards convergent validity, the SPQ showed good correlation with the global oral health question.Conclusion: These findings support that the Chinese version of SPQ can be used as a reliable and valid tool for Chinese patients with painful TMD.
# Background
Temporomandibular disorders (TMD) are "a group of biopsychosocial illnesses characterized by chronic painful conditions and dysfunction in the muscles of mastication and the temporomandibular joint" [bib_ref] Identifying potential predictors of pain-related disability in Turkish patients with chronic temporomandibular..., Ozdemir-Karatas [/bib_ref]. It affects 14.9 to 17.9% of Chinese population [bib_ref] Temporomandibular disorders in German and Chinese adolescents, Wu [/bib_ref] [bib_ref] Prevalence of temporomandibular joint dysfunction (TMJD) in Chinese children and adolescents. A..., Deng [/bib_ref]. Pain is one of the most common clinical symptoms of TMD [bib_ref] Chronic Temporomandibular disorders: disability, pain intensity and fear of movement, Gil-Martinez [/bib_ref]. TMD pain involves not only the masticatory muscles and TMD, but also affects the adjacent structures such as ears, teeth, head and neck muscles [bib_ref] Review of aetiological concepts of temporomandibular pain disorders: towards a biopsychosocial model..., Suvinen [/bib_ref]. Studies have shown that oral health-related quality of life (OHRQoL) is negatively affected by TMD pain [bib_ref] Oral health-related quality of life in patients with Temporomandibular disorders, Almoznino [/bib_ref] [bib_ref] TMD pain: the effect on health related quality of life and the..., Tjakkes [/bib_ref] [bib_ref] Validation of the Tampa scale for Kinesiophobia for Temporomandibular disorders (TSK-TMD) in..., He [/bib_ref].
In recent years, there is a growing interest in exploring the relationship between social support and pain behavior [bib_ref] Pain coping and social support as predictors of long-term functional disability and..., Evers [/bib_ref] [bib_ref] Physical and psychosocial predictors of functional trunk capacity in older adults with..., Ledoux [/bib_ref] [bib_ref] Illness behaviors in patients with unexplained chronic fatigue are associated with significant..., Romano [/bib_ref]. Several studies reported that social support plays an important role in overcoming pain-related disability [bib_ref] Pain coping and social support as predictors of long-term functional disability and..., Evers [/bib_ref] [bib_ref] Social support as a buffer to the psychological impact of stressful life..., Kornblith [/bib_ref]. Other studies, however, have shown that social support has negative effects on mobility and physical function [bib_ref] Hip fracture-related pain strongly influences functional performance of patients with an intertrochanteric..., Kristensen [/bib_ref] [bib_ref] Chronic pain patient-spouse behavioral interactions predict patient disability, Romano [/bib_ref]. To further explore the impact of social support on chronic pain may give us new insights into the treatment of pain-related diseases [bib_ref] Social support in chronic pain: development and preliminary psychometric assessment of a..., Van Der Lugt [/bib_ref]. Most measures on social support are mainly concerned with social support in daily-life situations, rather than its relationship to pain [bib_ref] Groningen Orthopaedic social support scale: validity and reliability, Van Den Akker-Scheek [/bib_ref] [bib_ref] The measurement of social support in the 'European research on incapacitating diseases..., Suurmeijer [/bib_ref]. The West Haven-Yale Multidimensional Pain Inventory (MPI) is the widely used instrument to evaluate the chronic pain experience from the cognitive-behavioral perspective [bib_ref] The West Haven-Yale Multidimensional pain inventory (WHYMPI), Kerns [/bib_ref] [bib_ref] Do psychological and behavioral factors classified by the West Haven-Yale Multidimensional pain..., Eklund [/bib_ref]. It does concern support related to pain by one person. However, it fails to capture the social support from others, like friends and family [bib_ref] Social support and social structure: a descriptive epidemiology, Turner [/bib_ref].
Recently, the Social support and Pain Questionnaire (SPQ) was developed by the Academic Centre for Dentistry Amsterdam (ACTA) to measure satisfaction with social support related to pain [bib_ref] Social support in chronic pain: development and preliminary psychometric assessment of a..., Van Der Lugt [/bib_ref]. It is a 6-item, one-factor structure, self-administered English instrument [bib_ref] Social support in chronic pain: development and preliminary psychometric assessment of a..., Van Der Lugt [/bib_ref]. The SPQ is a reliable and valid tool to evaluate the patient's satisfaction with pain-related social support [bib_ref] Social support in chronic pain: development and preliminary psychometric assessment of a..., Van Der Lugt [/bib_ref].
To make this instrument suitable for other languages and cultures, an internationally standardized evaluation procedure must be implemented. Therefore, the aim of our study is to validate a Chinese translation of the SPQ for patients with painful TMD.
# Methods
## Patients
A total of 118 consecutive patients were recruited from our university clinic between January 2015 and October 2015. The inclusion criteria were: at least 18 years of age, a diagnosis of painful TMD according to the Research Diagnostic Criteria for Temporomandibular Disorders [bib_ref] Diagnostic criteria for Temporomandibular disorders (DC/TMD) for clinical and research applications: recommendations..., Schiffman [/bib_ref] , complaint of pain for at least 3 months, and sufficient ability to fill in the questionnaire. The exclusion criteria included: had a history of psychiatric illness, had acute pain for less than 3 months, a systemic disease, dental pain, and were reluctant to sign informed consents. For the sample size, it was determined according to the quality criteria for health status questionnaires [bib_ref] Quality criteria were proposed for measurement properties of health status questionnaires, Terwee [/bib_ref]. The criteria suggested that a study should recruit at least 100 patients for reliability and validity analysis [bib_ref] Quality criteria were proposed for measurement properties of health status questionnaires, Terwee [/bib_ref]. Before completing the SPQ, patients were instructed how to fill in the questionnaire. When necessary, they can consult interviewer at any time.
All the patients signed the written informed consent, and the present study was approved by the Ethics Committee of Chongqing Medical University.
## The spq
The original SPQ is composed of 6 items: the support that I get from the people around me, the advice that I get from the people around me, how much opportunity I have to discuss the pain with the people around me, how much care I receive, how much understanding the people around me show and the practical help people around me give. The response is scored using a Likert scale from 0 to 4 (0 = very dissatisfied, 1 = dissatisfied, 2 = neutral, 3 = satisfied, 4 = very satisfied). In general, higher scores indicate more satisfaction with social support related to pain. Additionally, to assess the convergent validity, an extra global question ("In general, how would you rate your social support related to pain ") was added at the end of instrument. The question is scored with the following options: (1 = very good, 2 = good, 3 = fair, 4 = poor, 5 = very poor).
## Translation and cross-cultural adaptation
The process of translation and cross-cultural adaptation of the original SPQ was carried out according to Guillemin's guidelines [bib_ref] Cross-cultural adaptation of health-related quality of life measures: literature review and proposed..., Guillemin [/bib_ref].
## Translation into
# Statistical analysis reliability
A Cronbach's alpha between 0.70 and 0.95 and an Intraclass Correlation Coefficients (ICC) value over 0.70 were used to determine the internal consistency and test-retest reliability, respectively [bib_ref] Quality criteria were proposed for measurement properties of health status questionnaires, Terwee [/bib_ref]. The ICC vales were analyzed by asking 30 patients to complete the Chinese SPQ again after a period of 2-week.
## Validity
Validation of the Chinese SPQ included the assessment of the construct validity and convergent validity. Construct validity was examined using exploratory factor analysis (EFA). The Kaiser-Meyer-Olkin test and Bartlett's sphericity test were firstly carried out to determine if the data was suitable for performing the EFA [bib_ref] Tests of significance in factor analysis, Bartlett [/bib_ref]. The EFA followed by varimax rotation method was carried out to evaluate the dimensionality of the Chinese SPQ. Significance was defined as a loading higher than 0.40. All the statistical analysis were performed using version 20.0 of the SPSS software (IBM Corp. 2011; NY; USA). In the present study, the p value of <0.05 was considered statistically significant.
# Results
## Patient characteristics
A total of 118 patients with painful TMD were recruited from a university-affiliated dental clinic. All the patients declared that the questions were easy to understand and completed the questionnaire fully. Patient characteristics are shown in [fig_ref] Table 1: Characteristics of patients [/fig_ref]. Of the 118 patients selected, 56.8% were female, with the mean age of 46.4 ± 14.7 years old. The majority of patients were employed (72.9%), and half of them had been educated in middle school (54.2%). The pain were classified into joint pain (32.2%), muscle pain (36.4%) and mixed pain (31.4%). The mean scores, corrected item-total correlations and Cronbach's alpha if item deleted results for the SPQ are presented in [fig_ref] Table 3: Range, mean scores, corrected item-total correlations and factor analysis results for the... [/fig_ref].
## Reliability
As displayed in, Cronbach's alpha for the SPQ was 0.926. The test-retest reliability of the instrument was also acceptable, with the ICC value of 0.87.
## Validity
The result of the KMO test was 0.788 and Bartlett's test of sphericity was 716.2 (df = 15, P < 0.001). These results suggested that EFA of the sample was appropriate. [fig_ref] Table 3: Range, mean scores, corrected item-total correlations and factor analysis results for the... [/fig_ref] shows the results of the EFA for the SPQ. The factor loadings of all items had exceeded the recommended 0.40. A one-factor model, accounted for 74.8% of the explanatory variance, was retained.
In term of convergent validity, it was assessed by correlating the SPQ scores with the score of the global oral health question. The result showed that the correlation was good.
# Discussion
In the present study, the translation and validation of the Chinese version of SPQ is carried out. Statistical analysis showed that the 6-item Chinese version of SPQ with one-factor structure was reliable and valid. It can be used in Chinese population to evaluate the patient's satisfaction with pain-related social support.
In the current study, chronic TMD pain lasting for at least 3 months was used as a temporal criterion to recruit patients suffering from painful TMD. This was consistent with other previous studies [bib_ref] Social support in chronic pain: development and preliminary psychometric assessment of a..., Van Der Lugt [/bib_ref] [bib_ref] The Portuguese formal social support for autonomy and dependence in PAIN inventory..., Matos [/bib_ref]. In order to get a semantically and conceptually similar version to the original SPQ, the process of translation and cross-cultural adaptation of the original SPQ was carried out. The Chinese version of SPQ was overall well received by the patients, and no problem or language difficulty existed.
In terms of reliability, both the internal consistency and test-retest reliability were proved to be good. The Cronbach's alpha for the SPQ was 0.926, indicating a high level of internal consistency. This value indicates high correlations among items of SPQ. And when each item was deleted, the values of Cronbach's alpha remained stable. The value of ICC (0.784, 90% CI: 0.648-0.893) was [bib_ref] Social support in chronic pain: development and preliminary psychometric assessment of a..., Van Der Lugt [/bib_ref]. For the selection of the time interval, previous study suggested that a period of 2 days to 2 weeks was thought to be suitable [bib_ref] A comparison of two time intervals for test-retest reliability of health status..., Marx [/bib_ref]. A period of 2 weeks was chosen in the present study. And in the original study, a maximal time interval of 8 weeks was adopted. The lower ICC value may be explained by the longer time interval in the original study [bib_ref] Social support in chronic pain: development and preliminary psychometric assessment of a..., Van Der Lugt [/bib_ref]. Regarding factorial structure of the SPQ, the finding of EFA identified a one-factor structure of the SPQ. This result was consistent with previous study proposed by Van Der Lugt et al. [bib_ref] Social support in chronic pain: development and preliminary psychometric assessment of a..., Van Der Lugt [/bib_ref]. Similar as the original study, the factor loadings of all items were above 0.80 [bib_ref] Social support in chronic pain: development and preliminary psychometric assessment of a..., Van Der Lugt [/bib_ref]. As regards convergent validity, the Chinese version of SPQ had good correlation with the global oral health question. The finding was close to the values in the original study [bib_ref] Social support in chronic pain: development and preliminary psychometric assessment of a..., Van Der Lugt [/bib_ref]. Overall, the Chinese version of SPQ had adequate validity for using in patients with painful TMD.
However, the present study has two limitations which requiring further assessment. Firstly, all patients were enrolled from a single university-affiliated hospital, and thus may not be able to represent all Chinese population affected by painful TMD. Secondly, the current study does not contain long-term follow-up analyses, and therefore the sensitivity and responsiveness of the Chinese version of SPQ could not be evaluated.
# Conclusion
The present study supports that the Chinese version of Social support and Pain Questionnaire (SPQ) can be used as a reliable and valid tool for Chinese patients with painful TMD.
[table] Table 1: Characteristics of patients (n = 118) [/table]
[table] Table 3: Range, mean scores, corrected item-total correlations and factor analysis results for the SPQ higher than the values in the original study, indicating good test-retest reliability [/table]
|
Systemic capillary leak syndrome requiring fasciotomy for limb compartment syndrome: A case report and literature review
Background: Limb compartment syndrome (LCS), a rare but serious complication of systemic capillary leak syndrome (SCLS)related systemic edema, warrants prompt decompressive fasciotomy. We report a case of SCLS complicated by LCS of four extremities requiring emergent fasciotomies; furthermore, we reviewed existing published reports on SCLS with LCS.Case Presentation: A 36-year-old man was diagnosed with SCLS based on profound vascular permeability with no other underlying conditions. Within a few hours of admission, LCS was noted in the patient's lower legs and thighs and he was treated using decompressive fasciotomy. Additional forearm fasciotomy was subsequently carried out. After fluid management, vasopressor support, mechanical ventilation, and renal replacement therapy, the patient was discharged without any neuromuscular deficits caused by LCS. Literature review suggested that lower legs are prone to LCS in patients with SCLS.Conclusion: Limb compartment syndrome is a serious complication that clinicians must be aware of and requires prompt decompressive fasciotomy.
# Introduction
S YSTEMIC CAPILLARY LEAK syndrome (SCLS) is a rare disease characterized by recurrent hypovolemia, systemic edema, hypoalbuminemia, hypotension, and hemoconcentration due to life-threatening attacks on capillary hyperpermeability. [bib_ref] Narrative review: the systemic capillary leak syndrome, Druey [/bib_ref] Due to significant fluid and protein shifts to extravascular spaces, initial treatment of SCLS requires remarkable fluid resuscitation to maintain hemodynamic stability. Profound systemic edema could cause limb compartment syndrome (LCS), a rare but serious complication warranting decompressive fasciotomies. [bib_ref] Compartment syndrome as a result of systemic capillary leak syndrome. Case Rep, Kyeremanteng [/bib_ref] [bib_ref] Systemic capillary leak syndrome associated with hypovolemic shock and compartment syndrome, Saugel [/bib_ref] Herein, we report a case of SCLS complicated by LCS of four limbs that required emergent fasciotomies. We reviewed existing studies on SCLS with compartment syndrome as there were only sporadic cases previously reported. In addition to raising awareness of this rare disease as a differential diagnosis of significant hypovolemic shock, our aim was to emphasize the importance of appropriate management of compartment syndrome because delayed intervention could induce neuromuscular sequalae and impact the patient's functional prognosis.
## Case presentation
A 36-year-old man presented to our hospital emergency department with dizziness and vomiting. He had experienced mild flu-like symptoms including nasal discharge and sputum production the day prior. Initial physical examination revealed the following: mild disorientation; Glasgow coma scale score, 14 (E4V4M6); body temperature, 36.3°C; blood pressure, 71/41 mmHg; pulse rate, 113 b.p.m.; respiratory rate, 20 breaths/min; and oxygen saturation, 93% with 4 L/min oxygen. He had no fever, rash, or skin bruising. Initial arterial blood gas analysis showed metabolic acidosis [fig_ref] Table 1: Laboratory results on admission of a 36-year-old man with systemic capillary leak... [/fig_ref]. Laboratory results showed an increased hematocrit level of 71% and hemoglobin level of 23 g/dl [fig_ref] Table 1: Laboratory results on admission of a 36-year-old man with systemic capillary leak... [/fig_ref]. Laboratory results did not indicate cytopenia and imaging studies did not indicate any notifiable infections, malignancy, or trauma; indications of splenomegaly, lymphadenopathy, or intracorporeal fluid accumulation were absent.
He was admitted to the intensive care unit and required resuscitative treatment for hypovolemic shock. He received 6 L crystalloid fluid during the first 4 h, vasopressor support, and antibiotics including meropenem and vancomycin for suspected bacterial infection. Although aggressive resuscitation was performed, he remained hypotensive and continued to require fluids. Eight hours after admission, mechanical ventilation and continuous renal replacement therapy (CRRT) were initiated to support respiratory failure and metabolic acidosis with oliguria, respectively. An increase in hematocrit and decrease in total protein and albumin levels indicated significant capillary permeability. Because of the profound hypotensive shock, which required a large quantity of fluid, and progressive systemic edema, we considered a diagnosis of capillary leak syndrome. Given the absence of medication or underlying condition that might induce capillary leak syndrome, and laboratory and imaging findings suggestive of differential diagnoses such as hemophagocytic lymphohistiocytosis or multicentric Castleman disease, the patient was diagnosed with idiopathic SCLS. Plasma electrophoresis confirmed the presence of monoclonal immunoglobulin (kappa).
Within 1 h of admission and during the subsequent 6 h, he experienced pain with increased muscular tension in the lower extremities. After intubation, compartment syndrome was suspected in the lower limbs as the muscular tension worsened. The compartment pressure of the lower legs and thighs was 40-100 mmHg under a diastolic pressure of 70 mmHg. Physical examination findings, clinical course, and increased compartment pressures were consistent with compartment syndrome of the lower limbs. We performed decompressive fasciotomies on the lower legs and thighs. Over several subsequent hours both forearms developed compartment syndrome, also requiring fasciotomies. Vacuum-assisted dressings were placed on all extremities.
In addition to supportive treatments including fluid management, vasopressor support, mechanical ventilation, and CRRT, intravenous immunoglobulin and steroid therapy were administered for acute phase treatment. Theophylline and b-adrenergic agonist (salbutamol) were added for maintenance therapy.
By admission day 3, his hemodynamic state had gradually recovered, and he required small quantities of fluids and vasopressors. Continuous renal replacement therapy was discontinued on day 3, and the patient was extubated on day 5. The first closure surgery for fasciotomy wounds on the left forearm, right thigh, and left leg were carried out on day 4, and the remaining wounds were closed on day 7. His clinical course was good until he experienced another attack of hypotensive shock after the last closure surgery. Seven hours after the second closure surgery, he showed a drop in blood pressure (75/48 mmHg) and sinus tachycardia of 135 b.p.m.; he required fluid resuscitation and vasopressor support. During this transient hypotension and tachycardia, changes in hematocrit and albumin levels were minimal. As there was no pain and swelling of the extremities and creatine kinase levels did not increase, there were no indications of a compartment syndrome relapse. The patient showed no signs of infection and achieved negative blood culture results without requiring antibiotics. His clinical course indicated the possibility of a second attack of mild capillary leak syndrome, which occurred when the initial attack seemed to have resolved. He was discharged from the intensive care unit on day 11 and from the hospital on day 18 without any neuromuscular deficits caused by compartment syndrome [fig_ref] Figure 1: Clinical course of a 36-year-old man with systemic capillary leak syndrome complicated... [/fig_ref].
# Discussion
L IMB COMPARTMENT SYNDROME is known to occur after limb trauma; however, it might not be easily recognizable when caused by nontraumatic diseases. Recognizing SCLS as a nontraumatic condition that causes LCS and understanding its clinical time course could increase LCS awareness among clinicians and help them perform fasciotomies in a timely manner.
Systemic capillary leak syndrome episodes have an infectious trigger, especially mediated by viruses 1 ; however, the overall pathophysiology of SCLS remains unclear. Although our patient experienced mild flu-like symptoms, no viral etiologies were identified, and the event did not occur during the influenza season or COVID-19 pandemic. The clinical diagnosis of SCLS often requires the exclusion of other conditions such as sepsis, angioedema, anaphylactic shock, hemophagocytic lymphohistiocytosis, or multicentric Castleman disease. [bib_ref] Narrative review: the systemic capillary leak syndrome, Druey [/bib_ref] Although it is difficult to distinguish other differential diagnoses from the first impression, the current patient's clinical course, including profound vascular permeability with the absence of other underlying conditions, was more likely to fit the clinical picture of SCLS. Hemoconcentration and monoclonal gammopathy could be diagnostic clues for differentiating SCLS from other conditions. [bib_ref] Systemic capillary leak syndrome resulting in compartment syndrome and the requirement for..., Perry [/bib_ref] Patients with SCLS typically develop prodromal symptoms, such as weakness, malaise, myalgias, or abdominal pain, followed by a leakage phase represented by hypotensive shock and edema. This phase, which lasts for 1-3 days, 1 is the critical period where patients are prone to hypoperfusion-related multiorgan dysfunction and LCS as severe complications 2-10 ; patients require the administration of fluids (in massive quantities), vasopressors for resuscitation, and additional treatments for complications. The leakage phase is followed by the postleakage phase, in which vascular permeability starts to be restored with fluid recruitment into the intravascular circulation. [bib_ref] Narrative review: the systemic capillary leak syndrome, Druey [/bib_ref] Our patient experienced a typical course of SCLS, where the leakage phase lasted for approximately 3 days, with kidney dysfunction warranting CRRT and LCS requiring decompressive fasciotomies. Specific treatments for SCLS, including intravenous immunoglobulin, steroids, b-adrenergic agonists, and theophylline, have been reported for acute treatment and recurrence prevention. Although medical treatments for SCLS are based on reported cases and their effectiveness remains unproven, intravenous immunoglobulin is a promising therapy. [bib_ref] Capillary leak syndrome: etiologies, pathophysiology, and management, Siddall [/bib_ref] PubMed and Google Scholar searches using the keywords "systemic capillary leak syndrome" and "compartment syndrome" identified nine reported cases of adult SCLS complicated by LCS, published in English since 2000 [fig_ref] Table 2: Clinical review of reported cases of systemic capillary leak syndrome with limb... [/fig_ref]. Eight of these patients were men (average age, 45 years). Most patients presented with lower leg pain and swelling as the chief complaint, indicating that the diagnosis of LCS was made prior to or during the recognition of SCLS. Thus, LCS could be the primary presentation of SCLS patients seeking medical care. However, some patients present with nonspecific complaints and later develop LCS, as in our case. Although most patients survived, notable complications associated with compartment syndrome were neuropathies. Thus, lower legs are prone to LCS in patients with SCLS, and early recognition as well as timely decompressive fasciotomies are mandatory to avoid neuromuscular complications.
Limb compartment syndrome is a serious complication that clinicians must be aware of and requires prompt decompressive fasciotomy. Careful examination of the extremities is essential, as symptoms in the early phase, including pain or paresthesia, could be difficult to assess when the patient is under intensive care.
[fig] Ó 2022: The Authors. Acute Medicine & Surgery published by John Wiley & Sons Australia, Ltd on behalf of Japanese Association for Acute Medicine. [/fig]
[fig] Figure 1: Clinical course of a 36-year-old man with systemic capillary leak syndrome complicated by limb compartment syndrome during his stay in the intensive care unit. Alb, albumin; CRRT, continuous renal replacement therapy. Ó 2022 The Authors. Acute Medicine & Surgery published by John Wiley & Sons Australia, Ltd on behalf of Japanese Association for Acute Medicine. [/fig]
[table] Table 1: Laboratory results on admission of a 36-year-old man with systemic capillary leak syndrome complicated by limb compartment syndrome [/table]
[table] Table 2: Clinical review of reported cases of systemic capillary leak syndrome with limb compartment syndrome (LCS) in the last 22 years [/table]
|
Combined use of expression and CGH arrays pinpoints novel candidate genes in Ewing sarcoma family of tumors
Background: Ewing sarcoma family of tumors (ESFT), characterized by t(11;22)(q24;q12), is one of the most common tumors of bone in children and young adults. In addition to EWS/FLI1 gene fusion, copy number changes are known to be significant for the underlying neoplastic development of ESFT and for patient outcome. Our genome-wide high-resolution analysis aspired to pinpoint genomic regions of highest interest and possible target genes in these areas.Methods: Array comparative genomic hybridization (CGH) and expression arrays were used to screen for copy number alterations and expression changes in ESFT patient samples. A total of 31 ESFT samples were analyzed by aCGH and in 16 patients DNA and RNA level data, created by expression arrays, was integrated. Time of the follow-up of these patients was 5-192 months. Clinical outcome was statistically evaluated by Kaplan-Meier/ Logrank methods and RT-PCR was applied on 42 patient samples to study the gene of the highest interest.Results: Copy number changes were detected in 87% of the cases. The most recurrent copy number changes were gains at 1q, 2, 8, and 12, and losses at 9p and 16q. Cumulative event free survival (ESFT) and overall survival (OS) were significantly better (P < 0.05) for primary tumors with three or less copy number changes than for tumors with higher number of copy number aberrations. In three samples copy number imbalances were detected in chromosomes 11 and 22 affecting the FLI1 and EWSR1 loci, suggesting that an unbalanced t(11;22) and subsequent duplication of the derivative chromosome harboring fusion gene is a common event in ESFT. Further, amplifications on chromosomes 20 and 22 seen in one patient sample suggest a novel translocation type between EWSR1 and an unidentified fusion partner at 20q. In total 20 novel ESFT associated putative oncogenes and tumor suppressor genes were found in the integration analysis of array CGH and expression data. Quantitative RT-PCR to study the expression levels of the most interesting gene, HDGF, confirmed that its expression was higher than in control samples. However, no association between HDGF expression and patient survival was observed.Conclusion:We conclude that array CGH and integration analysis proved to be effective methods to identify chromosome regions and novel target genes involved in the tumorigenesis of ESFT.
# Background
The Ewing sarcoma family of tumors (ESFT) is a group of highly aggressive and often metastatic small round cell tumors characterized by specific t(11;22)(q24;q12) chromosomal rearrangements, which create the EWS/FLI1 gene fusion and thereby a chimeric, oncogenic transcription factor [bib_ref] Gene fusion with an ETS DNA-binding domain caused by chromosome translocation in..., Delattre [/bib_ref]. ESFT is one of the most common bone and soft tissue tumors in children and young adults arising generally during the second decade of life [bib_ref] Ewing's sarcoma: diagnostic, prognostic, and therapeutic implications of molecular abnormalities, Burchill [/bib_ref] [bib_ref] Overview of sarcomas in the adolescent and young adult population, Herzog [/bib_ref]. The ESFT tumors are divided into four subtypes according to the histopathological description: classical Ewing sarcoma in bones, extraskeletal Ewing sarcoma, peripheral neuroepithelioma (PNET), and Askin's tumor. Most of these ESFT cases manifest defects in the maintenance of genomic stability with subsequent DNA copy number alterations.
Conventional CGH and array CGH studies have shown that 63-84% of ESFT patient samples have copy number changes [bib_ref] Recurrent gains of 1q, 8 and 12 in the Ewing family of..., Armengol [/bib_ref] [bib_ref] Elomaa I: Clinical correlations of genetic changes by comparative genomic hybridization in..., Tarkkanen [/bib_ref] [bib_ref] CGH analysis of secondary genetic changes in Ewing tumors: correlation with metastatic..., Brisset [/bib_ref] [bib_ref] Dockhorn-Dworniczak B: Genetic imbalances revealed by comparative genomic hybridization in Ewing tumors, Ozaki [/bib_ref] [bib_ref] Molecular cytogenetic parameters in Ewing sarcoma, Amiel [/bib_ref] [bib_ref] Array CGH and gene-expression profiling reveals distinct genomic instability patterns associated with..., Ferreira [/bib_ref]. These copy number alterations play a significant role in the tumorigenesis and malignant progression of solid tumors. The diagnosis and clinical management of patients would substantially benefit from identification of these novel chromosomal targets and molecular markers involved in the tumorigenesis of ESFT, since secondary genetic alterations in ESFT have been shown to correlate with patient's outcome. In addition to overall number of chromosomal imbalances [bib_ref] Updates on cytogenetics and molecular genetics of bone and soft tissue tumors:..., Sandberg [/bib_ref] [bib_ref] Acquisition of secondary structural chromosomal changes in pediatric ewing sarcoma is a..., Zielenska [/bib_ref] , gains of 1q, 8 and 12 and losses of 9p21. [bib_ref] Overview of sarcomas in the adolescent and young adult population, Herzog [/bib_ref] and 16q have been associated with poor clinical outcome [bib_ref] Dockhorn-Dworniczak B: Genetic imbalances revealed by comparative genomic hybridization in Ewing tumors, Ozaki [/bib_ref] [bib_ref] Prognostic impact of deletions at 1p36 and numerical aberrations in Ewing tumors, Hattinger [/bib_ref] [bib_ref] Ewing sarcomas with p53 mutation or p16/p14ARF homozygous deletion: a highly lethal..., Huang [/bib_ref]. Rapid development of microarray technology has led to more sophisticated analyses, which can be utilized to find novel tumor specific genetic alterations. Further, numerous studies have demonstrated that integrating genomic data from different sources, e.g. at RNA and DNA level, can enhance the reliability of genetic analysis in understanding tumor progression. Our aim was to identify common regions of gain and loss and to define the influence of copy number alterations on gene expression to identify chromosomal areas and genes involved in malignant progression of Ewing sarcoma. We used high-resolution array-based CGH to screen simultaneously multiple loci for possible copy number imbalances in ESFT patient samples. This approach enables us to detect both largescale and gene-size copy number alterations down to ~35 kb in size. To investigate the impact of copy number imbalances on the gene expression levels of affected genes, we performed also an expression array analysis to combine RNA and DNA level data and validated the most interesting result by quantitative RT-PCR analysis.
# Methods
## Patient samples and clinical data
Fresh frozen samples (stored at -70°C) were collected from the archives of the Laboratory of Oncologic Research, Istituti Ortopedici Rizzoli (IOR), Bologna. A total of 31 tumor specimens of from ESFT patients treated at IOR between years 1992 and 2005 were available for the aCGH study. In order to study ESFT expression profiles, 42 patient samples were collected for RNA extraction. To validate the ESFT diagnosis, the presence of EWS/ FLI or EWS/ERG translocation was confirmed by RT-PCR for all samples with available RNA. Clinical data for 31 samples [fig_ref] Table 1: Clinical data of the ESFT patients included in array CGH and/or data... [/fig_ref] used in aCGH and in data integration analysis were collected from the patient records at IOR. All patients were treated within controlled prospective trials [bib_ref] Italian Cooperative Study for the treatment of children and young adults with..., Rosito [/bib_ref] [bib_ref] Adjuvant and neo-adjuvant chemotherapy for Ewing's sarcoma family tumors and osteosarcoma of..., Bacci [/bib_ref]. The mean age of the patients was 20.7 years, ranging from 5 to 41 years and the male-to-female ratio was 22:9 (2.4). Of the 31 samples used in aCGH analysis, 23 were primary tumors, two recurrencies, and six metastatic tumors. The majority of these patients were diagnosed with classical Ewing sarcoma, four with soft tissue Ewing sarcoma, three with Askin's tumor and two with PNET. Seven of the patients with primary tumors had metastases at the time of diagnosis. Sixteen tumors had Type 1 (exon 7 of EWS/exon 6 of FLI1) gene fusion, eight had other types of fusion (Type 2: exon 7 of EWS/exon 5 of FLI1 or Type 3: exon 10 of EWS/exon 6 of FLI1), three samples were negative for the most common fusion genes (EWS-FLI1 and EWS-ERG), and in four cases this information was not available. The sample set was handled in a coded fashion and the collected clinical and quality control data of the samples is publicly available in a microarray database at http://www.cangem.org [bib_ref] CanGEM: mining gene copy number changes in cancer, Scheinin [/bib_ref]. This study has been reviewed and approved by the Ethical Review Board of Helsinki University Central Hospital.
## Nucleic acid isolation
Genomic DNA from 31 samples was extracted using the standard phenol-chloroform method. Prior to extraction, the proportion of tumor cells was verified to exceed 75% in all samples by using hematoxylin and eosin-stained sections. Tissue necrosis was evaluated on the whole tumor mass. In cases with high percentages of necrosis, nucleic acids were isolated from the tissue samples in which viable cells were still present. Reference DNAs, male and female, were extracted from pooled blood samples (4 individuals) obtained from Blood Service, Red Cross, Finland. RNA from 42 ESFT samples was isolated using a TRIzol extraction kit (Invitrogen Ltd., Paisley, UK) according to the manufacturer's instructions. Both high quality genomic DNA and RNA were available for 16 patients after the nucleic acid extraction instead of 42 patients, due to insufficient amount of starting material. DNA and RNA concentrations were measured using a GeneQuant pro spectrophotometer (Amersham Pharmacia, Cambridge, UK), and RNA quality was assessed using Agilent's 2100 Bioanalyzer (Agilent, Palo Alto, CA).
## Array cgh hybridization, microarray image and data analysis
Digestion, labeling, hybridization, and data analysis of genomic DNA was performed according to Agilent's protocol version 2.0 for 44K arrays as described previously [bib_ref] Microdeletions in 9p21.3 induce false negative results in CDKN2A FISH analysis of..., Savola [/bib_ref] [bib_ref] CDKN2A deletions in acute lymphoblastic leukemia of adolescents and young adults-An array..., Usvasalo [/bib_ref]. In brief, the sample and reference DNAs, 7 μg each, were fragmented and 1.0-1.5 μg of the fragmented DNA was labeled by random priming using a BioPrime array labeling kit (Invitrogen, Carlsbad, CA) with Cy3-dUTP and Cy5-dUTP dyes (Perkin-Elmer, Wellesley, MA). Labeled samples were purified, combined, and hybridized for 48 h at 65°C, 10 rpm to Human Genome CGH 44B oligomicroarray slides (Agilent Technologies Santa Clara, CA) against gender matched reference DNAs. Then the arrays were washed and scanned [bib_ref] Microdeletions in 9p21.3 induce false negative results in CDKN2A FISH analysis of..., Savola [/bib_ref]. The array images were analyzed and data was extracted using Agilent's Feature Extraction (FE) Software version 8.1, providing dye normalization (Linear Lowess) and background substraction. The chromosomal imbalances were identified using Agilent's CGH Analytics software version 3.4. The altered chromosomal regions and breakpoints were detected using ADM-2 (threshold 8.0) with 1.0 Mb window size. Patient survival analysis was then performed by Kaplan-Meier and Logrank (Mantel-Cox) methods considering either event-free or overall survival.
## Expression array hybridizations
The ESFT RNA samples, 42 cases and control samples, a CD34+ cell line and a pool of normal muscle tissue samples were hybridized to Affymetrix Human Genome U133 Plus 2.0 oligonucleotide microarrays (Affymetrix, Santa Clara, CA) according to the manufacturer's GeneChip ® One-Cycle Target Labeling-protocol. In brief, 5 μg of total RNA was reverse transcribed to cDNA using One-Cycle cDNA Synthesis Kit (Affymetrix). Biotin-labeling of antisense cRNA was carried out using IVT Labeling Kit (Affymetrix). The labeled and fragmented cRNA (15 μg of each) was hybridized for 16 h at 45°C in a hybridization oven 640 (60 rpm). Washing and staining of the arrays with streptavidin-phycoerythin (SAPE) was completed in a Fluidics Station 450 (Affymetrix). The arrays were then scanned using a confocal laser GeneChip Scanner 3000 and images were analyzed using GeneChip Operating Software (GCOS; Affymetrix, Sacramento, CA). The expression measurements were preprocessed using Robust Multi-array Analysis (RMA) for the whole collection of 44 chips (42 ESFT patients and two hypothetical normal samples). While only 16 of these were used in the integration, running the preprocessing for the whole collection (n = 44) provides more accurate estimates of the true expression levels.
## Integration of gene copy number and expression data
In order to compare the measurements obtained on Affymetrix and Agilent platforms, the sequences used in the probes were matched to the NCBI36 human genome build, using the BLAST algorithm to provide a unique location for each Affymetrix probe set using the target sequences provided by Affymetrix and each Agilent probe. Multiple matches were combined to provide a single location covering all matches if the resulting sequence length was below 2,5 Mb. Note that the locations do not necessarily match the reference sequences of the NCBI36 genome, since they correspond to the locations of the probe sequences, not RefSeqs. In the joint analysis, each Affymetrix probe set was paired with the closest Agilent probe, measured as the distance between the mean points of the sequences. The Affymetrix probe sets that had no Agilent probes within 375 kb were ignored. Correlation between expression and gene copy number of different patients was measured separately for each gene (identified based on the Affymetrix probe set). Genes with high positive or negative correlation were chosen for further examination. The goal of this process was to detect genes where a copy number change and a change in expression are observed on the same patients. A similar analysis was conducted by first dividing patients into groups according to their copy number status and then testing whether these groups have a significant difference in their expression levels [bib_ref] Integrated gene copy number and expression microarray analysis of gastric cancer highlights..., Myllykangas [/bib_ref]. Here the correlation approach was chosen instead of the testing approach, because it can take into account also small amplification imbalances not detectable with the method described in Section "Array CGH hybridization, microarray image and data analysis". It also takes into account possible higher copy number changes. As a correlation measure we used Spearman's rank correlation, since the copy number data does not follow normal distribution. We used the algorithm of Best [bib_ref] Algorithm AS 89: The Upper Tail Probabilities of Spearman's rho, Best [/bib_ref] to compute the p-value against the correlation being zero, and corrected for multiple testing by computing false discovery rates using the q-value procedure [bib_ref] Statistical significance for genomewide studies, Storey [/bib_ref]. The correlation was computed only for genes located on chromosome arms where at least 20% of the full patient collection, including also samples not used in the integration analysis, showed copy number aberration, in order to focus on regions where associations would be likely.
## Quantitative rt-pct analysis by taqman low density arrays
Pre-designed TaqMan PCR probe and primer sets for HDGF were used: Assay ID Hs00610314-m1 (Applied Biosystems, Foster City, CA, USA). All PCRs were done by using ABI PRISM 7900 Sequence Detection System (Applied Biosystems) as recommended by the supplier. Thermal cycling conditions were: 50°C for 2 min, 95°C for 10 min, 95°C for 15 sec and 60°C for 1 min. Gene expression values were calculated based on the ΔΔCt method [bib_ref] Analysis of relative gene expression data using real-time quantitative PCR and the..., Livak [/bib_ref] , in which RNA from CD34+ cells derived from human bone marrow and pooled muscle normal tissues derived from three patients were the designated cali-brators for the analysis of all other samples. CD34+ positive cells and pooled muscle normal tissues were processed in the same way as tumor samples and used as separate calibrators for the RT-PCR experiments. For evaluating the prognostic value of HDGF, we calculated its median expression value, and patients were stratified as "high-expressors" or "low-expressors" relative to the median value. Patient survival analysis was then performed by Kaplan-Meier and Logrank methods considering either event-free survival or overall survival.
# Results
## Copy number changes
Results from a high-resolution analysis of copy number aberrations in ESFT (n = 31), using Agilent's 44K oligoarray platform and CGH analytics software are shown in [fig_ref] Table 2: Array CGH results [/fig_ref]. In all ESFT patient samples, 0-26 aberrations were detected per sample (mean: 7.2) and 27 of the 31 samples showed (87%) copy number changes. All samples without copy number changes (n = 4) were primary tumors. Metastases (mean: 11.8) showed more copy number changes than local recurrencies (mean: 9.5) and primary tumors (mean: 5.8). The sizes of these aberrations ranged from < 60 kb deletions to gains or losses of whole chromosomes. Among primary tumors, the samples with low copy number changes (≤ 3 copy number aberrations) showed a significantly better prognosis with respect to those with a high number of chromosomal alterations (> 3 copy number aberrations), both in terms of event-free and overall survival [fig_ref] Figure 1: Outcome of patients with low copy number changes [/fig_ref]. Indeed, only 3/11 patients (27%) with less than three copy number changes developed metastases within 6 years from diagnosis in contrast with 8/10 (80%) of those with a high number of chromosomal alterations (P = 0,03 Fisher's test), indicating how the number of chromosomal alterations may have a highly prognostic significance despite the low number of patients here considered. Recurrent aberrations were gains of 1q (32%), 2 (29%), 8 (67%), and 12 (29%) and losses at 9p (23%) and 16q (32%) as visualized in . The prominent deletion in 9p21.3 harboring CDKN2A tumor suppressor gene and microdeletions of these region have been previously described and discussed in a separate report by Savola et al. [bib_ref] Microdeletions in 9p21.3 induce false negative results in CDKN2A FISH analysis of..., Savola [/bib_ref]. The gain of chromosome 8 was the most prominent copy number change in our sample set (21 of 31 cases). Gain of 8q arm (minimal common overlapping area) was present in all samples with chromosome 8 aberration. The minimal common overlapping area of copy number gain in chromosome 1 was 1q22-qter. In chromosome 12 the smallest common region of gain was 12q13.2-q14.1, which harbors two known oncogenes, ERBB3 and CDK4.
Losses of 16q were observed in three cases together with 1q gain, suggesting the occurrence of an unbalanced t(1;16). Interesting copy number gains of 11q24.3-qter and 22q11.12-q12.1 starting or ending, respectively, at FLI1 and EWSR1 loci, were detected in patient samples D153 , D248, and D254 [fig_ref] Table 2: Array CGH results [/fig_ref]. Copy number imbalances affecting the same loci were detected also in samples D154 (uncontinuous amplification 22q12.1-q12.1) , D312 (+11q24.3-qter), and D315 (-22q12.1) (see [fig_ref] Table 2: Array CGH results [/fig_ref]. Original microarray data, scanned images and FE output text files, are available at the public repository CanGEM http://www.cangem.org [bib_ref] CanGEM: mining gene copy number changes in cancer, Scheinin [/bib_ref].
## Integration of gene copy number and expression data
Array CGH data and expression data were combined for a total of 16 patient samples [fig_ref] Table 1: Clinical data of the ESFT patients included in array CGH and/or data... [/fig_ref]. Matching of expression microarray probes to the corresponding copy number microarray probes using a 375 kb genomic window yielded 53,145 probe pairs. 10,115 of those located in chromosomal areas where at least 20% of patients showed a copy number aberration (1q, 2q, 8q, 12, and 16q). Several putative ESFT-related genes were pinpointed, differentially expressed due to copy number alteration in these chromosomal locations of highest interest. These novel putative oncogenes and tumor suppressor genes based on our data analysis include 20 genes (by q-value < 0.20), which previously have not been associated with ESFT [fig_ref] Table 3: Putative target genes for tumorigenesis and tumor progression in recurrent copy number... [/fig_ref]. For a supplementary table with all integration analysis results see Additional file 1.
## Microarray analysis and quantitative rt-pcr on hdgf
Array CGH and expression microarray results on showed clear evidence that patients with HDGF gain had higher HDGF expression [fig_ref] Figure 4: Correlation of HDGF copy number and expression by microarray analysis and validation... [/fig_ref] -C, correlation 0.81) than patients without HDGF gain. However, ESFT patients could not be divided unambiguously into two groups (see [fig_ref] Figure 4: Correlation of HDGF copy number and expression by microarray analysis and validation... [/fig_ref] and 4B) based on this data. To validate HDGF microarray results, the relative expression levels of HDGF were analysed by TaqMan Low Density arrays in all 42 available ESFT patient samples [fig_ref] Figure 4: Correlation of HDGF copy number and expression by microarray analysis and validation... [/fig_ref]. This analysis confirmed that ESFT patient samples express higher levels of HDGF than normal controls. No statistically significant correlation of HDGF expression with poor clinical outcome could be shown [fig_ref] Figure 4: Correlation of HDGF copy number and expression by microarray analysis and validation... [/fig_ref] and 4E), nor correlation with patient gender, age or location could be shown. Clinical data summary of these 42 ESFT patients included in the analysis can be viewed on Additional file 2.
# Discussion
In this study, we have performed a comprehensive genome wide array CGH analysis of 31 EFST patient samples. Our oligoarray CGH results, the recurrent gains of 1q, 2, 8, and 12, and losses at 9p and 16q that were present in more than 20% of the patient samples, are in agreement with previous ESFT studies by G-banding, conventional CGH [bib_ref] Recurrent gains of 1q, 8 and 12 in the Ewing family of..., Armengol [/bib_ref] [bib_ref] Elomaa I: Clinical correlations of genetic changes by comparative genomic hybridization in..., Tarkkanen [/bib_ref] [bib_ref] CGH analysis of secondary genetic changes in Ewing tumors: correlation with metastatic..., Brisset [/bib_ref] [bib_ref] Dockhorn-Dworniczak B: Genetic imbalances revealed by comparative genomic hybridization in Ewing tumors, Ozaki [/bib_ref] [bib_ref] Molecular cytogenetic parameters in Ewing sarcoma, Amiel [/bib_ref] and array CGH [bib_ref] Array CGH and gene-expression profiling reveals distinct genomic instability patterns associated with..., Ferreira [/bib_ref]. Our array CGH results revealed complex large-scale changes in several samples. Gains of DNA sequences were more prevalent
Outcome of patients with low copy number changes (≤ 3 copy number aberrations) and high copy number changes (> 3 copy number aberrations) than losses and most of the gains affected whole chromosomes or chromosome arms. Further, our analysis showed that patients with low copy number changes (≤ 3 copy number aberrations) showed a significantly better prognosis than patients with a high number of chromosomal alterations, both in terms of event-free and overall survival. [bib_ref] Prognostic impact of deletions at 1p36 and numerical aberrations in Ewing tumors, Hattinger [/bib_ref] [bib_ref] Cytogenetic characterization of Ewing tumors: further update on 20 cases, Udayakumar [/bib_ref]. Our results suggest that duplication of the der(22)t(11;22) is a common event in ESFT. Copy number gain of the fusion gene EWS-FLI1 may further increase the expression of this fusion product and possibly impair the prognosis. Amplification or gain of a chimeric fusion gene is relatively infrequent mechanism in both leukemia and solid tumors. However, rare cases of gain or amplification of the derivative chromosomes or episomes carrying the fusion gene have been reported [bib_ref] Novel formation and amplification of the PAX7-FKHR fusion gene in a case..., Weber-Hall [/bib_ref] [bib_ref] Genetics of dermatofibrosarcoma protuberans family of tumors: from ring chromosomes to tyrosine..., Sirvent [/bib_ref] and gene dosage effect of the fusion gene can improve the tumor growth resulting in more aggres-Chromosomal locations of copy number changes in ESFT patient sample (n = 31) Chromosomal locations of copy number changes in ESFT patient sample (n = 31). The ideogram shows the summary of gains and losses of DNA sequence copy numbers and their frequencies in ESFT tissue samples (n = 31) analyzed by array CGH. Gains (light green) and amplifications (dark green) are shown on the right of each chromosome and losses (red) on the left (number refer to the percentage per band). Chromosomal ideogram was generated using the PROGENETIX software [bib_ref] Progenetix.net: an online repository for molecular cytogenetic aberration data, Baudis [/bib_ref]. Correlation of HDGF copy number and expression by microarray analysis and validation of HDGF expression using RT-PCR partner of EWSR1 is at 20q [bib_ref] Detection and molecular cytogenetic characterization of a novel ring chromosome in a..., Szuhai [/bib_ref]. However, the specific chromosomal region in 20q remained unknown. Based on our results, the translocation partner of EWSR1 on chromosome 20 might reside in the amplification breakpoint, either at 20q11.23 or 20q13.12-q13.2 .
Putative translocation partners of EWSR1 are therefore genes assigned to the breakpoints of these amplifications:RPN2, BLCAP, CDH22, SLC13A3, EYA2, NCOA3, Kua-UEV, and NFATC2. Based on literature, the most interesting candidates are EYA2 (located at 20q13.12), which has been found to function as a transcriptional activator in ovarian cancer cells [bib_ref] Transcriptional coactivator Drosophila eyes absent homologue 2 is up-regulated in epithelial ovarian..., Zhang [/bib_ref] , and NFATC2 (located at 20q13.2), which functions in positive regulation of transcription [bib_ref] Foxp3 interacts with nuclear factor of activated T cells and NF-kappa B..., Bettelli [/bib_ref]. Both EYA2 and NFATC2 are oriented on the amplification breakpoints so that they are in the correct direction for transcription after the possible fusion event. In addition to the chromosome 20, genes on region 8p are interesting as putative fusion partners, since many of these genes are involved in carcinomas and sarcomas. Indeed the region of 8p11.21-p21.2 was gained in patent sample D154. However, the possible involvement of 8p11.21-p21.2 as a location of the translocation partner for EWSR1 was ruled out since this region was not amplified like EWSR1 was. We would assume that the fusion partners would be amplified on the same scale, since translocation is likely to take place before the amplification of the fusion gene.
According to our integrated analysis of array CGH and expression data including 16 ESFT patient samples, we selected as one of the most interesting putative target genes within the common 1q22-qter gain gene HDGF, which has been reported as a putative prognostic marker for several tumor types, e.g., gastrointestinal stromal tumors (GIST) [bib_ref] Expression of hepatoma-derived growth factor is correlated with lymph node metastasis and..., Yamamoto [/bib_ref] [bib_ref] Hepatoma-derived growth factor is a novel prognostic factor for gastrointestinal stromal tumors, Chang [/bib_ref] , hepatocellular carcinoma [bib_ref] Hepatoma-derived growth factor is a novel prognostic factor for hepatocellular carcinoma, Yoshida [/bib_ref] , non-small-cell lung carcinoma [bib_ref] Expression of hepatoma-derived growth factor is a strong prognostic predictor for patients..., Ren [/bib_ref] [bib_ref] Hepatoma-derived growth factor as a prognostic marker in completely resected non-small-cell lung..., Iwasaki [/bib_ref] and pancreatic ductal carcinoma [bib_ref] Hepatoma-derived growth factor is a novel prognostic factor for patients with pancreatic..., Uyama [/bib_ref]. HDGF has been shown to stimulate cell proliferation and growth after nuclear translocation [bib_ref] Nuclear targeting is required for hepatoma-derived growth factor-stimulated mitogenesis in vascular smooth..., Everett [/bib_ref] [bib_ref] Hepatoma-derived growth factor stimulates cell growth after translocation to the nucleus by..., Kishima [/bib_ref] , which makes it a likely target also in ESFT. Furthermore, our preliminary results from an aCGH analysis of ESFT cell lines showed that HDGF was inside the minimal common overlapping area of 1q21.1-q23.1 (Savola et al, unpublished results). Our RT-PCR analysis confirmed that Ewing's sarcoma cells expressed higher levels of HDGF with respect to putative normal controls (CD34 positive cells and normal muscle tissues). However, when we analyzed HDGF expression level correlation with patient survival, no significant association was seen. So HDGF can play a role in the tumorigenesis and tumor progression of EFST, but it shows no prognostic value. However, due to limitations in numbers of patients (n = 42) included in the HDGF expression study, no definitive conclusions of the outcome evaluation of HDGF expression in ESFT can be drawn.
Other interesting target genes pinpointed by integration analysis in 1q include TMEM63A (1q42.12), C1orf107 (1q32.2), HEATR1 (1q43), all relatively unknown genes in their functions and COG2 (1q42.2), gene involved in various Golgi functions [bib_ref] The interaction of two tethering factors, p115 and COG complex, is required..., Sohda [/bib_ref]. In chromosome 8 genes WDR67 (8q24.13) and GSDMDC1 (8q24.3) locating nearby each other and in chromosome 12 DDX47 (12p13.1) and CACNA1C (12p13.3) [bib_ref] Mapping of a human brain voltage-gated calcium channel to human chromosome 12p13-pter, Sun [/bib_ref] [bib_ref] Molecular and functional expression of voltage-operated calcium channels during osteogenic differentiation of..., Zahanich [/bib_ref] are interesting targets for further studies. Also potential oncogenes in ESFT at 12q are WSB2 (12q24.23), which takes part in the intracellular signalling cascades and has shown to be a potential biomarker in colorectal cancer [bib_ref] Differential expression of a novel colorectal cancer differentiation-related gene in colorectal cancer, Li [/bib_ref] , PPHLN1 (12q12), which controls cell cycle regulation by modifying expression of cdc7 involved in progression of DNA replication [bib_ref] Characterization of periphilin, a widespread, highly insoluble nuclear protein and potential constituent..., Kazerounian [/bib_ref] [bib_ref] Overexpression of CR/periphilin downregulates Cdc7 expression and induces S-phase arrest, Kurita [/bib_ref] and KRT79 (12q13.13), a member of human type II keratin gene family. Previously loss of 16q has been shown to be a sign of poor prognosis in ESFT [bib_ref] Dockhorn-Dworniczak B: Genetic imbalances revealed by comparative genomic hybridization in Ewing tumors, Ozaki [/bib_ref]. Our results suggest that the putative target gene within this chromosomal area is HEATR3 (16q12.1) or ANKRD11 (16q24.3), which has been recently identified to interact with p53 and act as a co-activator in the regulatory feedback loop with p53 [bib_ref] Identification of ANKRD11 as a p53 coactivator, Neilsen [/bib_ref]. Functional studies to confirm these results are warranted.
# Conclusion
This study adds new information regarding gene copy number changes and their relation to expression in ESFT providing valuable data for further analysis. In addition, array CGH showed to be efficient in the detection of a putative novel translocation in one patient sample and provided new information about copy number changes of the EWS/FLI1 fusion gene. Therefore we can conclude that array CGH analysis and integrated DNA microarray analysis of global gene expression patterns and gene copy number imbalances is a powerful method to identify novel molecular targets and chromosomal regions of highest interest in ESFT.
[fig] Figure 1: Outcome of patients with low copy number changes (≤ 3 copy number aberrations) and high copy number changes (> 3 copy number aberrations). Kaplan-Meier plots show A) event-free survival and B) overall survival of patients with low copy number changes (≤ 3 copy number aberrations detected in the sample by array CGH) in bold line and with high number copy number changes (> 3 copy number aberrations) in hatched line. [/fig]
[fig] Figure 4: Correlation of HDGF copy number and expression by microarray analysis and validation of HDGF expression using RT-PCR. A) HDGF copy number ratio and B) HDGF expression ratio by microarray analysis in ESFT patient samples (n = 16). The patients in figure A) and B) are in the same order, and labelled according to the codes of the DNA samples. C) Correlation of HDGF copy number and expression ratio (correlation 0.844, P < 0.001, q-value 0.024). D) HDGF expression in ESFT patient samples (n = 42) by RT-PCR analysis, on y-axis refers to log of fold-change in HDGF gene expression and x-axis to patients RNA code number. Kaplan-Meier plots of ESFT patient (n = 42) survival according to D) event-free survival and E) overall survival, patients with high expression of HDGF in bold line and patients with low expression in hatched line. [/fig]
[table] Table 1: Clinical data of the ESFT patients included in array CGH and/or data integration analysis. [/table]
[table] Table 2: Array CGH results (n = 31) in ESFT patient samples arranged by diagnosis. [/table]
[table] Table 3: Putative target genes for tumorigenesis and tumor progression in recurrent copy number changes of the 16 ESFT patients included in the integration analysis. [/table]
[bib_ref] Ewing's sarcoma: diagnostic, prognostic, and therapeutic implications of molecular abnormalities, Burchill [/bib_ref] |
Spatial variation of yield response and fertilizer requirements on regional scale for irrigated rice in China
A large number of on-farm experiments (n = 5556) were collected for the period 2000-2015 from the major rice (Oryza sativa L.) producing regions in China, to study the spatial variability of attainable yield, yield response, relative yield and fertilizer requirements at regional scale, by coupling geographical information system with the Nutrient Expert for Rice decision support system. Results indicated that average attainable yield was 8.8 t ha −1 across all sites, with 18.3% variation. There were large variations in yield response to nitrogen (N), phosphorus (P), and potassium (K) fertilizer application with coefficients of variation of 39.2%, 57.0%, and 53.4%, and the sites of 73.4%, 85.8%, and 87.6% in the study area ranged from 2.0 to 3.0, from 0.7 to 1.3, and from 0.7 to 1.3 t ha −1 , respectively. Mapping the spatial variability of relative yield to N, P, and K indicated that the sites of 78.6%, 92.4%, and 88.7% in the study area ranged from 0.65 to 0.75, from 0.80 to 0.92, and from 0.84 to 0.92, respectively. The high yield response and low relative yield to N and P were mainly located in the Northeast (NE), Northwest (NW), and north of the Middle and Lower Reaches of Yangtze River (MLYR) regions. The spatial distribution of N, P, and K fertilizer requirements ranged 140-160 kg N ha −1 , 50-70 kg P 2 o 5 ha −1 and 35-65 kg K 2 o ha −1 which accounted for 66.4%, 85.5% and 73.0% of sites in the study area, respectively. this study analyzed the spatial heterogeneity of attainable yield, soil nutrient supply capacity and nutrient requirements based on a large database at regional or national scale by means of geographical information systems and fertilizer recommendation systems, which provided a useful tool to manage natural resources, increase efficiency and productivity, and minimize environmental risk.Rice (Oryza sativa L.) yield has greatly increased in the past two decades and plays a vital role in guaranteeing food security and promoting economic development in China. The rice area harvested and production in China accounted for 19.1% and 28.5%, respectively, of the worlds' total 1 ; the equilibrium of rice supply and demand has great impact on the stability of the world grain market. However, there are unprecedented pressures on the current agricultural and natural resources to meet the increasing food demand. The problems of resource shortages and environmental pollution bring great challenges to sustainable development of agricultural production in China 2,3 .The rice yield has increased continuously to meet population growth while the area of arable land has decreased, because of variety improvement, fertilizer application, and optimization of cultivation techniques and management measures 4-6 . However, low fertilizer use efficiency is an outstanding and serious problem in current rice production. Studies showed that nitrogen (N) use efficiency was less than 30%, and phosphorus (P) recovery efficiency was below 20% for most famers' fertilization practices in many regions of China 7,8 . Low nutrient use efficiency means that more nutrients accumulated in the soil or were lost to the environment. Chen et al. 3 study showed that the N surplus was 46-280 kg N ha −1 in farmers' fields, and the Olsen P concentration in the top layer (0-20 cm) has exceeded 20 mg kg −1 in many regions for rice systems in China 9 . These high nutrient 1 institute of Agricultural Resources and Regional Planning, chinese Academy of Agricultural Sciences (cAAS),
www.nature.com/scientificreports www.nature.com/scientificreports/ concentrations are also implicated in the substantial greenhouse gas emissions and serious water eutrophication in rivers and lakes [bib_ref] Nitrogen fertilizer induced greenhouse gas emissions in China, Liu [/bib_ref] [bib_ref] An experiment for the world, Zhang [/bib_ref].
As an important component of agricultural production, optimized or balanced fertilization can improve crop yields, save resources and protect the environment. Many methods have been adopted to optimize fertilization rates and improve nutrient management to increase nutrient use efficiency and reduce environmental pollution. These include: fertilizer requirements based on the interactions of indigenous nutrient supply, yield target and crop nutrient demand [bib_ref] Optimizing nutrient management strategies for rice-wheat system in the Indo-Gangetic Plains of..., Singh [/bib_ref] , fertilizer management according to alternate wetting and drying irrigation [bib_ref] Alternate wetting and drying irrigation and controlled-release nitrogen fertilizer in late-season rice...., Ye [/bib_ref] , optimizing nutrient management strategies in combination with leaf color chart and optimum fertilizer placement [bib_ref] Integrating best management practices for rice with farmers' crop management techniques: A..., Alam [/bib_ref] , site-specific nutrient management with water management and optimized transplanting density [bib_ref] A preliminary precision rice management system for increasing both grain yield and..., Zhao [/bib_ref]. These methods have potentially increased rice yield by 10% and nutrient use efficiency by 30% over farmers' practices.
At present, fertilizer recommendations based on individual or few data cannot meet the demand of the current intensive agricultural production, but the spatial variability in climate, soil fertility and management practices must be considered when developing the most cost efficient nutrient management strategy. However, it is difficult to develop a variable rate technique by studying the spatial variability of soil nutrients in the family responsibility system in China and to communicate the great need to establish a fertilizer recommendation method based on crop responses [bib_ref] Methodology of fertilizer recommendation based on yield response and agronomic efficiency for..., Xu [/bib_ref]. Geostatistics has been adopted to assess spatial variability of grain yield and nutrient balance at regional scale to develop reasonable nutrient management [bib_ref] The yield gap of global grain production: a spatial analysis, Neumann [/bib_ref] [bib_ref] Spatial variability of the nutrient balance and related NPSP risk analysis for..., Wang [/bib_ref] , which will contribute to realize the rational allocation and efficient use of fertilizer resources [bib_ref] Spatial variability of soil nutrients and site-specific nutrient management in the, Jin [/bib_ref].
Rice production must increase to meet future food requirements amid strong competition for limited resources and environmental pollution. Understanding the distribution of attainable yield, yield response to fertilizer application and soil nutrient supply capacity are necessary in developing fertilizer recommendations, which will help to reform agricultural policies aimed at the characteristics of agricultural production in China. Therefore, we combined the Nutrient Expert for Rice system and geostatistical analysis to (1) analyze the current rice attainable yield distribution; (2) map the distribution of yield response and relative yield to fertilizer application; and (3) explore spatial variation of fertilizer requirements at regional scale.
# Materials and methods
study area and data source. The database used in this study was obtained from 5,556 field experiments conducted by the International Plant Nutrient Institute China Program and the results of these studies were published in scientific journals from 2000 to 2015, which covered all the main rice-producing regions including several distinct agroecosystems in China [fig_ref] Figure 1: Geographical distribution of the studied locations and rice experiment sites in the... [/fig_ref]. In the Wed of science, the following keywords were used to search the literature: rice, yield, nutrient use efficiency, nutrient uptake. The data in the articles must be from filed experiment under carefully controlled conditions, and harvest index less than 0.4 was excluded in building Nutrient Expert system because these data were treated as crop suffering stress from water, biotic or abiotic stress [bib_ref] A system for quantitative evaluation of the fertility of tropical soils (QUEFTS), Janssen [/bib_ref]. The four rice types in terms of season (early, middle, late and single-season rice) were farmed using intensive cultivation methods. The experiments included optimum fertilization experiments, long-term field experiments, and different fertilizer application rates, varieties, water and fertilizer management across six regions of China. Single-season rice is mainly cultivated in a mono-rice system in northeast (NE) and northwest China (NW) where a cool-temperate climate is dominant, and is grown from early or middle May to middle or late September. Middle rice is mainly cultivated in the middle and lower reaches of the Yangtze River (MLYR, such as Hubei, An'hui, Jiangsu and Zhejiang provinces), southwest (SW), and some provinces in north-central China (NC, Henan and Shandong provinces). These regions are dominated by temperature and subtropical climates, and rice crops are rotated with winter wheat, rape, or other crops. The rice crops in these regions are grown from late May or early June to late September or early October. Early and late rice are mainly concentrated in south China (SC) and in some southern provinces of the MLYR (Jiangxi, Hunan, and south of Hubei and Zhejiang provinces) that have a subtropical and tropical climate, where they are grown in an early/late rice rotation system, early rice www.nature.com/scientificreports www.nature.com/scientificreports/ is grown from late March or early April to early or middle July, late rice is grown from late July or early August to late October or middle November.
Nutrient Expert for Rice decision support system. The Nutrient Expert for Rice decision support system was used to recommend fertilizer rate for each experimental site in the current study. The principle of the method is based on agronomic efficiency (yield increase per unit of fertilizer) and yield response, which combine with the Quantitative Evaluation of the Fertility of Tropical Soils (QUEFTS) model [bib_ref] A system for quantitative evaluation of the fertility of tropical soils (QUEFTS), Janssen [/bib_ref] [bib_ref] Estimating nutrient uptake requirements for rice in China, Xu [/bib_ref] and "4R" principles (applying the right source of nutrient at the right rate and the right time in the right place) to determine fertilizer rate and develop nutrient management practices. Agronomic efficiency is related to yield response and fertilizer application rate, and determined by optimal amounts of added nutrients to provide a reasonable value. The relationships among yield response, agronomic efficiency, relative yield and soil indigenous nutrient supply were built to support fertilizer recommendation. The N recommendation was based on yield response to fertilizer and agronomic efficiency of applied N. While P and potassium (K) requirements considered target yield, yield response, nutrient balances and residual nutrients from the previous crop. The fertilization principle and parameterization of the Nutrient Expert system have been described in detail and reported by Pampolino et al. [bib_ref] Development approach and evaluation of the Nutrient Expert software for nutrient management..., Pampolino [/bib_ref] and Xu et al. [bib_ref] Methodology of fertilizer recommendation based on yield response and agronomic efficiency for..., Xu [/bib_ref] [bib_ref] Spatial variation of attainable yield and fertilizer requirements for maize at the..., Xu [/bib_ref]. Nutrient Expert system combines all the steps and guidelines in site-specific nutrient management into simple software, and gives a dynamic fertilizer management that are tailored to the specific field due to different environmental conditions, management practices from current and previous residual nutrient, in order to take full advantage of soil nutrient. In the current study, the system assumed that straw was not returned into soil for single-season rice because of low temperature, while all straw was incorporated into soil after harvest for early, middle and late rice planting regions.
Analysis. An optical remote sensing image based on terrestrial ecosystems was obtained to identify the rice fields from land cover classification. The software of GS + 5.3 (Plainwell, Michigan, USA) and ArcGIS 9.3 software (ESRI, Redlands, USA) were adopted to map the distribution of attainable yield, yield response, relative yield and fertilizer requirements across the study plots using semivariogram models in conjunction with kriging interpolation. For the early-late rice rotation system and the multiple-year experiments, the average value was adopted at the same site. In this study, attainable yield is the maximum yield derived in field experiments, which is determined from field trials or local experts' experience for a geographic region or growing environment according to site characteristics and farmers' actual yield; yield response to fertilizer N, P, and K is the yield gap between full NPK plots and omission plots in which one of the nutrients is omitted, yield response is determined by attainable yield and soil fertility and inversely related to the soil fertility, the higher soil fertility, the lower yield response; relative yield is defined as the ratio between nutrient-limited yield from the nutrient omission plot and attainable yield from the ample NPK plot, the percentile of relative yield was used to estimate yield response when yield response data are unavailable [bib_ref] Development approach and evaluation of the Nutrient Expert software for nutrient management..., Pampolino [/bib_ref]
[formula] K K K K K K K S = × + × × × + × − × × . [/formula]
Where the unit of fertilizer N, P, and K are fertilizer N, P 2 O 5 , and K 2 O requirements (kg ha −1 ), respectively, YR is yield response (kg ha −1 ), AE is agronomic efficiency (kg kg −1 ), RIE is nutrient uptake requirement per ton of grain yield (kg ha −1 ), RE is recovery efficiency to nutrient application (%), Ya is attainable yield (kg ha −1 ), HI is harvest index, X G % and X S % are the nutrient return proportion of grain and straw, respectively, and 2.292 and 1.205 are conversion constant for P and K, respectively.
# Results
Distribution of attainable yield. The attainable yield showed an obvious spatial distribution in regions and plant types. The attainable yield in middle and single-season rice planting regions was higher than in early and late rice planting regions [fig_ref] Figure 2: Distribution of rice attainable yield [/fig_ref]. On average, attainable yield was 8.8 t ha −1 across all sites, with 18.3% variation. The attainable yield values were 7.5, 9.3, 7.9, and 9.4 t ha −1 for early, middle, late and single-season rice, with 18.6%, 15.8%, 18.9%, and 15.1% variation, respectively [fig_ref] Figure 2: Distribution of rice attainable yield [/fig_ref]. Sites with attainable yield of <8 t ha −1 accounted for 22.4% of the study area and were mainly distributed in double-season rice planting regions, such as the south MLYR and SC; of these, 1.7% of sites in the study area had attainable yield of <7 t ha −1 . Sites with attainable yield in the range of 8 to 9 t ha −1 accounted for 42.0% of the study area, and were mainly distributed in the central MLYR, north SW and northwest NE, and east SC. Sites with attainable yield in the range of 9 to 10 t ha −1 accounted for 30.3% of the study area, and were mostly in the north MLYR, and SW and NE regions. We found that 5.3% of sites in the study area exceeded 10 t ha −1 across all regions, mainly distributed in the southeast SW, central NW and south NE regions. spatial distribution of yield response and relative yield. On average, the N yield response was 2.7 t ha −1 across all sites, with 39.2% variation. The coefficients of variation (CVs) of N yield response were 37.3%, 36.8%, 37.9%, and 37.3% for early, middle, late, and single-season rice, respectively [fig_ref] Figure 3: Distribution of yield response [/fig_ref]. The yield response in middle and single-season rice planting regions was higher than that in early and late rice planting regions, except for north SW. High N yield response (more than 3.0 t ha −1 ) was mostly observed in northeast MLYR, central NW www.nature.com/scientificreports www.nature.com/scientificreports/ and NE regions, which accounted for 18.7% of the study area. Low N yield response of <2.5 t ha −1 accounted for 47.1% of the study area and was mainly distributed in double-season rice planting regions and in the north SW. High N yield response means low N relative yield; sites with N relative yield of <0.7 were mostly located in the NE region, central NW, and northeast MLYR, with N relative yield values of <0.65 for 9.2% of sites in the study area [fig_ref] Figure 3: Distribution of yield response [/fig_ref]. In the current study, the average N relative yield was 0.70 with 13.7% variation. The majority of N relative yield values were between 0.65 and 0.75, which accounted for 78.6% of the study area. The N relative yield was above 0.75 for about 12.2% of sites in the study area and was mainly located in the SW and SC regions.
The average P yield response was 0.9 t ha −1 across all sites, with 57.0% variation, and 1.0, 0.9, 0.7, and 1.0 t ha −1 for early, middle, late, and single-season rice, with 56.9%, 53.6%, 62.7% and 54.5% CVs, respectively. Sites with P yield response values of between 0.7 and 1.3 t ha −1 accounted for 85.8% of the study area [fig_ref] Figure 4: Distribution of yield response [/fig_ref]. Sites with low P yield response (<0.7 t ha −1 ) accounted for 11.1% of the study area, and were mainly distributed in double-season rice planting regions, such as south and east MLYR and northeast SC regions. Sites with high P yield response of >1.3 t ha −1 were mainly located in the single-season rice planting regions. Sites with P relative yield values of between 0.88 and 0.92 accounted for 71.2% of the study area [fig_ref] Figure 4: Distribution of yield response [/fig_ref]. Sites with P relative yield values of >0.92 accounted for 7.1% of the study area and were mainly located in east MLYR region. Sites with P relative yield of <0.88 accounted for 21.7% of the study area and were mostly located in the middle and single-season rice planting regions. www.nature.com/scientificreports www.nature.com/scientificreports/ The average K yield response was 1.0 t ha −1 across all sites with 53.4% variation; and 87.6% of sites in the study area had K yield response values in the range of 0.7-1.3 t ha −1 [fig_ref] Figure 5: Distribution of yield response [/fig_ref]. The K yield response values for early, middle, late and single-season rice were similar at 1.0, 1.1, 0.9, and 0.9 t ha −1 , with CVs of 54.1%, 51.1%, 53.3%, and 54.3%, respectively. Higher values of K yield response (>1 t ha −1 ) were mainly located in the northeast MLYR and southeast SW, which corresponded to the lower K relative yield in these regions. Sites with K relative yield values of <0.84 accounted for 4.0% of the study area [fig_ref] Figure 5: Distribution of yield response [/fig_ref]. Sites with K relative yield values of 0.84-0.92 accounted for 88.7% of the study area. Sites with K relative yield of >0.92 accounted for 7.3% of the study area and were mainly located in the middle and single-season rice planting regions, such as central NW, central MLYR and north SW. spatial distribution of nutrient requirements. Nitrogen fertilizer requirement showed an obvious spatial distribution between regions, and middle and single-season rice had higher N fertilizer requirement than early and late rice [fig_ref] Figure 6: Distribution of fertilizer N [/fig_ref]. Of the study areas, the sites of 20.9% had N fertilizer requirements higher than 160 kg N ha −1 , mainly distributed in the NE, north MLYR and NW regions; the high N (>180 kg N ha −1 ) was needed for achieving high yield in some regions because of high attainable yield and N yield response [fig_ref] Figure 2: Distribution of rice attainable yield [/fig_ref]. Fertilizer N requirements of 140-160 kg ha −1 accounted for 66.4% of the study area, mainly located in the MLYR and SW, and north NE. In some early-late rice planting regions, a low N rate (130-140 kg ha −1 or less) in each season could meet the crop requirement, for example south MLYR and middle SC, which accounted for 12.8% of the study area. www.nature.com/scientificreports www.nature.com/scientificreports/ There was strong spatial variability in P fertilizer requirement among regions, with a CV of 25.1% [fig_ref] Figure 6: Distribution of fertilizer N [/fig_ref] , and the CVs of P fertilizer requirement were 27.6%, 22.4%, 25.1%, and 23.2% for early, middle, late, and single-season rice, respectively. Most of the study areas (85.5%) were within the range of 50-70 kg P 2 O 5 ha −1 , mainly in the MLYR, SC, north SW, and north NE regions. In some middle and single-season rice planting regions, such as southeast SW, central NE, and north MLYR, 70 kg P 2 O 5 ha −1 or more was needed to meet the crop requirement, which accounted for 6.7% of sites in the study area. For the remaining 7.8% of the study area, 50 kg P 2 O 5 ha −1 was enough; these sites were mainly distributed in the early-late rice planting regions, such as south MLYR and south SC regions.
There was large variation in K fertilizer requirement among regions, with a CV of 38.8% [fig_ref] Figure 6: Distribution of fertilizer N [/fig_ref]. The CVs of K fertilizer requirement were 35.4%, 31.4%, 32.9%, and 20.2% for early, middle, late, and single-season rice, respectively. The K fertilizer requirements in middle and single-season rice were higher than for early and late rice, especially in NE and NW regions, with 12.5% of sites in the study area requiring >80 kg K 2 O ha −1 . The average K fertilizer requirements needed to maintain the soil K balance in single-season rice planting regions were 99 kg K 2 O ha −1 . In the SW region, the K fertilizer requirement was mainly within the range of 65-80 kg K 2 O ha −1 , except in the north SW. In 73% of sites in the study area, the K fertilizer application rates were <65 kg K 2 O ha −1 ; these sites were mainly in the MLYR and SC, and north SW regions.
# Discussion
The rice yield has greatly increased worldwide in the past decades because of a series of agricultural practices, such as improved varieties, soil fertility management, and water, weed, pest and disease management 4,6 . The worldwide rice production could still be increased by 47% if 100% of attainable yield could be achieved [bib_ref] Closing yield gaps through nutrient and water management, Mueller [/bib_ref]. China accounts for more than one fifth of world rice production mainly because of improved productivity per hectare; the unit yield increased by 1 t ha −1 in the past two decades 1 . However, the decreasing rate of rice yield increase is largely interpreted as a decreasing relative contribution of fertilizer 25 , and concerns have been raised about the synergetic development of sustainable rice production and environmental pollution [bib_ref] Producing more grain with lower environmental costs, Chen [/bib_ref] [bib_ref] Grain yield, water productivity and CH 4 emission of irrigated rice in..., Liang [/bib_ref]. There was large spatial variability in attainable yield due to differences in rice varieties, plant type of season and soil fertility among regions, and multiple-year/site experiments were collected to attain reasonable target yield when estimating nutrient requirements. In the current study, the CV of attainable yield was 18.3%, and the average middle and single-season rice yields were higher than those of early and late rice [fig_ref] Figure 2: Distribution of rice attainable yield [/fig_ref] , the one of reasons is because longer growth duration (20-30 more days) and larger day/night temperature differences contributing to higher dry matter accumulation in middle and single-season rice than in early and late rice.
Excessive and imbalanced fertilizer application has led to very serious environmental problems, such as runoff, water eutrophication, and greenhouse gas emissions 2,3 . Integrated technologies and fertilizer recommendation methods should be more carefully selected and deployed to continue to increase food production and avoid negative effects on the environment 11 , such as integrative crop management including optimize nutrient input at different growth-stage, plant density and irrigation management, which increased grain yield and recovery efficiency by 10% and 20%, respectively 27 . Multi-split topdressing, controlled-release N fertilizer application, and integration of water and fertilizer management can all significantly enhance grain yield and N uptake [bib_ref] Crop management based on multi-split topdressing enhances grain yield and nitrogen use..., Chen [/bib_ref] [bib_ref] Long-term effects of controlled release urea application on crop yields and soil..., Geng [/bib_ref] [bib_ref] Effects of integrated high-efficiency practice versus conventional practice on rice yield and..., Cao [/bib_ref]. However, a reasonable parameter is necessary to express soil nutrient supply capacity to obtain a reasonable amount of fertilizer, because there are challenges in terms of a shortage of agricultural workers and high soil testing costs.
Fertilizer nutrients applied into soils will be eventually absorbed by plants and can be reflected by crop yield increase. The yield response to nutrient application is an important parameter to express soil nutrient supply capacity, and the 25th percentile, median, and 75th percentile of relative yield are used as coefficients to estimate yield response when yield response data for a particular field are not available [bib_ref] Methodology of fertilizer recommendation based on yield response and agronomic efficiency for..., Xu [/bib_ref] , and represent soil indigenous supply classes of 'low' , 'average' , and 'high' , respectively. High yield response corresponds to low relative yield and low soil nutrient supply capacity, and means more fertilizer was needed. The relative yield values to soil indigenous supply classes were determined by different databases according to rice plant types (early, middle, late and single-season rice) in the Nutrient Expert for Rice system. The differences in attainable yield and yield response www.nature.com/scientificreports www.nature.com/scientificreports/ resulted in the variation of fertilizer requirements between fields. The large CVs of the yield response in the current study indicated that specific fertilizer recommendations should be conducted according to the differences in climate, yield levels, and nutrient uptake [bib_ref] Estimating nutrient uptake requirements for rice in China, Xu [/bib_ref]. Understanding and analyzing spatial heterogeneity of yield response and relative yield will help to accurately determine fertilizer application rate.
Many studies have suggested that high yield can be attained under low fertilizer application rate or optimal nutrient management practices when compared with farmers' standard practices [bib_ref] Crop management based on multi-split topdressing enhances grain yield and nitrogen use..., Chen [/bib_ref] [bib_ref] Optimizing nutrient management strategies for rice-wheat system in the Indo-Gangetic Plains of..., Singh [/bib_ref] [bib_ref] Optimizing nitrogen supply increases rice yield and nitrogen use efficiency by regulating..., Sui [/bib_ref]. In the current study, high fertilizer application rates were mainly located in NE, north MLYR and central NW, in which the high N and P fertilizer application rates were mainly related to high yield, high yield response and low relative yield. High K fertilizer application rates in the single-season rice planting regions were mainly because straw was not returned to the fields, as 84% of K is concentrated in the straw [bib_ref] Estimating nutrient uptake requirements for rice in China, Xu [/bib_ref]. If the straw was returned into the soil for single-season rice planting regions, K fertilizer application rates could be reduced by 30% using Nutrient Expert for Rice. In addition, the residual nutrient from previous season also was considered to determine fertilizer requirements in the Nutrient Expert for Rice system.
With the continuous dynamic optimization of the Nutrient Expert for Rice system for each cropping season, the method increased grain yield and N recovery efficiency [bib_ref] Methodology of fertilizer recommendation based on yield response and agronomic efficiency for..., Xu [/bib_ref]. The policy support from the government plays a crucial role to control fertilizer consumption, such as "National Key Research and Development Program" was implemented from 2016 to develop high-efficient fertilizer recommendation method and establish nutrient-limits standard, in order to achieve "zero-growth" by the year of 2020. The fertilizer recommendation system combined with geographical information systems provides useful information for making agricultural policies or strategies, obtaining reasonable fertilizer distribution and supply, and ensuring food security [bib_ref] Yield gap analysis with local to global relevance-A review, Van Ittersum [/bib_ref]. These methods require holistic field information, including the spatial distribution of grain yield, soil nutrient supply capacity and nutrient requirements, so as to achieve sustainable agricultural development. Traditional training approach has a limited role in reducing farmers' fertilizer use 32 , integrated technologies, substantial investment, policy interventions and formulation of legislation should be conducive in field-guidance to resolve the contradiction between food security and environmental protection in agricultural production systems.
# Conclusions
Chinese agricultural research must keep pace with the growing demands for food, high-efficiency utilization of resources and environmental protection. Coupling geographical information systems and fertilizer recommendation systems based on strong agronomic foundations is a sound approach, to provide actionable guidance in developing agricultural strategies and policies. In the current study, the spatial distribution of grain yield, yield response to nutrient application and relative yield were analyzed using geographical information systems based on a large number of field experiments; the nutrient requirements at regional scale were mapped and combined with the Nutrient Expert for Rice decision support system across major rice producing regions. We found great spatial variations in grain yield, yield response and relative yield between and within regions. The CVs of yield response to N, P, and K fertilizer application reached 39.2%, 57.0%, and 53.4%, respectively, across all study regions, which resulted in the spatial variation of nutrient requirements. The results of this study will help to manage natural resources, increase nutrient efficiency and productivity, and minimize environmental risk at regional or national scale.
[fig] Figure 1: Geographical distribution of the studied locations and rice experiment sites in the six regions of China. (2019) 9:3589 | https://doi.org/10.1038/s41598-019-40367-2 [/fig]
[fig] Figure 2: Distribution of rice attainable yield (a) in China, and the box plots of grain yield for four rice planting types in this study (b), solid and dashed lines indicate median and mean, respectively. The box boundaries indicate the upper and lower quartiles, the whisker caps indicate 90th and 10th percentiles, and the circles indicate the 95th and 5th percentiles. [/fig]
[fig] Figure 3: Distribution of yield response (a) and relative yield (b) to nitrogen (N) fertilizer application for rice in China. Scientific RepoRts | (2019) 9:3589 | https://doi.org/10.1038/s41598-019-40367-2 [/fig]
[fig] Figure 4: Distribution of yield response (a) and relative yield (b) to phosphorus (P) fertilizer application for rice in China. [/fig]
[fig] Figure 5: Distribution of yield response (a) and relative yield (b) to potassium (K) fertilizer application for rice in China. [/fig]
[fig] Figure 6: Distribution of fertilizer N (a), P 2 O 5 (b) and K 2 O (c) requirements for rice in China. (2019) 9:3589 | https://doi.org/10.1038/s41598-019-40367-2 [/fig]
|
The current state of navigation in robotic spine surgery
The advent and widespread adoption of pedicle screw instrumentation prompted the need for image guidance in spine surgery to improve accuracy and safety. Although the conventional method, fluoroscopy, is readily available and inexpensive, concerns regarding radiation exposure and the drive to provide better visual guidance spurred the development of computer-assisted navigation.Contemporaneously, a non-navigated robotic guidance platform was also introduced as a competing modality for pedicle screw placement. Although the robot could provide high precision trajectory guidance by restricting four of the six degrees of freedom (DOF), the lack of real-time depth control and high capital acquisition cost diminished its popularity, while computer-assisted navigation platforms became increasingly sophisticated and accepted. The recent integration of real-time 3D navigation with robotic platforms has resulted in a resurgence of interest in robotics in spine surgery with the recent introduction of numerous navigated robotic platforms. The currently available navigated robotic spine surgery platforms include the ROSA Spine Robot (Zimmer ), Mazor X spine robot (Medtronic Navigation Louisville, CO;Medtronic Spine, Memphis, TN; formerly Mazor Robotics, Caesarea, Israel) and TiRobot (TINAVI Medical Technologies, Beijing, China). Here we provide an overview of these navigated spine robotic platforms, existing applications, and potential future avenues of implementation.
# Introduction
Pedicle screw instrumentation has remained the gold standard technique for spinal fixation since its popularization by Roy Camille in the 1970s [bib_ref] Internal Fixation of the Lumbar Spine with Pedicle Screw Plating, Roy-Camille [/bib_ref]. Despite the benefit of pedicle screw instrumentation, misplaced pedicle screws can result in serious neurovascular injury. Consequently, modalities of image-guidance in spine surgery have evolved over the years [bib_ref] Computer-Aided Fixation of Spinal Implants, Nolte [/bib_ref] [bib_ref] Computer-aided navigation in neurosurgery, Grunert [/bib_ref] [bib_ref] Next-Generation Robotic Spine Surgery: First Report on Feasibility, Safety, and Learning Curve, Khan [/bib_ref] [bib_ref] Comparing Next-Generation Robotic Technology with 3-Dimensional Computed Tomography Navigation Technology for the..., Khan [/bib_ref] [bib_ref] Frameless stereotaxis for the insertion of lumbar pedicle screws, Murphy [/bib_ref] [bib_ref] Application of frameless stereotaxy to pedicle screw fixation of the spine, Kalfas [/bib_ref] [bib_ref] Accuracy of pedicle screw placement: a systematic review of prospective in vivo..., Gelalis [/bib_ref]. Although the precision of imageguided screw placement is high [bib_ref] Accuracy of pedicle screw placement: a systematic review of prospective in vivo..., Gelalis [/bib_ref] [bib_ref] Functional outcome of computer-assisted spinal pedicle screw placement: a systematic review and..., Verma [/bib_ref] [bib_ref] Image-guided pedicle screw insertion accuracy: a meta-analysis, Tian [/bib_ref] , the constant pursuit to improve clinical outcomes has prompted the introduction of robotics into the field of spine surgery [bib_ref] Assessment of Pedicle Screw Placement Accuracy, Procedure Time, and Radiation Exposure Using..., Lieberman [/bib_ref] [bib_ref] Robotic-assisted pedicle screw placement: lessons learned from the first 102 patients, Hu [/bib_ref] [bib_ref] Perioperative course and accuracy of screw positioning in conventional, open robotic-guided and..., Kantelhardt [/bib_ref].
When utilizing robots for surgical procedures, surgeons can interface with robots in three ways. In the supervisory-control system, a surgeon plans the procedure and observes the execution of the plan by an autonomously acting robot. An example is the Cyberknife (Accuray, Sunnyvale, California, USA) radiosurgery robot employed by neurosurgeons for the treatment of tumors of the central nervous system [bib_ref] Image-guided robotic radiosurgery, Adler [/bib_ref]. The DaVinci (Intuitive, Sunnyvale, California, USA) robot is an example of a telesurgical robot system. These systems are characterized by Review Article on Current State of Intraoperative Imaging the translation of direct real-time surgeon input to robot output [bib_ref] Remote operative urology using a surgical telemanipulator system: preliminary observations, Bowersox [/bib_ref]. Finally, a shared-control system allows the robot to function in concert with the surgeon, who remains in primary control of the procedure. Currently, all FDA approved, commercially available spinal robotic systems are shared-control systems. These systems are theoretically capable of reducing human error through increased precision in execution, indefatigability, motion scaling, and tremor filtration via mechanical actuation.
The first shared-control robot designed for use in spine surgery, SpineAssist (Mazor Surgical Technologies, Caesarea, Israel), emerged in the early 2000s. The SpineAssist was developed contemporaneously with computer-assisted spinal navigation systems -both of which were promoted as a solution to the unsatisfactory screw malposition rate and increased radiation exposure associated with minimally invasive spinal instrumentation techniques [bib_ref] Assessment of Pedicle Screw Placement Accuracy, Procedure Time, and Radiation Exposure Using..., Lieberman [/bib_ref] [bib_ref] Robotic-assisted pedicle screw placement: lessons learned from the first 102 patients, Hu [/bib_ref] [bib_ref] Computer tomography assessment of pedicle screw insertion in percutaneous posterior transpedicular stabilization, Schizas [/bib_ref]. There are many reports of successful implementation of this robotic platform with evidence of increased screw accuracy and reduction in radiation exposure relative to traditional fluoroscopically-guided freehand technique [bib_ref] Assessment of Pedicle Screw Placement Accuracy, Procedure Time, and Radiation Exposure Using..., Lieberman [/bib_ref] [bib_ref] Robotic-assisted pedicle screw placement: lessons learned from the first 102 patients, Hu [/bib_ref] [bib_ref] Perioperative course and accuracy of screw positioning in conventional, open robotic-guided and..., Kantelhardt [/bib_ref]. Irrespective of radiographic results, these studies showed that clinical outcomes were not significantly impacted with the use of a robot compared to the fluoroscopically-guided freehand technique. This clinical equipoise combined with the high capital acquisition cost of the system stagnated its widespread adoption [bib_ref] Image-Guided Navigation and Robotics in Spine Surgery, Kochanski [/bib_ref].
Historically, robots required a certain level of trust from surgeons as the robotic systems did not provide any realtime visual feedback for instrument localization and depth gauging. However, non-robotic 3D computer assisted navigation (CAN) in spine surgery [bib_ref] Accuracy of pedicle screw placement: a systematic review of prospective in vivo..., Gelalis [/bib_ref] [bib_ref] Functional outcome of computer-assisted spinal pedicle screw placement: a systematic review and..., Verma [/bib_ref] [bib_ref] Image-guided pedicle screw insertion accuracy: a meta-analysis, Tian [/bib_ref] [bib_ref] Accuracy of Pedicle Screw Insertion Among 3 Image-Guided Navigation Systems: Systematic Review..., Du [/bib_ref] [bib_ref] Pedicle screw insertion accuracy with different assisted methods: a systematic review and..., Tian [/bib_ref] has led to the integration of computer-assisted navigation with modern spine robot platforms. This new combination of technologies appears to be driving the resurgence of spinal robotics. In this article we review the contemporary 3D navigation enabled robotic spine platforms and their current and future applications.
## Current integrated navigation and robotics platforms
The ROSA Spine Robot (Zimmer Biomet Robotics formerly Medtech SA, Montpellier, France) was launched in 2011 in Europe and has been FDA approved for spine surgery in the USA since 2016 . It is a fully automated robotic arm with 6 degrees of freedom (DOF) deployed from a floor-mounted base station that also houses an integrated CAN interface. A separate optical camera is used for registration and real-time tracking. A removable fiducial array is attached to the robot arm for the registration process. A standard posterior superior iliac spine (PSIS) or spinous process mounted DRB is used for anatomic reference. Pre-operative or intra-operative images are acquired and registered to the patient and robot. The base station is anchored and the robotic arm is deployed. The end effector allows drilling and subsequent insertion of a guide wire. Pedicle tapping and subsequent placement of the screw can all be performed with optional CAN assistance using a tracker array that is affixed to the instrument [bib_ref] Robot-assisted spine surgery: feasibility study through a prospective casematched analysis, Lonjon [/bib_ref] [bib_ref] Evaluation of the ROSATM Spine robot for minimally invasive surgical procedures, Lefranc [/bib_ref] [bib_ref] Evaluation of Screw Placement Accuracy in Circumferential Lumbar Arthrodesis Using Robotic Assistance..., Chenin [/bib_ref]. The ROSA spine robot is able to move with patient respiration and surgical manipulation as the optical camera constantly tracks the location of the patient DRB relative to a reference array mounted on the base station next to the arm. The ROSA platform received an FDA approved upgrade in March of 2019, and is now known as Rosa ONE. The upgrade includes a fully integrated navigation interface and Zimmer Biomet instrumentation package and is adaptable across cranial, spinal, and orthopedic knee arthroplasty procedures. The ExcelsiusGPS ® (Globus Medical, Inc., Audubon, PA, USA) launched in late 2017 and at the time was advertised as the first US-built robotic surgical spine and fully integrated navigation platform with real-time instrument tracking-allowing pedicle screw placement without K-wires . The system is anchored on a floor-mounted base station that supports the CAN interface as well as the robotic arm itself. The robotic end effector employs small wirelessly powered LED markers rather than larger standard reflective ball markers, and the instruments that pass through and are held by the end effector also have their own specific tracking arrays [fig_ref] Figure 3: Globus Excelsius GPS end effector with navigated instrument [/fig_ref]. A standard DRB is also mounted to the patient's PSIS or a spinous process. A separate surveillance DRB is bone mounted as well. An optical camera is used for registration and tracking. Intraoperative CT is the imaging modality of choice for registration for ease of work flow, although the robot is capable of fluoroscopic registration utilizing a pre-operative CT scan. The system can also be used for 3D CAN without the robotic arm, if desired [bib_ref] Ensuring navigation integrity using robotics in spine surgery, Crawford [/bib_ref].
The Mazor X spine robot (Medtronic Navigation Louisville, CO, USA; Medtronic Spine, Memphis, TN, USA; formerly Mazor Robotics, Caesarea, Israel) initially launched commercially in October 2016 [fig_ref] Figure 4: Mazor X stealth edition platform [/fig_ref]. The X represented an iterative improvement over the Renaissance (Mazor Robotics, Caesarea, Israel) as the robotic arm was upgraded to be fully automated and no longer requires a patient mounted track for deployment. Instead, the X mounts to both the surgical bed frame with rail adaptor attachments, and to the patient with a bone connection bridge from the robotic arm to a pin placed in the PSIS or spinous process. The X has 3 cameras on the arm itself, which first detect and volumetrically defines the operative field in order to prevent collision with the patient. A fiducial marker is then temporarily attached to the arm and registered either to the pre-operative CT scan with AP and oblique fluoroscopic images, or by intra-operative O-arm cone beam CT scan. The robot is then able to guide pedicle cannulation and K-wire insertion. The pedicle is tapped and screw is placed manually with fluoroscopy for depth guidance. Following the acquisition by Medtronic in December of 2018, the co-adaptation of Stealth Navigation and the X Platform was announced and upgrades were launched in early 2019. The O-arm acquisition dataset is reformatted separately by both the X Robotic guidance processing unit and the integrated Stealth Navigation software. This allows parallel integration of fully navigated Medtronic instruments for real-time instrument position feedback. The integrated technical hardware adds a new optical tracking camera as well as a new navigation-specific DRB that is mounted to the base of the robotic arm [fig_ref] Figure 5: Mazor X stealth edition navigation DRB [/fig_ref]. The X Stealth Edition otherwise performs identically to the non-navigated version.
The TiRobot (TINAVI Medical Technologies, Beijing, China) is the first multi-disciplinary orthopedic robot created entirely in China and received China FDA approval in 2016 [fig_ref] Figure 6: TINAVI TiRobot platform [/fig_ref]. It is the most popular surgical robot in China. It has a robotic arm with 6 DOF and registration is completed by cross-referencing a patient mounted (PSIS or spinous process clamp) DRB as well as a DRB on the robotic end effector after 3D iso-centric fluoroscopic image acquisition. An optical camera station tracks the relative positions of the patient and robotic arm. A third and separate CAN station houses the interface for screw planning and visual feedback. The end effector holds a guide tube for pedicle cannulation and implantation of a guidewire. An optical navigation pointer can be used to check trajectories and anatomical landmarks in real time. Pedicle screws can then be placed over this guidewire and implanted with standard techniques [bib_ref] Safety and accuracy of robotassisted versus fluoroscopy-assisted pedicle screw insertion in thoracolumbar..., Han [/bib_ref]. This robot can be used for other orthopedic applications as well [bib_ref] Percutaneous Sacroiliac Screw Placement: A Prospective Randomized Comparison of Robot-assisted Navigation Procedures..., Wang [/bib_ref] [bib_ref] A robot-assisted surgical system using a force-image control method for pedicle screw..., Tian [/bib_ref].
## Analysis of platform nuances and features
These current navigated robotic platforms all implement optical registration and tracking with 3D imaging for pedicle screw trajectory planning and real-time visual feedback. Although the original ROSA spine robot and the TiRobot are capable of real-time 3D navigation with optical pointer tracking, their systems are not capable of full k-wireless 3D image-guided navigation of instrumentation. Both robotic arms have end effector guide tubes that facilitate non-navigated drilling and subsequent placement of a guidewire. Pedicle tapping and screw implantation can then be performed with navigated instruments using attachable tracker arrays. The upgraded ROSA One should be capable of integrated k-wireless 3D guided robotic screw placement, but there is no published literature or media about it at the time of publication. The Mazor X Stealth Edition and Globus Excelsius both have fully independent navigation capabilities, which are unified with their own manufacturer specific instrumentation and allow for K-wireless screw placement.
Unlike the other platforms, which utilize optical tracking for both robot/patient co-localization and 3D navigation registration, the Mazor X Stealth Edition interfaces with the patient directly. As previously described, the robot is mounted to both the patient and the bed. Additionally, the fiducial marker for the robot arm is removed after registration. Together, this allows the robot to move with the patient and bed during respiration and surgical manipulation while still maintaining the target trajectory. This process is not dependent on optical tracking arrays that are susceptible to line-of-sight obscurations or inadvertent DRB movement.
Maintaining accuracy of registration is of utmost importance in 3D navigated spine surgery [bib_ref] Ensuring navigation integrity using robotics in spine surgery, Crawford [/bib_ref] [bib_ref] Intraoperative image-guided spinal navigation: technical pitfalls and their avoidance, Rahmathulla [/bib_ref]. The current navigated robotic platforms all have features to maintain navigation accuracy. The Rosa and TIRobot systems optically track and facilitate real-time end effector compensation for patient movement during respiration and surgical manipulation [bib_ref] Robot-assisted spine surgery: feasibility study through a prospective casematched analysis, Lonjon [/bib_ref] [bib_ref] Evaluation of the ROSATM Spine robot for minimally invasive surgical procedures, Lefranc [/bib_ref] [bib_ref] Evaluation of Screw Placement Accuracy in Circumferential Lumbar Arthrodesis Using Robotic Assistance..., Chenin [/bib_ref] [bib_ref] Safety and accuracy of robotassisted versus fluoroscopy-assisted pedicle screw insertion in thoracolumbar..., Han [/bib_ref] [bib_ref] Robotic navigation during spine surgery, Zhang [/bib_ref]. The Excelsius GPS also employs additional unique features to ensure navigation integrity [bib_ref] Ensuring navigation integrity using robotics in spine surgery, Crawford [/bib_ref]. To address unintentional DRB shift, which can go unnoticed and result in screw mal-positioning, an additional optical array with just one reflective marker ball is placed into the contralateral PSIS. Continuously updated in real-time, this surveillance marker serves to detect an offset greater than 1.0 mm from the DRB, which triggers an alert automatically. Additionally, the Excelsius GPS robotic arm is engineered to be rigid, capable of maintaining less than 1.0 mm of deflection under a lateral force of 200 N [bib_ref] Ensuring navigation integrity using robotics in spine surgery, Crawford [/bib_ref]. For unintended shift of the DRB, The Mazor X Stealth Edition features a rapid navigation re-registration feature without the need for additional O-arm image acquisition by referencing the original robotic guidance registration. In addition, the navigation DRB is placed on the base of the robot arm and is physically far from the operative field, reducing the chances of accidental displacement.
Skiving, resulting in lateral pedicle breach, is a well described phenomenon in robotic spine surgery. This occurs when the instrument-bone interface loses orthogonality due to a medial entry point with a lateral trajectory [bib_ref] Image-Guided Navigation and Robotics in Spine Surgery, Kochanski [/bib_ref]. To address this issue, the Globus Excelsius end effector was developed with a unique sensor that detects excessive lateral forces generated by instrument skive [bib_ref] Ensuring navigation integrity using robotics in spine surgery, Crawford [/bib_ref]. Visualization of instrument deflection by real time navigation can also help mitigate skive.
It is important to note that the ROSA One and TINAVI TiRobot are both unified platforms designed for use in multiple neurosurgery and orthopedic subspecialties. This ensures efficient resource utilization considering the high capital acquisition cost of these machines.
## Potential future adjunctive navigation modalities
As digital optics continue to improve, high resolution images of the surgical field may be used to register robotic platforms to pre-operative imaging rather than intraoperative fluoroscopy. 7D Surgical (Toronto, ON, Canada) implements digital stereoscopic topographical referencing of exposed bony elements for point-paired surface matching with a pre-operative CT. This allows for rapid registration, elimination of intraoperative radiation exposure, and ease of re-registration. One of the most significant disadvantages of this technology is the technical inability, at the current time, to utilize this method for MIS cases where minimal bony anatomy is exposed.
Augmented reality (AR) has generated significant interest within spine surgery. Augmedics Xvision System (Chicago, IL, USA), is a recently FDA approved AR platform that allows a 3D en face "tip of the spear" view of the bony anatomy to be projected to a user-worn translucent heads-up display. This platform allows the surgeon to maintain line-of-sight to the surgical field and all instruments while receiving 3D navigation feedback in real time. This study, which specifically evaluated radiation exposure with the use of this technology for pedicle screw placement showed that the use of AR surgical navigation resulted in minimal radiation exposure to operating room and staff, whereas the radiation exposure to the patient was equivalent to that reported in prior studies using intra-operative CT-based navigation platforms. Other outcomes including operative time, intra-or post-operative complications, accuracy of screw placement, etc. were not reported in this study.
## Pedicle screw accuracy in current platforms
The literature regarding pedicle screw accuracy with these navigated platforms is relatively sparse at the current time.
No literature exists yet on the pedicle screw accuracy with either the Stealth Edition of Mazor X or the Rosa One Spine platform, but one can infer that the rates will be similar to existing studies that report 100% Ravi Grades A or B [bib_ref] Comparing Next-Generation Robotic Technology with 3-Dimensional Computed Tomography Navigation Technology for the..., Khan [/bib_ref] [bib_ref] Clinical accuracy of computer-assisted two-dimensional fluoroscopy for the percutaneous placement of lumbosacral..., Ravi [/bib_ref].
The original Rosa Spine robot has achieved Gertzbein-Robbin (GR) grades A or B combined accuracy rates of 96.3%, 97.3%, and 98.3% in three different studies [bib_ref] Robot-assisted spine surgery: feasibility study through a prospective casematched analysis, Lonjon [/bib_ref] [bib_ref] Evaluation of the ROSATM Spine robot for minimally invasive surgical procedures, Lefranc [/bib_ref] [bib_ref] Evaluation of Screw Placement Accuracy in Circumferential Lumbar Arthrodesis Using Robotic Assistance..., Chenin [/bib_ref] [bib_ref] Spinal Cervical Meningiomas: The Challenge Posed by Ventral Location, Lonjon [/bib_ref].
The Excelsius GPS has been the subject of two large retrospective reviews. Jain et al. reported no screw related complications or returns to the operating room with 643 screws placed. Of the 66 screws that were reviewed with post-operative CT scanning, 100% of those were categorized as GR A or B. In their series of 600 screws all reviewed with post-operative CT, Wallace et al. reported a GR A or B grade for 98.2% of the screws. Grade C and D designations were given to 1.5% and 0.3% of screws. These were all laterally breached, and were repositioned successfully by utilizing a more lateral entry point and medial trajectory. The offset from the plan to the final position of the tips and tails of the screws were 1.7±1.3 and 1.8±1.2 mm, respectively. Deviation from planned angle was 2.0±1.6 degrees (36). Elswick et al. also reported 97.6% GR grades A and B for their study involving 125 screws [bib_ref] Robotic-Assisted Spinal Surgery: Current Generation Instrumentation and New Applications, Elswick [/bib_ref].
The TiRobot has been compared to fluoro-guided free hand (FFH) screw placement in a large randomized prospective trial by Han et al. [bib_ref] Safety and accuracy of robotassisted versus fluoroscopy-assisted pedicle screw insertion in thoracolumbar..., Han [/bib_ref]. A total of 1,116 screws were placed between the two groups, 532 in the robot guided (RG) cohort and 584 in the FFH group. The TiRobot demonstrated superior results across the board. Gertzbein-Robbin accuracy was 95.3% grade A and 98.7% combined grades A and B for the RG group versus 86.1% grade A and 93.5% grades A and B for the FFH group. Radiation exposure and blood loss were significantly lower in the robot guided group. There were no proximal facet violations in the RG group compared to 2.1% violation in the FFH group. Surgical times between the two groups were not statistically different.
Although no accuracy studies exist for robot navigated pedicle screws placed in the lateral position for single position circumferential thoracolumbar surgery, these data are undoubtedly coming [bib_ref] Single position spinal surgery for the treatment of grade II spondylolisthesis: A..., Quiceno [/bib_ref]. Navigated robotic placement of cortical trajectory lumbar screws has also been reported [bib_ref] Robot-Assisted Cortical Bone Trajectory Insertion of Pedicle Screws: 2-Dimensional Operative Video, Agyei [/bib_ref].
## Radiation exposure in navigated spine robots
Although non-navigated robotic systems are not commonly utilized, previous studies have shown a significant decrease in radiation exposure compared to fluoroscopic-assisted pedicle screw implantation techniques [bib_ref] Assessment of Pedicle Screw Placement Accuracy, Procedure Time, and Radiation Exposure Using..., Lieberman [/bib_ref] [bib_ref] Robotic-assisted pedicle screw placement: lessons learned from the first 102 patients, Hu [/bib_ref] [bib_ref] Perioperative course and accuracy of screw positioning in conventional, open robotic-guided and..., Kantelhardt [/bib_ref]. In a prospective randomized trial of FFH versus 3D navigation versus RG pedicle screw placement, Roser et al. reported half the radiation dose in the RG group vs. FFH. Interestingly, there was even less ionizing radiation emitted in the 3D navigation group, with 35% decreased exposure compared to the RG. This could be explained by the obviation of fluoroscopic guidance for depth control with real-time navigation. Still, conflicting evidence exists with Khan et al. reporting higher radiation exposure in the 3D navigation group compared to the non-navigated RG [bib_ref] Comparing Next-Generation Robotic Technology with 3-Dimensional Computed Tomography Navigation Technology for the..., Khan [/bib_ref]. Future studies comparing navigated robotic cohorts to non-navigated robotic cohorts, as well as those comparing navigated robotic cohorts to non-robotic CAN will help further define this potential advantage.
## Learning curve
Studies have shown little difference between novice and experienced surgeons with respect to procedural efficiency in placing robotically guided pedicle screws; despite this, a trend towards increasing efficiency after the performance of a certain number of cases or placement of a certain number of screws placed has been reported [bib_ref] Initial academic experience and learning curve with robotic spine instrumentation, Urakov [/bib_ref] [bib_ref] Learning Curve for Robot-Assisted Percutaneous Pedicle Screw Placement in Thoracolumbar Surgery, Kam [/bib_ref]. The only study in a full navigation enabled robotic platform that comments on learning curve showed no difference between the attending and the first participating fellow. They both needed to place 30 screws before a noticeable improvement in efficiency was quantified. Notably, the subsequent fellow in that study, now learning from the attending surgeon with newly acquired robotic experience, demonstrated immediate and sustained superior performance compared to the attending, suggesting that the learning curve is easily transferable through observation [bib_ref] Robot-Assisted Pedicle Screw Placement: Learning Curve Experience, Siddiqui [/bib_ref]. The visual feedback from real-time navigation could be responsible for this observation. However, further investigation is required to fully understand the learning process of adoption of robotic surgery in practice and how the surgeons' prior experience with other image guidance and CAN modalities impact this.
Other applications of navigated robotic guidance Navigated cervical spine applications including successful placement of odontoid screw [bib_ref] Robot-assisted Anterior Odontoid Screw Fixation: A Case Report, Tian [/bib_ref] and C1-2 trans-articular screwshave been reported of the TiRobot platform only, but not any of the other robots currently in use. In both the above reports, the DRB was placed on the Mayfield head fixation clamp.
Navigated robot-assisted laminectomy/osteotomy with full decompression will inevitably be a part of the future, with biomechanical studies having been conducted for this application [bib_ref] Force-based control of a compact spinal milling robot, Wang [/bib_ref] [bib_ref] Multilevel Fuzzy Control Based on Force Information in Robot-Assisted Decompressive Laminectomy, Qi [/bib_ref]. These preliminary studies demonstrate that the use of a robot for this purpose has the potential to provide stable and steady bony decompression, with the ability to program the robot to discontinue further drilling when the pre-planned thickness of remaining bone has been achieved in order to prevent complete penetration of the lamina. We envision that this could be implemented by using supervisory-control for bony drilling with automatic disengagement a few millimeters proximal to a predefined 3D area of interest such as the boundaries of the spinal canal. The surgeon could then resume control and complete the remainder of the decompression manually with navigated robotic assistance to avoid critical neural elements. Similarly, robot-guided osteotomies for en-bloc resection of primary spinal column tumors have already been reported for non-navigated spine robots. While these initial studies are promising, the evidence is currently limited due to the small number of cases reported. The integration of navigation with robotics will likely promote an increase in their utilization in this arena given the ability to determine depth with real-time navigation [bib_ref] Robotic guidance for en bloc sacrectomy: a case report, Bederman [/bib_ref] [bib_ref] Robotic Tissue Manipulation and Resection in Spine Surgery, Trybula [/bib_ref]. This should theoretically improve procedural efficiency and safety.
The strenuous physical demands of surgery also pose a significant occupational risk to surgeons (54). A robotic system with 3D navigation and real-time feedback with potential to perform automated sub-periosteal spine exposure may relieve manual hand strain. Automated pedicle screw trajectory planning [bib_ref] Automatic pedicle screw planning using atlas-based registration of anatomy and reference trajectories, Vijayan [/bib_ref] and automated insertion could be pre-programmed for execution by the robotic system. The platforms might also integrate telerobotic control, allowing the surgeon to perform the operation by remote control in an ergonomically optimized command station. Although the use of the DaVinci robot for posterior spine approaches in a porcine model for non-instrumented spinal surgery was reported over a decade ago, this platform did not gain traction for spinal surgery likely related to limitations with hardware and software. However, given the recent advances in and increased adoption of other robotic systems for spine surgery, a platform similar to the DaVinci robot, which allows direct translation of real-time surgeon input to robot output may be revisited.
# Conclusions
Although spine surgery robots were introduced nearly 20 years ago, their adoption was limited due to a number of factors. Meanwhile, the popularity of 3D navigation in spine surgery has grown considerably. The combination of robotics with real-time 3D navigation has led to a resurgence in the interest of these new robotic platforms and their potential future applications beyond pedicle screw insertion. Initial studies have shown that the use of technology is feasible, safe and effective. It has the potential to decrease occupational radiation exposure, improve accuracy of instrumentation, allow for more efficient surgery, and alleviate some of the physical and mental demands of surgery for the surgeon. Additionally, while a majority of current studies report on the use of this technology for pedicle screw instrumentation, numerous other applications are being explored. This will further add to the utility, versatility and cost-effectiveness of these platforms, and allow for more widespread use. Further studies are necessary to determine the efficacy, efficiency, safety and value of these new combined robotic navigation systems.
# Acknowledgments
We would like to acknowledge Kristopher Barberio, DC and Summer Shedd for providing technical educational resources pertaining to current robotic platforms. Funding: None.
## Footnote
Provenance and Peer Review: This article was commissioned by the Guest Editor (Dr. Sheeraz Qureshi) for the series "Current State of Intraoperative Imaging" published in Annals of Translational Medicine. The article was sent for external peer review organized by the Guest Editor and the editorial office.
Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at http://dx.doi. org/10.21037/atm-2020-ioi-07). The series "Current State of Intraoperative Imaging" was commissioned by the editorial office without any funding or sponsorship. BNS reports personal fees from Medtronic, outside the submitted work and is a member of the Medtronic Robotics Advisory Board. The authors have no other conflicts of interest to declare.
Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the noncommercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.
[fig] Figure 1 07, Figure 2: Rosa spine robot platform. Freely mobile floor-mounted base station with arm (left) and optical infrared tracking camera (right). © Annals of Translational Medicine. All rights reserved. Ann Transl Med 2021;9(1):86 | http://dx.doi.org/10.21037/atm-2020-ioi-Globus Excelsius GPS platform. Freely mobile floor-mounted base station with arm (left) and Optical infrared tracking camera (right). [/fig]
[fig] Figure 3: Globus Excelsius GPS end effector with navigated instrument. [/fig]
[fig] Figure 4: Mazor X stealth edition platform. Navigation tracking camera (left), base station (middle), bed (and patient) mounted robotic arm with end effector (right). [/fig]
[fig] Figure 5: Mazor X stealth edition navigation DRB (Black 6 marker array) and navigated instrument (Blue 4 marker array). [/fig]
[fig] Figure 6: TINAVI TiRobot platform. Base station (left), optical tracking camera (middle), and floor-mounted robotic arm (right). [/fig]
|
Hemangioma of the Ilium Simulating an Aggressive Bone Lesion on Imaging
We present the case of a 35-year-old man with a hemangioma in the iliac wing that simulated an aggressive bone lesion on CT, MRI, and bone scintigraphy. On CT, the lesion, 12 cm in in maximum dimension, was expansile and septated. On MRI, the lesion was isointense with respect to muscle on T1 weighted images, and hyperintense on T2 images, with focal regions of cortical breach. Radionuclide bone scan showed increased tracer accumulation. The diagnosis was made following surgical biopsy, and the lesion was treated conservatively.A 35-year-old man presented with a four-year history of right hip pain and disability exacerbated by physical Hemangioma is frequently detected and diagnosed by cross-sectional imaging as incidental finding in the thoracic and lumbar spine. Less frequently, hemangiomas are found in the calvaria, and rarely in the appendicular skeleton. We present a case of hemangioma in the iliac wing that simulated an aggressive bone lesion on CT, MRI, and bone scintigraphy.
# Introduction
activity and weightbearing. The patient denied any prior trauma. He did not have any fever, chill, night sweat or weight loss. Right hip radiographs showed an expansile and septated lesion affecting the right acetabulum. Pelvic CT confirmed the large and aggressively appearing acetabular lesion with diffuse trabecular destruction. The tumor had cortical bone expansion with multiples foci of osseous erosion and disruption , arrows). The tumor measured 8 x 5 x 12 cm extending from the inferior ilium to the ischium without definite adjacent soft tissue invasion. MRI showed a large lesion of the left iliac bone involving the acetabulum and ischium , arrows). The mass was isointense to muscles on T1-weighted images and hyperintense on T2-weighted sequences. Cortical breaches were detected at the medial aspect of the acetabulum and iliac bone. There was no hip joint mass or effusion. Since CT and MRI showed features of an aggressive tumor, bone scan was obtained for search of potential metastasis. The whole body technetium-99m MDP scintigraphy showed increased radiotracer uptake of the right acetabulum and right ischium , arrows) but did not detect any additional osseous lesion. An initial CT-guided bone biopsy showed rare fragments of reactive woven bone, several irregular aggregates of spindle cells with round to oval shaped nuclei, bland-appearing chromatin patterns and Osseous hemangiomas are rare and represent about 1% of all bone tumors [bib_ref] Solitary skeletal hemangioma of the extremities, Kaleem [/bib_ref] [bib_ref] Solitary osseous hemangioma outside the spinal and craniofacial bones, Ogose [/bib_ref]. The majority of these benign bone lesions, close to 80%, are encountered in the spine especially thoracic and lumbar, and cranio-facial bones. Tubular and long bone sites account for 10% of the total. Pelvic bone hemangioma is uncommon, seen only in 3-4% of all cases [bib_ref] Benign vascular lesions of bone: radiologic and pathologic features, Wenger [/bib_ref] [bib_ref] Rare, primary iliac, pubic and ischial tumours in children (report of 14..., Kozlowski [/bib_ref] [bib_ref] Acetabular protrusion secondary to pelvic hemangioma. A case report and review of..., Rao [/bib_ref]. Intraosseous hemangiomas are usually seen in a middle-aged patient population with female gender predilection even though they can occur at extreme ages of both sexes. Their pathogenesis may be secondary to congenital, developmental or acquired vascular proliferations [bib_ref] Benign bone tumors and tumor-like lesions: value of cross-sectional imaging, Woertler [/bib_ref]. Incidental finding is frequent on cross-sectional imaging for the majority of osseous hemangiomas, which are asymptomatic. A small percentage of these benign tumors are detected secondary to pain symptomatology related to their weight-bearing location such as pelvic girdle and lower extremities [bib_ref] Acetabular protrusion secondary to pelvic hemangioma. A case report and review of..., Rao [/bib_ref] [bib_ref] Intraosseous haemangioma of the proximal femur: imaging findings, Chawla [/bib_ref].
Classic radiographic patterns include bone demineralization with coarsely prominent vertical trabeculae and corduroy appearance. CT usually shows well-marginated focal areas of decreased bone attenuation with polka-dot or honeycomb appearance. MR frequently demonstrates high signal intensity on T2 weighted sequences and variable degrees of increased signal intensity on T1-weighted images depending of the vascular and fat components of the lesion [bib_ref] Solitary osseous hemangioma outside the spinal and craniofacial bones, Ogose [/bib_ref]. Sunburst and spoke-wheel features may be seen with calvarial hemangiomas, which are multiple in 15% of the cases [bib_ref] Benign vascular lesions of bone: radiologic and pathologic features, Wenger [/bib_ref]. Similar to this case presentation, osseous hemangiomas may mimic malignant neoplasms with aggressive looking patterns of cortical bone expansion and destruction. The unusual pelvic and appendicular bone sites with atypical imaging features and pronounced clinical symptomatology would heighten the suspicion for Discussion malignancy. Technetium-99m MDP bone scintigraphy is non-specific with either variable degrees of radiotracer uptake or absence of tracer accumulation [bib_ref] Vertebral body hemangioma showing increased uptake of Tc-99m MDP and decreased Tc-99m..., Halkar [/bib_ref] [bib_ref] Hemangioma, a rare cause of photopenic lesion on skeletal imaging, Makhija [/bib_ref] [bib_ref] Bone SPECT imaging of vertebral hemangioma: correlation with MR imaging and symptoms, Han [/bib_ref] [bib_ref] Vertebral hemangioma presenting as a cold defect on bone scintigraphy, Reader [/bib_ref].
As seen in other rare cases of hemangiomas around the hip joint, the differential diagnosis includes a large spectrum of expansile lytic lesions such as aneurysmal bone cyst, eosinophilic granuloma, fibrous dysplasia, giant cell tumor, plasmacytoma, brown tumor, malignant fibrous histiocytoma, and metastasis. Correct histological tumor identification is necessary for treatment. Routine biopsy may be challenging and non-contributory due to insufficient material. En-bloc open biopsy should provide adequate specimen showing dilated vascular channels lined with endothelial cells characteristic of hemangioma. Large skeletal hemangiomas may be prone to osseous deformities at weight-bearing sites, pathologic fractures, spontaneous bleed, iatrogenic hemorrhage secondary to biopsy or surgery, and malignant degeneration [bib_ref] Acetabular protrusion secondary to pelvic hemangioma. A case report and review of..., Rao [/bib_ref]. Treatment is conservative for asymptomatic and small hemangiomas. Surgery by curettage and bone graft is indicated for small symptomatic ones. Large symptomatic or incompletely resected hemangiomas may benefit from radiation therapy with long term improvement of symptoms and tumor no apparent mitotic figures. The spindle cells were in a fibromyxoid matrix with presence of scattered thin-walled blood vessels and rare giant cells. There was no histologic evidence of plasmacytoma or giant cell tumor. This CTguided biopsy was not diagnostic but favored a low-grade neoplasm. A subsequent surgical open biopsy reached the diagnosis of iliac bone hemangioma. Due to the large size of the tumor, a conservative approach was observed. Close follow-up for two years showed continuously increasing pain requiring radiation therapy. Additional two-year posttherapeutic monitoring witnessed clinical improvement with stability of the iliac hemangioma on cross-sectional imaging.
[fig] Figure 1A: 35-year-old man with hemangioma of the right iliac wing. A, Axial CT-guided biopsy image of the right pelvis showed an expansile and septated lesion involving the iliac bone. There are foci of cortical erosion and destruction at the medial aspect of the right acetabulum (arrows). [/fig]
|
Anesthetic Management for Multiple Family Members with Myotonic Dystrophy for Interventional Cardiac Procedures—A Case Series
Myotonic muscular dystrophy (MMD) is a rare autosomal dominant disorder that can complicate anesthetic management of patients. MMD is characterized by progressively worsening muscle loss and weakness, cardiac conduction abnormalities, cardiomyopathy, restrictive lung disease, obstructive sleep apnea, and delayed gastric emptying. Patients presenting with MMD for any surgical procedure present a management challenge to the anesthesiologist. Several reports of airway loss due to medicationmediated respiratory depression, sudden death due to dysrhythmias, aspiration of stomach contents, and prolonged intubation have been reported. We present a case series of three family members with MMD type 1 who presented for electrophysiologic assessment of the cardiac conduction system and possible pacemaker insertion. While there are reports of anesthetic management of patients with myotonic dystrophy for various procedures, our report is unique in that we were able to demonstrate variations of anesthetic management based on the procedure and variation in disease phenotype-differing severity between family members.Keywords: myotonic muscular dystrophy presentation, anesthetic care of patients with MMD, general anesthesia vs. monitored anesthesia care for patients with myotonic dystrophy, aspiration risk with myotonic dystrophy, pacemaker insertion in patients with myotonic dystrophy
# Introduction
Myotonic muscular dystrophy (MMD) is a rare autosomal dominant disorder that can complicate anesthetic management of patients. It is the most common form of adult-onset muscular dystrophy. MMD is characterized by progressively worsening muscle loss and weakness. It affects the skeletal muscle and the smooth muscle of the body (2), leading to prolonged muscle contractions and impaired relaxation. Other symptoms include cataract formation, cardiac conduction abnormalities, cardiomyopathy, restrictive lung disease, obstructive sleep apnea, delayed gastric emptying, infertility, and, in severe cases, cognitive dysfunction [bib_ref] Myotonic dystrophies, type 1 and 2: anesthetic care, Veyckemans [/bib_ref]. There are two most common types of MMD, type 1 and type 2. They differ in their genetic mutations as well as manifestations. MMD type 1 is caused by mutations in the DMPK gene, whereas type 2 is caused by mutations in the CNBP gene. The muscular weakness present in MMD type 1 usually affects the face, neck, hands, and lower legs, while the weakness associated with MMD type 2 usually affects the neck, shoulder, elbows, and hips.
Patients presenting with MMD for any surgical procedure present a management challenge to the anesthesiologist. There have been several reports of airway loss due to medication mediated respiratory depression, sudden death due to dysrhythmias, aspiration of stomach contents, and prolonged intubation. We present a case series of three family members with MMD type 1 who presented for electrophysiologic (EP) assessment of the cardiac conduction system and possible pacemaker insertion. While there are reports of anesthetic management of patients with myotonic dystrophy for various procedures, our report is unique in that we were able to demonstrate variations of anesthetic management based on the procedure and variation in disease phenotype-differing severity between family members.
# Methods
# Results/case series
Three siblings with known history of MMD type 1 were scheduled to undergo EP studies with possible device placement under anesthesia over a span of 3 weeks. Per family report, their mother, who also had MMD type 1, died from aspiration pneumonia after an endoscopy 3 years prior. As a result, the family was very anxious about undergoing this procedure. After consultation with the family, the decision was made to try to avoid the use of neuromuscular blockade.
The first patient was a 31-year-old female, presenting for an EP study with possible pacemaker placement. The patient was able to ambulate without the use of a walker or braces. She reported the ability to walk a considerable distance before becoming fatigued. As a part of the MMD disease process, the patient was also diagnosed with central sleep apnea and used a CPAP machine at night. She denied any history of gastroesophageal reflux disease (GERD). The decision was made to proceed with the EP study under monitored anesthesia care (MAC) and convert to general anesthesia (GA) if a pacemaker needed to be placed. The patient received Ondansetron 4 mg IV, Metoclopramide 10 mg IV, and Dexamethasone 4 mg IV at the start of the procedure in order to prevent nausea, vomiting, and possible aspiration. For sedation, she was given Fentanyl 50 mcg IV, Midazolam 2 mg IV, and was started on a Propofol infusion. She tolerated the study well.
When the EP study revealed a prolonged PR interval (217 ms) and a wide QRS (125 ms), the cardiology team decided to proceed with pacemaker placement. At this time, the decision was made to convert to GA and secure the airway. The patient was given an additional bolus of Propofol (80 mg IV) and Fentanyl (50 mcg IV) and was successfully intubated using direct laryngoscopy without the use of paralytic agents. She tolerated this well and was maintained on a Propofol infusion for the duration of the procedure. On emergence from GA, the patient received Ondansetron 4 mg IV and was extubated uneventfully after she was taking adequate tidal volumes, protecting her airway, and following commands. She was discharged home the next day with no acute events.
The second patient was a 25-year-old male, the younger brother of the first patient. He presented for an EP study with possible pacemaker placement. The patient reported generalized weakness and poorer functional status than his sister, necessitating the use of assistance to ambulate. However, in contrast with his sister, this patient denied a history of sleep apnea or the use of a CPAP machine (polysomnogram was reportedly negative). He also denied any history of GERD. This patient had a similar anesthetic plan prescribed, using MAC anesthesia for an EP study and converting to GA if a pacemaker needed to be placed. For sedation, the patient received a total of 100 mcg of Fentanyl IV (in 50 mcg boluses) and 2 mg of Midazolam IV (in 1 mg boluses). He required minimal sedation for the study and was comfortable throughout. His EP study did not reveal any ECG abnormalities (PR interval 200 ms, QRS 110 ms), and there was no need for pacemaker insertion at the time. The patient emerged from sedation without incident and recovered in the post-anesthesia care unit without evidence of respiratory compromise. He was discharged home the same day with recommendations for serial ECGs every 6-12 months.
The third patient, the middle sibling of the first two patients, was a 27-year-old male presenting for an EP study with possible pacemaker placement. The patient reported generalized weakness as well but was able to ambulate without any assistance. He had a history of central sleep apnea and used a CPAP machine at night, similar to his sister, the first patient. On physical exam, it was noted that the patient had extremely poor dentition with multiple erythematous areas, suspicious for infection and/or abscess. After discussion with the electrophysiology team, it was agreed that if the patient needed pacemaker insertion, he would need to obtain dental clearance prior to placement due to high risk of device infection. However, the EP study could still proceed. The plan was to use MAC anesthesia for the study and have the patient return for a subsequent anesthetic if a pacemaker was indicated. He received Dexamethasone 4 mg IV, Famotidine 20 mg IV, Metoclopramide 10 mg IV, and Ondansetron 4 mg IV for nausea and vomiting prophylaxis. For the procedure, the patient received a total of 100 mcg of Fentanyl IV (in 50 mcg boluses) and 2 mg of Midazolam IV (in 1 mg boluses). Interestingly, during the study, the patient developed atrial fibrillation with rapid ventricular response, and heart rate as high as the 150 s. His blood pressure was stable throughout the event. The decision was made to cardiovert the patient. Propofol 30 mg IV bolus was given, and cardioversion was performed without the need for intubation. The patient subsequently returned to normal sinus rhythm. The EP study showed abnormal infranodal conduction and pacemaker placement was recommended. However, due to poor dentition and concern for tooth abscess, pacemaker placement was deferred. After satisfactory recovery from sedation and no evidence of respiratory compromise or apneic events, the patient was discharged home.
# Discussion
This case series documents the anesthetic care of three siblings with MMD type 1. Each sibling had different degrees of disease manifestation and severity. The first and third patients both had generalized weakness but were able to ambulate without the assistance of a walker. The second patient, however, did require assistance with ambulation. In addition, the first and third patients were also diagnosed with central sleep apnea and used a CPAP machine at night. All of these factors contributed to their respective anesthetic management.
The anesthetic plan for all three siblings was to start with MAC sedation and convert to GA if a device was indicated. Only the first of the three patients required conversion to GA for pacemaker insertion. While the third patient would have benefited from device placement as well, his extremely poor dentition at the time of the procedure increased the risk of infecting a newly placed device significantly. The decision point between MAC and GA is a very important one for these patients, especially those with more severe disease burden. A retrospective study by Mathieu et al. of 219 patients demonstrated an odds ratio of 14.1 for postoperative respiratory complications in MMD patients with proximal limb weakness undergoing GA versus MMD patients without [bib_ref] Anesthetic and surgical complications in 219 cases of myotonic dystrophy, Mathieu [/bib_ref]. While GA has been reportedly administered safely in MMD patients with more benign disease, many common agents administered during a routine GA pose significant concerns-opioids, neuromuscular blockers [bib_ref] Anesthesia and myotonic dystrophy type 2: a case series, Weingarten [/bib_ref]. Thus, the risks and benefit assessment needs to be well communicated with the patient, anesthesia team, and the surgical or procedural team.
Monitored anesthesia care sedation has the advantage of avoiding intubation and maintaining protective airway reflexes; however, many procedures cannot be safely accomplished using this technique. Another complicating factor is that sedation for these patients has to be very minimal. If they are even mildly over-sedated without a secure airway, there is a significant risk of airway loss or aspiration due to lower baseline pharyngeal muscle tone [bib_ref] Myotonic dystrophies, type 1 and 2: anesthetic care, Veyckemans [/bib_ref].
In contrast, GA provides a secure airway for the procedure. However, these patients are at an increased risk of aspiration on induction of GA. In addition, there is an increased risk for prolonged intubation, which can lead to other complications such as ventilator-associated pneumonia, deconditioning, and chronic respiratory failure requiring a tracheostomy. As demonstrated by the first case, the conversion to GA was successfully accomplished with the use of propofol and fentanyl but no paralytic agents, in order to minimize any effects on the patient's musculoskeletal system. Neuromuscular blocking agents (paralytics), depolarizing and non-depolarizing, may have altered effects on this patient population, but case reports to date have been inconsistent. Succinylcholine has been reported to cause dose-dependent contractures in MMD patients and should be avoided [bib_ref] Myotonic dystrophies, type 1 and 2: anesthetic care, Veyckemans [/bib_ref]. While non-depolarizing agents are generally well-tolerated, the antagonizing agent, neostigmine, has been reported to incompletely reverse neuromuscular blockade (3) and even cause a myotonic response [bib_ref] Anaesthesia and myotonia, Russell [/bib_ref]. However, with the recent approval of sugammadex, safe and successful reversal of rocuronium has been reported with this patient population [bib_ref] The use of sugammadex in a patient with myotonic dystrophy, Matsuki [/bib_ref].
All three patients were very closely monitored for arrhythmias throughout the procedure. The third patient developed atrial fibrillation with rapid ventricular response during the study and had to undergo cardioversion. The other two patients did not develop any arrhythmias.
In summary, taking care of patients with myotonic dystrophy can pose a significant challenge to the anesthesiologist in the operating room but may be safely conducted with the proper considerations. Patients with this condition are at risk for aspiration, airway loss, prolonged intubation, and cardiac arrhythmias. Careful consideration must be given to the type of anesthetic required for the procedure. A careful and thorough risk and benefit assessment must be communicated with the patient as well as the surgical team. If MAC is chosen, it is imperative to avoid over-sedating the patient. If GA is chosen, it is preferable to avoid the use of paralytic agents. However, if muscle relaxation is required for the given procedure, then full and adequate reversal of the paralytic agent is required with demonstration of adequate muscular function prior to extubation. Additional monitoring for muscle weakness in the PACU or ICU can be beneficial as well.
# Ethics statement
This case series was IRB exempt. Written informed consent was obtained.
# Author contributions
LG and AF took care of two of the patients described in this case series. Both authors wrote and edited the manuscript.
## References
|
Regional employment and individual worklessness during the Great Recession and the health of the working-age population: Cross-national analysis of 16 European countries
Studies from single countries suggest that local labour market conditions, including rates of employment, tend to be associated with the health of the populations residing in those areas, even after adjustment for individual characteristics including employment status. The aim of this study is to strengthen the cross-national evidence base on the influence of regional employment levels and individual worklessness on health during the period of the Great Recession. We investigate whether higher regional employment levels are associated with better health over and above individual level employment. Individual level data (N = 23,078 aged 15-64 years) were taken from 16 countries (and United Kingdom) participating in the 2014 European Social Survey. Regional employment rates were extracted from Eurostat, corresponding with the start (2008) and end (2013) of the Great Recession. Health outcomes included self-reported heart or circulation problems, high blood pressure, diabetes, self-rated health, depression, obesity and allergies (as a falsification test). We calculated multilevel Poisson regression models, which included individuals nested within regions, controlling for potential confounding variables and country fixed effects. After adjustment for individual level socio-demographic factors, higher average regional employment rates (from 2008 to 2013) were associated with better health outcomes. Individual level worklessness was associated with worsened health outcomes, most strongly with poor self-rated health. In models including both individual worklessness and the average regional employment rate, regional employment remained associated with heart and circulation problems, depression and obesity. There was evidence of an interaction between individual worklessness and regional employment for poor self-rated health and depression. The findings suggest that across 16 European countries, for some key outcomes, higher levels of employment in the regional labour market may be beneficial for the health of the local population.
# Introduction
Since the onset of the global financial crisis in 2008 and the subsequent recession experienced in many countries, there has been renewed interest in the role of unemployment, job insecurity and worklessness in influencing population health. This study aims to strengthen the cross-national evidence base on the influence of individual worklessness and regional employment levels (as a measure of the local labour market) on health during the Great Recession.
2007), adverse health behaviours such as alcohol and tobacco consumption [bib_ref] Unemployment, cigarette smoking, alcohol consumption and body weight in young British men, Montgomery [/bib_ref] [bib_ref] Does unemployment lead to greater alcohol consumption?, Popovici [/bib_ref] , poor self-rated health [bib_ref] Welfare state regimes, unemployment and health: a comparative study of the relationship..., Bambra [/bib_ref] and elevated inflammatory biomarkers [bib_ref] Unemployment and inflammatory markers in England, wales and Scotland, 1998-2012: meta-analysis of..., Hughes [/bib_ref]. Links between unemployment and poor health have conventionally been explained through two inter-related concepts: the psychosocial effects of unemployment (e.g. stigma, isolation and loss of self-worth) and the material consequences of unemployment (e.g. wage loss and resulting changes in access to essential goods and services).
Other aspects of worklessness are also linked to poorer health and employment outcomes. For example, lone mothers are twice as likely as coupled mothers to describe their health as 'not good' and across Europe, a range of adverse health outcomes are experienced disproportionately by lone parents, including psychiatric disorders; attempted suicide; and alcohol and drugs-related disease. Mechanisms linking lone parenthood to poor health also include poverty, often due to non-employment, lack of support, and stigma [bib_ref] The self-reported health status of lone parents, Benzeval [/bib_ref]. Further, having a long-term health condition or disability is a significant risk factor for being out of the labour market. People who develop chronic health problems whilst in employment are twice as likely to become workless within a four year period as those who remain healthy, and women and men in poor health are 60% and 40% less likely to enter paid employment than those in good health [bib_ref] Effectiveness of a health promotion programme for long-term unemployed subjects with health..., Schuring [/bib_ref]. In combination with other labour market disadvantages such as low educational level, poor health further increases the risk of worklessness and there are substantial regional inequalities in health-related worklessness [bib_ref] Worklessness and regional differences in the social gradient in general health: evidence..., Bambra [/bib_ref].
## Local and regional labour market context
More recently, research has shown that local labour market conditions, including rates of employment and unemployment, tend to be associated with the physical and mental health outcomes of the populations residing in those areas, even after adjustment for individual characteristics including employment status [bib_ref] Neighbourhood environment and its association with self rated health: evidence from Scotland..., Cummins [/bib_ref] [bib_ref] Do local unemployment rates modify the effect of individual labour market status..., Flint [/bib_ref]. Understanding the role of local labour market conditions in influencing health is important because the health impacts of economic downturns in some places may be larger and more prolonged than in other places, which has significant policy implications. However, to date, the international evidence base examining the relationships between local labour market conditions and health is slight, and to our knowledge there are no international comparative studies investigating the individual and contextual influence of employment on individual level health status.
Local labour market conditions are likely to influence population health through a number of interrelated pathways including: heightened job insecurity; weakening community cohesion; increasing placebased stigma; health selective migration; stifling regional income levels; raising workplace stresses; accentuating local problems related to poverty such as crime, unhealthy behaviours (e.g. tobacco and alcohol use) and illicit drug use; and detrimentally affecting the local economy by restricting the availability of community resources such as shops [bib_ref] Long-term local area employment rates as predictors of individual mortality and morbidity:..., Riva [/bib_ref].
The evidence base linking contextual employment to health is mixed and it is clear that, for some health outcomes at least, the relationships between local employment opportunities and health are complex. Studies within the same country have found conflicting results; in Sweden two studies found that after adjusting for individual employment status, local unemployment levels were associated with adverse outcomes including coronary heart disease and smoking , whereas another study of psychological distress did not find an association [bib_ref] Mental health among the unemployed and the unemployment rate in the municipality, Strandh [/bib_ref]. However, psychiatric disorders, suicide and parasuicide have been found to be less prevalent for unemployed people in areas of high unemployment compared to low unemployment areas [bib_ref] Mental health of unemployed men in different parts of England and Wales, Jackson [/bib_ref] [bib_ref] Long term trends in parasuicide and unemployment in Edinburgh, Platt [/bib_ref] [bib_ref] Suicide and unemployment in Italy: description, analysis and interpretation of recent trends, Platt [/bib_ref] [bib_ref] Are there geographical variations in the psychological cost of unemployment in South..., Powdthavee [/bib_ref].
One interpretation of this is that in areas where unemployment levels are high, unemployment is less stigmatised, and the impact of an individual's own unemployment is diminished [bib_ref] Do local unemployment rates modify the effect of individual labour market status..., Flint [/bib_ref]. However, in a study exploring the impact of unemployment on subjective wellbeing in Germany and Switzerland, high regional unemployment levels did not act as a buffer for those who became unemployed [bib_ref] Does unemployment hurt less if there is more of it around? A..., Oesch [/bib_ref] , going against this hypothesis.
Importantly, recent work has emphasised that the long-term employment trajectories in local labour market conditions may be important for health. A longitudinal study of trends in local labour market conditions between 1981 and 2008 in England found that the risk of mortality or morbidity was greater in places where employment rates were persistently low or declining [bib_ref] Long-term local area employment rates as predictors of individual mortality and morbidity:..., Riva [/bib_ref]. Similarly, work using a Swedish cohort found that high levels of local unemployment had a detrimental impact on functional somatic symptoms (bodily complaints such as headaches, musculoskeletal pain, abdominal pain and dizziness) and this association was strongest in adulthood at age 30 [bib_ref] Does contextual unemployment matter for health status across the life course? A..., Brydsten [/bib_ref]. There is also evidence that the influence of local labour market conditions on health may vary between different sociodemographic groups. Local labour market conditions are likely to be particularly important for those who are economically inactive or job insecure, as well as those who are less mobile or tend to face greater work-based discrimination (e.g. women, ethnic minority groups and those with low levels of relevant skills or education). Considering the potential impact of local labour market conditions on health is vital during times of economic downturn as regions may have differing levels of resilience to economic shocks depending on their baseline level of employment and the type of employment which dominates (e.g. manufacturing and public sector), as well as the differential response to the recession across regions, including the implementation of austerity measures characterised by public expenditure cuts to reduce government debt [bib_ref] Regional unemployment impacts of the global financial crisis in the new member..., Blažek [/bib_ref] [bib_ref] Regional resilience in the 2008-2010 downturn: comparative evidence from European countries, Davies [/bib_ref].
## The great recession and austerity
National economic wealth (i.e. Gross Domestic Product) has long been considered as a major global determinant of population health, with the vast differences in mortality between developed and developing countries accounted for in terms of differences in economic growth. Changes in the economy therefore potentially have important implications for population health and inequalities in health. Economic recessions are characterised by instability (in terms of inflation and interest rates) and sudden reductions in production and consumption with corresponding increases in unemployment. The economic downturn which started in late 2007 is popularly referred to as the 'Great Recession' as it has been longer, wider and deeper than any previous economic downturns including the 'Great Depression' of the 1930s. For example, it was characterised by unemployment rates of around 8.5% in the UK and the USA, 10% in France and more than 20% in Spain.
The short term overall population health effects of recessions are rather mixed. The majority of international studies conclude that there are declines in all-cause mortality, deaths from cardiovascular disease and motor vehicle accidents, as well as decreases in hazardous health behaviours during economic downturns, whilst deaths from suicides, rates of mental ill health and chronic illnesses increase . Following the 2007/8 crisis, a worldwide an excess of 4884 suicides was observed in 2009 [bib_ref] Impact of the economic recession and subsequent austerity on suicide and self-harm..., Corcoran [/bib_ref] and over the next 3 years (2008-2010) an excess of 4750 suicides occurred in the USA, 1000 suicides in England, and 680 suicides in Spain (Lopez [bib_ref] The effect of the late 2000s financial crisis on suicides in Spain:..., Bernal [/bib_ref] [bib_ref] Increase in state suicide rates in the USA during economic recession, Reeves [/bib_ref]. There is also evidence of other increases in poor mental health and wellbeing after the 'Great Recession' including self-harm and psychiatric morbidity [bib_ref] Seeking help in times of economic hardship: access, experiences of services and..., Barnes [/bib_ref] [bib_ref] Trends in population mental health before and after the 2008 recession: a..., Katikireddi [/bib_ref].
However, studies have found that there are important variations in the effects of recessions and economic downturns on population health -depending on policy responses. In a wide ranging and well publicised analysis of the health effects of austerity,concluded that the overall effects of recessions on the health of different nations vary significantly by political and policy context, with those countries (such as Iceland or the USA) who responded to the financial crisis of 2007/8 with an economic stimulus, faring much better -particularly in terms of mental health and suicides -than those countries (e.g. Spain, Greece or UK) who chose to pursue austerity. Similarly, [bib_ref] Financial crisis, austerity, and health in Europe, Karanikolos [/bib_ref] found that across Europe, weak social protection systems increased the health and social crisis in Europe. found that in Thailand and Indonesia where social welfare spending was decreased during the Asian recession of the late 1990s, mortality rates increased. However, in Malaysia where no cut backs occurred, mortality rates were unchanged [bib_ref] Economic stability and health status: evidence from East Asia before and after..., Hopkins [/bib_ref]. Similarly, a study of 26 European countries concluded that greater spending on social welfare could considerably reduce suicide rates during periods of economic downturn [bib_ref] The public health effect of economic crises and alternative policy responses in..., Stuckler [/bib_ref]. In the UK, there is evidence that the pressures that austerity has placed on key social and health care services resulted in up to 10,000 additional deaths in 2018 compared to previous years [bib_ref] Rise in Mortality in England and Wales in First Seven Weeks, Hiam [/bib_ref]. However, there is little evidence on the differential regional impact of the recession. In the UK, areas with higher unemployment rates experienced greater increases in suicide [bib_ref] Impact of the recent recession on self-harm: longitudinal ecological and patient-level investigation..., Hawton [/bib_ref] and studies have demonstrated the mixed effect the recession has had on regions across Europe depending on political decisions and existing institutional frameworks [bib_ref] Regional resilience in the 2008-2010 downturn: comparative evidence from European countries, Davies [/bib_ref].
## Aims and objectives
The aim of this study is to strengthen the evidence base on the influence of regional employment levels (as a measure of the local labour market) and individual worklessness on health by examining these relationships across 16 European countries during the Great Recession. Specifically, we have four key objectives:
1. Examine the influence of regional employment levels and individual worklessness on the health of the working-age population across Europe. 2. Examine whether regional employment levels at different time points, the employment rate across the whole recessionary period, or the change in employment rate during a period of recession and austerity, are associatied with health. 3. Investigate whether regional employment levels are associated with health over and above individual worklessness. 4. Examine whether there is an interaction between individual level worklessness and the regional employment rate.
# Methods
## Data
Individual level data were taken from Round 7 (edition 2.1) of the European Social Survey (ESS) which was collected during 2014/15 (European Social Survey, 2014). Data are openly available and can be accessed by visiting https://www.europeansocialsurvey.org. We included data from 16 European countries (Austria, Belgium, Czech Republic, Denmark, Finland, France, Germany, Hungary, Ireland, Netherlands, Norway, Poland, Portugal, Spain, Sweden and United Kingdom). We excluded Estonia, Israel, Switzerland and Lithuania due to the lack of available and comparable regional level data. The regional level data were extracted from Eurostat (https://ec.europa.eu/ eurostat/) during 2016. All regional data were classified at the second Nomenclature of Territorial Units for Statistics (NUTS-2) level (with a typical population of 800,000 to 3 million) apart from Germany and United Kingdom, as for these countries the ESS data were only available at the NUTS-1 level (population 3-7 million). The NUTS is a hierarchical system for dividing up the economic territory of the European Union based on population, country administrative divisions or geographical units [bib_ref] Regional inequalities in self-reported conditions and non-communicable diseases in European countries: findings..., Thomson [/bib_ref]. The regional data were matched with the individual level ESS data using the corresponding NUTS code. We included people aged 15-64 years (N = 23,078).
## Outcomes
We included a range of health outcomes present in Round 7 of the ESS, these included: self-reported heart or circulation problems, high blood pressure, diabetes, self-rated health, depression and obesity. The regional distribution for these outcomes has been described previously [bib_ref] Regional inequalities in self-reported conditions and non-communicable diseases in European countries: findings..., Thomson [/bib_ref]. As a falsification test we also tested an outcome (allergies) that we did not expect to be associated with individual or regional level employment. The first three outcomes (heart or circulation problems, high blood pressure and diabetes) and the allergy falsification test variable were binary variables, whereby participants were coded yes if they reported having the condition over the past 12 months and no if they did not. Poor self-rated health was defined as those reporting their health as bad or very bad, as opposed to very good, good or fair. Depressive symptoms were measured via the Center for Epidemiologic Depression Scale (CESD-8) [bib_ref] The CES-D scale: a self-report depression scale for research in the general..., Radloff [/bib_ref] , which included eight questions relating to how often the participant felt a range of feelings, such as sadness, loneliness, and happiness, over the past week. Positively worded questions were reverse coded and a scale was derived from the sum of the eight items, which ranged from 0 to 24. Potential cases of depression were coded as those scoring a value of 10 or more [bib_ref] Regional inequalities in self-reported conditions and non-communicable diseases in European countries: findings..., Thomson [/bib_ref]. Obesity was defined as having a Body Mass Index of 30 or more, which was derived from self-reported height and weight.
## Exposures
Individuals were asked about their main activity in the past seven days: in paid work, in education, unemployed and actively looking for a job, unemployed and not actively looking for a job, permanently sick or disabled, retired, in community or military service, doing housework, or looking after children or other persons. Individuals who did not report being in paid work were defined as workless and this was used as our main individual level exposure variable. We chose to categorise individuals into those who were workless versus those who were not because we were interested in the combined influence of worklessness and the regional employment level on health, not just unemployment. We chose to keep students in the workless group as we hypothesised that some outcomes (e.g. depression) may be worse among students in areas of low regional employment and this is consistent with a previous study [bib_ref] Does non-employment contribute to the health disadvantage among lone mothers in Britain,..., Fritzell [/bib_ref]. At the regional level we included the employment rate of those aged 15-64 years during 2008, representing a time period near the beginning of the Great Recession to examine whether this had a lasting association with the health outcomes, and also during 2013, which represents the most recent period preceding the collection of the health outcome data, but also a time at the end of the Great Recession during a period of austerity in some European countries. We also examined the average employment rate between 2008 and 2013 and the difference in the employment rate between 2008 and 2013 to assess whether the change in the employment rate was important over and above the overall rate. We chose to use the employment rate rather than the unemployment rate because employment rates are a more robust indicator of the local labour market and have been used previously in a number of other key studies focused on the influence of the local labour market on health [bib_ref] Changing labour market conditions during the 'great recession' and mental health in..., Curtis [/bib_ref] [bib_ref] Long-term local area employment rates as predictors of individual mortality and morbidity:..., Riva [/bib_ref]. Unemployment rates alone only capture one aspect of recession and local labour market conditions, whereas employment rates also take into account potential increases in those out of work due to sickness, disability and caring responsibilities. Unemployment rates may also fall when there has been no improvement in the local labour market, as the definition of unemployment only covers those who are out of work and actively seeking employment. When unemployment rates are high, those who are disadvantaged for various reasons (e.g. long-term unemployed, low skills, disability) may become discouraged due to increased competition and cease actively searching for employment, resulting in a fall in the unemployment rate . Employment rates are also more likely to be comparable across countries as the definition of the unemployed is more variable between countries [bib_ref] Welfare state regimes, unemployment and health: a comparative study of the relationship..., Bambra [/bib_ref].
## Confounding variables
We included a number of potential individual level confounding factors: age (years), gender, education level (tertiary versus non-tertiary) and marital status (married/cohabiting versus not married/cohabiting). These individual level variables were considered to be associated with both individual level worklessness and the health outcomes, therefore potentially confounding these relationships. All models also included country dummy variables to control for factors at the country level which may be related to both worklessness and the outcome variables, such as the national employment rate.
# Statistical analysis
We firstly examined descriptive statistics, using post-stratification weights, for each variable. We then calculated multilevel random-intercept Poisson regression models for the binary outcome variables, which included individuals nested within NUTS regions. Poisson regression for binary outcome variables enables the calculation of Prevalence Risk Ratios (PRRs), allowing reliable comparison across models and samples, as well as performing well when the outcome is rare [bib_ref] Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models..., Barros [/bib_ref] [bib_ref] Employment status and income as potential mediators of educational inequalities in population..., Katikireddi [/bib_ref]. All models included country fixed effects, which control for the variance in the health outcomes attributed to the country level that may be due to factors such as the national unemployment rate, and therefore reducing the likelihood of omitted variable bias. Due to the relatively small number of countries included, it was not feasible to calculate three-level models. We also included robust standard errors (clustered at the regional level) to account for potential violations to heteroskedasticity. We excluded individuals with missing exposure data (N = 296). Models for each outcome variable may therefore contain a different number of individuals as we did not exclude those with missing values for each outcome. All analyses were performed using Stata/MP 15.1.
The following statistical models were calculated for each outcome variable: we first calculated multilevel regression models including only the regional level employment variables in turn, controlling for country fixed effects. This was followed by the calculation of multilevel regression models which included the individual level confounding variables (age, gender, marital status and education level) and country fixed effects. We then added individual worklessness status followed by the regional level employment variable, and then the interaction between the regional employment and individual worklessness. For any statistically significant interactions, we calculated predictive margins and graphed these to aid the interpretation of results. As a sensitivity analysis, we also tested the interaction between the regional employment rate and gender, as well as education level, to investigate whether there were potential moderating effects. For results of interest we broke the workless category down into employed, unemployed, permanently sick or disabled, retired, homemaker or other (e.g. in education, military or community service) to see if any specific group was driving the results, as a sensitivity analysis.
# Results
Across the 16 countries, allergies were the most prevalent health outcome (14.0%) and diabetes (2.9%) the least [fig_ref] Table 1: Descriptive statistics [/fig_ref].
Heart and circulation problems were most prevalent in Poland (10.5%) and the least in Ireland (2.1%). High blood pressure was highest in Germany (16.4%) and lowest in Ireland (6.0%). Diabetes was most frequently observed in Portugal (5.5%) and the least in Ireland (1.5%). Poor self-rated health was most often observed in Germany (7.7%) and the least in Ireland (1.7%). Depression was most common in Portugal (19.9%) and the least common in Finland (6.9%). Obesity was most prevalent in United Kingdom (18.0%) and the least in Austria (9.1%). Allergies were most common in Portugal (22.6%) and the least common in Hungary and Ireland (5.7%).
Overall, the prevalence of worklessness was 37.7% and it varied from 31.0% in Austria to 45.9% in Ireland . The lowest average regional employment rate in 2008 was observed in Hungary (56.4%), whereas in 2013 it was in Spain (54.7%), but Hungary displayed the lowest average employment rate across the six years. Norway exhibited the highest employment rate at both time periods (78.0% in 2008 but decreasing to 75.4% in 2013). The largest fall in the employment rate between 2008 and 2013 was seen for Spain (9.5%), whereas Germany saw the largest increase (3.6%).
There was little difference in the strength of the associations when comparing the relationship between the 2008 and 2013 regional employment rates and the health outcomes in the multilevel Poisson regression models . The associations observed were in the expected direction, whereby higher regional employment levels were associated with lower risk of poor health and a PRR of below one. The strongest association between the average regional employment rate and the health outcomes was observed for heart and circulation problems where the PRR was 0.971 (95% CI: 0.951 to 0.991) and also obesity with a PRR of 0.971 (95% CI: 0.960 to 0.982). The average regional employment rate was associated with all health outcomes , apart from allergies (0.995, 95% CI: 0.978 to 1.011). The change in the regional employment rate between 2008 and 2013 was not associated with any of the health outcomes under study, above that of the 2008 regional employment rate. As there was little difference between the different measures of regional employment, we chose to conduct the remaining analysis using the average regional employment rate from 2008 to 2013.
We identified gender differences across all health outcomes studied , Model 1). Women had higher risk of heart and circulation problems, poor self-rated health, depression and allergies, whereas men had greater risk of high blood pressure, diabetes and obesity. Being married or cohabiting with a partner was associated with lower risk of all health outcomes apart from obesity. Educational inequalities were apparent across all health outcomes apart from allergies, where the higher educated experienced greater risk. The largest educational inequality was seen for poor self-rated health. Adding individual worklessness to the model attenuated the educational inequalities in health , Model 2).
Worklessness was associated with higher risk of all adverse health outcomes; the strongest association was observed for poor self-rated health (3.554, 95% CI: 3.038 to 4.158) and the weakest for obesity (1.092, 95% CI: 1.015 to 1.174). When the average regional employment rate was added to the models the association between individual worklessness and the health outcomes decreased , Model 3). Regional employment retained an association with heart and circulation problems (0.978, 95% CI: 0.959 to 0.997), depression (0.984, 95% CI: 0.970 to 0.999) and obesity (0.975 , 95% CI: 0.965 to 0.984), over and above the association with individual level worklessness. Sensitivity analysis breaking the workless group down demonstrated that the regional employment level remained associated with both heart and circulation problems and obesity . The association with depression was weakened and no longer statistically significant at p < 0.05. This may be due to the strong association between depression and being out of work due to sickness, disability or unemployment.
When we examined the interaction between individual level worklessness and the regional employment rate, we found statistically significant interactions for depression and poor self-rated health , Model 4). For depression, at higher levels of regional employment the difference in depression and between the employed and workless was larger, with the employed experiencing lower risk of depression at higher levels of regional employment to a greater extent compared to the workless . The sensitivity analysis breaking the workless group down revealed that no specific group was driving the interaction found for poor self-rated health. However, for depression, increased regional employment was related to higher risk of depression for the workless groups who were permanently sick or disabled or classified as 'other' , which includes those in education. Whereas, the other groups experienced lower risk at higher rates of regional employment. In sensitivity analyses, we found no statistically significant interactions between gender, education level and regional employment levels .
# Discussion
Our analysis of worklessness, regional employment and health across 16 European countries revealed that higher regional employment levels were associated with lower risk of depression, heart and circulation problems and obesity, over and above the individual level association with worklessness, as well as country level factors. We also found that individual level worklessness was associated with each health outcome studied. Our results demonstrate that higher regional employment levels during 2008 and 2013, as well as the average regional employment rate across the period of the Great Recession, were associated with reduced risk of most health outcomes. The only outcome not associated with the average regional employment rate was allergies, which we included as a falsification test, and this therefore adds confidence to our findings. In addition, the change in regional employment did not seem to matter above that of the overall regional employment rate in the countries studied. When we investigated the interaction between the regional employment rate and individual level worklessness we found interesting results for the depression and poor self-rated health outcomes. For depression in particular, as regional employment levels increased the inequality in depression between the employed and workless increased, as the employed appeared to benefit more from the higher levels of regional employment. However, when we investigated this in further detail, the results seemed to be driven by the permanently sick or disabled group; the risk of depression increased as regional employment levels increased. Descriptive statistics for worklessness (weighted) and regional employment by country. N=Number of individuals; SD=Standard deviation.
C.L. The finding that higher regional employment levels may be protective against poor health highlights the importance of considering potential contextual influences on health at different geographic levels. Although our study, to our knowledge, is the first cross-national study to examine the influence of regional employment on a range of health outcomes, the results are generally consistent with previous research limited to single countries. Research from England and Scotland demonstrated that neighbourhood unemployment was related to poor self-rated health, in addition to individual level unemployment [bib_ref] Neighbourhood environment and its association with self rated health: evidence from Scotland..., Cummins [/bib_ref]. Other studies find that higher local area unemployment may confer some protection against psychological distress associated with being unemployed [bib_ref] Do local unemployment rates modify the effect of individual labour market status..., Flint [/bib_ref]. Similarly, the impact of unemployment on wellbeing was found to be less in areas of high unemployment in South Africa [bib_ref] Are there geographical variations in the psychological cost of unemployment in South..., Powdthavee [/bib_ref]. Evidence from Sweden demonstrated that higher vacancy rates at the municipal level was related to better mental health among the unemployed, but the unemployment rate had little influence [bib_ref] Mental health among the unemployed and the unemployment rate in the municipality, Strandh [/bib_ref]. Our findings for depression echo those of [bib_ref] The social norm of unemployment in relation to mental health and medical..., Buffel [/bib_ref] who found that the employed were more depressed in regions with high unemployment rates, resulting in a narrowing of the mental health gap between the unemployed and the employed.
Taken together, our findings and those of other studies (as detailed above) highlight that the local labour market may be just as important for health as individual employment status (particularly for more sensitive outcomes such as mental health and those related to unhealthy behaviours). Further, people who are employed but living in households where another household member is unemployed or out of work due to other factors (e.g. illness or caring responsibilities) may also be impacted indirectly by declining regional employment levels. This could be due to the stress and worry related to the local labour market that may lead to increases in depression and anxiety, as well as the adoption of less healthy behaviours (e.g. poor diet) as a coping mechanism, and lower physical activity due to a decrease in work-related exertion [bib_ref] Exercise, physical activity, and exertion over the business cycle, Colman [/bib_ref] , which could lead to obesity. Our findings are also supported by a recent study which used Scottish longitudinal data across the period of the Great Recession [bib_ref] Changing labour market conditions during the 'great recession' and mental health in..., Curtis [/bib_ref]. It found that for people living in areas that had experienced relatively high and stable levels of employment, the likelihood of reporting a mental illness was significantly lower when compared to similar people living in areas with persistently low employment rates. The authors also note that the trajectory of local area employment during the recession seems to have a stronger association with mental health compared to individual level employment status. Our study found that the change in regional employment during the time of the Great Recession did not have an additional association with the health outcomes under study, above that of the overall rate. However, the direction of the association between the change in regional employment and the more sensitive health outcomes in our study, including depression and obesity, was such that a rise in regional employment was suggestive of a protective association. These results reinforce the contextual effects literature which asserts that population health is shaped by both individual and area-level factors.
The lack of a clear association for the change in employment rates in this study may be due to the very mixed impact of, and response to, the recession on different regions across Europe [bib_ref] Regional resilience in the 2008-2010 downturn: comparative evidence from European countries, Davies [/bib_ref]. The economic downturn associated with the Great Recession caused regional employment to drop in most countries across Europe (the average drop in employment between 2008 and 2013 was 1.7%), the highest regional decline occurred in Spain (−9.5%) and increased the most in Germany (3.6%). Macroeconomic policies which influence the demand for labour may be consequently important for health. Recessions impact population health unevenly depending on whether an economic stimulus approach was followed or austerity . Further work is needed to explore the unequal regional consequences of austerity for health and other potential mechanisms through which the regional labour market may influence health and health behaviours.
## Table 3
Results from multilevel regression models investigating regional employment levels and health outcomes.
Heart or circulation problems High blood pressure
# Strengths and limitations
The key strength of our study was the use of comparable cross-national data that integrated individual and regional level variables and included a range of health outcomes, whilst controlling for country fixed effects. We also investigated the full range of employment statuses, unlike previous studies which have often been limited to the employed and unemployed. Although we investigated a broad workless group, we also examined specific employment status groups to see if any were dictating the key results. Our study is unfortunately limited by the use of self-reported data which may be subject to reporting bias. However, our research builds on existing cross-national studies in the area of employment and health in which self-rated health has often been the sole outcome variable [bib_ref] The impacts of job loss and job recovery on self-rated health: testing..., Huijts [/bib_ref] [bib_ref] To what extent do financial strain and labour force status explain social..., Shaw [/bib_ref] [bib_ref] Unemployment transitions and self-rated health in Europe: a longitudinal analysis of EU-SILC..., Tøge [/bib_ref]. Our study is also cross-sectional which precludes the inference of causal relationships. It also should be noted that the European Social Survey was never sampled for analysis at the regional level, so it may be possible that the respondents in some countries are not representative of the population at the sub-national level [bib_ref] Regional inequalities in self-reported conditions and non-communicable diseases in European countries: findings..., Thomson [/bib_ref]. Data for some of the countries included was also only available at the largest NUTS level and we had to exclude a number of countries from the analysis due to a lack of comparable regional data. In addition, as we only examined regional employment levels we cannot rule out the possibility that our results are affected by other regional level factors, such as air pollution, Gross Domestic Product, income inequality and working conditions. However, it is likely that these may be causally related to regional employment levels and may therefore represent potential mediators rather than confounders. Similarly, although we included country fixed effects in our analyses to control for potential country level heterogeneity and to help reduce the possibility of omitted variable bias, there is still a possibility of residual confounding due to factors at the national level, such as the welfare and healthcare systems.
Our study is one of the first to investigate the impact of regional employment levels on a variety of health outcomes across the European working-age population. There was evidence to suggest that higher regional employment levels may exhibit a protective influence on some health outcomes, such as obesity and depression. These outcomes may be considered more sensitive to changes in the local environment, whereas the other outcomes such as diabetes may not be as sensitive with longer lag times if any effect is present. Future research would benefit from taking a life course approach to explore whether there are particular phases of the life course in which worklessness and regional labour markets relate to poor health (e.g. upon leaving education or late working life) [bib_ref] Life course of place: a longitudinal study of mental health and place, Pearce [/bib_ref]. Further research is needed to test potential causal links between regional employment and health by examining other factors in the local and regional environment which may be affected by employment levels, such as job insecurity, social cohesion, income and welfare spending [bib_ref] Area-level unemployment and perceived job insecurity: evidence from a longitudinal survey conducted..., Milner [/bib_ref] [bib_ref] Social protection spending and inequalities in depressive symptoms across, Niedzwiedz [/bib_ref]. More in-depth investigation of the potential biological mechanisms would also be fruitful [bib_ref] Unemployment and inflammatory markers in England, wales and Scotland, 1998-2012: meta-analysis of..., Hughes [/bib_ref] [bib_ref] Economic insecurity during the Great Recession and metabolic, inflammatory and liver function..., Niedzwiedz [/bib_ref]. However, at present, there are a lack of comparable crossnational health data amongst the working-age population, with outcomes often limited to self-rated health. We also did not explore the impact of changing regional inequality in response to the recession. However, evidence suggests, at least for some Central and Eastern European countries, that regional variations in unemployment decreased following the Great Recession [bib_ref] Regional unemployment impacts of the global financial crisis in the new member..., Blažek [/bib_ref]. It will be important to fully explicate the long-term population health effects of the austerity measures implemented across some European countries and their likely uneven spatial distribution.
# Conclusion
This cross-European study emphasises the importance of macroeconomic processes in understanding population health in Europe, but that there can be distinct geographical differences in the impact of these processes. In particular, our findings emphasise that worklessness is likely to be an important determinant of a variety physical and mental health outcomes. In addition, the work suggests that macroeconomic factors might exert an additional influence on the health of Europeans through the contextual influence of regional employment levels. This study therefore emphasises the importance of examining the geographical specificities of the political economy, and in particular, the multiple socio-spatial pathways through which structural factors exert an influence on population health and inequalities. Further work in this area will not only provide greater clarity as to the pathways linking macroeconomic change and health, but will also help to identify . Results from the multilevel regression models examining the average regional employment rate from 2008 to 2013 and health outcomes (95% confidence intervals shown), PRR = prevalence risk ratio.
C.L. 112377 Results from multilevel regression models investigating individual worklessness, regional employment levels and health outcomes.
Heart or circulation problems geographically-specific factors that can exacerbate or mitigate against these processes. These insights can provide policy makers with new insights into how to foster greater resilience to structural changes including financial shocks, such as economic recession.
# Declarations of interest
None.
## Fig. 2.
Predicted probability of depression for employed and workless groups according to the average regional employment rate from 2008 to 2013 (95% confidence intervals shown).
[table] Table 1: Descriptive statistics (weighted) for each health condition by country. [/table]
|
In vitro germ cell induction from fertile and infertile monozygotic twin research participants
Graphical abstract Highlights d Monozygotic (MZ) monoamniotic (MA) twins with discordant POI are evaluated here d Each twin's hiPSCs lines make separate amniotic sacs with germ cells in vitro d Competency to generate germ cells is equivalent between twins with discordant POI SUMMARYHuman induced pluripotent stem cells (hiPSCs) enable reproductive diseases to be studied when the reproductive health of the participant is known. In this study, monozygotic (MZ) monoamniotic (MA) twins discordant for primary ovarian insufficiency (POI) consent to research to address the hypothesis that discordant POI is due to a shared primordial germ cell (PGC) progenitor pool. If this is the case, reprogramming the twin's skin cells to hiPSCs is expected to restore equivalent germ cell competency to the twins hiPSCs. Following reprogramming, the infertile MA twin's cells are capable of generating human PGC-like cells (hPGCLCs) and amniotic sac-like structures equivalent to her fertile twin sister. Using these hiPSCs together with genome sequencing, our data suggest that POI in the infertile twin is not due to a genetic barrier to amnion or germ cell formation and support the hypothesis that during gestation, amniotic PGCs are likely disproportionately allocated to the fertile twin with embryo splitting. investigation, E.C.P.; genomic analysis, F.-M.H., M.D., and E.C.P.; funding acquisition, A.T.C., E.C.P., and S.J.S.; writing -original draft, E.ReportllOPEN ACCESSProcessing aggregates for 10x genomics Aggregates at day 4 of differentiation from MZT01, MZT02 MZT04, MZT05, MZT06 hiPSC sublines (15 sublines) were dissociated to single cells using Trypsin (0.05%). The single cell suspensions were resuspended in 13 PBS with 0.04% Bovine Serum Albumin
## Correspondence [email protected]
In brief hiPSCs to evaluate germ cell and amnion formation in MZ MA twins discordant for primary ovarian insufficiency (POI). Results suggest a barrier to germ cell formation such that the infertile twin is unable to generate sufficient germ cells to overcome POI as young adult, which reprogramming to hiPSCs can reverse.
# Introduction
The earliest human embryonic stem cell (hESC) lines were derived from isolated inner cell masses of human blastocysts consented to research. [bib_ref] Embryonic stem cell lines derived from human blastocysts, Thomson [/bib_ref] Almost a decade later, human induced pluripotent stem cell (hiPSC) lines were derived. [bib_ref] Induction of pluripotent stem cells from adult human fibroblasts by defined factors, Takahashi [/bib_ref] [bib_ref] Induced pluripotent stem cell lines derived from human somatic cells, Yu [/bib_ref] Since then, thousands of patient-specific hiPSC lines from ''healthy'' or ''disease'' research participants have been generated, catalyzing stem cell science in exciting and collaborative ways. [bib_ref] EBiSC best practice: how to ensure optimal generation, qualification, and distribution of..., Steeg [/bib_ref] [bib_ref] Korea national stem cell bank, Kim [/bib_ref] [bib_ref] UK stem cell bank, O'shea [/bib_ref] Critical to the use of pluripotent stem cells is the informed consent process and a framework for scientific and ethical oversight. [bib_ref] Obtaining consent for future research with induced pluripotent cells: opportunities and challenges, Aalto-Setä Lä [/bib_ref] Certain areas of basic science research with pluripotent stem cells are considered sensitive, such as reproductive science research and the differentiation of germ cells and gametes, a technique called in vitro gametogenesis (IVG).
Bioethicists consider the differentiation of germ cells from pluripotent stem cells without the intent to make embryos ethically similar to differentiating somatic cells. [bib_ref] Obtaining consent for future research with induced pluripotent cells: opportunities and challenges, Aalto-Setä Lä [/bib_ref] However, patient and community perspectives in this area are starting to emerge, with a recent study revealing that research participants deem gonadal organoids (containing germ cells) as morally distinct from other types of organoid research. [bib_ref] Patients' perspectives on the derivation and use of organoids, Bollinger [/bib_ref] Balancing patient perspectives with the current state of the science, the 2021 update to the International Society for Stem Cell Research (ISSCR) ''Guidelines for Stem Cell Research and Clinical Translation'' recommend that differentiation of primordial germ cells (PGCs) from pluripotent stem cells in non-integrated stem cell-based embryo models be reportable to a specialized scientific and ethics oversight review process but not normally subject to further review. [bib_ref] Human embryo research, stem cell-derived embryo models and in vitro gametogenesis: considerations..., Clark [/bib_ref] In contrast, experiments involving use of IVG-derived gametes to generate embryos for research purposes should be subject to a specialized review process. [bib_ref] Human embryo research, stem cell-derived embryo models and in vitro gametogenesis: considerations..., Clark [/bib_ref] Deriving disease-specific hiPSC lines for research tends to focus on devastating lethal diseases. Absent from the disease lists in most iPSC repositories is infertility. Infertility is a disease of the reproductive system defined as a failure to achieve a pregnancy after 12 months of trying.Current estimates indicate that infertility affects between 48 million couples and 186 million individuals globally. [bib_ref] National, regional, and global trends in infertility prevalence since 1990: a systematic..., Mascarenhas [/bib_ref] [bib_ref] International estimates of infertility prevalence and treatment-seeking: potential need and demand for..., Boivin [/bib_ref] Causes of infertility are varied and diagnosed in all genders; however, a failure to specify PGCs will cause certain infertility given that gametes (which originate from PGCs) are the only cells in the body capable of fertilization.
In the adult ovary, a lack of germ cells causes ovarian failure, also referred to as primary ovarian insufficiency (POI). A lack of germ cells in a pre-pubescent child's ovaries will result in failure to transition through puberty and POI. POI is not a rare disease, with 1% of women experiencing ovarian failure before the age of 40. [bib_ref] Premature ovarian failure, Goswami [/bib_ref] A baby assigned female at birth does not generate new oocytes after birth,which is why establishment of an appropriate number of germ cells and oocytes during the prenatal window is a critical determinant of whether a person will experience POI.
POI is more common among twins than the general population, with monozygotic (MZ) twins affected at 3-to 5-fold higher rates than un-related individuals. [bib_ref] Ovarian transplantation between monozygotic twins discordant for premature ovarian failure, Silber [/bib_ref] [bib_ref] Ovarian transplantation in a series of monozygotic twins discordant for ovarian failure, Silber [/bib_ref] [bib_ref] Prevalence of premature ovarian failure in monozygotic and dizygotic twins, Gosden [/bib_ref] MZ, monochorionic (MC), monoamniotic (MA) twins (also called MA twins) are a rare subset of twins occurring in 1 in every 100 sets of MZ twin births. 20,21 MA twin pairs are especially useful for investigating potential epigenetic causes of discordant diseases, including POI, as they are genetically similar, and the twins shared the same in utero environment prior to birth. [bib_ref] The embryology of conjoined twins, Kaufman [/bib_ref] [bib_ref] Non-identical monozygotic twins, intermediate twin types, zygosity testing, and the non-random nature..., Machin [/bib_ref] An MA twin pair splits from a single MZ embryo between days 8-13 post-fertilization (pf) or at Carnegie stage (CS) 5b-c and thus share a placenta, an amnion, and a chorion. Importantly, MA twin splitting occurs around the time PGC specification, which begins at CS5b in primates, [bib_ref] Human primordial germ cells are specified from lineage-primed progenitors, Chen [/bib_ref] [bib_ref] The germ cell fate of cynomolgus monkeys is specified in the nascent..., Sasaki [/bib_ref] suggesting that these twins may also share a PGC progenitor pool located in the amnion. Given this, the hypothesis to be addressed is that reprogramming somatic cells from MA twin pairs discordant for POI into hiPSCs could reset the epigenome of the affected twin such that she is able to produce PGC-like cells (human PGC-like cells and amnion equivalent to her fertile sister.
# Results
## Human subject selection
Participants were consented into this research study from a cohort of MZ twins discordant for POI who were treated at the Infertility Center of St. Louis (MO, USA). Three sets of MZ twins (two of which are known to be MA) donated skin punch biopsies to this study [fig_ref] Figure 1: Derivation of HDFs and hiPSCs from twin pairs with discordant POI [/fig_ref]. MZT01 was diagnosed with early-onset menopause due to POI at age 25. Her twin sister, MZT02, a woman with normal fertility during her reproductive years, successfully gave birth without intervention. MZT01 underwent ovarian transplant surgery where she received an ovary from her fertile twin sister MZT02 at the Infertility Center of St. Louis, and following this procedure, she subsequently gave birth to two children . The second twin pair consists of MZT03, a woman diagnosed with early-onset menopause due to POI at age 31, and her twin sister, MZT05, a woman with normal fertility during her reproductive years who successfully gave birth without intervention. MZT03 underwent ovarian transplant surgery, receiving an ovary from her fertile twin sister MZT05 at the Infertility Center of St. Louis, and she subsequently gave birth to one child. MZT03 was also diagnosed with leukemia and underwent radiation and chemotherapy treatment prior to consenting to this study. The final twin pair consists of MZT04, a woman with Report ll OPEN ACCESS normal fertility during her reproductive years, and her twin sister, MZT06, a woman with early-onset menopause due to POI at age 22. MZT06 received an ovary transplant from her twin sister MZT04 and subsequently gave birth to three children. [bib_ref] A series of monozygotic twins discordant for ovarian failure: ovary transplantation (cortical..., Silber [/bib_ref] Following informed consent, the women donated a skin punch biopsy, and human dermal fibroblasts (HDFs) were successfully isolated. All HDF samples were karyotypically normal except MZT03, which did not yield karyotypically normal HDFs despite two skin biopsy donations. Therefore, this sample was not used in further experiments. The remaining five HDF samples were reprogrammed into hiPSCs, with three colonies picked for further analysis resulting in karyotypically normal, self-renewing pluripotent sublines that are genetically identical to their parental HDFs as confirmed by short tandem repeat (STR) analysis [bib_ref] Generation of three human induced pluripotent stem cell sublines (UCLAi004-A, UCLAi004-B, and..., Pandolfi [/bib_ref] [fig_ref] Figure 1: Derivation of HDFs and hiPSCs from twin pairs with discordant POI [/fig_ref]. Three hiPSC sublines from each research participant were used as biological replicates in the following experiments.
Whole-genome sequencing (WGS) suggests no causative POI mutations in the twins As STR does not evaluate POI candidate genes, we first sought to expand the genome analysis to include WGS (20 samples in total) . Jaccard indices 29 were used to calculate the relatedness of HDF samples, confirming high similarity between twin pairs MZT04/MZT06 and MZT01/MZT02 (values close to 1.0) as well as their unrelatedness to MZT05 [fig_ref] Figure 1: Derivation of HDFs and hiPSCs from twin pairs with discordant POI [/fig_ref]. Next, we queried the sequence of 22 genes previously associated with POI . No genomic discrepancies between twins were observed in the protein-coding regions of these 22 genes; however, single-nucleotide variants (SNVs) were identified in an intron of FSHR and PMM2 in MZT01 and MZT06 (POI participant) but not their fertile twin sisters MZT02 or MZT04. These SNVs were identified as common SNPs in the human population and have not been reported as pathological in dbSNP. The reprogramming process is known to introduce genomic alterations. 31-37 Given that multiple hiPSC sublines were derived from each HDF sample, we next sought to assess the number of genomic changes acquired by each hiPSC subline. We applied two approaches (GATK and Strelka), 38-40 standard bioinformatics approaches for analyzing cancer genomes, treating the hiPSCs as ''tumor'' and the matched HDFs as ''normal.'' To generate a stringent call set, we retained only genomic changes detected by both methods and an allelic frequency greater than 10%. [bib_ref] Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples, Cibulskis [/bib_ref] Consistent with previous reports, we found that the number of mutations varied between sublines following reprogramming [fig_ref] Figure 1: Derivation of HDFs and hiPSCs from twin pairs with discordant POI [/fig_ref] ; , with the average number of acquired mutations being 1.2 mutations/Mb [fig_ref] Figure 1: Derivation of HDFs and hiPSCs from twin pairs with discordant POI [/fig_ref]. The majority of mutations correspond to SNVs of the C > T or T > C type. Four hiPSC sublines (MZT02-G, MZT02-H, MZT01-N, and MZT04-J) acquired >10,000 mutations [fig_ref] Figure 1: Derivation of HDFs and hiPSCs from twin pairs with discordant POI [/fig_ref].
Induced reprogramming restores germ cell competency to the infertile twin In order to assess germ cell differentiation in each hiPSC subline, we induced the hiPSCs into incipient mesoderm-like cells (iM-eLCs) [bib_ref] Robust in vitro induction of human germ cell fate from pluripotent stem..., Sasaki [/bib_ref] followed by differentiation as three-dimensional (3D) aggregates in round-bottom low-adhesion 96-well plates in media containing BMP4 and other cytokines. [bib_ref] Robust in vitro induction of human germ cell fate from pluripotent stem..., Sasaki [/bib_ref] The percentage of hPGCLCs in the aggregates was quantified using fluorescence-activated cell sorting (FACS) at day 4 of aggregate differentiation with integrin alpha 6 (ITGA6) and epithelial cell adhesion molecule (EPCAM), two cell-surface markers of hPGCs [fig_ref] Figure 2: Percentage of hPGCLCs differentiated from hiPSCs The hiPSC sublines for each participant [/fig_ref]. [bib_ref] Robust in vitro induction of human germ cell fate from pluripotent stem..., Sasaki [/bib_ref] [bib_ref] Germline competency of human embryonic stem cells depends on eomesodermin, Chen [/bib_ref] There was no significant difference in hPGCLC percentage when comparing MZT01 with MZT02 or MZT04 with MZT06. Furthermore, hiPSCs derived from MZT05, the fertile twin sister of MZT03, also produced comparable percentages of hPGCLCs to the other participants. Secondary analysis comparing sublines derived from the same participant revealed no significant difference [fig_ref] Figure 2: Percentage of hPGCLCs differentiated from hiPSCs The hiPSC sublines for each participant [/fig_ref] , indicating that the especially high numbers of DNA mutations in MZT02-G, MZT02-H, MZT01-N, and MZT04-J did not alter hPGCLC competency. Immunofluorescence (IF) analyses for PGC markers SOX17, PRDM1, and TFAP2C [bib_ref] DNA demethylation dynamics in the human prenatal germline, Gkountela [/bib_ref] [bib_ref] The transcriptome and DNA methylome landscapes of human primordial germ cells, Guo [/bib_ref] [bib_ref] A unique gene regulatory network resets the human germline epigenome for development, Tang [/bib_ref] [bib_ref] SOX17 is a critical specifier of human primordial germ cell fate, Irie [/bib_ref] [fig_ref] Figure 2: Percentage of hPGCLCs differentiated from hiPSCs The hiPSC sublines for each participant [/fig_ref] further verified hPGCLC identity at day 4.
Germline and somatic gene expression is equivalent between the twins Next, we performed single-cell RNA sequencing (scRNA-seq) using 10X Genomics to evaluate gene expression profiles of hPGCLCs and somatic cells in all hiPSC sublines . Single cells from each subline were collected at day 4 of aggregate differentiation consistent with previous studies. [bib_ref] Human primordial germ cells are specified from lineage-primed progenitors, Chen [/bib_ref] [bib_ref] Germline competency of human embryonic stem cells depends on eomesodermin, Chen [/bib_ref] [bib_ref] Female human primordial germ cells display X-chromosome dosage compensation despite the absence..., Chitiashvili [/bib_ref] [bib_ref] TRIM28-Regulated transposon repression is required for human germline competency and not primed..., Tao [/bib_ref] [bib_ref] Human reproduction is regulated by retrotransposons derived from ancient Hominidae-specific viral infections, Xiang [/bib_ref] The hPGCLC population within the aggregate was defined as a clearly separated cluster expressing hPGC markers NANOG, SOX17, NANOS3, and PRDM1 [fig_ref] Figure 3: Differentiation of each MZ twin's hiPSCs yields hPGCLCs with similar identity [/fig_ref]. Absence of DAZL [fig_ref] Figure 3: Differentiation of each MZ twin's hiPSCs yields hPGCLCs with similar identity [/fig_ref] indicates that the hPGCLCs are in an early stage and have not undergone determination to create committed hPGCs (also called late-stage PGCs or gonocytes). In a principle-component analysis (PCA), all hPGCLCs clustered near hPGCs from CS7 human embryos 53 [fig_ref] Figure 3: Differentiation of each MZ twin's hiPSCs yields hPGCLCs with similar identity [/fig_ref] , reinforcing the notion that hPGCLCs in the current study are equivalent to early hPGCs rather than committed hPGCs in the embryonic gonad. Furthermore, the hPGCLCs from each twin clustered together regardless of their fertility states [fig_ref] Figure 3: Differentiation of each MZ twin's hiPSCs yields hPGCLCs with similar identity [/fig_ref] , with no statistically significant difference in hPGCLC gene expression between the twin pairs, including expression levels of key hPGCLC/hPGC genes [fig_ref] Figure 3: Differentiation of each MZ twin's hiPSCs yields hPGCLCs with similar identity [/fig_ref]. These data indicate that the stage and transcriptional identity of hPGCLCs differentiated from hiPSC lines derived from all research participants are very similar regardless of fertility diagnosis and correspond to hPGCs in vivo at CS7.
Transcriptome analysis of somatic cells in the aggregate revealed expression of somatic lineage markers [fig_ref] Figure 3: Differentiation of each MZ twin's hiPSCs yields hPGCLCs with similar identity [/fig_ref]. This includes rare cells expressing TBXT and MIXL1, which likely mark primitive streak-like cells. FOXA2+ cells, which likely mark endoderm, as well as a large fraction of HAND1+ cells, of which a subset express TFAP2A, IGFBP5, and GABRP, recently identified markers of amnionic ectoderm. [bib_ref] Capturing human trophoblast development with naive pluripotent stem cells in vitro, Io [/bib_ref] These cells, which we putatively call ''amnion-like cells,'' clustered with the amnion-like cells identified by Zheng et al. 55 [fig_ref] Figure 3: Differentiation of each MZ twin's hiPSCs yields hPGCLCs with similar identity [/fig_ref] , suggesting that 3D aggregate differentiation of hiPSCs in the presence of BMP4, LIF, and ROCKi generates cells by day 4 that resemble amniotic ectoderm.
Each twin is competent to generate her own amniotic sac containing hPGCLCs The gestational history of two MZ twin pairs in this study indicate that, while in utero, MZT01/MZT02 and MZT04/MZT06 shared an amnion and a chorion. Given that hPGCs in primates in vivo are specified around the time of amnion formation, [bib_ref] The germ cell fate of cynomolgus monkeys is specified in the nascent..., Sasaki [/bib_ref] we next asked whether each twin can generate her own amniotic saclike structure using a non-integrated embryo model. To achieve this, we used a microfluidics approach to generate posterior embryonic-like sacs, 55 which recapitulates in vivo 3D amniotic sac tissue architecture and spatiotemporal lineage development reminiscent of those in the early post-implantation human embryos at the time of PGC specification [fig_ref] Figure 4: Induction of hPGCLCs in an embryo model of the amniotic sac [/fig_ref]. To create this embryo model, hiPSCs from each twin were injected into microfluidic devices, and 30 h after exposure to BMP4, individual embryo models formed, containing a squamous amniotic [fig_ref] Figure 4: Induction of hPGCLCs in an embryo model of the amniotic sac [/fig_ref]. The presence of amniotic ectodermlike cells within the embryo model were confirmed using IF for TFAP2A [fig_ref] Figure 4: Induction of hPGCLCs in an embryo model of the amniotic sac [/fig_ref]. The emergence of hPGCLCs was evaluated using IF for NANOG, TFAP2C, and SOX17 with an average of 2-3 hPGCLCs identified as NANOG+ TFAP2C+ SOX17+ triple-positive cells, in the amniotic ectoderm-like cell layer [fig_ref] Figure 4: Induction of hPGCLCs in an embryo model of the amniotic sac [/fig_ref]. There was no statistical difference in the number of hPGCLCs in the amniotic ectoderm-like cell layer generated from each twin [fig_ref] Figure 4: Induction of hPGCLCs in an embryo model of the amniotic sac [/fig_ref]. We also identified hPGCLCs in the pre-primitive streak EPI [fig_ref] Figure 4: Induction of hPGCLCs in an embryo model of the amniotic sac [/fig_ref] as previously reported, 55 consistent with the porcine model of PGC specification. [bib_ref] Principles of early human development and germ cell program from conserved model..., Kobayashi [/bib_ref]
# Discussion
In this study, we consented MZ twins with discordant POI. Two sets of unrelated twins in this study had gestational histories indicating that each pregnancy involved a twin pair gestating within a single amnion. [bib_ref] Ovarian transplantation between monozygotic twins discordant for premature ovarian failure, Silber [/bib_ref] [bib_ref] Ovarian transplantation in a series of monozygotic twins discordant for ovarian failure, Silber [/bib_ref] [bib_ref] A series of monozygotic twins discordant for ovarian failure: ovary transplantation (cortical..., Silber [/bib_ref] Given recent evidence that specified primate PGCs are situated in the dorsal amnion at CS 5b prior to gastrulation, a possible scenario is that MA twins also share the amniotic PGC progenitor pool. Therefore, it could be hypothesized that at the time of embryo splitting, shared amniotic PGC progenitors disproportionally allocate to the fertile twin. Disproportionate allocation of cells during embryo cleavage is a phenomenon supported by the variable allelic frequency range observed between twin pairs after splitting. [bib_ref] Differences between germline genomes of monozygotic twins, Jonsson [/bib_ref] WGS was performed to exclude the possibility that the infertile twin had disproportionately acquired cells with mutations in POI-associated genes that could also explain the discordant POI phenotype. Instead, our data support the hypothesis that at the time of embryo splitting, an epigenetic barrier to PGC specification was likely established, and the infertile twin was unable to generate a sufficient cohort of additional PGCs in order to overcome POI as young adult.
Given that MA twins share an amniotic sac in utero, we used a non-integrated human embryo model to evaluate amniotic-like sac formation from each participant's cells. Using hiPSCs, we demonstrated that all research participants are competent to generate their own amniotic sac-like structures containing equivalent numbers of hPGCLCs. Thus, the single amnion in these MA twins is likely due to the timing of embryo splitting and not a genetic barrier to amnion formation. This analysis does not prove that hPGCs in vivo are specified exclusively in the amnion. Rather, our data indicate that specified hPGCs are consistently and reliably identified in the amniotic ectoderm-like cell layer in vitro, and we propose it is likely the extra-embryonic pool of PGCs that are disproportionally allocated to the fertile sister in the MA twins discordant for POI.
Prior to this study, we reported that the HDFs and hiPSCs used for this study were karyotypically normal. [bib_ref] Generation of three human induced pluripotent stem cell sublines (UCLAi004-A, UCLAi004-B, and..., Pandolfi [/bib_ref] WGS in the current study revealed variable subkaryotypic changes in each twin's hiPSCs relative to the original HDFs. These genomic changes occurred with reprogramming and corresponded to as few as 276 acquired SNVs in MZT01-F to as high as 332,530 acquired SNVs in MZT02-G. Our data corroborate previously published studies demonstrating that large differences in the number and type of genomic changes can be identified when comparing different hiPSC subclones derived from the same individual. Similar to other studies, the hiPSC sublines generated here also contained a higher fraction of T > C or C > T SNVs, an occurrence associated with high rates of hydrolytic deamination of cytosine bases. [bib_ref] Signatures of mutational processes in human cancer, Alexandrov [/bib_ref] Our study revealed that the hiPSC sublines containing the highest numbers of mutations do not show any difference in hPGCLC differentiation potential. Therefore, a high frequency of genomic mutations acquired with reprogramming does not serve as a barrier to hPGCLC differentiation in vitro or as selectable criteria for excluding hiPSC sublines from downstream studies. Genomic alterations in HDF-derived hiPSCs advances our understanding of the genomic impact of reprogramming and culturing on human cells, providing a reference point for discussions of tolerable mutation level when considering the safety of IVG for potential future reproductive purposes. Critically, in the current study, all hiPSC sublines had a higher number of genomic changes than would be anticipated during human germline development in vivo. [bib_ref] Whole genome characterization of sequence diversity of 15, Jó Nsson [/bib_ref] [bib_ref] ) antibodies. Double positive cells were collected using an ARIA-H Fluorescence Activated..., Wong [/bib_ref] We believe that this concern must be addressed before using gametes generated by IVG for reproductive purposes.
The work presented here shows that when epigenetic reprogramming is used to create hiPSCs from MA twins discordant for POI, germ cell differentiation from the resulting hiPSC sublines is equivalent regardless of whether aggregate or embryo models are used. Our ability to derive hPGCLCs from these hiPSC sublines with a similar transcriptome to hPGCs from CS7 embryos, suggests that these hPGCLCs may have the potential to differentiate further into oocytes under appropriate culture conditions.
## Limitations of the study
This work used hiPSCs to model developmental events occurring at the time of amnion formation and MA twinning, which could be used to explain the high incidence of discordant POI in MA twins. Although our work demonstrates that hiPSCs derived from each MA twin pair have equivalent capacity to induce germ cells and develop amniotic sacs, this study does not evaluate post-natal or adult stages of germ cell and follicle formation, where discordant POI phenotypes could also arise.
Unfortunately, the technologies for differentiating follicles from hiPSCs capable of folliculogenesis do not currently exist; however, once these technologies are established, future studies could address this. In addition, given the diversity in POI etiologies, extrapolation of these findings to other MA twins outside of this study is unclear.
# Star+methods
Detailed methods are provided in the online version of this paper and include the following:
[formula] d KEY RESOURCES [/formula]
# Materials availability
hiPSC lines used in this study [bib_ref] Generation of three human induced pluripotent stem cell sublines (UCLAi004-A, UCLAi004-B, and..., Pandolfi [/bib_ref] were generated at UCLA and are available upon request to the Lead Author with MTA and appropriate institutional approvals for working with human induced pluripotent stem cells.
Data and code availability scRNA-seq data in this paper have been deposited at GEO Database: GSE181205 and are publicly available as of the date of publication. The accession number is also listed in the key resources table.
Whole genome sequencing data have been deposited at NCBI BioProject Database: PRJNA759332 and are available upon the date of publication.
Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.
## Experimental model and subject details
## Human dermal fibroblasts (hdfs)
The karyotypically normal human dermal fibroblast (HDF) samples used in this study were previously published. [bib_ref] Generation of three human induced pluripotent stem cell sublines (UCLAi004-A, UCLAi004-B, and..., Pandolfi [/bib_ref] The names of the HDFs are; MZT01, MZT02, MZT04, MZT05 and MZT06. These fibroblasts originated from a skin punch biopsy donated from the five women who at the time of biopsy were aged 39-53 as indicated in . MZT01 and MZT06 were previously diagnosed with POI, their respective twin sisters MZT02 and MZT04, had normal fertility during their reproductive years. Upon thawing, HDFs are cultured at 37 C, 5.0% CO 2 on tissue culture treated plates coated with 0.1% gelatin (Sigma) in media consisting of 15% Fetal Bovine Serum; FBS (GE Healthcare), 1% Non-Essential Amino Acids (Invitrogen), 1% Glutamax (GibcoTM), 1% Penicillin-Streptomycin-Glutamine (Gibco) and Primocin (Invogen). Consent to a skin biopsy, generation of HDFs, generation of hiPSCs and differentiation of hiPSCs was approved and annually reviewed by the UCLA Institutional Review Board (IRB #16-001176) together with additional approval and annual review by the UCLA Human Pluripotent Stem Cell Research and Oversight (hPSCRO) Committee (hPSCRO #2016-003). Mycoplasma was regularly tested before banking using the MycoAlert kit from Lonza Catalog #LT07-318.
## Human induced pluripotent stem cell (hipsc) lines
For each HDF sample, we used n = 3 previously published hiPSC sublines, [bib_ref] Generation of three human induced pluripotent stem cell sublines (UCLAi004-A, UCLAi004-B, and..., Pandolfi [/bib_ref] which were generated under the same consent and approval process as the HDFs described above (IRB, [bib_ref] Robust in vitro induction of human germ cell fate from pluripotent stem..., Sasaki [/bib_ref] with minor modifications adopted by Chen et al. [bib_ref] Germline competency of human embryonic stem cells depends on eomesodermin, Chen [/bib_ref] The detailed methods described above use the modified protocol adopted by To compare the in vitro hPGCLC aggregates with in vivo PGCs, valid cells and UMIs were determined by UMI-tools to generate whitelist. Reads corresponding to valid barcodes were aligned to GRCh38 with STAR 2.7, and only uniquely mapped reads were kept for further analyses. Count matrices were generated by featureCounts v2.0.1 from the Subread R package, with UMIs info further appened to the alignment .bam file. Finally, the count matrix of all valid cells was generated with umi_tools count function.
Amnion-like cells within the aggregates derived from one hiPSC line from each human participant were identified based on expression of IGFBP5 and GABRP. These cells were then isolated and compared to amnion-like cells previously identified in Zheng. Differential expressed gene (DEG) analysis between PGCLCs from each twin were calculated using a Wilcox Rank Sum Statistical test using the parameters of a R 2-fold cut-off with genes expressed in >70% of PGCLCs. No statistically significant genes were identified.
# Smartseq analysis
Using the Carnegie Stage 7 (CS7) SMART-Seq raw data of annotated PGCs from Tyser, 53 the SMARTseq raw reads were trimmed with cutadapt 1.18 and reads with length over 30 bp were aligned to GRCh38 with STAR 2.7, and only uniquely mapped reads were kept for further analyses. Count matrices were generated by featureCounts v2.0.1 from the Subread R package.
## Principal component analysis (pca)
Each scRNAseq library (week 13 hPGCs, Carneigie 7 hPGCs and MZT hPGCLCs) were normalized with edgeR R package to acquire a CPM (Count Per Million) matrix. Top 2000 variable genes were extracted to perform PCA with the prcomp() function in R.
## Unsupervised hierarchical clustering (uhc)
The UHC was performed with scipy.cluster.hierarchy python function.
Whole genome sequencing Raw sequencing reads were aligned to the Human reference genome GRCh37 using BWA-MEM v0.7.17. Duplicated reads were marked and removed using MarkDuplicates (Picard), base quality score recalibration was performed with BaseRecalibrator followed by PrintReads. Variates of each individual sample were called through GATK haplotypecaller v4.2.2.0 and filtered with criteria: 80 R DP > 20, QUAL > 500, GQ > 50, MQ > 30 and MAF > 0.1, and the similarity coefficient (Jaccard index) were measured by pair-wise comparison to verify the genetic relationship of each biological sample. Two methods were applied and intersected to find high confidence variates in hiPSCs. Unique variates found by HaplotypeCaller were obtained from each hiPSC and fibroblast deduction. Secondly, each hiPSC and fibroblast comparison was verified though Strelka2 somatic mode, and the variates were filtered with criteria 80 R DP > 20, MQ > 30 and MAF >0.1. Only variates found in both GATK and Strelka2 were counted. Libraries were pooled and sequenced together in three separate sequencing runs.
## Quantification and statistical analysis
All statistical analysis was performed using R data analysis software. A p value of <0.05 was considered statistically significant. For all experiments, data are expressed as the mean ± SD. Unpaired two-tailed t tests are used in all cases unless otherwise stated. Power analyses were performed before experiments to determine n values. Experimenter was blinded to fertility phenotype of the participants, HDFs, and hiPSC lines for the duration of the study. For each experiment the number of technical replicates, and the number of biological replicates in each group is reported. Statistical details of experiments can be found in figure legends.
[fig] Figure 1: Derivation of HDFs and hiPSCs from twin pairs with discordant POI (A) Diagram depicting relationship of research participants and their HDFs and hiPSC sublines. Sublines are used as biological replicates for each participant. [/fig]
[fig] Figure 2: Percentage of hPGCLCs differentiated from hiPSCs The hiPSC sublines for each participant (n = 3 biological replicates) were each differentiated a minimum of three times. (A) Fluorescence-activated cell sorting (FACS) plots and quantification of hPGCLC percentages at day 4 of aggregate differentiation from twin pair MZT01 and MZT02 (t test, n = 10, p = 0.36). (B) Twin pair MZT04 and MZT06 (t test, n = 10, p = 0.73). (C) MZT05. The hPGCLC population is identified as double-positive for EPCAM and ITGA6 (circle). Data are represented as mean ± SEM with each circle on the graph, indicating the hPGCLC population. Statistical significance was calculated using a t test. [/fig]
[fig] Figure 3: Differentiation of each MZ twin's hiPSCs yields hPGCLCs with similar identity (A) The hiPSC sublines from each participant (n = 3 biological replicates) were analyzed by 10X Genomics to evaluate germline and somatic cell identity at day 4 of aggregate differentiation. In vivo hPGCs from CS7 human embryos 53 were used to stage hPGCLC development. (B-D) Germ cell gene expression was evaluated in the hPGCLCs from (B) twins MZT01 and MZT02, (C) twins MZT04 and MZT06, and (D) MZT05. The hPGCLC population analyzed on the right is indicated by the circled population of cells on the left. (E) Principle component analysis of the hPGCLC populations in this study with CS7 hPGCs and hPGCs from week 13 fetal gonads. The * indicates the affected twin. Cell Reports Medicine 3, like cell layer at the pole directly opposed to BMP4, columnar epiblast-like cells at the opposite pole, and hPGCLCs 55 ( [/fig]
[fig] Figure 4: Induction of hPGCLCs in an embryo model of the amniotic sac (A) Diagram of the embryo model. BMP4 was added 18 h after loading hiPSCs into the device. Thirty h after adding BMP4, amniotic sac-like embryo models develop, each containing an amniotic cavity, an amniotic ectoderm-like cell layer, pre-primitive streak epiblast (EPI)-like cells, and hPGCLCs (right). (B) The amniotic ectoderm-like cell layer is TFAP2A+ (n = 12 modeled embryos per participant were evaluated). Scale bar: 30 mm. (C) Representative images of hPGCLCs (triple positive for TFAP2C, NANOG, and SOX17) in the amniotic ectoderm-like cell layer (shown is MZT05-D). Arrows indicate hPGCLCs. (D) The number of specified hPGCLCs in the amniotic ectoderm-like cell layer was quantified from n = 8 embryo models from each participant's hiPSCs (MZT01 versus MZT02, p = 0.58; MZT04 versus MZT06, p = 0.30). Scale bar: 10 mm. Data are represented as mean ± SEM. Figure created with BioRender.com. Statistical significance was calculated using t test.6 Cell Reports Medicine 3, [/fig]
[fig] 27: Pandolfi, E.C., Rojas, E.J., Sosa, E., Gell, J.J., Hunt, T.J., Goldsmith, S., Fan, Y., Silber, S.J., and Clark, A.T. (2019). Generation of three human induced pluripotent stem cell sublines (MZT04D, MZT04J, MZT04C) for reproductive science research. Stem Cell Res. 40, 101576. https://doi. org/10.1016/j.scr.2019.101576. 28. Pandolfi, E.C., Sosa, E., Hunt, T.J., Goldsmith, S., Hurlbut, K., Silber, S.J., and Clark, A.T. (2021). Generation of six human induced pluripotent stem cell sublines (MZT01E, MZT01F, MZT01N and MZT02D, MZT02G and MZT02H) for reproductive science research. Stem Cell Res. 51, 102204. https://doi.org/10.1016/j.scr.2021.102204. 29. Besta, M., Kanakagiri, R., Mustafa, H., Karasikov, M., Ratsch, G., Hoefler, T., and andSolomonik, E. (2020). Communication-efficient Jaccard similarity for high-performance distributed genome comparisons. In Proceedings -2020 IEEE 34th International Parallel and Distributed Processing Symposium, IPDPS 2020. 30. Sherry, S.T., Ward, M., and Sirotkin, K. (1999). Database for single nucleotide polymorphisms and other classes of minor genetic variation. Genome Res. 9, 677-679. 31. Peterson, S.E., and Loring, J.F. (2014). Genomic instability in pluripotent stem cells: implications for clinical applications. J. Biol. Chem. 289, 4578-4584. https://doi.org/10.1074/jbc.R113.516419. 32. Bhutani, K., Nazor, K.L., Williams, R., Tran, H., Dai, H., D zakula, Z., Cho, E.H., Pang, A.W.C., Rao, M., Cao, H., et al. (2016). Whole-genome mutational burden analysis of three pluripotency induction methods. Nat. Commun. 7, 10536. https://doi.org/10.1038/ncomms10536. 33. Laurent, L.C., Ulitsky, I., Slavin, I., Tran, H., Schork, A., Morey, R., Lynch, C., Harness, J.V., Lee, S., Barrero, M.J., et al. (2011). Dynamic changes in the copy number of pluripotency and cell proliferation genes in human ESCs and iPSCs during reprogramming and time in culture. Cell Stem Cell 8, 106-118. https://doi.org/10.1016/j.stem.2010.12.003. 34. Gore, A., Li, Z., Fung, H.L., Young, J.E., Agarwal, S., Antosiewicz-Bourget, J., Canto, I., Giorgetti, A., Israel, M.A., Kiskinis, E., et al. (2011). Somatic coding mutations in human induced pluripotent stem cells. Nature 471, 63-67. https://doi.org/10.1038/nature09805. 35. D'Antonio, M., Benaglio, P., Jakubosky, D., Greenwald, W.W., Matsui, H., Donovan, M.K.R., Li, H., Smith, E.N., D'Antonio-Chronowska, A., and Frazer, K.A. (2018). Insights into the mutational burden of human induced pluripotent stem cells from an integrative multi-omics approach. Cell Rep. 24, 883-894. https://doi.org/10.1016/j.celrep.2018.06.091. 36. Hussein, S.M., Batada, N.N., Vuoristo, S., Ching, R.W., Autio, R., Nä rvä , E., Ng, S., Sourour, M., Hä mä lä inen, R., Olsson, C., et al. (2011). Copy number variation and selection during reprogramming to pluripotency. Nature 471, 58-62. https://doi.org/10.1038/nature09871.STAR+METHODS KEY RESOURCES [/fig]
[fig] 001176: and hPSCRO# 20016-003). All hiPSC subline cells were cultured at 37 C, 5.0% CO 2 on a feeder layer of mitomycin C-treated murine embryonic fibroblasts (MEFs) in pluripotent stem cell media (DMEM/F-12) (Life Technologies), 20% KSR (Life Technologies), 10 ng/mL bFGF (R&D Systems), 1% nonessential amino acids (Life Technologies), 2 mM L-glutamine (Life Technologies), Primocin TM (Invivogen), and 0.1 mM b-mercaptoethanol (Sigma). Media was changed daily and colonies were passaged with collagenase type IV (ThermoFisher, 17104019) every 7 days. For some experiments, cells were also maintained in feeder-free conditions on Matrigel (Fisher Scientific, 08-774-552) in mTeSR (STEMCELL Technologies 85850) at 37 C, 5.0% CO 2 . For mTeSR conditions, media was changed daily and colonies were passaged with ReLeSr (STEMCELL Technologies 05873) every 5 days. Mycoplasma was regularly tested using MycoAlert kit from Lonza Catalog #LT07-318.METHOD DETAILSInduction of hPGCLCs in aggregatesTo induce hiPSCs into hPGCLCs, the hiPSCs cultured for 7-days on mitomycin C-treated MEFs were trypsinized for 5 minutes (0.05% trypsin, Life Technologies) before resuspending in trypsin inhibitor (Life Technologies) to quench the trypsin. The MEFs weree depleted from the cell suspension by plating the cells onto tissue culture dishes, two times for 5 min each. The MEF-depleted cell suspension was then collected from the second plate, pelleted at 1,000 rpm using a centrifuge and resuspended in iMeLC media composed of (GMEM) (Life Technologies) containing 15% KSR (Life Technologies), 0.1 mM nonessential amino acids (Life Technologies), penicillin/streptomycin/L-glutamine (Life Technologies), Primocin TM (Invivogen), 0.1 mM b-mercaptoethanol (Sigma), sodium pyruvate (Life Technologies), activin A (PeproTech), CHIR99021 (Stemgent) and Y-27632 (Stemgent). After re-suspending in iMELC media, the cell suspension was filtered through a 40 mm cell strainer (Falcon) and plated at a density of 2.0 3 10 5 cells per well of a human plasma fibronectin (Invitrogen)-coated 12-well plate. After 24 h of incubation at 37 C with 5.0% CO 2 , the cells are referred to as iMeLCs. At this point, the cells are harvested from the 12-well plate by incubating in 0.05% trypsin for 5 minutes (Life Technologies) before resuspending in aggregate media consisting of (GMEM) (Life Technologies), 15% KSR (Life Technologies), 0.1 mM nonessential amino acids (Life Technologies), penicillin/streptomycin/L-glutamine (Life Technologies), Primocin TM (Invivogen), 0.1 mM b-mercaptoethanol (Sigma), and sodium pyruvate (Life Technologies) containing 10 ng/mL human LIF (EMD Millipore), 200 ng/mL BMP4 (R&D Systems), 50 ng/mL EGF (Fisher Scientific) and 10 mM Y-27632 (Stemgent). The single cell suspension is then added to low adherence 96-well plates (Corning) at a density of 3.0 3 10 3 cells per well to make the aggregates. All experiments analyzing hPGCLC induction in the aggregates are performed on day 4 (D4) after generating aggregates in 96 well plates. This technique was first described bySasaki et al., [/fig]
|
Novel niobium-doped titanium oxide towards electrochemical destruction of forever chemicals
Electrochemical advanced oxidative processes (EAOP) are a promising route to destroy recalcitrant organic contaminants such as per-and polyfluoroalkyl substances (PFAS) in drinking water. Central to EAOP are catalysis-induced reactive free radicals for breaking the carbon fluorine bonds in PFAS. Generating these reactive species electrochemically at electrodes provides an advantage over other oxidation processes that rely on chemicals or other harsh conditions. Herein, we report on the performance of niobium (Nb) doped rutile titanium oxide (TiO 2 ) as a novel EAOP catalytic material, combining theoretical modeling with experimental synthesis and characterization. Calculations based on density functional theory are used to predict the overpotential for oxygen evolution at these candidate electrodes, which must be high in order to oxidize PFAS. The results indicate a nonmonotonic trend in which Nb doping below 6.25 at.% is expected to reduce performance relative to TiO 2 , while higher concentrations up to 12.5 at.% lead to increased performance, approaching that of state-of-the-art Magnéli Ti 4 O 7 . TiO 2 samples were synthesized with Nb doping concentration at 10 at.%, heat treated at temperatures from 800 to 1100 °C, and found to exhibit high oxidative stability and high generation of reactive oxygen free radical species. The capability of Nb-doped TiO 2 to destroy two common species of PFAS in challenge water was tested, and moderate reduction by ~ 30% was observed, comparable to that of Ti 4 O 7 using a simple three-electrode configuration. We conclude that Nb-doped TiO 2 is a promising alternative EAOP catalytic material with increased activity towards generating reactive oxygen species and warrants further development for electrochemically destroying PFAS contaminants.OPEN
Per-and poly-fluoroalkyl substances (PFAS) are man-made chemicals that have a high stability in the environment due to the strength of the carbon-fluorine (C-F) bond 1 . Long-term studies reveal that PFAS can bioaccumulate differently than other contaminants owing to their high water solubility, thus entering the body through potable water and causing diverse chronic health effects 2 . Meanwhile, PFAS have dispersed globally in groundwater for over sixty years, reaching far from pollution sources 3 , and their thermal and chemical stability pose significant technical challenges for remediation. Separation-based technologies are the most common treatment method for PFAS-contaminated water, but this approach still requires destruction of the secondary PFAS waste stream, leading to other logistical, environmental, and health concerns [bib_ref] PFAS experts symposium: Statements on regulatory policy, chemistry and analtyics, toxicology, transport/fate,..., Simon [/bib_ref]. In addition to separation methods, there is an urgent need for more effective methods of PFAS destruction.
To accelerate the identification and experimental development of novel catalytic materials for PFAS destruction, computational tools based on first-principles density functional theory (DFT) are useful means 5 . DFT-based methods have frequently been leveraged for applications relating to hydrogen and oxygen evolution reactions (HER and OER, respectively) for water splitting [bib_ref] The genesis of molecular volcano plots, Wodrich [/bib_ref]. Wide ranges of both metals and oxides have been computationally screened and placed on "volcano plots" that exhibit a peak in catalytic activity at certain moderate binding energies of critical intermediate species [bib_ref] Universality in oxygen evolution electrocatalysis on oxide surfaces, Man [/bib_ref]. Accordingly, DFT catalyst screening has generally focused on optimizing materials toward the peaks of OER and HER volcano plots [bib_ref] Developments and perspectives of oxide-based catalysts for the oxygen evolution reaction, Fabbri [/bib_ref] [bib_ref] Design of efficient bifunctional oxygen reduction/evolution electrocatalyst: recent advances and perspectives, Huang [/bib_ref]. In this work, we adapt these methods for a different purpose, i.e., to screen materials for the breakdown of recalcitrant contaminants such as PFAS. In this case, we apply DFT calculations to seek materials far from the peak in the volcano plot and hence exhibiting suppressed OER, which is a prerequisite of electrode materials for electrochemical advanced oxidative processes (EAOP) [bib_ref] Critical review of electrochemical advanced oxidation processes for water treatment applications, Chaplin [/bib_ref]. www.nature.com/scientificreports/ EAOP have garnered intense interest due to the utilization of an electromotive force to induce the destruction of recalcitrant PFAS contaminants [bib_ref] Critical review of electrochemical advanced oxidation processes for water treatment applications, Chaplin [/bib_ref] [bib_ref] Electrochemical advanced oxidation processes (EAOP) to degrade per-and polyfluoroalkyl substances (PFASs), Fang [/bib_ref] [bib_ref] Energy-efficient electrochemical oxidation of perfluoroalkyl substances using a Ti4O7 reactive electrochemical membrane..., Le [/bib_ref] [bib_ref] A review of electrochemical reduction processes to treat oxidized contaminants in water, Mousset [/bib_ref] [bib_ref] Facing the challenge of poly-and perfluoroalkyl substances in water: is electrochemical oxidation..., Radjenovic [/bib_ref]. The key advantage of EAOP in destroying PFAS lies in its environmental compatibility, being based on electricity rather than additional chemicals that are themselves potentially toxic or require subsequent removal [bib_ref] Critical review of electrochemical advanced oxidation processes for water treatment applications, Chaplin [/bib_ref] [bib_ref] Electrochemical advanced oxidation processes (EAOP) to degrade per-and polyfluoroalkyl substances (PFASs), Fang [/bib_ref] [bib_ref] Opportunities for nanotechnology to enhance electrochemical treatment of pollutants in potable water..., Garcia-Segura [/bib_ref]. The underlying mechanism for contaminant destruction by EAOP has been the subject of debate, whether it is driven by direct oxidation at the electrode surface [bib_ref] Oxidative destruction of perfluorooctane sulfonate using boron-doped diamond film electrodes, Carter [/bib_ref] or by indirect oxidation via electrogenerated intermediates [bib_ref] Electrochemical oxidation of phenol at boron-doped diamond electrode, Iniesta [/bib_ref]. In the case of either hypothesized mechanism, an increased overpotential for OER is desired, hence suppressing it. In the first case, the rate-limiting step would be direct electron transfer between the electrode and the contaminant molecule. This mechanism has been supported in some work, for example, by comparison of experimentally measured kinetics with activation barriers calculated through DFT [bib_ref] Oxidative destruction of perfluorooctane sulfonate using boron-doped diamond film electrodes, Carter [/bib_ref]. In the second case, the rate-limiting step is presumed to be electrochemical production of hydroxyl free radicals ( - OH) as a byproduct of water oxidation at the anode surface, which then react unselectively with recalcitrant organic contaminants in water (Eq. 1): [bib_ref] Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms..., Buxton [/bib_ref] This EAOP can further cascade to generate other ROS free radicals such as superoxide radical (O 2- -), hydroperoxyl radical ( - OOH), etc. [fig_ref] Figure 1: Predicted trends in oxygen evolution overpotential relative to SHE based on DFT... [/fig_ref] [bib_ref] Critical review of electrochemical advanced oxidation processes for water treatment applications, Chaplin [/bib_ref] [bib_ref] Electrochemical advanced oxidation processes (EAOP) to degrade per-and polyfluoroalkyl substances (PFASs), Fang [/bib_ref] [bib_ref] Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms..., Buxton [/bib_ref]. These free radicals possess the capability to oxidize (e.g. break) strong C-F bonds (531 kJ mol -1 ) 11 . However, due to the short lifetime of these radicals (e.g. 2-4 μs) in aqueous solutions, novel materials that can generate copious amounts of radicals are needed [bib_ref] Electrochemical advanced oxidation processes (EAOP) to degrade per-and polyfluoroalkyl substances (PFASs), Fang [/bib_ref] [bib_ref] Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms..., Buxton [/bib_ref]. Of note, the fluorocarbon organics' stability demands high generation of - OH, and only a few EAOP electrode materials have demonstrated efficacy for this difficult operation [bib_ref] Electrochemical advanced oxidation processes (EAOP) to degrade per-and polyfluoroalkyl substances (PFASs), Fang [/bib_ref].
The current state-of-the-art EAOP electrode is based on a highly oxidatively stable, conductive material, boron-doped diamond (BDD) [bib_ref] Critical review of electrochemical advanced oxidation processes for water treatment applications, Chaplin [/bib_ref] [bib_ref] Surface science of diamond: Familiar and amazing, Ristein [/bib_ref] [bib_ref] Boron-doped diamond thin-film electrodes, Xu [/bib_ref]. Though diamond has a large bandgap (> 5 eV) that renders it as an electrical insulator, this material can be made semi-conductive (p-type semiconductor) by doping with boron atoms [bib_ref] Electrochemical oxidation of phenol at boron-doped diamond electrode, Iniesta [/bib_ref] [bib_ref] Electrochemical oxidation of water on synthetic boron-doped diamond thin film anodes, Michaud [/bib_ref] [bib_ref] Electrochemical incineration of organic pollutants on borondoped diamond electrode: Evidence for direct..., Zhi [/bib_ref] [bib_ref] Electrochemical oxidation of phenolic wastes with boron-doped diamond anodes, Cañizares [/bib_ref]. The stability of BDD results from carbon atoms being in sp 3 hybridization; however, these electrodes are still subject to failure due to delamination from the substrate [bib_ref] Doped diamond: a compact review on a new, versatile electrode material, Kraft [/bib_ref]. Moreover, the high cost associated with the preparation of this material has motivated the study of alternative candidates that also exhibit comparable activity and stability. The cost of diamond is the limiting factor in the usage of this EAOP electrode [bib_ref] Doped diamond: a compact review on a new, versatile electrode material, Kraft [/bib_ref]. Therefore, alternatives such as doped-tin oxides, lead oxides, and titanium oxides have been the subject of interest for identifying costeffective, stable, novel EAOP catalytic materials [bib_ref] Critical review of electrochemical advanced oxidation processes for water treatment applications, Chaplin [/bib_ref].
Among those alternatives to BDD, titanium oxides exist in a number of polymorphs (e.g. rutile, anatase, and brookite), and are promising owing to their high oxidative stability, low cost, and high electrical conductivity when expressed as a defective material (e.g., typically with oxygen vacancies) 10,25-27 . Stoichiometric TiO 2 , an insulator, is an inactive electrode and is not able to participate in the generation of ROS. Conductive Magnéli titanium oxide (Ti 4 O 7 ; n-type) has recently been implemented for PFAS destruction and has demonstrated high stability, coupled with high efficacy for destroying organic contaminants (< 10% PFAS concentration in permeate) [bib_ref] Energy-efficient electrochemical oxidation of perfluoroalkyl substances using a Ti4O7 reactive electrochemical membrane..., Le [/bib_ref] [bib_ref] Opportunities for nanotechnology to enhance electrochemical treatment of pollutants in potable water..., Garcia-Segura [/bib_ref]. Several sub-oxides of Ti 4 O 7 phases exist, based on the generic chemical formula, Ti n O 2n-1 , 4 ≤ n ≤ 10; the most conductive phases are Ti 4 O 7 and Ti 5 O 9 . Magnéli Ti 4 O 7 is also commercially available, which ensures a seamless transition to industrial-scale applications. The key disadvantage to this material is that molecular oxygen can likely be incorporated into the lattice structure upon anodic polarization due to the oxygen deficient nature of Ti 4 O 7 , thus forming an insulating TiO 2 layer at the surface [bib_ref] Critical review of electrochemical advanced oxidation processes for water treatment applications, Chaplin [/bib_ref] [bib_ref] Electrochemical oxidation of sulfide ion in synthetic sour brines using periodic polarity..., El-Sherif [/bib_ref]. Given the breadth of studies focused on understanding Ti 4 O 7 , identifying other TiO 2 -based electrodes with higher stability during anodic polarization would greatly benefit EAOP research toward applications such as PFAS destruction.
Another strategy for producing conductive TiO 2 involves doping with group V elements such as V, Nb, and Ta, which partially converts Ti(IV) to Ti(III) and produces n-type semi-conductivity [bib_ref] Electrochemical production of hydroxyl radical at polycrystalline Nbdoped TiO2 electrodes and estimation..., Kesselman [/bib_ref] [bib_ref] Photoemission and STM study of the electronic structure of Nbdoped TiO2, Morris [/bib_ref] [bib_ref] Development of supported bifunctional electrocatalysts for unitized regenerative fuel cells, Chen [/bib_ref]. Though the possibility of producing oxygen vacancies may still exist, low dopant concentrations (< 10 at%) should yield much less vacancies as compared with Ti 4 O 7 . A prior study in 1997 by Kesselman et al. demonstrated doping TiO 2 with niobium (Nb-TiO 2 ) generated stable and conductive electrodes that can produce - OH radicals with an anodic bias greater than 2 V vs. standard hydrogen electrode (SHE) [bib_ref] Electrochemical production of hydroxyl radical at polycrystalline Nbdoped TiO2 electrodes and estimation..., Kesselman [/bib_ref]. Nb-TiO 2 may therefore be an alternative EAOP catalytic material with stability, activity, and cost competitive with Magnéli Ti 4 O 7 and BDD. However, literature on Nb-TiO 2 for this application is scarce; thus, this material warrants further investigation on its potential use as a durable, EAOP catalytic material [bib_ref] Critical review of electrochemical advanced oxidation processes for water treatment applications, Chaplin [/bib_ref].
Herein, we hypothesize that Nb-doped TiO 2 electrodes may provide an effective means for PFAS destruction through the generation of ROS. We report the results of DFT calculations investigating the effect of Nb doping in TiO 2 on its predicted electrochemical properties. These results are compared with Magnéli phase Ti 4 O 7 , which has also been shown to effectively destroy perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS) [bib_ref] Energy-efficient electrochemical oxidation of perfluoroalkyl substances using a Ti4O7 reactive electrochemical membrane..., Le [/bib_ref]. To confirm the DFT results, we also synthesized Nb-doped TiO 2 and performed experimental measurements of its oxidative stability, generation of ROS, and destruction of PFOA and PFOS. The results confirm that Nb-doped TiO 2 may be considered as a potential EAOP material for PFAS destruction.
# Results
We first report the results of our DFT calculations with the goal of comparing TiO 2 -based materials for effective ROS production. As a proxy, we have predicted the activity of several materials based on established methods for the oxygen evolution reaction (OER) that must be suppressed in order to preferentially generate ROS, as shown in [fig_ref] Figure 1: Predicted trends in oxygen evolution overpotential relative to SHE based on DFT... [/fig_ref]. A high overpotential for OER is therefore an important criterion for catalytic materials for high ROS production [bib_ref] Critical review of electrochemical advanced oxidation processes for water treatment applications, Chaplin [/bib_ref]. This activity can be predicted in terms of a theoretical overpotential for OER based on a combined descriptor, G O * − G HO * , which is the difference between free energies of binding of the O* and HO* intermediates of the OER mechanism (asterisks denoting surface sites) [bib_ref] Universality in oxygen evolution electrocatalysis on oxide surfaces, Man [/bib_ref]. The results for rutile TiO 2 , Nb-doped TiO 2 (NTO), and Magnéli phase Ti 4 O 7 are shown in [fig_ref] Figure 1: Predicted trends in oxygen evolution overpotential relative to SHE based on DFT... [/fig_ref] , and the DFT results used to calculate the quantity www.nature.com/scientificreports/ G O * − G HO * for each material are reported in [fig_ref] Table 1: Total energies in eV of TiO 2 , Ti 4 O 7... [/fig_ref]. Calculated results from prior work by Man et al. [bib_ref] Universality in oxygen evolution electrocatalysis on oxide surfaces, Man [/bib_ref] are shown (gray circles) for a range of oxides, including TiO 2 which enables verification of our calculation procedure.
[formula] (1) H 2 O → · OH + H + + e − Scientific [/formula]
In the cases of 12.5 at.% and 6.25 at.% NTO, we generated multiple model structures with the same dopant concentration in different spatial configurations relative to the surface. Recent literature has shown that modeling binding energies on surfaces of materials with random defects requires particular care in this regard [bib_ref] An essential descriptor for the oxygen evolution reaction on reducible metal oxide..., Huang [/bib_ref]. The structures for all four material systems are shown in [fig_ref] Figure 2: Structural models of the [/fig_ref] , including three structures shown for NTO at each of the two dopant levels. In both cases, we found that structures with Nb substituted at surface sites (A in [fig_ref] Table 1: Total energies in eV of TiO 2 , Ti 4 O 7... [/fig_ref] result in much smaller values of G O * − G HO * than those with Nb only at subsurface sites (B and then C). The magnitude depends significantly and monotonically on the depth of the impurity atoms within the top three trilayers, with the variation diminishing to within 0.06 eV between the third and fourth trilayers. Therefore, for NTO at each loading level, the overpotential cannot be estimated from any single model structure, which would fail to capture the effects of random occupation in the experimentally synthesized materials. We therefore report the calculated values in [fig_ref] Figure 1: Predicted trends in oxygen evolution overpotential relative to SHE based on DFT... [/fig_ref] and [fig_ref] Table 1: Total energies in eV of TiO 2 , Ti 4 O 7... [/fig_ref] in terms of the mean and standard error among point calculations at the same Nb concentration.
To validate the DFT calculations on understanding the fundamental properties of NTO, this material was first synthesized using a solution process comprising niobium(V) oxalate hydrate and titanium(IV) isopropoxide as the metal precursor sources in a stoichiometric ratio of 1:9, leading to a theoretical 10 at.% Nb dopant stoichiometry (i.e. Nb 0.1 Ti 0.9 O x ). Carbon black was used as a sacrificial support to reduce agglomeration of the oxide powders. A series of post-heat treatment protocols was carried out on the as-derived powders to induce crystallization into the rutile phase. The calcination temperatures used were 800, 900, 1000, and 1100 °C, and the Nb-doped TiO 2 samples are denoted as NTO800, NTO900, NTO1000, and NTO1100, respectively. At 800 °C, a mixture of anatase and rutile phases are present, while at 900 °C, a predominately rutile phase is observed. At elevated heat-treatment temperatures, the diffraction peaks corresponding to the rutile phase increase, thus indicating the increase in particle size. The minor impurity peaks seen infor 900, 1000, and 1100 °C correspond to the titanium niobium oxide phase TiNb 2 O 7 . To quantify the phase fraction of TiNb 2 O 7 , Rietveld Higher loadings of Nb-doped TiO 2 (NTO) lead to reduction of the critical descriptor G O * − G HO * compared to pure rutile TiO 2 . This corresponds to an initial decrease in the magnitude of the OER overpotential prior to increasing again. The suppressed OER activity of 12.5 at.% NTO is predicted to approach that of Ti 4 O 7 , which was shown previously as a candidate for PFAS destruction. www.nature.com/scientificreports/ refinement was further performed to quantify the fractions between TiNb 2 O 7 and rutile TiO 2. Based on a goodness-of-fit factor (R wp ) of 9.146, we calculated that rutile TiO 2 comprises 98.8% of the total diffraction pattern, while TiNb 2 O 7 amounts only to 1.2%. Therefore, for the purposes of electrochemical characterization, we turned our attention to NTO900. Because TiNb 2 O 7 is present as an 'impurity' phase in the X-ray diffraction pattern, we further evaluated the distribution of Nb, while also imaging the particles using scanning electron microscopy (SEM). The SEM micrographof NTO900 shows a distribution of smaller (hundreds of nanometers) sized particles agglomerated that form larger primary particles on the order of 3 μm. Based on the collected scanning electron micrograph, energy-dispersive spectroscopy (EDS) was performed simultaneously to yield elemental mapping for Nb, Ti, and O shown in Figs. 4b-d. The elemental mapping confirms the primary elemental constituents are Ti and O, respectively, while also showing Nb distributed throughout the TiO 2 -rich particles. This suggests that the 900 °C heat-treatment generates a uniformly-distributed NTO material, and that the rutile pattern observed from X-ray diffraction ofcontains Nb within the crystal structure lattice. Furthermore, we performed X-ray absorption spectroscopy (XAS), and the corresponding XANES spectra for this series of NTO materials (NTO800, NTO900, NTO1000, and NTO1100) all yield the same spectra at the Nb K-edge [fig_ref] Figure 2: Structural models of the [/fig_ref] , thus validating the presence of the Nb dopants in the rutile TiO 2 structure.
To validate the generation of ROS, we first performed chronoamperometry at a potential of 3.07 V vs. SHE (pH ~ 7.0) for 10 min for Magnéli Ti 4 O 7 and this series of NTO-based EAOP electrode materials (800, 900, 1000, and 1100 °C). The supporting electrolyte consisted of 25 μL of 20 mM 2′-7′-dichlorofluorescin diacetate (DCFH-DA), a widely-used probe for detecting ROS, in 15 mL of de-ionized water [bib_ref] Evaluation of the probes 2' ,7'-dichlorofluorescin diacetate, luminol, and lucigenin as indicators..., Myhre [/bib_ref] [bib_ref] Measuring reactive oxygen and nitrogen species with fluorescent probes: challenges and limitations...., Kalyanaraman [/bib_ref]. Further details are provided in the Methods section. After electrochemical treatment at the respective potentials, this solution was analyzed for fluorescence, which showed that both Ti 4 O 7 and NTO900 exhibit appreciable amount of ROS. Of note, electrochemical treatment at 3.0 V for 30 min increases the fluorescence of the DCFH-DA probe dramatically for both Ti 4 O 7 and NTO900. Therefore, for the next series of experiments, a potential hold for 30 min was selected.
After validating the generation of ROS, we then performed practical electrochemical treatment of water solutions containing 20 ppb of PFOA and PFOS, the two important PFAS chemicals that are advised for monitoring by the United States Environmental Protection Agency. We performed oxidation of the solutions for both Ti 4 O 7 and NTO900 at 2.6, 2.8, 3.1, 3.3, and 3.6 V vs. SHE for 30 min. For all tests, we used a three-electrode configuration with a flooded beaker cell. www.nature.com/scientificreports/
# Discussion
The DFT results summarized in in [fig_ref] Figure 1: Predicted trends in oxygen evolution overpotential relative to SHE based on DFT... [/fig_ref] [bib_ref] An essential descriptor for the oxygen evolution reaction on reducible metal oxide..., Huang [/bib_ref]. These results suggest that NTO at sufficiently high loading levels can provide ROS generation characteristics superior to pure TiO 2 and competitive with Ti 4 O 7 . With respect to values of the descriptor G O * − G HO * , the sequence of TiO 2 , NTO of increasing concentration, and Ti 4 O 7 crosses over the "peak" of the volcano plot, resulting in a strongly non-monotonic trend in the OER overpotential, even accounting for the reported uncertainties in the calculation. As a result, based on a high OER overpotential as a proxy for activity for ROS generation, these calculations suggest in descending order of expected ROS generation: Ti 4 O 7 > 12.5 at.% NTO > TiO 2 > 6.25 at.% NTO. Notably, low Nb loading levels first exhibit reduced ROS generation due to enhanced OER relative to pure TiO [bib_ref] Toxicological effects of perfluoroalkyl and polyfluoroalkyl substances, Dewitt [/bib_ref] 10 , but further increased Nb loading is predicted to increase ROS generation. The results of the experimental characterization are qualitatively consistent with this picture. Initial assessment of the electrochemical stability of Ti 4 O 7 along with this series of NTO materials was first performed to identify oxidative stability [fig_ref] Figure 5: Electrochemical stability window of Ti 4 O 7 , and this series... [/fig_ref]. Based on a potential of 3.0 V vs. SHE, we set this as the oxidative stability limit for EAOP electrodes. Based on this threshold, all samples yield a current < 60 mA cm -2 . Based on prior oxygen [bib_ref] Benchmarking heterogeneous electrocatalysts for the oxygen evolution reaction, Mccrory [/bib_ref]. Based on this FOM, Ti 4 O 7 , NTO800, and NTO1100 all show that OER is suppressed well below 3.0 V vs. SHE. Though NTO900 and NTO1000 showed a slightly higher current density at the FOM, these materials may still be considered promising for EAOP applications. In addition, the specific surface area of this series of NTO materials all yielded values < 9 m 2 g -1 , which indicates that adsorption is highly unlikely to be responsible for significant reduction of PFAS contaminants. As a control, the specific surface area of Ti 4 O 7 was found to be 5.8 m 2 g -1 . Overall, the electrochemical stability window of all Ti-based materials, particularly when doped with Nb, show that OER extended into high potentials, thus indicating the likelihood for high and stable ROS generation. www.nature.com/scientificreports/ Qualitatively, we observe that Ti 4 O 7 yields comparably high fluorescence, indicative of high generation of ROS compared to high calcination-derived NTO1000 and NTO1100 . With lower calcination temperatures, NTO800 and NT900 both exhibit higher fluorescence, while NTO900 is shown to generate the highest amounts of ROS. By down-selecting NTO900 as the EAOP electrode-of-interest, we then performed chronoamperometry at 2.6, 2.8, and 3.07 V vs. SHE to understand the relationship between potential and ROS generation. NTO900 achieves the highest fluorescence at 3.1 V vs. SHE and much lower fluorescence at 2.6 and 2.8 V .
For PFOA, at 2.6, 2.8, and 3.1 V, NTO900 achieved a reduction in C f /C i to approximately 0.70, whereas Ti 4 O 7 achieved a reduction to approximately 0.60 . At higher potentials, the C f /C i was measured to be 0.80 for both NTO900 and Ti 4 O 7 , while NTO900 maintained that value at 3.6 V; for Ti 4 O 7 , a positive increase in C f was reflected. At higher potentials, the overall increase in the C f /C i value is attributed to the competing oxygen evolution reaction that may hinder the generation of ROS or the direct oxidation of the contaminant. For PFOS, we observe better performance for NTO900 compared to Ti 4 O 7 , where from 2.6 to 2.8 V, the C f /C i value was ~ 0.74, while Ti 4 O 7 gave ~ 0.8 . Similarly, at higher potentials, the reduction in PFOS concentration was not affected dramatically. Overall, the higher C f /C i values for PFOS vs. PFOA are consistent with that observed by Le et al. [bib_ref] Energy-efficient electrochemical oxidation of perfluoroalkyl substances using a Ti4O7 reactive electrochemical membrane..., Le [/bib_ref] Lastly, we note that a flow-through reactor is the ideal configuration for faster kinetics of PFAS electrochemical destruction, as evidenced by Carter et al. [bib_ref] Oxidative destruction of perfluorooctane sulfonate using boron-doped diamond film electrodes, Carter [/bib_ref] when they showed an order-of-magnitude faster decrease . Fluorescence data using 2'-7'-dichlorofluorescin diacetate (DCFH-DA) by performing chronoamperometry at (a) 3.07 V vs. SHE for 10 min for NTO800, NTO900, NTO1000, NTO1100, and Ti 4 O 7 ; and (b) 2.6, 2.8, and 3.07 V vs. SHE for 10 min of NTO900 to verify generation of reactive oxygen species. Practical electrochemical oxidation treatment of 20 ppb for 30 min of (c) perfluorooctanoic acid (PFOA) and (d) perfluorooctanesulfonic acid (PFOS) in de-ionized water for Ti 4 O 7 and NTO900. C f /C i is the ratio between the final and initial concentrations of the organic contaminants. Standard error for these treatments were within 5% based on replicate testing. www.nature.com/scientificreports/ in organic contaminant destruction with BDD. Our present results suggest that follow-up studies with NTO in such a flow-through system are warranted.
# Conclusions
Here we report on the performance of Nb-doped TiO 2 as a novel EAOP catalytic material, combining theoretical modeling with experimental synthesis and characterization. Our DFT calculations show that Nb doping below 6.25 at.% is expected to reduce reactive oxygen generation relative to TiO 2 , but higher concentrations up to 12.5 at.% should reverse this trend and increase performance to the point of being competitive with sub-stoichiometric Ti 4 O 7 . Nb-doped TiO 2 samples were synthesized and found to generate ROS, therefore showing them to be a promising candidate material for EAOP. We conclude that Nb-doped TiO 2 is a promising alternative EAOP catalytic material with increased activity relative to bare TiO 2 , lower cost than boron-doped diamond, and supporting comparable PFOA and PFOS reduction relative to sub-stoichiometric Ti 4 O 7 . These results point towards a promising direction for alternative TiO 2 -based EAOP catalyst materials via computation-guided synthesis and engineering of NTO.
# Methods
Density functional theory (DFT) calculations. All calculations were performed using density functional theory with a plane wave basis as implemented in Quantum Espresso version 6.6 36 using the revised Perdew-Burke-Ernzerhof functional of Hammer et al. [bib_ref] Improved adsorption energetics within density-functional theory using revised Perdew-Burke-Ernzerhof functionals, Hammer [/bib_ref] (RPBE) because of its optimization for adsorption calculations. Core electrons were treated using the projector augmented wave method. The plane wave kinetic energy cutoff was 40 Ry and electron density cutoff was 400 Ry. Geometry optimizations were performed with energy minimized within 10 -4 atomic units and forces within 10 -3 atomic units. Initial surface models were prepared using the Atomic Simulation Environment 38 starting with the experimental bulk crystal structures and cleaving slab models with the desired surface exposed. We constructed models of the (110) surface of rutile and 120 surface of Ti 4 O 7 , selected based on prior experimental work to be most stable surfaces of each material and therefore correspond to the greatest exposed area. The slab supercell model for rutile was constructed based on (2 × 2) tetragonal (110) surface unit cells with a thickness of 4 TiO 2 trilayers, resulting in a total supercell of 48 atoms with dimensions of 5.918 Å × 6.497 Å × 32.266 Å, corresponding to 20 Å of vacuum above the (110) surface 39 . Monkhorst-Pack grid of 4 × 4 × 1 k-points was used to sample the Brillouin zone, and occupations were treated with Methfessel-Paxton smearing of 0.001 Ry to account for electron donors into conductive states in Nb-doped TiO 2 and sub-stoichiometric Ti 4 O 7 relative to bulk rutile. The bottom two trilayers were frozen during geometry optimization to model the relaxation of a semi-infinite surface rather than a two-sided slab. The same procedure was adapted to prepare the model of the 120 surface of Ti [bib_ref] PFAS experts symposium: Statements on regulatory policy, chemistry and analtyics, toxicology, transport/fate,..., Simon [/bib_ref] Overpotential modeling. DFT calculations were used to estimate the overpotential η for OER for four different materials: TiO 2 , 6.25 at.% Nb-doped TiO 2 (NTO), 12.5 at.% NTO, and Ti 4 O 7 . For each material, η in V relative to SHE was calculated as a function of the descriptor G O * − G HO * using the "activity volcano" relationship which has been previously validated for a wide range of oxide surfaces: [bib_ref] Universality in oxygen evolution electrocatalysis on oxide surfaces, Man [/bib_ref] The descriptor represents the difference in free energies of binding of the critical intermediate species O* and HO*. This is calculated based on total energies E DFT O * and E DFT HO * of systems consisting of the respective O* and HO* species bound to each material surface obtained using DFT:
The optimized energy of gas-phase H 2 was also calculated using the respective DFT parameters within each system for consistent comparison. The zero-point energy and entropy corrections ZPE and TS 0 were the same as those used in Ref. [bib_ref] Universality in oxygen evolution electrocatalysis on oxide surfaces, Man [/bib_ref] [bib_ref] A review of the applications, environmental release, and remediation technologies of per-and..., Meegoda [/bib_ref]. In a typical procedure for the synthesis of 100 g of Nb 0.1 Ti 0.9 Ox with a final calcination temperature of 800 °C, niobium(V) oxalate hydrate (57.5210 g, 0.1173 mol) was dissolved in hot distilled water (500 mL) to form Solution I. Solution II was prepared by adding 97% titanium isopropoxide (309.6987 g, 1.0965 mol) to water (1 L), then oxalic acid dehydrate (315 g, 2.4986 mol) was added to dissolve the precipitate. Solution III was made by mixing I and II and was added to carbon black (500 g) to form a paste. The paste was first dried at 120 °C, then heated in air to 800 °C at a ramp rate of 2 °C min -1 . The dried paste stayed at 800 °C for several days until all the carbon black was removed. Four www.nature.com/scientificreports/ aliquots of 20 g each of the white powder were heat-treated in air for 8 h at temperatures of 800, 900, 1000 and 1100 °C to obtain the final products. The ramping rate to each temperature was 2 °C min -1 .
[formula] η OER = {max [(�G O * − �G HO * ), 3.2eV − (�G O * − �G HO * )]/e} − 1.23V G O * − G HO * = E DFT O * − E DFT HO * + 1 2 E DFT H 2 + ZPE − TS 0 Scientific Reports |( [/formula]
Materials characterization. SEM and energy dispersive X-ray spectroscopy (EDS) images were collected on a Thermo Scientific Scios DualBeam scanning electron microscope. SEM images and EDS mapping were obtained using a 5 kV accelerating voltage with an extraction current of 0.2 nA and a working distance of 7 mm. X-ray diffraction was performed using a PANalytical Empyrean diffractometer equipped with a CuKα (λ = 1.5416 Å) source to verify the phase of the synthesized Nb 0.1 Ti 0.9 Ox powders and Magnéli Ti 4 O 7 . The X-ray diffraction patterns of Nb 0.1 Ti 0.9 Ox powders calcined at 800, 900 1000 and 1100 °C were subsequently Rietveldrefined using the GSAS-II software package [bib_ref] The genesis of modern open-source all purpose crystallography software package, Toby [/bib_ref]. The specific surface area was measured by N 2 -porosimetry using a Quantachrome NOVA porosimeter. All samples were degassed with N 2 under vacuum for 12 h at 150 °C before obtaining the isotherms. The Brunauer-Emmett-Teller (BET) model was used to determine the specific surface area of all samples. X-ray absorption (XAS) was utilized to verify the presence of the Nb dopant. X-ray absorption near-edge structures (XANES) analysis was conducted at the Nb K-edge on the Beamline for Materials Measurement (6-BM) at the National Synchrotron Light Source II at Brookhaven National Laboratory (Upton, NY). The powders were spread as a thin-film onto Kapton tape and covered with x-ray clean polyfilm. Nb XANES were collected for each sample from −100 to −20 eV below the Nb K-edge (5.0 eV step size), from −20 to 50 eV above the Nb K-edge (0.2 eV step size), and 50 eV to 15 Å -1 above the Nb K-edge (0.05 Å -1 step size) all with a 0.5 s point -1 acquisition time. Spectra were collected in both transmission and fluorescence modes, and a standard ionization chamber and a four element silicon drift detector were used for each of these measurement types, respectively. Nb XANES standard spectra of Nb foil and NbO 2 were collected and used for calibration. Data processing was performed using the Athena software package [bib_ref] Hephaestus: Data analysis for X-ray absorption spectroscopy using IFEFFIT, Ravel [/bib_ref]. Spectra were normalized by fitting a firstorder polynomial to the pre-edge region and by fitting a second order polynomial to normalize the post-edge region to 1.0.
Electrochemical characterization. Electrochemical stability tests of this series of materials was carried out using a Gamry Reference 3000 potentiostat using a three-electrode configuration with 0.1 M HClO 4 as the supporting electrolyte (pH ~ 1.0). Platinum wire was used as the counter electrode, while all potentials were referenced to a frit-isolated Ag/AgCl electrode (+ 199 mV vs. NHE; saturated KCl). Glassy carbon (GC) electrodes (5 mm dia., geometric area = 0.196 cm 2 , PEEK casing; Gamry Instruments) were coated with a thin, uniform film of this series of Nb 0.1 Ti 0.9 Ox materials (800, 900, 1000, and 1100 °C) and Ti 4 O 7 . The film was deposited from a solution comprising 10 mg of electrocatalyst powder dispersed in a 5 wt% styrene-butadiene rubber (SBR) solution in water. An SBR polymer binder was selected, owing to its stability with the catalyst materials, as well as it being a F-free binder, which would eliminate any concerns of F leaching into the PFAS-destructed solutions. Before coating, the GC electrodes were polished with 0.5 μm alumina slurry, and thoroughly sonicated sequentially in de-ionized H 2 O and methanol, then dried under ambient conditions. Current-potential measurements for the oxidative stability of the Ti-based electrocatalysts were obtained using linear sweep voltammetry (LSV) with an imposed scan rate of 5 mV s -1 from 0 to 2.8 V vs. Ag/AgCl and 0 to −1.5 V vs. Ag/AgCl for assessing oxidative stability and reductive stability, respectively. All electrochemical potentials were iR compensated to account for the uncompensated resistance, as determined by measuring impedance at the open circuit potential.
To verify the generation of ROS, such as hydroxyl radicals, electrodes were prepared by drop-casting the same solution used to prepare GC electrodes, but at a higher mass loading, and on a 1 cm × 2 cm titanium foil substrate, with 1 cm 2 being the electroactive geometric area. The optimized mass loading of the electrodes was obtained by drop-casting 60 μL aliquots of the solution comprising 10 mg of electrocatalyst dispersed in 5 wt% SBR, which led to a mass loading of ~ 0.2 mg cm -2 . Chronoamperometry was performed where potentials of 2.6, 2.8, and 3.07 V vs. SHE (1.977, 2.117, and 2.447 V vs. Ag/AgCl; pH ~ 7.0) were imposed for 10, 30, and 60 min. The supporting electrolyte consisted of 25 μL of 20 mM 2′-7′-dichlorofluorescin diacetate (DCFH-DA), a widelyused probe for detecting ROS 32,33 , in 15 mL of de-ionized water. After electrochemical treatment at the respective potentials, this solution was analyzed for fluorescence using a Horiba Scientific Fluoromax + spectrometer. Each sample was placed into a 10 mL crystal cuvette and analyzed at an excitation wavelength (λ ex ) of 495 nm (slit 5 nm) and emission wavelengths (λ em ) from 500-600 nm (slit 2 nm).
[fig] Figure 1: Predicted trends in oxygen evolution overpotential relative to SHE based on DFT calculations. [/fig]
[fig] Figure 2: Structural models of the (a) TiO 2 (110) surface, the same surface with (b) 6.25 at.% and (c) 12.5 at.% Nb doping, and the (d) Ti 4 O 7 120 surface. Multiple structures were generated for each doped material system to capture the strong effect of proximity of the dopants to the surface. The models are periodic in all three directions, with 20 Å of vacuum normal to the surface. For each system, the shown geometry resulted from optimization with bound HO*. Images generated using Jmol version 14.29 40 . Scientific Reports | (2021) 11:18020 | https://doi.org/10.1038/s41598-021-97596-7 [/fig]
[fig] Figure 3: (a) X-ray diffraction patterns of Nb-TiO 2 calcined at various temperatures (800, 900, 1000, and 1100 °C). A and R labels correspond go anatase and rutile phases, respectively. (b) Rietveld-refined X-ray diffraction pattern of NTO900 with a yielded goodness-of-fit factor of 9.146.35 , a current density generation of 10 mA cm -2 yields a figure-of-merit (FOM) for OER [/fig]
[fig] Figure 4: (a) SEM micrograph of NTO900 showing overall particle morphology coupled with EDS elemental chemical mapping taken at the Nb (b), Ti (c), and O (d) edges, corroborating Nb doping into TiO 2 . Images generated using AZtec (Oxford Instruments). [/fig]
[fig] Figure 5: Electrochemical stability window of Ti 4 O 7 , and this series of NTO (NTO800, NTO900, NTO1000, and NTO1100) cycled with an imposed scan rate of 5 mV s -1 , with a rotation speed of 1600 rpm, in a 0.1 M HClO 4 supporting electrolyte. 3.0 V vs. SHE corresponds to the upper oxidative stability limit for EAOP electrodes. All electrochemical potentials were iR compensated for the uncompensated resistance, as determined by measuring impedance at the open circuit potential.Scientific Reports | (2021) 11:18020 | https://doi.org/10.1038/s41598-021-97596-7 [/fig]
[table] Table 1: Total energies in eV of TiO 2 , Ti 4 O 7 , and Nb-doped TiO 2 surfaces with O* and HO* adsorbates calculated using DFT and the resulting free energy differences used as predictor for the OER potential, η. [/table]
|
Tai Chi for Chronic Obstructive Pulmonary Disease (COPD): An Overview of Systematic Reviews
Objective: Since current systematic reviews (SRs) show that results of effectiveness on Tai Chi for chronic obstructive pulmonary disease (COPD) are inconsistent, the purpose of this study is to find the reasons of the disparity by comprehensively appraising the related SRs. Methods: Six databases were systematically searched from the inception date to April 17, 2021. The methodological quality, the risk of bias, the reporting quality, and the quality of evidence were independently assessed by two reviewers with the AMSTAR 2, ROBIS, PRISMA, and GRADE. Results: A total of 12 studies met the inclusion criteria: 10 SRs were rated critically low quality and two SRs were low quality by AMSTAR 2. By the ROBIS, four out of 12 SRs were rated as "low risk". According to PRISMA, nine out of 12 SRs were adequately reported over 80%. With the GRADE tool, three out of 12 SRs rated the FEV1 as "Moderate", one out of 12 SRs (1/12, 9%) rated the FEV1/FVC (%) as "Moderate", three out of 12 SRs assessed the 6MWD as "Moderate", three out of 12 SRs evaluated the SGRQ as "Moderate", and the remaining evidence was fully rated as "Low" or "Very Low". Conclusion: We found that the methodological quality, risk of bias, reporting quality, and quality of evidence of the included SRs on Tai Chi for COPD were suboptimal. These limitations may have a negative impact on SRs, consequently leading to inconsistent results. Further wellconducted SRs with less risk of bias, more rigorous methodology, normative reporting and highquality of evidence are needed to provide robust evidence on Tai Chi for COPD. Registration Number: This study has been registered in the PROSPERO International Prospective Register of Systematic Reviews (registration number: CRD42019126600).
# Background
Chronic obstructive pulmonary disease (COPD) is a common, preventable and treatable disease that is characterized by persistent respiratory symptoms and airflow limitation due to airway and/or alveolar abnormalities.Globally, there are around 3 million deaths annually caused by COPD. [bib_ref] Chronic obstructive pulmonary disease: current burden and future projections, Lopez [/bib_ref] By 2030, 3 more than 4.5 million people may die from COPD and related diseases every year. Currently, pharmacological therapies are considered as the first-line treatment for COPD. However, medications mainly focus on relieving symptoms rather than improving pulmonary function and quality-of-life. [bib_ref] Management Guidelines for Chronic Obstructive Pulmonary Disease, Celli [/bib_ref] Guidelines 5 for COPD indicate that exercise, as the core content of pulmonary rehabilitation, should be added to routine treatments for patients with moderate-to-severe COPD.
Tai Chi, as a traditional Chinese exercise with unique cultural characteristics, has been widely used for the rehabilitation in COPD. Clinical studies have manifested that Tai Chi was beneficial for COPD, due to its effectiveness of strengthening and stretching the body, improving the flow of blood and other fluids throughout the body, and improving balance.With the development of evidence-based medicine, systematic reviews (SRs) have become the standard approach for clinicians to make decisions. [bib_ref] Summing up evidence: one answer is not always enough, Lau [/bib_ref] [bib_ref] The importance and impact of evidence based medicine, Lewis [/bib_ref] [bib_ref] The role of systematic reviews and meta-analysis in dermatology, Katrina [/bib_ref] However, evidence of the effectiveness on Tai Chi for COPD is inconsistent. For example, Wu et al [bib_ref] Effectiveness of meditative movement on COPD: a systematic review and meta-analysis, Wu [/bib_ref] found that Tai Chi significantly increased the forced expired volume in one second per predicted (FEV1%), but Ngai et al [bib_ref] Tai Chi for chronic obstructive pulmonary disease (COPD), Ngai [/bib_ref] showed negative results. In such circumstances, the clinicians and policy-makers may be inundated with inconsistent evidence.
The disparity mainly comes from the methodology, risk of bias, reporting, and the quality of evidence of outcomes. Low methodologic quality may cause the results of SRs to be less reliable. [bib_ref] Methodological quality assessment of clinical trials in traditional Chinese medicine: the principles..., Liu [/bib_ref] Bias in the design or progress of SRs may lead to uncertain conclusions. Reporting quality is crucial for both authors and readers [bib_ref] Evaluations of the uptake and impact of the Preferred Reporting Items for..., Page [/bib_ref] to obtain the information, yet a large number of suboptimal reporting persists. [bib_ref] Epidemiology and reporting characteristics of systematic reviews of biomedical research: a Cross-Sectional..., Page [/bib_ref] Because of the inadequate reporting and low quality of evidence, readers cannot fully assess the strengths and weaknesses of SRs, which ultimately means that the findings of SRs cannot be implemented into clinical practice. [bib_ref] Reducing waste from incomplete or unusable reports of biomedical research, Glasziou [/bib_ref] [bib_ref] Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, David [/bib_ref] Therefore, a comprehensive overview of SRs, which can comply data from multiple SRs [bib_ref] A descriptive analysis of overviews of reviews published between, Hartling [/bib_ref] and reduce disparity between SRs, is more intuitive than numbers of SRs with low quality and inconsistent recommendations.
To date, no overview of related SRs on Tai Chi for COPD has been conducted. Our study aimed to conduct a comprehensive overview and to critically appraise the methodology, risk of bias, reporting quality, and quality of evidence of these SRs, by using the tools of A Measurement Tool to Assess Systematic Reviews 2 (AMSTAR 2), Risk of bias in systematic reviews (ROBIS), Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA), and the grading of recommendations assessment, development, and evaluation (GRADE), respectively. Besides, the effectiveness and safety of Tai Chi for COPD were narratively summarized.
# Method
## Protocol and registration
A priori protocol has been registered in the PROSPERO International Prospective Register of Systematic Reviews (registration number: CRD42019126600).
## Search strategy
We searched the following electronic bibliographic databases: PubMed, Embase, The Cochrane Library, China National Knowledge Infrastructure (CNKI), Wan Fang database and VIP database, using the keywords of Tai Chi, chronic obstructive pulmonary disease, and systematic review from inception to . Studies published in English and Chinese were included. The search strategies were listed in Appendix A. In addition, we manually searched the references and relevant domestic journals, including the Chinese Journal of Rehabilitation Medicine, Chinese Journal of Rehabilitation Theory and Practice, and the Chinese Journal of Physical Medicine and Rehabilitation.
## Inclusion and exclusion criteria
Inclusion criteria were established as follows: i) types of study: SRs containing more than one randomized controlled trial (RCT); ii) participants: Patients with COPD defined by spirometry; iii) interventions: Any type of Tai Chi with or without routine treatments; iv) comparison intervention: Routine treatments, such as drug therapy, routine activities, and respiratory training; v) outcome indicators: Pulmonary function including forced expiratory volume in 1 second (FEV1), FEV1% predicted normal values (FEV1%), the ratio of FEV1 to forced vital capacity [FEV1/FVC (%)], 6-minute walking distance (6MWD) and the St. George's Respiratory Questionnaire (SGRQ). Exclusion criteria were: i) Guidelines, review comments, overviews of SRs and editorials. ii) Full text unavailable after contacting the author. Literature retrieval was done independently by two researchers (LPY/LYX).
## Data management and data collection
Endnote X7 (BId 7072) was used to perform data management. Two researchers (YZ/YLZ) independently screened titles and abstracted for potentially relevant studies after eliminating duplications. Full texts were downloaded and read for eligible included studies based on inclusion and exclusion criteria. Then a cross-check after completion was performed to avoid mis-entry. Any discrepancies were discussed by a third reviewer (JRJ).
Data extraction was independently completed by two researchers (YZ/YLZ). Discrepancies were discussed with a third reviewer (JL). A data extraction form was developed in advance. The main contents of data extraction were as follows:
## Assessment of methodological quality
AMSTAR 2 is a common instrument to evaluate the methodology of SRs. [bib_ref] AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised..., Shea [/bib_ref] Two researchers (LPY/DLZ) independently assessed the methodology of SRs using AMSTAR2. AMSTAR2 has 16 items, including seven critical items (item 2/4/7/9/11/13/15) which can critically affect the validity of a SR and its conclusion. Each item was evaluated as "yes," "partial yes," and "no" according to the standard of AMSTAR 2 guideline. [bib_ref] AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised..., Shea [/bib_ref] The details of AMSTAR 2 items and the general rules for results of a SR are seen in Appendix B. Discrepancies were solved by team discussion.
## Assessment of risk of bias
ROBIS is a tool designed specifically to assess the risk of bias in SRs, [bib_ref] Clinical trial registration was not an indicator for low risk of bias, Farquhar [/bib_ref] which is comprised of three phases formed by signaling questions. Two researchers (WW/TYL) independently assessed the risk of bias of each SR using ROBIS. Phase 1 assesses the assessing relevance, which is optional. Phase 2, which is comprised of four domains formed by 21 signaling questions, aims to identify concerns with the review process. Phase 3, with three signaling questions, concentrates to judge risk of bias of the SR. All signaling questions were answered as "yes", "probably yes", "probably no", "no", and "no information". If all of signaling questions of phase 3 were answered as "yes," SR was judged as "low risk". Any of signaling question of phase 3 was answered as "probably no" or "no", SR was assessed as "high risk". If the information provided was insufficient to judge, SR was rated as "unclear risk". Discrepancies were solved by team discussion.
## Assessment of reporting quality
PRISMA is a reporting guideline designed to improve transparency of SRs, [bib_ref] Improving the quality of reports of meta-analyses of randomised controlled trials: the..., Moher [/bib_ref] [bib_ref] PRISMA harms checklist: improving harms reporting in systematic reviews, Zorzela [/bib_ref] [bib_ref] Preferred reporting items for systematic review and meta-analyses of individual participant data:..., Stewart [/bib_ref] which consisted of a 27-item checklist and a 4-phase flow diagram. Two researchers (YXL/LG) independently assessed reporting quality of included SRs using PRISMA. The highest score is 27. Each item was answered as "yes", "no", and "not applicable". We assessed the completion of the overall SRs' reporting by using a percentage, which was obtained through summing all items' score and dividing by the maximum score. The completion for each item was presented in ratio. Discrepancies were solved by team discussion.
## Assessment of quality of evidence
GRADE is widely used for rating the quality of evidence of each outcome in SRs.Two researchers (ZDL/LPY) independently utilized the GRADE tool to assess quality of each outcome in included SRs, including FEV1, FEV1/ FVC (%), FEV1%, 6MWD, and SGRQ. According to GRADE, evidence based on RCTs begin as high quality, but the quality of evidence may be downgraded dependent on five key factors (Risk of Bias, Inconsistency, Indirectness, Imprecision and Publication bias) of GRADE. Quality of evidence of each outcome was rated as "High", "Moderate", "Low", and "Very Low".
## Data synthesis and presentation
The results of AMSTAR 2, ROBIS, and PRISMA were summarized as a percentage of achievement per item. The results of AMSTAR 2, ROBIS, PRISMA, and the GRADE were summarized via tabulations. The characteristics of included SRs, the effectiveness and safety of Tai Chi for COPD were narratively summarized.
# Results
## Results on literature search and selection
We retrieved 101 citations and excluded 37 duplicates before screening. The 43 citations were excluded by title and abstract screening. After reading full-text, 12 eligible articles were included. A flow diagram of literature search was shown in [fig_ref] Figure 1: Flow diagram of literature search [/fig_ref].
## Characteristics of included reviews
Of these 12 SRs, six 24-29 studies were written in Chinese and six [bib_ref] Tai Chi for chronic obstructive pulmonary disease (COPD), Ngai [/bib_ref] [bib_ref] Chi for improving cardiopulmonary function and quality of life in patients with..., Guo [/bib_ref] [bib_ref] Effects of Tai Chi on exercise capacity and health-related quality of life..., Wu [/bib_ref] [bib_ref] Effects of Tai Chi in patients with chronic obstructive pulmonary disease: preliminary..., Jun-Hong [/bib_ref] [bib_ref] Effects of Tai Chi training on the physical and mental health status..., Guo [/bib_ref] [bib_ref] The effect of Tai Chi on four chronic conditions-cancer, osteoarthritis, heart failure..., Chen [/bib_ref] in English. They were published from 2013 to 2021. As for intervention, eight SRs took Tai Chi combined with other treatments (eg, drug treatment, routine rehabilitation therapy) as treatment, and four included Tai Chi alone. Several common treatments were used as comparison, including drug therapy, routine activities, respiratory training, aerobics, strength training, etc. The
## Dovepress
## 3019
addition, seven [bib_ref] Tai Chi for chronic obstructive pulmonary disease (COPD), Ngai [/bib_ref] [bib_ref] A meta-analysis of the effect of taijiquan exercise on pulmonary function rehabilitation..., Wei [/bib_ref] [bib_ref] Meta analysis of intervention effect of taijiquan on elderly patients with COPD, Wang [/bib_ref] [bib_ref] Meta-analysis of Tai Chi exercise on rehabilitation of middle-aged and elderly patients..., Dong [/bib_ref] [bib_ref] Chi for improving cardiopulmonary function and quality of life in patients with..., Guo [/bib_ref] [bib_ref] Effects of Tai Chi on exercise capacity and health-related quality of life..., Wu [/bib_ref] [bib_ref] Effects of Tai Chi training on the physical and mental health status..., Guo [/bib_ref] of these SRs used The Cochrane Collaboration's Risk of Bias to assess the risk of bias of the original RCTs. The remaining four [bib_ref] A meta-analysis of the effect of taijiquan on the rehabilitation of patients..., Tian [/bib_ref] [bib_ref] Effects of Tai Chi exercise on lung function and exercise endurance of..., Wang [/bib_ref] [bib_ref] Meta-analysis of the effects of taijiquan on lung function, exercise ability and..., Wang [/bib_ref] [bib_ref] Effects of Tai Chi in patients with chronic obstructive pulmonary disease: preliminary..., Jun-Hong [/bib_ref] SRs used Jadad scores and one SR 34 used the Physiotherapy Evidence Database (PEDro) scale. Details are shown in [fig_ref] Table 1: Characteristics of Systematic Reviews Tai Chi may represent an appropriate alternative or... [/fig_ref]. [fig_ref] Table 2: The Results of AMSTAR 2 [/fig_ref] shows the results of methodological quality of the included SRs using AMSTAR 2. Among the 12 included SRs, 10 SRs were considered critically low quality and two SRs 11,30 were low quality. AMSTAR 2 items(I) (ie, lower rates of "yes") with the lowest compliance rates were I2 ("a prior protocol provided", 8.33%), I3 (Did the review authors explain their selection of the study designs for inclusion in the review?, 0%), I5 (Did the review authors perform study selection in duplicate?, 42%), I7 (Did the review authors provide a list of excluded studies and justify the exclusions?, 17%), I12 ("the potential impact of RoB in individual studies on the results", 42%), I13 ("Did the review authors account for RoB in primary studies when interpreting/discussing the results of the review?", 42%), and I15 ("publication bias assessed", 42%).
## Results of methodological quality
## Results of risk of bias
According to [fig_ref] Table 3: Continued). [/fig_ref] , all SRs were in low risk in Phase 1. In phase 2, low risk of domain 1 was 100%, domain 2 was 33.33%, domain 3 was 66.67%, and domain 4 was 66.67%. Low risk of phase 3 was 33.33%. In addition, signaling questions (Q) of ROBIS with the highest concerns (ie, higher rates of "no") were Q6 (Did the search include an appropriate range of databases/electronic sources for published and unpublished reports?, no=33.33%), Q7 (Were methods additional to database searching used to identify relevant reports?, no=50%), Q20 (Were the findings robust, eg, as demonstrated through funnel plot or sensitivity analyses. no=41.67%), and Q22 (Did the interpretation of findings address all of the concerns identified in Domains 1 to 4?, no=50%). According to the final phase of ROBIS, four out 12 of SRs was rated as "low risk", 6SRs were "High risk", and two SRs were "Unclear risk".
[formula] I1 Y Y Y Y Y Y Y Y Y Y Y Y I2 Y N N Y Y N N N N N N N I3 N N N N N N N N N N N N I4 Y N Y Y Y Y Y Y N N Y N I5 PY PY PY PY PY PY Y Y Y Y Y N I6 Y Y PY Y Y Y Y Y Y Y N Y I7 N N N N Y Y N N N N N N I8 Y Y Y Y Y Y Y Y Y Y Y Y I9 Y Y Y Y Y Y Y Y Y Y Y Y I10 Y PY Y Y Y Y Y Y Y Y Y Y I11 Y Y Y Y Y Y Y Y Y Y Y Y I12 Y N N Y Y Y Y N N N N N I13 Y N N Y Y Y Y N N N N N I14 Y N N Y Y Y Y Y Y Y N Y I15 N Y N Y Y N N N Y Y N N I16 N PY Y Y Y Y Y Y Y N N N [/formula]
## Ranking of quality
[formula] Y Y Y Y Y Y Y Y Y Y Y Y 2.1.3(Q3) Were eligibility criteria unambiguous? Y Y Y Y Y Y Y Y Y Y YY)? Y Y Y Y Y Y Y Y Y Y YY [/formula]
## Results of reporting quality
We used PRISMA to assess the reporting quality of included SRs. We found that the section of title, abstract, introduction, study characteristics, limitation, conclusion, and funding were all well reported. We observed that nine out of 12 SRs were found adequately reported over 80%. However, several items (I) had a lower score (ie, lower rates of "yes"), that were I5 (Protocol and registration, yes=8.33%), I9 (Study selection, yes=41.67%), I11 (Data items, yes=25%), I16 (Additional analyses, yes=60%), and I22 (Risk of bias across studies, yes=25%). These items may account for the main reporting limitations and should be paid attention. The PRISMA checklists of each SR were presented in [fig_ref] Table 4: Reporting Quality Assessment of Systematic Reviews by PRISMA [/fig_ref].
## Results with grade tool
In our study, evidence of FEV1 in three SRs (3/12, 25%) 11,24,30 was rated as "Moderate", noevidence of FEV1% was rated as "Moderate", evidence of FEV1/ FVC (%) in one SR (1/12, 9%) 11 was rated as "Moderate", evidence of 6MWD in three SRs 11,30,31 was rated as "Moderate", evidence of SGRQ in two SRs 28,33 was rated as "Moderate", evidence of FEV1/ FVC (%) in one SR (1/12, 9%) 11 was rated as "Moderate", and the remaining evidence was rated as "Low" or "Very Low". We presentthem in . Deserved to be mentioned, Tai Chi is a physical exercise which consists of different forms and postures, it is not practical for participants to be blinded about group allocation. Lack of blinding may lead to bias. Thus, all studies were graded as "serious (−1)" in category of risk of bias. This may be the reason why there was no highquality evidence in our study. Besides, the quality of the same outcome among different the included SRs are inconsistent, a comprehensive and rigorous SR is needed to verify the effectiveness of Tai Chi for COPD. We are very uncertain about the estimate. ⊕⊕⊕⊝: The evidence is rated as "moderate" quality; ⊕⊕⊝⊝:
[formula] Y Y Y Y Y Y Y Y Y Y YYTitle I1 Title Y Y Y Y N Y N Y Y Y Y Y Abstract I2 Structured sumary Y Y Y Y Y Y Y Y Y Y Y Y Introduction I3 Rationale Y Y Y Y Y Y Y Y Y Y Y Y I4 Objectives Y Y Y Y Y Y Y Y Y Y Y Y Methods I5 Protocol and registration N N N Y N N N N N N N N I6 Eligibility criteria Y Y Y Y Y Y Y Y Y Y Y Y I7 Information sources Y Y Y Y Y Y Y Y NA NA Y Y I8 Search Y NA NA Y Y Y NA Y Y Y Y Y I9 Study selection Y NA N N Y N Y Y Y N N N I10 Data collection process Y Y Y Y Y Y Y Y Y Y Y Y I11 Data items NA NA NA NA Y NA NA NA N N Y Y I12 Risk of bias in individual studies Y Y Y Y Y Y Y Y Y Y Y Y I13 Summary measures Y Y N Y Y Y Y Y Y Y Y Y I14 Synthesis methods Y Y Y Y Y Y Y Y Y Y Y Y I15 Risk of bias across studies Y Y Y Y Y Y Y Y Y Y Y Y I16 Additional analyses Y N N NA Y Y Y N N Y NA Y Results I17 Study selection Y Y Y Y Y Y Y Y Y Y Y Y I18 Study characteristics Y Y Y Y Y Y Y Y Y Y Y Y I19 Risk of bias within studies Y Y Y Y Y Y Y Y Y Y Y Y 120 Results of individual studies Y Y Y Y Y Y Y Y Y Y Y Y I21 Synthesis of results Y Y Y Y Y Y Y Y Y Y Y Y I22 Risk of bias across studies N NA N Y Y N NA NA N N N Y I23 Additional analysis Y N NA Y Y Y Y Y Y Y Y Y I25 Limitations Y Y Y Y Y Y Y Y Y Y Y Y I26 Conclusions Y Y Y Y Y Y Y Y Y Y Y Y Funding I27 Funding Y N Y Y Y Y Y Y Y Y Y YTotal [/formula]
"low" quality; ⊕⊝⊝⊝:
"very low" quality. a No mention of blind and allocation concealment; b
The optimal sample size is not The Effectiveness and Safety of Tai Chi
We summarized the information from included SRs and found that their results were inconsistent. The details are shown in . FEV1, as we all know, has always been used as a primary metric of severity for obstructive lung disease. [bib_ref] Relationship between lung function impairment and health-related quality of life in COPD..., Berry [/bib_ref] [bib_ref] Minimal clinically important differences in COPD lung function, Donohue [/bib_ref] Among the SRs, seven out of nine SRs (78%) reported that FEV1 was significantly enhanced in the Tai Chi group, and one SR (11%) reported not. One SR (11%) reported that FEV1 was significantly improved after 6months of Tai Chi training, but with no significance less than 3 months or over 12 months. Both FEV1% and FEV1/FVC% were the main indicators for the diagnosis of COPD. [bib_ref] Relationship between lung function impairment and health-related quality of life in COPD..., Berry [/bib_ref] According to the results of FEV1%, three out of six SRs (50%) reported that FEV1% was significantly enhanced in the Tai Chi group, and three SRs (50%) reported not. As for FEV1/ FVC%, three out of eight SRs (38%) reported that FEV1/ FVC% was significantly enhanced in the Tai Chi group, five SRs (62%) reported that Tai Chi had no priority in FEV1/FVC%. 6MWD is a simple and reproducible test aiming to a global and integrated response of both physical (pulmonary and non-pulmonary factors) and psychological factors. [bib_ref] Determinants of poor 6-min walking distance in patients with COPD: the ECLIPSE..., Spruit [/bib_ref] [bib_ref] ATS statement: guidelines for the six-minute walk test, Crapo [/bib_ref] [bib_ref] Reproducibility of 6-minute walking test in patients with COPD, Hernandes [/bib_ref] It can be used to assess the functional exercise capacity. [bib_ref] Interpretation of treatment changes in 6-minute walk distance in patients with COPD, Puhan [/bib_ref] Of the 12 SRs, nine (75%) reported that 6MWD was significantly enhanced in the Tai Chi group. Only one SR (8%) reported that Tai Chi had no priority, and one SR (7%) reported that there was a significant increase at 3 months and 6 months, but no significance over 6 months.
SGRQ is a widely used instrument that is able to assess the health status and health-related quality-of-ife in COPD. [bib_ref] Quality of life measured by the St George's respiratory questionnaire and spirometry, Weatherall [/bib_ref] [bib_ref] Relationship between disease severity and quality of life in patients with chronic..., Sanchez [/bib_ref] Among the five included SRs, three SRs (60%) reported that SGRQ was significantly enhanced in the Tai Chi group. Two SRs (40%) reported that Tai Chi had no priority in SGRQ.
Only two SRs made subgroup analysis based on the duration and the remaining SRs had an ambiguous description of duration for Tai Chi. We found that the duration of Tai Chi affected the results of SR. For example, Guo et al 30 found that FEV1 was significant increased after 6-months of Tai Chi training, but had no significance less than 3 months or over 12 months. Guo et al 33 discovered that FEV1 could be improved after 3-months of Tai Chi training.
Among the 12 included studies, the intervention in the experimental group in nine SRs was Tai chi combined other rehabilitation therapy, only three SRs adopted Tai Chi alone. Wu et al [bib_ref] Effects of Tai Chi on exercise capacity and health-related quality of life..., Wu [/bib_ref] [bib_ref] Effects of Tai Chi on exercise capacity and health-related quality of life..., Wu [/bib_ref] Dong et al 29 also reported that FEV1 was not significantly enhanced in the Tai Chi group. In addition, among the 12 included SRs, no adverse effects of Tai chi on COPD were reported.
# Discussion
## The main findings
We narratively summarized the results of included SRs, and appraised the methodological quality, risk of bias, reporting quality, and quality of evidence of these SRs. Among the 12 SRs, 10 were considered as "critically low" in methodological quality and two [bib_ref] Tai Chi for chronic obstructive pulmonary disease (COPD), Ngai [/bib_ref] [bib_ref] Chi for improving cardiopulmonary function and quality of life in patients with..., Guo [/bib_ref] were low quality according to AMSTAR 2. By the ROBIS, six SRs were rated as "High" risk of bias, four were "Low" risk, and two were "Unclear risk". With the PRISMA checklist, we found most of the SRs (75%) were of relatively good reporting quality. However, the methodological quality, risk of bias, reporting quality, and quality of evidence of these SRs on Tai Chi for COPD were still suboptimal. Due to these limitations, the reliability of evidence for Tai Chi on COPD was weakened, and we should be cautious when recommending Tai Chi as an alternative treatment or alone treatment for improving pulmonary function and quality-of-life on COPD.
In addition, in this present study, we mainly focus on the methodology, risk of bias, and reporting quality of SRs about Tai Chi on COPD. We found that some items of AMSTAR 2, ROBIS, and PRISMA, may be repetitive. So, according to the results, we highlighted several common weaknesses of the included SRs as follows: 1) absence of a protocol registration or publication statement prior to the conduct of SR; 2) lack of a comprehensive literature search strategy and some additional methods for retrieving; 3) lack of appropriate instruments to assess the risk of bias (or methodological quality); 4) lack of prospectively describing the additional analyses in method; an 5) lack of appropriate methods to solve the heterogeneity and examine the robustness of the findings.
## Implications for clinicians and producers of further study for clinicians
Firstly, according to our results, Tai Chi was a safety treatment for COPD patients as no adverse events were reported in the included SRs. Secondly, we found that Tai Chi is prone to improve the functional capacity and quality-of-life in COPD patients, rather than pulmonary function. 75% of included SRs reported the significant increase for 6MWD in Tai Chi group, 60% for SGRQ, 78% for FEV1, 50% for FEV1%, and only 38% for FEV1/FVC%. Thirdly, due to the limitation of these SRs and the inconsistent results, we should be cautious when recommending Tai Chi as an alternative treatment or alone treatment on COPD.
## For producers of further sr
Most of these shortcomings can be avoided if the producers pay more attention to methodology, risk of bias, and standard reporting quality of SRs. Therefore, we have several suggestions for SR producers.
We suggest that the producers should register SR on registration platform like the international databases PROSPERO (https://www.crd.york.ac.uk/prospero/). Numerous studies have indicated that prospective registration may benefit the design, conduct, and reporting of SRs. [bib_ref] Registration in the international prospective register of systematic reviews (PROSPERO) of systematic..., Sideri [/bib_ref] [bib_ref] Protocol registration or development may benefit the design, conduct and reporting of..., Xu [/bib_ref] [bib_ref] Association between prospective registration and overall reporting and methodological quality of systematic..., Ge [/bib_ref] For producers of SRs, it is important to perform additional methods to ensure that more eligible studies can be retrieved, such as citation searches, contacting experts, reference checking, and handsearching. A list of excluded studies is necessary, otherwise there may be a risk that they remain invisible and the impact of their exclusion from the review is unknown. Risk of bias (or methodological quality) should be formally assessed using appropriate criteria. Of the 11 SRs, four, using the Jadad Scale, were all rated as "probably no" for absence of allocation concealment by AMSTAR2. The Jadad score does not include allocation concealment which has been shown to be a major potential risk of bias in RCTs. [bib_ref] Assessing the quality of reports of randomized clinical trials: is blinding necessary?, Jadad [/bib_ref] In addition, we suggest authors prospectively describe the additional analyses in statistical method, such as sensitivity analyses, subgroup analyses, and meta-regression.
For results of SR, the authors should focus on heterogeneity and robustness of findings. Heterogeneity may arise as a result of clinical or methodological differences between studies. [bib_ref] Heterogeneity in cochrane and non-cochrane meta-analyses in orthodontics, Koletsi [/bib_ref] Between-study variation might be inspected either visually or through statistical tests. Visual inspection mainly refers to the examination of the forest plots and the degree of overlap of the associated confidence intervals. [bib_ref] Systematic reviews: CRD's guidance for undertaking reviews in health care, Tacconelli [/bib_ref] Statistical heterogeneity may be confirmed through statistical tests, such as the I 2 statistic. [bib_ref] Quantifying heterogeneity in a metaanalysis, Higgins [/bib_ref] The Cochrane Collaboration has proposed the use of four categories when interpreting heterogeneity based on I 2 as follows: 0-40% unimportant heterogeneity, 30-60% moderate, 50-90% substantial, and 75-100% considerable heterogeneity.As for inevitable heterogeneity, we suggest the authors of SRs to perform meta-regression or subgroup analyses. In addition, robustness of the findings, which is often examined by funnel plots. Heterogeneity, reporting bias, and chance may all lead to asymmetry funnel plots. [bib_ref] Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised..., Sterne [/bib_ref] Therefore, we suggest the authors of SRs to perform a sensitivity analysis or use the funnel plots to examine relationships between effect size and study size.
# Strengths and limitations
There are some strengths in our study. First, it is the first overview which comprehensively appraised the SRs of Tai Chi for COPD with AMSTAR 2, ROBIS, PRISMA, and GRADE. It can help the physicians and patients to have a better understanding of the effectiveness on Tai Chi for COPD in clinical practice. Second, in order to reduce the risk of bias in the review, we prospectively registered on the PROSPERO platform to ensure the transparency of the study. Despite the advantages, this study also had some limitations. First, there might be some missing information since we only searched studies in English and Chinese language. Second, since there are overlapping studies between the included SRs, we only made narratively described the effectiveness and safety of Tai Chi on COPD, but did not synthesize the data from original RCT included in eligible SRs.
# Conclusion
There are more and more SRs concerning the Taiji treatment of COPD. However, through assessing the current SRs on Tai Chi for COPD, we found that the methodological quality, risk of bias, reporting quality, and quality of evidence of these SRs were suboptimal. Further well-conducted SRs with less risk of bias, more rigorous methodology, normative reporting, and high-quality of evidence are needed to provide robust evidence on Tai Chi for COPD.
## Abbreviations
## Dovepress
## 3031
Reviews and Meta-Analyses; GRADE, the grading of recommendations assessment, development and evaluation; PICO, participants, interventions, controls, and outcomes; PEDro, the physiotherapy evidence database; MCID, the minimum clinically important difference; EQ-5D, EuroQol-5 Dimension; SF-36, the 36-Item Short Form Health Survey.
## Ethics approval and consent to participate
Ethical assessment and informed consent were not required since our research did not involve individual data.
[fig] Figure 1: Flow diagram of literature search. https://doi.org/10.2147/IJGM.S308955 DovePress International Journal of General Medicine 2021:14 3020 [/fig]
[fig] 2.1. 5 Phase 3: Q5) Were any restrictions in eligibility criteria based on sources of information, appropriate (eg publication status or format, language, availability of data)2.1(Q6) Did the search include an appropriate range of databases/electronic sources for published and 2.4(Q9) Were restrictions based on date, publication format, or language appropriate? ://doi.org/10.2147/IJGM.S308955 DovePress International Journal of General Medicine 2021:14 2.3.1(Q11) Were efforts made to minimise error in 3.2(Q12) Were sufficient study characteristics available for both review authors and readers to be able to interpret the 3.4(Q14) Was risk of bias (or methodological quality) formally assessed using appropriate criteria? 4.1(Q16) Did the synthesis include all studies that it should? 4.3(Q18) Was the synthesis appropriate given the nature and similarity in the research questions, study designs and outcomes across included studies? 4.4(Q19) Was between-study variation (heterogeneity) minimal or addressed in the synthesis? 4.5(Q20) Were the findings robust, eg, as demonstrated through funnel plot or sensitivity analyses? Judging risk of bias 3.1(Q22) Did the interpretation of findings address all of the concerns identified in Domains 1 of General Medicine 2021:14 https://doi.org/10.2147/IJGM.S308955 DovePress 3025 [/fig]
[fig] 3. 3: Q24) Did the reviewers avoid emphasizing results on the basis of their statistical significance? : Y, yes; PY, probably yes; PN, probably no; N, no; NI, no information; Q, signaling questions. https://doi.org/10.2147/IJGM.S308955 DovePress International Journal of General Medicine 2021:14 [/fig]
[table] Table 1: Characteristics of Systematic Reviews Tai Chi may represent an appropriate alternative or complement to standard rehabilitation programs.However, whether Tai Chi is better than pulmonary rehabilitation exercise has not been determined.6MWTThe results demonstrated a favorable effect or tendency of Tai Chi to improve physical performance and showed that this type of exercise could be performed by individuals with different chronic conditions, including COPD, HF and OA.6MWD Tai Chi exercise can effectively improve the lung function of elderly patients, so that the patient's breathing is better and smoother, and the exercise capacity is significantly improved.6MWDTaijiquan exercise intervention in the stale period of the elderly patients with chronic obstructive pulmonary disease rehabilitation effect is significant.https://doi.org/10.2147/IJGM.S308955 International Journal of General Medicine 2021:14 [/table]
[table] Table 2: The Results of AMSTAR 2 [/table]
[table] 3023 Table 3: Abbreviations: Y, yes; PY, partial yes; N, no. International Journal of General Medicine 2021:14 https://doi.org/10.2147/IJGM.S308955 The Result of ROBIS [/table]
[table] Table 4: Reporting Quality Assessment of Systematic Reviews by PRISMA [/table]
|
Extracellular Superoxide Dismutase Is Associated With Left Ventricular Geometry and Heart Failure in Patients With Cardiovascular Disease
BACKGROUND: Extracellular superoxide dismutase (Ec-SOD) is a major scavenger of reactive oxygen species. However, its relationships with abnormal left ventricular (LV) geometry patterns and heart failure (HF) are still unknown in patients with cardiovascular disease.METHODS AND RESULTS:A cross-sectional study was carried out to evaluate the association of serum Ec-SOD activity with LV geometry, as well as HF in 1047 patients with cardiovascular disease. All participants underwent standard echocardiography examination and measurement of serum Ec-SOD activity. Overall, we found a significantly decreased trend of serum Ec-SOD activity from subjects with normal geometry (147.96±15.94 U/mL), subjects with abnormal LV geometry without HF (140.19±20.12 U/mL), and subjects with abnormal LV geometry and overt HF (129.32±17.92 U/mL) after adjustment for potential confounders (P for trend <0.001). The downward trends remained significant in the concentric hypertrophy and eccentric hypertrophy groups after stratification by different LV geometry patterns. Multinomial logistic regression analysis showed that each 10 U/mL increase in serum Ec-SOD activity was associated with a 16.5% decrease in the odds of concentric remodeling without HF (odds ratio [OR], 0.835; 95% CI, 0.736-0.948), a 40.4% decrease in the odds of concentric hypertrophy with HF (OR, 0.596; 95% CI, 0.486-0.730), a 16.1% decrease in the odds of eccentric hypertrophy without HF (OR, 0.839; 95% CI, 0.729-0.965) and a 34.0% decrease in the odds of eccentric hypertrophy with HF (OR, 0.660; 95% CI, 0.565-0.772).CONCLUSIONS: Serum Ec-SOD activity was independently associated with abnormal LV geometry patterns with and without overt HF. Our results indicate that Ec-SOD might be a potential link between LV structure remodeling and the development of subsequent HF in patients with cardiovascular disease.REGISTRATION: URL: https://www.clini caltr ials.gov; Unique identifier NCT03351907.
H eart failure (HF) is a complicated clinical syndrome caused by structural and/or functional cardiac abnormalities 1 and has become a rapidly growing public health issue throughout the world. In 2016, there were an estimated 37.7 million people living with HF globally, and the rising prevalence with advancing age still exists. [bib_ref] Epidemiology and risk profile of heart failure, Bui [/bib_ref] [bib_ref] Epidemiology and aetiology of heart failure, Ziaeian [/bib_ref] [bib_ref] Epidemiology of heart failure, Roger [/bib_ref] Although the mortality of HF has been reported to have decreased in recent decades, the 1-and 5-year mortality rates remain high, at near 20% and 50%, respectively, [bib_ref] Long-term trends in the incidence of and survival with heart failure, Levy [/bib_ref] partly attributable to the lack of effective therapy. Thus, early diagnosis, preventing the onset, and delaying the progression of HF, is of great importance.
Although the pathogenic mechanism of HF is complex and has not been fully elucidated, accumulating evidence suggests that enhanced oxidative stress contributes to cardiac ventricular and vascular remodeling and promotes the progression of HF. [bib_ref] Oxidative stress and heart failure, Tsutsui [/bib_ref] [bib_ref] Oxidative stress causes heart failure with impaired mitochondrial respiration, Nojiri [/bib_ref] [bib_ref] Oxidative stress and cardiac remodeling: an updated edge, Rababa'h Am [/bib_ref] Oxidative stress is an imbalanced state of reactive oxygen species production and the antioxidant defense system. [bib_ref] Is oxidative stress a therapeutic target in cardiovascular disease?, Münzel [/bib_ref] The superoxide dismutase (SOD) family are the first-line antioxidant enzymes in oxidative stress modulation. [bib_ref] Manganese superoxide dismutase and oxidative stress modulation, Bresciani [/bib_ref] Extracellular superoxide dismutase (Ec-SOD) is the predominant isoform, accounting for >70% of the total SOD activity in the human cardiovascular system. [bib_ref] Superoxide dismutase multigene family: a comparison of the CuZn-SOD (SOD1), Mn-SOD (SOD2),..., Zelko [/bib_ref] [bib_ref] Extracellular superoxide dismutase in vessels and airways of humans and baboons, Oury [/bib_ref] Previous studies have demonstrated that Ec-SOD activity is associated with endothelial function and long-term outcomes in patients with chronic HF with cardiomyopathy. [bib_ref] Serum uric acid correlates with extracellular superoxide dismutase activity in patients with..., Alcaino [/bib_ref] [bib_ref] Superoxide dismutase activity as a predictor of adverse outcomes in patients with..., Romuk [/bib_ref] Left ventricular (LV) remodeling is widely regarded as a crucial event in the progression of HF related to both cardiac geometry and function. [bib_ref] Cardiovascular remodelling in coronary artery disease and heart failure, Heusch [/bib_ref] [bib_ref] Heart failure, left ventricular remodeling, and circulating nitric oxide metabolites, Chirinos [/bib_ref] Compared with normal geometry (NG), concentric remodeling (CR), eccentric hypertrophy (EH), and concentric hypertrophy (CH) are LV structure remodeling phenotypes that have been well characterized via echocardiography examination. [bib_ref] Recommendations for cardiac chamber quantification by echocardiography in adults: an update from..., Lang [/bib_ref] [bib_ref] Prognosis of left ventricular geometric patterns in the Framingham Heart Study, Krumholz [/bib_ref] Although epidemiologic studies have reported that abnormal LV geometry phenotypes were associated with an increased risk of HF incidence, [bib_ref] The natural history of left ventricular geometry in the community: clinical correlates..., Lieb [/bib_ref] as well as worse outcomes in patients with HF independent of traditional measures of LV size and function, 20,21 the underlying mechanism from normal LV geometry to abnormal LV geometry and, finally, to HF remains unclear. In addition, substantial evidence indicates that oxidative stress contributes to cardiac remodeling via several mechanisms, [bib_ref] Oxidative stress and cardiac remodeling: an updated edge, Rababa'h Am [/bib_ref] [bib_ref] Activation of NADPH oxidase during progression of cardiac hypertrophy to failure, Li [/bib_ref] [bib_ref] Apoptosis signal-regulating kinase 1 plays a pivotal role in angiotensin II-induced cardiac..., Izumiya [/bib_ref] [bib_ref] Cardiomyocyte dimethylarginine dimethylaminohydrolase1 attenuates left-ventricular remodeling after acute myocardial infarction: involvement in..., Hou [/bib_ref] and Ec-SOD protects the heart against oxidative stress and ventricular remodeling in mice. [bib_ref] Extracellular superoxide dismutase protects the heart against oxidative stress and hypertrophy after..., Van Deel [/bib_ref] [bib_ref] Extracellular superoxide dismutase deficiency exacerbates pressure overload-induced left ventricular hypertrophy and dysfunction, Lu [/bib_ref] However, while Ec-SOD has been studied individually in cardiac remodeling animal models or patients with end-stage HF, few studies have directly explored the role of Ec-SOD in the progression from LV structure remodeling to HF status. Evidence from large-scale population studies of the association between Ec-SOD and the transition from LV structure remodeling to HF is scarce.
The present hospital-based observational study was conducted to evaluate the association between serum Ec-SOD activity and abnormal LV geometry patterns in patients with and without symptomatic HF and to determine whether Ec-SOD is a potential marker in the early stage of cardiac structure remodeling before symptomatic clinical HF is apparent.
# Methods
The data and study materials that support the findings of this study are available from the corresponding author upon reasonable request.
## Study population
The study participants were a subset from the Guangdong Cardiovascular Disease Cohort, a CLINICAL PERSPECTIVE What Is New?
- Serum extracellular superoxide dismutase activity was independently associated with abnormal left ventricular (LV) geometry patterns. - Serum extracellular superoxide dismutase activity presented a significant gradual downward trend from normal LV geometry to abnormal LV geometry without heart failure, and finally, to abnormal LV geometry with overt heart failure in patients with cardiovascular disease.
What Are the Clinical Implications?
- Our findings indicate that extracellular superoxide dismutase might be a potential link in the progression of normal LV structure to LV structure remodeling, and further to heart failure in patients with cardiovascular disease, and serum extracellular superoxide dismutase activity could be added to current biomarkers for risk assessment, as well as clinical management of heart failure.
## Nonstandard abbreviations and acronyms
## Data collection
All participants were interviewed face to face to collect demographic information, medical history, medication use, behavioral habits, and risk factor prevalence. Smoking habits were classified into 2 groups: never or past smoking and current smoking. Current smoking was defined as at least 1 cigarette per day regularly for more than 6 months before recruitment. Current alcohol drinking was defined as drinking any type of alcoholic beverage at least once a week for more than half a year before recruitment.
## Biomarker measurements
Overnight fasting venous blood specimens were sampled the next morning after hospital admission. Blood specimens were sent to the central clinical laboratory of Sun Yat-sen Memorial Hospital within 2 hours and measured by trained technicians. Serum Ec-SOD activity was tested using the autoxidation of the pyrogallol method (Superoxide Dismutase Assay Kit, Fuyuan Biotechnology Co. Ltd., Fujian, China), following the manufacturer's instructions. Creatinine, hsCRP (highsensitivity C-reactive protein), serum uric acid, and lactate dehydrogenase were determined using standard techniques by an automatic analyzer (Beckman Coulter chemistry analyzer AU5800, Beckman Coulter Co., Ltd, Tokyo, Japan). NT-proBNP and hsTNT (highsensitivity troponin T) were measured by a fully automated electrochemiluminescence immunoassay system (Roche Cobas e601, Hoffmann-La Roche Ltd, Basel, Switzerland). Glycated hemoglobin was measured by high-performance liquid chromatography (Variant II; Bio-Rad Laboratories, Hercules, CA). The estimated glomerular filtration rate was calculated using the Chronic Kidney Disease Epidemiology Collaboration Equation.
# Statistical analysis
Normally distributed data were expressed as the mean±SD, and variables with skewed distributions were reported as the median (interquartile range). Oneway ANOVA or the Kruskal-Wallis H test was used for overall comparisons for continuous variables, and the least significant difference t test was used for pairwise comparisons. Categorical variables were expressed by frequency and percentages, and intergroup comparisons were analyzed by the chi-square test. The bivariate correlations between serum Ec-SOD activity and echocardiographic parameters and laboratory biomarkers were determined by Spearman correlation analysis, and a partial correlation analysis on ranks (Spearman correlation) was further conducted to calculate the correlation coefficients after controlling for potential covariates. Afterward, a multivariable-adjusted general linear model (GLM) was used to compare the differences in Ec-SOD between NG and the other 3 types of LV geometry and determine the association of NG, abnormal LV geometry without HF, and abnormal LV geometry with HF, as well as to compare the differences of Ec-SOD in patients with different NYHA classes or ACCF/AHA stages of HF, only variables identified significant in univariate analysis would be incorporated in further multivariable-adjusted GLM. Multinomial logistic regression analysis was performed to estimate the odds ratios (ORs) per 10 U/mL increase in serum Ec-SOD activity for CR no HF, CR+HF, CH no HF, CH+HF, EH no HF, and EH+HF, with NG patients as the reference, as well as for NYHA class and ACCF/ AHA stages of HF, with the lowest class or stage as the references, using a forward stepwise procedure to select variables with the test level α=0.05. Binary logistic regression models were constructed to predict HF in patients with CR, CH, and EH, using a forward selection procedure to select variables with the test level α=0.05. After the patients were divided into the low and high Ec-SOD groups based on the median serum Ec-SOD activity in each type of LV geometry, 2 biomarkers reflecting myocardial stretch (NT-proBNP) and myocyte injury (hsTNT) were compared between the low and high Ec-SOD groups, and the Mann-Whitney U test and GLM were used to compare the differences of these 2 markers between the low and high Ec-SOD groups in different LV geometry patterns. Receiver operating characteristic (ROC) curve analysis was performed to test the potential ability of serum Ec-SOD activity to identify abnormal LV geometry with or without HF, using the bootstrap method to make comparisons of different areas under the curves (AUCs). All statistical analyses were performed using SPSS version 21.0 (SPSS Inc., Chicago, IL) and R (3.5.0). A 2-sided P<0.05 was considered statistically significant.
# Results
## Baseline characteristics of the study patients
The baseline demographic and clinical characteristics of the 1047 recruited patients with CVD in this study are summarized in [fig_ref] Figure 1: Serum extracellular superoxide dismutase [/fig_ref]. Compared with patients with abnormal LV geometry, patients in the NG group were younger, had lower systolic blood pressure and lower rates of obstructed coronary artery disease, hypertension, diabetes mellitus, dilated cardiomyopathy/hypertrophic cardiomyopathy, percutaneous coronary intervention history, atrial fibrillation, and valve disease. The serum Ec-SOD activities were normally distributed in the study patients with a mean level of 140.48±19.62 U/mL, which suggests that there were no carriers of R213G in our study population, since R213G carriers would have very high serum Ec-SOD activity and would appear as outliers . The serum Ec-SOD activities decreased in a stepwise manner from NG patients (147.96±15.74 U/mL) to CR patients (139.46±17.98 U/ mL), to EH patients (135.12±24.08 U/mL) and to CH patients (133.63±18.00 U/mL). Serum Ec-SOD activity was inversely associated with LVMI, left atrial diameter, end-diastolic interventricular septum, LVDd, LV mass, LV ejection fraction, NT-proBNP, hsTNT and hsCRP. NYHA class did not affect serum Ec-SOD activity after adjusting for covariates ; serum Ec-SOD activity was not an independent influencing factor for the ACCF/ AHA stages of HF [fig_ref] Figure 4: Receiver operating characteristic curves of serum Ec-SOD activity in the identification of... [/fig_ref].
## Serum ec-sod activity in different lv geometry patterns
Compared with the NG group, serum Ec-SOD activity in the CR, CH, and EH groups were markedly declined (P for trend <0.001) after adjusting for covariates. [fig_ref] Figure 1: Serum extracellular superoxide dismutase [/fig_ref]. Values are mean±SD, n (%), or median (interquartile range). Significance tests for comparisons by group based on one-way analysis of variance for normal distribution continuous variables, and Kruskal-Wallis H test for skewed distribution continuous variables; χ 2 test for categorical variables. ACEI indicates angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin-receptor blocker; BMI, body mass index; BSA, body surface area; CAD, coronary artery disease; DBP, diastolic blood pressure; DCM, dilated cardiomyopathy; Ec-SOD, extracellular superoxide dismutase; eGFR, estimated glomerular filtration rate; HbA 1c , glycosylated hemoglobin A 1 ; HCM, hypertrophic cardiomyopathy; HF, heart failure; hsCRP, high-sensitivity C-reactive protein; hsTNT, high sensitivity troponin T; IVSd, interventricular septum; LA, left atrial diameter; LDH, lactate dehydrogenase; LVDd, left ventricular end-diastolic diameter; LVEF, left ventricular ejection fraction; LVM, left ventricular mass; LVMI, left ventricular mass index; LVPW, left ventricular posterior wall thickness; NC, normal coronary artery; NOCAD, nonobstructive coronary artery disease; NT-proBNP, N-terminal pro-B type natriuretic peptide; OCAD, obstructive coronary artery disease; PCI, percutaneous coronary stent implantation; RTW, relative wall thickness; and SBP, systolic blood pressure.
## Comparison of hf markers in patients with different lv geometry with low and high ec-sod activity
In patients with NG, no difference was found in NT-proBNP or hsTNT between these 2 groups. In patients with CR without HF, NT-proBNP was elevated in the low Ec-SOD group (OR, 74.8; 95% CI, 27.8-166.0; P=0.030) compared with the high Ec-SOD group (OR, 46.6; 95% CI, 19.7-96.2) even after controlling for confounders by GLM. In patients with CH plus HF, NT-proBNP and hsTNT were significantly higher in the low Ec-SOD group (OR, 161.4; 95% CI, 62.7-951.4) before adjustment for covariates, but they were not significant after adjustment. In the EH group, we found that NT-proBNP was markedly elevated in the low Ec-SOD group, in subjects both with and without overt HF, before and after adjusting for covariates [fig_ref] Table 3: Comparisons of HF Markers in Subjects According to the Serum Ec-SOD Activity... [/fig_ref].
## Diagnostic ability of serum ec-sod activity for patients with abnormal lv geometry with and without overt hf
Serum Ec-SOD activity showed mild but significant diagnostic ability to distinguish patients in the CR no HF group (AUC, 0.626; 95% CI, 0.573-0.678; sensitivity, 49.4%; specificity, 71.6%), CH no HF group (AUC, 0.652; 95% CI, 0.595-0.712; sensitivity, 53.6%; specificity, 71.6%), and EH no HF group (AUC, 0.655; 95% CI, 0.600-0.711; sensitivity, 52.6%; specificity, 69.4%) from the patients in the NG group. When tested together with NT-proBNP and hsTNT, 2 conventional risk markers of cardiac function and damage, serum Ec-SOD activity added significant improvement in diagnosis performance beyond these 2 markers in distinguishing normal geometry from concentric remodeling without HF (AUC, 0.703; 95% CI, 0.566-0.879; P for bootstrap method=0.015) [fig_ref] Table 4: AUC of ROC Analysis in Identifying Abnormal LV Geometry Without HF [/fig_ref].In addition, compared with NG, patients with CR plus HF could be identified by Ec-SOD with a sensitivity of 64.7% and a specificity of 83.1% (AUC, 0.723; 95% CI, 0.566-0.879); patients with CH plus HF could be identified by Ec-SOD with a sensitivity of 68.9% and a specificity of 80.0% (AUC, 0.819; 95% CI, 0.769-0.868); and patients with EH plus HF could be identified by Ec-SOD with a sensitivity of 74.7% and a specificity of 71.6% (AUC, 0.789; 95% CI, 0.747-0.832). Furthermore, compared with CH no HF, patients with CH plus HF could be identified by Ec-SOD with a sensitivity of 58.1% and a specificity of 73.3% (AUC, 0.696; 95% CI, 0.618-0.775); compared with EH no HF, patients with EH plus HF could be identified by Ec-SOD with a sensitivity of 55.1% and a specificity of 78.1% (AUC, 0.692; 95% CI, 0.631-0.753) [fig_ref] Figure 4: Receiver operating characteristic curves of serum Ec-SOD activity in the identification of... [/fig_ref].
# Discussion
In this study, which included 1047 patients with CVD, we demonstrated that serum Ec-SOD activity presented a significant gradual downward trend from normal LV geometry to abnormal LV geometry without HF, and finally, to abnormal LV geometry with overt HF. After adjusting for demographic and clinical covariates, this association remained significant in patients with concentric hypertrophy and eccentric hypertrophy. To our knowledge, this is the first study to investigate the association of serum Ec-SOD The mean serum Ec-SOD activity in subject with normal geometry (NG, n=409), concentric remodeling (CR, n=171), concentric hypertrophy (CH, n=292) and eccentric hypertrophy (EH, n=175). **P<0.01 vs NG; ***P<0.001 vs NG. P values were calculated using the multivariable-adjusted general linear model, adjusted by age, sex, smoking, hypertension, diabetes mellitus, coronary artery disease, dilated cardiomyopathy/hypertrophic cardiomyopathy, PCI history, atrial fibrillation, valve disease, drug use (including angiotensin-converting enzyme inhibitor/ qngiotensin receptor blocker, beta-blocker, and diuretic agents), glycated hemoglobin, high-sensitivity C-reactive protein, estimated glomerular filtration rate, and uric acid, all covariates included in the multivariable-adjusted general linear model were significant in univariate analysis. activity with LV geometry and heart failure in a large population. Our study indicates that declining serum Ec-SOD activity might be an independent risk factor for the presence of abnormal LV geometry patterns and subsequent heart failure, especially in patients with LV hypertrophy.
Adverse LV structure remodeling is considered an intermediate phenotype of HF, given the high incidence of HF events observed among individuals with abnormal LV structure. [bib_ref] The natural history of left ventricular geometry in the community: clinical correlates..., Lieb [/bib_ref] [bib_ref] Prognostic implications of left ventricular mass and geometry following myocardial infarction: the..., Verma [/bib_ref] [bib_ref] Advances in the epidemiology of heart failure and left ventricular remodeling, Cheng [/bib_ref] de Simone et al 29 provided strong evidence that concentric LV hypertrophy is a risk factor for the development of HF independent of myocardial infarction and overload pressure, suggesting that mechanisms other than myocardial ischemia and hemodynamic load may play key roles in the development of HF in individuals with abnormal LV structure. However, the underlying pathophysiological mechanism in the progression of normal LV geometry to LV structure remodeling and eventually to HF remains to be elucidated.
Accumulating evidence derived from animal studies has demonstrated that Ec-SOD plays an important role in the development of HF. Ec-SOD gene-deficient mice developed more LV hypertrophy in response to . Serum Ec-SOD activity in patients with normal left ventricular (LV) geometry and abnormal LV geometry with and without HF. A, Differences of serum Ec-SOD activities among subjects with NG, subjects with all types of abnormal LV geometry without HF (abLVG, including CR, CH, and EH subject without HF), and subjects with abnormal LV geometry plus heart failure (abLVG+HF); B, differences of serum Ec-SOD activities among subjects with NG, subjects with CR but without HF (CR no HF), and subjects with CR and overt HF (CR+HF); C, differences of serum Ec-SOD activities among subjects with NG, subjects with CH but without HF (CH no HF), and subjects with CH and overt HF (CH+HF); D, differences of serum Ec-SOD activities among subjects with NG, subjects with EH but without HF (EH no HF) and subjects with EH and overt HF (EH+HF). abLVG indicates abnormal left ventricular geometry; CH, concentric hypertrophy; CR, concentric remodeling; Ec-SOD, extracellular superoxide dismutase; EH, eccentric hypertrophy; HF, heart failure; and NG, normal geometry. *P<0.05; **P<0.01; ***P<0.001; P values were calculated using the same multivariable-adjusted general linear model as [fig_ref] Figure 1: Serum extracellular superoxide dismutase [/fig_ref]. Least significant difference t test was used for pairwise comparison.
overload pressure and showed greater oxidative stress and myocardial fibrosis associated with activation of the mitogen-activated protein kinase signaling cascades. [bib_ref] Extracellular superoxide dismutase protects the heart against oxidative stress and hypertrophy after..., Van Deel [/bib_ref] [bib_ref] Extracellular superoxide dismutase deficiency exacerbates pressure overload-induced left ventricular hypertrophy and dysfunction, Lu [/bib_ref] Overexpression of Ec-SOD in the hearts of transgenic mice helps to protect cardiac function from ischemia-reperfusion injury. [bib_ref] Extracellular superoxide dismutase transgene overexpression preserves postischemic myocardial function in isolated murine..., Chen [/bib_ref] A genetic variant with a substitution in the heparin-binding domain of Ec-SOD (Ec-SOD-R213G) was associated with excessive oxidative stress, endothelial dysfunction, [bib_ref] Vascular effects of a common gene variant of extracellular superoxide dismutase in..., Iida [/bib_ref] increased risk of ischemic heart disease, and more severe HF. [bib_ref] Genetically reduced antioxidative protection and increased ischemic heart disease risk: the Copenhagen..., Juul [/bib_ref] The Ec-SOD R213G mutation is present in 4% to 6% of the Asian population, and the plasma EC-SOD levels are 10-fold or higher in mutation carriers than in noncarriers. [bib_ref] 10-fold increase in human plasma extracellular superoxide dismutase content caused by a..., Sandström [/bib_ref] [bib_ref] Two variants of extracellular-superoxide dismutase: relationship to cardiovascular risk factors in an..., Marklund [/bib_ref] It has been speculated that such an increase results from the accelerated release of EC-SOD from the interstitial matrix. [bib_ref] 10-fold increase in human plasma extracellular superoxide dismutase content caused by a..., Sandström [/bib_ref] [bib_ref] Substitution of glycine for arginine-213 in extracellular-superoxide dismutase impairs affinity for heparin..., Adachi [/bib_ref] In noncarriers, higher serum EC-SOD comes from higher tissue EC-SOD, which is the effective part to protect against oxidative stress in tissue. [bib_ref] Gene transfer of extracellular superoxide dismutase reduces arterial pressure in spontaneously hypertensive..., Chu [/bib_ref] [bib_ref] Vascular effects of the human extracellular superoxide dismutase R213G variant, Chu [/bib_ref] In the present study, the serum EC-SOD activities were normally distributed, and there were no patients with very high serum EC-SOD activity levels, suggesting that there was no carriers of the Ec-SOD R213G mutation in our study; thus, the negative correlation of serum EC-SOD activity and the severity of the heart phenotype may not be influenced by the Ec-SOD R213G mutation.
Previous population studies have reported that the serum activities of the SOD family, including manganese SOD, copper/zinc-containing SOD, and total SOD, were predictors of worse long-term clinical outcome in nonischemic dilated cardiomyopathy patients, which is a frequent cause of HF. 14 Reduced Ec-SOD activity was reported to be closely associated with increased vascular oxidative stress and endothelial dysfunction in patients with chronic HF. [bib_ref] Vascular oxidative stress and endothelial dysfunction in patients with chronic heart failure:..., Landmesser [/bib_ref] A later small sample case-control study including 38 patients with chronic HF and 12 controls validated this association and found that it might be related to serum uric acid. [bib_ref] Serum uric acid correlates with extracellular superoxide dismutase activity in patients with..., Alcaino [/bib_ref] However, only a few small-sample studies based on populations have paid attention to the association between . ORs of different LV geometry patterns with and without HF by 10 U/mL serum Ec-SOD activity increase. Multivariable-adjusted odds ratios (95% CI) for the presence of different LV geometry patterns with and without HF per 10 U/mL serum Ec-SOD activity increase were calculated through multinomial logistic regression model, using forward stepwise procedure to select variables. CH indicates concentric hypertrophy; CR, concentric remodeling; Ec-SOD, extracellular superoxide dismutase; EH, eccentric hypertrophy; HF, heart failure; NG, normal geometry; and ORs, odds ratios. circulating Ec-SOD activity and nonischemic dilated cardiomyopathy in patients with end-stage HF. There is still limited information about circulating Ec-SOD activity in patients with HF and its precursor condition, LV geometry remodeling. The present study extends this information in 4 aspects in a large Chinese population for the first time. First, serum Ec-SOD activity decreased in patients with CVD with abnormal LV geometry patterns without HF, including patients with concentric remodeling, concentric hypertrophy, and eccentric hypertrophy, and the declining levels presented more obviously in patients with LV hypertrophy than in patients with simple concentric remodeling. Second, serum Ec-SOD activity decreased more notably in patients with abnormal LV geometry plus overt HF than in those without HF. Third, we demonstrated that even after adjustment for demographic and clinical covariates, serum Ec-SOD activity gradually declined from normal LV geometry to abnormal LV geometry and finally to HF in patients with concentric hypertrophy and eccentric hypertrophy but not in patients with concentric remodeling. Fourth, serum Ec-SOD activity showed significant improvement [fig_ref] Figure 1: Serum extracellular superoxide dismutase [/fig_ref]. CH indicates concentric hypertrophy; CR, concentric remodeling; Ec-SOD, extracellular superoxide dismutase; EH, eccentric hypertrophy; HF, heart failure; hsTNT, high-sensitivity troponin T; NG, normal geometry; and NT-proBNP, N-terminal pro-B-type natriuretic peptide. *P<0.05. P value was calculated using bootstrap method to make comparisons of the AUC between NT-proBNP+hsTNT and Ec-SOD+NT-proBNP+hsTNT. AUC indicates area under the curve; CH, concentric hypertrophy; CR, concentric remodeling; Ec-SOD, extracellular superoxide dismutase; EH, eccentric hypertrophy; HF, heart failure; hsTNT, high-sensitivity troponin T; NG, normal geometry; NT-proBNP, N-terminal pro-B-type natriuretic peptide; and ROC, receiver operating characteristic curve. in the ability to distinguish patients with concentric remodeling without HF from those with normal LV geometry beyond NT-proBNP, which has been widely used for HF management, outcome prediction, and risk assessment. [bib_ref] Relation between serum N-terminal pro-brain natriuretic peptide and prognosis in patients with..., Coats [/bib_ref] Additionally, in patients with eccentric hypertrophy but without overt HF, subjects with lower serum Ec-SOD activity were in higher HF risk since they also had higher NT-proBNP. Taken together with previous studies, we speculate that serum Ec-SOD might be a link in the progression of normal LV geometry to LV structure remodeling and further HF, and decreased serum Ec-SOD activity might contribute to alterations in LV structure and the onset of HF in patients with CVD. Further studies are required to elucidate the mechanism behind these associations.
There were some limitations in this study. First, it is important to stress that the cross-sectional design could not determine causal relationships between declining serum Ec-SOD activity and LV structure alteration or HF. However, we analyzed the associations in a large population and the evidence was strengthened by considering a variety of established confounders. Additionally, since we analyzed Ec-SOD activity in venous blood instead of in heart tissue, the activity level might not be as specific it would have been in myocardial biopsies. It is impossible to obtain heart tissue in routine clinical practice; thus, testing Ec-SOD activity in circulation might be more receptive and helpful in evaluating these complex phenotypes.
In summary, we observed that serum Ec-SOD activity declined gradually and significantly from normal LV geometry to abnormal LV geometry without HF, and finally, to abnormal LV geometry with overt HF. Our results suggest that Ec-SOD may be an independent link between LV structure remodeling and the development of subsequent HF.
## Article information
Received April 3, 2020; accepted June 30, 2020. A, Rreceiver operating characteristic curve analysis of serum Ec-SOD activity for distinguishing subjects in CR+HF group from subjects in NG group; B, for distinguishing subjects in CH+HF group from subjects in NG group; C, for distinguishing subjects in EH+HF group from subjects in NG group; D, for distinguishing subjects in CR+HF group from subjects in CR no HF group; E, for distinguishing subjects in CH+HF group from subjects in CH no HF group; F, for distinguishing subjects in EH+HF group from subjects in EH no HF group. AUC indicates area under the curve; CH, concentric hypertrophy; CR, concentric remodeling; Ec-SOD, extracellular superoxide dismutase; EH, eccentric hypertrophy; HF, heart failure; and NG, normal geometry. . Distribution of serum Ec-SOD activity in total study population. . Association between serum Ec-SOD activity and NYHA class.
# Supplemental material
[fig] Figure 1: Serum extracellular superoxide dismutase (Ec-SOD) activity in subjects with different left ventricular geometry patterns. [/fig]
[fig] Figure 4: Receiver operating characteristic curves of serum Ec-SOD activity in the identification of abnormal LV geometry patients with HF. [/fig]
[fig] Figure S1: Flow chart of the study population in this study. [/fig]
[table] Table 1: Baseline Characteristics of Study Participants [/table]
[table] Table 2: Association Between Ec-SOD Activity and the Risk of HF Presence in CR, CH, and EH Patients procedure to select variables. CH indicates concentric hypertrophy; CR, concentric remodeling; Ec-SOD, extracellular superoxide dismutase; EH, eccentric hypertrophy; HF, heart failure; and OR, odds ratio. [/table]
[table] Table 3: Comparisons of HF Markers in Subjects According to the Serum Ec-SOD Activity Unadjusted P value was calculated by the Mann-Whitney U test. Adjusted P value was calculated by multivariable-adjusted general linear model adjusted by potential confounders (covariates were same as [/table]
[table] Table 4: AUC of ROC Analysis in Identifying Abnormal LV Geometry Without HF [/table]
[table] Table S1: Correlation coefficients between Ec-SOD and Echocardiographic parameters or laboratory biomarkers.Bivariate correlation coefficients were calculated using Spearman correlation analysis.Partial correlation coefficients were calculated using Partial Spearman correlation analysis controlling for age, sex, BMI, smoking, drinking, hypertension, diabetes, coronary artery disease, Dilated cardiomyopathy/Hypertrophic cardiomyopathy, PCI history, Atrial fibrillation, Valve disease, drug use (including ACEI/ARB, Beta-blocker, Anti-diabetes, Diuretic agents and statin). Abbreviations are consistent withTable 1. [/table]
|
WBC image classification and generative models based on convolutional neural network
Background: Computer-aided methods for analyzing white blood cells (WBC) are popular due to the complexity of the manual alternatives. Recent works have shown highly accurate segmentation and detection of white blood cells from microscopic blood images. However, the classification of the observed cells is still a challenge, in part due to the distribution of the five types that affect the condition of the immune system.Methods: (i) This work proposes W-Net, a CNN-based method for WBC classification. We evaluate W-Net on a realworld large-scale dataset that includes 6562 real images of the five WBC types. (ii) For further benefits, we generate synthetic WBC images using Generative Adversarial Network to be used for education and research purposes through sharing.Results: (i) W-Net achieves an average accuracy of 97%. In comparison to state-of-the-art methods in the field of WBC classification, we show that W-Net outperforms other CNN-and RNN-based model architectures. Moreover, we show the benefits of using pre-trained W-Net in a transfer learning context when fine-tuned to specific task or accommodating another dataset. (ii) The synthetic WBC images are confirmed by experiments and a domain expert to have a high degree of similarity to the original images. The pre-trained W-Net and the generated WBC dataset are available for the community to facilitate reproducibility and follow up research work.Conclusion:This work proposed W-Net, a CNN-based architecture with a small number of layers, to accurately classify the five WBC types. We evaluated W-Net on a real-world large-scale dataset and addressed several challenges such as the transfer learning property and the class imbalance. W-Net achieved an average classification accuracy of 97%. We synthesized a dataset of new WBC image samples using DCGAN, which we released to the public for education and research purposes.
organisms such as microorganisms and antigens, in order to remove them out of the body [bib_ref] Lymphocyte homing and homeostasis, Butcher [/bib_ref]. Monocytes phagocytose foreign substances in the tissues [bib_ref] Monocyte and macrophage heterogeneity, Gordon [/bib_ref]. The usual distribution of these five classes is 62%, 2.3%, 0.4%, 30% and 5.3% among WBCs in the body. This distribution of WBC describes the condition of the immune system. Considering the complexity of manually estimating the distribution of WBC, e.g., by consulting a human expert, many studies have introduced methods for automating the process through WBC segmentation, detection, and classification. Despite these numerous studies, which are greatly focused on the segmentation and detection tasks, less attention has been given to the WBC classification task and factors impacting the accuracy and performance of the task.
Accurate WBC classification is also beneficial for diagnosing leukemia, a type of blood cancer in which abnormal WBCs in the blood rapidly proliferate, decreasing the number of normal blood cells making the immune system vulnerable to infections In the US, around 60,000 people are diagnosed with leukemia every year, and around 20,000 people die of leukemia annually. From 2011 to 2015, leukemia was the sixth most common cause of cancer-caused death in the US. There are various types of leukemia, including ALL (Acute lymphocytic leukemia), AML (Acute myelogenous leukemia), CLL (Chronic lymphocytic leukemia), CML (Chronic myelogenous leukemia). Chronic leukemia progresses more slowly than acute leukemia which requires immediate medical care. Acute leukemia is characterized by proliferation of blasts, CLL is characterized by increased lymphocytes while CML shows markedly increased neutrophils and some basophils in the blood. Therefore, accurate classification of WBCs contributes to the diagnosis of leukemia.
Recent advancements in the field of computer vision and computer-aided diagnosis show a promising direction for the applicability of deep learning-based technologies to assist accurate classification and counting of WBC. Convolutional neural network (CNN) is one of the most common and successful deep learning architectures that have been utilized for analyzing and classifying medical imagery data [bib_ref] Deep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and..., Shin [/bib_ref] [bib_ref] Convolutional neural networks for medical image analysis: full training or fine tuning?, Tajbakhsh [/bib_ref] [bib_ref] CNN-based projected gradient descent for consistent CT image reconstruction, Gupta [/bib_ref] [bib_ref] Generative adversarial networks for noise reduction in low-dose CT, Wolterink [/bib_ref]. In this paper, we propose W-Net, a CNN-based network for WBC images classification. W-Net consists of three convolutional layers and two fully-connected layers, and they are responsible for extracting and learning features from WBC images and classifying them into five classes using a softmax classifier. In comparison to state-of-the-art methods, W-Net shows outstanding results in terms of accuracy. Further, we investigate the performance of several deep learning architectures in performing the WBC classification task. We applied and compared the performance of several architectures including W-Net, AlexNet [bib_ref] Imagenet classification with deep convolutional neural networks, Krizhevsky [/bib_ref] , VGGNet, ResNet [bib_ref] Deep residual learning for image recognition, He [/bib_ref] , and Recurrent Neural Network (RNN). Moreover, we compared the utilization of different classifiers such as softmax classifier and Support Vector Machine (SVM) on top of the adopted models. Moreover, we explore the effects of pre-training W-Net using public datasets, such as the LISC public [bib_ref] Automatic recognition of five types of white blood cells in peripheral blood, Rezatofighi [/bib_ref] , on its performance. Understanding the importance of large-scale datasets on the models' performance, we generate new WBC images using GAN [bib_ref] Generative adversarial nets, Goodfellow [/bib_ref] to augment current educational and research datasets.
# Contributions
The contributions of this paper are as follows. [bib_ref] W-Net: a CNN-based architecture for white blood cells image classification, Changhun [/bib_ref] We propose ❶ W-Net, a CNN-based network, designed to accurately classify WBCs while maintaining a high efficiency through minimal depth of the CNN architecture. ❷ We evaluate the performance of W-Net using a real-world large-scale dataset that consist of 6562 real images. ❸ We address and handle the problem of imbalanced classes of WBCs and achieve an average classification accuracy of 97% for all classes. ❹ We show how W-Net which consists of three convolutional layers stands among most popular CNN-based architectures, in the field of image classification and computer vision, in performing the WBCs classification task. ❺ Serving the purpose of advancing the task, we studied the applicability of transfer learning and generating larger datasets of WBC images using GAN for the public release. ❻ We generate and publicize synthetic WBC images using Generative Adversarial Network to be used for education and research purposes. The synthetic WBC images are verified by experiments and a domain expert to have a high degree of similarity to the original images. The pretrained W-Net and the generated WBC dataset are available for the public.
## Organization
The rest of the paper is organized as follows: in "Related works" section, we review literature. We introduce our model W-Net in "Methods" section. We evaluate W-Net through various experiments on WBC images in "Experiments" section. Our design choices and the experiment result are discussed in "Design considerations for W-Net" section. We release a new WBC dataset using GAN in "Dataset sharing" section. Finally, we conclude our study in "Conclusion" section.
# Related works
## Previous works
Analysis of white blood cells (WBC) has vital importance in diagnosing diseases. Distribution of the five WBC types, (basophils, eosinophils, lymphocytes, monocytes and neutrophils) reflects highly on the condition of the immune system. Analyzing the components of WBCs requires performing segmentation and classification processes. The traditional analysis of WBC includes observing a blood smear on a microscope and using the visible properties, such as shapes and colors, to classifing the blood cells. However, the accuracy of the WBCs analysis depends significantly on the knowledge and experience of the medical operator [bib_ref] A spectral and morphologic method for white blood cell classification, Wang [/bib_ref]. This makes the process of analyzing of WBCs using conventional methods timeconsuming and labor-intensive [bib_ref] A spectral and morphologic method for white blood cell classification, Wang [/bib_ref] [bib_ref] Semiautomatic white blood cell segmentation based on multiscale analysis, Dorini [/bib_ref] [bib_ref] Segmentation of white blood cells and comparison of cell morphology by linear..., Prinyakupt [/bib_ref]. Therefore, many studies have proposed computer-aided technologies to facilitate the WBC analysis through accurate cell detection and segmentation to reduce the manual efforts needed by human experts. For instance, Shitong and Min [bib_ref] A new detection algorithm (NDA) based on fuzzy cellular neural networks for..., Shitong [/bib_ref] have proposed an algorithm based on fuzzy cellular neural networks to detect WBCs in microscopic blood images as the first key step for automatic WBC recognition. Using mathematical morphology and fuzzy cellular neural networks, the authors achieved a detection accuracy of 99%. The detection of WBCs is followed by a segmentation process, which segments the image into nucleus and cytoplasm regions. This task has been pursued by several studies providing accurate segmentation using a variety of methods. The most common approach for nuclei segmentation is the clustering based on extracted features from pixels values [bib_ref] Recent computational methods for white blood cell nuclei segmentation: a comparative study, Andrade [/bib_ref] [bib_ref] Fuzzy c means detection of leukemia based on morphological contour segmentation, Viswanathan [/bib_ref]. The literature shows a successful nuclei segmentation using different clustering techniques, such as K-means, fuzzy K-means [bib_ref] Fuzzy c means detection of leukemia based on morphological contour segmentation, Viswanathan [/bib_ref] , C-means [bib_ref] Fuzzy c means detection of leukemia based on morphological contour segmentation, Viswanathan [/bib_ref] , and GK-means [bib_ref] Fuzzy based blood image segmentation for automated leukemia detection, Mohapatra [/bib_ref]. Among other unsupervised techniques for nuclei segmentation beside clustering, many studies utilized thresholding [bib_ref] Segmentation of white blood cells and comparison of cell morphology by linear..., Prinyakupt [/bib_ref] [bib_ref] Automatic segmentation, counting, size determination and classification of white blood cells, Nazlibilek [/bib_ref] [bib_ref] Unsupervised segmentation of leukocytes images using thresholding neighborhood valley-emphasis, Tosta [/bib_ref] [bib_ref] A novel algorithm for segmentation of leukocytes in peripheral blood, Cao [/bib_ref] , arithmetical operations [bib_ref] Development of a robust algorithm for detection of nuclei and classification of..., Hegde [/bib_ref] , edge-based detection [bib_ref] Fuzzy c means detection of leukemia based on morphological contour segmentation, Viswanathan [/bib_ref] , region-based detection, genetic algorithm [bib_ref] Leukocyte nucleus segmentation and nucleus lobe counting, Chan [/bib_ref] , watershed algorithm, and Gram-Schmidt orthogonalization [bib_ref] Automatic recognition of five types of white blood cells in peripheral blood, Rezatofighi [/bib_ref].
The literature on WBC segmentation process is very rich and provides valuable insights for the WBC identification. Andrade et al. [bib_ref] Recent computational methods for white blood cell nuclei segmentation: a comparative study, Andrade [/bib_ref] provides a survey and a comparative study on the performance of 15 segmentation methods using five public WBC databases. Some of these works are dedicated to the separation of adjacent cells, while many others addressed particularly the separation of overlapping cells. After the segmentation process, the WBC image classification or identification process is conducted. The distinction between the task of WBC identification and WBC image classification is the identification process aims to detect and identify leucocytes in an image, while the classification process aims to distinguish the different types of WBC. Even though many studies are dedicated to segmentation and identification task, fewer researches are addressed the classification of the WBCs. The literature shows that classification methods used for this purpose include the K-Nearest Neighbor (KNN) classifier [bib_ref] Semiautomatic white blood cell segmentation based on multiscale analysis, Dorini [/bib_ref] , Bayesian classifier [bib_ref] Segmentation of white blood cells and comparison of cell morphology by linear..., Prinyakupt [/bib_ref] [bib_ref] Scalable system for classification of white blood cells from Leishman stained blood..., Mathur [/bib_ref] , SVM classifier [bib_ref] Automatic recognition of five types of white blood cells in peripheral blood, Rezatofighi [/bib_ref] [bib_ref] A spectral and morphologic method for white blood cell classification, Wang [/bib_ref] [bib_ref] Fuzzy based blood image segmentation for automated leukemia detection, Mohapatra [/bib_ref] [bib_ref] A neural-network-based approach to WBC classification, Su [/bib_ref] , Linear Discriminant Analysis (LDA) [bib_ref] Isolation and two-step classification of normal white blood cells in peripheral blood..., Ramesh [/bib_ref] , decision trees and random forest classifier [bib_ref] Blood smear analyzer for white blood cell counting: a hybrid microscopic image..., Ghosh [/bib_ref] , and deep learning [bib_ref] Automatic recognition of five types of white blood cells in peripheral blood, Rezatofighi [/bib_ref] [bib_ref] Automatic segmentation, counting, size determination and classification of white blood cells, Nazlibilek [/bib_ref] [bib_ref] Development of a robust algorithm for detection of nuclei and classification of..., Hegde [/bib_ref] [bib_ref] A neural-network-based approach to WBC classification, Su [/bib_ref] [bib_ref] Automatic white blood cell classification using pre-trained deep learning models: Resnet and..., Habibzadeh [/bib_ref] [bib_ref] Application of ensemble artificial neural network for the classification of white blood..., Rawat [/bib_ref].
Recently, deep learning-based methods have been utilized for WBC classification and segmentation tasks [bib_ref] White blood cells image classification using deep learning with canonical correlation analysis, Patil [/bib_ref] [bib_ref] Classification of white blood cells using deep features obtained from convolutional neural..., Toğaçar [/bib_ref] [bib_ref] A review on traditional machine learning and deep learning models for WBCs..., Khan [/bib_ref]. Patil et al. [bib_ref] White blood cells image classification using deep learning with canonical correlation analysis, Patil [/bib_ref] incorporated canonical correlation analysis with CNN to extract and train on multiple nuclei patches with overlapping nuclei for WBC classification. Toğaçar et al. [bib_ref] Classification of white blood cells using deep features obtained from convolutional neural..., Toğaçar [/bib_ref] have utilized multiple CNN-based models, namely, AlexNet, GoogLeNet, and ResNet-50, for feature extraction and adopted quadratic discriminant analysis for classifying WBCs. Their method achieved an accuracy of 97.95% on a dataset of four categories: Neutrophil, Eosinophil, Monocyte, and Lymphocyte. Mohamed et al. [bib_ref] Improved white blood cells classification based on pre-trained deep learning models, Mohamed [/bib_ref] have investigated the use of deep CNN models over different shallow classifiers for WBC classification. For example, using a logistic regression classifier, extracting features using MobileNet-22 enabled a classification accuracy of 97.03%. Banik et al. [bib_ref] An automatic nucleus segmentation and CNN model based classification method of white..., Banik [/bib_ref] explored the use of combining features from different layers of CNN model to classify WBC in the BCCD dataset. Karthikeyan et al. [bib_ref] Interpolative Leishman-stained transformation invariant deep pattern classification for white blood cells, Karthikeyan [/bib_ref] proposed the LSM-TIDC approach to classify WBCs in blood smear images where a multi-directional model is used to extract texture and geometrical features that are then fed to a CNN model. Kutlu et al. [bib_ref] White blood cells detection and classification based on regional convolutional neural networks, Kutlu [/bib_ref] proposed using Regional-Based CNN model for WBC classification in blood smear images. Many other approaches have been proposed to tackle various challenges in the field of WBC using traditional machine learning and deep learning-based methods. Khan et al. [bib_ref] A review on traditional machine learning and deep learning models for WBCs..., Khan [/bib_ref] provided a comprehensive review of such practices and their impact on the field. [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] shows an overview of the performance and methods of the related works.
## Cnn with medical images
Due to the vast success in a variety of applications, CNN has been adopted in several medical applications where imagery inputs are analyzed for diagnosis or classification. In the field of medical imaging, CNN has been successfully utilized for histological microscopic image, pediatric pneumonia [bib_ref] Identifying medical diagnoses and treatable diseases by image-based deep learning, Kermany [/bib_ref] , diabetic macular edema [bib_ref] Identifying medical diagnoses and treatable diseases by image-based deep learning, Kermany [/bib_ref] , ventricular arrhythmias [bib_ref] Automated identification of shockable and non-shockable life-threatening ventricular arrhythmias using convolutional neural..., Acharya [/bib_ref] , thyroid anomalies, mitotic nuclei estimation [bib_ref] Identification of thyroid nodules in infrared images by convolutional neural networks, Moran [/bib_ref] [bib_ref] Mitotic nuclei analysis in breast cancer histopathology images using deep ensemble classifier, Sohail [/bib_ref] , neuroanatomy [bib_ref] DeepNAT: deep convolutional neural network for segmenting neuroanatomy, Wachinger [/bib_ref] , and others [bib_ref] Deep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and..., Shin [/bib_ref] [bib_ref] Convolutional neural networks for medical image analysis: full training or fine tuning?, Tajbakhsh [/bib_ref] [bib_ref] CNN-based projected gradient descent for consistent CT image reconstruction, Gupta [/bib_ref] [bib_ref] Generative adversarial networks for noise reduction in low-dose CT, Wolterink [/bib_ref] [bib_ref] Multi-feature representation for burn depth classification via burn images, Zhang [/bib_ref] [bib_ref] Transfer learning in medical image segmentation: new insights from analysis of the..., Karimi [/bib_ref] [bib_ref] A deep learning approach for mitosis detection: application in tumor proliferation prediction..., Nateghi [/bib_ref] [bib_ref] Image quality assessment of pediatric chest and abdomen ct by deep learning..., Yoon [/bib_ref] [bib_ref] Efficiency of a deep learning-based artificial intelligence diagnostic system in spontaneous intracerebral..., Wang [/bib_ref] [bib_ref] Hahn-PCNN-CNN: an end-to-end multimodal brain medical image fusion framework useful for clinical..., Guo [/bib_ref] [bib_ref] Application of a deep learning image reconstruction (DLIR) algorithm in head CT..., Sun [/bib_ref]. Kermany et al. [bib_ref] Identifying medical diagnoses and treatable diseases by image-based deep learning, Kermany [/bib_ref] showed that CNN can detect diabetic macular edema and age-related macular degeneration with high accuracy and with a comparable performance of human experts. The authors also demonstrated the applicability of CNN in diagnosing pediatric pneumonia from chest X-ray images. Alexander et al.have provided the state-of-the-art performance (by the publication date) using CNN for histopathological image classification on the dataset of the ICIAR 2018 Grand Challenge on Breast Cancer Histology Images. Acharya [bib_ref] Automated identification of shockable and non-shockable life-threatening ventricular arrhythmias using convolutional neural..., Acharya [/bib_ref] have shown that CNN can be used to accurately detect shockable and non-shockable life-threatening ventricular arrhythmias. Wachinger et al. [bib_ref] DeepNAT: deep convolutional neural network for segmenting neuroanatomy, Wachinger [/bib_ref] proposed DeepNAT, a CNN-based approach for automatic segmentation of Neu-roAnaTomy in magnetic resonance images. The authors showed that their approach provided comparable results to those of state-of-the-art methods.
# Methods
This section provides a description of the dataset used in this study, the pre-processing steps for the WBC images, and the proposed CNN-based architecture for WBC classification.
## Pre-processing of wbc images
Prior to the model creation and training, WBC images are pre-processed using three steps: ❶ image border cropping, ❷ image re-sizing, and ❸ image normalization. To eliminate the external borders of the image and to focus on the WBC, we remove the top 80 pixels, the bottom 81 pixels, the left 80 pixels, and the right 80 pixels of the image. The resulting cropped images, i.e., images with a size of 200 × 200 × 3, are then re-sized to 128 × 128 × 3 for properly fitting them into a GPU memory and for efficient processing. Samples of the processed images are shown in [fig_ref] Figure 1: Neutrophil, eosinophil, basophil, lymphocyte and monocyte from the left [/fig_ref]. The image normalization process was applied to reduce the heterogeneity of the RGB distribution in the images and to prevent over/underflow. This step is shown in .
## W-net: architecture and design
We introduce our CNN-based model architecture for WBC image classification. As illustrated in , W-Net consists of three convolutional layers and two fully-connected layers, and they are responsible for extracting and learning features from WBC images to accurately classifying them into five classes using a softmax classifier. Each convolutional layer has a kernel size of 3 × 3 with stride of size 1 and uses ReLU activation function f(x) = max(0, x). The first convolutional layer has 16 filters, the second has 32 filters and third has 64 filters. After each convolutional layer, there is a max-pooling layer of size 2 × 2 with stride of size 2 and zero padding. We also use dropout regularization with p = 0.6 [bib_ref] Dropout: a simple way to prevent neural networks from overfitting, Srivastava [/bib_ref] to prevent overfitting in each convolutional layer. The output of the third convolutional layer is flattened and fed into the first fully connected layer which has 1024 units. ReLU activation, and dropout with p = 0.6 are followed. The second fully connected layer has five units (five classes of WBC) and is followed by softmax classifier to map the output (features) to one of the five classes. The network has a total size of 16,806,949 trainable parameters. The model parameters were initialized using Xavier uniform initializer W ~ U[− x, x], where x = sqrt(6/(in + out)). The training of models is guided by minimizing the softmax-cross-entropy loss function
[formula] − x p(x)logq(x) , where q(x i ) = exp(x i )/ j exp(x j )) using Adam optimizer �θ t = −n ·m t / v t + ǫ [63] [/formula]
with a learning rate of 0.0001. The training process is conducted with different batch sizes and terminated by the conclusion of 500 training epochs. The evaluation of the models is conducted using a tenfold cross-validation approach [bib_ref] A survey of cross-validation procedures for model selection, Arlot [/bib_ref]. The structure is illustrated in [fig_ref] Table 3: The structure of five layers [/fig_ref]. The hyperparameters are described in [fig_ref] Table 4: Hyperparameters for all the models [/fig_ref]. Design choices for W-Net are discussed in "Design considerations for W-Net" section.
## Experiments
We show the performance of W-Net for WBC classification and compare the softmax classifier of W-Net with SVM. We show that W-Net provides remarkable results in the WBC classification by comparing it to the prior work. We also show how the number of layers affects performance. The comparison includes AlexNet, VGG-Net, ResNet and RNN models. For transfer learning, we provide insights on adopting pre-trained W-Net to gain higher WBC classification performance on public datasets. ROC curve and AUC are a useful method for evaluating a system in medical area and are usually used to classify a binary task such as a diagnosis. However, we remark that our results are only based on an accuracy, because the output of our model is multiple-class not the binary. [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] in "Appendix 1" shows the accuracy achieved by W-Net using tenfold cross-validation approach.
## W-net performance
Conducting the experiments required 33.87 h of model's training time. For the neutrophil, 1800 images were used for training and 206 images were used for testing in each fold, and the average accuracy was 98%. For the eosinophil, 1179 images were used for training and 131 images were used for testing in each fold, and the average accuracy was 97%. For the basophil, 340 images were used for training and 37 images were used for testing in each fold, and the average accuracy was 95%. For the lymphocyte, 1509 images were used for training and 167 images were used for testing in each fold, and the average accuracy was 97%. For the monocyte, 1074 images were used for training and 119 images were used for testing in each fold, and the average accuracy was 97%. The average overall accuracy of the five WBC classes was 97%. As shown in [fig_ref] Figure 3 a: Since we the WBC images are 128 × 128 × 3 images,... [/fig_ref] , it provides ROC curve and Precision-Recall (PR) curve in (a) and (b) respectively, based on the idea of one vs rest for multi-class classification. Each line in (a) represents each class of the five WBC classes, and our W-Net model achieved an AUC of 0.97 on average on ROC curve. On the one hand, it achieved an AUC of 0.98 on average on PR curve.
## W-net versus w-net-svm performance
We compared softmax classifier of W-Net with SVM to demonstrate classifier's abilities in performing the WBC
[formula] l(y) = max(0, 1 − tㆍy) instead of softmax (W-Net). [/formula]
We followed the same experimental settings adopted in previous experiment including the training parameters, dataset, pre-processing steps, workstation environment, and tenfold cross-validation approach for the evaluation. The network has a total of 16,806,949 trainable parameters. [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] in "Appendix 1" shows the performance of W-Net-SVM using tenfold cross-validation in the WBC classification task. The training time of W-Net-SVM was 33.79 h. The achieved results for the neutrophil, eosinophil, basophil, lymphocyte, and monocyte classes are 98%, 97%, 87%, 98%, and 97%, respectively. The overall average accuracy of the five classes was 95%.
## Wbc classification with alexnet
This experiment adopts AlexNet architecture in the WBC classification task. AlexNet network consists of five convolutional layers and three fully-connected layers which apply ReLU activation function (in all layers except the last (softmax) layer). The training of AlexNet model is conducted by minimizing the softmax-cross-entropy loss function using the momentum optimizer θ t = −γ ν t−1 − ηg t [bib_ref] On the importance of initialization and momentum in deep learning, Sutskever [/bib_ref]. Using a cross-validation approach, the best training hyperparameters that achieved the best WBC classification accuracy are described in [fig_ref] Table 4: Hyperparameters for all the models [/fig_ref]. We follow the same experimental settings adopted in previous experiments by using same dataset, pre-processing steps (except for the image size, we re-sized the images to 224 × 224 × 3 for AlexNet), workstation environment, and the tenfold cross-validation evaluation approach. The AlexNet-based network has a total of 46,767,493 trainable parameters. [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] in "Appendix 1" shows the performance of AlexNet using a tenfold cross-validation approach in the WBC classification task. The overall average accuracy is 84% (see [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] for details).
## Wbc classification with vggnet
We compared W-Net with VGGNet to demonstrate the effectiveness of W-Net in the WBC image classification. We trained a VGGNet-based model that consists of 16 convolutional layers and three full-connected layers, which followed with ReLU activation function. The model training is conducted using the minimization of the softmax-cross-entropy loss though Adam optimizer. Using a cross-validation method, the best training hyperparameters are described in [fig_ref] Table 4: Hyperparameters for all the models [/fig_ref]. This experiment followed the same experimental settings adopted in previous experiments. The VGGNet-based model includes a total of 121,796,165 trainable parameters. The training of the VGGNet-based model required 510.59 h of training time. [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] in "Appendix 1" shows the results of the tenfold cross-validation of VGGNet-based model in the WBC classification. The overall average accuracy of the five classes is 44% (see [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] for details).
## Wbc classification with resnet
We adopt ResNet50 and ResNet18 networks for WBC classification, which consists of 50 and 18 convolutional layers, respectively. Both models are trained by minimizing the softmax-cross-entropy loss using momentum optimizer. Using a cross-validation approach, the best training hyperparameters to achieve the highest accuracy in the WBC classification task are described in [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] in "Appendix 1" shows the classification accuracy obtained by of ResNet50 model using the tenfold cross-validation approach, while [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] shows the results of ResNet18. The overall average accuracy of the five classes for ResNet50 is 51%. On the one hand, ResNet18 achieved the overall average accuracy of the five classes is 79% (see [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] for details, respectively).
## Wbc classification with rnn
We explore the capabilities of RNN in the WBC classification task. Using RNN for WBC image classification, we adopted the common approach by considering the image rows as sequences and the columns as timesteps. [fig_ref] Table 4: Hyperparameters for all the models [/fig_ref]. Once hyperparameters are selected, we conducted a new training process using a tenfold evaluation approach, where 10 different models are trained and evaluated using ten fold splits (each time a model is trained on nine folds and tested on one fold). The achieved accuracy for the individual classes are as follows: neutrophil 89%, eosinophil 88%, basophil 57%, lymphocyte 93%, and monocyte 90%.
The results are shown in [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] , "Appendix 1". The average accuracy of the five classes is 83%. show that W-Net outperforms other RNN-and CNNbased model's architectures and an architecture with a small number of layers is also better than an architecture with many layers. The detailed results of tenfold cross validation for all experiments are in "Appendix 1".
## Models comparison for wbc classification
## Further training with public data
The LISC public dataset [bib_ref] Automatic recognition of five types of white blood cells in peripheral blood, Rezatofighi [/bib_ref] includes WBC images of size 720 × 576 × 3 that were collected from peripheral blood of eight normal people. The images are classified by a hematologist into five types of WBC: neutrophils, eosinophils, basophils, lymphocytes and monocytes. For pre-processing the public dataset, we cropped the WBC images (nucleus and cytoplasm regions) in the original images, and then re-sized the images to 128 × 128 × 3 for training. We used a total of 254 WBC images as our dataset: 56, 39, 55, 56 and 48 images for neutrophil, eosinophil, basophil, lymphocyte and monocyte, respectively. Using the LISC public data, this experiment shows the performance of W-Net when adopted for different datasets. Moreover, we show the performance of W-Net using transfer learning when a pre-trained W-Net is fine-tuned to classify WBCs from different dataset or used for different WBC-related tasks. To this end, we conducted two experiments as follows: ❶ W-Net architecture is used for building a WBC classifier trained using only the LISC public data, ❷ a pre-trained W-Net with softmax classifier from "W-Net performance" section is fine-tuned to classify WBCs from LISC public data. Except the training epochs, the training hyperparameters are set to be identical in both experiments. In the first experiment, W-Net-based model was trained from scratch using 4000 training epochs (254 × 4000/5 iterations) on the LISC public data. The training process was concluded after 10.33 h. In the second experiment, we establish a pretrained W-Net-based model (trained on our dataset for 500 training epochs.) to be used on the LISC public data. The pre-trained W-Net-based model was fine-tuned for 4000 epochs (254 × 4000/5 iterations) on the public data. The training process was concluded after 10.83 h. [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] in "Appendix 1" shows the result of the first experiment where W-Net is used to classify WBCs from the LISC public data. The achieved results is an average accuracy of 91%. [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] in "Appendix 1" shows the result of the experiment. The average accuracy achieved using a pre-trained W-Net model is 96%.
In the results, the second experiment shows a better performance. This result shows that training a model in large-scale dataset (such as the one used for this study) can benefit other transfer learning tasks, where the model is fine-tuned to other dataset or performing other WBC-related tasks. We share our pre-trained model on GitHuband believe that using the transfer learning property (transfer learning in the same domain) of deep learning models can help other researchers in the field.
## Design considerations for w-net
Design choices for our deep learning architecture are described in this section. There are two challenging issues to consider in choosing a specific architecture in the large design space for WBC classification problem: One is how to figure out the data imbalance problem, and the other is to classify similar-looking images into the relatively small number of classes. In many datasets in real world, data imbalance is quite common and WBC images resembles way more each other compared to objects in traditional image classification problems. Also, the number of classes is quite limited compared with the traditional object identification problems such as Ima-geNet challenge. Therefore, it is necessary to take a different approach to the classification problem.
## Handling data imbalance: large batch and sampling
The results show that W-Net performs well despite the dataset's imbalance, which is observed by the number of samples for each class. Even though the least-represented class in the dataset (basophil with 6% of the dataset) show the least accuracy of 95% in comparison to other classes, this accuracy is still higher than the results achieved by other methods, e.g., CNN-based and RNNbased models, for the same class. This performance can be due to several reasons. For instance, the evaluation of all experiments follows a stratified k-folds crossvalidation approach, which preserves the percentage of samples across all folds. Using this approach allows the sampling from all classes in different ratios in each fold, which dictates the inclusion of all classes in both the training and testing phases. When using a small batch size, e.g., five samples as adopted during the training of W-Net, the error resulting from misclassifying one class, especially from underrepresented classes, highly impacts the average cost of the learning epoch and contributes in an effective learning process for these classes. In contrast, using a large batch size and considering a random sampling scheme for batching could result in minimizing the effect of misclassification of underrepresented classes since performing well on other classes could out-weigh the misclassification of small, if any at all, samples from classes with small ratios in the dataset.
Having different distributions of image samples per class is a hard part to classify WBC images. W-Net achieves an accuracy of 95% for identifying the basophil class which are represented with the least number of samples (377 samples and a ratio of 6% of the dataset). This result is remarkable knowing that all other CNNbased and RNN-based models achieved an accuracy below 56% and 57%, respectively, for the same class. The overall average accuracy of W-Net is 97%, which is the highest among other methods for WBC classification. Considering the results for this large-scale dataset, W-Net presents a state-of-the-art performance.
Furthermore, the result of W-Net with softmax classifier is 97%, the result of W-Net with SVM classifier is 95% and they seem similar. However, for the basophil class that has 6% distribution of our dataset, the accuracy of W-Net with SVM is only 87% and it is lower than 95% the result of softmax. The W-Net-SVM uses the hinge loss function, while W-Net uses the softmax cross-entropy loss function. The nature of optimization under these functions differs since the optimization using the hinge loss concludes when finding parameters that satisfy the classification with the predefined margin. However, using softmax cross-entropy loss keeps the optimization going beyond a specific margin pushing the decision boundaries further. This allows the model to maintain robust generalization capabilities, hence the better performance of W-Net over W-Net-SVM. AlexNet has many layers than our W-Net, however, the average accuracy is 84%, and especially the average accuracy of the basophil class that has 6% distribution of our dataset is 33%. This means the SVM classifier and the network of AlexNet are not appropriate to address the unbalanced dataset. As a result, we can claim that W-Net with softmax classifier is more effective than AlexNet and W-Net with SVM classifier in WBC image classification area.
## Wbc dedicated architecture with shallow depth
In the tenfold cross-validation evaluation of W-Net, the minimum average accuracy is 91% (basophil, Fold-9) and maximum average accuracy is 100%. However, in the case of VGGNet and ResNet50 architectures which have more depth (considering the number of layers), the variance between the folds is from 0 to 100% resulting in 44% tenfold average accuracy, and from 0 to 100% resulting in 51% tenfold average accuracy, respectively. In a comparison between ResNet50 and ResNet18, since ResNet18 consists of a shallower layer than ResNet50, the overfitting problem seems to occur less. It leads that ResNet18 shows better performance with 79% on average than ResNet50. This means that very deep networks may not be the optimal choice for WBC image classification. Most of the state-of-the-art CNN-based models (e.g., AlexNet, VGGNet, and ResNet) use larger receptive fields, (e.g., 7 × 7 in case ResNet and 11 × 11 in the case of AlexNet), which seem to work better on larger images with larger objects (classes). However, handling the WBC classification task requires adopting smaller filters to bring attention to finer receptive fields that hold relevant features.
The results of this research show that architectures such as W-Net's, which has five layers (three convolutional and two fully-connected.), can be sufficient and more effective in the WBC classification task in comparison to other deeper networks such as VGGNet, ResNet50 and ResNet18. In general, deep networks are known to perform well for the image classification, the VGGNet and ResNet with deep networks show good performance in ILSVRC. However, they did not show good performance in WBC image dataset. We claim that our dataset to be classified is different from the dataset aimed by those deep networks in two aspects: (1) the ILSVRC dataset has 1000 classes, but our WBC dataset has only five classes, and (2) The images of the ILSVRC dataset are very different from each other (For example, they are dog, bird, flower and food etc.), while our dataset has very high visual similarity.
To support this claim we conducted two simple experiments, ❶ the first experiment was to run W-Net on 200 classes (bird, ball and car etc.) of images from Tiny Ima-geNet dataset [68] and ❷ the second experiment was to run W-Net on five classes without visual similarity (fish, clothes, chair, car and teddy bear) from Tiny Ima-geNet dataset with the same (imbalance) distribution of our WBC dataset. In these two experiments, we only used different dataset with our WBC dataset, and used same network, parameters (learning rate and training epoch etc.) and tenfold cross-validation approach with our W-Net. In the first experiment, we used the dataset with 200 classes, and each class had 500 images. We used total 100,000 images. The result from the first experiment showed 100% accuracy for the 200th class, but 0% accuracy for the other 199 classes. The average accuracy was 0.5%, it showed that the model was not trained at all. In the second experiment, we used the dataset with 5 classes, and each class had 500, 333, 100, 433, 300 (making them have the same distribution with our dataset) images. We used total 1666 images. The result from the second experiment showed 34% accuracy for the third class (100 chair images), and 84%, 78%, 90% and 65% for other classes, respectively. The average accuracy was 79%, which was not as good as the results of W-Net using our dataset. Therefore, we claim that a simple network may be better to classify our WBC dataset with data distribution imbalance, small number of classes, and visual similarity.
## Why not rnn?
RNN-based models perform well in sequential data and show remarkable results in capturing temporal dependencies within the input data. There are different variations of RNN, and for our experiments we used LSTM models for their abilities to handle long-term dependencies (e.g., 128 sequences in our application) and the vanishing gradient problem. The average achieved results when using one-layer LSTM model with 32 hidden units is 83%. This result is far from the results achieved by W-Net (97%).
However, it outperforms other CNN-based models such as VGGNet (44%) and ResNet50 (51%). Karol et al. [bib_ref] Draw: a recurrent neural network for image generation, Gregor [/bib_ref] have also shown that RNN can encode independent scenes within an image instead of processing the entire image as a single input. Adopting sequential processing of white blood images via LSTM, enables the model to extract/adapt to patterns/changes in the scene to build a more robust model than following single-shot processing.
## Dataset sharing
Recent advances in big data have also led to advances in deep learning, accordingly having a good dataset has become important. In this section, we generate new WBC image samples using Generative Adversarial Networks (GAN) [bib_ref] Generative adversarial nets, Goodfellow [/bib_ref] then release them in public for education and research to help other researchers. GAN is a deep learning architecture for generating new artificial samples, it composes of two deep networks: ❶ the generator G, and ❷ the discriminator D. The G generates new samples from the domain, and the D classifies whether the samples are real or fake. The output of the D is used to update both the model weights of the D itself and the G. Accordingly, the performance of the G depends on how well the D performs. GAN can be expressed by: min
[formula] G max D V (D, G) = E x∼p data (x) [logD(x)]+ E z∼p z (z) [log(1 − D(G(z)))] [/formula]
, where x ~ p data (x) and z ~ p z (z) indicate the distribution of a real data and a fake data respectively, the D aims to maximize logD(x) and G aims to minimize log(1 − D(G(z))), to maximize the chance to recognize real images as real and generated images as fake. This expression defines GAN as a minimax game.
## Experimental settings
We use the same dataset (6562 WBC images of size of 128 × 128 × 3), similar experimental settings of previous experiments, and Deep Convolutional Generative Adversarial Network (DCGAN)to train (G and D) models for generating images. For the network of D, six convolutional layers, one fully connected layer, LeakyReLU [bib_ref] Rectifier nonlinearities improve neural network acoustic models, Maas [/bib_ref] activation, sigmoid activation and dropout are used. For the network of G, six convolutional layers, one fully connected layer, ReLU activation, sigmoid activation, dropout, and batch normalizationare used. The training hyperparameters are set as follows: alpha 0.2, momentum 0.9, batch size 1, learning rate 0.00001, dropout 0.6, and training epochs 10,000. The network of G and D have a total of 2,780,099 and 69,878,401 trainable parameters. It took 191. h to train G and D models for five WBC classes, and it took an average of 18 min to generate 1000 images per each class. We generated 1000 plausible WBC images of size of 128 × 128 × 3 for each class (a total of 5000 images). shows the samples of both the original images (left side) for training DCGAN model and the generated images (right side) by trained DCGAN model. The first row of the is the neutrophil class, followed by the eosinophil, the basophil, the lymphocyte, and the monocyte.
## Generated image quality
To see how similar images were generated from the original images, we verified the generated WBC images using ❶ baseline-W-Net, ❷ generative-W-Net (i.e., W-Net trained on generated synthetic dataset), ❸ cosine similarity, and ❹ domain-expert experiment with a medical laboratory specialist. First, we experimented to classify the generated images using W-Net. [fig_ref] Table 6: The confusion matrix for classification experiment result with generated WBC images using... [/fig_ref] shows the confusion matrix for the results achieved for the classification of the generated WBC im-ages. The second column indicates true classes, the second row indicates predicted classes, and the images are well-classified with 100% accuracy by W-Net. Second, we trained W-Net model using the 5000 generated synthetic images. For the training, we follow the same experimental settings of creating the baseline-W-Net. Then, we evaluated the generative-W-Net for classifying the 6562 real WBC images. [fig_ref] Table 7: The confusion matrix for classification experiment result with real WBC images using... [/fig_ref] shows the confusion matrix for the results achieved for the classification of real WBC images using generative-W-Net. The images are classified with an accuracy of 95%, precision of 93%, recall of 95%, and F1-score of 94%. Third, we measure the similarity between the original images and the generated images using cosine similarity. We first measure the cosine similarity between the original images and the original images for each class (e.g., 377 vs. 377 for the basophil class), then we measure the cosine similarity between the original images and the 1000 generated WBC images for each class (e.g., 377 vs. 1000 for the basophil class) and then we compare them. the cosine similarity between the original images and generated images. It was 4% for the neutrophil, 3% for the eosinophil, 7% for the basophil, 6% for the lymphocyte, and 6% for the monocyte with average 5% for five classes. Fourth, we conducted a domain-expert experiment on how well a medical laboratory specialist could classify the generated WBC images. The dataset used in this experiment consists of 10 random original images and 10 random generated images for each class, i.e., a total of 100 images. Without informing the medical laboratory specialist of the source of the WBC images in the dataset, we asked for the classification of provided images. [fig_ref] Table 9: The confusion matrix for the user experiment result with the medical laboratory... [/fig_ref] shows the confusion matrix for this experiment.
The results show that the specialist well-classified the given WBC samples with an accuracy of 95%. Among the five misclassified images, there are three original images and only two generated images. The results of all verification methods for the generated images show that the generated images are similar to the original images. We released the generated (labeled) WBC images on GitHubfor the education and research purposes.
# Conclusion
Analysis of WBC images is essential for diagnosing leukemia. Although there are several methods for detecting and counting WBCs from microscopic images of a blood smear, the classification of the five types of WBCs is still a challenge in real-life applications, which we addressed in this work. The rapid growth in the area of computer vision and machine/deep learning have provided feasible solutions to classification tasks in many domains. This work proposed W-Net, a CNN-based architecture with a small number of layers, to accurately classify the five WBC types. We evaluated W-Net on a real-world largescale dataset and addressed several challenges such as the transfer learning property and the class imbalance. W-Net achieved an average classification accuracy of 97%. Moreover, we compared the result of W-Net with W-Net with SVM, AlexNet, VGGNet, ResNet and RNN architectures to show the superiority of W-Net which consists of three layers over other architecture. We synthesized a dataset of new WBC image samples using DCGAN, which we released to the public for education and research purposes. Even though our W-Net model provides good performance with an average classification accuracy of 97%, it still remains an error of 3%. In the future work, we will conduct the dataset augmentation using our generative model based on DCGAN, to address the dataset imbalance. Then, we will carry out additional experiments to further increase the accuracy performance of the classification model with the balanced dataset.
## Appendix 1: the detailed results for all experiments
In this section, we show the detailed results of tenfold cross validation for W-Net [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] , W-Net-SVM [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] , AlexNet [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] , VGGNet [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] , ResNet [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] , RNN [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref]. [fig_ref] Table 1: Related work highlighting the used datasets, their size, number of classes [/fig_ref] The result of ResNet18 for classification using tenfold cross-validation
[fig] Figure 1: Neutrophil, eosinophil, basophil, lymphocyte and monocyte from the left. These were cropped and rescaled with 128 × 128 × 3 for efficient trainingFig. 2 An overview of the pre-processing and the proposed CNN-based architecture for WBC image classification. The pre-processing consists of cropping, re-sizing and normalizing. Three convolutional layers (including three pooling layers) are in charge of extracting and learning features, and two fully connected layers are in charge of classification [/fig]
[fig] Figure 3 a: Since we the WBC images are 128 × 128 × 3 images, we feed the model with batches of 128 sequences of size 128 × 3. The RNN model adopted in this experiment consists of only one single hidden layer. The experimental settings for the training process are set with the following search space: learning rate = 0.0001, 0.001, 0.003, 0.01, 0.1, 0.3, batch size = 16, 32, 64, 128 and hidden units = 16, Provides ROC curve of our W-Net model based on the idea of one versus rest for multi-class classification, and b shows Precision-Recall curve. In a, each class achieves an AUC of 0.97 on average and achieves an AUC of 0.98 on average in b [/fig]
[table] Table 1: Related work highlighting the used datasets, their size, number of classes (C), employed methods, and accuracyThe parts in bold mean our model [/table]
[table] Table 2: The number of five type samples in the dataset [/table]
[table] Table 3: The structure of five layers (Conv. and FC.) for W-Net [/table]
[table] Table 4: Hyperparameters for all the models [/table]
[table] Table 5: The result of accuracy, precision, recall, F1-score on average and the number of layers for all experimentsThe parts in bold mean our model [/table]
[table] Table 8: shows the difference inFig. 4 Left side: the original images of size of 128 × 128 × 3 for training DCGAN model. Right side: the synthesized images of size of 128 × 128 × 3 by trained DCGAN model. The first row is the neutrophil class, followed by the eosinophil, the basophil, the lymphocyte, and the monocyte classes [/table]
[table] Table 6: The confusion matrix for classification experiment result with generated WBC images using W-Net modelThe images were well-classified with 100% accuracy [/table]
[table] Table 7: The confusion matrix for classification experiment result with real WBC images using the fake W-Net modelThe images were classified with 95% accuracy [/table]
[table] Table 9: The confusion matrix for the user experiment result with the medical laboratory technologistThe technologist classified the generated WBC images with 95% accuracy [/table]
|
A Confirmed Case of SARS-CoV-2 Pneumonia with Routine RT-PCR Negative and Virus Variation in Guangzhou, China
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pneumonia is a newly recognized disease, and its diagnosis is primarily confirmed by routine RT-PCR detection of SARS-CoV-2. However, we report a confirmed case of SARS-CoV-2 pneumonia with routine RT-PCR negative. This case has been finally diagnosed by Nanopore sequencing combined with antibody of SARS-CoV-2. Simultaneously, the ORF and NP gene variation of SARS-CoV-2 were found. This case has highlighted that false negative results could be present in the routine RT-PCR diagnosis, especially with virus variation. At the moment, Nanopore pathogen sequencing and antibody detection have been found effective in clinical diagnosis.
A c c e p t e d M a n u s c r i p t
# Background
Recently, coronavirus disease 2019 (COVID-has become a global epidemic, up to June 25, 2020, there have been 9,296,202 confirmed cases and 479,133 deaths of COVID-19 around the world [bib_ref] Clinical Characteristics of Coronavirus Disease 2019 in China, Guan [/bib_ref]. According to WHO and Chinese interim guidance, the confirmation of diagnosis is usually via routine real-time reverse transcriptase polymerase chain reaction (RT-PCR) detection of SARS-CoV-2. However, some studies found that the routine RT-PCR may present false negative results [bib_ref] Correlation of Chest CT and RT-PCR Testing in Coronavirus Disease 2019 (COVID-19)..., Ai [/bib_ref] , but these reports didn't clarify the cause of the false negative results and the significance of its clinical and epidemic impact.
Coincidentally, we have found a routine RT-PCR negative, but confirmed case of SARS-CoV-2 pneumonia in Guangzhou, China. This report mainly describes the clinical diagnosis process and the new diagnostic methodology, the evolution and mutation analysis of viruses, and the management and some unexpected discovery of this case.
## Case report
The patient is a fifty-seven years old women, who returned to Guangzhou from her hometown, Xiantao city of Hubei Province with her families on January 20, 2020 [fig_ref] Figure 1: Timeline of the patient's clinical course [/fig_ref].
She had no contact with patients with fever or COVID-19, or wild animals. And she has no chronic disease or history of smoking.
On January 30, 2020, she developed fever and so did her husband. She had a maximum temperature of 37.8°C with symptoms of chills, chest and back pain, and no other respiratory or digestive symptoms, without treatment. Due to no easing of the symptoms, she and her husband went to see a doctor in the heat outpatient of our hospital (hospital A) two days later (February 1). Considering their clinical symptoms and history of stay in the epidemic area of COVID-19, the routine RT-PCR detection of SARS-CoV-2 was primarily carried out for them with the oropharyngeal swab sample and sent to Guangdong Centers for Disease Control and Prevention (CDC). Unfortunately, the detection result from CDC was positive of SARS-CoV-2 for her husband, but negative for her. Subsequently, her husband was taken to hospital B, which was the designated treatment hospital of COVID-19, for further treatment, and she was sequentially admitted to hospital A as a suspected patient.
A c c e p t e d M a n u s c r i p t [bib_ref] Correlation of Chest CT and RT-PCR Testing in Coronavirus Disease 2019 (COVID-19)..., Ai [/bib_ref] On admission (Hospital Day 1) of this patient in our case, the physical examination revealed a body temperature of 37.3°C, pulse of 87 beats per minute, respiratory rate of 20 breaths per minute, blood pressure of 114/78 mm Hg, and oxygen saturation of 95% (breath ambient air). She preserved sanity, and lung auscultation revealed the sound of breath was rough, but without rhonchi or moist crackles. Laboratory examinations were performed which showed that leukocyte was 2.60X10 9 /L, lymphocyte was 0.9X10 9 /L. Procalcitonin (PCT), liver and kidney function, enzymatic indicators and D-dimer were normal .
Chest computed tomographic (CT) scan performed at that time reported the two lungs being scattered with ground glass lesions, which obviously appeared on the lower right dorsal segment/outer basal segment of the lung [fig_ref] Figure 2: Imaging of chest computed tomographic [/fig_ref]. Based on the above mentioned, we classified her as a highly suspected patient of SARS-CoV-2 pneumonia, which made as a clinical diagnosis. Thus she was isolated in a single room, and was given a treatment with low-flow oxygen and moxifoxacin (oral, 400mg qd), according to the treatment guideline of community-acquired pneumonia [bib_ref] Diagnosis and treatment of community-acquired pneumonia in adults: 2016 clinical practice guidelines..., Cao [/bib_ref]. However, her fever was rising day by day, and the thermal spike presented on Hospital Day 3, when the maximum temperature reached 38.5℃.
Fortunately, other symptoms did not deteriorate. The routine RT-PCR detection of SARS-CoV-2 was repeatedly carried out with oropharyngeal swab sample again on Hospital Day 2, and the result remained negative. Subsequently, a nasopharyngeal swab specimen was obtained and examined for respiratory pathogens using a rapid nucleic acid amplification test (NAAT; QIAstat-Dx, Respiraotry Panel, LOT 190255); and this was reported back in about 1 hour to show negative for all pathogens tested, including influenza A and B, parainfluenza, RSV, rhinovirus, adenovirus and other coronaviruses etc.
Given she was diagnosed as a highly suspected patient of SARS-CoV-2 pneumonia. Thus, the routine RT-PCR detection of SARS-CoV-2 was used as a conventional monitor from However, all of these test results were still negative [fig_ref] Table 2: The Results of Nucleic Acid and Antibody Detection of SRSA-CoV-2 [/fig_ref]. As her fever went on, the antiviral drug Lianhuaqingwen capsule, which is one kind of natural herbal medicine that has shown to have antiviral effect [bib_ref] Natural herbal medicine Lianhuaqingwen capsule anti-influenza A (H1N1) trial: a randomized, double..., Duan [/bib_ref] , was used for treatment of the patient on Hospital Day 7 [fig_ref] Figure 1: Timeline of the patient's clinical course [/fig_ref] , and the temperature gradually dropped to normal on Hospital Day 12. The sputum sample was sent to the lab for clinical metagenomic next-generation sequencing (mNGS) testing on Hospital Day 8. Unexpectedly, there was a SARS-CoV-2 genomic segment detected, however, the genome coverage was only 75 bp (aboundacne, 0.05%) A c c e p t e d M a n u s c r i p t 6 (Supplementary . In order to further validate the existence of SARS-CoV-2, the same sputum sample was tested by Nanopore. To our excitement, SARS-CoV-2 genomic sequence was again detected. After that, another sputum sample collected on Hospital Day 12 has continued to show SARS-CoV-2 gene detection by Nanopore [fig_ref] Figure 3: Length and quality distributions of the nanopore data and Taxonomic annotation results... [/fig_ref]. The chest CT was retested on February 12 (Hospital Day 12), which presented that the inflammatory exudation in the field of the right lower lung was increased and became dense [fig_ref] Figure 2: Imaging of chest computed tomographic [/fig_ref]. Moreover, the antibody of mycoplasma pneumoniae was 1:320 (positive range> 1:40), which hinted the infection of mycoplasma pneumoniae . Hence, the Moxifoxacin and Lianhuaqingwen capsule were still prescribed [fig_ref] Figure 1: Timeline of the patient's clinical course [/fig_ref]. The IgM and IgG antibody were retested and the results were positive on Hospital Day 13, and there was an approximate tenfold (0.067 vs 0.673) increase of IgG . Furthermore, the micro-neutralization antibody (IgM and IgG) of SARS-CoV-2 from Guangdong CDC also was positive (Hospital Day 20) [fig_ref] Table 2: The Results of Nucleic Acid and Antibody Detection of SRSA-CoV-2 [/fig_ref]. Up to this point, taking into consideration of the epidemiological history, clinical features, imaging findings, the positive results of SARS-CoV-2 gene sequencing of Nanopore and the antibody of SARS-CoV-2, this patient was finally diagnosed for "SARS-CoV-2 pneumonia, mild case", although the routine RT-PCR detection kept to be negative all the time [fig_ref] Table 2: The Results of Nucleic Acid and Antibody Detection of SRSA-CoV-2 [/fig_ref]. The clinical symptoms of this patient were completely in remission on Hospital Day 24. The laboratory examinations were by and large normal . The chest CT showed that the inflammation effusion of lung was significantly absorbed compared to that of February 12 (Hospital Day 12). Therefore, the patient was discharged from hospital upon recovery on February 25 (Hospital Day 25; Illness
## Day 27).
A c c e p t e d M a n u s c r i p t 7
# Methods
## Specimen collection
Clinical specimens, including oropharyngeal and anal swab, serum, sputum, urine and stool were obtained in accordance with WHO and Chinese guidelines. Specimens were stored between 2°C to 8°C until ready for shipment to the Guangdong CDC and the lab of hospital A.
## Nucleic acid isolation and routine rt-pcr detection of sars-cov-2
Total RNA was extracted from 200μL specimen with automatic nucleic acid extractor
## Rna and gene detection for sars-cov-2
1) Method of Nanopore pathogen sequencing On the basis of previous research methods [bib_ref] Rapid Confirmation of the Zaire Ebola Virus in the Outbreak of the..., Mbala-Kingebeni [/bib_ref] , our team has developed a Platform of New Generation of Pathogenic Gene Sequencing (the fourth generation) -Nanopore Sequencing and Analysis Platform for this study. Furthermore, the Nanopore sequenced reads were aligned to the complete SARS-CoV-2 genome published on NCBI (NC_045512) [bib_ref] A novel coronavirus associated with a respiratory disease in Wuhan of Hubei..., Wu [/bib_ref].
# 2) bioinformatic methods
With the raw sequencing data, we performed data filtration with NanoFilt (version 1.7.0) [bib_ref] NanoPack: visualizing and processing long-read sequencing data, Coster [/bib_ref] , and specie annotation with Kraken [bib_ref] Improved metagenomic analysis with Kraken 2, Wood [/bib_ref]. Then, genome alignment (NCBI:
MN908947.3) was carried out with Minimap2 (version 2.17-r941) [bib_ref] Minimap2: pairwise alignment for nucleotide sequences, Li [/bib_ref] , and genome variations were called with bcftools (version 1.8) [bib_ref] BCFtools/RoH: a hidden Markov model approach for detecting autozygosity from next-generation sequencing..., Narasimhan [/bib_ref]. With SARS genome (NC_004718.and SARS-CoV-2 genomes (one from Guangzhou, China and the other one from A c c e p t e d M a n u s c r i p t 8 Washington, USA), SNPs were detected using Mummer (version: 3.23) [bib_ref] Versatile and open software for comparing large genomes, Kurtz [/bib_ref] , and the phylogenetic tree was constructed with Mega X (version 10.0.4) [bib_ref] Molecular Evolutionary Genetics Analysis across Computing Platforms, Kumar [/bib_ref].
## 3) rna detection and sanger sequencing
To identify the mutation of NP gene, the amplification of NP gene was performed by PCR with forward primer: 5'-GACCTACACAGGTGCCATCAA -3' and reverse primer: 5'-CCATCTGCCTTGTGTGGTCT -3'. The product of PCR was sequenced by Sangon Biotech (Shanghai, China). The gene sequence of Sanger sequencing was shown as supplementary materials (Supplementary [fig_ref] Table 2: The Results of Nucleic Acid and Antibody Detection of SRSA-CoV-2 [/fig_ref].
## Igm/igg antibody of sars-cov-2 was detected with elisa assay
Anti-Human IgM (μ-chain specific) antibody or N protein of SARS-CoV-2 (IgG) was used as the coating. The plasma of patients was diluted at 1:100 for testing. HRP labeled N protein of SARS-CoV-2 (IgM) or anti-human IgG (H + L) antibody labeled with HRP was used as the secondary antibody. The color was developed by TMB and terminated by H 2 SO 4 . Then OD450 was tested. The positive and negative control were set at the same time.
# Results
## The sequence detected by nanopore was sars-cov-2, and the nucleocapsid (np) and orf of sars-cov-2 have variation
The sputum specimens obtained from this patient on Hospital Day 8 and 12 were tested by Nanopore sequencing. A total of 242,889 reads were obtained for the samples [fig_ref] Figure 3: Length and quality distributions of the nanopore data and Taxonomic annotation results... [/fig_ref] , B), and the aligning ratio was 53.96% when they were mapped to SARS-CoV-2 genome. With specie annotation, virus was the dominant domain with a relative ratio of 60.20% [fig_ref] Figure 3: Length and quality distributions of the nanopore data and Taxonomic annotation results... [/fig_ref]. Moreover, 99.99% of the reads in virus domain were from SARS-CoV-2 [fig_ref] Figure 3: Length and quality distributions of the nanopore data and Taxonomic annotation results... [/fig_ref].
After Nanopore reads alignment, 172,62bp of SARS-CoV-2 genome were covered with a depth of 876.3 on average [fig_ref] Figure 4: Genome comparison and phylogenetic relationships between sample and SARS-CoV-2 [/fig_ref]. However, the alignment depth was not balanced [fig_ref] Figure 4: Genome comparison and phylogenetic relationships between sample and SARS-CoV-2 [/fig_ref] and [fig_ref] Figure 1: Timeline of the patient's clinical course [/fig_ref] , and 5' end of NP gene (29,380-29,533) was uncovered. In addition, the sample exhibited 6 single nucleotide polymorphism variations (SNPs, [fig_ref] Figure 4: Genome comparison and phylogenetic relationships between sample and SARS-CoV-2 [/fig_ref] , which were distributed in ORF1ab, ORF6, and ORF8 [fig_ref] Figure 4: Genome comparison and phylogenetic relationships between sample and SARS-CoV-2 [/fig_ref] , detailed in . After the construction of phylogenetic tree, the closer phylogenetic A c c e p t e d M a n u s c r i p t 9 relationship was discovered between the sample 1 and EPI_ISL_412967, which were both isolated from Guangzhou patients [fig_ref] Figure 4: Genome comparison and phylogenetic relationships between sample and SARS-CoV-2 [/fig_ref]. Because the 5' end of NP gene was uncovered, we used Sanger sequencing to identify the segment. PCR primers was designed to cover 29,380-29,533 region of SARS-CoV-2 genome and the amplified length was 490bp. This sequence only had 6% query cover with SARS-CoV-2, and it could not match any sequence in NCBI database [fig_ref] Figure 4: Genome comparison and phylogenetic relationships between sample and SARS-CoV-2 [/fig_ref].
## Igm and igg antibody turned from negative to positive, and micro-neutralization
## Antibody of sars-cov-2 was positive for this patient
The initial blood specimen (plasma) obtained from this patient on Hospital Day 8 was negative for the total IgM/IgG antibody of SARS-CoV-2, and the OD value for IgM and IgG [fig_ref] Figure 5: The IgM and IgG antibody of SARS-CoV-2 detected in this patient [/fig_ref].
# Discussion
Herein, we report a confirmed case of SARS-CoV-2 pneumonia in Guangzhou, China, which was finally confirmed by Nanopore sequencing and SARS-CoV-2 antibody detection combining with clinical features and chest CT, but negative by routine RT-PCR Furthermore, we found the newly discovered virus variation in NP and ORF of SARS-CoV-2, which may lead to the routine RT-PCR result negative.
To the situation, we think the best way is to efficiently identify the COVID-19 cases, and provide effective quarantine and clinical treatment to the patients. However, with only clinical features, such as fever and imaging features [fig_ref] Figure 2: Imaging of chest computed tomographic [/fig_ref] , it is difficult to differentiate COVID-19 from other viral infections. Although virus nucleic acid RT-PCR, CT imaging and some hematology parameters are adopted for clinical diagnosis of the infection A c c e p t e d M a n u s c r i p t 10 [bib_ref] A rapid advice guideline for the diagnosis and treatment of 2019 novel..., Jin [/bib_ref] , yet there proved to be a certain amount of false negative results [bib_ref] Correlation of Chest CT and RT-PCR Testing in Coronavirus Disease 2019 (COVID-19)..., Ai [/bib_ref] , especially for highly suspected cases as our case [fig_ref] Table 2: The Results of Nucleic Acid and Antibody Detection of SRSA-CoV-2 [/fig_ref] , which would make the epidemic worse.
Therefore, more effective methods should be applied for the clinical diagnosis. The Nanopore sequencing could be an efficient method, which has made nanotechnology achieve a practical breakthrough in single-molecule detection for the first time [bib_ref] Nanopore sequencing technology: nanopore preparations, Rhee [/bib_ref]. This technology has provided rapid detection of South America Zika virus, African Ebola virus and other new viruses [bib_ref] Nanopore Sequencing as a Rapidly Deployable Ebola Outbreak Tool, Hoenen [/bib_ref]. After COVID-19 outbreak, we have established a new Nanopore sequencing method of SARS-CoV-2 [fig_ref] Figure 3: Length and quality distributions of the nanopore data and Taxonomic annotation results... [/fig_ref] , and it was able to correctly provide positive test results of SARS-CoV-2 where the routine RT-PCR was negative [fig_ref] Table 2: The Results of Nucleic Acid and Antibody Detection of SRSA-CoV-2 [/fig_ref] In this study, we have discovered that most regions on SARS-CoV-2 genome could be covered [bib_ref] A novel coronavirus associated with a respiratory disease in Wuhan of Hubei..., Wu [/bib_ref] , which provided us with a sufficient method for the diagnosis of patients. In combination with Sanger sequencing [fig_ref] Figure 4: Genome comparison and phylogenetic relationships between sample and SARS-CoV-2 [/fig_ref] , the general genomic features and variations could be detected for the virus isolated from the patient. This virus exhibited 4 nonsynonymous mutations on gene ORF1ab and ORF8 [fig_ref] Figure 3: Length and quality distributions of the nanopore data and Taxonomic annotation results... [/fig_ref].
Since ORF1ab involved in the transcription and replication of viral RNA, the mutations could be a signal that the evolution of the virus is still underway. That said, it would cause increasing difficulties for the traditional detection methods of SARS-CoV-2 which mainly A c c e p t e d M a n u s c r i p t 11 adopt ORF1ab as the specified target region for the virus examination . In addition, the non-synonymous mutation on ORF8 (location: 28,144) has been recognized as an important virus typing mutation . According to previous reports, we knew that the virus of this patient could be the older SARS-CoV-2 strain (S typing). The phylogenetic analysis on this virus and previously published SARS-CoV-2 illustrated that its genome exhibited closer relationship with the virus isolated from the same region (Guangzhou, China), and was separated from the virus isolated from Wuhan region [fig_ref] Figure 4: Genome comparison and phylogenetic relationships between sample and SARS-CoV-2 [/fig_ref]. Therefore, the evolution process and the origin of SARS-CoV-2 still need to be explored. Furthermore, Sanger sequencing has also proved that the NP gene in this virus has variation [fig_ref] Figure 5: The IgM and IgG antibody of SARS-CoV-2 detected in this patient [/fig_ref] Supplementary .Since the examination of SARS-CoV-2 requires that the NP and ORF1ab genes of SARS-CoV-2 in the same specimen are both positive for RT-PCR [24], the gene variations in these genes might lead to the RT-PCR negative results for the patient.
Admittedly, there are some deficiencies in this paper, the biggest drawback is the lack of the virus culturing for this patient due to a low viral load and laboratory condition and qualification, and the unavailability of the full viral sequence. However, most part of virus genome was detected by Nanopore sequencing, and more samples from other patients in Guangzhou will be tested and reported in subsequent studies.
We report the clinical features, clinical diagnosis and virus mutation of a confirmed patient of SARS-CoV-2 pneumonia with RT-PCR negative in Guangzhou, China. The study of this case highlights that the Nanopore sequencing could be used in the clinical pathogenic diagnosis of SARS-CoV-2 pneumonia, especially when the virus mutation leads to routine RT-PCR negative. This report also demonstrates that we should be focusing on the evolution and variation of virus with epidemic development, which hold the possibility to lead to false negative test results and hence to increase the difficulty of epidemic prevention and control in practice. Finally, it suggests that multiple detection methods, including the antibody detection, should be used to make clinical diagnosis, especially when the routine RT-PCR results are found negative for highly suspected patients.
## Potential conflicts of interest. the authors report no conflicts of interest.
A c c e p t e d M a n u s c r i p t M a n u s c r i p t A c c e p t e d M a n u s c r i p t
[fig] February 5 to 24: Hospital Day 5 to 24), with samples including oropharyngeal swab, sputum, urine, stool, anal swab and whole blood test in hospital A and Guangdong CDC. [/fig]
[fig] 20190001 ,: GenAct NE-48, Shanghai GeneoDx Biotech Co., LTD, China) according to the manufacturer's instructions; a 50 μL elution volume was obtained for each sample. A 2μL aliquot of RNA was used for real-time RT-PCR, which targeted the ORF1ab and NP gene using a RT-PCR probe kit (GZ-D2RM, Shanghai GeneoDx Biotech Co., LTD, China). Realtime RT-PCR was performed under the following conditions: 42°C for 5 min and 95°C for 10s, followed by 40 cycles of amplification at 95°C for 10 s and 60°C for 45 s. Criteria for judging results: CT value<37 positive; 37≤CT value≤40 suspicious positive, and ≥40 negative. However, the positive should meet both ORF1ab and NP gene positive simultaneously. [/fig]
[fig] Furthermore: we designed new primers focusing on the non-mutation area of NP and ORF of SARS-CoV-2. The RT-PCR result was positive (Supplementary Table 4), which suggested that the variant virus was SARS-CoV-2. [/fig]
[fig] Figure 1: Timeline of the patient's clinical course. Symptom and treatment according to day of illness and day of hospitalization, January 20 to February 25, 2020. [/fig]
[fig] Figure 2: Imaging of chest computed tomographic (CT) scan. Panel A shows pulmonary and mediastinal window of chest CT on disease onset (Hospital Day 1; Illness Day 3). The two lungs were scattered with ground glass lesions, particularly in the dorsal/outer basal segments of the right lower lung (pulmonary window), suggesting a high possibility of viral pneumonia, and no abnormalities were noted in mediastinal window. The recheck was on Hospital Day 12, it showed ground-glass opacity and consolidation, and the lesion range was enlarged (Panel B). And the exudation of inflammation was well absorbed on Hospital Day 24 after treatment (Panel C). [/fig]
[fig] Figure 3: Length and quality distributions of the nanopore data and Taxonomic annotation results of sample. A. The distribution of read length. In this plot, x coordinate represents the read length, and the y coordinate represents the base number for the reads with different length. B. The distributions of read quality and read length. In this plot, each read was represented by a dot, and the x and y coordinates represent their length and average quality respectively. The upper and right-side histograms exhibited the distributions of reads length and quality respectively. C. The relative abundances of domains in the sample. D, E and F plots exhibited the compositions and relative abundances of Viruses, Eukaryota and Bacteria at the specie level respectively. [/fig]
[fig] Figure 4: Genome comparison and phylogenetic relationships between sample and SARS-CoV-2. A. Genome alignment between the consensus sequence from the sample and SARS-CoV-2 genome. In this plot, the red and blue dots represent the forward and backward alignment respectively, and the x and y coordinates represent the genome of SARS-CoV-2 and the sample respectively. B. The distributions of SNPs in sample as compared with SARS-CoV-2. Gene names are list on the x coordinate, and the corresponding SNP numbers are list on the y coordinate. The blue and red bars represent the synonymous and non-synonymous mutations respectively. C. Alignment of sequencing reads from sample on SARS-CoV-2 genome. The x and y coordinates represent the locations of SARS-CoV-2 genome and the aligned depth respectively. D. Phylogenetic relationships among the samples. In this phylogenetic tree, MN908947 was taken as the representative genome for SARS-CoV-2, and the SNPs in other genomes were detected using Mummer software. EPI_ISL_412967 and EPI_ISL_412970 were the published SARS-CoV-2 genomes from Guangzhou (China) and Washington (USA) patients, while NC_004718 was the representative genome for SARS. E. The 5' end of Nucleocapsid gene of SARS-CoV-2 had about 490bp mutation，proved by Sanger Sequencing. [/fig]
[fig] Figure 5: The IgM and IgG antibody of SARS-CoV-2 detected in this patient. The plasma obtained from patient on Hospital Day 8, 14 and 24. The level of IgM and IgG antibody were tested with ELISA method, and the results showed with OD value. [/fig]
[table] Table 2: The Results of Nucleic Acid and Antibody Detection of SRSA-CoV-2. [/table]
|
Safety, pharmacokinetics, and immunological activities of multiple intravenous or subcutaneous doses of an anti-HIV monoclonal antibody, VRC01, administered to HIV-uninfected adults: Results of a phase 1 randomized trial
# Ethical considerations
Multiple candidate HIV vaccines will need to be studied simultaneously in different populations around the world before a successful HIV preventive vaccine is found. It is critical that universally accepted ethical guidelines are followed at all sites involved in the conduct of these clinical trials. The HIV Vaccine Trials Network (HVTN) has addressed ethical concerns in the following ways:
HVTN trials are designed and conducted to enhance the knowledge base necessary to find a preventive vaccine, using methods that are scientifically rigorous and valid, and in accordance with Good Clinical Practice (GCP) guidelines.
HVTN scientists and operational staff incorporate the philosophies underlying major codes, declarations, and other guidance documents relevant to human subjects research into the design and conduct of HIV vaccine clinical trials.
HVTN scientists and operational staff are committed to substantive community input-into the planning, conduct, and follow-up of its research-to help ensure that locally appropriate cultural and linguistic needs of study populations are met. Community Advisory Boards (CAB) are required by DAIDS and supported at all HVTN research sites to ensure community input.
HVTN clinical trial staff counsel study participants routinely on how to reduce HIV risk. Participants who become HIV-infected during the trial are provided counseling on notifying their partners and about HIV infection according to local guidelines. Staff members will also counsel them about reducing their risk of transmitting HIV to others.
Participants who become HIV-infected during the trial are referred to medical practitioners to manage their HIV infection and to identify potential clinical trials they may want to join.
The HVTN provides training so that all participating sites similarly ensure fair participant selection, protect the privacy of research participants, and obtain meaningful informed consent. During the study, participants will have their wellbeing monitored, and to the fullest extent possible, their privacy protected. Participants may withdraw from the study at any time.
Prior to implementation, HVTN trials are rigorously reviewed by scientists who are not involved in the conduct of the trials under consideration.
HVTN trials are reviewed by local and national regulatory bodies and are conducted in compliance with all applicable national and local regulations.
The HVTN designs its research to minimize risk and maximize benefit to both study participants and their local communities. For example, HVTN protocols provide enhancement of participants' knowledge of HIV and HIV prevention, as well as counseling, guidance, and assistance with any social impacts that may result from research participation. HVTN protocols also include careful medical review of each research participant's health conditions and reactions to study products while in the study.
HVTN research aims to benefit local communities by directly addressing the health and HIV prevention needs of those communities and by strengthening the capacity of the communities through training, support, shared knowledge, and equipment. Researchers involved in HVTN trials are able to conduct other critical research in their local research settings.
The HVTN recognizes the importance of institutional review and values the role of in country Institutional Review Boards (IRBs) and Ethics Committees (ECs) as custodians responsible for ensuring the ethical conduct of research in each setting. This protocol minimizes risks to participants by (a) correctly and promptly informing participants about risks so that they can join in partnership with the researcher in recognizing and reporting harms; (b) respecting local/national blood draw limits; (c) performing direct observation of participants postinfusion and collecting information regarding side effects for several days postinfusion; (d) having staff properly trained in administering study procedures that may cause physical harm or psychological distress, such as blood draws, infusions, HIV testing and counseling and HIV risk reduction counseling; (e) providing HIV risk reduction counseling and checking on contraception use (for women); and (f) providing safety monitoring. [bib_ref] Infusion reactions: diagnosis, assessment, and management, Vogel [/bib_ref] In all public health research, the risk-benefit ratio may be difficult to assess because the benefits to a healthy participant are not as apparent as they would be in treatment protocols, where a study participant may be ill and may have exhausted all conventional treatment options. However, this protocol is designed to minimize the risks to participants while maximizing the potential value of the knowledge it is designed to generate.
## Irb/ec review considerations
## Reasonable risk/benefit balance
## Equitable subject selection 45 cfr 46.111 (a) 3 and 21 cfr 56.111 (a) 3: subject selection is equitable
This protocol has specific inclusion and exclusion criteria for investigators to follow in admitting participants into the protocol. Participants are selected because of these criteria and not because of positions of vulnerability or privilege. Investigators are required to maintain screening and enrollment logs to document volunteers who screened into and out of the protocol and for what reasons. [bib_ref] Infusion reactions: diagnosis, assessment, and management, Vogel [/bib_ref] The protocol specifies that informed consent must be obtained before any study procedures are initiated and assessed throughout the trial (see Section 9.1). Each site is provided training in informed consent by the HVTN as part of its entering the HVTN. The HVTN requires a signed consent document for documentation, in addition to chart notes or a consent checklist. This protocol has extensive safety monitoring in place (see . Safety is monitored daily by HVTN Core and routinely by the HVTN 104 Protocol Safety Review Team (PSRT). In addition, the HVTN Safety Monitoring Board (SMB) or a Data and Safety Monitoring Board (DSMB) periodically reviews study data. [bib_ref] Infusion reactions: diagnosis, assessment, and management, Vogel [/bib_ref] Privacy refers to an individual's right to be free from unauthorized or unreasonable intrusion into his/her private life and the right to control access to individually identifiable information about him/her. The term "privacy" concerns research participants or potential research participants as individuals whereas the term "confidentiality" is used to refer to the treatment of information about those individuals. This protocol respects the privacy of participants by informing them about who will have access to their personal information and study data (see Appendix A). The privacy of participants is protected by assigning unique identifiers in place of the participant's name on study data and specimens. In the United States, research participants in HVTN protocols are protected by a Certificate of Confidentiality from the US NIH, which can prevent disclosure of study participation even when that information is requested by subpoena. Participants are told of the use and limits of the certificate in the study consent form. In addition, each staff member at each study site in this protocol signs a Confidentiality Agreement with the HVTN and each study site participating in the protocol is required to have a standard operating procedure on how the staff members will protect the confidentiality of study participants.
## Appropriate informed consent
## Protect privacy/confidentiality
## Overview title
A phase 1 clinical trial to evaluate the safety and drug levels of a human monoclonal antibody, VRC-HIVMAB060-00-AB (VRC01) administered in multiple doses intravenously and subcutaneously in different dosing schedules to healthy, HIVuninfected adults
## Primary objectives
Primary objective 1:
To evaluate the safety and tolerability of VRC01, administered intravenously (IV) and subcutaneously (SC), at multiple timepoints.
## Primary objective 2 (groups 1-3):
To evaluate the serum levels of VRC01, administered IV and SC in 3 different regimens, at Month 6.
## Primary objective 3 (groups 4 and 5):
To evaluate the serum levels of VRC01 at 2 timepoints after each IV administration.
## Study products and routes of administration
VRC01: human monoclonal antibody (mAb) VRC-HIVMAB060-00-AB in formulation buffer at pH 5.8. Administered IV in 100 mL of normal saline (Sodium Chloride for Injection 0.9%, USP) or administered SC by needle and syringe injection.
SC placebo for VRC01: the placebo for VRC01 (VRC-PLAMAB068-00-AB) is a sterile, buffered aqueous solution of 25 mM Sodium Citrate, 50 mM Sodium Chloride, 150 mM L-Arginine Hydrochloride, 10% Dextran 40 (w/w), and 0.005% Polysorbate 80 (w/w) at pH 5.8 administered SC by needle and syringe injection.
IV placebo for VRC01: Sodium Chloride for Injection 0.9%, USP administered IV in 100 mL of normal saline (Sodium Chloride for Injection 0.9%, USP).
# Background
Although the global incidence of new HIV infections peaked in the mid-1990s, UNAIDS has reported that 2.5 million new HIV infections occurred last year, for a global total of more than 34 million people living with HIV. The wider availability of antiretroviral (ARV) therapy, mother-to-child transmission prevention programs, and a diverse array of other prevention programs have all contributed to turning the tide of the epidemic, but the magnitude of new infections remains a major concern, warranting the need for the development of safe and effective preventive vaccines.
The search for a globally effective HIV vaccine, as well as elucidation of biomarkers predictive of vaccine efficacy, continues. Modest success in preventing HIV acquisition has been achieved by the RV144 trial in Thailand with 31% efficacy [bib_ref] Vaccination with ALVAC and AIDSVAX to prevent HIV-1 infection in Thailand, Rerks-Ngarm [/bib_ref]. Neutralizing antibodies (nAbs) were not substantially elicited by this vaccine regimen and were not found to be a correlate of risk in this modestly effective vaccine regimen [bib_ref] Immune-correlates analysis of an HIV-1 vaccine efficacy trial, Haynes [/bib_ref]. However, most licensed vaccines elicit protective nAbs that correlate with vaccine efficacy. For HIV, it is possible that multiple immune mechanisms will be needed to prevent acquisition. In nonhuman primate (NHP) models, the presence of nAbs has been shown to prevent SHIV acquisition [bib_ref] Complete protection of neonatal rhesus macaques against oral exposure to pathogenic simian-human..., Ferrantelli [/bib_ref]. However, to date HIV vaccines have not been successful in generating nAbs effective against a wide variety of infecting strains [bib_ref] HIV vaccine design and the neutralizing antibody problem, Burton [/bib_ref] [bib_ref] Aiming to induce broadly reactive neutralizing antibody responses with HIV-1 vaccine candidates, Haynes [/bib_ref] [bib_ref] Neutralizing antibodies generated during natural HIV-1 infection: good news for an HIV-1..., Stamatatos [/bib_ref]. In recent years, research has made progress in the discovery of broad and potent nAbs found in the sera of chronically HIV infected donors [bib_ref] Broad HIV-1 neutralization mediated by CD4-binding site antibodies, Li [/bib_ref] [bib_ref] Human immunodeficiency virus type 1 elite neutralizers: individuals with broad and potent..., Simek [/bib_ref] [bib_ref] Rational design of envelope identifies broadly neutralizing human monoclonal antibodies to HIV-1, Wu [/bib_ref] [bib_ref] Structural basis for broad and potent neutralization of HIV-1 by antibody VRC01, Zhou [/bib_ref]. Knowledge gained from such discoveries holds promise for the development of new immunogens capable of eliciting broadly nAbs (bnAbs) at titers which could be potentially effective in protection against HIV-1 [bib_ref] Rational design of vaccines to elicit broadly neutralizing antibodies to HIV-1, Kwong [/bib_ref].
The Vaccine Research Center (VRC), NIAID, NIH has developed VRC01, a broadly neutralizing human mAbs which is targeted against the HIV-1 CD4 binding site [bib_ref] Rational design of envelope identifies broadly neutralizing human monoclonal antibodies to HIV-1, Wu [/bib_ref]. This mAb was originally discovered in a participant infected with HIV-1 for more than 15 years who maintained viral control without use of ARV therapy [bib_ref] Selection pressure on HIV-1 envelope by broadly neutralizing antibodies to the conserved..., Wu [/bib_ref]. By applying a novel method of isolating B cells that produce a specific antibody, and using recombinant DNA technology, the heavy and light chains encoding VRC01 were cloned and sequenced, allowing the synthetic production of codon-optimized genes encoding the variable region that was inserted into proprietary immunoglobulin G1 (IgG1) background sequences [bib_ref] Rational design of envelope identifies broadly neutralizing human monoclonal antibodies to HIV-1, Wu [/bib_ref]. In the interim since the isolation of the VRC01 antibody, subsequent work evaluating longitudinal serum collected from HIV-1 infected individuals has indicated that although antibodies capable of binding to VRC01-like epitopes may be induced during HIV-1 infection, they occur in only a minority of HIV-infected individuals and may take years to develop [bib_ref] The Development of CD4 Binding Site Antibodies During HIV-1 Infection, Lynch [/bib_ref].
The structure of VRC01 bound to HIV gp120 core has been determined. It binds to the HIV-1 gp120 envelope protein. VRC01 displays several unusual structural features. It is highly affinity-matured, has a disulfide link between complementarity-determining region (CDR) H1 and H3 and has a glycan in the variable (V) region of the light chain. However, none of these features appears to be required for binding affinity or neutralization [bib_ref] Structural basis for broad and potent neutralization of HIV-1 by antibody VRC01, Zhou [/bib_ref]. VRC01 does not have an unusually long CDR-H3 region like some other HIV-1 nAbs. It is not self-or poly-reactive and lacks anti-phospholipid antibody activity [bib_ref] Mechanism of neutralization by the broadly neutralizing HIV-1 monoclonal antibody VRC01, Li [/bib_ref] ; these features are consistent with the hypothesis that VRC01 will be safe for human administration.
By in vitro testing, VRC01 has a half-maximal inhibitory concentration (IC 50 ) of < 50 mcg/mL against 91% of primary isolates of various HIV-1 clades and < 1 mcg/mL against 72% of these isolates. Several proof-of concept studies have been conducted to determine whether the in vitro neutralization capabilities of VRC01 translate into the ability to protect NHPs from challenge with virulent chimeric simian-human immunodeficiency virus (SHIV), which contains the HIV envelope in an SIV background. Protection against a single high-dose SHIV-SF162P3 (a CCR5 tropic strain of HIV) rectal challenge was demonstrated at a 20 mg/kg dose level and partial protection at 5 mg/kg in Rhesus macaques. Protection against vaginal challenge was also demonstrated at 20 mg/kg dose level, given intravenously (IV) [NHP studies by J. . As tested in the SHIV model in infant macaques, passive transfer of IgG1b12 nAbs was protective, supporting the use of neutralizing antibodies in perinatal settings [bib_ref] Passive neutralizing antibody controls SHIV viremia and enhances B cell responses in..., Ng [/bib_ref]. In addition, VRC01 at 20 mg/kg protected infant macaques from an oral SHIV challenge [Nancy Haigwood, personal communication].
The clinical use of mAbs to prevent the establishment of viral infections has been previously demonstrated. Monthly injections of palivizumab, a mAb that blocks respiratory syncytial virus (RSV) binding to pulmonary epithelia were found to be safe and well-tolerated, and effective in protecting neonates and infants with underlying pulmonary disease from developing clinically significant RSV infection necessitating hospital admission [bib_ref] Humanized Respiratory Syncytial Virus Monoclonal Antibody, Reduces Hospitalization From Respiratory Syncytial Virus..., Palivizumab [/bib_ref].
## Rationale for trial concept
The rationale for HVTN 104 has evolved during the course of its development and implementation. Initially, in Version 1.0 with Groups 1-3, this trial was intended to collect safety and VRC01 trough level data from multiple doses of VRC01 administered over 6 months in dosing regimens hypothesized to result in drug levels consistently within the protective range, based on preclinical data. These goals are in support of potential future efforts in HIV prevention in infants (PMTCT) and adults. Subsequently, further clarity on the potential and distinct contributions that VRC01 may allow for the HIV vaccine field have come to the forefront and drive the modification of HVTN 104 to Version 2, with the inclusion of Groups 4 and 5 as well as the addition of more mucosal sampling timepoints. In these new groups, rather than maintain a consistently high VRC01 level, the goal is to allow the bnAb levels (i.e. neutralization titers) to drop to levels that are hypothesized to be within a range potentially inducible by a vaccine. For these groups, optimizing inter-subject variability between dosing intervals while maintaining VRC01 neutralization titers estimated within an IC50 <1mcg/ml (see Section 4.7) is the goal.
VRC01 drug levels and functional activity in serum and mucosal fluids will be evaluated by using multiple assays. Assessment of VRC01 drug levels in serum for the primary endpoint analysis will use an anti-idiotype assay specific for VRC01 (described in Section 10.4.1). Secondary endpoint analyses will assess VRC01 binding to Env antigens using the validated Binding Antibody Multiplex Assay (BAMA) (also described in Section 10.4.1). Additionally, we plan to assess the functional activity of VRC01 using the validated TZM-bl neutralization assay (described in Section 10.5.2).
Rationale for Version 1.0 (Groups 1-3):
VRC01 has successfully prevented SHIV acquisition in the NHP challenge protection model. The next step is testing the concept in humans. The VRC is preparing for a large scale international test-of-concept study in infants assessing prevention of mother-tochild transmission of HIV-1. The VRC is conducting initial first-in-human dose escalation studies for safety, tolerability and pharmacokinetic assessments in HIV infected (VRC 601) and HIV un-infected (VRC 602) adults. Upon confirmation of maximum dose acceptability, additional safety and pharmacokinetic (PK) data collection from multiple doses in a larger cohort of adults interrogating a range of dosing regimens is then needed to support future test of concept studies.
HVTN 104 will provide additional supportive data for planned and postulated uses of VRC01 for prevention of HIV acquisition. The primary goal is to further validate the safety and tolerability of VRC01 administered multiple times over 6 months. The data from HVTN 104 will help support the dosing schedule and rationale for a planned test of concept infant study; which will be a staged phase 1/2b, randomized, double-blind, placebo-controlled, multi-dose, international trial in breastfed infants born to HIV infected mothers. The first stage of the infant study will assess safety and tolerability and PK parameters of multiple dosing while the second stage will assess efficacy with a primary endpoint being the prevention of HIV transmission at 24 weeks of life. The preliminary dosing planned is monthly SC administration of 20mg/kg after an initial loading dose of 40mg/kg administered shortly after birth.
The NIH Vaccine Research Center's VRC 601 and 602 clinical trials will provide initial safety and pharmacokinetic data for 1 mg/kg, 5 mg/kg, 20mg/kg, and 40mg/kg IV doses as well as 5 mg/kg SC doses in a small number of adult participants. In adults, the SC dose evaluation is limited to 5 mg/kg because the volume of product that would be needed to be delivered for a dosage ≥ 20 mg/kg SC is too large to be accommodated by the SC route. HVTN 104 is being conducted in a larger cohort of adults and began when safety data from the maximum tolerated dose tested in VRC 602 was available. Group 1 in HVTN 104 is testing the same regimen proposed in the phase 1/2b infant study, but is being administered IV in adults. Groups 2 and 3 in HVTN 104 will also test alternate dosing regimens that may be considered for later studies in infants or adults.
Another goal of this study is to assess the VRC01 trough levels reached by administering different doses of VRC01 in different frequencies and routes. Based on the preclinical studies, achievement of trough serum levels of approximately 40 or 50 mcg/mL may be sufficient for HIV prevention. Assessing the kinetics of a higher dose administered less frequently (bimonthly) alongside a moderate dose administered more frequently (monthly) by IV is useful to determine if either or both schedules can achieve trough levels in this range. There are obvious logistical and operational advantages to less frequent dosing intervals if this can achieve trough levels hypothesized to provide protection, for any route method, but especially for IV. For this objective, we propose Group 1 to be dosed at 40 mg/kg IV at Month 0 and then 20 mg/kg at Months 1, 2, 3, 4, 5, and Group 2 to be dosed at 40 mg/kg at Months 0, 2, 4. Due to the expected long halflife for VRC01 in humans, a loading dose is anticipated to be required to allow trough levels to reach adequate levels earlier than without a loading dose. Therefore, a loading dose is planned for Groups 1 and 3. Since the maximum dose planned for testing in humans is 40mg/kg, a higher loading dose will not be administered for Group 2.
In addition, evaluation of the kinetics of a smaller dosage of VRC01 given SC at more frequent intervals merits exploration. Studies on the kinetics of IV and SC administration of IgG immunotherapy in patients with immunodeficiency disorders have compared high doses given IV every 3 weeks to smaller doses given SC weekly. IVIG dosing results in high peak serum levels (up to 1000 mg/dL) that fall rapidly over few days and with a further slow decline from catabolism. However, use of smaller SC doses allowing a slow diffusion of IgG into the vasculature and lymphatics results in stable higher trough IgG serum levels which remain constant between consecutive SC IG infusions [bib_ref] Subcutaneous immunoglobulin: opportunities and outlook, Misbah [/bib_ref]. Both routes of administration display similar half-lives. Trough levels, rather than peak levels, are of greatest importance, rather than peak, for sufficient prevention of HIV acquisition during passive immunotherapy, and thus smaller doses given more frequently via the SC route may have the potential to result in sufficient and consistent trough levels as compared to higher IV doses given less frequently. For this objective, we are proposing a third active arm of 5mg/kg SC biweekly (after a loading dose of 40mg/kg IV).
VRC01 is formulated at 100 ± 10 mg/mL. Subcutaneous dosing in infants is the expected route of administration and is feasible at the doses planned as the volume administered will be sufficiently small to be administered in 1 or 2 SC injections. Due to the concentration of the VRC01, larger volumes are required for adult dosing at the higher doses; therefore IV administration will be used for the higher dosing in these initial studies.
Considering long term that SC dosing can be potentially self-administered (as it is with IgG immunotherapy), then biweekly SC dosing may be more accessible, less burdensome, less expensive, and preferred by patients (as it is for many IgG immunotherapy patients) than monthly or even bimonthly IV dosing. Reactogenicity has not been a significant concern with SCIG [bib_ref] Subcutaneous immunoglobulin: opportunities and outlook, Misbah [/bib_ref].
## Rationale for version 2.0 (groups 4 and 5):
Major questions with respect to bnAbs and development of HIV vaccines include:
What is the range of protection afforded by a bnAb that blocks the binding of HIV gp120 to the cellular CD4 binding site? What is the dynamic range in concentration of antibody and neutralizing activity associated with protection? Can lower levels of neutralization activity, including levels potentially inducible by an HIV vaccine, afford protection or does in vivo protection require only high concentrations of antibodies to HIV CD4 binding site?
Are non-neutralizing effector functions also predictive of efficacy in addition to neutralizing activity? What are the kinetics and functional activities (nonneutralizing) that are seen at low neutralization titers for VRC01?
What is the PK profile of VRC01 in relation to different levels of neutralization over time? Do compartment-specific differences in blood and mucosal fluids affect the neutralizing and non-neutralizing functions of VRC01? These levels will allow us to model the HIV acquisition rates through a wide dynamic range of neutralization levels for VRC01.
To answer these questions, it will be necessary to evaluate bnAbs in later phase trials, evaluating protection and HIV acquisition endpoints at varying antibody levels in at-risk populations. Neutralizing antibodies to HIV have been shown to protect against experimental challenge in the NHP SHIV model [bib_ref] Protection of macaques against vaginal transmission of a pathogenic HIV-1/SIV chimeric virus..., Mascola [/bib_ref] [bib_ref] Protection of Macaques against pathogenic simian/human immunodeficiency virus 89.6PD by passive transfer..., Mascola [/bib_ref] [bib_ref] Defining the protective antibody response for HIV-1, Mascola [/bib_ref] [bib_ref] Neutralizing antibody directed against the HIV-1 envelope glycoprotein can completely block HIV-1/SIV..., Shibata [/bib_ref] [bib_ref] Human neutralizing monoclonal antibodies of the IgG1 subtype protect against mucosal simian-human..., Baba [/bib_ref] [bib_ref] Postnatal pre-and postexposure passive immunization strategies: protection of neonatal macaques against oral..., Hofmann-Lehmann [/bib_ref] [bib_ref] Determination of a statistically valid neutralization titer in plasma that confers protection..., Nishimura [/bib_ref]. In these studies the degree of protection has varied with the neutralizing potency of the antibodies and with the dose, route, and sensitivity of the challenge stocks. Low antibody concentrations have in some instances been quite protective, especially against repeat low dose mucosal challenges [bib_ref] Broadly neutralizing human anti-HIV antibody 2G12 is effective in protection against mucosal..., Hessell [/bib_ref] [bib_ref] Effective, low-titer antibody protection against low-dose repeated mucosal SHIV challenge in macaques, Hessell [/bib_ref]. Recent studies have suggested that VRC01 concentrations as low as ~1-2 mcg/ml can achieve sterilizing immunity (see Section 4.9.4 and [bib_ref] Highly potent HIV-specific antibody neutralization in vitro translates into effective protection against..., Moldt [/bib_ref] [bib_ref] Neutralizing antibodies to HIV-1 envelope protect more effectively in vivo than those..., Pegu [/bib_ref] [bib_ref] Enhanced neonatal Fc receptor function improves protection against primate SHIV infection, Ko [/bib_ref] [bib_ref] Enhanced potency of a broadly neutralizing HIV-1 antibody in vitro improves protection..., Rudicell [/bib_ref]. By adding Groups 4 and 5 to the HVTN 104 protocol, the study team will be able to assess the levels of VRC01 in blood and mucosal secretions when administered intravenously at 10 mg/kg or 30 mg/kg. Therefore, Groups 4 and 5 will inform the dose-selection for a phase 2b correlates trial aiming to evaluate the efficacy of a bnAb in high risk adults and to establish markers of passive immunoprophylaxis that correlate with protection against HIV infection. By using lower doses every 2 months (10 mg/kg IV in Group 4 and 30 mg/kg IV in Group 5), these newly added groups are evaluating dosing regimens that are expected to result in lower trough levels with large inter-subject variability between dosing intervals (see Section 4.7), necessary to accomplish these correlates goals.
## Mucosal specimen collection
Mucosal samples will be collected in order to assess the dynamics of transudation of VRC01 into mucosal secretions, where the protective effects of the mAb against founder virions would be expected to be most significant. Collection will include salivary and rectal secretions in men and women, the collection of semen in men and cervicovaginal secretions in women. The mucosal specimens will be collected on day 84, to assess early levels of VRC01 in the mucosal compartments, as well as at day 168 and day 224, which correspond to final trough and late timepoints. With Version 2.0, mucosal secretion sampling is being added at additional timepoints: 3, 14 and 21 days after the 2 nd or last VRC01 dose administration (the 21 day collection timepoint for all groups except Group 3). These additional samplings will allow for better potential semi-quantitative analysis of serum:mucosal surface VRC01 levels and intra-as well as inter-individual variability for mucosal expression. Additionally, functional and neutralizing Ab activity within the mucosal secretions may potentially be assessed at differing VRC01 levels.
## Anti-vrc01 antibodies
Production of anti-VRC01 antibodies may occur in some participants, yet are not expected to elicit hypersensitivity reactions. The package insert for Synagis ® (palivizumab), the licensed mAb product most analogous to the intended clinical use of the VRC mAb [bib_ref] Humanized Respiratory Syncytial Virus Monoclonal Antibody, Reduces Hospitalization From Respiratory Syncytial Virus..., Palivizumab [/bib_ref] , describes the incidence of anti-palivizumab antibody production following the fourth injection to be 1.1% in the placebo group and 0.7% in the active group in the single season clinical trial of 1502 children. In children receiving Synagis ® for a second season, one of the fifty-six children had transient, low titer reactivity. This reactivity was not associated with adverse events or alteration in serum concentrations. Immunogenicity of Synagis ® was also assessed in a trial involving another 379 children that compared liquid to lyophilized formulations and observed 0.3% incidence of antipalivizumab antibodies. These data reflect the percentage of children whose test results were considered positive for antibodies to palivizumab using an enzyme-linked immunosorbent assay (ELISA) which has substantial limitations in detecting antipalivizumab antibodies in the presence of palivizumab. Immunogenicity samples tested with the ELISA assay were likely to have contained palivizumab at levels that may have interfered with the detection of anti-palivizumab antibodies. An electrochemical luminescence (ECL) based immunogenicity assay, with a higher tolerance for palivizumab presence compared to the ELISA, was used to evaluate the presence of antipalivizumab antibodies in subject samples from two additional clinical trials. The rates of anti-palivizumab antibody positive results in these trials were 1.1% and 1.5%. One objective of HVTN 104 is to assess for the presence of anti-VRC01 antibodies and if present, determine if they interfere with VRC01 levels. An electrochemical luminescence assay rather than ELISA will be used for the detection of anti-VRC01 antibodies (see Section 10.5.1). Briefly, using polymerase chain reaction (PCR) amplification and cloning of the heavy and light chain variable region genes, a mAb was initially isolated from a single B cell from an HIV-1 infected subject who displayed broadly neutralizing antibodies. VRC01 is an IgG1 antibody and is highly somatically mutated from the germline precursor. The heavy chain CDR3 region is 14 amino acids long, which is an average length relative to natural antibodies and the glycosylation pattern is derived from its production in a CHO cell line.
## Vrc01: vrc-hivmab060-00-ab
The potency and breadth of neutralization by VRC01 compared with b12 (mAb derived from an HIV-1 clade B-infected donor) were assessed on a comprehensive panel of viral envelope pseudoviruses. The panel of 190 viral strains represented the major circulating HIV-1 genetic subtypes (clades) and included viruses derived from acute and chronic stages of HIV-1 infection. The ability of VRC01 and b12 mAbs to neutralize such pseudoviruses is reported on the neutralization dendrograms in -1 [bib_ref] Rational design of envelope identifies broadly neutralizing human monoclonal antibodies to HIV-1, Wu [/bib_ref].
VRC01 has an IC50 of <50 mcg/mL against 91% of primary HIV-1 isolates and IC50 <1 mcg/mL against 72% of HIV-1 isolates tested in vitro Of Clade B and Clade C isolates, 95% and 97%, respectively, are sensitive in vitro at a VRC01 concentration of 50 mcg/ml, 94% and 96% respectively at 10 mcg/ml, and 80% and 82% respectively at 1 mcg/ml [bib_ref] Rational design of envelope identifies broadly neutralizing human monoclonal antibodies to HIV-1, Wu [/bib_ref]. More details on VRC01 composition and manufacturing can be found in the IB. More details on the placebo composition and manufacturing can be found in the IB.
## Placebo for vrc01: vrc-plamab068-00-ab
## Sodium chloride placebo
Sodium Chloride for Injection 0.9%, USP administered IV in 100 mL of normal saline (Sodium Chloride for Injection 0.9%, USP).
## Trial design rationale
HVTN 104 will evaluate the safety profiles and trough levels of 5 different regimens for the IV and SC administration of VRC01. For adults, the advantage of the IV regimens is they allow administration of a larger volume of VRC01 to achieve the higher dosage levels that may be needed for prevention of HIV acquisition. However, an advantage of the SC regimens is the shorter infusion time, potential for reduced technical burden of a needle and syringe injection compared to IV product administration. This is important in a resource-constrained environment, as well as for product administration to infants. In the long term development of the product, a SC injection regimen allows potential for self-administration. Testing both IV and SC administered regimens of VRC01 is also important because each route may have particular advantages in different prevention indications. To prevent mother-to-child HIV transmission in breast feeding infants, repeated SC dosing might be beneficial to maintain consistent levels given frequent exposure to HIV, while for postexposure prophylaxis after a sexual assault in an adult, a single higher dose administration may be of most value.
The dose considerations for the evaluation of drug levels were informed primarily by prior studies with Synagis ® (palivizumab), which is the only FDA-licensed mAb for which the target is a viral pathogen, but consideration was also given to other investigational mAbs directed at pathogens and preclinical studies with VRC01. For prevention of severe RSV, infants are treated with 15 mg/kg IM every month throughout the RSV season [bib_ref] Humanized Respiratory Syncytial Virus Monoclonal Antibody, Reduces Hospitalization From Respiratory Syncytial Virus..., Palivizumab [/bib_ref]. In the NHP preclinical studies with VRC01, dosing at 20 mg/kg was associated with prevention of SHIV infection (see [fig_ref] Table 4: How the in vitro data correlates with in vivo VRC levels and... [/fig_ref] -3 and the IB). Other investigational mAbs directed at pathogens that have been safely taken into efficacy trials include a mAb directed at Clostridium difficile toxin administered at a 10 mg/kg dosage [bib_ref] Treatment with monoclonal antibodies against Clostridium difficile toxins, Lowy [/bib_ref] and a mAb directed at hepatitis C virus administered at a 50 mg/kg dosage [bib_ref] Human monoclonal antibody MBL-HCV1 delays HCV viral rebound following liver transplantation: a..., Chung [/bib_ref]. Thus, the dose escalation plan in the first Phase 1 study (VRC 601) starts at least 10-fold lower than typical dosages for other mAb directed at pathogens and increases in a 5-fold, 4-fold, 2-fold dosage increase plan to the 40 mg/kg dose level, which is twice the level associated with VRC01 prevention of SHIV in an NHP model. Importantly, with regard to the potential use of VRC01 to further reduce mother-to-child HIV transmission, the use of palivizumab for prevention of RSV in infants also provides a well-established record of dosage levels that are effective against a viral pathogen in infants and a model for safe evaluation of a mAb in an infant population.
In adult volunteers Synagis ® had a pharmacokinetic profile similar to a human IgG1 antibody with regard to the volume of distribution and the half-life (mean 18 days). In pediatric patients, monthly intramuscular doses of 15 mg/kg achieved mean ±SD 30 day trough serum drug concentrations of 37 ± 21 mcg/mL after the first injection, 57 ± 41 mcg/mL after the second injection, 68 ± 51 mcg/mL after the third injection and 72 ± 50 mcg/mL after the fourth injection [bib_ref] Humanized Respiratory Syncytial Virus Monoclonal Antibody, Reduces Hospitalization From Respiratory Syncytial Virus..., Palivizumab [/bib_ref]. Thus, the selected monthly IV administration of 20 mg/kg or bimonthly IV administration of 40 mg/kg of VRC01 for evaluation in human clinical trials was selected based on an expectation that this would maintain trough concentrations at levels hypothesized to be in the effective range in humans.
Because SC administration of VRC01 is the planned route for the future infant efficacy trials which will utilize a placebo group, the placebo for VRC01 was developed only for SC administration to match the viscosity of VRC01 to prevent unblinding. The SC arm in HVTN 104 will include a placebo control group to enhance the safety and tolerability evaluations of this route and of both of these products. Normal saline is the placebo for IV administration because the higher doses of IV infused VRC01 will be diluted in normal saline. In order to be as parsimonious as possible, for Group 3 the randomization will be 5:1, i.e. 20 participants receiving VRC01 and 4 participants receiving the comparable placebo for VRC01. Since VRC01 levels will only be present in VRC01 recipients, a placebo group is not necessary for the PK goals of this trial.
However, a placebo group for each of the IV arms in this study is not considered necessary or particularly informative for safety assessments of the IV administrations.
The VRC 601 and 602 trials (see Section 4.10) will provide the initial safety data in participants receiving IV infusions and this trial will serve to confirm those findings.
## Rationale for dose selections for groups 1-5:
Mathematical simulations were conducted to inform dosing regimen selection for Groups 1-5. These simulations used a standard two compartment PK model with parameter values estimated from the combined VRC 601 and VRC 602 data of 19 IV-infusion participants and 6 SC-injection participants. These simulations assumed linear pharmacokinetics and PK parameters proportional to body weight. These assumptions appear to be reasonable in the preliminary individual-subject PK modeling of the drug concentration data in VRC 601 and VRC 602. -2 shows the predicted median drug concentration for dosing regimens evaluated in Groups 1-3. -3 shows the predicted drug concentration levels (5 th , 50 th and 95 th percentiles) for Group 4 dosing regimen (10 mg/kg IV every 8 weeksno loading dose). These simulation results demonstrated that reasonable trough levels could be achieved in all five groups; the most sparse regimen with 10 mg/kg IV every 2 months (Group 4) has approximately 90% chance of reaching a trough level > 1 mcg/mL (Dr. Edmund Capparelli, personal communication). In support of Group 5, interim PK data collected after one or two infusions of 20 mg/kg IV from a small number of subjects in VRC 602 suggest that trough levels are maintained above 10 mcg/mL for more than 8 weeks (Dr. Barney Graham, personal communication and . At 30 mg/kg IV every 8 weeks, the trough level is anticipated to be intermediate compared to 20 mg/kg IV and 40 IV mg/kg every 8 weeks and spares product relative to the highest dose.
With respect to the loading dose considerations, Groups 1-3 have included the same loading dose of 40 mg/kg IV. This was done in order to allow the VRC01 drug levels to reach steady state more quickly to reflect application of a bnAb in a preventive setting. However, for Groups 4 and 5, in which the purpose is to understand the kinetics of drug concentrations and corresponding neutralization titers in the adult population, a slower achievement to a certain trough level is allowable and therefore no loading dose is included in Groups 4 or 5. In addition, the effect of a loading dose is abrogated in dosing regimens with the longer 8 week dosing interval (in Groups 2, 4 and 5). As noted in Section 4.4, VRC01 has an IC50 of <50 mcg/mL against 91% of primary HIV-1 isolates and IC50 <1 mcg/mL against 72% of HIV-1 isolates tested in vitro. Also, the median IC50 against viruses neutralized with an IC50 is 0.32 for all clades, but varies somewhat by clade and by study [fig_ref] Table 4: How the in vitro data correlates with in vivo VRC levels and... [/fig_ref] (www.bnaber.org).
## Dose and schedule
The proposed dosages for VRC 601 and 602 are based on the preclinical proof-ofconcept studies performed to date with VRC01, as well as publicly available data from human clinical trials experience with other mAb developed for use in prevention or treatment of human diseases, with attention to those directed at viral pathogens. In particular, the licensed product Synagis ® is directed against the viral pathogen, respiratory syncytial virus (RSV) [bib_ref] Humanized Respiratory Syncytial Virus Monoclonal Antibody, Reduces Hospitalization From Respiratory Syncytial Virus..., Palivizumab [/bib_ref] [bib_ref] Safety and tolerance of palivizumab administration in a large Northern Hemisphere trial...., Groothuis [/bib_ref] , and the method and schedule by which it is administered to infants provide a model for the product development plan.
Group 1 comprises an IV infusion of 40 mg/kg VRC01 administered in 100 mL of normal saline at week 0, followed by IV infusions of 20 mg/kg of VRC01 administered in 100 mL of normal saline over 1 hour at weeks 4, 8, 12, 16, and 20.
Group 2 comprises an IV infusion of 40 mg/kg VRC01 administered in 100 mL of normal saline at weeks 0, 8, and 16.
Group 3 comprises an IV infusion of 40 mg/kg VRC01 or sodium chloride placebo administered in 100 mL of normal saline at week 0, followed by SC injection of 5 mg/kg VRC01 or placebo for VRC01 administered by injection every 2 weeks for 20 weeks.
Group 4 comprises an IV infusion of 10 mg/kg VRC01 administered in 100 mL of normal saline at weeks 0, 8, and 16.
Group 5 comprises an IV infusion of 30 mg/kg VRC01 administered in 100 mL of normal saline at weeks 0, 8, and 16.
## Choice of control
The placebo for VRC01, VRC-PLAMAB068-00-AB, is comprised of components that are generally recognized as safe (GRAS), but the fully constituted placebo for VRC01 has not been tested in humans prior to the initiation of the VRC 602 study. This study agent is intended for use only as the placebo for SC evaluation of VRC01 and was developed to match the viscosity characteristics of VRC01 when formulated as a 100 mg/mL product.
The sodium chloride placebo used for the first IV infusion timepoint in the Group 3 placebo recipients will contain Sodium Chloride for injection 0.9%, USP serving as an inert control.
## Plans for future product development and testing
The data that will be generated from this study will help to establish the safety of VRC01 in HIV-uninfected men and women and will help to inform the decisions about appropriate dosing regimens for future studies of the use of VRC01 to prevent mother-tochild HIV transmission, design of future vaccine trials, or other immuno-prophylaxis indications. Ultimately, data from future passive immunization trials of highly selective mAbs will further inform future HIV vaccine development by determining if HIVspecific antibodies are capable of preventing HIV acquisition in humans, the levels or titers necessary for doing so and the quality of the antibodies that may be successful. In addition, research is ongoing to explore the design of increasingly specific immunogens strategically employed in a series of successive vaccinations in order to expeditiously mimic the development of broadly nAbs as observed in HIV chronic infection [bib_ref] HIV-1 antibodies from infection and vaccination: insights for guiding vaccine design, Bonsignori [/bib_ref].
The collection of simultaneous blood and anogenital secretion samples will advance the understanding of the diffusion of VRC01 between compartments, which may be relevant for selection of specific vaccine dosing regimens designed to generate broadly nAbs. It is conceivable that VRC01 could be part of a combined non-vaccine prevention modality (eg, antiretroviral prophylaxis) and vaccine approach, which would provide additional protection against HIV until an effective host immune response against HIV is generated. Finally, it is anticipated that there will be enough product for a phase 2b study in high risk/high incidence adult populations, and there is interest within the HVTN to participate in this endeavor.
## Preclinical studies
## Preclinical toxicology study of vrc01 in rats
A repeat dose toxicity study of IV and SC administration and a single dose pharmacokinetics (PK) study was performed by SRI International (Menlo Park, CA) with VRC01 in male and female Sprague-Dawley rats in accordance with U.S. FDA "Good Laboratory Practice (GLP) for Nonclinical Laboratory Studies." This study was conducted with a pre-GMP pilot lot of VRC01 manufactured at smaller scale using a similar purification process to that of the GMP clinical grade drug product.
For the safety assessment, various doses of VRC01 (4 mg/kg, 40 mg/kg, or 400 mg/kg) or a comparable vehicle was administered by tail vein infusion on Days 1 and 8 to Groups 1 through 4, respectively. An additional group (Group 5) received 40 mg/kg VRC01 via SC administration to the dorsal scapular region on Days 1 and 8. Each group contained 10 male and 10 female rats. Five animals per sex were sacrificed on Day 9, one day after the second administration, and the remaining animals were sacrificed on Day 30, 22 days after the second administration.
Results obtained showed that both routes of administration were well tolerated in the rats. All animals survived until their scheduled necropsy. No findings or changes were seen in clinical observation, body weight, food consumption, body temperature, infusion site irritation, hematology, coagulation, or organ weight evaluations that are attributed to administration of VRC01. VRC01 administration resulted in small, transient, dosedependent increases in aspartate aminotransferase (AST) and alkaline phosphatase (ALP) on Day 9. By Day 30, AST values had returned to normal, and ALP values were returning to normal.
Other than red discoloration of the administration site in one male in the SC group on Day 9, there were no other gross necropsy observations attributable to VRC01 administration. There were no histopathology findings that were considered related to IV administration of VRC01. However, histopathology evaluation revealed sub-acute inflammation at the SC injection site on Day 9, one day after infusion in all 10 rats administered VRC01 SC; dermal inflammation was usually minimal or mild while subcutaneous inflammation was usually mild, moderate, or marked. By Day 30, this inflammation had completely resolved, and the SC dose site was normal in all rats.
The pre-specified IV dose studied in rats was 400 mg/kg and SC was 40 mg/kg, which will greatly exceed the dose levels in the adult clinical studies. A "no observed effect level" (NOEL) was not determined in this study because transient elevations of AST and ALP were observed on Day 9 after IV administration and transient inflammation at the dose site was observed on Day 9 after SC administration. Because the elevated AST and ALP levels were transient and minor and did not correlate with histopathology findings, the no observed adverse effect level (NOAEL) for VRC01 by the IV route of administration in rats was 400 mg/kg, the highest dose used in this study. The systemic NOAEL for the SC route of administration of VRC01 in rats was 40 mg/kg, the only SC dose level examined in this study.
For the PK analysis, a separate cohort of rats received VRC01 on Day 1 at 4 mg/kg followed by 40 mg/kg by the IV route of administration and at 40 mg/kg by the SC route of administration. VRC01 levels in serum were determined using an enzyme-linked immunosorbent assay (ELISA) with samples collected predose from each animal and from an additional 3 males and 3 females to provide untreated control serum. Blood was collected from 3 rats/sex/PK group for a total of 4-5 collections per PK animal at each of the following postdose time points: After the peak concentration of VRC01 was achieved in the SC group, the serum levels decreased much more rapidly from 7 to 14 days than they did in the IV groups, and VRC01 concentrations in the SC group were not quantifiable at time points after 14 days. These data indicate that clearance of VRC01 in rats was markedly enhanced when it was administered by the SC route. The development of anti-drug antibodies that contribute to an increased rate of clearance is often observed in preclinical safety studies of proteinbased test articles when they are not tested in the species of origin. Although immunogenicity was not examined in this study, the presence of such antibodies might have possibly contributed to the increased rate of clearance of VRC01 after SC administration that was observed in this study [bib_ref] Immunogenicity of biologically-derived therapeutics: assessment and interpretation of nonclinical safety studies, Ponce [/bib_ref] [bib_ref] Antibody pharmacokinetics and pharmacodynamics, Lobo [/bib_ref].
## Tissue cross reactivity glp study of vrc01 with human tissues in vitro
A tissue cross-reactivity study of VRC01 using normal adult and neonatal human tissues in vitro [bib_ref] Postnatal pre-and postexposure passive immunization strategies: protection of neonatal macaques against oral..., Hofmann-Lehmann [/bib_ref]." In addition, the tissue cross-reactivity study used additional neonate/infant tissues suggested by the FDA to support future trials in infants.
To determine the cross-reactivity of VRC01 binding, VRC01 was applied to cryosections from a full panel of tissues from normal human adults and a limited panel of human neonatal tissues, immunohistochemically detected using a biotinylated rabbit anti-human IgG secondary antibody, and binding visualized with a streptavidin-horseradish peroxidase complex and a diaminobenzidine chromogen substrate. VRC01 binding was evaluated at concentrations of 5 and 50 mcg/mL. Specific VRC01 staining was not observed in any normal adult human or neonatal human tissues evaluated. Therefore, in vitro evaluation of cross-reactivity in tissue specimens did not identify potential tissue sites or organ systems to more thoroughly evaluate in subsequent preclinical studies, and it supports the possible future use of VRC01 in humans.
## Other toxicity studies
Several in vitro studies were conducted to assess for the demonstration of antibody activity against self-antigens by VRC01. Several anti-HIV neutralizing mAbs will crossreact to lipid or nuclear antigens or Hep-2 cells [bib_ref] Mechanism of neutralization by the broadly neutralizing HIV-1 monoclonal antibody VRC01, Li [/bib_ref] [bib_ref] Cardiolipin polyspecific autoreactivity in two broadly neutralizing HIV-1 antibodies, Haynes [/bib_ref]. Anti-lipid binding activity is understandable when considering that the HIV-1 gp41 protein is membrane-spanning and the epitopes (MPER: Membrane-Proximal External Region) recognized by some mAbs (e.g. 4E10 and 2F5) are membrane-proximal and likely extend into the membrane itself. Therefore, the ability (or lack thereof) of VRC01 to cross-react with lipids was assessed in collaboration with Dr. Barton Haynes of Duke University. Binding of antibody to cardiolipin was assessed in a luminescent assay, expressed in relative units. VRC01 was compared to 4E10, an anti-gp41 mAb known to bind to cardiolipin and nuclear antigens, and Synagis ® , a licensed anti-respiratory syncytial virus antibody used as a negative control. Synagis ® is included because it is the licensed mAb product most analogous to the intended clinical use of the VRC mAb [bib_ref] Humanized Respiratory Syncytial Virus Monoclonal Antibody, Reduces Hospitalization From Respiratory Syncytial Virus..., Palivizumab [/bib_ref].
Individual studies are summarized in [fig_ref] Table 4: How the in vitro data correlates with in vivo VRC levels and... [/fig_ref] -2. Unlike other anti-HIV neutralizing mAbs, VRC01 does not react to phospholipids or anti-nuclear antigens or Hep-2 cells. Additional details are provided in the IB. Several non-GLP studies of VRC01 have been completed in NHP to assess for plasma and secretion concentrations and for preclinical evidence of potential efficacy for prevention of HIV infection. [fig_ref] Table 4: How the in vitro data correlates with in vivo VRC levels and... [/fig_ref] -3 is a brief summary of the studies performed and supports the plan to evaluate up to 40 mg/kg dose administered IV as a dose range of potential interest for a preventive indication. The current assay being used to detect VRC01 in serum has a lower limit of detection in the range of 1.8-2 mcg/mL. Detectable concentrations of VRC01 were measured in vaginal, rectal, nasal and saliva samples after IV administration of 20 mg/kg of VRC01. Concentrations of VRC01 were lower in mucosal samples after IV administration of 5 mg/kg of VRC01. Please refer to the VRC01 IB for more details. SHIV BaL is a more neutralization sensitive strain used in the rectal challenge study and showed partial efficacy at dosing of 0.3 mg/kg and serum VRC01 levels of 1-2 mcg/ml. SHIV SF162P3 is a more neutralization resistant strain, and results of that study indicated that a higher dose and serum level of VRC01 is needed to completely protect NHP from a rectal challenge [fig_ref] Table 4: How the in vitro data correlates with in vivo VRC levels and... [/fig_ref]. While the in vitro neutralization data and the NHP challenge studies are intriguing, the applicability of these results remains yet to be determined in humans.
## Clinical studies
DAIDS is the sponsor of the investigational new drug application (IND) to evaluate the potential clinical uses of VRC01. VRC01's first evaluations in humans began in September 2013. The VRC01 initial development plan is directed towards an intended indication of prevention of HIV-1 infection through maternal transmission at birth or during breastfeeding. The safety, tolerability, dose effect, and PK of the VRC01 are being assessed in phase 1 trials in HIV-1 infected adults (in VRC 601) and in healthy HIVuninfected adults (in VRC 602) prior to evaluating the safety in infants in the US and internationally. Once the initial safety and PK data have been collected from these smaller dose escalation studies in adults, collection of additional safety data in larger numbers of adults from repeat doses will be used to support the infant studies. VRC 601 is the first study of the VRC01 mAb in HIV-infected participants. It is a doseescalation study to examine safety, tolerability, dose, PK, and anti-antibody immune responses. The hypothesis is that VRC01 will be safe for administration to HIV-1infected adults by the IV and SC routes and will not elicit hypersensitivity reactions. A secondary hypothesis is that VRC01 will be detectable in human sera and mucosal secretions with a definable half-life.
As shown in [fig_ref] Table 4: How the in vitro data correlates with in vivo VRC levels and... [/fig_ref] -5, there are 4 dose escalation groups for IV administration and 1 group for SC administration at 5 mg/kg planned for the study. During the dose escalation part of the study, each VRC01 infusion was administered in an inpatient unit and followed by intensive collection of samples for pharmacokinetic (PK) analysis. The study plan includes clinical blood tests for safety, clinician assessment of local reactions, solicitation of systemic symptoms for 3 days after each administration, HIV viral load, CD4, and blood samples to assess whether any human anti-VRC01 antibody is induced. In Groups 2, 3, 4 and 5, subjects who agree may have oral and rectal fluid samples collected and women may also have cervical fluid samples collected to determine if VRC01 is detectable in these mucosal samples. VRC 602 is the first study in healthy, HIV uninfected, adults of the VRC01 mAb. It is a dose-escalation study to examine safety, tolerability, dose, and pharmacokinetics of VRC01. The hypothesis is that VRC01 will be safe for administration to healthy adults by the IV and SC routes and will not elicit hypersensitivity reactions. A secondary hypothesis is that VRC01 will be detectable in human sera and mucosal secretions with a definable half-life. The SC route evaluation will be placebo-controlled and conducted in a single-blind manner to evaluate safety and tolerability of VRC01 and placebo (VRC-PLAMAB068-00-AB).
As shown in [fig_ref] Table 4: How the in vitro data correlates with in vivo VRC levels and... [/fig_ref] -6, there are 3 open-label, dose escalation groups (Groups 1, 2, and 3) for IV administration and 1 double-blinded, placebo-controlled group (Group 4) for SC administration. Enrollment started with subject randomization to Groups 1 and 4 in a 1:2 ratio. Within Group 4, subjects are randomized to SC infusions of VRC01 or placebo in a 1:1 ratio. After establishing the tolerability of the initial SC dose administered by a slow infusion with a controlled rate pump, the VRC 602 protocol was amended to allow the option of SC administration by needle and syringe injection with dividing the volume, as needed, into 2 or 3 SC injection sites. The study includes clinical blood tests for safety, clinical assessment of local reactions, solicitation of systemic symptoms for 3 days after each administration, and blood samples to assess whether any human anti-VRC01 antibody is induced. In all groups, when the subject agrees, oral and rectal fluid samples are obtained and women may also have cervical fluid samples collected to determine if VRC01 is detectable in these mucosal samples. The dose escalation to 40 mg/kg IV in the VRC 601 and VRC 602 studies was completed in a similar timeframe. As of January 9, 2015, cumulatively there have been 86 product administrations in 49 subjects in these two trials; including 21 subjects in VRC 601and 28 subjects in VRC 602. Of these 49 individuals, 9 subjects completed the one dose schedule, and of 40 subjects on a two dose schedule, 37 received two doses and 3 discontinued after one dose. Of the 3 subjects discontinued from study product administration, 1 was lost to follow-up, 1 was unable to comply with the study schedule and 1 experienced an unrelated adverse event (Streptococcal pharyngitis) that precluded timely administration.
Data from both HIV-infected and HIV-uninfected adults informs the questions of how to dose VRC01 to maintain a blood level in a range thought to be associated with a biological effect and to assess the safety of repeated exposure to VRC01. Both the IV and SC routes of administration are being evaluated. Ultimately, the best route for clinical use may depend upon the age of the recipient (adult or infant), stage of product development, formulation and important considerations related to volume needed and maintenance of a target VRC01 blood level considered to be in the therapeutic range.
## Current status of hvtn 104
The first participant enrolled in HVTN 104 on September 9, 2014. As of January 20, 2015, a total of 42 participants (27 in Groups 1 and 2 and 15 in Group 3) have been enrolled across 6 sites in the US. The study product has been well tolerated with most participants experiencing no or mild reactions and there have been no SAEs.
Two participants in Group 3 (VRC01 or placebo) had mild symptoms resulting in study product discontinuation. One participant had a mild itchy rash in several places a few days after the injection considered related to study product. The rash lasted a few hours and went away on its own without any other symptoms. The other participant experienced a tight feeling in the chest with wheezing upon auscultation that started shortly after the injection and went away on its own and was considered related to study product.
## Safety summary of vrc01
Cumulatively, across all studies as of January 9, 2015, there were no expedited safety reports to the FDA or study safety pauses for adverse events and, no reactions during the VRC01 or placebo product administration that resulted in an incomplete administration.
VRC01 SC administrations are sometimes associated with mild local reactions during the infusions, including pruritus (itchiness), redness and swelling, which resolve within a few minutes to a few hours after the administration is completed. The largest diameter for erythema or swelling events observed during infusions ranged up to about 5 cm.
Solicited local and systemic signs and symptoms following administration of VRC01 or placebo are generally none to mild. Less than 20% of subjects reported any moderate or greater solicited reactions after any product administration.
Adverse events attributed to study product administration for which study product administration was discontinued included one subject with chest discomfort and one subject with rash.
Other adverse events attributed to study product administration have included AST, ALT and creatinine elevation, decreased neutrophil count and pruritus at the administration site. The mild or moderate elevated transaminases were reported in about 20% of HIVinfected subjects (in association with ARVs, as well as, strenuous exercise and/or alcohol intake), and infrequently in HIV-uninfected subjects. These laboratory changes and pruritus events resolved and did not require discontinuation of study product administration.
Overall, VRC01 administration in the dose range from 1 to 40 mg/kg IV and at 5 mg/kg SC have been assessed as well-tolerated and safe for further evaluation.
## Future studies of vrc01
The data from HVTN 104 will help inform the dosing schedule and rationale for several studies including a phase 2b trial exploring the range of protection from HIV infection by a bnAb in high risk adults. In addition, a trial to assess the safety of VRC01 administered to infants born to HIV-infected mothers in the US is being planned by the IMPAACT network. Another study of VRC01 is being planned by USMHRP in Thai participants who are acutely-infected with HIV. An efficacy trial evaluating VRC01 in infants born to HIV-infected African women is planned assuming that early studies of VRC01 demonstrate an adequate safety profile.
## Potential risks of study products and administration
There is limited human experience with administration of the VRC01. VRC 601 and VRC 602 will evaluate the study agent prior to the start of HVTN 104.
In a preclinical study performed in rats, there was a small dose-dependent, but transient, increase in aspartate aminotransferase (AST) and alkaline phosphatase (ALP), but not in alanine aminotransferase (ALT) following IV administration. In rats, there were no histopathology findings following IV administration, but following SC administration there was minimal to marked dermal and SC irritation at the SC infusion site.
Thus far in VRC 601, there have been no SAEs or other safety concerns. Administration of mAb may have a risk of immune reactions such as acute anaphylaxis, serum sickness and the generation of antibodies, however, these reactions are rare and more often associated with mAb targeted to human proteins or with the use of murine mAbs which would have a risk of human anti-mouse antibodies [bib_ref] The safety and side effects of monoclonal antibodies, Hansel [/bib_ref]. In this regard, as VRC01 is targeted to a viral antigen and is a human mAb, it is expected to have a low risk of such side effects.
Typically, the side effects of mAbs are mild but may include fever, flushing, chills, rigors, nausea, vomiting, pain, headache, dizziness, shortness of breath, bronchospasm, hypotension, hypertension, pruritus, rash, urticaria, angioedema, diarrhea, tachycardia or chest pain. Clinical use of mAbs that are targeted to cytokines or antigens associated with human cells may be associated with an increased risk of infections [bib_ref] The safety and side effects of monoclonal antibodies, Hansel [/bib_ref] ; however, this is not expected to be a risk for a mAb targeted to a viral antigen.
It is known from published experience with human mAb directed against the cell surface targets on lymphocytes, that infusion of a mAb may be associated with cytokine release, causing a reaction known as "cytokine release syndrome" (CRS) [bib_ref] Monoclonal antibody-induced cytokine-release syndrome, Bugelski [/bib_ref]. Most infusionrelated events occur within the first 24 hours after beginning administration. Severe reactions; such as anaphylaxis, angioedema, bronchospasm, hypotension, and hypoxia, are infrequent and more often associated with mAbs targeted to human proteins or when a non-human mAb, such as a murine mAb, is used [bib_ref] The safety and side effects of monoclonal antibodies, Hansel [/bib_ref]. Specifically, with regard to CRS reactions, these most commonly occur within the first few hours of beginning the infusion and are more common with the first mAb infusion received. This is because the cytokine release is associated with lysis of the cells targeted by the mAb and the burden of target cells is greatest at the time of the first mAb treatment. With licensed therapeutic mAbs, CRS is managed by temporarily stopping the infusion, administration of histamine blockers and restarting the infusion at a slower rate [bib_ref] Infusion reactions: diagnosis, assessment, and management, Vogel [/bib_ref].
Delayed allergic reactions to a mAb may include a serum sickness type of reaction, which is characterized by urticaria, fever, lymph node enlargement, and joint pains. These symptoms may not appear until several days after the exposure to the mAb and is noted to be more common with chimeric types of mAb [bib_ref] The safety and side effects of monoclonal antibodies, Hansel [/bib_ref].
There are several FDA-licensed mAb for which reactions related to the rate of infusion have been described. Some symptoms may be treated by slowing or stopping the infusion. Supportive treatment may also be indicated for some signs and symptoms.
The published experience with mAb administered by the SC route is limited but there is experience with the SC route of administration of immunoglobulins, such as Hizentra® (CSL Behring LLC, Kankakee, IL), 20% immune serum globulin, for patients with primary immunodeficiency diseases. Comparison of the safety and PK of IVIG to subcutaneous immunoglobulin (SCIG) has been reported. The SC route of administration has a good safety profile. Tissue reactions are common but usually mild and tend to decline over time with repeated administrations [bib_ref] Home therapy with subcutaneous immunoglobulins for patients with primary immunodeficiency diseases, Haddad [/bib_ref]. The package insert for Hizentra ® notes that the most common adverse reactions (observed in ≥5% of study subjects) in clinical trials were local reactions (such as swelling, redness, heat, pain and itching at the infusion site), headache, vomiting, pain, and fatigue.
VRC01 is an antibody to an HIV protein. Therefore, it may be theoretically possible for a standard antibody-based HIV diagnostic test to detect VRC01 for a short time period postinfusion or postinjection. However, based on HVTN laboratory testing of HIV uninfected plasma samples spiked with VRC01 in the range of concentrations that mimic the likely range of concentrations (200 mcg/mL, 50 mcg/mL, and 1 mcg/mL) that may be observed in a clinical trial, VRC01 did not cause a positive test result in several standard antibody-based HIV-1/2 diagnostic tests used in the US.
Risks of Blood Drawing: Blood drawing may cause pain and bruising and may, infrequently, cause a feeling of lightheadedness or fainting. Rarely, it may cause infection at the site where the blood is taken. Problems from use of an IV for blood drawing are generally mild and may include pain, bruising, minor swelling or bleeding at the IV site and rarely, infection, vein irritation (called phlebitis), or blood clot.
## Risks of mucosal sample collection:
Collection of samples by swabs and wicks by rubbing them over the mucosal surfaces can cause momentary discomfort and, in some cases, minor bleeding.
## Risks of intravenous infusion:
The placement of an intravenous catheter can allow for the development of bacteremia because of the contact between the catheter and unsterile skin when it is inserted. This will be prevented through careful decontamination of local skin prior to catheter placement and through the use of infection control practices during infusion. Product contamination will be prevented by the use of aseptic technique in the pharmacy and universal precautions during product administration.
## Objectives and endpoints
## Primary objectives and endpoints
Primary objective 1:
To evaluate the safety and tolerability of VRC01 administered IV and SC at multiple timepoints.
## Primary endpoints 1:
Local and systemic reactogenicity signs and symptoms, laboratory measures of safety, and AEs and SAEs.
Early discontinuation of infusions and reason(s) for discontinuation and early study termination.
## Primary objective 2 (groups 1-3):
To evaluate the serum levels of VRC01 at Month 6 administered IV and SC in 3 different regimens.
## Primary endpoint 2:
Serum concentration of VRC01 in Groups 1-3 at Month 6.
## Primary objective 3 (groups 4 and 5):
To evaluate the serum levels of VRC01 at 2 timepoints after each IV administration.
## Primary endpoint 3:
Serum concentration of VRC01 28 and 56 days after each IV administration in Groups 4 and 5.
## Secondary objectives and endpoints
Secondary objective 1:
To evaluate the kinetics of in vitro neutralization in serum of a single VRC01 sensitive virus isolate
## Secondary endpoint 1:
Magnitude of serum neutralization of a single VRC01 sensitive virus isolate as measured in the TZMbl assay at multiple timepoints
## Secondary objective 2:
To further assess the serum levels of VRC01.
## Secondary endpoint 2:
Serum concentration of VRC01 in each group at multiple timepoints.
## Secondary objective 3:
To determine whether anti-idiotypic antibody (AIA) can be detected and whether there is a correlation of VRC01 levels and AIA levels in serum.
## Secondary endpoint 3:
Serum concentration of anti-VRC01 antibodies in each group at multiple timepoints compared to corresponding VRC01 levels.
## Secondary objective 4:
To determine if measurable levels of VRC01 can be found in genital, rectal, and oral secretions.
## Secondary endpoint 4:
Mucosal levels of VRC01 in each group at multiple timepoints.
## Secondary objective 5:
To evaluate the kinetics of in vitro neutralization in mucosal secretions of a single VRC01 sensitive virus isolate.
## Secondary endpoint 5:
Magnitude of neutralization in genital, rectal, and oral secretions of a single VRC01 sensitive virus isolate as measured in the TZMbl assay at multiple timepoints.
## Secondary objective 6:
To assess binding of VRC01 to multiple Env proteins.
## Secondary endpoint 6:
Binding antibody multiplex assay will be used to assess VRC01 binding in serum and genital, rectal, and oral secretions to multiple Env proteins in each group at multiple timepoints.
## Exploratory objectives
Exploratory objective 1:
To evaluate the breadth of in vitro neutralization in the serum and mucosal secretions if detected with the primary VRC01 sensitive isolate.
## Exploratory objective 2:
To compare the serum levels of VRC01 between groups at multiple timepoints
## Exploratory objective 3:
To evaluate additional functional humoral activities in serum and genital, rectal, and oral secretions.
## Exploratory objective 4:
To evaluate the effect of participant characteristics (i.e. sex and BMI) on serum levels of VRC01.
## Exploratory objective 5:
To describe the acceptability of study product administration on all study participants 6 Statistical considerations [bib_ref] Immune-correlates analysis of an HIV-1 vaccine efficacy trial, Haynes [/bib_ref]
## .1 accrual and sample size calculations
Recruitment will target enrolling 88 healthy, HIV-uninfected adult participants into five groups. Groups 1, 2, 4, and 5 entail IV administration of VRC01 on two different schedules with a randomization ratio of 1:1 for n=20 participants in Groups 1 and 2 and n=12 participants in Groups 4 and 5. Group 3 entails one initial IV administration followed by multiple SC administrations of VRC01 (n=20) or placebo (n=4) with a treatment:control randomization ratio of 5:1. Sites will be encouraged to enroll at least approximately 40% of each sex within Groups 1 and 2 combined, Groups 4 and 5 combined, and separately within Group 3. Groups 1 and 2 are randomized together, Groups 4 and 5 are randomized together. Group 3 undergoes a separate randomization process from the other groups because Group 3 participants will need to specifically agree to participate in a more intense dosing and visit schedule.
Since enrollment is concurrent with receiving the first complete or incomplete study infusion, all participants will provide some safety data. However, for PK analyses, it is possible that data may be missing for various reasons, such as participants terminating from the study early or problems in shipping specimens. For this reason, the sample size calculations in Section 6.1.2 account for 10% enrolled participants having missing data for the primary lab data endpoint.
## Sample size calculations for safety
The goal of the safety evaluation for HVTN 104 is to identify safety concerns associated with product administration. The ability of the study to detect SAEs can be expressed by the true event rate above which at least 1 SAE would likely be observed and the true event rate below which no events would likely be observed. Specifically, for an active arm of n=12, there is a 90% chance of observing at least 1 event if the true rate of such an event is 17.46% or more; and there is a 90% chance of observing no events if the true rate is 0.88% or less. For each active arm of n=20 , there is a 90% chance of observing at least 1 event if the true rate of such an event is 10.87% or more; and there is a 90% chance of observing no events if the true rate is 0.52% or less. For active IV administration arms combined (n=64), there is a 90% chance of observing at least 1 event if the true rate of such an event is 3.53% or more; and there is a 90% chance of observing no events if the true rate is 0.17% or less. For all active arms combined (n=84), there is a 90% chance of observing at least 1 event if the true rate of such an event is 2.70 % or more; and there is a 90% chance of observing no events if the true rate is 0.13 % or less. As a reference, in HVTN vaccine trials from December 2000 through December 2012, about 4% of participants who received placebos experienced an SAE.
Probabilities of observing 0, 1 or more, and 2 or more events among single arms or combined arms of sizes 12, 20, 64, and 84 are presented in-1 for a range of possible true AE or SAE rates. These calculations provide a more complete picture of the sensitivity of this study design to identify potential safety problems with the product. An alternative way of describing the statistical properties of the study design is in terms of the 95% confidence interval for the true rate of an adverse event based on the observed data.-2 shows the 2-sided 95% confidence intervals for the probability of an event based on a particular observed rate. Calculations are done using the score test method [bib_ref] Approximate is better than "exact" for interval estimation of binomial proportions, Agresti [/bib_ref]. If none of 12 participants in an active arm (groups 4 or 5) experience a safety event, the 95% 2-sided upper confidence bound for the true rate of such events in the total vaccinated population is 24.25%. If none of 20 participants in an active arm (Groups 1, 2, or 3) experience a safety event, the 95% 2-sided upper confidence bound for the true rate of such events is 16.11%. If none of the 64 combined participants in Groups 1, 2, 4, and 5 experience a safety event, the 95% 2-sided upper confidence bound for the true rate of such events is 5.66%. For the total 84 participants in the active arms, the 95% 2-sided upper confidence bound for this rate is 4.37%.
## Sample size calculations for drug levels of vrc01
The main goal of HVTN 104 regarding PK drug levels involves a preliminary estimation of the mean trough drug concentration after the administration series for each active arm. The precision with which the true concentration level can be estimated from the observed data depends on the standard deviation of measurements and the sample size. Two-sided 95% confidence intervals for the true mean drug level based on different observed average drug concentration levels in the active arms of size 12 and 20 are shown in-3, assuming a normal distribution for the log-transformed levels. The confidence intervals in the table are based on arm sizes of 10 and 18, which accounting for an assumed 10% loss of data from the original arms sizes. For example, if an average drug concentration of ln(15) and a standard deviation of 1.0 are observed based on the logtransformed drug levels among 18 participants, the 95% 2-sided upper confidence bound for the true mean drug concentration level is 24 mcg/mL. These calculations assume an approximate normal distribution for the mean of log-transformed drug concentration levels. The assumed mean and standard deviation are based on limited unpublished data from a study of a similar mAb built on the same isotype construct (personal communications). Since no individual-level data are available, first order Taylor expansions are used to approximate the mean and variance of the log-transformed data based on the reported mean and variance of the original scale data. Specifically, the natural log of the mean of the original scale data approximates the mean of the natural log-transformed data; the squared ratio between the variance and mean of the original scale data approximates the variance of the natural log-transformed data.
## Randomization
The randomization sequence will be obtained by computer-generated random numbers and provided to each HVTN CRS through the SDMC via a Web-based randomization system. Groups 1-3 will be enrolled simultaneously. With the implementation of Version 2.0, Groups 4 and 5 will enroll simultaneously. The randomization will be done in blocks to ensure balance across Groups 1 and 2, across Groups 4 and 5, and within Group 3. Because the schedule for Group 3 is more intensive, participants will be given the choice of being randomized in Group 3 or, separately, into one of Groups 1, 2, 4, or 5. At each institution, the pharmacist with primary responsibility for dispensing study products is charged with maintaining security of the treatment assignments.
## Blinding
Participants and site staff (except for site pharmacists) will be unblinded as to participant treatment assignments between Groups 1, 2, 4, and 5, but blinded as to treatment assignment (active vs. placebo) for Group 3. Study product assignments are accessible to those HVTN CRS pharmacists, DAIDS protocol pharmacists and contract monitors, and SDMC staff who are required to know this information in order to ensure proper trial conduct. Any discussion of study product assignment between pharmacy staff and any other HVTN CRS staff is prohibited. The HVTN SMB members also are unblinded to treatment assignment in order to conduct review of trial safety.
When a participant leaves the trial prior to study completion, the participant will be told he or she must wait until all participants are unblinded to learn his or her treatment assignment.
Emergency unblinding decisions will be made by the site investigator. If time permits, the HVTN 104 PSRT should be consulted before emergency unblinding occurs.
# Statistical analysis
This section describes the final study analysis, unblinded as to treatment arm assignment. All data from enrolled participants will be analyzed according to the initial randomization assignment regardless of how many infusions they received. The analysis is a modified intent-to-treat analysis in that individuals who are randomized but not enrolled do not contribute data and hence are excluded. Because of blinding and the brief length of time between randomization and enrollment-typically no more than 4 working days-very few such individuals are expected.
Analyses will be performed using SAS and R. Other software may be used to perform additional exploratory pharmacokinetics analyses..
No formal multiple comparison adjustments will be employed for multiple safety endpoints.
# Analysis variables
The analysis variables consist of baseline participant characteristic, safety, and laboratory measurements for primary-and secondary-objective analyses.
## Baseline comparability
Treatment arms will be compared for baseline participant characteristics using descriptive statistics.
# Safety/tolerability analysis
Since enrollment is concurrent with receiving the first infusion, all participants will have received at least 1 infusion and therefore will provide some safety data.
## Reactogenicity
The number and percentage of participants experiencing each type of reactogenicity sign or symptom will be tabulated by severity and treatment arm and the percentages displayed graphically by arm. For a given sign or symptom, each participant's reactogenicity will be counted once under the maximum severity for all infusion visits. In addition, to the individual types of events, the maximum severity of local pain or tenderness, induration or erythema, and of systemic symptoms will be calculated. Wilcoxon rank-sum tests will be used to test for differences in severity between Groups 1 and 2.
## Aes and saes
AEs will be summarized using MedDRA System Organ Class and preferred terms. Tables will show by treatment arm the number and percentage of participants experiencing an AE within a System Organ Class or within preferred term category by severity or by relationship to study product. For the calculations in these tables, a participant with multiple AEs within a category will be counted once under the maximum severity or the strongest recorded causal relationship to study product. Formal statistical testing comparing arms is not planned since interpretation of differences must rely heavily upon clinical judgment.
A listing of SAEs reported to the DAIDS Regulatory Support Center (RSC) Safety Office will provide details of the events including severity, relationship to study product, time between onset and last infusion, and number of infusions received.
## Local laboratory values
Boxplots of local laboratory values will be generated for baseline values and for values measured during the course of the study by treatment arm and visit. Each boxplot will show the first quartile, the median, and the third quartile. Outliers (values outside the boxplot) will also be plotted. If appropriate, horizontal lines representing boundaries for abnormal values will be plotted.
For each local laboratory measure, summary statistics will be presented by treatment arm and time point, as well as changes from baseline for postenrollment values. In addition, the number (percentage) of participants with local laboratory values recorded as meeting Grade 1 AE criteria or above as specified in the DAIDS AE Gradingwill be tabulated by treatment arm for each postinfusion time point. Reportable clinical laboratory abnormalities without an associated clinical diagnosis will be included in the tabulation of AEs described above.
## Reasons for discontinuation of study product administration and early study termination
The number and percentage of participants who discontinue study product administration and who terminate the study early will be tabulated by reason and treatment arm.
## Acceptability of study product or procedure
Acceptability of study product administration and injection procedures will be tabulated by reason and treatment arm.
## Analysis of antibody-level endpoints
## General approach
For the statistical analysis of endpoints, data from enrolled participants will be used according to the initial randomization assignment regardless of how many infusions they received. Additional analyses may be performed, limited to participants who received all scheduled infusions per protocol. Assay results that are unreliable, from specimens collected outside of the visit window, or from HIV-infected participants postinfection will be excluded. Since the exact date of HIV infection is unknown, any assay data from blood draws 4 weeks prior to an infected participant's last seronegative sample and thereafter may be excluded. If an HIV-infected participant does not have a seronegative sample postenrollment, then all data from that participant may be excluded from the analysis.
For continuous assay data (eg, serum concentration of VRC01), graphical and tabular summaries of the distributions by treatment arm and timepoint will be made. The difference of continuous assay data at the primary time-points of interest (Month 6) between groups will be tested with a nonparametric Wilcoxon rank sum test if the data are not normally distributed and with a 2-sample t-test if the data appear to be normally distributed. An appropriate data transformation (e.g., log transformation) may be applied prior to testing to better satisfy assumptions of symmetry and homoscedasticity (constant variance).Inference from these analyses would be limited by the small sample sizes of the groups.
More sophisticated analyses of drug-level data and other assay data collected over time employing repeated measures methodology (for example, nonlinear mixed effects models or generalized estimating equations) may be utilized to incorporate outcome responses over several timepoints and to account for subject heterogeneity. In addition, noncompartment PK analysis will be performed to estimate PK parameters from individual time-concentration curves. All statistical tests will be 2-sided and will be considered statistically significant if p 0.05.
For qualitative assay variables (e.g., positive or negative), the analyses will be performed by tabulating the frequency of positive responses for each assay by group at each timepoint at which an assessment is performed. Crude response rates will be presented with their corresponding 95% confidence interval estimates calculated using the score test method [bib_ref] Approximate is better than "exact" for interval estimation of binomial proportions, Agresti [/bib_ref].
For the analysis of correlation between two continuous assay variables over time, graphical summary and tabular summary of the sample correlation at each given timepoint will be made. Cross-correlation of the two variables with different time-lags may also be calculated and visually displayed if there are at least 10 participants with no missing data over time from both variables. More details of the statistical analysis approaches will be described in a separate Statistical Analysis Plan document.
Based upon previous AIDS Vaccine Evaluation Group (AVEG) and HVTN trials, missing 10% of research samples' results for a specific assay is common due to study participants terminating from the study early, problems in shipping specimens, or low cell viability of processed peripheral blood mononuclear cells (PBMCs). To achieve unbiased statistical estimation and inferences with nonparametric tests and generalized linear models fit by generalized estimating equation (GEE) methods, missing data need to be missing completely at random (MCAR). MCAR assumes that the probability of an observation being missing does not depend upon the observed responses or upon any unobserved covariates but may depend upon covariates included in the model (eg, missing more among whites than nonwhites). When missing data are minimal (specifically if no more than 20% of participants are missing any values), then nonparametric tests and GEE methods will be used, because violations of the MCAR assumption will have little impact on the estimates and hypothesis tests. These models will include as covariates all available baseline predictors of the missing outcomes.
If a substantial amount of antibody-level data are missing (at least 1 value missing from more than 20% of participants), then using the methods that require the MCAR assumption may give misleading results. In this situation, analyses of the endpoints at a specific timepoint will be performed using parametric generalized linear models fit by maximum likelihood. These methods provide unbiased estimation and inferences under the parametric modeling assumptions and the assumption that the missing data are missing at random (MAR). MAR assumes that the probability of an observation being missing may depend upon the observed responses and upon observed covariates, but not upon any unobserved factors. Generalized linear models for response rates will use a binomial error distribution and for quantitative endpoints, a normal error distribution. For assessing repeated immunogenicity measurement, linear mixed effects models will be used. If the immunological outcomes are left-and/or right-censored, then the linear mixed effects models of Hughes [bib_ref] Mixed effects models with censored data with application to HIV RNA levels, Hughes [/bib_ref] will be used, because they accommodate the censoring. In addition, secondary analyses of repeated immunogenicity measurements may be done using weighted GEE [bib_ref] Analysis of semi-parametric regression models with non-ignorable non-response, Rotnitzky [/bib_ref] methods, which are valid under MAR. All of the models described above will include as covariates all available baseline predictors of the missing outcomes.
## Analyses prior to end of scheduled follow-up visits
Any analyses conducted prior to the end of the scheduled follow-up visits should not compromise the integrity of the trial in terms of participant retention or safety or immunogenicity endpoint assessments. In particular, early unblinded analyses by treatment assignment require careful consideration and should be made available on a need to know basis only.
## Safety
During the course of the trial, unblinded analyses of safety data will be prepared approximately every 4 months during the main study, as defined in Section 3, for review by the SMB. Ad hoc safety reports may also be prepared for SMB review at the request of the HVTN 104 PSRT. The HVTN leadership must approve any other requests for unblinded safety data prior to the end of the scheduled follow-up visits.
## Anti-vrc01 and other laboratory assessments
For Group 3, an unblinded statistical analysis by treatment assignment of a primary laboratory endpoint may be performed when all participants have completed the corresponding visit and data are available for analysis from at least 80% of these participants. Similarly, an unblinded statistical analysis by treatment assignment of a secondary or exploratory endpoint may be performed when all participants have completed the corresponding visit and data are available for analysis from at least 80% of these participants. The Laboratory Program will review the analysis report prior to distribution to the protocol chairs, DAIDS, vaccine developer, and other key HVTN members and investigators. Distribution of reports will be limited to those with a need to know for the purpose of informing future trial-related decisions. The HVTN leadership must approve any other requests for HVTN laboratory data analyses prior to the end of the scheduled follow-up visits.
## Selection and withdrawal of participants
Participants will be healthy, HIV-uninfected (seronegative) adults who comprehend the purpose of the study and have provided written informed consent. Volunteers will be recruited and screened; those determined to be eligible, based on the inclusion and exclusion criteria, will be enrolled in the study. Final eligibility determination will depend on results of laboratory tests, medical history, physical examinations, and answers to self-administered and/or interview questions.
Investigators should always use good clinical judgment in considering a volunteer's overall fitness for trial participation. Some volunteers may not be appropriate for enrollment even if they meet all inclusion/exclusion criteria. Medical, psychiatric, occupational, or other conditions may make evaluation of safety and/or immunogenicity difficult, and some volunteers may be poor candidates for retention.
Determination of eligibility, taking into account all inclusion and exclusion criteria, must be made within 56 days prior to enrollment unless otherwise noted in sections 7.1 and 7.2. Negative urine glucose, and Negative or trace urine protein, and Negative or trace urine hemoglobin (if trace hemoglobin is present on dipstick, a microscopic urinalysis with red blood cells levels within institutional normal range).
## Inclusion criteria
## Reproductive status
## 21.
Volunteers who were born female: negative serum or urine beta human chorionic gonadotropin (β-HCG) pregnancy test performed prior to infusion on the day of initial infusion. Persons who are not of reproductive potential due to having undergone total hysterectomy with bilateral oophorectomy (verified by medical records), are not required to undergo pregnancy testing.
## Reproductive status: a volunteer who was born female must:
Agree to consistently use effective contraception (see Appendix A) for sexual activity that could lead to pregnancy from at least 21 days prior to enrollment through the last required protocol clinic visit. Effective contraception is defined as using any of the following methods: 10. Serious adverse reactions to vaccines or to vaccine components including history of anaphylaxis and related symptoms such as hives, respiratory difficulty, angioedema, and/or abdominal pain. (Not excluded: a volunteer who had a nonanaphylactic adverse reaction to pertussis vaccine as a child.) [bib_ref] Broad HIV-1 neutralization mediated by CD4-binding site antibodies, Li [/bib_ref]. Immunoglobulin received within 90 days before first infusion, unless eligibility for earlier enrollment is determined by the HVTN 104 PSRT.
[formula] [/formula]
12. Autoimmune disease (Not excluded: Volunteer with mild, stable and uncomplicated autoimmune disease that does not require immunosuppressive medication and that, in the judgment of the site investigator, is likely not subject to exacerbation and likely not to complicate reactogenicity and AE assessments) [bib_ref] Rational design of envelope identifies broadly neutralizing human monoclonal antibodies to HIV-1, Wu [/bib_ref]. Immunodeficiency
## Clinically significant medical conditions
## Untreated or incompletely treated syphilis infection
15. Clinically significant medical condition, physical examination findings, clinically significant abnormal laboratory results, or past medical history with clinically significant implications for current health. A clinically significant condition or process includes but is not limited to:
A process that would affect the immune response, A process that would require medication that affects the immune response,
Any contraindication to repeated infusions or blood draws, A condition that requires active medical intervention or monitoring to avert grave danger to the volunteer's health or well-being during the study period, A condition or process for which signs or symptoms could be confused with reactions to vaccine, or Any condition specifically listed among the exclusion criteria below. [bib_ref] Selection pressure on HIV-1 envelope by broadly neutralizing antibodies to the conserved..., Wu [/bib_ref]. Any medical, psychiatric, occupational, or other condition that, in the judgment of the investigator, would interfere with, or serve as a contraindication to, protocol adherence, assessment of safety or reactogenicity, or a volunteer's ability to give informed consent [bib_ref] The Development of CD4 Binding Site Antibodies During HIV-1 Infection, Lynch [/bib_ref]. Psychiatric condition that precludes compliance with the protocol. Specifically excluded are persons with psychoses within the past 3 years, ongoing risk for suicide, or history of suicide attempt or gesture within the past 3 years.
## Current anti-tuberculosis (tb) prophylaxis or therapy
19. Asthma other than mild, well-controlled asthma. (Symptoms of asthma severity as defined in the most recent National Asthma Education and Prevention Program (NAEPP) Expert Panel report).
Exclude a volunteer who:
Uses a short-acting rescue inhaler (typically a beta 2 agonist) daily, or Uses moderate/high dose inhaled corticosteroids, or In the past year has either of the following:
Greater than 1 exacerbation of symptoms treated with oral/parenteral corticosteroids;
Needed emergency care, urgent care, hospitalization, or intubation for asthma.
20. Diabetes mellitus type 1 or type 2, including cases controlled with diet alone. (Not excluded: history of isolated gestational diabetes.)
## Hypertension:
If a person has been found to have elevated blood pressure or hypertension during screening or previously, exclude for blood pressure that is not well controlled. Wellcontrolled blood pressure is defined as consistently ≤ 140 mm Hg systolic and ≤ 90 mm Hg diastolic, with or without medication, with only isolated, brief instances of higher readings, which must be ≤ 150 mm Hg systolic and ≤ 100 mm Hg diastolic. For these volunteers, blood pressure must be ≤ 140 mm Hg systolic and ≤ 90 mm Hg diastolic at enrollment.
If a person has NOT been found to have elevated blood pressure or hypertension during screening or previously, exclude for systolic blood pressure ≥ 150 mm Hg at enrollment or diastolic blood pressure ≥ 100 mm Hg at enrollment.
## 22.
Bleeding disorder diagnosed by a doctor (eg, factor deficiency, coagulopathy, or platelet disorder requiring special precautions) [bib_ref] Protection of macaques against vaginal transmission of a pathogenic HIV-1/SIV chimeric virus..., Mascola [/bib_ref]. Malignancy (Not excluded: Volunteer who has had malignancy excised surgically and who, in the investigator's estimation, has a reasonable assurance of sustained cure, or who is unlikely to experience recurrence of malignancy during the period of the study) [bib_ref] Protection of Macaques against pathogenic simian/human immunodeficiency virus 89.6PD by passive transfer..., Mascola [/bib_ref]. Seizure disorder: History of seizure(s) within past three years. Also exclude if volunteer has used medications in order to prevent or treat seizure(s) at any time within the past 3 years.
25. Asplenia: any condition resulting in the absence of a functional spleen 26. History of hereditary angioedema, acquired angioedema, or idiopathic angioedema.
## Participant departure from the study product administration schedule or withdrawal
This section concerns an individual participant's departure from the study product administration schedule. Pause rules for the trial as a whole are described in Section 11.3.
## Delaying study product administration for a participant
Under certain circumstances, a participant's scheduled study product administration will be delayed. The factors to be considered in such a decision include but are not limited to the following:
Within 10 days prior to any study product administration Prestudy product administration abnormal vital signs or clinical symptoms that may mask assessment of a study product reaction.
Intercurrent illness that is not expected to resolve prior to the next scheduled study product administration which is assessed by the site principal investigator (or designee) to require delay or withdrawal from the study product administration schedule. The investigator may consult the HVTN 104 PSRT.
Study product administration should not be administered outside the visit window period specified in the HVTN 104 Study Specific Procedures.
In order to avoid study product administration delays and missed study product administrations, participants who plan to receive live attenuated licensed vaccines other than influenza vaccine, or systemic glucocorticoids should be counseled to schedule receipt of these substances, when possible, outside the intervals indicated above. The effects of these substances on safety and immunogenicity assessments and their interactions with study products are unknown.
## Participant departure from study product administration schedule
Every effort should be made to follow the study product administration schedule per the protocol. If a participant misses a study product administration and the visit window period for the study product administration has passed, that study product administration cannot be given. The participant should be asked to continue study visits. The participant should resume the study product administration schedule with the next study product administration unless there are circumstances that require further delay or permanent discontinuation of study product administration (see Sections 7.3.1 and 7.3.3).
## Discontinuing study product administration for a participant
Under certain circumstances, an individual participant's study product administrations will be permanently discontinued. Specific events that will result in stopping a participant's study product administration schedule include:
Co-enrollment in a study with an investigational research agent (rare exceptions allowing for the continuation of study product administration may be granted with the unanimous consent of the HVTN 104 PSRT).
Clinically significant condition (ie, a condition that affects the immune system or for which continued study product administration and/or blood draws may pose additional risk), including but not limited to the following:
Pregnancy (regardless of outcome);
Any grade 4 local or systemic reactogenicity symptom, lab abnormality, or AE that is subsequently considered to be related to study product administration;
Any grade 3 lab abnormality or other clinical AE (exception: fever or vomiting and subjective local and systemic symptoms) that is subsequently considered to be related to study product administration; or Clinically significant type 1 hypersensitivity reaction associated with study product administration. Consultation with the HVTN 104 PSRT is required prior to subsequent study product administration following any type 1 hypersensitivity reaction associated with study product administrations; or Investigator determination in consultation with HVTN 104 Protocol Team leadership (eg, for repeated nonadherence to study staff instructions).
Such participants should be counseled on the importance of continuing with the study and strongly encouraged to participate in follow-up visits and protocol-related procedures per the protocol for the remainder of the trial, unless medically contraindicated.
In addition, study product administration will be stopped for participants diagnosed with HIV infection. HIV-infected participants will not continue in the trial after completing a 28 day safety follow-up period to the last prior study product administration, but will be monitored through other HVTN protocols (see Sections 7.3.4).
## Participant termination from the study
Under certain circumstances, an individual participant may be terminated from participation in this study. Specific events that will result in early termination include: Any condition where termination from the study is required by applicable regulations.
[formula] [/formula]
## Study product preparation and administration
CRS pharmacists should consult the Pharmacy Guidelines and Instructions for DAIDS Clinical Trials Networks for standard pharmacy operations. The protocol schema is shown in-1. See the IB for further information about study products.
## Study product regimen
The schedule of infusion is shown in Section 3 and additional information is given below.
## Group 4 (open-label)
Treatment 4 (T4): VRC-HIVMAB060-00-AB 10 mg/kg to be administered IV in 100 mL of Sodium Chloride for Injection USP, 0.9% at Months 0, 2, and 4.
## Group 5 (open-label)
Treatment 5 (T5): VRC-HIVMAB060-00-AB 30 mg/kg to be administered IV in 100 mL of Sodium Chloride for Injection USP, 0.9% at Months 0, 2, and 4. VRC01 is a highly concentrated protein solution and may develop white-to-translucent particles after thawing. These particles have been observed in approximately 1-3% of the vials and generally disappear over a few hours at room temperature. This particle formation has no effect on product quality. If these particles are observed after the minimum of one hour thaw time as described in Section 8.3, the vials should be held at room temperature for another 30 to 60 minutes to enhance the rate of particle dissolution. Product, free of particles, stored at room temperature, or 2-8° C, should be administered within 12 hours.
## Study product formulation
## Note: all calculations should be based on a concentration of vrc-hivmab060-00-ab of 100 mg/ml and a volume of 2 ml (equal to 200 mg) can be withdrawn from a vial.
The product label designates the long-term storage temperature as -35°C to -15°C. However, the Investigator's Brochure states that "Clinical site storage in a qualified, continuously monitored, temperature-controlled freezer with a temperature range of -45°C to -10°C is acceptable". The study products are described in further detail in the Investigator's Brochure (IB). The product label designates the long-term storage temperature as -35°C to -15°C. However, the Investigator's Brochure states that "Clinical site storage in a qualified, continuously monitored, temperature-controlled freezer with a temperature range of -45°C to -10°C is acceptable". The study products are described in further detail in the Investigator's Brochure (IB).
## Sc placebo for vrc01 [labeled as
## Iv placebo for vrc01 [labeled as sodium chloride for injection usp, 0.9%]
Sodium Chloride for Injection USP, 0.9% will be used as the IV Placebo for VRC01. It must be stored as directed by the manufacturer.
## Preparation of study products
A new prescription MUST be sent to the pharmacy after the participant is weighed on the day of each visit during which study product will be administered. The prescription MUST contain the subject's weight and dose level prior to being sent to the pharmacy (this may NOT be communicated verbally). If this information is NOT on the prescription, the prescription will be returned to the clinic from the pharmacy to be completed appropriately prior to the pharmacist beginning preparation of study product. To prepare an IV infusion, the pharmacist will verify the dose [total milligrams needed (40 mg x participant's weight in kg)] and thaw the minimum number of vials needed to obtain the full dose. The pharmacist will determine the number of vials by dividing the total milligrams needed by 200. The result will be rounded UP to the next whole number. Note: It is expected that each vial may be used for about 2 mL withdrawal volume (200 mg VRC01).
Each vial must be thawed for a minimum of 1 hour at room temperature after removal from the freezer. After the 1 hour minimum, the pharmacist should gently swirl the vials before inspecting for particles. DO NOT SHAKE VIAL. Once swirled, the pharmacist, using aseptic technique, will add the calculated total milligrams needed to a 100 mL bag of normal saline (Sodium Chloride for Injection USP, 0.9%). Up to 50 mL of VRC01 may be added to a 100 mL bag of normal saline.
The pharmacist must note on the IV bag label the dose calculated (in mg) for the participant (eg, 2800 mg for a 70kg individual) as well as the final total volume of fluid in the IV bag (eg, 131 mL for the 70 kg individual). The IV bag should be labeled with a 12-hour expiration date and time from the time the vial is removed from the freezer. The IV bag should be stored at room temperature or refrigerated (2 o -8 o C) until administration.
Note: The Sodium Chloride for Injection bags referred to as "100 mL bags" in the IV administration instructions will typically have 103 mL volume before any VRC01 is added and this is acceptable in the context of the instructions.
Any empty vials, unused portion of entered vials, or unused IV solution which contains study product should be discarded in a biohazard containment bag and incinerated or autoclaved in accordance with institutional or pharmacy policy.
## Vrc-hivmab060-00-ab (20mg/kg iv) -open label only (group 1 months 1 through 5)
To prepare an IV infusion, the pharmacist will verify the dose [total milligrams needed (20 mg x participant's weight in kg)] and thaw the minimum number of vials needed to obtain the full dose. The pharmacist will determine the number of vials by dividing the total milligrams needed by 200. The result will be rounded UP to the next whole number.
Note: It is expected that each vial may be used for about 2 mL withdrawal volume (200 mg VRC01).
Each vial must be thawed for a minimum of 1 hour at room temperature after removal from the freezer. After the 1 hour minimum, the pharmacist should gently swirl the vials before inspecting for particles. DO NOT SHAKE VIAL. Once swirled, the pharmacist, using aseptic technique will add the calculated total milligrams needed to a 100 mL bag of Sodium Chloride for Injection USP, 0.9%. Up to 50 mL of VRC01 may be added to a 100 mL bag of normal saline. Note: The Sodium Chloride for Injection bags referred to as "100 mL bags" in the IV administration instructions will typically have 103 mL volume before any VRC01 is added and this is acceptable in the context of the instructions.
Any empty vials, unused portion of entered vials, or unused IV solution which contains study product should be discarded in a biohazard containment bag and incinerated or autoclaved in accordance with institutional or pharmacy policy.
## Vrc-hivmab060-00-ab (40mg/kg iv) -double blind (group 3 t3)
To prepare an IV infusion, the pharmacist will verify the total dose [milligrams needed (40 mg x participant's weight in kg)] and thaw the minimum number of vials needed to obtain the full dose. The pharmacist will determine the number of vials by dividing the total milligrams needed by 200. The result will be rounded UP to the next whole number. Note: It is expected that each vial may be used for about 2 mL withdrawal volume (200 mg VRC01).
Each vial must be thawed for a minimum of 1 hour at room temperature after removal from the freezer. After the 1 hour minimum, the pharmacist should gently swirl the vials before inspecting for particles. After thawing, the vials should be gently swirled for 30 seconds to avoid foaming. DO NOT SHAKE VIAL. Once swirled, the pharmacist, using aseptic technique will add the calculated total milligrams needed to a 100 mL bag of normal saline (Sodium Chloride for Injection USP, 0.9%). Up to 50 mL of VRC01 may be added to a 100 mL bag of normal saline.
The pharmacist must note on the IV bag label the volume of study product added to the bag as well as the final total volume of fluid in the IV bag (eg, 28 mL VRC01 or Placebo Note: The Sodium Chloride for Injection bags referred to as "100 mL bags" in the IV administration instructions will typically have 103 mL volume before any VRC01 is added and this is acceptable in the context of the instructions.
Any empty vials, unused portion of entered vials, or unused IV solution which contains study product should be discarded in a biohazard containment bag and incinerated or autoclaved in accordance with institutional or pharmacy policy.
## Iv placebo for vrc01 -double blind (group 3 c3)
To prepare an IV infusion, the pharmacist will calculate the volume of Sodium Chloride for Injection USP, 0.9% needed by multiplying the participant's weight in kg by 0.4 mL. The pharmacist, using aseptic technique will add the calculated volume of Sodium Chloride for Injection USP, 0.9% needed to a 100 mL bag of normal saline (Sodium Chloride for Injection USP, 0.9%). Up to 50 mL of Sodium Chloride for Injection USP, 0.9% may be added to a 100 mL bag of normal saline.
The pharmacist must note on the IV bag label the volume of study product added to the bag as well as the final total volume of fluid in the IV bag (eg, 28 mL VRC01 or Placebo (IV) and Total Volume of 131 mL for the 70kg individual). The IV bag should be labeled with a 12-hour expiration date and time. The IV bag should be stored at room temperature or refrigerated (2 o -8 o C) until administration.
Note: The Sodium Chloride for Injection bags referred to as "100 mL bags" in the IV administration instructions will typically have 103 mL volume before any Sodium Chloride for Injection USP, 0.9% is added and this is acceptable in the context of the instructions.
Any unused IV solution which contains study product should be discarded in a biohazard containment bag and incinerated or autoclaved in accordance with institutional or pharmacy policy.
## Vrc-hivmab060-00-ab (5 mg/kg sc) double blind (group 3 t3)
To prepare a SC administration, the pharmacist will verify the total dose [milligrams needed (5 mg x participant's weight in kg)] and thaw the minimum number of vials needed to obtain the full dose. The pharmacist will determine the number of vials by dividing the total milligrams needed by 200. The result will be rounded UP to the next whole number. Note: It is expected that each vial may be used for about 2 mL withdrawal volume (200 mg VRC01).
Each vial must be thawed for a minimum of 1 hour at room temperature after removal from the freezer. After the 1 hour minimum, the pharmacist should gently swirl the vials before inspecting for particles. DO NOT SHAKE VIAL. Once swirled, the pharmacist, using aseptic technique will draw the calculated total milliliters needed into sterile syringe. If more than one syringe is needed, then the total dose should be divided as equally as possible between 2-4 syringes. The pharmacist will apply an overlay to each syringe.
The pharmacist must note on the label the final total volume of fluid in each syringe. The syringe(s) should be labeled with a 12-hour expiration date and time from the time the vial is removed from the freezer . The syringe should be stored at room temperature or refrigerated (2 o -8 o C) until administration.
Any empty vials, unused portion of entered vials, or expired prefilled syringes which contain study product should be discarded in a biohazard containment bag and incinerated or autoclaved in accordance with institutional or pharmacy policy.
## Sc placebo for vrc01 (vrc-plamab068-00-ab) double blind (group 3 c3)
To prepare a SC administration, the pharmacist will calculate the total volume needed (0.05 mL x participant's weight in kg) and thaw the minimum number of vials needed to obtain the full dose. Note: It is expected that each vial may be used for about 2 mL withdrawal volume.
Each vial must be thawed for a minimum of 1 hour at room temperature after removal from the freezer. After the 1 hour minimum, the pharmacist should gently swirl the vials before inspecting for particles. DO NOT SHAKE VIAL. Once swirled, the pharmacist, using aseptic technique will draw the calculated total milliliters needed into a sterile syringe. If more than one syringe is needed, then the total dose should be divided as equally as possible between 2-4 syringes. The pharmacist will apply an overlay to each syringe.
The pharmacist must note on the label the final total volume of fluid in each syringe. The syringe should be labeled with a 12-hour expiration date and time from the time the vial is removed from the freezer . The syringe should be stored at room temperature or refrigerated (2 o -8 o C) until administration.
Any empty vials, unused portion of entered vials, or expired prefilled syringes which contain study product should be discarded in a biohazard containment bag in accordance with institutional or pharmacy policy.
## Vrc-hivmab060-00-ab (10mg/kg iv) -open label only (group 4 months 0, 2, and 4)
To prepare an IV infusion, the pharmacist will verify the dose [total milligrams needed (10 mg x participant's weight in kg)] and thaw the minimum number of vials needed to obtain the full dose. The pharmacist will determine the number of vials by dividing the total milligrams needed by 200. The result will be rounded UP to the next whole number. Note: It is expected that each vial may be used for about 2 mL withdrawal volume (200 mg VRC01).
Each vial must be thawed for a minimum of 1 hour at room temperature after removal from the freezer. After the 1 hour minimum, the pharmacist should gently swirl the vials before inspecting for particles. DO NOT SHAKE VIAL. Once swirled, the pharmacist, using aseptic technique will add the calculated total milligrams needed to a 100 mL bag of Sodium Chloride for Injection USP, 0.9%. Up to 50 mL of VRC01 may be added to a 100 mL bag of normal saline. Note: The Sodium Chloride for Injection bags referred to as "100 mL bags" in the IV administration instructions will typically have 103 mL volume before any VRC01 is added and this is acceptable in the context of the instructions.
Any empty vials, unused portion of entered vials, or unused IV solution which contains study product should be discarded in a biohazard containment bag and incinerated or autoclaved in accordance with institutional or pharmacy policy.
## Vrc-hivmab060-00-ab (30mg/kg iv) -open label only (group 5 months 0, 2, and 4)
To prepare an IV infusion, the pharmacist will verify the dose [total milligrams needed (30 mg x participant's weight in kg)] and thaw the minimum number of vials needed to obtain the full dose. The pharmacist will determine the number of vials by dividing the total milligrams needed by 200. The result will be rounded UP to the next whole number. Note: It is expected that each vial may be used for about 2 mL withdrawal volume (200 mg VRC01).
Each vial must be thawed for a minimum of 1 hour at room temperature after removal from the freezer. After the 1 hour minimum, the pharmacist should gently swirl the vials before inspecting for particles. DO NOT SHAKE VIAL. Once swirled, the pharmacist, using aseptic technique will add the calculated total milligrams needed to a 100 mL bag of Sodium Chloride for Injection USP, 0.9%. Up to 50 mL of VRC01 may be added to a 100 mL bag of normal saline. Note: The Sodium Chloride for Injection bags referred to as "100 mL bags" in the IV administration instructions will typically have 103 mL volume before any VRC01 is added and this is acceptable in the context of the instructions.
Any empty vials, unused portion of entered vials, or unused IV solution which contains study product should be discarded in a biohazard containment bag and incinerated or autoclaved in accordance with institutional or pharmacy policy.
## Administration
For additional information regarding study product administration refer to the HVTN 104 Study Specific Procedures.
## Vrc01 (intravenously)
## For groups 1, 2, 4, and 5 (open label):
The IV bag prepared by the pharmacy will include the total amount (mg) of VRC01 added to the 100 mL normal saline bag and the final volume of the bag. The clinician responsible for administration and another clinician will each check the bag label and confirm that the identifier is correct and that the correct total milligrams to be administered is shown based on subject weight and dosage level before beginning the IV administration. The investigational study product solution will typically be administered IV over 60 minutes using a volumetric pump. The rate of infusion (mL/hr) will vary based on the total volume needed to administer the full dose. The total time needed to administer the dose may be longer based on factors such as subject tolerance.
## Vrc01 or iv placebo (intravenously)
For Group 3 Month 0 (blinded): The IV bag prepared by the pharmacy will include the total amount (mL) of VRC01 or IV Placebo for VRC01 (Sodium Chloride for Injection 0.9%, USP) added to the 100 mL normal saline bag and the final volume of the bag. The clinician responsible for administration and another clinician will each check the bag label and confirm that the identifier is correct and that the correct total milliliters to be administered is shown based on subject weight and dosage level before beginning the IV administration.
The investigational study product solution will typically be administered IV over 60 minutes using a volumetric pump. The rate of infusion (mL/hr) will vary based on the total volume needed to administer the full dose. The total time needed to administer the dose may be longer based on factors such as subject tolerance.
## Vrc01 or sc placebo (subcutaneously)
The SC syringe(s) prepared by the pharmacy will include the total amount (mL) of VRC01 or SC Placebo (VRC-PLAMAB068-00-AB). At the time of administration, the clinician and participant will again discuss the administration site(s) (abdomen, upper arm, or thigh) and deem the one selected to be acceptable.
If more than one syringe is needed, the ziplock bag containing the syringes will show the total volume to be administered. The clinician responsible for administration and another clinician will each check the syringe/ziplock bag label and confirm that the identifier is correct and that the correct total milliliters to be administered is shown based on the participant's weight and dosage level before administering the SC injection(s).
All SC injections must be at least 2 inches apart.
When preparing a dose in a syringe and administering the dose, consideration should be given to the volume of solution in the needle before and after the dose is administered. Particularly if the needle used to withdraw the product is replaced prior to product administration, consideration should be given to conserving the full dose of product. The pharmacy and clinic staff members are encouraged to work together to administer the dose specified in the protocol.
## Acquisition of study products
VRC-HIVMAB060-00-AB and VRC-PLAMAB068-00-AB are provided by the VRC/DAIDS/NIAID.
IV Placebo for VRC01 (Sodium Chloride for Injection USP, 0.9%) will not be provided through the protocol and must be obtained by the site.
Once an HVTN CRS is protocol registered, the pharmacist can obtain study products from the NIAID Clinical Research Products Management Center (CRPMC) by following the ordering procedures given in Pharmacy Guidelines and Instructions for DAIDS Clinical Trials Networks.
## Pharmacy records
The HVTN CRS pharmacist is required to maintain complete records of all study products. The pharmacist of record is responsible for maintaining randomization codes and randomization confirmation notices for each participant in a secure manner.
## Final disposition of study products
All unused study products must be returned to the CRPMC after the study is completed or terminated unless otherwise instructed by the CRPMC. The procedures and relevant form are included in the Pharmacy Guidelines and Instructions for DAIDS Clinical Trials Networks.
## Clinical procedures
The schedules of clinical procedures are shown in Appendix H and Appendix I and Appendix J.
## Informed consent
Informed consent is the process of ensuring that participants fully understand what will and may happen to them while participating in a research study. The HVTN informed consent form documents that a participanthas been informed about the potential risks, benefits, and alternatives to participation, and (2) is willing to participate in an HVTN study. Informed consent encompasses all written or verbal study information HVTN CRS staff provide to the participant, before and during the trial. HVTN CRS staff will obtain informed consent of participants according to HVTN policies and procedures.
The informed consent process continues throughout the study. Key study concepts should be reviewed periodically with the participant and the review should be documented. At each study visit, HVTN CRS staff should consider reviewing the procedures and requirements for that visit and for the remaining visits. Additionally, if any new information is learned that might affect the participants' decisions to stay in the trial, this information will be shared with trial participants. If necessary, participants will be asked to sign revised informed consent forms.
An HVTN CRS may employ recruitment efforts prior to the participant consenting. For example, some HVTN CRSs use a telephone script to prescreen people before they come to the clinic for a full screening visit. Participants must sign a screening or protocolspecific consent before any procedures are performed to determine eligibility. HVTN CRSs must submit recruitment and prescreening materials to IRB/EC and any applicable Regulatory Entity (RE) for human subjects protection review and approval.
Note: As defined in the DAIDS Protocol Registration Manual, an RE is "Any group other than the local IRB/EC responsible for reviewing and/or approving a clinical research protocol and site-specific ICFs prior to implementation at a site." CRSs are responsible for knowing the requirements of their applicable REs.
## Screening consent form
Without a general screening consent, screening for a specific study cannot take place until the site receives protocol registration from the DAIDS Protocol Registration Office's Regulatory Support Center (RSC).
Some HVTN CRSs have approval from their IRB/EC and any applicable RE to use a general screening consent form that allows screening for an unspecified HIV prevention trial. In this way, HVTN CRS staff can continually screen potential participants and, when needed, proceed quickly to obtain protocol-specific enrollment consent. Sites conducting general screening or prescreening approved by their IRB/EC and any applicable RE may use the results from this screening to determine eligibility for this protocol, provided the tests are conducted within the time periods specified in the eligibility criteria.
## Protocol-specific consent forms
The protocol-specific consent forms describe the study products to be used and all aspects of protocol participation, including screening and enrollment procedures. A sample protocol-specific consent form for the main study is located in Appendix A. A separate sample consent form for other uses of specimens is located in Appendix C.
Each HVTN CRS is responsible for developing a protocol-specific consent form(s) for local use, based on the sample protocol-specific consent forms in Appendix A and Appendix C. The consent form(s) must be developed in accordance with requirements of the following: Study sites are strongly encouraged to have their local CABs review their sites-specific consent forms. This review should include, but should not be limited to, issues of cultural competence, local language considerations, and the level of understandability.
The sample informed consent form includes instructions throughout the document for developing specific content.
Sites should follow the instructions in the Protocol-specific Official Memo distributed along with this protocol regarding when they may begin using their site-specific protocol consent forms.
Regarding protocol registration, sites should follow procedures outlined in the current version of the DAIDS Protocol Registration Manual.
## Assessment of understanding
Study staff are responsible for ensuring that participants fully understand the study before enrolling them. This process involves reviewing the informed consent form with the participant, allowing time for the participant to reflect on the procedures and issues presented, and answering all questions completely.
An Assessment of Understanding is used to document the participant's understanding of key concepts in this clinical trial. The participant must complete the Assessment of Understanding before enrollment. Staff may provide assistance in reading and understanding the questions and responses, if necessary. Participants must verbalize understanding of all questions answered incorrectly. This process and the participant's understanding of the key concepts should be recorded in source documentation at the site.
IRB/EC and any applicable RE may require that a participant has signed either a screening or protocol-specific consent document prior to administering the Assessment of Understanding. The consent process (including the use of the Assessment of Understanding) should be explained thoroughly to the IRB/EC and any applicable RE, whose recommendations should be followed.
## Pre-enrollment procedures
Screening may occur over the course of several contacts/visits, up to and including before infusion on day 0. All inclusion and exclusion criteria must be assessed within 56 days before enrollment, unless otherwise specified in the eligibility criteria (or below in this section).
After the appropriate informed consent has been obtained and before enrollment, the following procedures are performed: Discussion of pregnancy prevention. A pregnant or breastfeeding person may not be enrolled in this trial. Specific criteria and assessment of contraception and pregnancy status are described in study inclusion criteria. Discussion of pregnancy prevention includes advising a participant who was born female and who reports no current sexual activity that could lead to that participant becoming pregnant to have a plan to begin adequate birth control. This plan would be put to use if, during the study, the participant becomes sexually active in a way that could lead to that participant becoming pregnant.
[formula] [/formula]
## Use of screening results from another hvtn study
If a participant screens for an HVTN study at the same HVTN CRS but then does not join that study, screening results from that effort may be applied to the screening for this protocol, as long as the screening was done under participant consent, the participant has signed a consent form to begin screening for this study, and the tests were conducted within the time periods specified in the eligibility criteria (see Sections 7.1 and 7.2).
## Enrollment and study product administration visits
Enrollment is simultaneous with first infusion. The time interval between randomization and enrollment should not exceed 4 working days. The HVTN CRS registers the participant by scheduling the day 0 visit (enrollment) via the Web-based randomization system, and requests the randomization assignment. Circumstances may require a participant's enrollment visit to be changed. This may exceed the 4-day randomization time limit.
At all study product administration visits, the following procedures are performed before study product administration:
Abbreviated physical examination, including weight, vital signs, and a symptomdirected evaluation by history and/or appropriate physical exam based on participant self-reported symptoms or complaints;
Assessment of baseline reactogenicity parameters;
Assessment of concomitant medications (as described in Section 9.2);
Assessment of any new or unresolved AEs/intercurrent illnesses; and Urine or serum pregnancy test (for participants who were born female). Persons who are not of reproductive potential due to having undergone total hysterectomy with bilateral oophorectomy (verified by medical records), are not required to undergo pregnancy testing.
Following completion of all procedures in the preceding list and results indicate that study product administration may proceed, the study products are prepared and administered (see Sections 8.3 and 8.4).
Immediately following study product administration, the participant remains in the clinic for observation. An initial reactogenicity assessment is made at a target of 60 minutes after all IV infusions and the first SC injection (visit 3) in Group 3, with an acceptable range of 60-120 minutes. For subsequent SC injections in Group 3, an initial reactogenicity assessment is made at a target of 30 minutes, with an acceptable range of 25-60 minutes. Before leaving the clinic, the participant is given the postproduct symptom log and is instructed on how to complete it. The site will make arrangements to obtain daily reports of reactogenicity events from the participant during the reactogenicity period (as described in Section 9.9).
The following procedures will be performed at all study product administration visits. These procedures may be performed prior to or following study product administration:
Risk reduction counseling (as described in Section 9.6);
Pregnancy prevention assessment (as described in Section 9.2 and 9.7); and Assessment of new or unresolved social impacts (site staff will ask participant about the status of any unresolved social impacts and if s/he has experienced any new social impacts as a result of the trial participation).
The following procedure will be performed at all study product administration visits. This procedure must be performed following study product administration:
Administration of the acceptability questionnaire (level of discomfort with the procedures, willingness to undergo procedures in a "real world" setting; level of pain, and level of anxiety).
Additional procedures will be performed at scheduled visits as specified in Appendix H and Appendix I and Appendix J:
Confirm that participants received HIV test results from previous visit. If not, provide test results and post-test counseling as appropriate; and Specimen collection (to be completed prior to study product administration).
## Follow-up visits
The following procedures are performed at all scheduled follow-up visits:
Risk reduction counseling (as described in Section 9.6);
Pregnancy prevention assessment (as described in Section 9.2 and 9.7); and Assessment of new or unresolved social impacts (site staff will ask participant about the status of any unresolved social impacts and if s/he has experienced any new social impacts as a result of the trial participation);
Administration of the acceptability questionnaire (level of discomfort with the procedures, willingness to undergo procedures in a "real world" setting; level of pain, and level of anxiety); Assessment of new or continuing concomitant medications (as described in Section 9.2); and Assessment of new or unresolved AEs/intercurrent illnesses.
Additional procedures will be performed at scheduled follow-up visits as specified in Appendix H and Appendix I and Appendix J:
Administration of the social impact assessment questionnaire (types of impacts assessed involve personal relationships, health insurance, life insurance, educational or employment opportunities, housing, immigration, or travel);
Participants in Group 3 will also be asked whether they believe they received the active study product or the control;
HIV infection assessment including pretest counseling. A subsequent follow-up contact is conducted to provide posttest counseling and to report results to participant; Urine or serum pregnancy test (for participants who were born female); Persons who are not of reproductive potential due to having undergone total hysterectomy with bilateral oophorectomy (verified by medical records), are not required to undergo pregnancy testing.
## Mucosal secretion sampling
Mucosal secretion samples will be collected from all study participants who agree to these procedures at timepoints indicated Appendix H and Appendix I and Appendix J (see also HVTN 104 Study Specific Procedures). These samples include saliva, rectal secretions, semen (only for people born male), or cervical secretions (only for people born female).
Mucosal sampling must be performed prior to study product administration.
For participants born female providing cervical and/or rectal samples, a pregnancy test must be performed and must be negative prior to any mucosal sampling. Persons who are not of reproductive potential due to having undergone total hysterectomy with bilateral oophorectomy (verified by medical records), are not required to undergo pregnancy testing.
The following conditions apply to mucosal sampling. Study participants are advised of these in the study informed consent form (see Appendix A):
Salivary fluid sampling: Participants willing to provide salivary samples should abstain from smoking, eating, or drinking anything other than water, brushing their teeth, using mouthwash, chewing gum or tobacco, or engaging in intimate oral activity for one hour prior to sample collection.
Rectal fluid sampling: Participants who consent to provide rectal samples are subject to the following exclusion criteria:
Participant is pregnant. For participants born female, a pregnancy test must be performed and be negative prior to any rectal mucosal sampling.
Participant has engaged in receptive anal intercourse or insertion of any foreign object or substance into the anus for 48 hours prior to sample collection. Foreign objects include, but are not limited to cleaning products (creams, gels, lotions, pads, etc.), lubricant, enemas, and douching, even with water.
Participant has used any perianal or intra-anal steroid or other anti-inflammatory creams for 48 hours prior to sample collection.
Participant has any active infection or inflammation of the colorectal area (such as an HSV-2 outbreak or inflamed hemorrhoids or colitis/diarrhea).
Participant has any active Genital Tract Infections (GTI), including untreated syphilis. If participant has received treatment for syphilis, documentation of resolution must be obtained prior to sampling. Semen sampling: Participants born male who consent to provide semen samples are requested to refrain from ejaculation for at least 48 hours prior to specimen collection. In addition, participants are subject to the following exclusion criteria:
Participant has any active Genital Tract Infections (GTI), including untreated syphilis. If participant has received treatment for syphilis, documentation of resolution must be obtained prior to sampling.
## Hiv counseling and testing
HIV counseling will be performed in compliance with the CDC's guidelines or other local guidelines for HIV counseling and referral. HIV testing will be performed in accordance with the current HVTN HIV testing algorithm following enrollment.
Participants will be counseled routinely during the trial on the avoidance of HIV infection and on the potential negative social impacts of the theoretical possibility of testing antibody positive due to the study product. They will also be counseled on the risks of HIV antibody testing outside of the HVTN CRSs and will be discouraged from doing so during study participation and/or during any period of study product reactive serology.
Study staff will take particular care to inform study participants of the likelihood of routine HIV testing being offered or performed outside the study CRS at emergency rooms, clinics, and medical offices. Such testing has become more likely due to the CDC's revised guidelines for HIV counseling and testing, as well as policy changes in many countries to make HIV testing more frequent and routine. CRS staff should inform participants of their right to opt out of HIV testing outside the study site. CRS staff should inform study participants if local and/or state policies and regulations permit medical providers to perform HIV testing without first informing patients. If this is the case, then CRS staff should advise study participants that they may decline testing preemptively. CRS staff should also inform participants if positive results must be reported to local public health authorities. CRS staff should also inform participants of the need to maintain study blinding by getting HIV testing only at the study CRS. CRS staff should provide participants with CRS contact information and should encourage participants to ask medical providers to contact the CRS. The CRS can verify that the participant is in an HIV vaccine clinical trial and should only be tested at the study CRS.
Potential participants identified as being HIV infected during screening are not enrolled. All participants who become HIV infected during the study will be terminated from this study. Potential and enrolled participants identified as HIV infected will be referred for medical treatment, counseling, and management of the HIV infection. These individuals may also be referred to appropriate ongoing clinical trials or observational studies.
## Distinguishing intercurrent hiv infection from study product reactive serology
The following procedures will be conducted in order to assess study product reactive serology:
Participants will have physical examinations at visits specified in Appendix H and Appendix I and Appendix J. Signs or symptoms of an acute HIV infection syndrome, an intercurrent illness consistent with HIV-1 infection, or probable HIV exposure would prompt a diagnostic workup per the HVTN algorithm for Recent Exposure/Acute Infection Testing to determine HIV infection.
HIV testing will be performed at multiple timepoints throughout the study (see Appendix E and Appendix F and Appendix G). The Laboratory Program (or approved diagnostic laboratory) will follow the HVTN HIV testing algorithm (as described in the HVTN Site Lab Reference Manual), which is able to distinguish study product reactive serology responses from actual HIV infections.
All participants can receive HIV-1 diagnostic testing from the site following their last scheduled visit until they are told that they did not receive VRC01 or that antibodybased HIV diagnostic tests are not reactive.
All participants who received VRC01 and who have VRC01-induced positive or indeterminate HIV-1 serology (as measured by the standard anti-HIV antibody screening tests) at or after the study is unblinded will be offered poststudy HIV-1 diagnostic testing (per the HVTN poststudy HIV-1 testing algorithm) periodically and free of charge as medically/socially indicated (approximately every 6 months).
## Contraception status
Contraception status is assessed and documented at every scheduled clinic visit for a participant who was born female and who is sexually active in a way that could cause that participant to become pregnant. Prior to enrollment and throughout the study, staff will ask participants to verbally confirm their use of adequate contraceptive methods. A participant who was born female and is sexually active in a way that could cause that participant to become pregnant should be reminded at all scheduled clinic visits of the importance of using contraception and should be referred to specific counseling, information, and advice as needed. (Specific contraception requirements are listed in Section 7.1). This reminder should be documented in the participant's study record.
Self-reported infertility-including having reached menopause (no menses for 1 year) or having undergone hysterectomy, bilateral oophorectomy, or tubal ligation-must be documented in the participant's study record.
## Urinalysis
Dipstick testing may be performed in the clinic or the lab, as long as the required elements (glucose, protein, and hemoglobin) are tested. The examination is performed on urine obtained by clean catch.
If the screening dipstick is transiently abnormal due to menses or infection, document this issue in the participant's source documentation. For infection, provide appropriate treatment and/or referral. Following resolution, repeat the dipstick and, if within the eligibility limits specified in the protocol, the participant may be enrolled.
Follow-up urinalysis should be deferred if a participant is menstruating, but should be performed as soon as possible. If a follow-up dipstick is abnormal due to a participant's menstrual period, document in the comment section of the case report form (CRF) and repeat the dipstick once the participant is no longer menstruating; a micro-urinalysis is not required.
## Assessments of reactogenicity
For all participants, baseline assessments are performed before and reactogenicity assessments are performed after each study product administration. All reactogenicity symptoms are followed until resolution and graded according to the Division of AIDS The reactogenicity assessment period is 3 full days following each study product administration per the assessment schedule shown in-1. Participants are instructed to record symptoms using a post product symptom log and to contact the site daily during the assessment period. Clinic staff will follow new or unresolved reactogenicity symptoms present at day 3 to resolution. Participants are instructed to contact the clinic for events that arise during the period between study product administration and the next scheduled visit. In general, a participant who self-reports any post administration reaction greater than mild is seen by a clinician within 48 hours after onset, unless the reaction is improving and/or has completely resolved.
Reactogenicity events are reported using CRFs that correspond to the time of assessment in-1. Reactogenicity assessments include assessments of systemic and local symptoms, and study product-related lesions. Events not listed on a CRF, or with an onset after the reactogenicity assessment period (day of study product administration and 3 full days after), or those meeting SAE/adverse events requiring expedited reporting to DAIDS criteria, are recorded on an adverse experience log form. Between 12:00am and 11:59pm day 3 HVTN CRS staff or participant a Day of study product administration b New or unresolved reactogenicity symptoms present on day 3 are followed until resolution
## Assessment of systemic and local symptoms
Typical study product administration site reactions are erythema/induration/swelling/edema. The maximum horizontal and maximum vertical measurements for all study product administrationsite reactions are recorded.
Body temperature is measured by oral or infrared thermometry and reported in degrees Celsius. If temperature is measured in Fahrenheit, the conversion to Celsius should be documented in the participant's chart note. A measurement is taken once daily during the assessment period and should be repeated if participant is feeling feverish.
## Assessment of the study product administration site
Typical study product administration site reactions are erythema/induration/swelling/edema. The maximum horizontal and maximum vertical measurements for all study product administration site reactions are recorded.
All study product administration site reactions are monitored until resolution. Areas greater than 25 cm 2 are followed daily; otherwise, the frequency of follow-up is based on clinician judgment.
## Visit windows and missed visits
Visit windows are defined in HVTN 104 Study Specific Procedures. For a visit not performed within the window period, a Missed Visit form is completed. If the missed visit is one that required safety assessments or local safety labs, HVTN CRS staff should attempt to bring the participant in for an interim visit as soon as possible.
Procedures performed at an interim visit are usually toxicity/safety assessments (including local safety labs) and HIV testing. With the exception of HIV testing, these procedures are performed only if they were required at the missed visit or if clinically indicated. HIV testing may be performed as deemed appropriate by the study staff. Blood samples for immunogenicity assays are not typically collected at interim visits.
If a missed visit required study product administration, please refer to Section 7.3.2 and Section 7.3.3 for resolution.
## Early termination visit
In the event of early participant termination, site staff should consider if the following assessments are appropriate: a final physical examination, clinical laboratory tests (including urine dipstick, CBC with differential, platelet count, and chemistry panel), pregnancy testing, social impact assessment, and HIV test.
## Pregnancy
If a participant becomes pregnant during the course of the study, no more administrations of study product will be given, but remaining visits and study procedures should be completed unless medically contraindicated or applicable regulations require termination from the study. In case of required termination, enrollment in an observational study should be offered to the participant. If the participant terminates from the study prior to the pregnancy outcome, the site should make every effort to keep in touch with the participant in order to ascertain the pregnancy outcome.
## Laboratory
## Hvtn crs laboratory procedures
The HVTN Site Lab Reference Manual provides further guidelines for operational issues concerning the clinical and processing laboratories. The manual includes guidelines for general specimen collection, special considerations for phlebotomy, specimen labeling, whole blood processing, HIV screening/diagnostic testing, and general screening and safety testing.
Tube types for blood collection are specified in Appendix E and Appendix F and Appendix G. For tests performed locally, the local lab may assign appropriate tube types.
In specific situations, the blood collection tubes may be redirected to another laboratory or may require study-specific processing techniques. In these cases, laboratory special instructions will be posted on the protocol-specific section of the HVTN website.
## Total blood volume
Required blood volumes per visit are shown in Appendix E and Appendix F and Appendix G. The FHCRC laboratory will further specify the tube type and collection volumes in special instructions posted to the protocol-specific section of the HVTN website.) Not shown is any additional blood volume that would be required if a safety lab needs to be repeated, or if a serum pregnancy test needs to be performed. The additional blood volume would likely be minimal. The total blood volume drawn for each participant will not exceed 500 mL in any 56-day (8-week) period.
## Primary timepoint
The primary timepoint in this study is at visit 19 (day 168) (ie, 4 weeks after the 6 th study product administration) for Group 1; visit 19 (day 168) (ie, 8 weeks after the 3 rd study product administration) for Groups 2, 4, and 5; and visit 19 (day 168) (2 weeks after the 12 th study product administration) for Group 3. Other time points may also be examined to better understand how serum concentrations of VRC01 change. The schedules are shown in Appendix E and Appendix F and Appendix G.
## Drug levels
## Vrc01 mab levels
VRC01 levels will be measured in serum and mucosal secretions including saliva, semen, cervical secretions, and rectal secretions. An ELISA will be used to determine the concentration of the VRC01 antibody in the serum. The ELISA employs the VRC01 Fabspecific 5C9 mAb, which is an anti-idiotype antibody cloned from a single B cell that was sorted by flow cytometry using a VRC01 scFv probe. ELISA will also be used to detect the presence of VRC01 antibody in various mucosal secretions. The 4-parameter logistic curve regression of a standard curve of VRC01 covering the range from 0.031 to 1.0 mcg/mL is utilized in this assay to quantitate the sample concentrations based upon the average of sample dilutions within the range of the assay. This assay has been qualified but not formally validated. The functional sensitivity for the generation of the ELISA assay format, which is currently used at NVITAL, is 2 mcg/ml; and as the technology for this assay continues to develop, an updated assay may be utilized
Binding multiplex antibody assay (BAMA) may also be used to assess binding of VRC01 from serum and mucosal secretions to various Env proteins, including Consensus gp120, Consensus gp140 and CD4 binding site proteins. The lower limit of detection will be determined for the matched lot of VRC01 study product prior to assessment of VRC01 binding in HVTN 104 participants. However, historical control assays carried out under GCLP validated assay conditions indicate that the lower limit of detection for VRC01 is at least 1-10 ng/ml.
## Endpoint assays: humoral
## Anti-vrc01 antibody assay
Assessment for development of anti-VRC01 antibodies in subjects will be performed using the Forte Bio Octet BioLayer Interferometry (BLI) technology. The assay uses VRC01 immobilized to a biosensor. The biosensor is dipped into patient serum samples and antibodies against VRC01 are directly measured. The binding response is directly proportional to the anti-drug concentration as determined against a calibration curve using the 5C9 antibody.
## Neutralizing antibody assay
Depending upon the concentrations measured in collected specimens, further evaluation of the research samples to assess for functional capacity to neutralize HIV in blood and mucosal secretions may be evaluated by an in vitro cell-based virus neutralization assay [bib_ref] Measuring HIV neutralization in a luciferase reporter gene assay, Montefiori [/bib_ref] [bib_ref] Development and implementation of an international proficiency testing program for a neutralizing..., Todd [/bib_ref] [bib_ref] International technology transfer of a GCLP-compliant HIV-1 neutralizing antibody assay for human..., Ozaki [/bib_ref] using pseudotyped viruses.
One or more viruses that are among the most sensitive to VRC01 (e.g. MN.3 and MW965.26) will be assayed. The IC50 of VRC01 against both of these viruses is 0.01 -0.03 mcg/ml and the IC90 of VRC01 against both of these viruses is 0.2 mcg/ml. The TZM-bl assay is validated for this level of sensitivity.
## Genotyping
Molecular human leukocyte antigen (HLA) typing may be performed on enrolled participants using cryopreserved PBMC collected at baseline. Other participants (including control recipients) may be HLA-typed to support future studies of immunological interest at the discretion of the HVTN Laboratory Program. Other markers, such as genes associated with immune responses or HIV-1 disease progression may also be assessed.
## Exploratory studies
Samples may be used for other testing and research related to furthering the understanding of HIV immunology or vaccines. In addition, cryopreserved samples may be used to perform additional assays to support standardization and validation of existing or newly developed methods.
## Other use of stored specimens
The HVTN stores specimens from all study participants indefinitely, unless a participant requests that specimens be destroyed or if required by IRB/EC, or RE.
Other use of specimens is defined as studies not described in the protocol.
This research may relate to HIV, vaccines, the immune system, and other diseases. This could include limited genetic testing and, potentially, genome-wide studies. This research is done only to the extent authorized in each study site's informed consent form, or as otherwise authorized under applicable law. Other testing on specimens will occur only after review and approval by the HVTN, the IRB/EC of the researcher requesting the specimens, and the CRS's IRBs/ECs if required.
The protocol sample informed consent form is written so that the participant either explicitly allows or does not allow their samples to be used in other research when they sign the form. Participants who initially agree to other use of their samples may rescind their approval once they enter the study; such participants will remain in this study and their samples will only be used for the studies described in this protocol. If a participant decides against allowing other research using his or her samples, or at any time rescinds prior approval for such other use, the study site investigator or designee must notify HVTN Regulatory Affairs in writing. In either case, HVTN Regulatory Affairs directs the HVTN Lab Program not to use samples from these participants for such other uses.
CRSs must notify HVTN Regulatory Affairs if institutional or local governmental requirements pose a conflict with or impose restrictions on other use of specimens.
## Biohazard containment
As the transmission of HIV and other blood-borne pathogens can occur through contact with contaminated needles, blood, and blood products, appropriate precautions will be employed by all personnel in the drawing of blood and shipping and handling of all specimens for this study, as currently recommended by the CDC and the NIH or other applicable agencies.
All dangerous goods materials, including Biological Substances, Category A or Category B, must be transported according to instructions detailed in the International Air Transport Association Dangerous Goods Regulations. The clinician members of HVTN 104 PSRT are responsible for decisions related to participant safety.
The Protocol Team clinic coordinator, project manager, vaccine developer representative, clinical trial manager, and others may also be included in HVTN 104 PSRT meetings.
## Hvtn smb
The SMB is a multidisciplinary group consisting of biostatisticians, clinicians, and experts in HIV vaccine research that, collectively, has experience in the conduct and monitoring of vaccine and drug trials. Members of the SMB are not directly affiliated with the protocols under review.
The SMB reviews safety data, unblinded as to treatment arm, approximately every 4 months. The reviews consist of evaluation of cumulative reactogenicity events, AE, laboratory safety data, and individual reports of adverse events requiring expedited reporting to DAIDS. To increase the sensitivity for detecting potential safety problems, the SMB will review safety data aggregated across multiple protocols that use the same or similar vaccine candidates. The SMB conducts additional special reviews at the request of the HVTN 104 PSRT.
Study sites will receive SMB summary minutes and are responsible for forwarding them to their IRB/EC and any applicable RE.
## Sdmc roles and responsibilities in safety monitoring
The roles and responsibilities of the SDMC in relation to safety monitoring include:
Maintaining a central database management system for HVTN clinical data;
Providing reports of clinical data to appropriate groups such as the HVTN 104 PSRT and HVTN SMB (see Section 11.1.2).
## Hvtn core roles and responsibilities in safety monitoring
The roles and responsibilities of HVTN Core in relation to safety monitoring include:
Daily monitoring of clinical data for events that meet the safety pause and HVTN 104 PSRT AE review criteria (see Section 11.3);
Notifying HVTN CRSs and other groups when safety pauses or planned holds are instituted and lifted (see Section 11.3);
Querying HVTN CRSs for additional information regarding reported clinical data; and
Providing support to the HVTN 104 PSRT.
## Safety reporting
## Submission of safety forms to sdmc
Sites must submit all safety forms (eg, reactogenicity, adverse experience, urinalysis, local lab results, and concomitant medications) before the end of the next business day after receiving the information. The forms should not be held in anticipation of additional information at a later date. If additional information is received at a later date, the forms should be updated and refaxed before the end of the next business day after receiving the new information.
## Ae reporting
An AE is any untoward medical occurrence in a clinical investigation participant administered a study product/procedure(s) and which does not necessarily have a causal relationship with this treatment. An AE can therefore be any unfavorable and unintended sign (including an abnormal laboratory finding), symptom, or disease temporally associated with the use of an investigational study product/procedure(s), whether or not related to the investigational study product/procedure(s). All AEs are graded according to the Division of AIDS (2) if the AE meets the criteria for a safety pause/prompt AE review (Section 11.3).
Sites are expected to notify the CSS of any serious safety concern requiring their attention (see [fig_ref] Table 11 - 1: AE notification and safety pause/AE review rules [/fig_ref] In the case of email notification, the CSS will reply during working hours (US Pacific Time) to confirm that the email has been received and reviewed. If email service is not available, the HVTN CRS should notify the CSS of the event by telephone, then submit CRFs.
In addition, site investigators are required to submit AE information in accordance with IRB/EC and any applicable RE requirements.
## Expedited reporting of adverse events to daids
Requirements, definitions and methods for expedited reporting of AEs are outlined in Version 2.0 (January 2010) of the Manual for Expedited Reporting of Adverse Events to DAIDS (DAIDS EAE Manual), which is available on the RSC website at http://rsc.techres.com/safetyandpharmacovigilance/. The SAE Reporting Category will be used for this study.
The internet-based DAIDS Adverse Event Reporting System (DAERS) must be used for expedited AE reporting to DAIDS. In the event of system outages or technical difficulties, expedited AE reports may be submitted via the DAIDS EAE Form. For questions about DAERS, please contact [email protected] or from within the DAERS application itself.
Sites where DAERS has not been implemented will submit expedited AE reports by documenting the information on the current DAIDS EAE Form. This form is available on the RSC website: http://rsc.tech-res.com/safetyandpharmacovigilance/. For questions about expedited AE reporting, please contact the RSC ([email protected]).
The study products for which expedited reporting are required are:
[formula] VRC01 Placebo for VRC01 Sodium chloride placebo [/formula]
While the participant is in the study reporting period (See Section 3), the SAE Reporting Category will be used.
After the protocol-defined AE reporting period for the study, unless otherwise noted, only Suspected, Unexpected Serious Adverse Reactions as defined in Version 2.0 of the DAIDS EAE Manual must be reported to DAIDS, if the study staff become aware of the events.
The NIAID/DAIDS will report all unexpected SAEs related to the study products observed in this clinical trial to the FDA in accordance with 21 CFR 312.32 (IND Safety Reports). However, because safety is a primary study endpoint, the Sponsor Medical Officer will not be unblinded to study treatment assignment when there is an assessment of relatedness of the SAE with the study product(s); and the safety report will be sent to the FDA based on the blinded attribution assessment.
If the PSRT believes unblinding of the site PI to treatment assignment will assist with the clinical management of the SAE, the PSRT will consult the independent HVTN SMB for a recommendation. In the event the HVTN SMB determines that unblinding is indicated, the SMB will inform the site physician of the participant's treatment assignment in such a manner as to maintain the study blind of the PSRT and study team. For additional impact and management of SAEs on the study, refer to Section 11.3.
## Safety pause and prompt psrt ae review
When a trial is placed on safety pause, all enrollment and administration with the product related to the event that triggered the pause will be held until further notice. The AEs that will lead to a safety pause or prompt HVTN 104 PSRT AE review are summarized in Once a trial is paused, the HVTN 104 PSRT reviews safety data and decides whether the pause can be lifted or permanent discontinuation of study product administration is appropriate, consulting the SMB if necessary. HVTN Core notifies the participating HVTN CRSs, HVTN Regulatory Affairs, DAIDS PAB, DAIDS RAB, and DAIDS SPT of the decision regarding resumption or discontinuation of study product administrations. Based on the HVTN 104 PSRT assessment, DAIDS RAB notifies the FDA as needed.
If an immediate HVTN 104 PSRT notification or prompt HVTN 104 PSRT AE review is triggered, HVTN Core notifies the HVTN 104 PSRT as soon as possible during working hours (US Pacific Time)-or, if the information was received during off hours, by the morning of the next work day. If a prompt HVTN 104 PSRT AE review cannot be completed within 72 hours of notification (excluding weekends and US federal holidays), an automatic safety pause occurs.
The HVTN requires that each CRS submit to its IRB/EC protocol-related safety information (such as IND safety reports, notification of vaccine holds due to the pause rules, and notification of other unplanned safety pauses). CRSs must also follow all applicable RE reporting requirements.
In addition, all other AEs are reviewed routinely by the HVTN 104 PSRT (see Section 11.4.2).
## Review of cumulative safety data
Routine safety review occurs at the start of enrollment and then throughout the study.
Reviews proceed from a standardized set of protocol-specific safety data reports. These reports are produced by the SDMC and include queries to the HVTN CRSs. Events are tracked by internal reports until resolution.
## Daily review
Blinded daily safety reviews are routinely conducted by HVTN Core for events requiring expedited reporting to DAIDS, and events that meet safety pause criteria or prompt HVTN 104 PSRT AE review criteria.
## Weekly review
During the study product administration phase of the trial, the HVTN 104 PSRT reviews clinical safety reports on a weekly basis and conducts calls to review the data as appropriate. After the study product administrations and the final 2-week safety visits are completed, less frequent reporting and safety reviews may be conducted at the discretion of the HVTN 104 PSRT. HVTN Core reviews reports of clinical and laboratory AEs. Events identified during the review that are considered questionable, inconsistent, or unexplained are referred to the HVTN CRS clinic coordinator for verification.
## Study termination
This study may be terminated early by the determination of the HVTN 104 PSRT, HVTN SMB, FDA, NIH, Office for Human Research Protections (OHRP), or study product developer. In addition, the conduct of this study at an individual HVTN CRS may be terminated by the determination of the IRB/EC and any applicable RE.
## Protocol conduct
This protocol and all actions and activities connected with it will be conducted in compliance with the principles of GCP (ICHe6), and according to DAIDS and HVTN policies and procedures as specified in the HVTN Manual of Operations, DAIDS Clinical Research Policies and Standard Procedures Documents including procedures for the following:
Protocol registration, activation, and implementation;
Informed consent, screening, and enrollment;
Study participant reimbursement;
Clinical and safety assessments;
Safety monitoring and reporting;
Data collection, documentation, transfer, and storage;
Participant confidentiality;
Study follow-up and close-out;
Unblinding of staff and participants;
Quality control;
Protocol monitoring and compliance;
Advocacy and assistance to participants regarding negative social impacts associated with the vaccine trial;
Risk reduction counseling;
Specimen collection, processing, and analysis;
Ancillary studies, and Destruction of specimens.
Any policies or procedures that vary from DAIDS and HVTN standards or require additional instructions (eg, instructions for randomization specific to this study) will be described in the HVTN 104 Study Specific Procedures.
## Social impacts
Participants in this study risk experiencing discrimination or other personal problems, resulting from study participation. The HVTN CRS is obliged to provide advocacy for and assistance to participants regarding these negative social impacts associated with the vaccine trial. If HVTN CRS staff have questions regarding ways to assist a participant dealing with a social impact, a designated NIAID or HVTN Core representative can be contacted.
Social harms are tabulated by the SDMC and are subjected to descriptive analysis. The goal is to reduce their incidence and enhance the ability of study staff to mitigate them when possible.
Summary tables of social impact events will be generated weekly, and made available for review by the protocol chairs, protocol team leader, and the designated NIAID representative.
## Emergency communication with study participants
As in all clinical research, this study may generate a need to reach participants quickly to avoid imminent harm, or to report study findings that may otherwise concern their health or welfare.
When such communication is needed, the CRS will request that its IRB/EC and any applicable RE expedite review of the message. If this review cannot be completed in a timeframe consistent with the urgency of the required communication, the site should contact the participant first, and then notify the IRB/EC and any applicable RE of the matter as soon as possible.
## Version history
The Protocol Team may modify the original version of the protocol. Modifications are made to HVTN protocols via clarification memos, letters of amendment, or full protocol amendments.
The version history of, and modifications to, Protocol HVTN 104 are described below. Thank you for your interest in our research study. Please read this consent form or ask someone to read it to you. If you decide to join the study, we will ask you to sign or make your mark on this form. We will offer you a copy to keep. We will ask you questions to see if we have explained everything clearly. You can also ask us questions about the study.
Research is not the same as treatment or medical care. The purpose of a research study is to answer scientific questions.
## About the study
The HIV Vaccine Trials Network (HVTN) and [Insert site name] are doing a study to test an antibody against HIV. HIV is the virus that causes AIDS. Antibodies are one of the ways the human body fights infection. Antibodies are natural proteins that the body can make to prevent infectious agents such as bacteria and viruses from making you sick. Antibodies can also be manufactured like a drug and infused or injected into the body. This approach has been successfully used to prevent or treat some other diseases.
About 88 people will take part in this study at multiple sites. The researcher in charge of this study at this clinic is [Insert name of site PI]. The US National Institutes of Health (NIH) is paying for the study.
## We are doing this study to answer several questions.
Are the study products safe to give to people?
Are people able to take the study products without becoming too uncomfortable?
How do people's bodies respond to the study products?
How much of the antibody remains in your body as time passes?
How does the body's response to the study products change depending on the amount and timing of the doses?
Does the method of giving the antibody change the body's response?
## The study products cannot give you hiv.
It is impossible for the study products to give you HIV. Also, they cannot cause you to give HIV to someone else. However, we do not know if the study products will decrease, increase, or not change your chance of becoming infected with HIV if you are exposed to the virus.
## These study products are experimental.
The antibody being tested is called VRC-HIVMAB060-00-AB. It is an antibody against the HIV virus. From here on, we will call it VRC01 or the antibody. We will also be testing the placebo. The placebo is made from inactive ingredients made to look like the antibody. Together we will call them the study products.
They are experimental. That means we do not know if they will be safe to use in people, or if the antibody will work to prevent HIV infection. They are used only in research studies.
The study products were developed by the NIH. They were both made using the controlled, sterile conditions used for drug manufacturing.
In laboratory and animal studies, VRC01 attached to and disabled many kinds of HIV viruses. We do not know if the antibody will act the same way when given to people. It will take many studies to learn if the products will be useful for prevention of HIV or treatment of HIV. This study alone will not answer these questions.
## Risks of vrc01:
VRC01 has been given to more than 40 participants in clinical trials at the NIH Clinical Center. The study products have been tested in one study with HIV-infected participants and in one study with HIV-uninfected participants. Over 40 participants have received VRC01 in HVTN 104 also. So far the antibody is generally well tolerated; the majority of participants had no or mild side effects; a few reported moderate or severe subjective symptoms such as malaise/fatigue or muscle aches. When injected into skin there may be mild redness, swelling, and itching at the injection site that resolves within a few minutes to hours. Some participants had a temporary change in a laboratory test that required additional tests. These lab changes did not cause symptoms and went away on their own.
Two participants had some mild side effects after injections. One person had a mild itchy rash in several places a few days after the injection. The rash lasted a few hours and went away on its own. The other person had a tight feeling in their chest and some wheezing that started shortly after the injection and went away on its own. We do not know yet if these participants got VRC01 or the placebo, so we decided not to give them any more of the study products.
With antibody products, most side effects tend to occur within the first 24 hours.
In general, side effects of antibody products are mild, but may include fever, chills, shaking, nausea, vomiting, pain, headache, dizziness, trouble breathing, high or low blood pressure, itchiness, rash, hives, lip or face swelling, diarrhea, racing heartbeat or chest pain.
In a study where VRC01 was given to animals, there was a small increase in two liver lab tests. This increase lasted for a short period of time. No sign of liver damage was seen in the animals. There were no abnormal findings in the animals' organs except for irritation at the location where the antibody was given.
VRC01 may have other side effects that we do not know about yet.
Some antibody products have a small risk of causing serious drug reactions. These reactions may be life-threatening.
One type of reaction may occur soon after an antibody product is given. It includes difficulty breathing possibly leading to low blood oxygen, low blood pressure, hives or rash, and/or swelling in the mouth and face.
A second type of reaction may occur several days to three weeks after an antibody product is given. It includes having hives or a rash, fever, big lymph nodes, muscle pains, joint pains, chest discomfort and shortness of breath.
When antibodies are given to a person by infusion or injection they do not last in the body more than a few months. Any antibody given to you in this study will be gone from your body several months after your last dose.
## Risks of placebo:
The placebo does not contain antibodies and is made of inactive ingredients. These are generally recognized as safe but there may be unknown risks associated with the placebo.
Joining the study 4. It is completely up to you whether or not to join the study.
Take your time in deciding. If it helps, talk to people you trust, such as your doctor, friends or family. If you decide not to join this study, or if you leave it after you have joined, your other care at this clinic and the benefits or rights you would normally have will not be affected.
If you join this study, you may not be allowed to join other HIV vaccine or HIV prevention studies now or in the future. You cannot be in this study while you are in another study where you receive a study product. Also during the study, you should not donate blood or tissue.
If you choose not to join this study, you may be able to join another study.
Site: Remove item 5 if you use a separate screening consent that covers these procedures.
## 5.
If you decide to join the study, we will screen you to see if you are eligible.
Screening involves a physical exam, HIV test and health history. A physical exam may include, but is not limited to:
Checking your weight, and height, temperature and blood pressure Looking in your mouth and throat Listening to your heart and lungs Feeling your abdomen (stomach and liver)
Checking your veins to assess how easy it might be to start an IV We will also do blood and urine tests. These tests tell us about some aspects of your health, such as how healthy your kidneys, liver, and immune system are. We will ask you about medications you are taking. We will also test you for syphilis, Hepatitis B, and Hepatitis C. We will ask you about behaviors that might put you at risk for getting HIV. If you were born female, we will test you for pregnancy. People who have had a complete hysterectomy (removal of the uterus and ovaries, verified by medical records), are not required to have a pregnancy test.
We will review the screening results with you, and offer you counseling and referral if you need medical care. We will not pay for this medical care. The screening results may show you are not eligible to join the study, even if you want to.
6. If you were born female and could become pregnant, you must agree to use birth control to join this study.
## Site: list approved birth control methods here if you do not want to hand out the separate approved birth control methods sheet.
You should not become pregnant during the study because we do not know how the study products could affect the developing baby. You must agree to use effective birth control from 21 days before you first receive study products until your last required clinic visit. We will talk to you about effective birth control methods. They are listed on a handout that we will give to you. Site: Delete the preceding sentence if you include the birth control sheet in this consent form. If you join the study, we will test you for pregnancy at some visits.
## Being in the study
If you meet the study requirements and want to join, here is what will happen:
## You will come to the clinic for scheduled visits about 15-21 times over about 8 months depending on which group you are in.
## Site: insert number of visits and range of visit lengths. (there is site-specific variation in screening protocols and in the number of possible follow-up visits between protocolmandated visits.)
## Visits can last from [#] to [#] hours.
You may have to come for more visits if you have a lab or health issue.
We may contact you after the main study ends (for example, to tell you about the study results).
The high, medium, and low doses of the study products that are used in all groups will be adjusted for your body weight. We will weigh you on the day of each dose to determine the amount you will get. The night after each infusion or injection, and for three more days, you will need to write down how you are feeling and if you have any symptoms. Contact the clinic staff if you have any issues or concerns after receiving an infusion or injection. If you have a problem, we will continue to check on you until it goes away.
10. We will give you either the antibody or the placebo.
Not everyone in this study will get the antibody.
Groups 1, 2, 4, and 5: All of the people in these groups will get the antibody.
Group 3: 20 of the people in group 3 will get the antibody. The other 4 people in group 3 will get the placebo.
In Group 3, there is one placebo for the infusion given by IV, and another placebo for the SC injection. Neither you nor the clinic staff will know if you are getting the antibody or placebo. Only the pharmacist at your site will have this information while the study is going on. We will compare the results from people who got the placebo with results from people who got the antibody.
We will also compare the results from people who got the antibody in all 5 groups.
If you are in Group 3, you will have to wait until everyone completes their final study visits to find out whether you got the antibody or the placebos. This could be more than a
year. But, if you have a serious medical problem and need to know what you got before the end of the study, we can tell you.
## You will be assigned to a group.
Based on your screening visit and schedule, you and the study staff will discuss which group is best for you. Group 3 has more study visits, so you will need to decide if this fits your schedule. If you prefer a less frequent visit schedule, we will assign you to Group 1, 2, 4, or 5 randomly, like flipping a coin. You will be in the same group the whole time you are in the study.
12. In addition to giving you the antibody or placebo, we will: Ask questions about your health, including medications you may be taking;
[formula] [/formula]
Ask questions about your experience of getting the antibody or placebo;
Ask questions about any personal problems or benefits you may have from being in the study.
When we take blood, the amount will depend on the lab tests we need to do. It will be some amount between 20 mL and 120 mL (a little more than 1 tablespoon to about ½ cup). Your body will make new blood to replace the blood we take out. We will be looking for side effects. We will review the results of these procedures and tests with you at your next visit, or sooner if necessary. If any of the results are important to your health, we will tell you. We will also offer you counseling and referral for needed care.
13. If you agree, we will also collect saliva, rectal and semen or cervical samples.
Because most people are exposed to HIV in their mouth, or on their rectum, penis, or vagina, it is important to learn if antibodies are found in these locations after getting infusions or injections. For this reason, we want to collect saliva, rectal and semen or cervical samples to look for antibodies. Depending on which group you are in, we will collect these samples at 7 or 8 visits. We will do this only if you agree and are able to provide the samples.
At the end of this consent form, we will ask if you allow us to collect these samples. You can decide not to give us these samples and still be in the study. You can decide to provide some of these samples and not others. If you agree to provide these samples, you can change your mind at any time during the study.
## About saliva samples
For participants who agree to give saliva samples, we will ask you to spit into a container.
We want to collect about a teaspoon of saliva. Please avoid the following for 1 hour before your visit:
smoking, eating, chewing gum or tobacco, drinking anything but water, intimate oral activity, and brushing your teeth or using mouthwash.
## About rectal samples
For participants who agree to give rectal samples, we will collect rectal fluids by placing a small absorbent sponge in the rectum using a plastic tube about as wide as a pencil. This will take about 5 minutes. We will not collect the sample if you have an active infection, inflamed hemorrhoids, or colitis/diarrhea. We will:
perform a pregnancy test for participants born female; people who have had a total hysterectomy (removal of the uterus and ovaries, verified by medical records), are not required to undergo pregnancy testing. We do not do most study procedures on pregnant women, so we will do a pregnancy test first.
not collect the sample if you have had receptive anal intercourse or inserted anything into your anus for 48 hours (2 days) before the visit.
not collect the sample if you have used steroids or other anti-inflammatory creams in or around your anus for 48 hours (2 days) before the visit.
## About semen samples
For participants who were born male and who agree to give semen samples, you may provide the samples at home or at the clinic. We will ask you:
not to ejaculate for 48 hours (2 days) before the visit. This will help make sure the samples you provide give accurate lab readings.
to ejaculate into a plastic cup that we will give to you.
to bring the semen sample to the clinic within 2 hours after collection, if the sample is collected outside of the clinic.
## About cervical samples
For participants who were born female and who agree, we will collect cervical fluid. To collect cervical fluid, we will insert a speculum (a device that opens the vagina) into your vagina. Then we will place a small sponge in the opening of the cervix for about a minute to absorb the fluid. We will not collect the sample if you have any active genital infections or sores. We will:
perform a pregnancy test for participants born female; persons who have had a total hysterectomy (removal of the uterus and ovaries, verified by medical records), are not required to undergo pregnancy testing. We do not do most study procedures on pregnant women, so we will do a pregnancy test first.
not collect the sample if you are menstruating.
not collect the sample if you have had vaginal intercourse or inserted anything into your vagina for 48 hours (2 days) before the visit.
not collect the sample if you have used any spermicide, lubricants or topical/intravaginal medications (such as topical yeast infection treatments), including douching, within 48 hours (2 days) before the visit.
require a Pap smear if you have not had one within 3 years before enrollment (depending on your age), with the latest result reported as normal. We can give you a Pap smear if you have not had one within that timeframe.
14. We will counsel you on avoiding HIV infection.
We will ask you personal questions about your HIV risk factors such as sexual behavior and drug use. We will talk with you about ways to keep your risk of getting HIV low. Some topics we may discuss include:
What you think may cause risky behavior for you.
Methods to avoid getting HIV.
These may include not having sex, using condoms, or behavior changes, such as cutting down on alcohol. We will talk with you about new methods of HIV prevention and can give you information on how to access them.
## We will test your samples for this study.
We will send your samples (without your name) to a lab to see how your body responds to the antibody or placebo. The researchers may:
Take cells from your samples and grow more of them. We may grow more of your cells over time, so that they can continue to contribute to this study.
Do limited genetic testing. Your genes are passed to you from your birth parents. They affect how you look and how your body works. The difference in people's genes can help explain why some people get a disease while others do not. Limited genetic testing involves only some of your genes, not all of your genes (your genome). The researchers will not look at all of your genes, only the genes related to the immune system and diseases.
These tests are for research purposes only. The lab will not give the results to you or this clinic, and the results will not become part of your study record. What information is shared with other researchers? The samples and limited information will be labeled with a code number. Your name will not be part of the information. However, some information that we share may be personal, such as your race, ethnicity, gender, health information from the study, and HIV status. We may share information about the study products you received and how your body responded to the study products.
## What kind of studies might be done with my extra samples and information?
The studies will be related to HIV, vaccines, the immune system and other diseases. The researchers may:
Take cells from your samples and grow more of them. This means the researchers may keep your cells growing over time.
Do limited genetic testing, which involves only looking at some of your genes, not all of your genes.
If you agree, your samples could also be used for genome wide studies. In these studies, researchers will look at all of your genes (your genome). The researchers compare the genomes of many people, looking for common patterns of genes that could help them understand diseases. The researchers may put the information from the genome-wide studies into a protected database so that other researchers can access it. Usually, no one would be able to look at your genome and link it to you as a person. However, if another database exists that also has information on your genome and your name, someone might be able to compare the databases and identify you. If others found out, it could lead to discrimination or other problems. The risk of this is very small.
## Who will have access to my information in studies using my extra samples?
People who may see your information are:
Researchers who use your stored samples and limited information for other research Government agencies that fund or monitor the research using your samples or information The researcher's Institutional Review Board or Ethics Committee The people who work with the researcher All of these people will do their best to protect your information. The results of any new studies that use your extra samples or information may be published. No publication will use your name or identify you personally.
## We will do our best to protect your private information.
Sites: Check HIPAA authorization for conflicts with this section.
Your study records and samples will be kept in a secure location. We will label all of your samples and most of your records with a code number, not your name or other personal information. However, it is possible to identify you, if necessary. We will not share your name with the lab that does the tests on your samples, or with anyone else who does not need to know your name.
Clinic staff will have access to your study records. Your records may also be reviewed by groups who watch over this study to see that we are protecting your rights, keeping you safe, and following the study plan. These groups include: We have a Certificate of Confidentiality from the US government, to help protect your privacy. With the certificate, we do not have to release information about you to someone who is not connected to the study, such as the courts or police. Sometimes we can't use the certificate. Since the US government funds this research, we cannot withhold information from it. Also, you can still release information about yourself and your study participation to others.
The results of this study may be published. No publication will use your name or identify you personally.
We may share information from the study with other researchers. We will not share your name or information that can identify you.
When the study is done, we may share the information from the study with others so they can see it and use it. We will not share any information that will let someone identify you.
## We may stop your infusions or injections or take you out of the study at any time. we may do this even if you want to stay in the study and even if you were scheduled for additional infusions or injections.
This may happen if:
you do not follow instructions,
the researcher thinks that staying in the study might harm you, you get HIV, you enroll in a different research study where you receive another study product, or the study is stopped for any reason.
If we stop your infusions or injections, we may ask you to stay in the study to complete other study procedures.
19. If you become pregnant during the study, we will continue with some procedures but not infusions or injections.
We will do this for as long as it is safe for you and your developing baby.
If you leave the study while you are still pregnant, we will contact you after your due date to ask some questions about your pregnancy and delivery.
20. If you get infected with HIV during the study, we will help you get care and support.
You will not be able to stay in this study. We will counsel you about your HIV infection and about telling your partner(s). We will tell you where you can get support and medical care, and about other studies you may want to join. Site: Modify the following sentence as appropriate. We will not provide or pay for any of your HIV care directly.
## Other risks
## There are other risks to being in this study.
This section describes the other risks and restrictions we know about. There may also be unknown risks, even serious ones. We will tell you if we learn anything new that may affect your willingness to stay in the study.
## Risks of routine medical procedures:
In this study, we will do some routine medical procedures like taking blood. These procedures can cause bleeding, bruising, pain, fainting, soreness, redness, swelling, itching, muscle damage, and (rarely) infection where the needle was inserted or blood clot. Taking blood can cause a low blood cell count (anemia), making you feel tired.
## Risks of iv infusion and sc injection procedures:
Receiving an infusion or injection through a needle may cause stinging, discomfort, pain, soreness, redness, bruising, itching, rash and swelling at the location where the needle goes into the skin. Rarely, needlesticks can result in infections.
## Risks of sampling saliva, rectal and genital fluids
We will ask you to stop some behaviors related to your mouth, rectum, and genitals for a short time before we collect samples from these areas. You may find this inconvenient. These sample collections may also cause some anxiety, temporary discomfort, and embarrassment. For women, the collection of cervical fluid may cause discomfort similar to what happens during a Pap smear. We will try to make you as comfortable as possible.
Personal problems/discrimination/testing HIV antibody positive:
About 10 to 20% of people who join HVTN studies report personal problems or discrimination because of joining an HIV vaccine study. Although this is not a vaccine study, it may raise similar concerns. Family or friends may worry, get upset or angry, or assume that you are infected with HIV or at high risk and treat you unfairly as a result. Rarely, a person has lost a job because the study took too much time away from work, or because their employer thought they had HIV.
## Risks of disclosure of your personal information:
We will take several steps to protect your personal information. Although the risk is very low, it is possible that your personal information could be given to someone who should not have it. If that happened, you could face discrimination, stress, and embarrassment. We can tell you more about how we will protect your personal information if you would like it.
## Risks of genetic testing:
The genetic testing could show you may be at risk for certain diseases. If others found out, it could lead to discrimination or other problems. However, it is almost impossible for you or others to know your test results from the genetic testing. The results are not part of your study records and are not given to you.
In the very unlikely event that your genetic information becomes linked to your name, a federal law called the Genetic Information Nondiscrimination Act (GINA) helps protect you. GINA keeps health insurance companies and employers from seeing results of genetic testing when deciding about giving you health insurance or offering you work. GINA does not help or protect you against discrimination by companies that sell life, disability or long-term care insurance.
## Unknown risks:
We do not know if the study products will increase, decrease, or not change your risk of becoming infected with HIV if you are exposed to the virus. If you get infected with HIV, we do not know how the study products might affect your HIV infection or how long it takes to develop AIDS.
We do not know if getting these study products will affect how you respond to a future approved HIV vaccine. It could be that a future HIV vaccine may not work as well for you because you got these study products. Currently, no HIV vaccine has been approved for use.
We do not know how the study products will affect a pregnant participant or a developing baby.
## Benefits
## The study may not benefit you.
We do not know whether getting the study products might benefit you in any way. However, being in the study might still help you in some ways. The counseling that you get as part of the study may help you avoid getting HIV. The lab tests and physical exams that you get while in this study might detect health problems you don't yet know about.
This study may help in the search for a way to prevent HIV. However, if the study products later become approved and sold, there are no plans to share any money with you.
## Your rights and responsibilities
23. If you join the study, you have rights and responsibilities.
You have many rights that we will respect. You also have responsibilities. We list these in the Participant's Bill of Rights and Responsibilities. It was written for participants in HIV vaccine studies. We will give you a copy of it and will tell you how some of the rights and responsibilities are different because you are not getting a vaccine in this study.
## Leaving the study
## Tell us if you decide to leave the study.
You are free to leave the study at any time and for any reason. Your care at this clinic and your legal rights will not be affected, but it is important for you to let us know.
We will ask you to come back to the clinic one last time for a physical exam, and we may ask to take some blood and urine samples. We will also ask about any personal problems or benefits you have experienced from being in the study. We believe these steps are important to protecting your health, but it is up to you whether to complete them.
## Injuries
## If you get sick or injured during the study, contact us immediately.
Your health is important to us. We will help you get the medical care you need.
You could get sick or injured by the study products and/or procedures. If this happens, the HVTN has limited funds from the U.S. government to pay for your treatment.
If someone gets sick or injured in an HVTN study, the HVTN decides whether the injury is probably related to the study products and/or procedures. If the HVTN decides it was more likely due to the study products and/or procedures than any other cause, then the HVTN will use its funds to pay for treatment. The HVTN expects to cover the entire costs for the treatment of simple, temporary study related injuries. If your injuries are more severe or chronic, the HVTN funds may not be enough. If needed, the HVTN will seek more funds, but cannot guarantee them. If the HVTN cannot pay the entire cost of your treatment, you or your health insurance company would be responsible for any additional costs. Some health insurance companies will not pay for study related injuries.
[Sites: insert locale-appropriate medical insurance language in the preceding paragraph.]
If the HVTN decides the injury is likely not due to the study products and/or procedures, then you or your health insurance would be responsible for treatment costs.
[Sites: insert locale-appropriate medical insurance language in the preceding sentence.]You may disagree with the decision the HVTN makes about your injuries. At your request the HVTN will ask experts who are not connected with the HVTN to review its decision. No matter what, you still have the right to use the court system to address payment for your injuries if you are not satisfied.
Some injuries are not physical. For example, someone might be harmed psychologically or emotionally by being in an HIV related study. Or they might lose wages from injuries because they could not go to work. No funds have been set aside to pay for nonphysical injuries, even if they are related to participation in the study.
## Questions
## If you have questions or problems at any time during your participation in this study, use the following important contacts.
If you have questions about this study, contact [name and telephone number of the investigator or other study staff].
If you have any symptoms that you think may be related to this study, contact [name and telephone number of the investigator or other study staff].
If you have questions about your rights as a research participant, or problems or concerns about how you are being treated in this study, contact [name/title/phone of person on IRB or other appropriate organization].
If you want to leave this study, contact [name and telephone number of the investigator or other study staff].
Your permissions and signature [bib_ref] Human neutralizing monoclonal antibodies of the IgG1 subtype protect against mucosal simian-human..., Baba [/bib_ref]. In Section 13 of this form, we told you about collecting saliva, rectal, and semen or cervical samples. Please write your initials or make your mark in the boxes next to the options you choose.
I agree to provide rectal samples.
I do not agree to provide rectal samples.
I agree to provide semen or cervical samples.
I do not agree to provide semen or cervical samples.
I agree to provide saliva samples.
I do not agree to provide saliva samples. 29. If you agree to join this study, you will need to sign or make your mark below. Before you sign or make your mark on this consent form, make sure of the following:
You have read this consent form, or someone has read it to you.
You feel that you understand what the study is about and what will happen to you if you join. You understand what the possible risks and benefits are.
You have had your questions answered and know that you can ask more.
You agree to join this study.
You will not be giving up any of your rights by signing this consent form.
## How do other researchers get my samples and information?
When a researcher wants to use your samples and/or information, their research plan must be approved by the HVTN. Also, the researcher's institutional review board (IRB) or ethics committee (EC) will review their plan.
## What information is shared with other researchers?
The samples and limited information they receive will be labeled with a code number. Your name will not be part of the information. However, some information that we share may be personal, such as your race, ethnicity, gender, health information from the study, and HIV status. We may share information about the study products you received and how your body responded to the study products.
## What kind of studies might be done with my extra samples and information?
The studies will be related to HIV, vaccines, the immune system and other diseases. The researchers may also:
Take cells from your samples and grow more of them. This means the researchers may keep your cells growing over time.
Do limited genetic testing, which involves only looking at some of your genes, not all of your genes.
If you agree, your samples could also be used for genome wide studies. In these studies, researchers will look at all of your genes (your genome). The researchers compare the genomes of many people, looking for common patterns of genes that could help them understand diseases. The researchers may put the information from the genome-wide studies into a protected database so that other researchers can access it. Usually, no one would be able to look at your genome and link it to you as a person. However, if another database exists that also has information on your genome and your name, someone might be able to compare the databases and identify you. If others found out, it could lead to discrimination or other problems. The risk of this is very small.
## What are the risks of genetic testing?
The genetic testing could show you may be at risk for certain diseases. If others found out, it could lead to discrimination or other problems. However, it is almost impossible for you or others to know your test results from the genetic testing. The results are not part of your study records and are not given to you.
In the very unlikely event that your genetic information becomes linked to your name, a federal law called the Genetic Information Nondiscrimination Act (GINA) helps protect you. GINA keeps health insurance companies and employers from seeing results of genetic testing when deciding about giving you health insurance or offering you work. GINA does not help or protect you against discrimination by companies that sell life, disability or long-term care insurance.
## Who will have access to my information in studies using my extra samples?
Some people will be able to see the research records from any new study that uses your extra samples and information. Remember that your name will not be part of the information.
People who may see your information are:
Researchers who use your stored samples and limited information for other research Government agencies that fund or monitor the research using your samples or information The researcher's Institutional Review Board or Ethics Committee The people who work with the researcher All reviewers will take steps to keep your records private. The results of any new studies that use your extra samples or information may be published. No publication will use your name or identify you personally.
## Questions
## If you have questions or problems about allowing your samples and information to be used in other studies, use the following important contacts.
If you have questions about the use of your samples or information or if you want to change your mind about their use, contact [name and telephone number of the investigator or other study staff].
If you think you may have been harmed because of studies using your samples or information, contact [name and telephone number of the investigator or other study staff].
If you have questions about your rights as a research participant, contact [name/title/phone of person on IRB or other appropriate organization].
## Please write your initials or make your mark in the box next to the option you choose.
I allow my extra samples combined with limited information for other studies related to HIV, vaccines, the immune system, and other diseases. This may include limited genetic testing and keeping my cells growing over time.
OR I agree to the option above and also to allow my extra samples combined with limited information to be used in genome wide studies. HIV diagnostic algorithm [bib_ref] HIV vaccine design and the neutralizing antibody problem, Burton [/bib_ref] UW-VSL
## Or
[formula] - - - - - - - - - - - - - - - - - - - - - - 17 [/formula]
Humoral [fig_ref] Table 4: How the in vitro data correlates with in vivo VRC levels and... [/fig_ref] Safety labs CBC/ Diff/ platelets Local lab Local lab EDTA 5mL 5 -
[formula] - - X - - - - - - - X - - - - - - - - X - - - - Inf#1- 5 - - - - 5 - - 5 - 5 - - - - - - 5 - - - - 30 [/formula]
Chemistry Signed screening consent (if used)Signed screening consent (if used)Signed screening consent (if used)
[formula] - - - - - - - - - - - - - - - - - - - - - - 17- - X - - - - - - - X - - - - - - - - X - - - -- - - - - - - - - - - - - - - - - - - - - - 17- - X - - - - - - - X - - - - - - - - X - - - -X - - - - - - - - - - - - - - - - - - - - - - - - - Assessment of understanding X - - - - - - - - - - - - - - - - - - - - - - - - - Signed protocol consent X - - - - - - - - - - - - - - - - - - - - - - - - - Medical history X - - - - - - - - - - - - - - - - - - - - - - - - - Complete physical exam X - - - - - - - - - - - - - - - - - - - - - - - X - PostX - - - - - - - - - - - - - - - - - - - - - - - - - Assessment of understanding X - - - - - - - - - - - - - - - - - - - - - - - - - Signed protocol consent X - - - - - - - - - - - - - - - - - - - - - - - - - Medical history X - - - - - - - - - - - - - - - - - - - - - - - - - Complete physical exam X - - - - - - - - - - - - - - - - - - - - - - - X - PostX - - - - - - - - - - - - - - - - - - - - - - - - - Assessment of understanding X - - - - - - - - - - - - - - - - - - - - - - - - - Signed protocol consent X - - - - - - - - - - - - - - - - - - - - - - - - - Medical history X - - - - - - - - - - - - - - - - - - - - - - - - - Complete physical exam X - - - - - - - - - - - - - - - - - - - - - - - X - [/formula]
[fig] 2. 5: Adequate safety monitoring 45 CFR 46.111 (a) 6 and 21 CFR 56.111 (a) 6: There is adequate provision for monitoring the data collected to ensure the safety of subjects. [/fig]
[fig] Figure 4 - 1: Analysis of neutralization by VRC01 and b12 antibodies against a panel of 190 pseudoviruses representing the major circulating clades of HIV-1. [/fig]
[fig] Figure 4 - 2: Simulated Model of VRC01 levels in Cohorts 1-3. Cohort 1 entails 20mg/kg IV infusion every 4 weeks with a loading dose of 40 mg/kg, Cohort 2 40mg/kg IV infusion every 8 weeks, and Cohort 3 5mg/kg SC injection every 2 weeks. Initial PK estimates based on the combined VRC 601 and VRC 602 data of 19 IV-infusion participants and 6 SC-injection participants were used in the simulations. [/fig]
[fig] Figure 4 - 3: Simulated Model of VRC01 levels in Group 4. Group 4 entails 10mg/Kg IV infusion every 8 weeks with no loading dose. Initial PK estimates based on the combined VRC 601 and VRC 602 data of 19 IV-infusion participants were used in the simulations. [/fig]
[fig] Figure 4 - 4: VRC01 levels from 3 participants in VRC 602 given a single dose of 20mg/kg IV at Day 0 and followed for 8 weeks. [/fig]
[fig] Virology 17: Negative HIV-1 and -2 blood test: US volunteers must have a negative FDA-approved enzyme immunoassay (EIA). 18. Negative Hepatitis B surface antigen (HBsAg) 19. Negative anti-Hepatitis C virus antibodies (anti-HCV), or negative HCV polymerase chain reaction (PCR) if the anti-HCV is positive Urine 20. Normal urine: [/fig]
[fig] Group 1: OPEN-LABEL)Treatment 1 (T1): VRC-HIVMAB060-00-AB 40 mg/kg to be administered IV in 100 mL of Sodium Chloride for Injection USP, 0.9% at Month 0.THEN VRC-HIVMAB060-00-AB 20 mg/kg to be administered IV in 100 mL of Sodium Chloride for Injection USP, 0.9% at Months 1, 2, 3, 4, and 5.Group 2 (OPEN-LABEL)Treatment 2 (T2): VRC-HIVMAB060-00-AB 40 mg/kg to be administered IV in 100 mL of Sodium Chloride for Injection USP, 0.9% at Months 0, 2, and 4.Group 3 (DOUBLE-BLIND / PLACEBO CONTROLLED) Treatment 3 (T3): VRC-HIVMAB060-00-AB 40 mg/kg to be administered IV in 100 mL of Sodium Chloride for Injection USP, 0.9% at Month 0 THEN VRC-HIVMAB060-00-AB 5 mg/kg to be administered SC every 2 weeks at Months 0.Placebo 3 (P3): IV Placebo for VRC01 (Sodium Chloride for Injection USP, 0.9%) to be administered IV in 100 mL of Sodium Chloride for Injection USP, 0.9% at Month 0 THEN SC Placebo for VRC01 (VRC-PLAMAB068-00-AB) to be administered SC (subcutaneously) every 2 weeks at Months 0. [/fig]
[fig] VRC-HIVMAB060- 00 -: AB [VRC01, Labeled as VRC01 HIV MAb Drug Product VRC-HIVMAB060-00-AB] VRC01 will be provided in a 3 mL clear glass vial containing 2.25 mL (± 0.1 mL) of a sterile clear, colorless to yellow isotonic solution with no visible particles. Each mL contains 100 mg (± 10 mg) of VRC-HIVMAB060-00-AB in formulation buffer. The formulation buffer is composed of 25 mM sodium citrate, 50 mM sodium chloride, and 150 mM L-arginine hydrochloride at pH 5.8. Vials are intended for single use only and do NOT contain a preservative. [/fig]
[fig] 8.3. 1: VRC-HIVMAB060-00-AB (40mg/kg IV) -OPEN LABEL ONLY (Group 1 Day 0 only and Group 2 at Months 0, 2 and 4) [/fig]
[fig] : CRS's IRB/EC, CRS's institution and any applicable REs, and Elements of informed consent as described in Title 45, Code of Federal Regulations (CFR) Part 46 and Title 21 CFR, Part 50, and in the International Conference on Harmonisation (ICH) E6, Good Clinical Practice: Consolidated Guidance 4.8. [/fig]
[table] Table 4: How the in vitro data correlates with in vivo VRC levels and nAb function is unknown, however, some limited NHP challenge data suggest that VRC01 serum levels between 75 mcg/ml and 10 mcg/ml would span a wide range of HIV virus sensitivities in vivo (John Mascola, VRC, personal communication). A dosing schedule of 10 mg/kg IV every 8 weeks is anticipated to result in a sufficiently variable range of VRC01 levels and corresponding nAb titers to inform dosing schedules for future phase 2b trials. [/table]
[table] Table 6 - 3: Two-sided 95% confidence intervals based on observing a particular average drug level in participants in any of the active arms (n=10, n=18) [/table]
[table] Table for: Grading the Severity of Adult and Pediatric Adverse Events (DAIDS AE Grading Table), Version 1.0, December 2004 (Clarification August 2009). [/table]
[table] Table 11 - 1: AE notification and safety pause/AE review rules [/table]
[bib_ref] Vaccination with ALVAC and AIDSVAX to prevent HIV-1 infection in Thailand, Rerks-Ngarm [/bib_ref] [bib_ref] Immune-correlates analysis of an HIV-1 vaccine efficacy trial, Haynes [/bib_ref] [bib_ref] Complete protection of neonatal rhesus macaques against oral exposure to pathogenic simian-human..., Ferrantelli [/bib_ref] [bib_ref] HIV vaccine design and the neutralizing antibody problem, Burton [/bib_ref] [bib_ref] Aiming to induce broadly reactive neutralizing antibody responses with HIV-1 vaccine candidates, Haynes [/bib_ref] [bib_ref] Neutralizing antibodies generated during natural HIV-1 infection: good news for an HIV-1..., Stamatatos [/bib_ref] [bib_ref] Broad HIV-1 neutralization mediated by CD4-binding site antibodies, Li [/bib_ref] [bib_ref] Vaccination with ALVAC and AIDSVAX to prevent HIV-1 infection in Thailand, Rerks-Ngarm [/bib_ref] [bib_ref] Immune-correlates analysis of an HIV-1 vaccine efficacy trial, Haynes [/bib_ref] [bib_ref] Complete protection of neonatal rhesus macaques against oral exposure to pathogenic simian-human..., Ferrantelli [/bib_ref] [bib_ref] HIV vaccine design and the neutralizing antibody problem, Burton [/bib_ref] [bib_ref] Aiming to induce broadly reactive neutralizing antibody responses with HIV-1 vaccine candidates, Haynes [/bib_ref] [bib_ref] Neutralizing antibodies generated during natural HIV-1 infection: good news for an HIV-1..., Stamatatos [/bib_ref] [bib_ref] Broad HIV-1 neutralization mediated by CD4-binding site antibodies, Li [/bib_ref] |
Alcohol consumption and PSA-detected prostate cancer risk—A case-control nested in the ProtecT study
Alcohol is an established carcinogen but not an established risk factor for prostate cancer, despite some recent prospective studies suggesting increased risk among heavy drinkers. The aim of this study was to investigate the role of alcohol on prostate-specific antigen (PSA) levels and prostate cancer risk. Two thousand four hundred PSA detected prostate cancer cases and 12,700 controls matched on age and general practice were identified through a case-control study nested in the PSA-testing phase of a large UK-based randomized controlled trial for prostate cancer treatment (ProtecT). Linear and multinomial logistic regression models were used to estimate ratios of geometric means (RGMs) of PSA and relative risk ratios (RRRs) of prostate cancer by stage and grade, with 95% confidence intervals (CIs), associated with weekly alcohol intake and drinking patterns. We found evidence of lower PSA (RGM 0.98, 95% CI: 0.98-0.99) and decreased risk of low Gleason-grade (RRR 0.96; 95%CI 0.93-0.99) but increased risk of high-grade prostate cancer (RRR 1.04; 95%CI 0.99-1.08; p difference 50.004) per 10 units/week increase in alcohol consumption, not explained by current BMI, blood pressure, comorbidities, or reverse causation. This is the first large population-based study to find evidence of lower PSA levels for increasing alcohol consumption, with potential public health implications for the detection of prostate cancer. Our results also support a modestly higher risk of high-grade disease for heavy drinkers, but require independent replication to establish the nature of the association of alcohol with low-grade disease, preferably in cohorts with a heterogeneous case-mix.
Alcohol is a known human carcinogen, 1 but there is limited evidence on whether it affects prostate cancer.Alcohol use and abuse are common in developed countries like the UK and USA and are rapidly increasing in many developing countries, especially amongst the urban poor. [bib_ref] Global burden of disease and injury and economic cost attributable to alcohol..., Rehm [/bib_ref] More research into a possible causal link between alcohol and prostate cancer is justified by the potential to discover a target for disease prevention on the one hand, and on the other by the need to better characterize the long-term effects of alcohol consumption over the life-course.
The evidence from epidemiological studies on the role alcohol plays in prostate cancer is inconclusive, despite the many publications to date, including 14 prospective studies, which have been pooled into one of two meta-analyses,,5 and at least six new studies published after the searches for those two reviews were last updated. [bib_ref] Red wine consumption and risk of prostate cancer: the California men's health..., Chao [/bib_ref] [bib_ref] Alcohol consumption, finasteride, and prostate cancer risk results from the prostate cancer..., Gong [/bib_ref] [bib_ref] Alcohol consumption and the risk for prostate cancer in the European prospective..., Rohrmann [/bib_ref] [bib_ref] A prospective cohort study of red wine consumption and risk of prostate..., Sutcliffe [/bib_ref] [bib_ref] Alcohol use and the risk of prostate cancer: results from the VITAL..., Velicer [/bib_ref] [bib_ref] Alcoholic beverages and prostate cancer in a prospective US cohort study, Watters [/bib_ref] The two meta-analyses based on systematic reviews yielded remarkably similar pooled estimates, with random-effects relative risks of 1.02 (95% confidence interval (CI): 0.85-1.23) 4 per additional drink/day, and of 1.03 (95% CI: 1.00-1.07)per additional unit/day. Three of the six most recent prospective studies published after these review found some evidence of an increased risk of high-grade disease ranging from a 20% increase to a doubling of the risk in the heaviest drinkers (consuming 1þ to 5þ drinks/day) as compared to light or nondrinkers, [bib_ref] Alcohol consumption, finasteride, and prostate cancer risk results from the prostate cancer..., Gong [/bib_ref] [bib_ref] Alcohol consumption and the risk for prostate cancer in the European prospective..., Rohrmann [/bib_ref] [bib_ref] A prospective cohort study of red wine consumption and risk of prostate..., Sutcliffe [/bib_ref] whereas two studies, one on advanced and fatal disease, [bib_ref] Alcoholic beverages and prostate cancer in a prospective US cohort study, Watters [/bib_ref] did not find increased risks, [bib_ref] Red wine consumption and risk of prostate cancer: the California men's health..., Chao [/bib_ref] [bib_ref] Alcoholic beverages and prostate cancer in a prospective US cohort study, Watters [/bib_ref] and another did not report results for heavy drinkers. [bib_ref] Alcohol use and the risk of prostate cancer: results from the VITAL..., Velicer [/bib_ref] Evidence from the Prostate Cancer Prevention Trial of an interaction between heavy drinking and the antiandrogen finasteride further suggests the possibility of a real and important effect of alcohol on prostate carcinogenesis. [bib_ref] Alcohol consumption, finasteride, and prostate cancer risk results from the prostate cancer..., Gong [/bib_ref] Although the available evidence does not offer any clear indication for a clinically meaningful increase in prostate cancer risk, mainly due to substantial between-study heterogeneity, an effect of heavier alcohol drinking on high-grade disease in particular cannot be excluded. Between-study heterogeneity is likely to be generated by methodological limitations at individual study level, which could either mask true underlying effects, or generate spurious results. These limitations include the sick-quitter effect (a form of reverse causation), associative selection bias, residual confounding and measurement error (typical of all observational studies of alcohol effects), as well as detection bias (specific to prostate cancer, since prostate-specific antigen (PSA) screening is associated with health-related behaviors, [bib_ref] Overdiagnosis due to prostate-specific antigen screening: lessons from U.S. prostate cancer incidence..., Etzioni [/bib_ref] [bib_ref] Prostate cancer association studies: pitfalls and solutions to cancer misclassification in the..., Platz [/bib_ref] and PSA itself might be affected by alcohol independently of prostate cancer). Moreover, it is possible that many studies were underpowered to detect small effects.
The aim of this study was to investigate the association of alcohol consumption from self-reported data with PSA levels and risk of PSA-detected prostate cancer, both overall and subgrouped according to stage and grade. We analyzed results from a large population-based case-control study nested within the PSA-testing phase of a randomized controlled trial for the treatment of localized prostate cancerthe Prostate testing for cancer and Treatment (ProtecT) trial. This study design overcomes some of the problems complicating the interpretation of the current epidemiological evidence: (i) it minimizes reverse causation in the forms of both the sick-quitter effect and recall bias or other differential information biases (because exposure information is collected before PSA test results become available); (ii) the fact that all cases are PSA-detected protects against gross levels of detection bias (caused for example by more health-conscious men undergoing (more frequent) PSA testing, whereas here all participants underwent PSA testing); and (iii) the large sample size allows the investigation of modest effects.
# Material and methods
## Study design and outcomes definition
Participants in this study were recruited in the PSA-testing phase of the ProtecT study. Details of the protocol have been described elsewhere. 14 Briefly, men aged 50-69 years in general practices located around nine UK cities were invited to attend a nurse-led prostate check clinic and, if they consented, to have a PSA test. Participants with a single raised PSA test result (!3.0 ng/ml) were invited to attend the center's urology department for digital rectal examination, repeat PSA test, and 10-core trans-rectal ultrasound-guided biopsy.
Eligible cases were the 3,324 men aged 50-69 with histologically-confirmed prostate cancer, detected among the 111,348 men undergoing PSA-testing between [of which 10,297 (11%) had raised PSA]. Prostate cancer cases were classified into localized or advanced stage according to TNM clinical staging, [bib_ref] The new American Joint Committee on Cancer and International Union Against Cancer..., Ohori [/bib_ref] and into low-and highgrade according to Gleason scores following review of histological material, 16 as previously described. [bib_ref] Height and prostate cancer risk: a large nested case-control study (ProtecT) and..., Zuccolo [/bib_ref] All participants in the ProtecT cohort who had no evidence of prostate cancer were eligible for selection as controls. These included all men with a PSA<3.0 ng/ml or PSA !3.0 ng/ml and a negative biopsy. Cases were frequency-matched (incidence-density sampling) to 6 controls by age (5-year bands) and general practice. As the clinics were held and completed in each general practice in turn, matching for general practice automatically matched for the calendar date of recruitment. Men of ethnicity other than white were a small minority and were excluded. Overall, 2,386 (69% of the total) men with histologically confirmed prostate cancer and 12,727 controls (63% of the total) returned the questionnaire and completed the section on alcohol during the prostate check clinic [fig_ref] Figure 1: Flow-chart of exclusions and final data-sets of study participants-ProtecT nested case-control [/fig_ref].
Study participants gave fully informed consent in writing for the use of their data for research purposes. The Trent Multicentre Research Ethics Committee approved the study.
## Data on alcohol and related phenotypes (potential confounding factors)
Participants were asked to complete a health and lifestyle questionnaire after the check clinic but before their PSA result was available. This included a section on alcohol, with questions on usual frequency and amount consumed over the previous year, changes in consumption over the previous 10 years, and amount drunk in the previous week, and distinguished between types of alcoholic drinks. Usual weekly alcohol intake was obtained by combining intake amount and frequency separately for beer and spirits/wine (weekly number of half pints of beer and weekly number of glasses of wine and units of spirits). This was then converted into standard UK units (1 unit ¼ 8 g of ethanol, corresponding to $ 10 ml of alcohol) by assuming the following volumes and alcohol contents: half pint of beer-284 ml, 3.5%; glass of wine-125 ml, 12%; standard pub measure of spirit-35 ml, 40% (http://www.cks.nhs.uk/alcohol_problem_drinking/background_information/definition/unit_of_alcohol).
## What's new?
Alcohol is not an established risk factor for prostate cancer; however, the current work suggests that heavy drinking could cause a small increase in risk of the more aggressive forms. If the results are confirmed to be causal, prostate cancer risk will be added to the many long-term health risks of heavy drinking, and public health strategies will then also reduce high-risk, poorer prognosis prostate cancer. The authors also found that heavy drinkers have lower PSA levels, suggesting that heavy alcohol consumption could be used as a marker to identify men in whom some cancers might be missed.
## Epidemiology
## Zuccolo et al.
People self-reporting only occasional alcohol consumption (special occasions) were excluded from the computation of usual weekly alcohol intake. The continuous variable ''weekly units of alcohol consumed'' only took positive values and was used in analyses investigating the dose-response effect of alcohol on PSA levels and on PSA-detected prostate cancer among current drinkers. For the purpose of analysis, another variable with five categories was derived that distinguished men who did not drink, men drinking on special occasions only, and men who were current drinkers, divided into thirds of weekly alcohol consumption (based on the distribution of weekly drinking among controls). Additional alcohol behavior variables were drinking most days of the week over the past year (versus more occasional/sporadic drinking), referred to as ''regular drinking,'' and binge drinking, defined as usually consuming at least 5 drinks per drinking session. The traditional threshold of 5 or more drinks (¼ 5þ units) per drinking occasions for the classification of binge drinking was chosen based on recent evidence confirming its validity in predicting alcohol-related risk. [bib_ref] Heavy episodic drinking: determining the predictive utility of five or more drinks, Jackson [/bib_ref] Self-reported data on current diet, lifestyle, history of comorbidities and occupation were also obtained from the questionnaires. Information on skin color, as proxy for ethnicity, was extracted during the nurse-led clinic, and >95% of the ProtecT cohort were recorded as ''White''. During this visit, the nurse measured participants' standing and sitting height, their weight, waist and hips circumference, and diastolic and systolic blood pressure, following standard operating procedures.
# Statistical analysis
To test the association between PSA levels and dimensions of alcohol behavior capturing patterns and levels of consumption (e.g., binge drinking, regular drinking, and amount consumed), we fitted linear regressions to data from randomly selected controls from the ProtecT study. The outcome was first log-transformed; therefore, effect estimates of the association are expressed as ratios of geometric means (RGMs). For example, an RGM of 0.96 corresponds to a (geometric) mean PSA difference of À4%. To test the association between alcohol and risk of PSA-detected prostate cancer, we fitted conditional logistic regressions to the case-control data set, and effect estimates are expressed as odds ratios (OR). Finally, to test the association between alcohol and prostate cancer risk by stage and grade sub-types, we used multinomial logistic regressions and ran two separate sets of analyses, each with the outcome variable grouped into 3 categories: (i) controls, localized cases (stage T1 or T2; NXM0), and advanced cases (stage T3 or T4); (ii) controls, low-grade Gleason cases (Gleason sum < 7), and high-grade Gleason cases (Gleason sum ! 7). Heterogeneity in associations of alcohol with localized as compared with advanced stage or low-grade compared with high-grade cancers were tested Epidemiology using Wald tests. Effect estimates of the association are expressed as relative risk ratios (RRRs).
We investigated the possible confounding of current BMI and blood pressure, as well as history of hypertension, hypercholesterolemia and diabetes, on the alcohol-PSA and alcohol-prostate cancer associations by performing likelihood ratio tests between nested models with and without them as covariates.
All analyses were matched on (prostate cancer risk) or adjusted for (PSA levels, prostate cancer sub-types) the design variables used for matching controls to cases (age in 5-year groups, center and date of recruitment and clinic visit), and additionally for age in years as a continuous variable, since both alcohol behavior and the outcomes are strongly age-dependant. All statistical tests were two-sided.
## Sensitivity analyses
Sensitivity analyses were performed to assess robustness of the alcohol-PSA and alcohol-prostate cancer associations based on self-reported data. Possible differential effects of recent versus usual drinking and of different beverage types on outcomes were investigated by comparing (i) estimates based on questions referring to the last 7 days only versus usual consumption over the past year, and (ii) estimates referring to total consumption, alcohol from beer only, and alcohol from wine and spirits combined (over past year). We also attempted to detect the presence of and adjust for potential bias caused by former heavy drinkers reducing their alcohol intake in middle-age because of comorbidities that may be alcohol-related (sick-quitter effect). To explore the possible impact of such bias, we performed sensitivity analyses by stratifying the analyses of the association between alcohol and PSA and prostate cancer risk according to whether men reported changing their drinking behavior in the last 10 years.
# Results
A description of selected characteristics is presented in [fig_ref] Table 1: Characteristics of cases and controls included in the analysis of alcohol and... [/fig_ref]. Cases were much more likely to have a positive family history of prostate cancer, more likely to have a positive history of hypertension, and less likely to have been diagnosed with diabetes, engage in strenuous exercise, and drink most days of the week.
## Patterns of consumption-binge drinking and regular drinking
Men drinking most days had lower PSA values than men drinking less frequently (RGM 0.96, 95% CI: 0.93-0.98), and the same was true for those binge drinking regularly [fig_ref] Table 2: Associations of dimensions of alcohol drinking behavior with PSA and prostate cancer... [/fig_ref]. Results for risk of total prostate cancer were in the same direction, with lower risk for those binge drinking and drinking regularly, although confidence intervals (CI) were wider and also compatible with the opposite effect [fig_ref] Table 2: Associations of dimensions of alcohol drinking behavior with PSA and prostate cancer... [/fig_ref]. On the basis of results of multinomial logistic regressions and tests for the difference in effect estimates for different cancer sub-types, there was no clear evidence of an association of binge drinking or regular drinking with prostate cancer clinical stage at diagnosis [fig_ref] Table 2: Associations of dimensions of alcohol drinking behavior with PSA and prostate cancer... [/fig_ref]. Men drinking on most days seemed less likely to be diagnosed with advanced disease as compared to those who did not (RRR 0.77, 95% CI: 0.58-1.04), but this was based on just over 200 cases.
There was a suggestion that binge drinking was associated with lower risks of low-grade disease, but no substantial change or perhaps a small increase in risk of high-grade disease, and a similar trend was noted for drinking most days [fig_ref] Table 2: Associations of dimensions of alcohol drinking behavior with PSA and prostate cancer... [/fig_ref].
## Dose-response: alcohol intake in drinkers
There was evidence of a small dose-response type reduction in PSA levels with increasing alcohol consumption (RGM 0.98, 95% CI: 0.98-0.99 per 10 units/week increase) [fig_ref] Table 3: Association of weekly alcohol consumption with PSA and prostate cancer risk Test... [/fig_ref]. Results for total prostate cancer risk were similar in direction but less precisely estimated [fig_ref] Table 3: Association of weekly alcohol consumption with PSA and prostate cancer risk Test... [/fig_ref]. Also, the risk of both localized and advanced cancer seemed to reduce with increasing alcohol consumption when comparing men in the top third of the distribution to men in the bottom third [fig_ref] Table 3: Association of weekly alcohol consumption with PSA and prostate cancer risk Test... [/fig_ref].
There was strong evidence that the dose-response effects for weekly alcohol consumption and prostate cancer risk differed by grade at diagnosis (p ¼ 0.004), with a suggestion that high-grade cancer risk was higher in heavier drinkers (RRR 1.04, 95% CI: 0.99-1.08), but the opposite was true for lowgrade disease (RRR 0.96, 95% CI: 0.93-0.99).
## Adjustment for potential confounding factors
In general, adjustment for current BMI and systolic blood pressure did not substantially change the effect estimates (model 2 when compared to model 1); however, adjustment for pre-existing conditions such as history of hypertension, hypercholesterolemia and diabetes (model 3 when compared to model 2) did attenuate the association of binge drinking with PSA levels [fig_ref] Table 4: Association of alcohol drinking with PSA levels in controls [/fig_ref] , and prostate cancer risk [fig_ref] Table 5: Association of alcohol drinking with prostate cancer risk comparing cases and controls [/fig_ref]. Moreover, fully-adjusted models showed evidence of lower prostate cancer risk but not of lower PSA in nondrinkers as compared to drinkers in the first thirds of the distribution (OR 0.79, 95% CI: 0.64-0.99), but this effect was less pronounced in models 1 and 2 [fig_ref] Table 5: Association of alcohol drinking with prostate cancer risk comparing cases and controls [/fig_ref].
## Sensitivity analyses
Estimates of the association of PSA levels and prostate cancer risk with weekly alcohol consumption, derived through either 7-day recall questions or average quantity and frequency questions referring to the previous 12 months, were in good agreement with each other, with 95% CIs largely overlapping (Figs. S1 and S2, Supporting Information online). Dose-response effect estimates for total consumption, alcohol from beer only, and alcohol from wine and spirits combined were all similar and suggested that alcohol could lower PSA levels by 2-3% per 10 units/week increase in consumption (Table S1a, Supporting Information online). For prostate cancer risk, point estimates were in the same direction, but there was limited statistical evidence to show decreased risks specific to any beverage type (Table S1b, Supporting Information online).
Compared to analyses including the entire sample, restricting analyses to men who did not change drinking habits left most results unchanged. As expected, there was a suggestion of attenuated effects in men who changed drinking behavior in the last 10 years, as compared to men who did not, for both PSA levels and prostate cancer risk [fig_ref] Table 6: Association of alcohol drinking with [/fig_ref].
# Discussion
# Summary of results
On the basis of this large case-control study nested in the ProtecT trial, there was some evidence that increasing alcohol consumption and frequency of drinking were associated with lower PSA and weaker evidence that they were also associated with lower risk of PSA-detected prostate cancer among current drinkers, with generally wide confidence intervals. This was attributable to a decrease in low Gleason grade tumors, the most common cancer sub-type detected through PSA testing, whereas a small increase was observed for the more aggressive highgrade tumors, in line with the suggestion of a small increase in risk of high-grade cancer and a decrease for low-grade cancer for binge drinkers. There was limited evidence of a differential association with alcohol according to tumor TNM stage. There was a tenuous suggestion of a nonlinear association of alcohol with the outcomes, as nondrinkers tended to have lower PSA and lower risk of PSA-detected prostate cancer as compared to light drinkers. However, confidence intervals were somewhat overlapping and occasionally compatible with no effect. Estimates were almost unchanged when restricted to men reporting not to have changed drinking habits in the past 10 years (i.e., no evidence of sick-quitter effect).
## Comparison with previous literature and interpretation of results
The small reduction in risk of PSA-detected prostate cancer observed in this study (OR 0.985, 95% CI: 0.96-1.01 per 10 units/week increase) contrasts with most of the published The current study is the only one in the literature on alcohol and prostate cancer risk whose prostate cancer cases were identified based on a PSA threshold of 3 ng/ml. Discrepancies between our result and previous prospective studies could be due to lower chances of a cancer diagnosis for the heaviest drinkers, because of their lower PSA levels. This is one of the first population-based studies, and the largest to date, to investigate the alcohol-PSA association. Previously two cross-sectional studies did not find strong evidence in support of an association. [bib_ref] Alcohol consumption, finasteride, and prostate cancer risk results from the prostate cancer..., Gong [/bib_ref] [bib_ref] Ethnic differences in diet and associations with clinical markers of prostate disease..., Harris [/bib_ref] The observed negative association is in line with published evidence of an inverse association of alcohol with benign prostate hyperplasia (BPH), [bib_ref] Alcohol consumption is associated with a decreased risk of benign prostatic hyperplasia, Parsons [/bib_ref] possibly underpinned by mechanisms independent of prostate cancer. It is not clear whether the relationship is causal, or due to the presence of confounding, or to reverse causation. The latter might be an issue since most observations of the alcohol-PSA and alcohol-BPH association are cross-sectional, and it is possible that men with increased PSA because of BPH and urinary symptoms might reduce their alcohol consumption. However, the prevalence of lower urinary tract symptoms in this unselected population-based sample was low, [bib_ref] Associations of lower urinary tract symptoms with prostate-specific antigen levels, and screendetected..., Collin [/bib_ref] and therefore unlikely to have substantially impacted on men's drinking behavior.
Results from this study are still compatible with the hypothesis that alcohol could increase prostate cancer risk itself, as suggested by the observations of an increase in high-grade disease in heavy drinkers. The stronger evidence of association with alcohol drinking observed for high-grade tumors could be due to the fact that they are less likely to be affected by detection bias, or to an underlying different etiology of less aggressive and more aggressive disease. The ProtecT data showed a suggestion of a threshold effect in the association of alcohol with high-grade prostate cancer, with a RRR of 1.11 (95% CI: 0.92-1.34) comparing the top third to the bottom third of alcohol consumption. This is in line with what had been suggested by recent prospective studies examining the risk of high-grade disease. [bib_ref] Alcohol consumption, finasteride, and prostate cancer risk results from the prostate cancer..., Gong [/bib_ref] [bib_ref] Alcohol consumption and the risk for prostate cancer in the European prospective..., Rohrmann [/bib_ref] [bib_ref] A prospective cohort study of red wine consumption and risk of prostate..., Sutcliffe [/bib_ref] The highest drinking category in this study comprised of men drinking 20þ alcohol units/week, equivalent to $ 24þ g of ethanol/day. This group therefore includes more moderate drinkers as compared to the top category used in the European Prospective Investigation into Cancer and Nutrition study (60þ g/day, RR 1.17, 95% CI: 0.81-1.67) [bib_ref] Alcohol consumption and the risk for prostate cancer in the European prospective..., Rohrmann [/bib_ref] and in the Prostate Cancer Prevention Trial (50þ g/day, RR 2.01, 95% CI: 1.33-3.05) 7 , but is comparable to the top category of the Health Professionals Follow-up Study (20þ g/day, RR 1.72, 95%CI: 0.88-3.36). 9
## Potential mechanisms
The existence of a threshold effect with increased risk observed in heavy drinkers but not in light or moderate drinkers is biologically plausible. High acetaldehyde concentrations could promote carcinogenesis in prostatic epithelial tissue. [bib_ref] Mechanisms in prostate damage by alcohol, Castro [/bib_ref] Acetaldehyde is mostly produced in the liver by alcohol dehydrogenases, and can reach the prostate through the bloodstream. Once in the prostate, it would be difficult to clear acetaldehyde because acetaldehyde dehydrogenase enzymatic activity is low in prostatic epithelial tissue (http://bioinfo.wilmer.jhu.edu/tiger/db_gene/ALDH2-index.html).
Only very heavy drinking could result in high enough intra-prostatic acetaldehyde concentrations to be carcinogenic (i.e., 40-1000 lM, based on in vitro experiments,with in vitro acetaldehyde levels correlating well with in vivo levels [bib_ref] High acetaldehyde levels in saliva after ethanol consumption: methodological aspects and pathogenetic..., Homann [/bib_ref]. Whereas systemic acetaldehyde concentrations following alcohol ingestion are usually in the range of 1-5 lM, 25 in alcoholics with blood alcohol concentrations >80 mM they have been found to reach 40 lM, even 10-12 h after the last drink. [bib_ref] Polymorphisms of alcohol dehydrogenase-1B and aldehyde dehydrogenase-2 and the blood and salivary..., Yokoyama [/bib_ref] Also, it is currently believed that chronic exposure to oxidative stress, which can be the result of chronic heavy drinking in combination with antioxidant deficiency or malabsorption, may have a carcinogenic effect in the prostate gland and affect both tumor aggressiveness and rate of progression. [bib_ref] Oxidative stress in prostate cancer, Khandrika [/bib_ref] Both these mechanisms arise from (chronic) heavy alcohol drinking, and therefore suggest a threshold effect rather than a dose-response relationship.
# Strengths and limitations
Strengths of this study design include adequate power provided by the large sample size, no recall bias, minimal sickquitter effect and reduced detection bias. Reverse causation, in the form of the sick-quitter effect or otherwise, could explain why moderate and heavy drinkers have lower PSA than light drinkers. It is in fact possible that men with high PSA caused by an enlarged prostate and at higher risk of urinary tract symptoms would have cut down on alcohol, since alcohol is one of the bladder irritants, which could cause urinary retention. [bib_ref] Alcohol consumption is associated with a decreased risk of benign prostatic hyperplasia, Parsons [/bib_ref] However, in analyses where the sample was restricted to men who had not modified their drinking habit in the past 10 years, the results were almost unchanged, suggesting that reverse causation is an unlikely explanation of our observations, and the prevalence of lower urinary tract symptoms was low. 21 A further advantage is that complete data on PSA levels for all participants allowed estimation of the alcohol-PSA association.
The use of a PSA threshold (3 ng/ml) for initiating the diagnostic process could complicate the interpretation of results. On the one hand, the possibility that some of the controls might have undiagnosed prostate cancer might attenuate the estimated effect. On the other, since alcohol behavior and associated factors could affect PSA levels independently of prostate cancer, this could introduce a bias resulting in apparent changes in risk of PSA-detected prostate cancer, in particular for low-grade tumors, as discussed in the above paragraph. However, the impact of detection bias is likely to be limited for associations between alcohol and risk of high Gleason-grade prostate cancer. Where all individuals undergo PSA testing, like in ProtecT, high-grade tumors will be less affected by a possible effect of alcohol on PSA, since they are usually characterized by high levels of the marker. [bib_ref] Prevalence of prostate cancer among men with a prostate-specific antigen level <¼..., Thompson [/bib_ref] Also, in situations where not everyone is PSA tested, as is the case in many studies from the literature, detection bias is minimized for these tumor sub-types as they are more likely to progress to advanced and metastatic stage and become symptomatic quicker, [bib_ref] 20-year outcomes following conservative management of clinically localized prostate cancer, Albertsen [/bib_ref] [bib_ref] Combining longitudinal studies of PSA, Inoue [/bib_ref] and to be diagnosed regardless of PSA testing. Therefore, associations found between alcohol and high-grade tumors are generally more robust than those for total prostate cancer or low-grade disease.
Limitations of this study include confounding, selection bias, measurement error and to some degree recall bias. Confounding offers a particularly important challenge in the context of prostate cancer, as on the one hand alcohol drinking clusters with many other behaviors detrimental to health, and on the other little is known about environmental and lifestyle causes of prostate cancer, 5 other than, for example, smoking does not seem to be one of these.In this scenario, speculation about residual confounding is common. We attempted adjustment for potentially confounding factors and generally found similar effect estimates; however, only quasi-experimental designs such as Mendelian randomization could reasonably exclude a major impact of confounding on the results. [bib_ref] Mendelian randomization: can genetic epidemiology contribute to understanding environmental determinants of disease?, Smith [/bib_ref] Selection bias might occur if heavy drinking study participants were self-selected and generally healthier (implying lower cancer risk) than nonparticipants, but generally such selection would not apply to moderate drinkers. However, this does not fit in with the observed dose-response association between alcohol and PSA, but would probably result in a scenario with a marked reduction of risk in heavy drinkers only as compared to light drinkers, and similar risk for moderate and light drinkers. If nonparticipation in the ProtecT study depended on previous PSA testing and/or prostatic symptoms, this could introduce another source of bias; however, PSA testing is not common place in the UK. Nondifferential measurement error was likely to be minimized by the choice of light drinkers over current nondrinkers as the reference category to improve internal validity of comparisons between groups of drinkers, and by the fact that participants displayed regular drinking behaviors, as evidenced by similar effects observed for recent and usual drinking across the entire distribution of alcohol intake. As for recall bias, there remains the possibility that men with prostate cancer symptoms reported their alcohol consumption differently from healthier men, or were more/less likely to fill in their questionnaires, although it is yet to be investigated whether in the ProtecT study response rates depended on PSA levels, a proxy for symptoms.
## Impact/implications
The nature of the association between alcohol consumption and prostate cancer is extremely important, as robust evidence on potentially modifiable risk factors for this disease is currently lacking. [bib_ref] Screening for prostate cancer, Frankel [/bib_ref] This is despite the fact that age, ethnicity and genetic/inherited factors can only predict 65% of disease, [bib_ref] Genetic variants and family history predict prostate cancer similar to prostate-specific antigen, Zheng [/bib_ref] whereas international comparisons show higher incidence and mortality in wealthier countries and a positive correlation with elements of Western diets, [bib_ref] Environmental factors and cancer incidence and mortality in different countries, with special..., Armstrong [/bib_ref] and migrant studies suggest that low-risk populations quickly acquire higher prostate cancer incidence rates when moving to highrisk countries. [bib_ref] Cancer incidence patterns among Chinese immigrant populations in Alberta, Luo [/bib_ref] [bib_ref] Cancers of the prostate and breast among Japanese and white immigrants in..., Shimizu [/bib_ref] If quasi-experimental studies can confirm the present results to be causal, prostate cancer risk will be added to the many long-term health risks of heavy drinking. Strategies to reduce the prevalence of drinking over the recommended amount will then also result in a small reduction in highrisk, poorer prognosis prostate cancer. For example, if the 30% of British men in this general population sample who currently drink more than 21 units/week were to cut down to less than 10 units/week, there could be a reduction in high-grade prostate cancer of around 3%, assuming a RR of 1.1 such as the one estimated in the present study. This is a modest effect, however, keeping alcohol consumption low throughout adulthood would be more effective in terms of prostate cancer prevention, given the long latency of prostate cancer and the suggestion that alcohol might act as a tumor initiator through procarcinogens activation, as well as a promoter. [bib_ref] Cocarcinogenic effects of alcohol in hepatocarcinogenesis, Stickel [/bib_ref] Establishing whether alcohol drinking causally influences PSA would help answering the question of how many more or fewer prostate cancer cases would be identified should drinking habits change in the population, at least as long as PSA testing remains widespread. On the other hand, the observation that men drinking heavily have lower PSA than men drinking lightly might have public health implications, even if not causal, provided that it is confirmed by independent studies. In the clinical setting, heavy alcohol consumption could be used as a marker identifying a group of men in which some cancers might be missed. The corresponding increase in mortality could be evaluated in current PSA screening trials, and these results might lead to recommending different PSA thresholds depending on usual and recent alcohol consumption.
# Conclusion
These results support the hypothesis that heavy alcohol drinking causes a small increase in risk of high-grade prostate cancer, and are generalizable to European-origin populations with widespread use of PSA testing. However, no firm conclusion can be reached on the nature of the effect of alcohol on low-grade cancers and on PSA in the absence of independent replication, preferably among large cohorts of men with a variety of grades and stages of prostate cancer.
[fig] Figure 1: Flow-chart of exclusions and final data-sets of study participants-ProtecT nested case-control. [/fig]
[table] Table 1: Characteristics of cases and controls included in the analysis of alcohol and prostate cancer risk Number with complete data.2 Nonmanual occupation includes codes for: professional, managerial, nonmanual, and skilled nonmanual. Manual occupation includes codes for: manual and skilled manual, semiskilled, and unskilled manual.3 Regularly consuming 5þ units/occasion.4 Geometric mean. Abbreviations SD, standard deviation; BMI, body mass index. [/table]
[table] Table 2: Associations of dimensions of alcohol drinking behavior with PSA and prostate cancer risk Test for difference in the effect estimates for localized versus advanced and low-grade versus high-grade prostate cancer.2 Ratios of geometric means and 95% confidence intervals from linear regression, adjusted for the design variables on which controls and cases were matchedadditionally adjusted for age as continuous variable. [/table]
[table] Table 4: Association of alcohol drinking with PSA levels in controls [/table]
[table] Table 3: Association of weekly alcohol consumption with PSA and prostate cancer risk Test for difference in the effect-estimates for localized versus advanced and low-grade versus high-grade prostate cancer.3 Ratios of geometric means and 95% confidence intervals from linear regression, adjusted for the design variables on which controls and cases were matched-additionally adjusted for age as continuous variable.4 Odds ratios and 95% confidence intervals from conditional logistic regression models, additionally adjusted for age as a continuous variable.5 Relative risk ratios and 95% confidence intervals from logistic regression, adjusted for the design variables on which controls and cases were matched-additionally adjusted for age as continuous variable. For definitions of localized/advanced stage and low/high-grade, see methods section. Abbreviations PSA, prostate-specific antigen. [/table]
[table] Table 6: Association of alcohol drinking with (a) PSA levels in controls, stratified according to changes in drinking behavior in the last 10 years Adjusted for age, center and date of recruitment and clinic visit.2 Regularly consuming 5þ units/occasion. Abbreviations PSA, prostate-specific antigen; RGM, ratio of geometric means; CI, confidence interval; N co, number of controls. [/table]
[table] Table 5: Association of alcohol drinking with prostate cancer risk comparing cases and controls [/table]
[bib_ref] Global burden of disease and injury and economic cost attributable to alcohol..., Rehm [/bib_ref] [bib_ref] Global burden of disease and injury and economic cost attributable to alcohol..., Rehm [/bib_ref] |
National, regional, and global causes of mortality in 5–19-year-olds from 2000 to 2019: a systematic analysis
## Webappendix 1. country classification
High-quality VR countries were based on WHO criteria on the quality, completeness of vital registration (VR) data for ages 15 and above, and usability.We used a 10/1000 mortality rate in 2010 from 5 to 20 years of age from UN-IGME 2 as the threshold that separates low and high mortality model countries. The mortality rate between ages 5 and 20 years, denoted as , is defined as the probability of dying between exact ages 5 and 20, expressed per 1,000 population aged 5.We identified percentiles aligned with those used for the country classification for children under-five: [bib_ref] Global, regional, and national causes of under-5 mortality in 2000-15: an updated..., Liu [/bib_ref] 10/1000 is the 70 percentile of for all countries, and the 50 percentile of for all countries excluding those with high-quality VR data in 2010.
Below is the list of the 195 countries for which we provide cause-of-death estimates. The table also includes their regional classification, model category, and whether they have been flagged as a fragile state.These are the same 195 countries for which UN-IGME provides estimates, which corresponds to the 194 WHO member states plus the West Bank and Gaza Strip. 2
## Webappendix 2. cause categorization and icd codes
Specific causes of death (COD) that made up at least 3% of global deaths in 2016 among any of the adolescent 5-year-age-and-sex groups according to the existing estimates 4,5 were considered for modeling. The final causes were influenced by model stability, which was driven by the frequency and magnitude of the cause fractions in the empirical data. Remaining causes were grouped into the respective "other" categories (Other communicable, maternal, perinatal and nutritional conditions -Other CMPN; Other non-communicable diseases -Other NCD; and Other injuries). Diseases of the digestive system K20-K92
Other NCD Remainder of non-communicable diseases
[formula] Injuries V01-Y09, Y35-Y36, Y40-Y86, Y88-Y89, Y871 Road traffic injuries V01-V04, V06, V09-V80, V87, V89, V99 Drownings W65-W74 Natural disaster X30-X39 Interpersonal violence X85-Y09, Y871 [/formula]
Collective violence: legal intervention Y35
Collective violence: war Y36
Self-harm X60-X84, Y871
## Other injuries remainder of injuries
Ill-defined R00-R99, Y10-Y34, Y872
## Webappendix 3. model input data procurement and preparation
## Webappendix 3.1 covariates and their preparation
Each covariate had at least 200 empirical country data points between 1980 and 2017, and covariates with high age-sex resolution were prioritized. Covariates were matched to the study population using the following hierarchy: location, year, age, and sex.
For covariates, empirical data was first taken directly from their source without adjustments. A complete time-series for each country was then generated using linear interpolation between existing empirical data points, extrapolation assuming a flat trend for years preceding or succeeding the last available empirical data points, and imputation based on region and country's lag distributed GDP where there was no empirical data. These time series were used for country-years in the input database that did not have subnational covariate data. A smoothed times series was also generated using a 7-year average for model prediction.
## Webappendix 3.2 cod model inputs preparation
## Systematic review for verbal autopsy studies
We conducted a systematic review for verbal autopsy studies published between January 1, 1980, and December 31, 2017 to be used as model inputs. We used search terms related to COD, ages 5-19, and HMM countries in major search engines covering global health and epidemiology journals (Yeung D, Feng Y, Hong J, et al.; unpublished data). Articles were screened with pre-set inclusion and exclusion criteria and abstracted by two independent researchers using DistillerSR software.We started with 55,471 articles and eventually included 62 in the final analysis.
## Search items for systematic review
We searched on Pubmed, Scopus, Embase, Web of Science, Global Health Index Medicus, Global Health OVID, IndMed, PAHO, Popline, and Cochrane for studies published between January 1, 1980 and December 31, 2017, using the following search terms:
(("Adolescent" OR "adolescents" OR "adolescence" OR "teen" OR "teens" OR "teenager" OR "teenagers" OR "youth" OR "youths" OR "young adult" OR "young adults" OR "young person" OR "young people" OR "preteen" OR "preteens") OR (("child" OR "children") AND ("6 years" OR "7 years" OR "8 years" OR "9 years" OR "10 years" OR "11 years" OR "12 years" OR "13 years" OR "14 years" OR "15 years" OR "school age" OR "age 6" OR "ages 6" OR "age 7" OR "ages 7" OR "age 8" OR "ages 8" OR "age 9" OR "ages 9" OR "age 10" OR "ages 10" OR "age 11" OR "ages 11" OR "age 12" OR "ages 12" OR "age 13" OR "ages 13" OR "age 14" OR "ages 14" OR "age 15" OR "ages 15"))) AND (("Cause of death" OR "causes of death") OR (("cause" OR "causes" OR "autopsy" OR "autopsies" OR "disease burden" OR "survey" OR "surveys" OR "surveillance" OR "register" OR "registers" OR "registration" OR "vital statistics" OR "report" OR "reports") AND ("Mortality" OR "Mortalities" OR "fatal" OR "fatality" OR "fatalities" OR "death" OR "deaths"))) AND ("Afghanistan" OR "Algeria" OR "Angola" OR "Azerbaijan" OR "Bangladesh" OR "Benin" OR "Bhutan" OR "Bolivia" OR "Botswana" OR "Burkina Faso" OR "Burundi" OR "Cambodia" OR "Cameroon" OR "Central African Republic" OR "Chad" OR "China" OR "Comoros" OR "Congo" OR "Cote d'Ivoire" OR "Democratic People's Republic of Korea" OR "Congo" OR "Djibouti" OR "Dominican Republic" OR "Equatorial Guinea" OR "Eritrea" OR "Ethiopia" OR "Gabon" OR "Gambia" OR "Ghana" OR "Guatemala" OR "Guinea" OR "Guinea-Bissau" OR "Haiti" OR "India" OR "Indonesia" OR "Iran" OR "Iraq" OR "Kazakhstan" OR "Kenya" OR "Kiribati" OR "Kyrgyzstan" OR "Laos" OR "Lesotho" OR "Liberia" OR "Madagascar" OR "Malawi" OR "Mali" OR "Marshall Islands" OR "Mauritania" OR "Micronesia" OR "Mongolia" OR "Morocco" OR "Mozambique" OR "Myanmar" OR "Namibia" OR "Nauru" OR "Nepal" OR "Niger" OR "Nigeria" OR "Pakistan" OR "Papua New Guinea" OR "Philippines" OR "Rwanda" OR "Sao Tome and Principe" OR "Senegal" OR "Sierra Leone" OR "Solomon Islands" OR "Somalia" OR "South Africa" OR "South Sudan" OR "Sudan" OR "Swaziland" OR "Tajikistan" OR "Timor-Leste" OR "Togo" OR "Turkmenistan" OR "Uganda" OR "Tanzania" OR "Uzbekistan" OR "Yemen" OR "Zambia" OR "Zimbabwe" OR "developing country" OR "developing countries" OR "developing nation" OR "developing nations" OR "developing population" OR "developing populations" OR "developing world" OR "less developed country" OR "less developed countries" OR "less developed nation" OR "less developed nations" OR "less developed population" OR "less developed populations" OR "less developed world" OR "lesser developed country" OR "lesser developed countries" OR "lesser developed nation" OR "lesser developed nations" OR "lesser developed population" OR "lesser developed populations" OR "lesser developed world" OR "under developed country" OR "under developed countries" OR "under developed nation" OR "under developed nations" OR "under developed population" OR "under developed populations" OR "under developed world" OR "underdeveloped country" OR "underdeveloped countries" OR "underdeveloped nation" OR "underdeveloped nations" OR "underdeveloped population" OR "underdeveloped populations" OR "underdeveloped world" OR "middle income country" OR "middle income countries" OR "middle income nation" OR "middle income nations" OR "middle income population" OR "middle income populations" OR "low income country" OR "low income countries" OR "low income nation" OR "low income nations" OR "low income population" OR "low income populations" OR "lower income country" OR "lower income countries" OR "lower income nation" OR "lower income nations" OR "lower income population" OR "lower income populations" OR "underserved country" OR "underserved countries" OR "underserved nation" OR "underserved nations" OR "underserved population" OR "underserved populations" OR "underserved world" OR "under served country" OR "under served countries" OR "under served nation" OR "under served nations" OR "under served population" OR "under served populations" OR "under served world" OR "deprived country" OR "deprived countries" OR "deprived nation" OR "deprived nations" OR "deprived population" OR "deprived populations" OR "deprived world" OR "poor country" OR "poor countries" OR "poor nation" OR "poor nations" OR "poor population" OR "poor populations" OR "poor world" OR "poorer country" OR "poorer countries" OR "poorer nation" OR "poorer nations" OR "poorer population" OR "poorer populations" OR "poorer world" OR "developing economy" OR "developing economies" OR "less developed economy" OR "less developed economies" OR "lesser developed economy" OR "lesser developed economies" OR "under developed economy" OR "under developed economies" OR "underdeveloped economy" OR "underdeveloped economies" OR "middle income economy" OR "middle income economies" OR "low income economy" OR "low income economies" OR "lower income economy" OR "lower income economies" OR "low gdp" OR "low gnp" OR "low gross domestic" OR "low gross national" OR "lower gdp" OR "lower gnp" OR "lower gross domestic" OR "lower gross national" OR "lmic" OR "lmics" OR "third world" OR "lami country" OR "lami countries" OR "transitional country" OR "transitional countries" OR "Africa" OR "Africa South of the Sahara" OR "Africa, Central" OR "Africa, Eastern" OR "Africa, Southern" OR "Africa, Western" OR "Africa, Northern" OR "Caribbean Region" OR "West Indies" OR "Central America" OR "Latin America" OR "South America" OR "Asia, Central" OR "Asia, Northern" OR "Asia, southeastern" OR "Asia, western" OR "middle east" OR "Asia" OR "far east" OR "Transcaucasia" OR "USSR" OR "Atlantic Islands" OR "Indian Ocean Islands" OR "Pacific Islands" OR "Micronesia" OR "Melanesia" OR "province" OR "provinces" OR "district" OR "districts" OR "prefecture" OR "prefectures" OR "county" OR "counties" OR "municipality" OR "municipalities")
Exclusion criteria for systematic review
The following criteria were used to screen articles captured in the systematic review. Articles were excluded if they fulfilled any of the following:
1. Did not disaggregated data between 5-19-year-olds. 2. Aggregated data beyond 24-year-olds. 3. Did not present cause of death data. 4. Included less than two specific causes of death. 5. Was not conducted in a high-mortality country without high-quality vital registration data. 6. Was not conducted using standardized verbal autopsy methods. 7. Had more than 25% of deaths with undefined causes. 8. Was not conducted in a general population. 9. Completed surveillance before 1980. 10. Did not have a reference period that was 12 months (±1 month) or did not have a study duration of more than 24 months. 11. Did not present primary data.
## Collapsing into study input data points
In both LMM and HMM, datapoints were collapsed hierarchically by year, sex, and/or age, to contain at least 15 deaths and reduce idiosyncratic errors, and retained if they included at least two causes. To limit the influence of extremely large but less informative data points, in the HMM studies were dropped if they included more than 5,000 total deaths or had 25% or more deaths categorized as "undetermined". As a result, 12 data points were dropped among 5-9 (4·9%), 9 among 10-14 (3·6%), and 8 among 15-19 (2·9%). Moreover, data points with more than 1,000 deaths in which one cause represented more than 50% of total deaths were also dropped, but this only applied to 4 datapoints for 15-19 males with large fractions of interpersonal violence.
## Vr data used in the hmm
To enhance model stability, we included 30 high-quality VR data points in the HMM input. In particular, we identified countries with high-quality VR data in which the all-cause mortality rate between ages 5 and 20 ( ) was above 10 per 1,000 population (threshold between LMM and HMM countries) at some point between 2000-2019.Next, for each of these countries, we selected the last year with empirical high-quality VR data in which ≥ 10. The table below summarizes the high-quality VR data points included in the HMM input.
## The bayesian lasso
A thorough description of the Bayesian LASSO model used here can be found elsewhere. [bib_ref] A Bayesian hierarchical model with integrated covariate selection and misclassification matrices to..., Mulick [/bib_ref] A brief overview is provided in the following.
Suppose there exist mutually exclusive causes of death, and that we have a sample of deaths from a given study , each of which is (correctly) classified into one and only one of the categories. If we denote the distribution of true COD in the sample as , , , , … , , and if the sample is random, we can assume that these observations come from a multinomial distribution, where , refers to the probability that a death is due to cause in the population in which study is conducted.
Suppose the probabilities , can be predicted by the values of a set of explanatory variables , , , , … , , . In a multinomial regression framework, we assume that the logarithm of the odds of each cause of death relative to a reference cause are linearly dependent on these explanatory variables. This is expressed as a system of − 1 linear equations corresponding to each cause of death (excluding the reference category , ), log , / , = , + , , + , , + ⋯ + , , log , / , = , + , , + , , + ⋯ + , , ⋮ log , / , = , + , , + , , + ⋯ + , , Note that the -coefficients (including the intercepts) do not have the study subindex . This is a fixedeffects model that assumes the associations of the explanatory variables with the causes of death are constant across all studies. We relax this assumption by adding study-specific normally distributed random effects to the intercepts, with mean 0 and cause-specific standard deviations. These standard deviations have uniformly distributed priors bounded between 0 and parameter .
We implemented LASSO covariate selection by penalizing large -coefficients in a subset of the fixedeffect parameters that could potentially result in overfitting the data. We did this by imposing a double exponential (also referred to as Laplace) prior distribution on them in the model specification. [bib_ref] The Bayesian Lasso, Park [/bib_ref] This shrinks the magnitude of the parameters without completely reducing them to zero, and has the additional advantage of stabilizing the model if convergence is slow or difficult. The intercepts and any -coefficient we did not want to be constrained in the LASSO were given non-informative normally distributed priors. The remaining -coefficients had a Laplace prior with mean 0 and precision > 0, the penalty imposed by the LASSO method. We used out-of-sample cross-validation to select the optimal and parameters (see .
Once the model has estimated the -coefficients and study-specific random effects, we estimated the expected distribution of true causes of death in any country for which we have covariate data as
[formula] , = exp , + × 1 + exp , + × + ⋯ + exp , + × , [/formula]
where , denotes the cause-and study-specific random effect, is a vector with -coefficients for cause , and is the vector of covariates in study .
## The base category
For each age-sex group and model, selection of the base cause was guided by its global burden and input data availability. Specifically, we selected the base category by identifying the cause with the largest number of deaths in the input database.
## Webappendix 4.2 estimation of the sex-specific deaths and rates for 15-19
Country-level estimates on all-cause mortality (envelopes) were borrowed from the United Nations Inter-agency Group for Child Mortality Estimation (UN-IGME).For the period 1990-2019 and ages 5 to 24, UN-IGME provides annual estimates on the number of deaths and mortality rates for 195 countries in 5-and 10-year age groups. [bib_ref] Global, regional, and national mortality trends in youth aged 15-24 years between..., Masquelier [/bib_ref] However, UN-IGME has yet to publish sex-specific mortality envelopes for ages 5+. We used sexspecific life tables from the United Nations (UN)to obtain all-cause sex-specific mortality rates and number of deaths for the 15-19 age group as follows:
1. For each country-year available, we recovered from the UN life tables the central death rates from 15 to 19 years for both sexes together ( ), for males ( ) and females ( ). Using these rates, for all country-years we calculated a male ratio = / and a female ratio = / . 2. We transformed the UN-IGME probabilities of dying between ages 15 and 19 ( ) into central death rates by applying the formula = /(5 × (1000− × 0.5)) , which assumes that deaths are equally distributed within each of the corresponding age intervals. Note that UN-IGME reports probabilities of dying, even though they refer to them as 'mortality rates', 2 which justifies this intermediate step.
3. Next, we applied the ratios and calculated from the UN life tables to the estimated central death rates from UN-IGME ( ), obtaining the sex-specific death rates = × and = × . 4. Finally, we back transformed these rates into sex-specific probabilities of dying, given by = 5000 × /(1 + 2.5 × ) and = 5000 × /(1 + 2.5 × ).
Similarly, we used the population shares by age and sex from UN World Population Prospects (UN-WPP)to calculate the sex-specific number of deaths in the 15-19 age group.
5. For each country-year, we divided the reported number of deaths ( ) from UN-IGME by the estimated death rates defined above, obtaining the population at risk for both sexes = / . 6. Let and denote the country-and year-specific shares by sex from UN-WPP for the 15-19 age group. We calculated the sex-specific population at risk for each country-year as = × and = × . 7. Finally, we estimated the sex-specific number of deaths for the 15-19 age group multiplying the sex-specific populations at risk by the sex-specific death rates from UN-IGME estimated above:
[formula] = × and = × . [/formula]
## Webappendix 4.3 point estimates and uncertainty
We run 10,000 iterations of the eight Bayesian models (LMM and HMM in each of the four age-sex groups) in four parallel chains. To assess convergence, we used an initial burn-in sequence of 4,000 and a thinning interval of 20 and calculated potential scale reduction factor.Hence, for each of the model parameters-and for each of the eight models-we got 1,204 sets of estimates after burn-in. We used these sets to calculate mean values of the multinomial regression parameters and obtain point estimates of the cause-specific mortality fractions, mortality rates and death counts. To estimate uncertainty, we used the 1,204 sets of multinomial regression parameters from the Bayesian model to obtain 1,204 sets of mortality estimates for each country-year and calculate uncertainty intervals. More specifically, for each draw we 1. Calculated mortality fractions for all countries in the period 2000-2019; 2. Randomly drew all-cause mortality estimates from the posterior distribution of the mortality envelopes; 2 3. Incorporated single-cause estimates. For each country-year, values of the single-cause estimates were drawn from the corresponding uncertainty intervals, with the following assumptions: a. For HIV/AIDS, we assumed data were normally distributed but truncated at 0, to avoid negative values. We used the point estimates as the means an estimated the standard deviations by dividing the range of the corresponding 95% uncertainty intervals by 3.92. b. For TB and measles, the upper bounds of the corresponding 95% uncertainty intervals tended to be large, indicating skewed distributions with long upper tails. To account for that, we drew values from log-normal distributions with matching means and standard deviations. Let and denote the lower and upper bounds of the 95% uncertainty intervals, respectively, and the point estimate. We defined = ( − )/3.92, and used standard formulae to obtain the usual log-normal parameters mean and variance , given by = log √ + and = log 1 + .
## Webappendix 4.4 transparency and replicability
We carried out our analyses using the open-source statistical software Rand Bayesian modelling was implemented in JAGS 15 with wrapper functions from the R2jags package.The source code, primary inputs and cause of death data collected and estimated are publicly available for research purposes from the GitHub repository https://github.com/panchoVG/Mort5to19. Measles estimates were taken from WHO Immunization, Vaccines and Biologicals Department. [bib_ref] Assessment of the 2010 global measles mortality reduction goal: results from a..., Simons [/bib_ref] These estimates assume zero deaths for ages 10 years and older. For 5-9 year-olds, we split measles deaths into endemic and epidemic deaths in countries where the measles caused at least 5% of total deaths in any year between 2000-2019 or more in this age group. Endemic measles was identified by fitting either a log-linear or loess model to the number of measles deaths and were accounted for within the UN-IGME all-cause mortality estimates, whereas epidemic measles was the difference between the total and endemic measles deaths, and were added outside the UN-IGME all-cause mortality estimates. Then, they were recombined to get the final measles cause of death fractions presented here.
## Data inputs
For all data inputs from multiple sources that are synthesized as part of the study:
3
Describe how the data were identified and how the data were accessed. Methods
## 4
Specify the inclusion and exclusion criteria. Identify all ad-hoc exclusions. Methods; Webappendix 3
## 5
Provide information on all included data sources and their main characteristics.
For each data source used, report reference information or contact name/institution, population represented, data collection method, year(s) of data collection, sex and age range, diagnostic criteria or measurement method, and sample size, as relevant.
Methods; Webappendix 3
6
Identify and describe any categories of input data that have potentially important biases (e.g., based on characteristics listed in item 5).
# Methods; webappendix 3
For data inputs that contribute to the analysis but were not synthesized as part of the study:
## 7
Describe and give sources for any other data inputs.
## Methods; webappendices 3 and 6
For all data inputs:
8
Provide all data inputs in a file format from which data can be efficiently extracted (e.g., a spreadsheet rather than a PDF), including all relevant meta-data listed in item 5. For any data inputs that cannot be shared because of ethical or legal reasons, such as third-party ownership, provide a contact name or the name of the institution that retains the right to the data.
Primary inputs and cause of death data collected are publicly available from the GitHub repository https://github.com/panchoVG/Mort5to19.
# Data analysis 9
Provide a conceptual overview of the data analysis method. A diagram may be helpful.
Methods; Webappendix 5
## 10
Provide a detailed description of all steps of the analysis, including mathematical formulae. This description should cover, as relevant, data cleaning, data pre-processing, data adjustments and weighting of data sources, and mathematical or statistical model(s).
Methods; Webappendices 4 and 6
## 11
Describe how candidate models were evaluated and how the final model(s) were selected.
Methods; Webappendix 9
12
Provide the results of an evaluation of model performance, if done, as well as the results of any relevant sensitivity analysis. Webappendix 9
## 13
Describe methods for calculating uncertainty of the estimates. State which sources of uncertainty were, and were not, accounted for in the uncertainty analysis.
Methods; Webappendix 4
## Webappendix 9. model selection process
Identification of LASSO precision parameter through cross-validation by estimation methods. Based on out-of-sample prediction with a random effect standard deviation set at = 0 · 07, the best , defined as the one that gives the least root mean squared error in the out-of-sample prediction, is marked red in the figures below. Webappendix 10. Global and regional cause-specific mortality estimates by age-sex group, 2000-2019
Webappendix 10.1 Data files with global and regional cause-specific mortality estimates
The following CSV (comma separated value) files with mortality estimates are available from the GitHub repository https://github.com/panchoVG/Mort5to19.
## Mortality estimates without uncertainty (point estimates)
'PointEstimates5to9-Regional.csv': Global and regional all-cause number of deaths, all-cause mortality rates, and cause-specific mortality fractions for 5-9 years, 2000-2019. 'PointEstimates10to14-Regional.csv': Global and regional all-cause number of deaths, allcause mortality rates, and cause-specific mortality fractions for 10-14 years, 2000-2019. 'PointEstimates15to19-Regional.csv': Global and regional all-cause number of deaths, allcause mortality rates, and cause-specific mortality fractions for 15-19 years, 2000-2019. 'PointEstimates5to19-Regional.csv': Global and regional all-cause number of deaths, all-cause mortality rates, and cause-specific mortality fractions for 5-19 years, 2000-2019.
Mortality estimates with 95% uncertainty intervals 'AllCauseUncert-Regional.csv': Global and regional all-cause number of deaths and all-cause mortality rates with 95% uncertainty intervals, 5-9, 10-14 and 15-19 years, 2000-2019. 'Uncertainty5to9-Regional.csv': Global and regional cause-specific number of deaths, mortality rates, and mortality fractions with 95% uncertainty intervals for 5-9 years, 2000-2019. 'Uncertainty10to14-Regional.csv': Global and regional cause-specific number of deaths, mortality rates, and mortality fractions with 95% uncertainty intervals for 10-14 years, 2000-2019. 'Uncertainty15to19-Regional.csv': Global and regional cause-specific number of deaths, mortality rates, and mortality fractions with 95% uncertainty intervals for 15-19 years, 2000-2019. 'Uncertainty5to19-Regional.csv': Global and regional cause-specific number of deaths and mortality fractions with 95% uncertainty intervals for 5 to 19 years, 2000-2019. 'GlobalAARR.csv': Global cause-specific annual average rate of reduction (AARR) 2000-2019 with 95% uncertainty intervals, 5-9, 10-14 and 15-19 years.
Webappendix 10.2 Global and regional leading causes of death 5-19 years in 2019
## Webappendix 10.3 global and regional causes of mortality fractions by age-sex group in 2019
The following table contains the data for from the main manuscript. Time trends of the cause-specific mortality rates for the three countries with highest burden in 2019.
[fig] Webappendix 1: Country classification .................................................................................. 3 Webappendix 2. Cause categorization and ICD codes ......................................................... 7 [/fig]
[fig] A00- B99 ,: D50-D53, D64·9, E00-E02, E40-E46, E50-E68, G00, G03-G04, H65-H66, J00-J22, N70-N73, O00-O99, P00-C97, D00-D48, D55-D89 (except D64.9), E03-E07, E10-E34, E65-E88, F01-F99, G06-G98, H00-H61, H68-H93, I00-I99, J30-J98, K00-K92, L00-L98, M00-M99, N00-N64, N75-N98, [/fig]
[fig] Webappendix 5: Modeling and estimation strategies VR: Vital Registration; VA: Verbal Autopsy; LASSO: Least Absolute Shrinkage and Selection Operator; CSMF: Cause-Specific Mortality Fraction; UN-IGME: United Nations Inter-agency Group for Child Mortality Estimation Webappendix 6. Additional details on single cause estimates Webappendix 6.1 Splitting measles into endemic and epidemic [/fig]
[table] Table S1: List of the 195 countries and their classification [/table]
[table] Table S2: Cause list and the ICD code mapping [/table]
[table] Table S3: 2. Vital registration data used in high-mortality model [/table]
[table] Table S4: 1. Base category for each age-sex group and model in the Bayesian LASSO [/table]
[bib_ref] Trends in mild, moderate, and severe stunting and underweight, and progress towards..., Stevens [/bib_ref] [bib_ref] The MERRA-2 aerosol reanalysis, 1980 onward. Part II: Evaluation and case studies, Buchard [/bib_ref] [bib_ref] A Bayesian hierarchical model with integrated covariate selection and misclassification matrices to..., Mulick [/bib_ref] [bib_ref] The Bayesian Lasso, Park [/bib_ref] [bib_ref] Global, regional, and national mortality trends in youth aged 15-24 years between..., Masquelier [/bib_ref] [bib_ref] A program for analysis of Bayesian graphical models using Gibbs sampling, Plummer [/bib_ref] [bib_ref] Assessment of the 2010 global measles mortality reduction goal: results from a..., Simons [/bib_ref] |
Challenging Occam’s Razor: Dual Molecular Diagnoses Explain Entangled Clinical Pictures
# Introduction
The ultimate goal of Medical Genetics is the identification of the molecular cause of genetic disorders. These conditions, albeit singularly rare, present a population prevalence of 3.5-5.9% [bib_ref] Estimating Cumulative Point Prevalence of Rare Diseases: Analysis of the Orphanet Database, Nguengang Wakap [/bib_ref] , thus posing a major burden from a medical, social, and economical point of view. In this light, achieving a correct molecular diagnosis has a positive impact not only on patients and their families but also on the healthcare system. In fact, understanding the genetic basis of a given disorder could implement specific indications for correct follow-up and potential treatment, avoid ineffective medical care, provide prognosis information, and identify familial recurrence risk.
Usually, medical geneticists hypothesise clinical diagnoses of Mendelian diseases by recognising peculiar syndromic patterns. As a consequence, whenever signs and symptoms do not precisely fit into a known model, either an apparently new condition or a phenotypic expansion of a single disorder are usually considered [bib_ref] Resolution of Disease Phenotypes Resulting from Multilocus Genomic Variation, Posey [/bib_ref]. However, the introduction of highthroughput sequencing technologies, such as Whole Exome Sequencing (WES), that allow the simultaneous analysis of thousands of genes, have highlighted another fascinating possibility: the coexistence of dual molecular diagnoses [bib_ref] Resolution of Disease Phenotypes Resulting from Multilocus Genomic Variation, Posey [/bib_ref] [bib_ref] Clinical Whole-Exome Sequencing for the Diagnosis of Mendelian Disorders, Yang [/bib_ref] [bib_ref] Molecular Diagnostic Experience of Whole-Exome Sequencing in Adult Patients, Posey [/bib_ref]. These are defined as the presence of pathogenic variants at two distinct and independently segregating loci that cause two different Mendelian conditions [bib_ref] Clinical Whole-Exome Sequencing for the Diagnosis of Mendelian Disorders, Yang [/bib_ref]. Since the introduction of WES in the clinical setting, several works have reported the identification of subjects affected by dual molecular diagnoses, with an overall rate of 5-7% [bib_ref] Resolution of Disease Phenotypes Resulting from Multilocus Genomic Variation, Posey [/bib_ref] [bib_ref] Clinical Whole-Exome Sequencing for the Diagnosis of Mendelian Disorders, Yang [/bib_ref] [bib_ref] Molecular Diagnostic Experience of Whole-Exome Sequencing in Adult Patients, Posey [/bib_ref].
In consideration of the growing number of reports of double or even triple genetic diagnoses [bib_ref] Triple Diagnosis of Wiedemann-Steiner, Waardenburg and DLG3-Related Intellectual Disability Association Found by..., Matis [/bib_ref] [bib_ref] Dual Diagnosis of Ellis-van Creveld Syndrome and Hearing Loss in a Consanguineous..., Vona [/bib_ref] [bib_ref] 22q and Two: 22q11.2 Deletion Syndrome and Coexisting Conditions, Cohen [/bib_ref] [bib_ref] HDAC8 Loss of Function and SHOX Haploinsufficiency: Two Independent Genetic Defects Responsible..., Severi [/bib_ref] [bib_ref] Novel MED12 Variant in a Multiplex Fragile X Syndrome Family: Dual Molecular..., Lahbib [/bib_ref] [bib_ref] Complex Presentation of Hao-Fountain Syndrome Solved by Exome Sequencing Highlighting Co-Occurring Genomic..., Priolo [/bib_ref] [bib_ref] Dual Diagnosis of Osteogenesis Imperfecta (OI) and Short Stature and Advanced Bone..., Ye [/bib_ref] , increasing attention must be paid to entangled clinical pictures, and atypical features should prompt awareness of possible multiple causes. This is of the utmost importance to provide patients with the best and safest standards of care, allowing for a more personalised and appropriate clinical management [bib_ref] 22q and Two: 22q11.2 Deletion Syndrome and Coexisting Conditions, Cohen [/bib_ref].
To further investigate the peculiar scenario of multilocus genomic variation, here we present the identification of patients affected by dual molecular diagnoses, within a cohort of 342 patients who underwent WES analysis in the last 24 months.
# Materials and methods
# Ethical statement
All the analyses have been performed following relevant guidelines and regulations. Written informed consent was obtained from all participants or their legal guardians. The study was conducted in accordance with the tenets of the Helsinki Declaration and was approved by the Ethics Committee of the I.R.C.C.S. "Burlo Garofolo" of Trieste.
## Clinical evaluation
In the last 24 months, 342 unrelated patients with complex clinical pictures and suspicion of underlying genetic disorders have been referred to the Medical Genetics Unit of the I.R.C.C.S. "Burlo Garofolo" (Trieste, Italy).
All participants were characterised through a detailed anamnesis, a dysmorphological examination and further investigations. A familial anamnesis has been collected in order to identify possible other affected family members and a personal medical history has been obtained to highlight potential confounding factors (i.e., infections, trauma, or other non-genetic causes of congenital abnormalities). A physical examination has been carried out to identify dysmorphic features, with particular attention to facial, ectodermal, skeletal, and genital features. Ancillary clinical tests have been performed when appropriate and included electroencephalogram (EEG), brain Magnetic Resonance Imaging (MRI) and Computed Tomography (CT), electrocardiogram (ECG) and echocardiography, abdominal ultrasound, neurological, ophthalmological, otorhinolaryngoiatric, and cardiological evaluations.
## Dna extraction and quality control
Genomic DNA was extracted from peripheral whole blood samples of patients and, whenever available, both their parents using the QIAsymphony ® SP instrument with QIAsymphony ® Midi Kit (Qiagen, Venlo, The Netherlands), following manufacturer's instructions. DNA quality was verified with 1% agarose gel electrophoresis and its concentration was measured using Nanodrop ND 1000 spectrophotometer (NanoDrop Technologies Inc., Wilmington, DE, USA).
## Wes and data analysis
WES was performed on an Illumina NextSeq 550 instrument (Illumina Inc., San Diego, CA, USA) using the Twist Human Core Exome and Human RefSeq Panel kit (Twist Bioscience, South San Francisco, CA, USA), according to the manufacturer's protocol. Briefly, genomic DNA was enzymatically fragmented, ligated to a universal adapter and amplified using the Unique Dual Index primers (Twist Bioscience, South San Francisco, CA, USA). Samples were therefore hybridised with the Twist Human Core Exome and the Human RefSeq Panel kit, which cover 99% of protein coding genes. Hybridised fragments have been captured, amplified, and sequenced. Sequencing coverage statistics for all samples and, in detail, for cases affected by dual molecular diagnoses are reported in [fig_ref] Table 1: List of dual molecular diagnoses identified through WES [/fig_ref].
The process allows the production of FASTQ files that were analysed through a custom pipeline (Germline-Pipeline), developed by enGenome s.r.l. (https://www.engenome. com/) [fig_ref] Figure 1: Dual molecular diagnoses patterns [/fig_ref]. This pipeline permits the identification of germline variants, including Single Nucleotide Variants (SNVs), short Insertions/Deletions (INDELs), and exon-level Copy Number Variations (CNVs) starting from sequence reads. The secondary analysis leads to the generation of final VCF files that contains SNVs, INDELs and CNVs. VCF files were analysed through the enGenome Expert Variant Interpreter (eVai) software (https://evai.engenome.com/), that allows variant annotation, interpretation, and prioritisation, exploiting both artificial intelligence and the American College of Medical Genetics and Genomics/Association for Molecular Pathology (ACMG/AMP) guidelines to analyse and classify genomic variants [bib_ref] Standards and Guidelines for the Interpretation of Sequence Variants: A Joint Consensus..., Richards [/bib_ref].
Variant frequency was verified both in NCBI dbSNP build 155 (https://www.ncbi.nlm. nih.gov/snp/) and gnomAD (https://gnomad.broadinstitute.org/) to exclude variants previously reported as polymorphisms. Pathogenicity of already-reported variants was assessed through The Human Gene Mutation Database (https://www.hgmd.cf.ac.uk/ac/ index.php) and ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/). All databases have been accessed lastly on 9 October 2022. The effect of all identified variants was evaluated through several in silico prediction tools, as PolyPhen-2 [bib_ref] Predicting Functional Effect of Human Missense Mutations Using PolyPhen-2, Adzhubei [/bib_ref] , Sorting Intolerant From Tolerant (SIFT) [bib_ref] Predicting Amino Acid Changes That Affect Protein Function, Ng [/bib_ref] , Pseudo Amino Acid Protein Intolerance Variant Predictor (PaPI score) [bib_ref] Pseudo Amino Acid Composition to Score Human Protein-Coding Variants, Limongelli [/bib_ref] , Deep Neural Network Variant Predictor (DANN score), and dbscSNV score [bib_ref] Silico Prediction of Splice-Altering Single Nucleotide Variants in the Human Genome, Jian [/bib_ref]. SNVs leading to synonymous amino acid substitutions not predicted as damaging, not affecting splicing or highly conserved residues were excluded; furthermore, variants with a quality score (QUAL) < 20 or called in off-target regions were excluded as well .
Variants were discussed within a multidisciplinary team to verify whether they could be possibly matched to each patient's phenotypic characteristics; finally, all variants of interest were confirmed by Sanger sequencing. Familial segregation was also performed by Sanger sequencing.
# Protein-protein interaction analysis
The Ingenuity Pathway Analysis (IPA) system (version 81348237, Ingenuity Systems; QIAGEN Inc., Hilden, Germany-https://digitalinsights.qiagen.com/IPA, (accessed on 21 October 2022)) was used to identify relationships between genes of interest based on associated functions and data mining from experimental studies reported in the literature. In particular, My Pathways-Build-PathExplorer tool was used for the protein-protein interaction analysis, taking into account only direct interactions and excluding non-coding genes from the analysis.
# Results
Three hundred and forty-two unrelated patients underwent WES analysis in the last 24 months and a molecular diagnosis has been achieved for 138 of them (40.4%). Seven patients presented a dual molecular diagnosis (5.1% of all solved cases), and their clinical and molecular characteristics are detailed in the following paragraphs.
After a careful examination of both their clinical and molecular features, patients affected by a dual molecular diagnosis may be divided into two different groups [fig_ref] Figure 1: Dual molecular diagnoses patterns [/fig_ref] , [fig_ref] Table 1: List of dual molecular diagnoses identified through WES [/fig_ref] :
A. Patients who present distinct phenotypes, due to each of the two different underlying genetic diseases. Five patients belong to this category (i.e., Patients 1, 2, 3, 4 and 5). B. Patients with overlapping clinical features that may be underpinned by both the identified genetic variations. Two patients belong to this group (i.e., Patients 6 and 7). . Dual molecular diagnoses with distinct phenotypes: patients present a blended phenotype due to the simultaneous presence of two Mendelian disorders caused by variants at two independently segregating loci. Each condition is characterised by a different and specific set of signs and symptoms, and they may be recognised as discrete upon reverse phenotyping. (B). Dual molecular diagnoses with overlapping phenotypes: patients' clinical picture may be ascribed to either of the two underlying Mendelian disorders. In the exemplified cases, one condition is inherited from both patients in an autosomal recessive manner, and the other has a de novo origin in the proband; nevertheless, all inheritance patterns are possible. . Dual molecular diagnoses with distinct phenotypes: patients present a blended phenotype due to the simultaneous presence of two Mendelian disorders caused by variants at two independently segregating loci. Each condition is characterised by a different and specific set of signs and symptoms, and they may be recognised as discrete upon reverse phenotyping. (B). Dual molecular diagnoses with overlapping phenotypes: patients' clinical picture may be ascribed to either of the two underlying Mendelian disorders. In the exemplified cases, one condition is inherited from both patients in an autosomal recessive manner, and the other has a de novo origin in the proband; nevertheless, all inheritance patterns are possible. [fig_ref] Figure 2: Family pedigrees of patients affected by dual molecular diagnoses [/fig_ref] is a 20-day-old girl born after an apparently uneventful pregnancy. indicate each of the seven patients described in this study. Half black filling of father and mother indicates healthy carrier parents. Three-quarters black filling of proband indicates that he/she is affected by an autosomal recessive condition. One-quarter or half chessboard filling of proband or parent indicates that subject is affected by an autosomal dominant disorder. Half diagonal-line filling of proband or parent indicates that subject is affected by another autosomal dominant disorder. Half vertical-line filling of a male proband indicates that he is affected by an X-linked disorder. Half vertical-line filling of a female subject indicates a healthy carrier of an X-linked condition.
## Father
WES was carried out on the family trio and revealed the presence of two novel likely pathogenic variants in two different genes. The first involves the SCN1A gene, known to be associated with several autosomal dominant epileptic disorders (e.g., Dravet syndrome-MIM: #607208-, Developmental and epileptic encephalopathy 6b-MIM: #619317-, and Generalized epilepsy with febrile seizures plus, type 2-MIM: #604403). The second involves MMP21, whose biallelic mutations cause Autosomal visceral heterotaxy 7 (MIM: #616749), that includes the transposition of the great arteries.
As regards SCN1A (NM_001165963.1), the c.2591_2593delTGC, p.(Leu864del) heterozygous inframe deletion is predicted as damaging by the in silico tool PaPI score, is not reported in any public database, and familial segregation confirmed its de novo origin in the proband.
As regards MMP21 (NM_147191.1), the c.903G>A, p.(Met301Ile) homozygous missense variant affects a highly conserved residue, is predicted as damaging by all in silico tools employed in the analysis, and is reported with a very low frequency in the gnomAD database (Minor Allele Frequency (MAF): 0.0016%). Familial segregation confirmed that both parents are heterozygous carriers and, as expected, do not present any clinical fea- At birth, facial asymmetry, hands abnormalities (i.e., short and enlarged thumbs bilaterally and partial duplication of the distal phalanx of the fifth finger), and feet syndactyly were identified, thus prompting a genetic evaluation. The clinical examination highlighted marked facial and cranial asymmetry, anteverted nares, low-set ears, exaggerated Cupid's bow, downturned oral commissures, and pointed chin. Furthermore, the newborn showed extensive xerosis cutis with multiple desquamative areas. A head CT scan highlighted a complex craniosynostosis with ossification of the right coronal hemi-suture and of both parietotemporal sutures; accordingly, a marked asymmetry of cerebral hemispheres volume and right eye proptosis were identified.
WES was carried out on the proband's and both her parents' DNA and data analysis highlighted the presence in the girl of a missense heterozygous variant in the FGFR2 gene (NM_000141.4) (c.940G>T, p.(Ala314Ser)). Pathogenic variants in FGFR2 are associated with several autosomal dominant diseases characterised by overlapping clinical features, as Apert syndrome (MIM: #101200), Crouzon syndrome (MIM: #123500) and Saethre-Chotzen syndrome (MIM: #101400), that are collectively known as FGFR2-associated craniosynostoses. The identified variant has already been described as pathogenic [bib_ref] The Mutations in FGFR2-Associated Craniosynostoses Are Clustered in Five Structural Elements of..., Steinberger [/bib_ref] and familial segregation confirmed its de novo origin in the proband.
Furthermore, we identified two compound heterozygous nonsense variants in the FLG gene (NM_002016.1). Pathogenic variants in FLG cause Ichthyosis Vulgaris (MIM: #146700), both autosomal dominant and autosomal recessive. Both variants (c.7339C>T, p.(Arg2447*) and c.3191G>A, p.(Trp1064*)) have already been reported in association with this phenotype [bib_ref] Comprehensive Analysis of the Gene Encoding Filaggrin Uncovers Prevalent and Rare Mutations..., Sandilands [/bib_ref] [bib_ref] Improving the Diagnostic Yield for Filaggrin: Concealed Mutations in the Dutch Population, Van Leersum [/bib_ref].
## Patient 2: epilepsy and congenital heart disease
Patient 2 [fig_ref] Figure 2: Family pedigrees of patients affected by dual molecular diagnoses [/fig_ref] is an eight-year-old girl that came to the geneticist's attention due to congenital heart malformation (i.e., levo-transposition of the great arteries and pulmonary valve stenosis identified at five months of age), drug-resistant epilepsy (i.e., tonic-clonic seizures, firstly appeared at four months of age), and intellectual disability. Her family history is unremarkable, with the exception of the maternal grandmother who is reported to having experienced tonic-clonic seizures from the age of seven to the age of 14.
WES was carried out on the family trio and revealed the presence of two novel likely pathogenic variants in two different genes. The first involves the SCN1A gene, known to be associated with several autosomal dominant epileptic disorders (e.g., Dravet syndrome-MIM: #607208-, Developmental and epileptic encephalopathy 6b-MIM: #619317-, and Generalized epilepsy with febrile seizures plus, type 2-MIM: #604403). The second involves MMP21, whose biallelic mutations cause Autosomal visceral heterotaxy 7 (MIM: #616749), that includes the transposition of the great arteries.
As regards SCN1A (NM_001165963.1), the c.2591_2593delTGC, p.(Leu864del) heterozygous inframe deletion is predicted as damaging by the in silico tool PaPI score, is not reported in any public database, and familial segregation confirmed its de novo origin in the proband.
As regards MMP21 (NM_147191.1), the c.903G>A, p.(Met301Ile) homozygous missense variant affects a highly conserved residue, is predicted as damaging by all in silico tools employed in the analysis, and is reported with a very low frequency in the gnomAD database (Minor Allele Frequency (MAF): 0.0016%). Familial segregation confirmed that both parents are heterozygous carriers and, as expected, do not present any clinical features associated with Autosomal visceral heterotaxy 7.
## Patient 3: syndromic intellectual disability and hypertrophic cardiomyopathy
Patient 3 [fig_ref] Figure 2: Family pedigrees of patients affected by dual molecular diagnoses [/fig_ref] is an 18-year-old boy affected by severe intellectual disability, spastic tetraplegia, and non-progressive hypertrophic cardiomyopathy. The boy presents severe neurodevelopmental delay: he acquired the sitting position at three years of age, started to walk at four years old and lost this ability at 12 years of age, and never learnt to talk. Furthermore, the proband is affected by conductive hearing loss, bilateral strabismus, and bilateral hip dysplasia. From a phenotypical point of view, he displays deeply set eyes, thick upper lip vermillion, widely spaced teeth, macroglossia, and mandibular prognathism.
As a single-gene defect underlying all his clinical features was suspected, in trio WES was performed. However, data analysis highlighted the presence of two pathogenic variants in two distinct genes. In particular, the boy carries a homozygous missense variant already described as causative of Intellectual developmental disorder, autosomal recessive 58 (MIM: #617270) in the ELP2 gene (NM_001242875.1) (c.1580G>A, p.(Arg527Gln)) [bib_ref] A Novel ELP2 Compound Heterozygous Mutation in a Boy with Severe Intellectual..., Turkyilmaz [/bib_ref] , and a heterozygous frameshift variant inherited from the father and already reported in association with Hypertrophic cardiomyopathy (MIM: #115197) in MYBPC3 (NM_000256.3) (c.913_914delTT, p.(Phe305Profs*27)) [bib_ref] Myofilament Protein Gene Mutation Screening and Outcome of Patients with Hypertrophic Cardiomyopathy, Olivotto [/bib_ref]. Upon the genetic finding, the father underwent a detailed cardiological examination that did not reveal the presence of a hypertrophic cardiomyopathy, possibly in accordance with the incomplete penetrance and variable expressivity that characterise this disorder.
3.1.4. Patient 4: Syndromic Intellectual Disability and Congenital Heart Disease Patient 4 [fig_ref] Figure 2: Family pedigrees of patients affected by dual molecular diagnoses [/fig_ref] is 37-year-old woman who sought genetic counselling because of a complex congenital heart malformation and intellectual disability. Her heart malformation had already been detected in the prenatal period and was characterised by interatrial and interventricular defects together with tricuspid valve atresia. In addition, she presented an arrhythmic phenotype, with atrial tachycardia as its main feature. Furthermore, the proband is affected by mild intellectual disability, submucosal cleft palate, moderate bilateral sensorineural hearing loss, and mild facial dysmorphisms (i.e., synophrys, thick upper and lower vermillion, and right earlobe malformation). Concerning the woman's family history, the mother is also reported to present mild intellectual disability and the same minor facial features.
WES of the family trio revealed the presence in both the proband and the mother of two novel likely pathogenic variants at the heterozygous state, one in the ZMIZ1 gene (NM_020338.3) and the other in the DSG2 gene (NM_001943.3). Variants in ZMIZ1 have been associated with autosomal dominant Neurodevelopmental disorder with dysmorphic facies and distal skeletal anomalies (MIM: #618659), which partially overlaps the clinical features of the proband. The identified missense variant (c.1984A>G, p.(Asn662Asp)) affects a highly conserved residue, is predicted as damaging by the in silico tools used during the analysis, and is not reported any public database. As regards DSG2, variants in this gene are known to cause autosomal dominant Arrhythmogenic right ventricular dysplasia 10 (MIM: #610193). WES data analysis showed the presence of a nonsense variant, c.621_626delTCCTCC, p.(Tyr207_Pro209delinsTer), predicted as damaging by the in silico tool PaPI score and already reported as pathogenic in ClinVar (RCV000544613). In consideration of this finding, a specific cardiological follow-up has been performed also in the proband's mother: an ECG, echocardiogram, and cardiac MRI have been performed and resulted all normal. This may reflect the incomplete and age-dependent penetrance of this disorder, that is estimated to be around 30-50% [bib_ref] Arrhythmogenic Right Ventricular Cardiomyopathy-Associated Desmosomal Variants Are Rarely De Novo, Van Lint [/bib_ref].
## Patient 5: multiple congenital malformations and autism spectrum disorder
Patient 5 [fig_ref] Figure 2: Family pedigrees of patients affected by dual molecular diagnoses [/fig_ref] is a four-year-old boy who came to the geneticist's attention due to the presence of anal atresia, identified at birth and surgically treated, lumbosacral transitional vertebra, recognised in the neonatal period via spine radiography and confirmed by spine MRI, and autism spectrum disorder (ASD), diagnosed at two years of age. A careful clinical evaluation revealed only some mild facial dysmorphisms (i.e., protruding and crumpled ears) and a wide toe bilaterally.
WES was then performed and revealed the presence in the proband of the c.460C>T, p.(Arg154*) heterozygous nonsense variant in the BMP2 gene (NM_001200.2). Pathogenic variants in BMP2 are associated with autosomal dominant Short stature, facial dysmorphism, and skeletal anomalies with or without cardiac anomalies 1 (MIM: #617877) and the identified variant has already been reported as causative of this phenotype [bib_ref] Monoallelic BMP2 Variants Predicted to Result in Haploinsufficiency Cause Craniofacial, Skeletal, and..., Tan [/bib_ref]. WES data analysis suggested that the variant was paternally inherited, but Sanger sequencing segregation did not confirm its presence in the father's DNA. WES data was examined again with particular attention to the variant's coverage in both the proband and the father: in the former, the alternative allele had a frequency of 50%, thus being compatible with the heterozygous state, whereas in the latter, the alternative allele had a frequency of 26%, thus prompting the suspicion of mosaicism. An allele-specific PCR was performed on the father's DNA and allowed to detect both the wild-type allele and the mutated one, hence confirming the mosaic hypothesis.
In addition, a novel inframe deletion has been detected in the INTS6L gene (NM_182540.4). Variants in this gene have recently been described in two studies reporting patients affected by X-linked ASD and developmental disorder [bib_ref] The Contribution of X-Linked Coding Variation to Severe Developmental Disorders, Martin [/bib_ref] , thus designating INTS6L as a new candidate for this type of diseases. The identified variant (c.2552_2554delACA, p.(Asn851del)) is predicted as damaging by the in silico tool PaPI score and is not reported in any public database. Segregation analysis confirmed that the variant was maternally inherited and present at the hemizygous state in the proband. To our knowledge, Patient 5 is the third reported subject that presents a neurodevelopmental disorder and also carries a variant within the INTS6L gene. In this light, the identification of further patients together with the implementation of functional studies are mandatory to corroborate this genotype-phenotype correlation and confirm the specific role of this gene in neurodevelopmental disorders. [fig_ref] Figure 2: Family pedigrees of patients affected by dual molecular diagnoses [/fig_ref] is a two-years-old boy with a history of drug-resistant epilepsy, firstly appeared at two months of age. Video-EEG recording coupled to Electromyography (EMG) polygraphy recorded several nonfebrile ictal episodes, characterised by loss of consciousness, fixed stare, and variable motor manifestations, thus fostering the initiation of anticonvulsant therapy, that, however, resulted ineffective.
## Patients who present
WES analysis of the trio revealed the presence in the proband of two variants in epilepsy-associated genes, SCN1A (NM_001202435.1) and CSNK2B (NM_001320.5). As regards SCN1A, the splice-site c.695-1G>A variant was identified at the heterozygous state; it affects a splice acceptor site, is predicted as damaging by the in silico tools DANN and db-scSNV scores, and has already been reported as possibly associated with Dravet syndrome (MIM: #607208) [bib_ref] Clinical and neuroimaging features of acute encephalopathy after status epilepticus in Dravet..., Tian [/bib_ref]. Conversely, variants in CSNK2B have recently been associated with the Poirier-Bienvenu neurodevelopmental syndrome (MIM: #618732), a neurological disorder characterised by early-onset and possibly refractory seizure, intellectual disability of various degree, and ASD. The c.384_394delAGGTGAAGCCA, p.(Pro128fs) frameshift variant was identified in the proband at the heterozygous state and has recently been reported as pathogenic in a small cohort of Poirier-Bienvenu patients [bib_ref] Expanding Phenotype of Poirier-Bienvenu Syndrome: New Evidence from an Italian Multicentrical Cohort..., Orsini [/bib_ref]. Familial segregation analysis confirmed that both variants originated de novo in the proband.
## Patient 7: epileptic encephalopathy
Patient 7 [fig_ref] Figure 2: Family pedigrees of patients affected by dual molecular diagnoses [/fig_ref] is a four-year-old girl affected by epileptic encephalopathy. She has a positive family history, since her mother reported to suffer from tonic-clonic seizure from the age of 12 and one of her mother's sisters presents epilepsy and intellectual disability.
The proband has experienced her first absence seizure at five months of age: brain MRI highlighted a thinning of the arcuate fasciculus that connects the Broca and Wernicke's areas, and the EEG revealed the presence of interictal epileptiform discharges, thus fostering the diagnosis of epileptic encephalopathy. Furthermore, the girl also presented a complex neurodevelopmental disorder, characterised by ataxic gait, global developmental delay with severe speech impairment, ASD, and sleep disorder with frequent awakenings. A careful clinical evaluation revealed mild dysmorphisms, as a square face and small ears.
DNA of the proband's father was not available since he passed away prior to the analysis, therefore WES was performed only on the girl's DNA and identified two heterozygous variants in two different epilepsy-associated genes, CACNB4 (NM_000726.5) and ZEB2 (NM_014795.4). The missense variant in CACNB4 (c.1418G>A, p.(Arg473His)) has already been reported as causative of Generalised idiopathic epilepsy (MIM: #607682). The novel missense variant in ZEB2 (c.905G>A, p.(Arg302Gln)) affects a highly conserved residue, is predicted as damaging by all in silico tools employed in the analysis, is not present in the gnomAD database, and has been reported three times as Variant of Uncertain Significance (VUS) in ClinVar (RCV000717537, RCV001209725, RCV000159444). Additionally, a different variant affecting the same residue (c.904C>T, p.(Arg302*)) has already been reported in the literature in patients affected by Mowat-Wilson syndrome (MIM: #235730) [bib_ref] Clinical and Molecular Analysis of Mowat-Wilson Syndrome Associated with ZFHX1B Mutations and..., Ishihara [/bib_ref] and epilepsy and neurodevelopmental disorders [bib_ref] Diagnostic Outcomes for Genetic Testing of 70 Genes in 8565 Patients with..., Lindy [/bib_ref]. Segregation of both variants was performed in the proband's mother and did not identify any of them.
# Discussion
The introduction in the clinical setting of high-throughput sequencing technologies, such as WES, has enormously implemented our ability to achieve a precise molecular diagnosis even in the most complex cases. WES offers the possibility to simultaneously analyse all protein-coding genes, thus covering the majority of known pathogenic variants and providing a detailed analysis of the genetic makeup of a patient in a timely and costeffective manner [bib_ref] A Three-Year Follow-up Study Evaluating Clinical Utility of Exome Sequencing and Diagnostic..., Fung [/bib_ref]. The huge amount of available data also opens up the opportunity to apply a genotype-first approach to identify a molecular diagnosis and this appears particularly useful in patients with entangled clinical pictures. Indeed, whenever a patient's signs and symptoms do not fit into a known syndromic pattern, two different possibilities are generally considered: a phenotypic expansion, with more severe or new characteristics added to a well-recognised condition, or the identification of an apparently novel disease [bib_ref] Phenotypic Expansion Illuminates Multilocus Pathogenic Variation, Karaca [/bib_ref]. Syndromes present a multisystemic and multiorgan involvement, but not all patients affected by the same condition display identical characteristics and equal severity. As a consequence, a certain interindividual variability and a clinical heterogeneity are usually expected. In this light, whenever a patient displays the peculiar features of a known syndrome but also some additional phenotypic characteristics, geneticists usually consider the latter as an ancillary clinical manifestation of the specific disorder, thus identifying a phenotypic expansion [bib_ref] The Genetic Basis of Mendelian Phenotypes: Discoveries, Challenges, and Opportunities, Chong [/bib_ref]. On the other hand, whenever patients present an association of signs and symptoms that has never been reported before, the clinical suspicion of a novel genetic disorder arises. In these cases, WES grants the possibility to simultaneously analyse a huge amount of data and, as already discussed, possibly identify new disease-causing genes never implicated in Mendelian disorders before [bib_ref] De Novo Truncating Variants in ASXL2 Are Associated with a Unique and..., Shashi [/bib_ref].
In this context, the opportunity to screen the entire exome in a hypothesis-free manner has highlighted another compelling scenario: the presence of dual molecular diagnoses [bib_ref] Resolution of Disease Phenotypes Resulting from Multilocus Genomic Variation, Posey [/bib_ref] [bib_ref] Clinical Whole-Exome Sequencing for the Diagnosis of Mendelian Disorders, Yang [/bib_ref] [bib_ref] Phenotypic Expansion Illuminates Multilocus Pathogenic Variation, Karaca [/bib_ref] [bib_ref] A Retrospective Review of Multiple Findings in Diagnostic Exome Sequencing: HalF.A.Re Distinct..., Smith [/bib_ref]. Although patients affected by multiple Mendelian disorders are being increasingly recognised and several case reports have recently described patients with multilocus genetic variation [bib_ref] Dual Diagnosis of Ellis-van Creveld Syndrome and Hearing Loss in a Consanguineous..., Vona [/bib_ref] [bib_ref] HDAC8 Loss of Function and SHOX Haploinsufficiency: Two Independent Genetic Defects Responsible..., Severi [/bib_ref] [bib_ref] Novel MED12 Variant in a Multiplex Fragile X Syndrome Family: Dual Molecular..., Lahbib [/bib_ref] [bib_ref] Dual Diagnosis of Osteogenesis Imperfecta (OI) and Short Stature and Advanced Bone..., Ye [/bib_ref] [bib_ref] Dual Molecular Diagnosis Contributes to Atypical Prader-Willi Phenotype in Monozygotic Twins, Jehee [/bib_ref] [bib_ref] More than an 'Atypical' Phenotype: Dual Molecular Diagnosis of Autoimmune Lymphoproliferative Syndrome..., Saettini [/bib_ref] [bib_ref] Dual Molecular Diagnosis of Tricho-Rhino-Phalangeal Syndrome Type I and Okur-Chung Neurodevelopmental Syndrome..., Xu [/bib_ref] , this occurrence is still often overlooked. Indeed, there is a lack of comprehensive studies that systematically investigate the possible presence of pathogenic variants in multiple genes in large cohorts of unselected patients. In this light, our study underlines how a thorough and unbiased analysis of WES data could lead to the identification, especially in complex patients, of more than one genetic disorder, thus fostering geneticists to be aware of such occurrence and be prone to consider this hypothesis whenever examining a subject with atypical features.
Our study confirms that the identification of patients affected by dual molecular diagnoses is not a rare event: in our cohort they represent 5.1% of all molecularly solved cases, in line with reported literature data [bib_ref] Resolution of Disease Phenotypes Resulting from Multilocus Genomic Variation, Posey [/bib_ref] [bib_ref] Clinical Whole-Exome Sequencing for the Diagnosis of Mendelian Disorders, Yang [/bib_ref] [bib_ref] Molecular Diagnostic Experience of Whole-Exome Sequencing in Adult Patients, Posey [/bib_ref]. Our patients may be divided into two different groups: A) subjects presenting distinct phenotypes due to two different variants in two independently segregating loci; B) subjects presenting overlapping characteristics that may be explained by either of the two identified variants. Regarding the first category, it is worth noting that only one patient was suspected ab initio to be affected by two different diseases. Indeed, Patient 1 presented a Syndromic craniosynostosis that prompted a gestalt diagnosis of Apert syndrome. However, upon a detailed analysis of the clinical features associated with this condition, it emerged that her other peculiar sign, namely xerosis cutis, is not part of the well-established characteristics of FGFR2-associated craniosynostosis. This awareness suggested to further evaluate WES data to verify the possible presence of pathogenetic variants in ichthyosis-associated genes. Only coupling a detailed phenotypical characterisation with a deep analysis of sequencing data, it was possible to identify in this patient both molecular diagnoses. Concerning all other patients belonging to group A, the first clinical hypothesis was always a single-gene defect that could explain all signs and symptoms, considering that usually, in medicine, a single pathogenetic cause is more common than multiple ones [bib_ref] The Diagnostic Approach in Complex Patients: Parsimony or Plenitude?, Kelly [/bib_ref]. This "parsimony principle" has indeed been challenged by Patients 2, 3, 4, and 5, whose clinical characteristics have eventually been explained by two different variants in two different genes, both responsible of distinct syndromes that were properly recognised only upon reverse-phenotyping.
patients. In this light, our study underlines how a thorough and unbiased analysis of WES data could lead to the identification, especially in complex patients, of more than one genetic disorder, thus fostering geneticists to be aware of such occurrence and be prone to consider this hypothesis whenever examining a subject with atypical features.. Diagnostic work-up of patients presenting complex clinical pictures. The first step consists in a detail clinical characterisation that is achieved through an anamnesis collection, a dysmorphological examination, and ancillary laboratory and imaging investigations. This leads to the formulation of a diagnostic hypothesis, that could immediately consider the possible presence of a dual molecular diagnosis or, on the contrary, contemplate the existence of a single syndromic condition.
Our study confirms that the identification of patients affected by dual molecular diagnoses is not a rare event: in our cohort they represent 5.1% of all molecularly solved cases, in line with reported literature data [bib_ref] Resolution of Disease Phenotypes Resulting from Multilocus Genomic Variation, Posey [/bib_ref] [bib_ref] Clinical Whole-Exome Sequencing for the Diagnosis of Mendelian Disorders, Yang [/bib_ref] [bib_ref] Molecular Diagnostic Experience of Whole-Exome Sequencing in Adult Patients, Posey [/bib_ref]. Our patients may be divided into two different groups: A) subjects presenting distinct phenotypes due to two different variants in two independently segregating loci; B) subjects presenting overlapping characteristics that may be explained by either of the two identified variants. Regarding the first category, it is worth noting that only one patient was suspected ab initio to be affected by two different diseases. Indeed, Patient 1 presented a Syndromic craniosynostosis that prompted a gestalt diagnosis of Apert syndrome. However, upon a detailed analysis of the clinical features associated with this condition, it emerged that her other peculiar sign, namely xerosis cutis, is not part of the well-established characteristics of FGFR2-associated craniosynostosis. This awareness suggested to further evaluate WES data to verify the possible presence of pathogenetic variants in ichthyosis-associated genes. Only coupling a detailed phenotypical characterisation with a deep analysis of sequencing data, it was possible to identify in this patient both molecular diagnoses. Concerning all other patients belonging to group A, the first clinical hypothesis was always a single-gene defect that could explain all signs and symptoms, considering that usually, in medicine, a single pathogenetic cause is more common than multiple ones [bib_ref] The Diagnostic Approach in Complex Patients: Parsimony or Plenitude?, Kelly [/bib_ref]. This "parsimony principle" has indeed been challenged by Patients 2, 3, 4, and 5, whose clinical characteristics have eventually been explained by two different variants in two different genes, both responsible of distinct syndromes that were properly recognised only upon reverse-phenotyping.
As regards group B, it is interesting to notice how both patients presented with epilepsy, albeit of different types. Epilepsy is one of the most common neurological disorders and is characterised by a high clinical and genetic heterogeneity, with more than 900 genes associated with this phenotype. In this light, if a patient is found to be a carrier of more than one variant in more than one epilepsy-associated gene, it could be extremely challenging to precisely define the role and the contribution of all identified variants in epileptogenesis [bib_ref] Improving Molecular Diagnosis in Epilepsy by a Dedicated High-Throughput Sequencing Platform, Mina [/bib_ref] [bib_ref] Diagnostic Exome Sequencing in Non-Acquired Focal Epilepsies Highlights a Major Role of..., Krenn [/bib_ref]. As a consequence, in order to delve deeper into the possible underlying mechanisms, an in silico pathway analysis has been performed to assess proteinprotein interactions. In Patient 6, affinity chromatography and sequencing-on-chip assays . Diagnostic work-up of patients presenting complex clinical pictures. The first step consists in a detail clinical characterisation that is achieved through an anamnesis collection, a dysmorphological examination, and ancillary laboratory and imaging investigations. This leads to the formulation of a diagnostic hypothesis, that could immediately consider the possible presence of a dual molecular diagnosis or, on the contrary, contemplate the existence of a single syndromic condition.
As regards group B, it is interesting to notice how both patients presented with epilepsy, albeit of different types. Epilepsy is one of the most common neurological disorders and is characterised by a high clinical and genetic heterogeneity, with more than 900 genes associated with this phenotype. In this light, if a patient is found to be a carrier of more than one variant in more than one epilepsy-associated gene, it could be extremely challenging to precisely define the role and the contribution of all identified variants in epileptogenesis [bib_ref] Improving Molecular Diagnosis in Epilepsy by a Dedicated High-Throughput Sequencing Platform, Mina [/bib_ref] [bib_ref] Diagnostic Exome Sequencing in Non-Acquired Focal Epilepsies Highlights a Major Role of..., Krenn [/bib_ref]. As a consequence, in order to delve deeper into the possible underlying mechanisms, an in silico pathway analysis has been performed to assess protein-protein interactions. In Patient 6, affinity chromatography and sequencingon-chip assays showed that the proteins encoded by both SCN1A and CSNK2B directly interact with the protein encoded by the KMT2A gene [bib_ref] Neuronal Kmt2a/Mll1 Histone Methyltransferase Is Essential for Prefrontal Synaptic Plasticity and Working..., Jakovcevski [/bib_ref]. The latter is involved in the methylation of histone H3 lysine 4 (H3K4), which mediates chromatin remodelling. Variants in the KMT2A gene are associated with Wiedemann-Steiner syndrome (MIM: #605130), a neurodevelopmental disorder characterised by intellectual disability, hypertrichosis cubiti, short stature, and typical facial features. It is interesting to note that also patients affected by this condition may present epilepsy. As a consequence, it may be argued that all three genes belong to a common molecular pathway involved in epileptogenesis. Additionally, in Patient 7, affinity chromatography assay showed that the proteins encoded by both CACNB4 and ZEB2 directly interact with the protein encoded by the TNIK gene [bib_ref] Spatiotemporal Profile of Postsynaptic Interactomes Integrates Components of Complex Brain Disorders, Li [/bib_ref]. TNIK encodes a serine/threonine kinase that activates the Wnt signalling pathway. The canonical Wnt/β-catenin pathway is necessary for processes involved in early brain development and its dysregulation has been implicated in several neurological disorders, including epilepsy. It could therefore be hypothesised that both CACNB4 and ZEB2 may be involved in the same epileptogenic pathway through Wnt signalling alteration. Indeed, the activation of the Wnt/β-catenin pathway has been associated to seizure-induced changes in the brain, as hippocampal neurogenesis, apoptosis pathway activation, hippocampal sclerosis, and mossy fibers sprouting [bib_ref] Wnt/β-Catenin Signaling as a Potential Target for Novel Epilepsy Therapies, Hodges [/bib_ref]. However, further functional in vitro and in vivo studies are mandatory, both to verify the pathogenicity of previously unreported variants (e.g., the novel missense variant in ZEB2 identified in Patient 7) and to verify the exact molecular mechanisms underlying these predicted interactions towards the determination of the epileptic phenotype.
An additional interesting consideration concerns the inheritance patterns of multilocus variations in our patients. In line with literature data [bib_ref] Resolution of Disease Phenotypes Resulting from Multilocus Genomic Variation, Posey [/bib_ref] , pathogenic or likely pathogenic variants in autosomal dominant (AD) disease genes were the most common, being identified in ten out of 14 diagnoses (71.43%). Among them, four variants where inherited from a parent (40%), including the peculiar case of Patient 5, whose father was a mosaic carrier of a known pathogenic variant in the BMP2 gene and did not display any associated clinical features. Additionally, four variants in AD genes (40%) had a de novo origin in the proband, supporting the hypothesis that private variants are a major cause of genetic disorders [bib_ref] Clan Genomics and the Complex Architecture of Human Disease, Lupski [/bib_ref]. Finally, for two variants in genes associated with AD disorders (20%), it was not possible to determine the pattern of inheritance: they were both identified in the same proband, whose father's DNA was not available for familial segregation analysis. In three out of 14 cases (21.43%), a disease-causing variant has been identified in autosomal recessive disease genes: in one case two compound heterozygous variants were detected (Patient 1) whereas in two cases homozygous variants were recognised (Patients 2 and 3). In the last two families, consanguinity was not overtly reported but, in both cases, the proband's parents declared to come from the same small town. Finally, only one diagnosis (7.14%) could be ascribed to an X-linked disease gene (Patient 5), highlighting the relevance of maternally inherited X-linked variants in the etiopathogenesis of Neurodevelopmental disorders [bib_ref] The Contribution of X-Linked Coding Variation to Severe Developmental Disorders, Martin [/bib_ref].
Overall, as it indeed appears from the cases reported in this study, the presence of dual molecular diagnoses represents a challenge to the clinician, since complex phenotypes might be misleading and drive to a wrong interpretation. In this light, only a detailed analysis of patients' clinical features together with a careful examination of WES data can provide the correct molecular diagnosis. It is therefore mandatory to analyse sequencing data within a multidisciplinary team that gathers all healthcare professionals involved in the clinical management, as geneticists, cardiologists, neurologists, otorhinolaryngologists, ophthalmologists, and radiologists. Delivering integrated care is also fundamental to establish the most appropriate follow-up and treatment, especially in neonatal and paediatric patients, since the presence of multiple molecular diagnoses might have an impact on medication response and interactions, [bib_ref] 22q and Two: 22q11.2 Deletion Syndrome and Coexisting Conditions, Cohen [/bib_ref]. Lastly, the presence of more than one genetic condition demands a careful examination of all at-risk family members to offer appropriate counselling and precisely define familial recurrence risk, thus guaranteeing an informed and responsible family planning to all involved subjects.
# Conclusions
In conclusion, our study underlines the importance of challenging the classical concept of "one explanation covers them all" whenever trying to identify the molecular defect underlying complex phenotypical presentations. Dual molecular diagnoses represent a fascinating model of complex inheritance and should always be suspected when patients present atypical clinical features, thus fostering a deeper and more detailed analysis of sequencing data. This is because the identification of a correct molecular diagnosis is the indisputable starting point for a truly personalised and safe clinical management.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/genes13112023/s1, [fig_ref] Figure 1: Dual molecular diagnoses patterns [/fig_ref] : Bioinformatic pipeline for the selection of the final list of variants; [fig_ref] Table 1: List of dual molecular diagnoses identified through WES [/fig_ref] : Coverage statistics; : Variant calling and filtering. Funding: This research was supported by D70-RESRICGIROTTO, BENEFICENTIA Stiftung and Ricerca Corrente L3_14\22 to G.G. and by the Italian Ministry of Health, through the contribution given to the Institute for Maternal and Child Health I.R.C.C.S. "Burlo Garofolo"-Trieste, Italy (RCR-2020-23670068_001).
## Institutional review board statement:
The study was conducted in accordance with the tenets of the Helsinki Declaration and was approved by the Ethics Committee of the I.R.C.C.S. "Burlo Garofolo" of Trieste.
Informed Consent Statement: Written informed consent was obtained from all participants involved in the study, or their legal guardians.
Data Availability Statement: Data presented in this study are available upon request to the corresponding author. Data are not publicly available due to privacy restriction.
[fig] Figure 1: Dual molecular diagnoses patterns. (A) [/fig]
[fig] Figure 2: Family pedigrees of patients affected by dual molecular diagnoses. The figure illustrates the main clinical features of each patient and the genes associated with their phenotypes. (A-G) [/fig]
[fig] Figure 3: Diagnostic work-up of patients presenting complex clinical pictures. The first step consists in a detail clinical characterisation that is achieved through an anamnesis collection, a dysmorphological examination, and ancillary laboratory and imaging investigations. This leads to the formulation of a diagnostic hypothesis, that could immediately consider the possible presence of a dual molecular diagnosis or, on the contrary, contemplate the existence of a single syndromic condition. [/fig]
[table] Table 1: List of dual molecular diagnoses identified through WES. [/table]
|
Concurrent and Subsequent Co-Infections of Clostridioides difficile Colitis in the Era of Gut Microbiota and Expanding Treatment Options
Citation: Trunfio, M.; Scabini, S.; Rugge, W.; Bonora, S.; Di Perri, G.; Calcagno, A. Concurrent and Subsequent Co-Infections of Clostridioides difficile Colitis in the Era of Gut Microbiota and Expanding Treatment Options. Microorganisms 2022, 10, 1275. https://doi.org/ 10.3390/microorganisms10071275 Academic Editor: Lynne V. Mcfarland
# Introduction
Clostridioides difficile (previously known as Clostridium difficile, Cdiff) is a spore-forming, obligate anaerobe, Gram-positive bacterium found in the intestinal tract of both humans and animals, from where its spores are shed in the environment and survive in variable and extreme conditions [bib_ref] Clostridium Difficile-Associated Diarrhea: A Review, Mylonakis [/bib_ref]. Cdiff is well recognized as one of the main causes of healthcareassociated (HA) diarrhea, and this "superbug" has recently emerged also as a less common cause of community-acquired diarrhea in younger individuals lacking traditional HA risk factors. The European Centre for Disease Prevention and Control reported in 2016 a total of 7711 cases of Cdiff infection (CDI) from 556 hospitals in Europe, of which 5765 (74.6%) were HA infections. In the US, after an initial increase in Cdiff incidence, a reduction in HA-Cdiff cases was observed between 2011 and 2017 [bib_ref] Antimicrobial Use in European Acute Care Hospitals: Results from the Second Point..., Plachouras [/bib_ref]. An important rate reduction in HA-CDI was also observed during the COVID-19 pandemic in parallel with the implementation of contact and droplet preventive measures [bib_ref] Impact of the Coronavirus Disease 2019 (COVID-19) Pandemic on Nosocomial Clostridioides difficile..., Ponce-Alonso [/bib_ref]. However, in the last may predispose to microbial translocation (defined as the migration of bacteria, fungi, and/or their products from the gut lumen to extraintestinal space and systemic circulation), blood-stream infections (BSIs), or reactivation of other gut pathogens, as summarized in [fig_ref] Figure 1: Schematic representation of predisposing factors and mechanisms underlying bacterial, fungal, and viral... [/fig_ref]. Highlighting common pathogenic pathways may be helpful in improving the clinical management of CDI by prompt recognition and management of such co-infections. We searched for articles indexed in Embase, MedLine, and PubMed and published in English up to January 2022 by using the following keywords: "Clostridium/Clostridioides difficile", "co-infection", "blood-stream infection", "fungemia", "Candida", "Cytomegalovirus", "probiotics", and "microbial translocation". In this narrative review, we provided a summary of the existing evidence on concurrent and subsequent co-infections associated with CDI trying to describe the overall burden and mechanisms of this phenomenon, to assess whether co-infections have a relevant impact on the outcomes and management of patients suffering from CDI, and to provide some practical clinical considerations.
## Bacterial blood-stream infections and clostridioides difficile
It is well established that the hematogenous translocation of bacteria residing in the gut is favored by some conditions: the loss of integrity of the intestinal mucosal barrier, the alterations of mucosal immunity, and the colonization of gut by overgrowing pathogens. During CDI, all these mechanisms take place along with an important mucosal inflammatory response [bib_ref] Clostridium Difficile Toxin B Causes Epithelial Cell Necrosis through an Autoprocessing-Independent Mechanism, Chumbler [/bib_ref]. Cdiff toxins A and B, which are primarily responsible for tissue damage and associated symptoms, stimulate inflammatory responses in the colonic lining by inducing cytoskeletal changes that compromise the epithelial barrier and stimulate inflammatory cytokines production. The disruption of enterocyte tight junctions allows toxins to cross the epithelium, where they can further induce immune responses in the cells residing in the lamina propria, leading to marked neutrophil recruitment and further destruction of the intestinal lining; indeed, the final pathological hallmark of CDI is the formation of pseudomembranes [fig_ref] Figure 1: Schematic representation of predisposing factors and mechanisms underlying bacterial, fungal, and viral... [/fig_ref] [bib_ref] Clostridium Difficile Toxins: Mediators of Inflammation, Shen [/bib_ref].
While human data are scarce, murine models have already described a potential role of all these Cdiff-induced intestinal alterations in favoring microbial translocation and subsequent BSIs [bib_ref] Interleukin-22 Regulates the Complement System to Promote Resistance against Pathobionts after Pathogen-Induced..., Hasegawa [/bib_ref]. It has been shown that the development of CDI may be favored by perturbations in gut microbiota which in turn is further altered by CDI, starting a vicious circle [fig_ref] Figure 1: Schematic representation of predisposing factors and mechanisms underlying bacterial, fungal, and viral... [/fig_ref]. In this setting, Bacteroidetes and Bifidobacterium spp. play an important role in the mechanism of resistance to Cdiff colonization. Lower concentrations of Bacteroidetes and higher relative amounts of Firmicutes and Proteobacteria were found in the gut of patients with CDI compared to controls [bib_ref] Microbe-Microbe Interactions during Clostridioides difficile Infection, Abbas [/bib_ref]. Similarly, few studies have shown how Cdiff can alter the composition of gut microbiota and promotes colonization with multidrug-resistant (MDR) organisms. Furthermore, MDR organisms' selection, gut dysbiosis, and eventually microbial translocation and BSIs can also be favored by some of the predisposing mechanisms leading to CDI, such as specific patterns of residing gut microbiota and antibiotics administration, amplifying the vicious circle that links CDI to BSIs [fig_ref] Figure 1: Schematic representation of predisposing factors and mechanisms underlying bacterial, fungal, and viral... [/fig_ref]. To further aggravate this event, even the use of anti-Cdiff therapies may contribute to alterations in the intestinal flora, to the overgrowth of bacterial populations, and, eventually, to the facilitation of bacterial translocation, microbe dissemination, and sepsis [fig_ref] Figure 1: Schematic representation of predisposing factors and mechanisms underlying bacterial, fungal, and viral... [/fig_ref] [bib_ref] Fidaxomicin Preserves the Intestinal Microbiome during and after Treatment of Clostridium Difficile..., Louie [/bib_ref].
Despite all these possible mechanisms involved in BSI development, the incidence and impact of BSIs complicating CDI have not been properly characterized to date. In [fig_ref] Table 1: Summary of literature studies reporting period prevalence data on bacterial and/or fungal... [/fig_ref] , we have summarized the available studies from the literature that reported specific period prevalence of either or both bacterial and fungal blood-stream infections after CDI. Falcone et al. described for the first time a significant association between CDI and subsequent HA-BSI [bib_ref] Risk Factors and Outcomes for Bloodstream Infections Secondary to Clostridium Difficile Infection, Falcone [/bib_ref]. In this retrospective analysis of 393 cases, 18.3% developed HA-BSI within 30 days from the onset of CDI. BSIs were caused by enteric pathogens (Candida spp, Enterobacteriaceae, and Enterococcus spp.), and 68.4% of the microbial isolates were MDR. Thirty-day mortality was significantly higher in the CDI plus BSI group compared to the CDI-only group (38.9% vs. 13.1%), as well as the incidence of intensive care unit (ICU) admission and longer hospitalization length. Higher oral vancomycin dosage (>500 mg/day), infection by Cdiff ribotype 027, Cdiff recurrence, and severe colitis were found to be independent risk factors for HA-BSI [bib_ref] Risk Factors and Outcomes for Bloodstream Infections Secondary to Clostridium Difficile Infection, Falcone [/bib_ref].
Similarly, in a study investigating non-staphylococcal BSIs in relation to the time from the first Cdiff-positive fecal sample, bacteremia from unrecognized sources (occult BSI) occurred more frequently from 3 days before to 10 days after Cdiff toxin positivity compared to the pre-Cdiff period. Of note, during the Cdiff period, positive blood cultures were characterized by a greater percentage of enterococci (50%), and the majority of occult BSI resolved without treatment [bib_ref] Bacteraemia from an Unrecognized Source (Occult Bacteraemia) Occurring during Clostridium Difficile Infection...., Thomas [/bib_ref]. In line with these findings, CDI has been identified as a risk factor for vancomycin-resistant enterococci (VRE) bacteremia in a small cohort (n = 59) of patients with acute leukemia [bib_ref] Clostridium Difficile Infection Is a Risk Factor for Bacteremia Due to Vancomycin-Resistant..., Roghmann [/bib_ref]. On the other side, it is not clear whether VRE gut colonization increases the risk of Cdiff colitis, or it can be favored by vancomycin-based treatments for CDI. To date, we can only observe that Cdiff co-colonization (and eventually co-infection) is more common in patients with VRE infection/colonization [bib_ref] Clostridium Difficile and Vancomycin-Resistant Enterococcus: The New Nosocomial Alliance, Poduval [/bib_ref] [bib_ref] Gut Colonization with Vancomycin-Resistant Enterococcus and Risk for Subsequent Enteric Infection, Axelrad [/bib_ref].
Conversely, another group observed only 86 cases of BSIs in a cohort of 570 patients with CDI (7.6%). Enterococcus and Klebsiella spp. were the most common bacterial isolates (14% for both) [bib_ref] Is Clostridium Difficile Infection a Risk Factor for Subsequent Bloodstream Infection?, Ulrich [/bib_ref]. Patients with BSIs showed a higher prevalence of comorbidities, and they were more likely to be immunosuppressed, critically ill, and to have a central venous catheter (CVC) in place. Surprisingly, CDI appeared protective against subsequent BSIs at the multivariate model after adjusting for gender, Charlson Comorbidity score, systemic inflammatory response syndrome, and CVC. The authors hypothesized that systemic immune activation and inflammation triggered by colitis itself could favor the clearance of blood-stream pathogens. However, although the large sample, this study was limited by the retrospective design, the length of follow up limited by the average hospital stay, and the possibility of very complex multifactorial confounders, since they did not rule out infectious sources other than the gastrointestinal tract despite the presence of staphylococcal BSI and CVC [bib_ref] Is Clostridium Difficile Infection a Risk Factor for Subsequent Bloodstream Infection?, Ulrich [/bib_ref]. In opposition to this hypothesis, Oliva et al. found that among 45 subjects hospitalized for CDI, of whom 17.7% developed BSIs, markers of microbial translocation, inflammation, and intestinal damage were increased during CDI, decreased after treatment, and did not normalize compared to healthy controls after CDI resolution. Subjects developing BSIs had higher microbial translocation and maintained it at higher degree after CDI resolution compared to those not complicating with BSIs, suggesting that local and systemic inflammation associate with intestinal barrier disruption, microbial translocation, and eventually increased risk of co-infections.
Cdiff bacteremia is also possible, and in the case of a prominent intestinal barrier injury as the main underlying mechanism for BSI co-occurrence, it should be expected at a similar rate to BSIs. Nevertheless, in the literature, cases of BSIs directly due to Cdiff isolated in blood cultures are extremely rare and solely reported in patients with underlying relevant gastrointestinal disorders. In these cases, BSIs are mixed bacterial co-infections with Cdiff and other gut bacteria, suggesting that massive intestinal barrier dysfunction is required for the translocation of Cdiff. Indeed, it is also possible that being Cdiff an anaerobe, some BSIs diagnoses may miss this blood co-infection due to a relatively higher difficulty in cultivating this germ [bib_ref] Bacteremia Due to Clostridium Difficile-Review of the Literature, Libby [/bib_ref].
Mortality in CDI is likely multifactorial. It is possible that concurrent bacterial translocation contributes to this, but it is hard to precisely estimate its effect, considering that BSIs seem to mainly complicate the more severe CDI only. Considering that both CDI infection and subsequent BSIs represent complex interactions between pathogens, host, native microbiota, and its perturbations and that most risk factors for CDI and BSI due to gut translocation overlap, further studies are needed to investigate this relationship and plan strategies to prevent BSIs in high-risk patients. In the latest guidelines on the management of Cdiff in adults released by the Infectious Diseases Society of America (IDSA) and the Society for Healthcare Epidemiology of America (SHEA), as well as in the latest from the European Society of Clinical Microbiology and Infectious Diseases (ESCMID), fidaxomicin is suggested as the preferred agent for initial CDI and for the first recurrent episode [bib_ref] European Society of Clinical Microbiology and Infectious Diseases: 2021 Update on the..., Van Prehn [/bib_ref] [bib_ref] Clinical Practice Guideline by the Infectious Diseases Society of America (IDSA) and..., Johnson [/bib_ref]. Fidaxomicin significantly reduces the recurrence rate in most patients compared to vancomycin, while it is non-inferior in terms of clinical cure [bib_ref] Clinical Practice Guideline by the Infectious Diseases Society of America (IDSA) and..., Johnson [/bib_ref]. This update in recommendations could also embrace the call to reduce the risk of concurrent and subsequent BSIs, thanks to the minimal impact of fidaxomicin, an oral macrolide, on gut microbiota compared to metronidazole and oral vancomycin. While the choice of fidaxomicin as first-line therapy for CDI is still limited among physicians partially due to its higher cost, the potential shorter length of hospitalization and reduced incidence of complicating BSIs (that have still to be demonstrated) may be cost-effective.
## Candida spp. and clostridioides difficile
Candidemia is defined as the presence of Candida spp. in blood, and it is invariably a pathological condition that requires proper evaluation, prompt treatment, and risk factor management as it cannot ever be considered a simple contamination.
Candida spp. is part of the normal gut microbiota, and translocation through the gastrointestinal wall is probably the most common mechanism by which Candida spp. enters the blood-stream in fragile patients, such as neutropenic subjects or those admitted to ICUs. The physiopathology behind the occurrence of candidemia in CDI should be the same as the one underlying bacterial BSIs. Raponi et al. have highlighted how CDI can predispose to Candida spp. overgrowth and its subsequent spread into the blood. In this prospective case-control study, they found that CDI was significantly associated with gut colonization by Candida spp. (83% in CDI-positive vs. 67% in CDI-negative), with Candida albicans being the species most often implicated [bib_ref] Clostridium Difficile Infection and Candida Colonization of the Gut: Is There a..., Raponi [/bib_ref]. Once again, the more plausible reasons for this phenomenon can be attributed to antibiotics use against both Cdiff and/or the concomitant infections that precede Cdiff colitis by reducing gut commensal competitors, as well as to direct interactions between Candida spp. and Cdiff. As for CDI-associated antibiotics, Nerandzic et al. analyzed the differences in the alterations of intestinal flora following different antibiotic treatments for CDI. After fidaxomicin treatment, there was a significant reduction in the risk of colonization by VRE or in the overgrowth of Candida spp. compared to patients receiving oral vancomycin; a finding likely attributable to the different spectrum of activity on the intestinal anaerobic flora [bib_ref] Reduced Acquisition and Overgrowth of Vancomycin-Resistant Enterococci and Candida Species in Patients..., Nerandzic [/bib_ref]. On the other side, the immunological alterations and changes in gut microbiota induced by Candida spp. overgrowth can modulate the susceptibility to CDI. In an experimental animal study, after oral inoculation of Cdiff spores, a lower rate of death due to CDI was observed in infected mice pre-colonized with C. albicans compared to those without. The subsequent growth of Cdiff in the gastrointestinal tract, the production of its toxins, and the presence of inflammation and tissue damage were similar in both groups, but the expression of specific inflammatory cytokines (such as IL-17A) in the infected tissues differed (being higher in pre-colonized mice, suggesting a different host response to Cdiff according to the presence and amount of Candida spp.) [bib_ref] Pre-Colonization with the Commensal Fungus Candida albicans Reduces Murine Susceptibility to Clostridium..., Markey [/bib_ref]. On the contrary, another mouse model detected higher serum 1-3 β-D-glucan (BDG), a fungal cell wall component, spontaneous Gram-negative BSIs, and gastrointestinal leakage markers in Cdiff infected mice that died compared to those surviving [bib_ref] → 3)-β-D-Glucan in a Clostridium Difficile Murine Model, Leelahavanichkul [/bib_ref]. In this model, BDG resulted as the best prognostic biomarker for 7-day mortality, and its levels, along with CDI severity, were attenuated by pre-emptive treatment with Lactobacillus rhamnosus [bib_ref] → 3)-β-D-Glucan in a Clostridium Difficile Murine Model, Leelahavanichkul [/bib_ref]. In another experimental study on the physical and chemical interactions between Cdiff and C. albicans, Cdiff was able to thrive at ambient oxygen levels when co-cultured with C. albicans, and it could secrete a compound with inhibitory activity against two virulence factors of C. albicans that modulate the transition from yeast (the invasive form) to hyphae and biofilm formation. Therefore, further studies are required to clarify what type of interactions between Cdiff and Candida spp. may occur and which are the resulting consequences to the host, as apparently discordant preliminary data point towards enhanced severity of the clinical co-infectious episode, but also to a concurrent increase in Candida spp. invasiveness and reduced virulence in the presence of actively replicating Cdiff.
Overall, Candida spp. seems to be the single most common microbial isolate in blood during CDI, and the prevalence of candidemia following CDI varied considerably among studies (0.8-8.6%), but co-infection is associated with substantially increased mortality [bib_ref] Candida Co-Infection Among Adults With Clostridium Difficile Infection in Metropolitan Atlanta, Vallabhaneni [/bib_ref]. Candidemia-related mortality is approximately 40%, but when candidemia is secondary to CDI, mortality can reach up to 60% [bib_ref] Management of Candidemia in Patients with Clostridium Difficile Infection, Falcone [/bib_ref]. This may be due to a higher translocating microbial burden, increased mucosal injury, an exacerbated inflammatory gut milieu, and the fact that candidemia is more commonly found in severe colitis among CDI cases. Indeed, severe CDI (aOR 4.4), including those by 027 ribotype (aOR 4.5), relapsing CDI (aOR 5.9), the treatment with high doses of vancomycin (≥1000 mg/day, aOR 2.1), immunosuppressive therapy (aOR 2.2), and the number of CDI relapses (aOR 3.1) have been previously recognized as independent risk factors for candidemia [bib_ref] Risk Factors and Clinical Outcomes of Candidaemia in Patients Treated for Clostridium..., Russo [/bib_ref]. Similarly, treatment with vancomycin plus metronidazole (usually prescribed for more severe cases) and severe CDI (based on clinical evaluation and Cdiff-specific complications) were associated with higher odds of developing candidemia up to 120 days after a Cdiff episode [bib_ref] Candida Co-Infection Among Adults With Clostridium Difficile Infection in Metropolitan Atlanta, Vallabhaneni [/bib_ref]. Therefore, we believe that a high index of suspicion for invasive candidiasis should be recommended, especially after a severe CDI episode.
## Cytomegalovirus and clostridioides difficile
Cytomegalovirus (CMV) infection is an important cause of morbidity and mortality among immunocompromised patients in contexts such as organ transplantation, chemotherapy, inflammatory bowel diseases (IBD) receiving immunosuppressive agents, and HIV infection [bib_ref] Cytomegalovirus Disease in Immunocompetent Adults, Kocak [/bib_ref]. Gastrointestinal involvement by CMV, namely CMV colitis, remains a rare occurrence in an immunocompetent host, but it is increasingly recognized in apparently immunocompetent subjects with some immune-modulating conditions such as advanced age, chronic renal failure, diabetes mellitus, and prolonged ICU stay [bib_ref] Coexisting Cytomegalovirus Infection in Immunocompetent Patients with Clostridium Difficile Colitis, Chan [/bib_ref].
Like the clinical manifestations of CDI, CMV colitis can manifest with symptoms such as diarrhea, abdominal pain, weight loss, intestinal bleeding, or fever and lead to complications such as toxic megacolon or even bowel perforation. Despite the aforementioned non-traditional risk factors for CMV colitis (partially overlapping with those of CDI) very few cases of this co-infection have been reported to date, so the incidence of this phenomenon cannot be inferred. Nevertheless, in the context of IBD, a close relationship between disease flares and CMV replication has been documented as gut inflammation seems to favor herpetic reactivation from compartmentalized intestinal sites of latent infection. The mechanisms underlying potential reactivation of intestinal latently residing CMV during Cdiff infections are not clear, but the shift in the mucosal immunologic balance, both in terms of cells and of cytokine patterns, could be hypothesized as it has been initially described in IBD and lead us to include CMV among the co-infections that should be considered. Moreover, it is possible that this co-infection is underestimated since the diagnosis of CMV colitis can be easily missed in patients with severe diarrhea and positive Cdiff toxin without apparent immunologic conditions requiring further investigations. In line with this hypothesis, the co-infection has been mainly described in case reports on CDI refractory to proper treatments that underwent further diagnostic work-up. In a patient with a squamous cell carcinoma of the lip and pancolitis secondary to CDI, persisting diarrhea despite appropriate treatment and proven bacteriological cure for Cdiff led to testing the stool and blood for CMV-DNA, which resulted in positive and dramatically improved after ganciclovir administration [bib_ref] A Rare Case Intractable Diarrhea Secondary to Clostridium Difficile and Cytomegalovirus Coinfection, John [/bib_ref]. Florescu et al. reported two cases of Cdiff and CMV co-infection in solid organ transplant recipients and analyzed seven previously published reports. Surprisingly, the authors observed a 100% rate of positive blood PCR for CMV, raising the possibility that CDI may increase the chance of developing a detectable viremia [bib_ref] Clostridium Difficile and Cytomegalovirus Colitis Co-Infection: Search for the Hidden "Bug, Florescu [/bib_ref]. Unfortunately, no values of the detected viremia were reported, so it cannot be assumed a real pathogenic role of plasma viremia; indeed, the role of low CMV viremia is uncertain in immunocompetent subjects or in HIV-positive patients with no evidence of organ involvement by CMV [bib_ref] CMV in Critically Ill Patients: Pathogen or Bystander?, Limaye [/bib_ref].
Few cases of CMV colitis following successful therapy for CDI have also been described: one patient admitted to ICU had colic ulceration due to CMV colitis three weeks after the resolution of a properly treated CDI, while a second elderly case showed CMV colitis as the cause of persisting bloody diarrhea at the end of oral vancomycin treatment for CDI [bib_ref] Cytomegalovirus Colitis in Intensive Care Unit Patients: Difficulties in Clinical Diagnosis, Chan [/bib_ref].
Although the co-infection of Cdiff and CMV seems to be a rare entity, considering the synergistic activity of these pathogens in increasing the risk of lethal intestinal perforation and toxic megacolon, clinicians should have a high index of suspicion to rule out CMV reactivation even in immunocompetent hosts, especially when other comorbidities or refractory disease are observed.
## Other co-infections in clostridioides difficile colitis
Co-infections with other enteric pathogens such as Salmonella spp., Cryptosporidium spp., Giardia spp., Enterocytozoon spp., and Campylobacter spp. have been described both in community and hospital-acquired cases of CDI [bib_ref] A Prospective Study of Community-Associated Clostridium Difficile Infections: The Role of Antibiotics..., Taori [/bib_ref] [bib_ref] Concurrent Infections of Giardia Duodenalis, Enterocytozoon Bieneusi, and Clostridium Difficile in Children..., Wang [/bib_ref]. In a prospective study on adult patients in Scotland, 13.3% of the tested HA-CDI cases were found to be co-infected with norovirus, which is a common pathogen causing nosocomial outbreaks [bib_ref] A Prospective Study of Community-Associated Clostridium Difficile Infections: The Role of Antibiotics..., Taori [/bib_ref].
Very few data are available about the co-infection of Cdiff with other common intestinal parasites. For instance, Entamoeba histolytica and Cdiff may present with similar clinical features or endoscopic findings and considering that empiric use of metronidazole for colitis treatment is widely practiced in low-and middle-income countries without testing for infectious causes, it may be speculated that amoebiasis is over-diagnosed and Cdiff infection underestimated as the latter is seldom considered and probably treated unknowingly with metronidazole [bib_ref] Clostridium Difficile and Entamoeba Histolytica Infections in Patients with Colitis in the..., Warren [/bib_ref]. Existing data suggest that the burden of CDI in low/middle-income countries is similar to high-income countries, but in the former, the diagnosis is hampered by both the lack of available testing and a low index of clinical suspicion [bib_ref] Assessing the Burden of Clostridium Difficile Infection in Low-and Middle-Income Countries, Roldan [/bib_ref]. Considering that the prevalence of intestinal parasites is higher in this setting, co-infections with Cdiff may not be an infrequent occurrence.
Co-infection with Cdiff and other gastrointestinal pathogens may also be common in children suffering from diarrhea, with a reported pooled rate of co-infections of 20.7% in Cdiff-positive children [bib_ref] Co-Infection as a Confounder for the Role of Clostridium Difficile Infection in..., De Graaf [/bib_ref]. Viral co-infections seem to be the most commonly found (46.0%), while bacterial and parasitic co-infections accounted for 14.9% and 0.01% of the cases, respectively [bib_ref] Co-Infection as a Confounder for the Role of Clostridium Difficile Infection in..., De Graaf [/bib_ref]. Unfortunately, the included studies were not conclusive regarding the impact of co-infections on CDI severity, and none evaluated causal relationships. Very scarce data also exist on these co-infections in adults. A case of co-infection with Giardia lamblia and Cdiff was described in a 49-year-old man taking ranitidine [bib_ref] Coinfection with Giardia Lamblia and Clostridium Difficile after Use of Ranitidine, Khatami [/bib_ref]. In another case, a patient with colorectal cancer was reported to have co-existing Cdiff and intestinal amebiasis infection [bib_ref] A 21 Year-Old Male Colorectal Cancer Patient with Clostridium Difficile and Intestinal..., Lugito [/bib_ref]. Due to the limited data available to date, no evidence can suggest an increased incidence or severity of parasitic intestinal infections in the presence of Cdiff/CDI.
## Probiotics and clostridioides difficile
Probiotics may play a role in the prevention of CDI by several mechanisms, including colonization resistance through maintaining a healthy gut flora, enhancing the clearance of Cdiff at the end of treatment, and inactivating the toxin receptor sites before the germination and growth of spores in the colon [bib_ref] Probiotics for Prevention of Clostridium Difficile Infection, Mills [/bib_ref]. Nevertheless, the use of probiotics in routine clinical practice remains debated.
A Cochrane meta-analysis of 39 randomized clinical trials concluded that probiotics reduce the incidence of CDI by 70% in adult and pediatric patients undergoing antibiotics for any reason, providing moderate quality of evidence in support of probiotics use in preventing Cdiff colitis. Post hoc analysis indicated that probiotics actually show preventive efficacy in patients with at least mild-moderate baseline risk of CDI and no benefit for subjects characterized by low risk [bib_ref] Probiotics for the Prevention of Clostridium Difficile-Associated Diarrhea in Adults and Children, Goldenberg [/bib_ref]. Conversely, a recent retrospective study on more than 3000 adults hospitalized patients observed that patients who received antibiotics with concurrent administration of probiotics (mainly Lactobacillus spp.) were more likely to develop CDI compared with those who did not receive probiotics (HR 2.7) [bib_ref] The Effect of Probiotics on the Incidence of Clostridioides difficile: Retrospective Cohort..., Saltzman [/bib_ref]. Similarly, the use of probiotics was not associated with decreased incidence of CDI among hospitalized adults aged 50 and above who received antibiotics in a recent multicentric study [bib_ref] A Multicenter Evaluation of Probiotic Use for the Primary Prevention of Clostridioides..., Heil [/bib_ref].
It is also recognized that these preparations, containing living microorganisms, can uncommonly cause different forms of invasive infections, particularly in critically ill or severely immunocompromised patients. Cases of Saccharomyces cerevisiae fungemia have been reported in patients with a history of probiotic use [bib_ref] Saccharomyces Cerevisiae Fungemia in a Critically Ill Patient with Acute Cholangitis and..., Fadhel [/bib_ref]. Furthermore, there is evidence of potential nosocomial development of fungemia in wards where probiotics were used: the contamination at the sites of vascular access was identified as the probable mechanism by which probiotics caused BSIs [bib_ref] Saccharomyces Cerevisiae Fungemia, a Possible Consequence of the Treatment of Clostridium Difficile..., Santino [/bib_ref]. As for other bacterial and fungal BSIs, also intesti-nal barrier impairment and concomitant administrations of broad-spectrum antibiotics have been acknowledged as possible risk factors for probiotics bacteremia, raising the hypothesis of an intestinal source that may play a role even in CDI [bib_ref] Mitigating Risk of Bloodstream Infection Related to Inpatient Probiotic Use, Polito [/bib_ref]. In line with this, some reports suggest the possibility of developing S. cerevisiae or Lactobacillus rhamnosus fungemia/bacteremia when probiotics containing such microorganisms were administered during CDI. In the latter case, bacteria could have been selected by the prolonged oral vancomycin therapy the patient received along with live yogurt as the administered probiotic (since Lactobacillus spp. are intrinsically resistant to vancomycin) [bib_ref] Yoghurt Biotherapy: Contraindicated in Immunosuppressed Patients?, Macgregor [/bib_ref]. Accordingly, cases of sepsis due to Lactobacillus spp. have been reported in neutropenic patients after oral vancomycin [bib_ref] Lactobacillus Species as Emerging Pathogens in Neutropenic Patients, Fruchart [/bib_ref].
While probiotics may play a role in preventing CDI in patients at risk that have not yet developed the infection, the administration of probiotics during overt and ongoing Cdiff colitis is therefore controversial and may be risky. Since the combination of enhanced intestinal permeability, altered gut microbiota, and immunosuppression is present in a large proportion of patients affected by CDI, the use of probiotics in this setting should be further evaluated to properly balance risks and benefits, especially when an extended duration of vancomycin is administered.
Finally, after a severe episode of CDI, immunological perturbations and cell damage in the intestinal mucosal barrier can occur and persist for several weeks. This functional disbalance in the gut and systemic immunity, already described in many other severe infections (such as malaria and septic shock), can also cast shadows on the opportunity and safety of probiotics to recover the normal gut microbiome following a severe episode of CDI. In a recent phase 3, double-blind RCT, the oral administration of SER-109, an investigational microbiome therapeutics composed of purified Firmicutes spores, to patients healed from a third or further episode of CDI (after standard-of-care antibiotic treatment) reduced the relative risk of recurrent infection by about 70% compared to placebo [bib_ref] SER-109, an Oral Microbiome Therapy for Recurrent Clostridioides difficile Infection, Feuerstadt [/bib_ref]. The study population was represented by 99% of outpatient subjects; therefore, it is likely that these promising results can be reliably applied to non-severe recurrent CDI (rCDI), and future studies are required to assess the safety and usefulness of probiotics in post-severe CDI.
## Fecal microbiota transplantation in clostridioides difficile colitis
Fecal microbiota transplantation (FMT), defined as the transfer of fecal microorganisms from healthy donors into the gut of recipient patients, has been associated with robust efficacy in the treatment of rCDI [bib_ref] Clinical Practice Guideline by the Infectious Diseases Society of America (IDSA) and..., Johnson [/bib_ref]. One or two FMT can be sufficient to cure rCDI in 90% of cases [bib_ref] Mechanistic Insights in the Success of Fecal Microbiota Transplants for the Treatment..., Baktash [/bib_ref]. In a randomized clinical study including 64 adult patients with rCDI, FMT delivered by colonoscopy or naso-jejunal tube after a short course of vancomycin was superior to fidaxomicin and standard-dose vancomycin monotherapies, based on the endpoints of clinical and microbiological resolution or clinical resolution alone [bib_ref] Fecal Microbiota Transplantation Is Superior to Fidaxomicin for Treatment of Recurrent Clostridium..., Hvas [/bib_ref].
Recent data suggest that FMT may be an alternative to antibiotic therapy also in the first CDI episode. In a small trial investigating the efficacy of FMT as a treatment for primary CDI, a clinical cure after initial treatment with no evidence of recurrence was achieved in seven patients in the transplantation group (78%) as compared with five in the metronidazole group (45%) [bib_ref] Fecal Microbiota Transplantation for Primary Clostridium Difficile Infection, Juul [/bib_ref]. Additionally, a phase three trial to assess FMT as a first-line treatment for severe primary CDI is ongoing (NCT02301000).
Interestingly, in a nonrandomized prospective single-center study, compared to antibiotics use, FMT reduced by 23% the incidence of BSIs in rCDI [bib_ref] Incidence of Bloodstream Infections, Length of Hospital Stay, and Survival in Patients..., Ianiro [/bib_ref]. A higher proportion of patients had sustained cure of CDI after treatment in the FMT group than in the antibiotic group (97% vs. 38%); no patient in the FMT group required surgery for severe CDI compared with 14 subjects in the antibiotic group (0% vs. 8%), and the incidence of BSIs during the 90-day follow-up was lower in the FMT group compared to antibiotic group (5% vs. 22%, and 1% vs. 6% polymicrobial infections) [bib_ref] Incidence of Bloodstream Infections, Length of Hospital Stay, and Survival in Patients..., Ianiro [/bib_ref]. No patient in the FMT developed fungal BSIs, while 12 (7.0%) cases of fungemia occurred in the antibiotic group [bib_ref] Incidence of Bloodstream Infections, Length of Hospital Stay, and Survival in Patients..., Ianiro [/bib_ref]. Moreover, patients in the FMT group had a significantly shorter length of hospitalization (13.4 vs. 27.8 days) [bib_ref] Incidence of Bloodstream Infections, Length of Hospital Stay, and Survival in Patients..., Ianiro [/bib_ref]. These results can be explained by several potential differences between the mechanisms underlying FMT compared to antibiotics: the earlier restoration of healthy gut microbiota, the avoidance of vancomycin, the increase in gut commensal competition, and the decrease in gut resistome, intended as the expression of antibiotic resistance genes by the gut microbiota. Similarly, despite the small sample and the tinier number of subjects receiving FMT for CDI (6 patients), compared to other treatments (vancomycin, fidaxomicin, and metronidazole), none of the subjects undergoing FMT developed BSIs.
Another possible favorable mechanism of prevention of co-infection operated by FMT is represented by the modulation of fecal bile acid composition; secondary bile acids, which are products of microbial metabolism, have been shown to inhibit Cdiff germination, growth, and toxin activity. The loss of beneficial Firmicutes bacteria induced by antibiotics leads to increased primary bile-acid concentrations, which on the contrary, enables Cdiff spore germination [bib_ref] Harnessing Microbiota to Kill a Pathogen: Fixing the Microbiota to Treat Clostridium..., Taur [/bib_ref].
Further studies are warranted to assess the potential multiple properties and mechanisms by which FMT may affect intestinal damage, inflammation, microbial translocation phenomena, and eventually, co-infection rates compared to standard of care for primary and rCDI episodes.
## Practical considerations
Based on currently available evidence, as well as on the relevant gap of data, it should be important to draw physicians' attention to the possibility of concurrent and subsequent co-infections in CDI, as bacterial BSIs and Candidemia may occur in up to 20% and 9% of the cases, respectively [fig_ref] Figure 2: Clinical changes, microbial epidemiology, and risk factors of co-infections in Cl tridioides... [/fig_ref]. Before, during, or up to one month after Cdiff treatment, fungal and bacterial coinfections should be suspected in the event of any clinical change or an unexpected variation in blood tests. As an example, high fever is an infrequent sign in CDI; it should trigger an assessment to rule out common co-infections, especially in non-severe cases (or whit deteriorating clinical status). Fulminant colitis is a life-threatening complication of CDI, occurring in about 3% of patients, and it is characterized by a clinical picture resembling that of septic shock: hypotension with or without the use of vasopressors, ileus, toxic megacolon, mental status changes, serum lactate levels >2.2 mmol/l, or any evidence of end-stage organ failure [bib_ref] Risk Factors for Recurrence, Complications and Mortality in Clostridium Difficile Infection: A..., Abou Chakra [/bib_ref]. Of note, a substantial number of patients with fulminant CDI (36% to 75%) have a history of recent surgery [bib_ref] Clinical Review of the Management of Fulminant Clostridium Difficile Infection, Jaber [/bib_ref] , which is a common risk factor for invasive candidiasis. All these three diagnoses (fulminant colitis, MDR-BSI, and candidemia) should be considered and potentially empirically addressed in case of shock development. Similarly, the mean duration of CDI symptoms is variable, also depending on the antibiotic therapy, and early surgical consultation is recommended in patients who do not respond to conventional therapy within 3 days [bib_ref] Initial Therapy Affects Duration of Diarrhoea in Critically Ill Patients with Clostridioides..., Manthey [/bib_ref]. The median length of Cdiff-related diarrhea was reported to be shorter in patients treated with vancomycin (about 3 days) compared with those given metronidazole (about 5 days) [bib_ref] Diarrhoea Caused by Clostridium Difficile: Response Time for Treatment with Metronidazole and..., Wilcox [/bib_ref]. The persistence or the new onset of signs and symptoms after 3-5 days from treatment initiation should also prompt the physician to rule out co-infections.
Clinical worsening after the resolution of a CDI episode should be interpreted as a warning sign to consider other explanations than Cdiff recurrence. Up to 25% of patients can experience recurrence of CDI within 30 days of completing treatment when antibioticinduced microbiota disruption facilitates Cdiff spore germination, especially in the elderly, or persisting use of antibiotics and proton pump inhibitors after the diagnosis. After the second CDI episode, the risk of multiple recurrences increases to 40-65% [bib_ref] Can We Identify Patients at High Risk of Recurrent Clostridium Difficile Infection?, Kelly [/bib_ref]. In the event of signs or symptoms of infection or of relapsing diarrhea, the differential diagnosis should address both subsequent co-infections and Cdiff recurrence. In case of worsening or recurrent diarrhea during or after treatment, detection of CMV-DNA by real-time PCR in fecal and blood samples or microbiological investigations in stools may be worthy in selected cases presenting traditional and non-traditional risk factors for CMV colitis or for rarer gastrointestinal pathogens (see [fig_ref] Figure 2: Clinical changes, microbial epidemiology, and risk factors of co-infections in Cl tridioides... [/fig_ref]. Clinical worsening after the resolution of a CDI episode should be interpreted a warning sign to consider other explanations than Cdiff recurrence. Up to 25% of patie can experience recurrence of CDI within 30 days of completing treatment when ant otic-induced microbiota disruption facilitates Cdiff spore germination, especially in elderly, or persisting use of antibiotics and proton pump inhibitors after the diagno After the second CDI episode, the risk of multiple recurrences increases to 40-65% [bib_ref] Can We Identify Patients at High Risk of Recurrent Clostridium Difficile Infection?, Kelly [/bib_ref] Prolonged diarrhea (defined as >5 days after the beginning of proper anti-Cdiff therapy) or relapsing diarrhea after initial resolution should also require the evaluation for causes other than recurrency and co-infections. After ruling out these events, in the absence of alternative diagnoses, patients suffering from persisting diarrhea may have refractory CDI or inflammatory colitis such as post-infectious irritable bowel syndrome; while the latter may complicate CDI in 4-25% of cases [bib_ref] Risk Factors for Recurrence, Complications and Mortality in Clostridium Difficile Infection: A..., Abou Chakra [/bib_ref] , it is extremely hard to exactly quantify the incidence and prevalence of refractory CDI as in most of the studies the temporal detection and definition criteria limit the possibility to distinguish it from reinfection. Detecting Cdiff toxins in the stool of these patients may not always be informative on the real ongoing pathological process, and colonoscopy should be considered. A practical scheme for the management of CDI and co-infections is depicted in [fig_ref] Figure 3: Practical flow-chart on Clostridioides difficile and co-infections clinical management [/fig_ref].
Together with clinical monitoring, a few common blood tests may help in assessing co-infections risk. Procalcitonin (PCT) remains at relatively low levels in CDI, although PCT concentration >0.5 μg/mL has been proposed as a reliable marker to identify severe CDI [bib_ref] The Role of Procalcitonin Levels in Assessing the Severity of Clostridium Difficile..., Dazley [/bib_ref]. Monitoring PCT in addition to blood culture collection from febrile patients may be helpful since a significant PCT elevation can predict the presence of BSIs after the start of anti-Cdiff therapy. Furthermore, Gram-BSIs have significantly higher PCT concentrations than Gram+ BSIs and candidemia, allowing for potential microbial etiology stratification [bib_ref] Serum Procalcitonin Levels Distinguish Gram-Negative Bacterial Sepsis from Gram-Positive Bacterial and Fungal..., Li [/bib_ref] [bib_ref] Procalcitonin Levels in Candidemia versus Bacteremia: A Systematic Review, Cortegiani [/bib_ref]. Strict monitoring of BDG levels can also be useful to rule out candidemia in emergent or persistent fever after initiation of anti-Cdiff therapy, especially in Prolonged diarrhea (defined as >5 days after the beginning of proper anti-Cdiff therapy) or relapsing diarrhea after initial resolution should also require the evaluation for causes other than recurrency and co-infections. After ruling out these events, in the absence of alternative diagnoses, patients suffering from persisting diarrhea may have refractory CDI or inflammatory colitis such as post-infectious irritable bowel syndrome; while the latter may complicate CDI in 4-25% of cases [bib_ref] Risk Factors for Recurrence, Complications and Mortality in Clostridium Difficile Infection: A..., Abou Chakra [/bib_ref] , it is extremely hard to exactly quantify the incidence and prevalence of refractory CDI as in most of the studies the temporal detection and definition criteria limit the possibility to distinguish it from reinfection. Detecting Cdiff toxins in the stool of these patients may not always be informative on the real ongoing pathological process, and colonoscopy should be considered. A practical scheme for the management of CDI and co-infections is depicted in [fig_ref] Figure 3: Practical flow-chart on Clostridioides difficile and co-infections clinical management [/fig_ref].
Together with clinical monitoring, a few common blood tests may help in assessing co-infections risk. Procalcitonin (PCT) remains at relatively low levels in CDI, although PCT concentration >0.5 µg/mL has been proposed as a reliable marker to identify severe CDI [bib_ref] The Role of Procalcitonin Levels in Assessing the Severity of Clostridium Difficile..., Dazley [/bib_ref]. Monitoring PCT in addition to blood culture collection from febrile patients may be helpful since a significant PCT elevation can predict the presence of BSIs after the start of anti-Cdiff therapy. Furthermore, Gram-BSIs have significantly higher PCT concentrations than Gram+ BSIs and candidemia, allowing for potential microbial etiology stratification [bib_ref] Serum Procalcitonin Levels Distinguish Gram-Negative Bacterial Sepsis from Gram-Positive Bacterial and Fungal..., Li [/bib_ref] [bib_ref] Procalcitonin Levels in Candidemia versus Bacteremia: A Systematic Review, Cortegiani [/bib_ref]. Strict monitoring of BDG levels can also be useful to rule out candidemia in emergent or persistent fever after initiation of anti-Cdiff therapy, especially in the setting of negative blood cultures. Lastly, marked leukocytosis can be seen in both candidemia and CDI, while leucopenia is more common in Gram-septicemia.
CDI is an extreme ominous example of negative pharmacoenosis, where the potentiality of beneficial drug combos is still unsatisfactorily exploited both as treatment and prevention. In the event of BSIs, the need to use additional antibiotics (other than those for CDI) can increase the risk of prolonged diarrhea and CDI recurrence. Thus, antibiotics associated with a lower risk of CDI should be preferred (such as macrolides, aminoglycosides, sulfonamides, vancomycin, or tetracyclines) and discontinued as soon as possible. Some authors have suggested prolonging treatment with antibiotics acting against Cdiff for a week after another broad-spectrum therapy has been withdrawn [bib_ref] Efficacy of Fidaxomicin Versus Vancomycin as Therapy for Clostridium Difficile Infection in..., Mullane [/bib_ref]. Notably, differently from what occurred with other cephalosporins, ceftobiprole was shown to have no significant ecological impact on the human microbiota and to exhibit some inhibitory activity against Cdiff in experimental models [bib_ref] Effect of Ceftobiprole Treatment on Growth of and Toxin Production by Clostridium..., Nerandzic [/bib_ref] [bib_ref] Ceftobiprole: Drug Evaluation and Place in Therapy, Giacobbe [/bib_ref]. Despite the role of tigecycline in CDI is controversial and not recommended by current guidelines [bib_ref] Tigecycline for the Treatment of Patients with Clostridium Difficile Infection: An Update..., Kechagias [/bib_ref] , it may be considered as a potential therapeutic option for patients with severe CDI, in addition to standard therapies, and as a relatively safe treatment for sensitive bacterial co-infections. As an anti-CDI booster, we may consider the administration of high-dose tigecycline (200 mg loading dose, followed by 100 mg every 12 h) in severe and fulminant colitis, as recommended in critically ill patients with severe infections due to multidrug-resistant organisms [bib_ref] Successful High-Dosage Monotherapy of Tigecycline in a Multidrug-Resistant Klebsiella Pneumoniae Pneumonia-Septicemia Model..., Van Der Weide [/bib_ref].
Potential successful strategies to reduce the risk of co-infections should primarily rely on the use of targeted anti-Cdiff therapies, which have minimal impact on gut microbiota, such as fidaxomicin, bezlotoxumab, and FMT. This last approach might not be just considered as the salvage treatment following repeated failures of antibiotics but as a reliable therapeutic option, although adoption of FMT as a first-line treatment is not yet recommended, and it would require further clinical and cost-effectiveness assessments; indeed, promising but limited evidence on hypothesized relevant benefits of FMT over microbial translocation and risk of BSIs in CDI is there [bib_ref] Frozen vs. Fresh Fecal Microbiota Transplantation and Clinical Resolution of Diarrhea in..., Lee [/bib_ref].
As for bezlotoxumab, none of the registration trials nor the few post-marketing studies reported on the incidence of BSIs in the study arms [bib_ref] Bezlotoxumab for Prevention of Recurrent Clostridium Difficile Infection, Wilcox [/bib_ref] [bib_ref] Real-World Comparison of Bezlotoxumab to Standard of Care Therapy for Prevention of..., Johnson [/bib_ref]. By blocking Cdiff toxin B, it is expected to act also dampening the gut epithelial injury and thereby the mechanical integrity of the mucosal barrier, reducing microbial translocation. On the opposite, vancomycin can be entero-toxic at high oral doses and inevitably increases gut dysbiosis and the risk of VRE colonization. Unfortunately, to date, oral vancomycin and bezlotoxumab do not occur at the same level in the management cascade of CDI, and further studies are required to endorse any reconsideration for the licensed prescription criteria of this anti-toxin B monoclonal antibody.
Prophylaxis with oral nystatin has been proposed in severe CDI at high risk for candidemia, but its role in preventing Candida translocation during and after Cdiff colitis has not been clearly established yet [bib_ref] Management of Candidemia in Patients with Clostridium Difficile Infection, Falcone [/bib_ref]. Probiotics seem beneficial in patients receiving antibiotic therapy at high risk for CDI, but evidence supporting their use as adjuvant therapy during overt CDI is scarce, and it is the opinion of the authors that during CDI, they may be potentially deleterious by increasing the risk of probiotics gut translocation and superinfections of blood-line access. Despite incomplete evidence, probiotics administration after the resolution of the episode may once again be promising in preventing at least CDI recurrences. Lastly, very few recent data point towards a potential application of prebiotics (either alone or in combination with probiotics) in preventing germination of Cdiff spores, modulating Cdiff adhesion [bib_ref] Effect of Prebiotics on Bacteroides Sp. Adhesion and Biofilm Formation and Synbiotic..., Piotrowski [/bib_ref] [bib_ref] Fructooligosaccharides and Mannose Affect Clostridium Difficile Adhesion and Biofilm Formation in a..., Piotrowski [/bib_ref] , and in stimulating competing gut commensals [bib_ref] Next-Generation Prebiotic Promotes Selective Growth of Bifidobacteria, Suppressing Clostridioides difficile, Hirano [/bib_ref] in both in vitro and animal models; thereby, prebiotics may indeed reduce CDI burden, but unfortunately to date, there are no studies reporting data on plausible effects of prebiotics on the modulation of the risk of microbial translocation and co-infections.
# Conclusions
In conclusion, further efforts are needed to better detail, quantify, prevent, and effectively manage the risk of concurrent and subsequent co-infections, mainly BSIs, in patients with CDI. Although it is not properly quantifiable to date, this complication likely contributes to the global burden of CDI by increasing the length of hospital stay and CDI-related mortality. Gut microbiota represents the main source of BSIs in CDI and its preservation, and eventually, restoration is pivotal to facing this clinical complication as well as to preventing Cdiff recurrences. Promising data on post-CDI-specific probiotics and FMT bodes well for the future. In the era of MDR bacteria, but also of expanding knowledge on gut microbiota and evolving treatment options, a multi-step approach to CDI and co-infections management is warranted; this should include CDI prevention by antimicrobial stewardship and modulation of risk factors, the adoption of microbiota "conciliating" drugs, the prompt recognition of this underestimated co-occurrence, and proper treatments able to avoid or limit the vicious circle that arises between Cdiff, antibiotics and gut environment.
## Conflicts of interest:
The authors declare no conflict of interest.
[fig] Figure 1: Schematic representation of predisposing factors and mechanisms underlying bacterial, fungal, and viral infections that develop concomitantly or subsequently to Clostridioides difficile colitis. Legend: Th cells, T helper lymphocytes; APC, antigen presenting cells; CMV, Cytomegalovirus; VRE, Vancomycin-resistant Enterococcus spp. [/fig]
[fig] Figure 2: Clinical changes, microbial epidemiology, and risk factors of co-infections in Cl tridioides difficile colitis. Legend: PCR, C reactive protein; MOF, multi organ failure; Cdiff, C tridioides difficile; spp., species; PCT, procalcitonin; ICU, intensive care unit; CVC, central ven catheter; MDRs, multi-drug resistant bacteria; IBD, inflammatory bowel diseases; HIV, Hum immunodeficiency virus. [/fig]
[fig] Figure 3: Practical flow-chart on Clostridioides difficile and co-infections clinical management. Legend: WBC, white blood cells; PCR, C reactive protein; PCT, procalcitonin; BDG, beta-D-glucan; MDR, multi-drug resistant; MAbs, anti-toxin B monoclonal antibodies (bezlotoxumab). * If possible prefer "microbiota-preserving" treatments, such as fidaxomicin. [/fig]
[fig] Author: Contributions: Conceptualization, M.T. and S.S.; methodology, M.T., S.S. and W.R.; writing-original draft preparation, M.T., S.S., W.R. and A.C.; writing-review and editing, all the authors; supervision, S.B., G.D.P. and A.C. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. [/fig]
[table] Table 1: Summary of literature studies reporting period prevalence data on bacterial and/or fungal blood-stream infection during and following Clostridioides difficile colitis. [/table]
|
Humoral and Cellular Immune Response in Asymptomatic Dogs with Visceral Leishmaniasis: A Review
# Introduction
Visceral leishmaniasis (VL), caused by the protozoa Leishmania infantum (syn. L. chagasi, in Latin America) and transmitted by the bite of female phlebotomine sand flies [bib_ref] Insights on adaptive and innate immunity in canine leishmaniosis, Hosein [/bib_ref] , is the most severe and progressive form of leishmaniasis [bib_ref] Canine Leishmaniasis: An Overview of the Current Status and Strategies for Control, Ribeiro [/bib_ref]. This important zoonotic disease affects both humans and dogs in endemic areas of the Mediterranean basin, Asia, and Latin America, and it is emerging in North America [bib_ref] Cytokine and Phenotypic Cell Profiles of Leishmania infantum Infection in the Dog, Maia [/bib_ref] [bib_ref] Cytokine profiles in canine visceral leishmaniasis, Carrillo [/bib_ref]. VL is one of the deadliest parasitic diseases in the world, causing an estimated 20,000 to 40,000 human deaths and 0.2-0.4 million new cases each year [bib_ref] Current Visceral Leishmaniasis Research: A Research Review to Inspire Future Study, Bi [/bib_ref]. Infected dogs, whose prevalence in the canine population may reach up to 80% in highly endemic areas, are the main reservoir of the parasite in urban zones and play a key role in the transmission cycle of L. infantum to humans [bib_ref] Cytokine profiles in canine visceral leishmaniasis, Carrillo [/bib_ref]. Indeed, it has been observed how an increase in canine visceral leishmaniasis (CVL) cases precedes a rise in human cases [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref]. This highlights the importance of CVL control strategies to manage the spread of the parasite in human and canine populations and the need for early and comprehensive diagnosis of dogs infected with L. infantum [bib_ref] Methods for diagnosis of canine leishmaniasis and immune response to infection, Maia [/bib_ref].
Effective control of CVL includes vector control, prevention, treatment, or culling of infected dogs [bib_ref] Current Visceral Leishmaniasis Research: A Research Review to Inspire Future Study, Bi [/bib_ref]. However, currently, there are still barriers to overcome, such as the early screening of infected dogs. Likewise, since there is no scientific evidence that supports that seropositive dog culling could reduce the incidence of VL, this control measure should be revised or sufficiently argued [bib_ref] Canine Leishmaniasis: An Overview of the Current Status and Strategies for Control, Ribeiro [/bib_ref] [bib_ref] Current Visceral Leishmaniasis Research: A Research Review to Inspire Future Study, Bi [/bib_ref]. Drug treatment of these infected dogs is also expensive, and although the available protocols can promote clinical cure, improve quality of life and life expectancy, and reduce the parasite load and infectiousness to sand fly vectors, parasitological cures are rarely achieved, and the rates of relapse are high [bib_ref] Canine Leishmaniasis: An Overview of the Current Status and Strategies for Control, Ribeiro [/bib_ref] [bib_ref] Systemic and compartmentalized immune response in canine visceral leishmaniasis, Reis [/bib_ref]. Vaccination of these animals is one of the most promising tools for the effective control of this disease. There are currently three commercially available vaccines against CVL: Leish-Tec ® (Ceva Animal Health, Brazil), CaniLeish ® (Virbac Santé Animale, France), and Letifend ® (Laboratorios Leti, Spain) [bib_ref] The balancing act: Immunology of leishmaniosis, Toepp [/bib_ref]. Despite the reductions in infectiveness and disease progression that these vaccines showed in some trials [bib_ref] Insights on adaptive and innate immunity in canine leishmaniosis, Hosein [/bib_ref] [bib_ref] An Overview of Immunotherapeutic Approaches Against Canine Visceral Leishmaniasis: What Has Been, Gonçalves [/bib_ref] , their protective efficacy is still low [bib_ref] Canine Leishmaniasis: An Overview of the Current Status and Strategies for Control, Ribeiro [/bib_ref]. Since the levels of protection offered by immunization alone are not considered satisfactory for preventing L. infantum infection, the commercially available vaccines themselves recommend simultaneous administration with topical insecticides [bib_ref] Commercially approved vaccines for canine leishmaniosis: A review of available data on..., Velez [/bib_ref]. Furthermore, the application of low-moderate efficacy vaccines in endemic areas could have a negative impact on the diagnosis and control of Leishmaniasis, since these vaccines, by only reducing the appearance or severity of clinical signs, may mask the disease in infected individuals, thus constituting an important reservoir of the parasite, which in turn could indirectly induce an increase in the incidence of infection [bib_ref] Veterinary Parasitology Vaccination with LiESP/QA-21 (CaniLeish ® ) reduces the intensity of..., Bongiorno [/bib_ref] [bib_ref] Novel Areas for Prevention and Control of Canine Leishmaniosis, Miró [/bib_ref]. On the other hand, currently available studies on licensed vaccines are considered insufficient and do not allow for comparative studies between them due to the lack of standardization in study design, methodological shortcomings, and substantial differences in the characteristics of the populations evaluated [bib_ref] Commercially approved vaccines for canine leishmaniosis: A review of available data on..., Velez [/bib_ref].
The precise diagnosis of CVL may also be complex, as not all infected animals develop clinical manifestations [bib_ref] Methods for diagnosis of canine leishmaniasis and immune response to infection, Maia [/bib_ref]. In 1988, Mancianti et al. classified CVL into three clinical forms on the basis of major features observed in seropositive infected dogs: asymptomatic dogs (AD), who do not show signs of the disease; oligosymptomatic dogs (OD), with a maximum of three clinical signs, including opaque bristles and/or localized alopecia and/or moderate loss of weight; and symptomatic dogs (SD), who show some or all severe signs of the disease, such as opaque bristles, severe loss of weight, onychogryphosis, cutaneous lesions, apathy, and keratoconjunctivitis [bib_ref] Studies on canine leishmaniasis control. 1. Evolution of infection of different clinical..., Mancianti [/bib_ref]. However, classification of dogs based solely on physical examination without considering clinical pathological abnormalities or the possibility of undetectable organ dysfunction may be insufficient and misleading. On this basis, more recently, different authors recommend following LeishVet guidelines, which have been developed from exhaustive review of evidence-based studies, clinical expertise, and consensus of opinions derived from critical debates [bib_ref] When is an "asymptomatic" dog asymptomatic?, Dantas-Torres [/bib_ref]. In addition, some asymptomatic dogs cannot be detected by conventional serological tests, while through direct methods, such as PCR, the percentage of these seronegative asymptomatic dogs is considerably high in endemic areas [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref]. This fact implies a major problem, since asymptomatic dogs have shown to be highly competent to transmit the parasite to the vector [bib_ref] Asymptomatic dogs are highly competent to transmit Leishmania (Leishmania) infantum chagasi to..., Laurenti [/bib_ref]. This is of great relevance because these animals that do not show clinical signs of CVL constitute an undetectable source of infection [bib_ref] Experimental infection of Phlebotomus perniciosus and determination of the natural infection rates..., Guarga [/bib_ref] [bib_ref] Canine leishmaniosis-New concepts and insights on an expanding zoonosis: Part one, Baneth [/bib_ref]. Although several studies have reported that symptomatic dogs infected with L. infantum are highly infectious to their sand fly vector compared to oligosymptomatic and asymptomatic dogs [bib_ref] Experimental infection of Phlebotomus perniciosus and determination of the natural infection rates..., Guarga [/bib_ref] [bib_ref] Infectivity of dogs infected with Leishmania infantum to colonized Phleobotomus perniciosus, Molina [/bib_ref] [bib_ref] Canine leishmaniasis transmission: Higher infectivity amongst naturally infected dogs to sand flies..., Guarga [/bib_ref] [bib_ref] Visceral leishmaniasis in Teresina, State of Piauí, Brazil: Preliminary observations on the..., Vexenat [/bib_ref] [bib_ref] Canine visceral leishmaniasis: Relationships between clinical status, humoral immune response, haematology and..., Da Costa-Val [/bib_ref] [bib_ref] Infectivity of seropositive dogs, showing different clinical forms of leishmaniasis, to Lutzomyia..., Michalsky [/bib_ref] [bib_ref] Canine visceral leishmaniasis in Colombia: Relationship between clinical and parasitological status and..., Travi [/bib_ref] , there is no clear relationship between the stage of clinical evolution and infectivity to sand flies [bib_ref] Experimental infection of Phlebotomus perniciosus and determination of the natural infection rates..., Guarga [/bib_ref] [bib_ref] Infectivity of dogs infected with Leishmania infantum to colonized Phleobotomus perniciosus, Molina [/bib_ref] [bib_ref] Canine leishmaniasis: Clinical, parasitological and entomological follow-up after chemotherapy, Alvar [/bib_ref] [bib_ref] Parasite load in the blood and skin of dogs naturally infected by..., Borja [/bib_ref]. In this context, it has been reported that other factors such as the nutritional status of the dogs, the virulence of the parasite strain, the vectorial capacity of the sandfly species, and the strains involved in transmission could also play a role in the intensity of transmission.
The evolution of Leishmania infection and its clinical manifestations are the result of the complex interactions between the host immune system and the parasite [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref]. As in many diseases caused by protozoan parasites, infection is the first contact between the parasite and its host; the host may kill the parasite due to innate and/or acquired immunity, or the parasite may survive due to an efficient mechanism that evades the host response. If the parasite survives, an intriguing dynamic relationship between host and parasite may result; when in balance, the host becomes an 'asymptomatic carrier', and when out of balance, the result is disease [bib_ref] Implications of asymptomatic infection for the natural history of selected parasitic tropical..., Alvar [/bib_ref]. Hereby, there is a correlation between the distinct clinical forms of CVL and certain immunopathological characteristics during the progression of the disease. In this context, it has been observed that while symptomatic dogs display a susceptible profile response, asymptomatic dogs exhibit a resistance pattern [bib_ref] Canine visceral leishmaniasis biomarkers and their employment in vaccines, Giunchetti [/bib_ref]. On this basis, the study of the immune response in asymptomatic dogs and biomarkers associated with resistance profiles might guide the development of more effective vaccines and more precise diagnostic/prognosis tests. The identification of biomarkers related to asymptomatic disease maintenance would also allow us to evaluate the immunoprotection induced by vaccination or treatment strategies [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref] [bib_ref] Canine visceral leishmaniasis biomarkers and their employment in vaccines, Giunchetti [/bib_ref] [bib_ref] Immunology of canine leishmaniasis, Barbiéri [/bib_ref]. Learning about the immune response and biomarkers associated with these dogs would widen the knowledge we have about CVL and the capacity to create better diagnostic tools to control the disease. Thus, the aim of this work was to review the systemic immune response of asymptomatic dogs by including studies focused on serum and/or peripheral blood and to discuss the biological meaning of these findings and the gaps that need to be addressed.
# Materials and methods
The literature search was performed in the Medline database (PubMed) in accordance with PRISMA guidelines and using determined keywords that are described in [fig_ref] Table 1: Studies that analyzed the immune profiles of PBMCs from infected and control... [/fig_ref]. The search was limited to articles published from 1990 to 2022. After the initial search and review of the references contained in each of the articles, reviews were excluded, and original research articles, reports, comparative studies, and short communications were selected.
Articles were included in this review according to the following inclusion criteria:
- Studies performed in dogs.
## -
Studies that analyze humoral and/or cellular immune responses in serum and/or peripheral blood of asymptomatic dogs infected with L. donovani, L. chagasi, or L. infantum.
## -
Studies that include not only a group of asymptomatic dogs but also a group of healthy noninfected control dogs and/or symptomatic dogs with visceral leishmaniasis.
Conversely, articles were excluded when they fell within one or more of the following categories:
- Studies unrelated to visceral leishmaniasis.
## -
Studies that focused on human and nondog animal models. - Studies in which dogs were vaccinated, immunized, or treated.
## -
Studies that did not focus on the immune response of dogs. - Studies in which the immune response was analyzed in tissues.
## -
Studies in which the clinical classification was incorrect or inexistent.
A more exhaustive literature review focused on the immune response was conducted using the clusters of terms described in . The articles were selected as described in [fig_ref] Figure 1: Flow diagram showing the steps followed to select the articles included in... [/fig_ref].
It is worth mentioning that when extracting the information from each individual study, we took into account the following variables to avoid any kind of bias: Leishmania species, type of infection (experimental or natural), fluid analyzed (serum or peripheral blood), number of dogs studied, geographical location, source of origin and inoculation route of parasites in case of experimental infection, diagnostic methods, clinical evaluation, techniques employed for cytokine and immunoglobulin detection, and type of antibody used for immunoglobulin detection (monoclonal or polyclonal antibody). It is worth mentioning that when extracting the information from each individual study, we took into account the following variables to avoid any kind of bias: Leishmania species, type of infection (experimental or natural), fluid analyzed (serum or peripheral blood), number of dogs studied, geographical location, source of origin and inoculation route of parasites in case of experimental infection, diagnostic methods, clinical evaluation, techniques employed for cytokine and immunoglobulin detection, and type of antibody used for immunoglobulin detection (monoclonal or polyclonal antibody).
# Results
## Phenotypic characterization of the lymphocyte population
Several studies have been conducted on the immunological profile of peripheral blood mononuclear cells (PBMCs) in the context of CVL [fig_ref] Table 1: Studies that analyzed the immune profiles of PBMCs from infected and control... [/fig_ref]. In most of them, asymptomatic dogs (AD) showed increased concentrations of total T lymphocytes [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref] [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref] [bib_ref] Phenotypic features of circulating leucocytes as immunological markers for clinical status and..., Reis [/bib_ref] [bib_ref] Cytokine Gene Expression in the Tissues of Dogs Infected by Leishmania infantum, Barbosa [/bib_ref] [bib_ref] Clinical forms of canine visceral leishmaniasis in naturally Leishmania infantum-infected dogs and..., Nicolato [/bib_ref] [bib_ref] Specific IgG1 and IgG2 antibody and lymphocyte subset levels in naturally Leishmania..., Bourdoiseau [/bib_ref] , CD4 and CD8 T cells [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref] [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref] [bib_ref] Phenotypic features of circulating leucocytes as immunological markers for clinical status and..., Reis [/bib_ref] [bib_ref] Specific IgG1 and IgG2 antibody and lymphocyte subset levels in naturally Leishmania..., Bourdoiseau [/bib_ref] [bib_ref] Detection of intracellular IFN-γ and IL-4 cytokines in CD4+ and CD8+ T..., Matralis [/bib_ref] , and B cells [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref] [bib_ref] Phenotypic features of circulating leucocytes as immunological markers for clinical status and..., Reis [/bib_ref] [bib_ref] Specific IgG1 and IgG2 antibody and lymphocyte subset levels in naturally Leishmania..., Bourdoiseau [/bib_ref] in comparison with symptomatic dogs (SD). When a distinction was made between asymptomatic dogs with negative (AD-I) and positive serology (AD-II), some authors observed that both AD-I and AD-II had higher levels of CD5 + and CD4 + T cells, whereas CD8 + T cells and CD21 + B cells were found only high in AD-II, decreasing the CD4 + /CD8 + ratio in seropositive dogs [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref]. In contrast, other researchers reported decreased values of CD5 + , CD4 + , CD8 + , and CD21 + cells in groups of seropositive dogs (AD-II and SD) when compared to seronegative dogs (control healthy dogs (CD) and AD-I), despite the occurrence of clinical signs of CVL, with
# Results
## Phenotypic characterization of the lymphocyte population
Several studies have been conducted on the immunological profile of peripheral blood mononuclear cells (PBMCs) in the context of CVL [fig_ref] Table 1: Studies that analyzed the immune profiles of PBMCs from infected and control... [/fig_ref]. In most of them, asymptomatic dogs (AD) showed increased concentrations of total T lymphocytes [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref] [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref] [bib_ref] Phenotypic features of circulating leucocytes as immunological markers for clinical status and..., Reis [/bib_ref] [bib_ref] Cytokine Gene Expression in the Tissues of Dogs Infected by Leishmania infantum, Barbosa [/bib_ref] [bib_ref] Clinical forms of canine visceral leishmaniasis in naturally Leishmania infantum-infected dogs and..., Nicolato [/bib_ref] [bib_ref] Specific IgG1 and IgG2 antibody and lymphocyte subset levels in naturally Leishmania..., Bourdoiseau [/bib_ref] , CD4 and CD8 T cells [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref] [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref] [bib_ref] Phenotypic features of circulating leucocytes as immunological markers for clinical status and..., Reis [/bib_ref] [bib_ref] Specific IgG1 and IgG2 antibody and lymphocyte subset levels in naturally Leishmania..., Bourdoiseau [/bib_ref] [bib_ref] Detection of intracellular IFN-γ and IL-4 cytokines in CD4+ and CD8+ T..., Matralis [/bib_ref] , and B cells [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref] [bib_ref] Phenotypic features of circulating leucocytes as immunological markers for clinical status and..., Reis [/bib_ref] [bib_ref] Specific IgG1 and IgG2 antibody and lymphocyte subset levels in naturally Leishmania..., Bourdoiseau [/bib_ref] in comparison with symptomatic dogs (SD). When a distinction was made between asymptomatic dogs with negative (AD-I) and positive serology (AD-II), some authors observed that both AD-I and AD-II had higher levels of CD5 + and CD4 + T cells, whereas CD8 + T cells and CD21 + B cells were found only high in AD-II, decreasing the CD4 + /CD8 + ratio in seropositive dogs [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref]. In contrast, other researchers reported decreased values of CD5 + , CD4 + , CD8 + , and CD21 + cells in groups of seropositive dogs (AD-II and SD) when compared to seronegative dogs (control healthy dogs (CD) and AD-I), despite the occurrence of clinical signs of CVL, with higher CD4 + /CD8 + ratios in AD-I and SD [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref]. However, this relationship between clinical status and lymphocyte profile was not observed in all studies. Some authors detected higher percentages of CD4 + and CD8 + T cells in all infected groups of naturally infected dogs, regardless of their clinical classification [bib_ref] Regulatory T cells, Cytotoxic T lymphocytes and a Th1 cytokine profile in..., Cortese [/bib_ref] , and other studies did not even find significant differences between noninfected and infected dogs [bib_ref] Flow cytometric analysis of cellular immune responses in dogs experimentally infected with..., Rosypal [/bib_ref] [bib_ref] Canine leishmaniosis. Immunophenotypic profile of leukocytes in different compartments of symptomatic, asymptomatic..., Alexandre-Pires [/bib_ref]. With regard to B cells, although some authors also found no significant differences among clinical groups, they did report a higher population of regulatory (IgD hi ) B cells in SD compared to CD and AD [bib_ref] Regulatory IgD hi B Cells Suppress T Cell Function via IL-10 and..., Schaut [/bib_ref]. Concerning the role of regulatory T cells (Tregs), very few studies focused on Vaccines 2022, 10, 947 5 of 17 their involvement in asymptomatic disease of L. infantum-infected dogs. The percentage of Tregs (Foxp3 + CD4 + ), which play a key role in T-cell activation, was decreased in all infected dogs, while the expression of Foxp3 in CD8 + cells was barely detected in any group [bib_ref] Regulatory T cells, Cytotoxic T lymphocytes and a Th1 cytokine profile in..., Cortese [/bib_ref]. Likewise, although the activation status of lymphocytes in PBMCs has been little studied, asymptomatic disease has been associated with a higher activation status of circulating lymphocytes, as these animals showed higher expression of MHC-II and an increased CD45RB/CD45RA expression index in comparison to SD and CD [bib_ref] Phenotypic features of circulating leucocytes as immunological markers for clinical status and..., Reis [/bib_ref]. However, another study observed this upregulation of MHC-II + in lymphocytes from peripheral blood in both asymptomatic and symptomatic animals compared to CD, although this increase was only significant in SD [bib_ref] Canine leishmaniosis. Immunophenotypic profile of leukocytes in different compartments of symptomatic, asymptomatic..., Alexandre-Pires [/bib_ref]. Concerning other markers involved in the regulation of T-cell activation, some authors have also studied the coinhibitory molecule PD-1, which was found to be overexpressed in CD4 + and CD8 + T cells from SD and AD compared to CD, with higher expression in SD [bib_ref] Programmed Death 1-Mediated T Cell Exhaustion during Visceral Leishmaniasis Impairs Phagocyte Function, Esch [/bib_ref]. No differences between AD and SD for any marker. ↓% CD4 + CD3 + cells and ↑ % CD8 + CD3 + cells in AD and SD versus CD (****). ↓ Treg CD4 + cells in AD and SD versus CD (****). [bib_ref] Regulatory T cells, Cytotoxic T lymphocytes and a Th1 cytokine profile in..., Cortese [/bib_ref] MHC-II AD (n = 10) SD (n = 10) TD (n = 10) CD (n = 10)
## Unknown
Natural FC ↓ CD3 + T cells in SD versus AD and CD (*). Similar numbers of CD4 + and CD8 + T cells between groups. ↑ % MHC-II + lymphocytes in SD versus CD (*). ↓ CD21 + cells in SD versus AD (**) and CD (*). ↑ MHC-II expression and CD45RB/CD45RA ratio in lymphocytes from AD versus SD, OD, and CD (*). ↑ CD8 + cells in AD-II versus SD and CD (*). ↑ CD21 + cells in AD-II and CD versus SD (*).
[formula] AD (n = 7) SD (n = 7) CD (n = 7) [/formula]
USA and Natal (Brazil) Natural FC and FACS Similar % CD19 + and CD21 + cells between groups. ↑ IgD hi B cells in SD versus AD and CD (***).
## Cytolytic activity of pbmcs
Even though resistance to Leishmania infection is associated with the cytokine profile of PBMCs, as we will examine in depth in this review, other mechanisms of action could be involved in the protective effect of T cells, such as cytolysis [bib_ref] Leishmania infantum-specific T cell lines derived from asymptomatic dogs that lyse infected..., Pinelli [/bib_ref]. The research conducted by Pinelli et al. showed that PBMCs from asymptomatic, but not from symptomatic, dogs lysed autologous L. infantum-infected macrophages and that cytotoxic CD8 + T and even CD4 + T cells in some animals were involved and essential for this cytolytic response [bib_ref] Leishmania infantum-specific T cell lines derived from asymptomatic dogs that lyse infected..., Pinelli [/bib_ref].
## Nitric oxide production
One of the mechanisms by which macrophages exert their antileishmanial activity is through the production of nitric oxide (NO). Given the role of this molecule in parasite clearance, it is also necessary to address the relationship between NO production in PBMCs and clinical signs of CVL. Panaro et al. analyzed NO production by PBMC-derived macrophages infected in vitro by L. infantum at 4 and 8 months after diagnosis. The authors observed that, while at the first follow-up (4 months), macrophages from symptomatic dogs released higher levels of NO, the opposite occurred 8 months after diagnosis, when NO production in asymptomatic dogs suffered a substantial increase that was not observed in symptomatic dogs. The correlation between NO production and asymptomatic disease was previously reported in naturally infected dogs from endemic areas of Brazil [bib_ref] Recombinant cysteine proteinase from Leishmania (Leishmania) chagasi implicated in human and dog..., Da Costa Pinheiro [/bib_ref]. Souza et al., in a recent manuscript, also described the presence of higher NO production in all infected dogs compared to noninfected controls and higher NO levels in asymptomatic versus symptomatic dogs [bib_ref] A potential link among antioxidant enzymes, histopathology and trace elements in canine..., Souza [/bib_ref].
## Lymphoproliferative response of pbmcs
The capacity of PBMCs from asymptomatic dogs to proliferate upon exposure to Leishmania-specific antigens has been extensively discussed . Multiple studies have found an association between the protective immunity of asymptomatic dogs and a strong proliferative response of PBMCs to several Leishmania antigens: soluble Leishmania antigen (SLA), frozen and thawed (f/t) antigens, amastigote extracts (AM), the purified promastigote protein gp63, several recombinant cysteine proteinases from L. infantum (rCPA, rCPB, CTE) and L. chagasi (rLdccys), amastigote protein P-8, and L. infantum recombinant antigens HSP-70, PFR-2, KMP-11, LeIF, and Ldp23 [bib_ref] Programmed Death 1-Mediated T Cell Exhaustion during Visceral Leishmaniasis Impairs Phagocyte Function, Esch [/bib_ref] [bib_ref] Leishmania infantum-specific T cell lines derived from asymptomatic dogs that lyse infected..., Pinelli [/bib_ref] [bib_ref] Recombinant cysteine proteinase from Leishmania (Leishmania) chagasi implicated in human and dog..., Da Costa Pinheiro [/bib_ref] [bib_ref] Early suppression of lymphoproliferative response in dogs with natural infection by Leishmania..., De Luna [/bib_ref] [bib_ref] Leishmania-specific lymphoproliferative responses and IgG1/IgG2 immunodetection patterns by Western blot in asymptomatic,..., Fernández-Pérez [/bib_ref] [bib_ref] Expression of cysteine proteinase type I and II of Leishmania infantum and..., Rafati [/bib_ref] [bib_ref] Humoral and cellular immune responses against Type I cysteine proteinase of Leishmania..., Nakhaee [/bib_ref] [bib_ref] Immunologic indicators of clinical progression during canine Leishmania infantum infection, Boggiatto [/bib_ref] [bib_ref] Leishmania-specific isotype levels and their relationship with specific cell-mediated immunity parameters in..., Rodríguez-Cortés [/bib_ref] [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] Cellular and humoral immune responses in dogs experimentally and naturally infected with..., Pinelli [/bib_ref] [bib_ref] The immunology of canine leishmaniosis: Strong evidence for a developing disease spectrum..., Cabral [/bib_ref] [bib_ref] Immune response against Leishmania antigens in dogs naturally and experimentally infected with..., Rhalem [/bib_ref]. SLA is one of the most commonly used antigens in the lymphoproliferative assays. [fig_ref] Figure 2: Graphical representation of the percentage of dogs who exhibited positive lymphoproliferative response... [/fig_ref] shows a comparative analysis of the lymphoproliferative response after SLA stimulus in AD and SD dogs. As observed in [fig_ref] Figure 2: Graphical representation of the percentage of dogs who exhibited positive lymphoproliferative response... [/fig_ref] , and even though in several studies the SLA stimulus induced a higher specific proliferative response in PBMCs from AD [bib_ref] Leishmania-specific lymphoproliferative responses and IgG1/IgG2 immunodetection patterns by Western blot in asymptomatic,..., Fernández-Pérez [/bib_ref] [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] Cellular and humoral immune responses in dogs experimentally and naturally infected with..., Pinelli [/bib_ref] , in some studies, other authors observed similar lymphoproliferative responses to this antigen in both AD and SD [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] Cell mediated immunity and specific IgG1 and IgG2 antibody response in natural..., Leandro [/bib_ref] [bib_ref] Clinical, parasitologic, and immunologic evolution in dogs experimentally infected with sand fly-derived..., Travi [/bib_ref]. PBMCs from AD, which did not proliferate upon stimulation with SLA, were shown to proliferate after exposure to other recombinant antigens, such as LeIF and Ldp23 [bib_ref] Early suppression of lymphoproliferative response in dogs with natural infection by Leishmania..., De Luna [/bib_ref]. PBMCs from experimentally infected AD also proliferate upon induction by HSP-70, PFR-2, or KMP-11 recombinant proteins, although at a lower level Vaccines 2022, 10, 947 7 of 17 than that induced by SLA. Moreover, the f/t lysate produced stronger proliferation in PBMCs from AD than the recombinant antigens rCPA and rCPB and the synthetic peptide CTE [bib_ref] Expression of cysteine proteinase type I and II of Leishmania infantum and..., Rafati [/bib_ref] [bib_ref] Humoral and cellular immune responses against Type I cysteine proteinase of Leishmania..., Nakhaee [/bib_ref] but slightly lower proliferation than gp63 [bib_ref] Immune response against Leishmania antigens in dogs naturally and experimentally infected with..., Rhalem [/bib_ref]. Likewise, L. chagasi cysteine proteinase Ldccys induced a higher PBMC proliferative response in AD than that obtained with amastigote extracts [bib_ref] Recombinant cysteine proteinase from Leishmania (Leishmania) chagasi implicated in human and dog..., Da Costa Pinheiro [/bib_ref]. As expected, most of the reports showed that the response of PBMCs to mitogens (polyclonal activators), such as Con A, PHA, or PWM, was high and similar between infected and noninfected animals [bib_ref] Expression of cysteine proteinase type I and II of Leishmania infantum and..., Rafati [/bib_ref] [bib_ref] Humoral and cellular immune responses against Type I cysteine proteinase of Leishmania..., Nakhaee [/bib_ref] [bib_ref] Immunologic indicators of clinical progression during canine Leishmania infantum infection, Boggiatto [/bib_ref] [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] The immunology of canine leishmaniosis: Strong evidence for a developing disease spectrum..., Cabral [/bib_ref] [bib_ref] Immune response against Leishmania antigens in dogs naturally and experimentally infected with..., Rhalem [/bib_ref] [bib_ref] Cell mediated immunity and specific IgG1 and IgG2 antibody response in natural..., Leandro [/bib_ref]. However, some authors have described a distinct mitogenic proliferative response between infected and noninfected animals [bib_ref] Early suppression of lymphoproliferative response in dogs with natural infection by Leishmania..., De Luna [/bib_ref]. This was the case of Carrillo et al., who reported a decreased proliferative response upon mitogens with disease progression.
comparative analysis of the lymphoproliferative response after SLA stimulus in AD and SD dogs. As observed in [fig_ref] Figure 2: Graphical representation of the percentage of dogs who exhibited positive lymphoproliferative response... [/fig_ref] , and even though in several studies the SLA stimulus induced a higher specific proliferative response in PBMCs from AD [bib_ref] Leishmania-specific lymphoproliferative responses and IgG1/IgG2 immunodetection patterns by Western blot in asymptomatic,..., Fernández-Pérez [/bib_ref] [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] Cellular and humoral immune responses in dogs experimentally and naturally infected with..., Pinelli [/bib_ref] , in some studies, other authors observed similar lymphoproliferative responses to this antigen in both AD and SD [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] Cell mediated immunity and specific IgG1 and IgG2 antibody response in natural..., Leandro [/bib_ref] [bib_ref] Clinical, parasitologic, and immunologic evolution in dogs experimentally infected with sand fly-derived..., Travi [/bib_ref]. PBMCs from AD, which did not proliferate upon stimulation with SLA, were shown to proliferate after exposure to other recombinant antigens, such as LeIF and Ldp23 [bib_ref] Early suppression of lymphoproliferative response in dogs with natural infection by Leishmania..., De Luna [/bib_ref]. PBMCs from experimentally infected AD also proliferate upon induction by HSP-70, PFR-2, or KMP-11 recombinant proteins, although at a lower level than that induced by SLA. Moreover, the f/t lysate produced stronger proliferation in PBMCs from AD than the recombinant antigens rCPA and rCPB and the synthetic peptide CTE [bib_ref] Expression of cysteine proteinase type I and II of Leishmania infantum and..., Rafati [/bib_ref] [bib_ref] Humoral and cellular immune responses against Type I cysteine proteinase of Leishmania..., Nakhaee [/bib_ref] but slightly lower proliferation than gp63 [bib_ref] Immune response against Leishmania antigens in dogs naturally and experimentally infected with..., Rhalem [/bib_ref]. Likewise, L. chagasi cysteine proteinase Ldccys induced a higher PBMC proliferative response in AD than that obtained with amastigote extracts [bib_ref] Recombinant cysteine proteinase from Leishmania (Leishmania) chagasi implicated in human and dog..., Da Costa Pinheiro [/bib_ref]. As expected, most of the reports showed that the response of PBMCs to mitogens (polyclonal activators), such as Con A, PHA, or PWM, was high and similar between infected and noninfected animals [bib_ref] Expression of cysteine proteinase type I and II of Leishmania infantum and..., Rafati [/bib_ref] [bib_ref] Humoral and cellular immune responses against Type I cysteine proteinase of Leishmania..., Nakhaee [/bib_ref] [bib_ref] Immunologic indicators of clinical progression during canine Leishmania infantum infection, Boggiatto [/bib_ref] [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] The immunology of canine leishmaniosis: Strong evidence for a developing disease spectrum..., Cabral [/bib_ref] [bib_ref] Immune response against Leishmania antigens in dogs naturally and experimentally infected with..., Rhalem [/bib_ref] [bib_ref] Cell mediated immunity and specific IgG1 and IgG2 antibody response in natural..., Leandro [/bib_ref]. However, some authors have described a distinct mitogenic proliferative response between infected and noninfected animals [bib_ref] Early suppression of lymphoproliferative response in dogs with natural infection by Leishmania..., De Luna [/bib_ref]. This was the case of Carrillo et al., who reported a decreased proliferative response upon mitogens with disease progression.
## Cytokine profile
The studies that analyzed the cytokine profile of PBMCs from asymptomatic dogs are listed in . Resistance to Leishmania infection has been generally associated with a Th1 response characterized by the secretion of proinflammatory cytokines [bib_ref] Leishmania infantum-specific T cell lines derived from asymptomatic dogs that lyse infected..., Pinelli [/bib_ref]. Thus, naturally and experimentally infected asymptomatic dogs have been shown to secrete high levels of IFN-γ [bib_ref] Programmed Death 1-Mediated T Cell Exhaustion during Visceral Leishmaniasis Impairs Phagocyte Function, Esch [/bib_ref] [bib_ref] Leishmania infantum-specific T cell lines derived from asymptomatic dogs that lyse infected..., Pinelli [/bib_ref] [bib_ref] Recombinant cysteine proteinase from Leishmania (Leishmania) chagasi implicated in human and dog..., Da Costa Pinheiro [/bib_ref] [bib_ref] Immunologic indicators of clinical progression during canine Leishmania infantum infection, Boggiatto [/bib_ref] [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref] , TNF-α [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref] , IL-2, and IL-18 [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref] [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref]. In fact, some authors suggested that high expression of both IFN-γ and IL-2 in dogs who stay symptomless may indicate protection against disease progression [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref]. Moreover, symptomatic dogs with severe CVL (clinical score higher than 7) showed the absence of IFN-γ, TNF-α, IL-2, IL-7, and IL-15 inflammatory mediatorsand increased levels of IL-10, CCL2, and CXCL1. Thus, dogs with clinical signs of the disease exhibit a predominant Th2 response, characterized by elevated levels of the anti-inflammatory cytokines IL-4 and IL-10 [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref] [bib_ref] Detection of intracellular IFN-γ and IL-4 cytokines in CD4+ and CD8+ T..., Matralis [/bib_ref] [bib_ref] Programmed Death 1-Mediated T Cell Exhaustion during Visceral Leishmaniasis Impairs Phagocyte Function, Esch [/bib_ref] [bib_ref] Recombinant cysteine proteinase from Leishmania (Leishmania) chagasi implicated in human and dog..., Da Costa Pinheiro [/bib_ref] [bib_ref] Immunologic indicators of clinical progression during canine Leishmania infantum infection, Boggiatto [/bib_ref] [bib_ref] Cytokine expression in dogs with natural Leishmania infantum infection, Panaro [/bib_ref] [bib_ref] Immunologic progression of canine leishmaniosis following vertical transmission in United States dogs, Vida [/bib_ref] , although significant levels of IL-4 have also been found by other authors in asymptomatic dogs that have been naturally [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref] [bib_ref] Cytokine Gene Expression in the Tissues of Dogs Infected by Leishmania infantum, Barbosa [/bib_ref] and experimentally infected [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref]. Similar findings have been reported in AD concerning IL-10, whose expression has been detected in unstimulated PBMCs [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref] , whole blood [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref] , serum [bib_ref] Immune response dynamics and lutzomyia longipalpis exposure characterize a biosignature of visceral..., Da Solcà [/bib_ref] , and Leishmania antigen-stimulated PBMCs [bib_ref] Regulatory IgD hi B Cells Suppress T Cell Function via IL-10 and..., Schaut [/bib_ref] [bib_ref] Clinical, parasitologic, and immunologic evolution in dogs experimentally infected with sand fly-derived..., Travi [/bib_ref] [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref] [bib_ref] Cytokine expression in dogs with natural Leishmania infantum infection, Panaro [/bib_ref]. However, some studies described that specific antigen stimulation did not induce IL-10 expression in any group of infected dogs (AD and SD), suggesting that this cytokine may not have a predominant negative immunoregulatory role in CVL.
Yet, the relationship between Th1 cytokines and protection against Leishmania in CVL is not clear. Thus, it has also been reported that fresh PBMCs isolated from asymptomatic dogs experimentally infected with L. infantum showed expression of TNF-α, IL-2, IFN-γ, IL-10 and IL-18 mRNAs similar to those from noninfected dogs [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref] [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref]. Likewise, some studies have called into question the role of IFN-γ alone as a resistance marker, as similar or higher levels of this cytokine were found in symptomatic dogs than in asymptomatic dogs [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref] [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref] [bib_ref] Cytokine Gene Expression in the Tissues of Dogs Infected by Leishmania infantum, Barbosa [/bib_ref] [bib_ref] Detection of intracellular IFN-γ and IL-4 cytokines in CD4+ and CD8+ T..., Matralis [/bib_ref] [bib_ref] Regulatory T cells, Cytotoxic T lymphocytes and a Th1 cytokine profile in..., Cortese [/bib_ref] [bib_ref] Clinical, parasitologic, and immunologic evolution in dogs experimentally infected with sand fly-derived..., Travi [/bib_ref] [bib_ref] Cytokine expression in dogs with natural Leishmania infantum infection, Panaro [/bib_ref] [bib_ref] Immune response dynamics and lutzomyia longipalpis exposure characterize a biosignature of visceral..., Da Solcà [/bib_ref]. Similarly, higher levels of IL-2 [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref] [bib_ref] Cytokine Gene Expression in the Tissues of Dogs Infected by Leishmania infantum, Barbosa [/bib_ref] and TNF-α [bib_ref] Cytokine expression in dogs with natural Leishmania infantum infection, Panaro [/bib_ref] have been detected in SD, as well as IL-18, although an association of IL-18 with resistance or susceptibility could not be established [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref].
## Analysis of the humoral immune response
The humoral response in CVL and the immunoglobulin subclasses that predominate in asymptomatic or symptomatic forms of the disease have been quite controversial. One of the most discussed topics is the correlation between IgG subclasses (IgG1 and IgG2) and disease progression, whose literature has been reviewed in . Thus, while some researchers observed a correlation between IgG1 and asymptomatic infection [bib_ref] Isotype patterns of immunoglobulins: Hallmarks for clinical status and tissue parasite density..., Reis [/bib_ref] [bib_ref] IgG subclass profile of serum antibodies to Leishmania chagasi in naturally infected..., Oliveira [/bib_ref] [bib_ref] Relationship of Leishmania-specific IgG levels and IgG avidity with parasite density and..., Neto [/bib_ref] [bib_ref] Profile of anti-Leishmania antibodies related to clinical picture in canine visceral leishmaniasis, De Freitas [/bib_ref] , others have proposed IgG1 as a susceptibility marker associated with the appearance of clinical signs [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] Immunoglobulin G and E responses in various stages of canine leishmaniosis, Iniesta [/bib_ref] [bib_ref] Serological and infection statuses of dogs from a visceral leishmaniasis-endemic area, Laranjeira [/bib_ref] [bib_ref] Analysis of the humoral immune response against total and recombinant antigens of..., Nieto [/bib_ref] [bib_ref] Leishmania infantum virulence factor A2 protein: Linear B-cell epitope mapping and identification..., Campos [/bib_ref] [bib_ref] Identification of immunoreactive Leishmania infantum protein antigens to asymptomatic dog sera through..., Agallou [/bib_ref] [bib_ref] Leishmania infantum-specific IgG, IgG1 and IgG2 antibody responses in healthy and ill..., Solano-Gallego [/bib_ref]. While IgG2 has also been associated with immune protection in AD, presenting high levels in these dogs [bib_ref] Specific IgG1 and IgG2 antibody and lymphocyte subset levels in naturally Leishmania..., Bourdoiseau [/bib_ref] [bib_ref] Humoral and cellular immune responses against Type I cysteine proteinase of Leishmania..., Nakhaee [/bib_ref] [bib_ref] Leishmania infantum virulence factor A2 protein: Linear B-cell epitope mapping and identification..., Campos [/bib_ref] , some studies detected higher levels of IgG2 in symptomatic animals [bib_ref] Canine visceral leishmaniasis: Relationships between clinical status, humoral immune response, haematology and..., Da Costa-Val [/bib_ref] [bib_ref] Leishmania-specific lymphoproliferative responses and IgG1/IgG2 immunodetection patterns by Western blot in asymptomatic,..., Fernández-Pérez [/bib_ref] [bib_ref] Cell mediated immunity and specific IgG1 and IgG2 antibody response in natural..., Leandro [/bib_ref] [bib_ref] Profile of anti-Leishmania antibodies related to clinical picture in canine visceral leishmaniasis, De Freitas [/bib_ref] [bib_ref] Leishmania infantum-specific IgG, IgG1 and IgG2 antibody responses in healthy and ill..., Solano-Gallego [/bib_ref] [bib_ref] Identification of antibodies to Leishmania silent information regulatory 2 (SIR2) protein homologue..., Cordeiro-Da-Silva [/bib_ref] [bib_ref] Antileishmanial antibody profile in dogs naturally infected with Leishmania chagasi, Almeida [/bib_ref] [bib_ref] Anti-Leishmania humoral and cellular immune responses in naturally infected symptomatic and asymptomatic..., Cardoso [/bib_ref] [bib_ref] Idiotype expression of IgG1 and IgG2 in dogs naturally infected with Leishmania..., Iniesta [/bib_ref]. However, this dichotomy in the levels of IgG1 and IgG2 responses in symptomatic and asymptomatic dogs was not observed in other studies, in which no correlation was found between these subclasses and the clinical status of Leishmania-infected dogs [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref] [bib_ref] Immunologic indicators of clinical progression during canine Leishmania infantum infection, Boggiatto [/bib_ref] [bib_ref] Clinical, parasitologic, and immunologic evolution in dogs experimentally infected with sand fly-derived..., Travi [/bib_ref] [bib_ref] IgG subclass responses in a longitudinal study of canine visceral leishmaniasis, Quinnell [/bib_ref] [bib_ref] Comparison of monoclonal and polyclonal antibodies for the detection of canine IgG1..., Carson [/bib_ref] [bib_ref] Use of ELISA employing homologous and heterologous antigens for the dectection of..., Ribeiro [/bib_ref] [bib_ref] Comparative analysis of the Leishmania infantum-specific antibody repertoires and the autoantibody repertoires..., Chaabouni [/bib_ref]. Interestingly, a recent study suggests the possibility that both IgG1 and IgG2 subclasses were associated with immune-protective mechanisms against Leishmania infection [bib_ref] Leishmania infantum nucleoside triphosphate diphosphohydrolase 1 (NTPDase 1) B-domain: Antibody antiproliferative effect..., Maia [/bib_ref].
The use of polyclonal or monoclonal antibodies that recognize the different IgG subclasses could be the reason for the published contradictory results regarding their association with susceptibility or resistance in CVL. Quinnell et al., using monoclonal antisera, reported significant increases in IgG1, IgG2, IgG3, and IgG4 in polysymptomatic dogs from an endemic area of Brazil [bib_ref] IgG subclass responses in a longitudinal study of canine visceral leishmaniasis, Quinnell [/bib_ref]. Oliveira et al., however, observed increased levels of IgG1 and IgG4 in AD versus SD using monoclonal antibodies [bib_ref] IgG subclass profile of serum antibodies to Leishmania chagasi in naturally infected..., Oliveira [/bib_ref]. Studies carried out in endemic zones of Brazil and in Iowa (USA) showed higher levels of IgG, IgG1, and IgG2 in symptomatic naturally infected dogs using polyclonal antibodies [bib_ref] Immunologic indicators of clinical progression during canine Leishmania infantum infection, Boggiatto [/bib_ref] [bib_ref] Use of ELISA employing homologous and heterologous antigens for the dectection of..., Ribeiro [/bib_ref]. This variability in the obtained results using both monoclonal and polyclonal antibodies was also described by Marcondes et al. [bib_ref] Temporal IgG subclasses response in dogs following vaccination against Leishmania with Leishmune..., Marcondes [/bib_ref]. These authors found no significant differences in the levels of IgG1 and IgG2 between AD and SD dogs when using polyclonal antibodies and instead observed higher levels of all IgG subclasses except IgG2 in SD when using monoclonal antibodies [bib_ref] Temporal IgG subclasses response in dogs following vaccination against Leishmania with Leishmune..., Marcondes [/bib_ref]. In addition, Travi et al. used polyclonal antisera and did not find any significant difference in the IgG1 and IgG2 levels from experimentally infected asymptomatic and symptomatic dogs [bib_ref] Clinical, parasitologic, and immunologic evolution in dogs experimentally infected with sand fly-derived..., Travi [/bib_ref]. Interestingly, given the inconsistencies in the results despite the type of antisera used, a group of researchers pointed out the use of whole promastigote extracts or SLA in ELISA as a source of nuclear and cytoplasmic components that could create nonspecific binding of IgG subclasses [bib_ref] Comparative analysis of the Leishmania infantum-specific antibody repertoires and the autoantibody repertoires..., Chaabouni [/bib_ref]. For that purpose, they tested both SLA and the recombinant proteins LACK and LeIF in dogs from endemic areas in Tunisia and found higher levels of IgG, IgG1, and IgG2 in SD dogs, regardless of the antigen used [bib_ref] Comparative analysis of the Leishmania infantum-specific antibody repertoires and the autoantibody repertoires..., Chaabouni [/bib_ref].
IgM, IgA, and IgE levels have also been discussed in the literature. Symptomatic dogs have shown higher expression of IgE [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref] [bib_ref] Isotype patterns of immunoglobulins: Hallmarks for clinical status and tissue parasite density..., Reis [/bib_ref] [bib_ref] Immunoglobulin G and E responses in various stages of canine leishmaniosis, Iniesta [/bib_ref] [bib_ref] Antileishmanial antibody profile in dogs naturally infected with Leishmania chagasi, Almeida [/bib_ref] , IgA [bib_ref] Leishmania-specific isotype levels and their relationship with specific cell-mediated immunity parameters in..., Rodríguez-Cortés [/bib_ref] [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] Isotype patterns of immunoglobulins: Hallmarks for clinical status and tissue parasite density..., Reis [/bib_ref] , and IgM [bib_ref] Leishmania-specific isotype levels and their relationship with specific cell-mediated immunity parameters in..., Rodríguez-Cortés [/bib_ref] [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] Profile of anti-Leishmania antibodies related to clinical picture in canine visceral leishmaniasis, De Freitas [/bib_ref] than asymptomatic dogs, which implies the failure of these isotypes to provide immunoprotection against L. infantum. In contrast, other studies reported no correlation between IgE and IgM and clinical status [bib_ref] Isotype patterns of immunoglobulins: Hallmarks for clinical status and tissue parasite density..., Reis [/bib_ref] [bib_ref] Profile of anti-Leishmania antibodies related to clinical picture in canine visceral leishmaniasis, De Freitas [/bib_ref] [bib_ref] Serological and infection statuses of dogs from a visceral leishmaniasis-endemic area, Laranjeira [/bib_ref] [bib_ref] Identification of antibodies to Leishmania silent information regulatory 2 (SIR2) protein homologue..., Cordeiro-Da-Silva [/bib_ref] and showed high levels of IgA and IgM in both symptomatic and asymptomatic dogs [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref] [bib_ref] Profile of anti-Leishmania antibodies related to clinical picture in canine visceral leishmaniasis, De Freitas [/bib_ref].
# Discussion
Infection caused by Leishmania infantum induces host defense reactions to infection involving effector mechanisms of the innate and acquired immune responses. Given the status of L. infantum as an obligate intracellular infectious agent, the cellular response, mainly mediated by T lymphocytes, plays a critical role in infection control [bib_ref] The dialogue of the host-parasite relationship: Leishmania spp. and Trypanosoma cruzi Infection, De Morais [/bib_ref]. Thus, these cells of the immune system recognize parasite antigens and promote the specific functions necessary for their elimination. Resistance to canine visceral leishmaniasis (CVL) seems to be associated with higher levels of total T lymphocytes (CD4 + and CD8 + ) and B cells [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref] [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref] [bib_ref] Phenotypic features of circulating leucocytes as immunological markers for clinical status and..., Reis [/bib_ref] [bib_ref] Clinical forms of canine visceral leishmaniasis in naturally Leishmania infantum-infected dogs and..., Nicolato [/bib_ref] [bib_ref] Specific IgG1 and IgG2 antibody and lymphocyte subset levels in naturally Leishmania..., Bourdoiseau [/bib_ref] [bib_ref] Detection of intracellular IFN-γ and IL-4 cytokines in CD4+ and CD8+ T..., Matralis [/bib_ref] , whereas susceptibility to the disease is related to a decreased number of these cells. CD8 + T cells have an important protective role during CVL, not only because of their ability to mount a protective Th1 response during the early stages of infection but also due to their cytolytic activity against L. infantum-infected macrophages [bib_ref] Leishmania infantum-specific T cell lines derived from asymptomatic dogs that lyse infected..., Pinelli [/bib_ref]. In the asymptomatic stage (AD) of visceral leishmaniasis, CD4 + T cells also play a role by being able to lyse infected macrophages, although the relevance of this subtype of cytotoxic activity for visceral leishmaniasis in vivo is not yet well known [bib_ref] Leishmania infantum-specific T cell lines derived from asymptomatic dogs that lyse infected..., Pinelli [/bib_ref]. Circulating lymphocytes of asymptomatic dogs also presented elevated expression of MHC-II and a higher CD45RB/CD45RA ratio [bib_ref] Phenotypic features of circulating leucocytes as immunological markers for clinical status and..., Reis [/bib_ref] [bib_ref] Canine leishmaniosis. Immunophenotypic profile of leukocytes in different compartments of symptomatic, asymptomatic..., Alexandre-Pires [/bib_ref]. CD45RB has previously been related to CD4 + and CD8 + T cells activated by protozoan infection [bib_ref] Dynamics of CD62L/CD45RB CD4+ and CD8+ lymphocyte subsets in hepatic and splenic..., Gomes-Pereira [/bib_ref] , whereas CD45RA seems to be highly expressed by naïve canine T cells, helper T cells secreting IFN-γ, and a wide range of B cells [bib_ref] Phenotypic features of circulating leucocytes as immunological markers for clinical status and..., Reis [/bib_ref]. These results suggest enhanced antigen presentation ability and the effective activation of T cells. The reduced presence of regulatory CD4 + T cells (Treg) found in infected dogs would also optimize T-cell activation and effector functions during infection [bib_ref] Regulatory T cells, Cytotoxic T lymphocytes and a Th1 cytokine profile in..., Cortese [/bib_ref] , as these Treg cells can suppress the antiparasitic CD4 + T-cell response [bib_ref] Distinct Roles for CD4 + Foxp3 + Regulatory T Cells and IL-10-Mediated..., Bunn [/bib_ref]. However, although low levels of Tregs in asymptomatic dogs would allow stronger control of parasite growth, the activation of this immunoregulatory mechanism is critical to protect tissues from damage caused by excessive inflammation [bib_ref] Regulatory T cells, Cytotoxic T lymphocytes and a Th1 cytokine profile in..., Cortese [/bib_ref] [bib_ref] Distinct Roles for CD4 + Foxp3 + Regulatory T Cells and IL-10-Mediated..., Bunn [/bib_ref]. The regulation of the immune response is also influenced by regulatory (IgD hi ) B cells, which, unlike B cells, were found at low levels in AD dogs but were increased in symptomatic dogs. It has been described that during symptomatic infection by Leishmania, IgD hi B cells produce IL-10 and suppress IFN-γ production in T cells through the PD-L1/PD-1 and IL-10 pathways, leading to the suppression of the T-cell response and cellular exhaustion in these dogs [bib_ref] Regulatory IgD hi B Cells Suppress T Cell Function via IL-10 and..., Schaut [/bib_ref]. In fact, the inhibitory receptor PD-1 has been found to be expressed at higher levels in SD than in AD [bib_ref] Programmed Death 1-Mediated T Cell Exhaustion during Visceral Leishmaniasis Impairs Phagocyte Function, Esch [/bib_ref]. This receptor, which is involved in the negative regulation of T-cell activation, seems to partially mediate CD8 + and CD4 + T-cell exhaustion in CVL [bib_ref] Programmed Death 1-Mediated T Cell Exhaustion during Visceral Leishmaniasis Impairs Phagocyte Function, Esch [/bib_ref].
Protective immunity against L. infantum in dogs has also been associated with a strong lymphoproliferative response of PBMCs to Leishmania antigen [bib_ref] Leishmania infantum-specific T cell lines derived from asymptomatic dogs that lyse infected..., Pinelli [/bib_ref]. Most of the studies analyzed in this review showed a higher proliferation of PBMCs from experimental and naturally infected asymptomatic dogs in response to several L. infantum antigens [bib_ref] Programmed Death 1-Mediated T Cell Exhaustion during Visceral Leishmaniasis Impairs Phagocyte Function, Esch [/bib_ref] [bib_ref] Leishmania infantum-specific T cell lines derived from asymptomatic dogs that lyse infected..., Pinelli [/bib_ref] [bib_ref] Recombinant cysteine proteinase from Leishmania (Leishmania) chagasi implicated in human and dog..., Da Costa Pinheiro [/bib_ref] [bib_ref] Early suppression of lymphoproliferative response in dogs with natural infection by Leishmania..., De Luna [/bib_ref] [bib_ref] Leishmania-specific lymphoproliferative responses and IgG1/IgG2 immunodetection patterns by Western blot in asymptomatic,..., Fernández-Pérez [/bib_ref] [bib_ref] Expression of cysteine proteinase type I and II of Leishmania infantum and..., Rafati [/bib_ref] [bib_ref] Humoral and cellular immune responses against Type I cysteine proteinase of Leishmania..., Nakhaee [/bib_ref] [bib_ref] Immunologic indicators of clinical progression during canine Leishmania infantum infection, Boggiatto [/bib_ref] [bib_ref] Leishmania-specific isotype levels and their relationship with specific cell-mediated immunity parameters in..., Rodríguez-Cortés [/bib_ref] [bib_ref] The immunology of canine leishmaniosis: Strong evidence for a developing disease spectrum..., Cabral [/bib_ref] [bib_ref] Immune response against Leishmania antigens in dogs naturally and experimentally infected with..., Rhalem [/bib_ref]. In contrast, PBMCs from symptomatic dogs failed to respond to these parasite antigens or showed lower cell proliferative responses. In asymptomatic dogs, it has been suggested that those that show a poor cellular response are more prone to progress in the disease than dogs with stronger cell-mediated immunity [bib_ref] Leishmania-specific isotype levels and their relationship with specific cell-mediated immunity parameters in..., Rodríguez-Cortés [/bib_ref]. Likewise, the inability to mount an effective and specific proliferative response observed in SD dogs would be indicative of the immune suppression that has been reported in these animals [bib_ref] Expression of cysteine proteinase type I and II of Leishmania infantum and..., Rafati [/bib_ref]. Some authors have described that this unresponsiveness occurred only in later stages of the disease and that, in early infection, symptomatic dogs were able to develop a proliferative cellular response to leishmanial antigens [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] Cell mediated immunity and specific IgG1 and IgG2 antibody response in natural..., Leandro [/bib_ref] [bib_ref] Clinical, parasitologic, and immunologic evolution in dogs experimentally infected with sand fly-derived..., Travi [/bib_ref] , although these studies have been conducted in a low number of dogs.
Concerning the cytokine profile, asymptomatic dogs seem to have a predominant Th1 response. In most of the studies published, these AD dogs showed high levels of one or more of these proinflammatory cytokines: IFN-γ [bib_ref] Programmed Death 1-Mediated T Cell Exhaustion during Visceral Leishmaniasis Impairs Phagocyte Function, Esch [/bib_ref] [bib_ref] Leishmania infantum-specific T cell lines derived from asymptomatic dogs that lyse infected..., Pinelli [/bib_ref] [bib_ref] Recombinant cysteine proteinase from Leishmania (Leishmania) chagasi implicated in human and dog..., Da Costa Pinheiro [/bib_ref] [bib_ref] Immunologic indicators of clinical progression during canine Leishmania infantum infection, Boggiatto [/bib_ref] [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref] , TNF-α [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref] , IL-2 [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref] [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref] , IL-18 [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref] , IL-6, IL-15, and IL-7. The IFN-γ, TNF-α, and IL-2 cytokines have been proven to activate macrophages to keep the parasite under control and to avoid its dissemination through the production of nitric oxide (NO) and reactive oxygen species (ROS). Conversely, the decreased NO levels in SD may be related to the inhibitory effect on signal transduction for iNOS and NO production induced by the anti-inflammatory cytokines released in the symptomatic active form of the disease, such as IL-4, IL-10, IL-13, or TGF-β. In fact, both IL-4 and IL-10 have been shown to inhibit the expression of the enzyme iNOS2, downregulating macrophage activity and allowing the persistence of parasites in blood and their transmission [bib_ref] Cytokine expression in dogs with natural Leishmania infantum infection, Panaro [/bib_ref]. IFN-γ and TNF-α participate in the regulation and activation of inducible nitric oxide synthase (iNOS), the enzyme responsible for NO production in macrophages. NO expression may have a protective role in asymptomatic dogs, although this molecule cannot be considered a resistance marker since high levels of NO expression have been detected in both symptomatic and asymptomatic animals [bib_ref] A potential link among antioxidant enzymes, histopathology and trace elements in canine..., Souza [/bib_ref]. IL-2, in addition to being involved in macrophage activation, may also be implicated in decreasing the adverse effects of the inflammatory response, as this cytokine has been shown to regulate the production of immunoglobulins by B cells and the differentiation of regulatory T cells [bib_ref] Cytokine Gene Expression in the Tissues of Dogs Infected by Leishmania infantum, Barbosa [/bib_ref]. IL-18 expression has also been associated with resistance to canine leishmaniasis by inducing Th1-cell development, IFN-γ production, and the activation of T-and NK-cell cytotoxicity [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref]. The cytokine IL-6 has been described to be involved in the regulation of IFN-γ receptor expression [bib_ref] Interferon (IFN)-γ, tumor necrosis factor-α, interleukin-6, and IFN-γ receptor 1 are the..., Ansari [/bib_ref] , and IL-7R seems to play a relevant role in T-cell survival [bib_ref] IL-7 receptor expression provides the potential for long-term survival of both CD62Lhigh..., Colpitts [/bib_ref]. The cytokine IL-15 has also been shown to participate in the control of CVL infection in asymptomatic dogs. In fact, IL-15 has the ability, in association with IL-12, to activate a strong proliferative response, promoting a decrease in programmed cell death protein 1 (PD-1) expression in lymphocytes as well as increases in the expression levels of the cytokines IFN-γ and TNF-α [bib_ref] Combined in vitro il-12 and il-15 stimulation promotes cellular immune response in..., Costa [/bib_ref] [bib_ref] IL-15 in human visceral leishmaniasis caused by Leishmania infantum, Milano [/bib_ref].
The relationship between the cellular Th1/Th2 response in CVL and its association with resistance and susceptibility may not be as clear in dogs as in other species [bib_ref] Immunoglobulin G and E responses in various stages of canine leishmaniosis, Iniesta [/bib_ref]. Thus, Th1 cytokines, usually related to asymptomatic infection, have been detected in symptomatic dogs [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref] [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref] [bib_ref] Cytokine Gene Expression in the Tissues of Dogs Infected by Leishmania infantum, Barbosa [/bib_ref] [bib_ref] Regulatory T cells, Cytotoxic T lymphocytes and a Th1 cytokine profile in..., Cortese [/bib_ref] [bib_ref] Clinical, parasitologic, and immunologic evolution in dogs experimentally infected with sand fly-derived..., Travi [/bib_ref] [bib_ref] Cytokine expression in dogs with natural Leishmania infantum infection, Panaro [/bib_ref] , and Th2 cytokines, usually associated with clinical disease, have been detected in asymptomatic dogs [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref] [bib_ref] Cytokine Gene Expression in the Tissues of Dogs Infected by Leishmania infantum, Barbosa [/bib_ref] [bib_ref] Regulatory IgD hi B Cells Suppress T Cell Function via IL-10 and..., Schaut [/bib_ref] [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref]. The Th2-type cytokine IL-10, which is associated with the suppression of cytokine production by Th1 cells and consequently with the development of a Th2 immune response [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref] , has been described could be overexpressed in IFN-γ-producing dogs as a negative feedback mechanism to control proinflammatory cytokines and reduce their detrimental effects on dog health. Likewise, the expression of proinflammatory cytokines such as IL-12, IFN-γ, and TNF-α in dogs with CVL has also been called into question, as some studies have also shown that they may be involved in disease progression [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref] [bib_ref] Cytokine Gene Expression in the Tissues of Dogs Infected by Leishmania infantum, Barbosa [/bib_ref] [bib_ref] Cytokine expression in dogs with natural Leishmania infantum infection, Panaro [/bib_ref].
In the context of the evolution of CVL in dogs and in addition to cell-mediated immunity, another important factor to be considered is the humoral response and, particularly, the correlation between IgG subclasses (IgG1 and IgG2) specific to Leishmania antigens. Numerous studies have pointed to increased levels of anti-Leishmania IgG1 as a determinant factor for the symptomatic evolution of the disease [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] Immunoglobulin G and E responses in various stages of canine leishmaniosis, Iniesta [/bib_ref] [bib_ref] Serological and infection statuses of dogs from a visceral leishmaniasis-endemic area, Laranjeira [/bib_ref] [bib_ref] Analysis of the humoral immune response against total and recombinant antigens of..., Nieto [/bib_ref] [bib_ref] Leishmania infantum virulence factor A2 protein: Linear B-cell epitope mapping and identification..., Campos [/bib_ref] [bib_ref] Identification of immunoreactive Leishmania infantum protein antigens to asymptomatic dog sera through..., Agallou [/bib_ref] [bib_ref] Leishmania infantum-specific IgG, IgG1 and IgG2 antibody responses in healthy and ill..., Solano-Gallego [/bib_ref]. It has been postulated that the ability of IgG1 to activate complement could contribute to increased pathologic manifestation of CVL in dogs, as this immune mechanism mediates inflammatory reactions [bib_ref] Analysis of the humoral immune response against total and recombinant antigens of..., Nieto [/bib_ref]. However, other authors found no association between the presence of symptomatology and high IgG1 levels. IgG2 isotype antibodies have been associated with immune protection mechanisms against L. infantum infection [bib_ref] Specific IgG1 and IgG2 antibody and lymphocyte subset levels in naturally Leishmania..., Bourdoiseau [/bib_ref] [bib_ref] Humoral and cellular immune responses against Type I cysteine proteinase of Leishmania..., Nakhaee [/bib_ref] [bib_ref] Leishmania infantum virulence factor A2 protein: Linear B-cell epitope mapping and identification..., Campos [/bib_ref] and have also been correlated with clinical symptoms of CVL [bib_ref] Canine visceral leishmaniasis: Relationships between clinical status, humoral immune response, haematology and..., Da Costa-Val [/bib_ref] [bib_ref] Leishmania-specific lymphoproliferative responses and IgG1/IgG2 immunodetection patterns by Western blot in asymptomatic,..., Fernández-Pérez [/bib_ref] [bib_ref] Cell mediated immunity and specific IgG1 and IgG2 antibody response in natural..., Leandro [/bib_ref] [bib_ref] Isotype patterns of immunoglobulins: Hallmarks for clinical status and tissue parasite density..., Reis [/bib_ref] [bib_ref] Relationship of Leishmania-specific IgG levels and IgG avidity with parasite density and..., Neto [/bib_ref] [bib_ref] Profile of anti-Leishmania antibodies related to clinical picture in canine visceral leishmaniasis, De Freitas [/bib_ref] [bib_ref] Leishmania infantum-specific IgG, IgG1 and IgG2 antibody responses in healthy and ill..., Solano-Gallego [/bib_ref] [bib_ref] Identification of antibodies to Leishmania silent information regulatory 2 (SIR2) protein homologue..., Cordeiro-Da-Silva [/bib_ref] [bib_ref] Antileishmanial antibody profile in dogs naturally infected with Leishmania chagasi, Almeida [/bib_ref] [bib_ref] Anti-Leishmania humoral and cellular immune responses in naturally infected symptomatic and asymptomatic..., Cardoso [/bib_ref] [bib_ref] Idiotype expression of IgG1 and IgG2 in dogs naturally infected with Leishmania..., Iniesta [/bib_ref]. The study of anti-Leishmania IgG subclass antibody production in a cohort of naturally infected dogs showed that the levels of all IgG subclasses were strongly intercorrelated and particularly elevated in sick dogs in which the presence of the parasite was detected by PCR. Thus, these results suggest that the evolution of CVL may be associated with the upregulation of antigen-specific antibodies of all IgG subclasses, particularly IgG1, IgG3, and IgG4 [bib_ref] IgG subclass responses in a longitudinal study of canine visceral leishmaniasis, Quinnell [/bib_ref].
This strong but nonprotective humoral response observed by several authors in symptomatic dogs [bib_ref] Specific IgG1 and IgG2 antibody and lymphocyte subset levels in naturally Leishmania..., Bourdoiseau [/bib_ref] [bib_ref] Immunologic indicators of clinical progression during canine Leishmania infantum infection, Boggiatto [/bib_ref] [bib_ref] Use of ELISA employing homologous and heterologous antigens for the dectection of..., Ribeiro [/bib_ref] [bib_ref] Comparative analysis of the Leishmania infantum-specific antibody repertoires and the autoantibody repertoires..., Chaabouni [/bib_ref] would also be expected, as one of the characteristics of symptomatic disease is polyclonal B-cell activation [bib_ref] Immunologic indicators of clinical progression during canine Leishmania infantum infection, Boggiatto [/bib_ref]. It should be noted, however, that most of the articles that correlated IgG subclasses to susceptibility or resistance to sickness used polyclonal antisera for antibody detection. In contrast, the studies in which a monoclonal panel of antibodies was employed did not report a polarized response of the IgG subclass in dogs but showed a general increase in all IgG isotypes with disease progression [bib_ref] IgG subclass responses in a longitudinal study of canine visceral leishmaniasis, Quinnell [/bib_ref] [bib_ref] Comparison of monoclonal and polyclonal antibodies for the detection of canine IgG1..., Carson [/bib_ref] [bib_ref] Temporal IgG subclasses response in dogs following vaccination against Leishmania with Leishmune..., Marcondes [/bib_ref]. Although some authors related these contradictory results with the use of polyclonal or monoclonal anti-Igs, the presence of cross-reactive autoantibodies, largely produced in this disease, may also influence the degree of specificity of the Leishmania-antigenic preparations and the results obtained [bib_ref] Comparative analysis of the Leishmania infantum-specific antibody repertoires and the autoantibody repertoires..., Chaabouni [/bib_ref]. Thus, the repertoires of autoantibodies against extracts of HEp-2, ds-DNA, human albumin, and transferrin as autoantigens indicated that in AD dogs, there are higher levels of IgG1 autoantibodies and a higher seroprevalence than in SD dogs, in which there are lower levels and lower seroprevalences of total IgG and IgG2 [bib_ref] Comparative analysis of the Leishmania infantum-specific antibody repertoires and the autoantibody repertoires..., Chaabouni [/bib_ref]. Moreover, data from competitive HEp-2-ELISA using total leishmanial antigens as inhibitors showed that in AD, IgG1 antibodies are predominantly autoantibodies to self-antigens, whereas in SD, they are mainly cross-reactive (Leishmania/self-antigens) [bib_ref] Comparative analysis of the Leishmania infantum-specific antibody repertoires and the autoantibody repertoires..., Chaabouni [/bib_ref]. Regarding the role of other immunoglobulins in CVL, such as IgE, IgA, and IgM, most of the studies reported high levels in symptomatic dogs. IgE, which is considered a serum marker of Th2 in different parasite infections, seems to be correlated with the symptomatic stage of CLV. In fact, higher expression of IgG1 and IgE was only present in symptomatic animals. This correlation between the expression of IgG1 and IgE and the pathology of leishmaniasis points to their potential role as markers of active disease [bib_ref] Immunoglobulin G and E responses in various stages of canine leishmaniosis, Iniesta [/bib_ref] [bib_ref] Antileishmanial antibody profile in dogs naturally infected with Leishmania chagasi, Almeida [/bib_ref]. The profile of anti-Leishmania antibodies in different clinical forms of canine visceral leishmaniasis (CVL) in naturally infected dogs was studied by Freitas et al. [bib_ref] Profile of anti-Leishmania antibodies related to clinical picture in canine visceral leishmaniasis, De Freitas [/bib_ref] , who showed that both asymptomatic and symptomatic dogs presented increased levels of total IgG, IgA, and IgE in addition to IgG1 and IgG2. Moreover, IgG2 and IgM presented positive correlations with the clinical signs of the disease, while total IgG, IgG1, and IgA showed negative correlations. The increase in IgE did not show a correlation with the clinical changes in infected dogs [bib_ref] Profile of anti-Leishmania antibodies related to clinical picture in canine visceral leishmaniasis, De Freitas [/bib_ref]. However, demonstrated a positive correlation of patterns of IgA with the clinical status of naturally infected animals [bib_ref] Isotype patterns of immunoglobulins: Hallmarks for clinical status and tissue parasite density..., Reis [/bib_ref]. Increased production of IgA, which is involved in mucosal immunity, was described in infected dogs showing symptomatology, suggesting that the worsening clinical condition in dogs has also been linked to elevated IgA levels [bib_ref] Leishmania-specific isotype levels and their relationship with specific cell-mediated immunity parameters in..., Rodríguez-Cortés [/bib_ref] [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref]. Additionally, it suggests that dogs developing a high T-cell response are probably able to avoid the dissemination of the parasite to mucosal surfaces and, as a consequence, to produce low or background specific IgA levels [bib_ref] Leishmania-specific isotype levels and their relationship with specific cell-mediated immunity parameters in..., Rodríguez-Cortés [/bib_ref] [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref]. Further studies are needed to investigate the relationship between specific IgA and parasite load, especially at mucosal sites. Furthermore, the discovery of IgA deposits in the kidneys of infected dogs has suggested that this immunoglobulin may contribute to the generation of glomerulonephritis associated with this disease [bib_ref] Isotype patterns of immunoglobulins: Hallmarks for clinical status and tissue parasite density..., Reis [/bib_ref]. Regarding IgM, although this immunoglobulin has been typically associated with the acute forms of infectious diseases, significant levels have been detected in the chronic phases of CVL [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref] [bib_ref] Leishmania-specific isotype levels and their relationship with specific cell-mediated immunity parameters in..., Rodríguez-Cortés [/bib_ref] [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] Isotype patterns of immunoglobulins: Hallmarks for clinical status and tissue parasite density..., Reis [/bib_ref] [bib_ref] Profile of anti-Leishmania antibodies related to clinical picture in canine visceral leishmaniasis, De Freitas [/bib_ref]. It has also been described that serum levels of anti-Leishmania IgM from naturally infected dogs (AD and SD) remain, with no significant differences compared to those from the noninfected control group [bib_ref] Profile of anti-Leishmania antibodies related to clinical picture in canine visceral leishmaniasis, De Freitas [/bib_ref] , although, despite this finding, the authors observed a positive correlation with respect to the association with symptomatology [bib_ref] Profile of anti-Leishmania antibodies related to clinical picture in canine visceral leishmaniasis, De Freitas [/bib_ref].
Although there are numerous findings that point to the existence of a different immune response in asymptomatic versus symptomatic dogs, there are relevant discrepancies in the results obtained in the different studies carried out in this regard. Thus, it is essential to continue the research in this context and to take into account variables such as the breed of the dog, the genotype of the infectious agent, the coexistence with other pathologies, the type of infection (natural or experimental), as well as the sensitivity of the techniques used to evaluate the antigen-specific immune response induced by the parasite infection. Likewise, we also consider it necessary to carry out more longitudinal studies in infected dogs in order to evaluate the kinetics of the immune response throughout the infection and its association with the control of the pathology. Altogether, it will make it possible to find clear patterns capable of predicting the outcome of the infection and that are useful as biomarkers of evolution and as activation molecules for the design of therapeutic and/or preventive vaccines. In this context, we believe that the identification of biomarkers associated with asymptomatic or symptomatic CVL would allow us to monitor the efficacy of therapies and vaccines and to develop better diagnostic tests.
It is also relevant to highlight that most of the reported studies establish the clinical stage of the infected dog based solely on physical examination. This is of limited value as dogs without apparent clinical manifestations may be classified as asymptomatic despite having relevant alterations in serum and urinary biochemical parameters and/or some organ dysfunction [bib_ref] When is an "asymptomatic" dog asymptomatic?, Dantas-Torres [/bib_ref] [bib_ref] Canine leishmaniosis-A challenging zoonosis, Solano-Gallego [/bib_ref]. Another aspect to consider is that most of the studies only included asymptomatic dogs with positive serological tests, excluding infected animals with low titers of anti-Leishmania antibodies, whose infection status can only be detected by PCR [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref]. Some studies have described that asymptomatic dogs with positive or negative serology showed a differential humoral and cellular responses to Leishmania antigens [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref] [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref]. Thus, the inclusion of the asymptomatic dogs with low titers of antibodies would be critical for understanding both the complex immune response triggered by infection and the factors involved in the symptomatology progression of canine visceral leishmaniasis. Overall, all these findings reinforce the idea that CVL is a complex multifactorial disease that is affected by a set of factors that are correlated and should not be evaluated in an isolated manner.
Supplementary Materials: The following supporting information can be downloaded at https://www. mdpi.com/article/10.3390/vaccines10060947/s1: [fig_ref] Table 1: Studies that analyzed the immune profiles of PBMCs from infected and control... [/fig_ref] : Keywords used for systematic searches in the Medline database (PubMed); : Keywords used in the secondary literature review in the Medline database (PubMed); : Studies that analyze the lymphoproliferative response of PBMCs to Leishmania antigens from infected and control dogs; : Studies that analyze the cytokine production profiles of infected and control dogs; and : Studies that analyze the humoral responses in infected and control dogs. (See [bib_ref] Immunological profile of resistance and susceptibility in naturally infected dogs by Leishmania..., Leal [/bib_ref] [bib_ref] Humoral and cellular immune responses in dogs with inapparent natural Leishmania infantum..., Coura-Vital [/bib_ref] [bib_ref] Canine visceral leishmaniasis: Relationships between clinical status, humoral immune response, haematology and..., Da Costa-Val [/bib_ref] [bib_ref] Leishmania DNA load and cytokine expression levels in asymptomatic naturally infected dogs, Manna [/bib_ref] [bib_ref] Cytokine Gene Expression in the Tissues of Dogs Infected by Leishmania infantum, Barbosa [/bib_ref] [bib_ref] Specific IgG1 and IgG2 antibody and lymphocyte subset levels in naturally Leishmania..., Bourdoiseau [/bib_ref] [bib_ref] Detection of intracellular IFN-γ and IL-4 cytokines in CD4+ and CD8+ T..., Matralis [/bib_ref] [bib_ref] Regulatory T cells, Cytotoxic T lymphocytes and a Th1 cytokine profile in..., Cortese [/bib_ref] [bib_ref] Regulatory IgD hi B Cells Suppress T Cell Function via IL-10 and..., Schaut [/bib_ref] [bib_ref] Programmed Death 1-Mediated T Cell Exhaustion during Visceral Leishmaniasis Impairs Phagocyte Function, Esch [/bib_ref] [bib_ref] Leishmania infantum-specific T cell lines derived from asymptomatic dogs that lyse infected..., Pinelli [/bib_ref] [bib_ref] Recombinant cysteine proteinase from Leishmania (Leishmania) chagasi implicated in human and dog..., Da Costa Pinheiro [/bib_ref] [bib_ref] Early suppression of lymphoproliferative response in dogs with natural infection by Leishmania..., De Luna [/bib_ref] [bib_ref] Leishmania-specific lymphoproliferative responses and IgG1/IgG2 immunodetection patterns by Western blot in asymptomatic,..., Fernández-Pérez [/bib_ref] [bib_ref] Expression of cysteine proteinase type I and II of Leishmania infantum and..., Rafati [/bib_ref] [bib_ref] Humoral and cellular immune responses against Type I cysteine proteinase of Leishmania..., Nakhaee [/bib_ref] [bib_ref] Immunologic indicators of clinical progression during canine Leishmania infantum infection, Boggiatto [/bib_ref] [bib_ref] Leishmania-specific isotype levels and their relationship with specific cell-mediated immunity parameters in..., Rodríguez-Cortés [/bib_ref] [bib_ref] A long term experimental study of canine visceral leishmaniasis, Rodríguez-Cortés [/bib_ref] [bib_ref] Cellular and humoral immune responses in dogs experimentally and naturally infected with..., Pinelli [/bib_ref] [bib_ref] The immunology of canine leishmaniosis: Strong evidence for a developing disease spectrum..., Cabral [/bib_ref] [bib_ref] Immune response against Leishmania antigens in dogs naturally and experimentally infected with..., Rhalem [/bib_ref] [bib_ref] Cell mediated immunity and specific IgG1 and IgG2 antibody response in natural..., Leandro [/bib_ref] [bib_ref] Clinical, parasitologic, and immunologic evolution in dogs experimentally infected with sand fly-derived..., Travi [/bib_ref] [bib_ref] Semi-quantitative analysis of cytokine expression in asymptomatic canine leishmaniasis, Chamizo [/bib_ref] [bib_ref] Cytokine expression in dogs with natural Leishmania infantum infection, Panaro [/bib_ref] [bib_ref] Demonstration of Leishmania specific cell mediated and humoral immunity in asymptomatic dogs, Cabral [/bib_ref].
# Data availability statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.
## Conflicts of interest:
The authors declare no conflict of interest.
[fig] Figure 1: Flow diagram showing the steps followed to select the articles included in this review. [/fig]
[fig] ↑: CD5 + , CD4 + and CD8 + cells in AD versus SD (**). [/fig]
[fig] ↓: CD5 + , CD4 + , CD8 + and CD21 + cells in AD-II and SD versus AD-I and CD (****). [/fig]
[fig] Figure 2: Graphical representation of the percentage of dogs who exhibited positive lymphoproliferative response to SLA. In this figure, all studies that specified the number of dogs who gave a [/fig]
[fig] Author: Contributions: Conceptualization, M.C.L.; methodology, A.G.-C., A.E. and M.C.L.; investigation, A.G.-C. and A.E.; writing-original draft preparation, A.G.-C.; writing-review and editing, A.G.-C., A.E., M.C.T. and M.C.L.; supervision, A.E., M.C.T. and M.C.L.; funding acquisition, M.C.T. and M.C.L. All authors have read and agreed to the published version of the manuscript. [/fig]
[fig] Funding: This work was supported by the Network of Tropical Diseases Research-RICET (RD16/0027/0005) and grant PID2019-109090RB-I00 from the Programa Estatal I+D+i (MINECO). AGC was supported by the Introduction to Research Fellowships (Ref. JAEINT_20_02108). Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. [/fig]
[table] Table 1: Studies that analyzed the immune profiles of PBMCs from infected and control dogs. [/table]
|
Galectin-4 levels in hospitalized versus non-hospitalized subjects with obesity: the Malmö Preventive Project
Background: Obesity is strongly associated with the development of cardiovascular disease (CVD). However, the heterogenous nature of obesity in CVD-risk is still poorly understood. We aimed to explore novel CVD biomarkers and their possible association with presumed unhealthy obesity, defined as hospitalized subjects with obesity (HO).Methods: Ninety-two proteins associated with CVD were analyzed in 517 (mean age 67 ± 6 years; 33.7% women) individuals with obesity (BMI ≥30 kg/m 2 ) from the Malmö Preventive Project cohort, using a proximity extension array technique from the Olink CVD III panel. Individuals with at least one recorded hospitalization for somatic disease prior to study baseline were defined as HO phenotypes. Associations between proteins and HO (n = 407) versus non-hospitalized subjects with obesity (NHO, n = 110), were analyzed using multivariable binary logistic regression, adjusted for traditional risk factors.Results:Of 92 analyzed unadjusted associations between biomarkers and HO, increased levels of two proteins were significant at a false discovery rate < 0.05: Galectin-4 (Gal-4) and insulin-like growth factor-binding protein 1 (IGFBP-1). When these two proteins were included in logistic regression analyses adjusted for age and sex, Gal-4 remained significant. Gal-4 was independently associated with the HO phenotype in multivariable logistic regression analysis (OR 1.72; CI95% 1.16-2.54). Post-hoc analysis revealed that this association was only present in the subpopulation with diabetes (OR 2.26; CI95% 1.25-4.07). However, an interaction analysis was performed, showing no significant interaction between Gal-4 and prevalent diabetes (p = 0.16).Conclusions:In middle-aged and older individuals with obesity, increased Gal-4 levels were associated with a higher probability of HO. This association was only significant in subjects with diabetes only, further implying a role for Gal-4 in diabetes and its complications.
# Introduction
Obesity (body mass index, BMI ≥30 kg/m 2 ) contributes to health complications and reduces life expectancy with up to approximately 20 years [bib_ref] Years of life lost due to obesity, Fontaine [/bib_ref]. This is mainly due to the significantly increased risk of developing numerous noncommunicable diseases, such as type 2 diabetes (DM2), cardiovascular disease (CVD) and certain types of cancer Open Access Cardiovascular Diabetology *Correspondence: [email protected] [bib_ref] Body-mass index and mortality among 1.46 million white adults, Berrington De Gonzalez [/bib_ref] [bib_ref] Body-mass index and cause-specific mortality in 900 000 adults: collaborative analyses of..., Studies [/bib_ref]. Even more troublesome, the global prevalence of obesity has been steadily increasing since the 1970s, especially among adolescents and children, today reaching pandemic levels [bib_ref] Obesity: global epidemiology and pathogenesis, Bluher [/bib_ref]. Even though the link between obesity and increased CVD risk is not a matter of debate per se, there have long been speculations regarding how certain individuals with obesity possess a lower risk of developing CVD and diabetes type 2 (DM2), thus showing a heterogeneity of obesity as a risk factor [bib_ref] The degree of masculine differentiation of obesities: a factor determining predisposition to..., Vague [/bib_ref].
Furthermore, although the cardiometabolic complications of obesity are well established from an epidemiological perspective, the underlying pathophysiological mechanisms are not fully understood, particularly when taking into consideration the heterogeneity of obesity [bib_ref] Metabolically healthy obesity-heterogeneity in definitions and unconventional factors, Brandao [/bib_ref]. Recently, there have been considerable technological advances in the incorporation of multiomics into exploring alterations in specific cell types and identifying modifications in signaling events that promote disease development [bib_ref] Relevance of Multi-Omics Studies in Cardiovascular Diseases, Leon-Mimila [/bib_ref]. To better understand the mechanisms behind disease progression in obesity, we applied proximity extension assay (PEA) technology to measure 92 proteins (biomarkers) associated with inflammation and CVD [bib_ref] Homogenous 96-plex PEA immunoassay exhibiting high sensitivity, specificity, and excellent scalability, Assarsson [/bib_ref]. This represents an appealing approach to explore associations between multiple proteins and biological systems, which could in turn present possible diagnostic, prognostic, and therapeutic implications.
The aim of this cross-sectional, population-based study was to explore possible novel associations between CVD biomarkers and a phenotype of unhealthy obesity, namely obese subjects with a history of hospitalization for somatic disease up until late mid-life, [bib_ref] Metabolically healthy obesity (MHO) in the Malmo diet cancer study -epidemiology and..., Korduner [/bib_ref] [bib_ref] Proteomic and metabolomic characterization of metabolically healthy obesity: a descriptive study from..., Korduner [/bib_ref] [bib_ref] Antibodies against phosphorylcholine in hospitalized versus non-hospitalized obese subjects, Jujic [/bib_ref] using a multiplex proteomic platform consisting of 92 proteins linked to cardiovascular disease.
# Methods
## Study population
In the 1970s, the Malmö Preventive Project (MPP) cohort was established at the University Hospital, Malmö, Sweden, for the purpose of investigating cardiovascular risk factors in the general population [bib_ref] Orthostatic hypotension predicts all-cause mortality and coronary events in middle-aged individuals (The..., Fedorowski [/bib_ref]. A total of 33,346 individuals were included at baseline (71% attendance rate, 2/3 men), and survivors of the original cohort were re-examined between 2002 and 2006 (n = 18,240) in the MPP Re-Examination cohort (MPP-RES, attendance rate 72%) [bib_ref] Proteomic exploration of common pathophysiological pathways in diabetes and cardiovascular disease, Molvin [/bib_ref]. Furthermore, from this MPP-RES cohort, a sub-sample of 1,792 participants was selected to undergo echocardiography and electrocardiogram (ECG) recordings. These individuals were randomly chosen from groups based on their glucometabolic status. Oversampling was performed within the groups with glucometabolic disturbances (impaired fasting glucose, IFG (≥ 6.1 mmol/L or a single measurement of 7.0-11.0 mmol/l of fasting plasma glucose (FPG); new onset diabetes; and prevalent diabetes) to ensure numerical balance, as described previously, [bib_ref] Myocardial structure and function by echocardiography in relation to glucometabolic status in..., Leosdottir [/bib_ref] resulting in approximately 1/3 normoglycemic subjects, 1/3 with IFG, and 1/3 with diabetes). Prevalent diabetes was defined as either newonset diabetes (defined by two separate measurements of FPG ≥ 7.0 mmol/l or one measurement ≥ 11.1 mmol/l) or previously known diabetes (obtained through participant self-reporting and/or reporting of current anti-diabetic medication) [bib_ref] Myocardial structure and function by echocardiography in relation to glucometabolic status in..., Leosdottir [/bib_ref].
From the MPP-RES echocardiography sub-cohort, a total of 517 individuals with obesity and complete biomarker data were included in the present study. This subsample was further sub-divided into two different categories based on hospitalization history. Individuals with obesity with at least one recorded history of hospitalization prior to study baseline (n = 407) were defined as hospitalized subjects with obesity (HO). Correspondingly, individuals who had no history of hospitalization for somatic disease up until inclusion at MPP-RES baseline (n = 110) in late mid-age were defined as non-hospitalized subjects with obesity (NHO), [fig_ref] Figure 1: Flow-chart of the MPP-RES cohort stratified for individuals with and without obesity,... [/fig_ref]. Data on prior hospitalization was obtained through the Swedish National Hospital Inpatient Register. Normal deliveries were considered non-hospitalization; otherwise, all diagnoses were included. A detailed list of included/excluded diagnoses can be found in Additional file 1: [fig_ref] Table 1: Characteristics of the study population AHT antihypertensive, BMI body mass index, DBP... [/fig_ref].
As described in previous publications, [bib_ref] Proteomic exploration of common pathophysiological pathways in diabetes and cardiovascular disease, Molvin [/bib_ref] [bib_ref] Myocardial structure and function by echocardiography in relation to glucometabolic status in..., Leosdottir [/bib_ref] data on medical history and lifestyle (including physical activity, alcohol consumption, dietary habits, and smoking status) were acquired through a self-administered questionnaire. Weight (kg) and height (m) were measured in light indoor clothing, and BMI (kg/m 2 ) was subsequently calculated. Blood pressure (mmHg) was measured twice using a validated sphygmomanometer with a mercury manometer in the supine position by trained nurses after 10-minutes of rest-the mean values were then recorded. No intra-and/or inter-observed variability calculations were performed; however, the sphygmomanometer used was validated and continuously calibrated according to research standards at Malmö University hospital. Blood samples were acquired after an overnight fast and stored at − 80 °C [bib_ref] Long-term outcome of the Malmo preventive project: mortality and cardiovascular morbidity, Berglund [/bib_ref].
## Proteomic profiling
Plasma samples were analyzed by the Proximity Extension Assay (PEA) technique, using the Proseek Multiplex CVD III 96 × 96 reagents kit (Olink Bioscience, Uppsala, Sweden). The technique uses two antibodies that bind pairwise to each specific protein, creating a polymerase chain reaction sequence which then can be detected and quantified. The CVD III panel consists of 92 markers with established or proposed involvement in metabolism, inflammation, or cardiovascular disease (Additional file 2: [fig_ref] Table 2: Logistic regression models displaying associations of Galectin-4 levels and probability of being... [/fig_ref]. One protein was below the limit of detection in > 15% samples (N-terminal pro-B-type natriuretic peptide, NT-proBNP) and thus excluded; instead, NT-proBNP measurement with an electrochemiluminescence immunoassay was used. The mean intra-and inter-assay variations were 8.1% and 11.4%, respectively. Further information on the assays is available on the Olink homepage (www. olink. com).
## Laboratory analyses
Fasting serum total cholesterol, serum triglycerides, serum high-density lipoprotein and FPG were analyzed using Beckman Coulter LX20 (Beckman Coulter Inc., Brea, USA). Serum low-density lipoprotein concentration (LDL-C) was calculated through Friedewald's formula [bib_ref] Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use..., Friedewald [/bib_ref]. NT-proBNP was measured with an electrochemiluminescence immunoassay (Elecsys; Roche Diagnostics, Basel, Switzerland) at the Department of Clinical Chemistry, Akershus University Hospital, Lorenskog, Norway.
# Statistical analysis
Continuous variables are presented as means (± standard deviation, SD) or medians (25th-75th percentiles). A stratified random sample was created for identification of eligible study subjects. HO and NHO subjects were compared using one-way ANOVA test for normally distributed continuous variables, Mann-Whitney U-test for continuous variables with non-normal distribution, and χ2 test for binary variables. Prior to analysis, skewed variables (FPG) were log-transformed. Unadjusted binary logistic regression models exploring associations between each of the 92 proteins and HO were carried out applying the Benjamini-Hochberg multiple testing correction [bib_ref] Selective inference in complex research, Benjamini [/bib_ref] (false discovery rate, FDR, < 0.05). Significant associations were carried forward to analyses according to Model 1 (age-and sex-adjusted), and further adjusted according to Model 2 (total cholesterol, current smoking, hypertension, BMI, prevalent diabetes of any type, and log(FPG)). Hypertension was defined as a measured systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg and/or currently on antihypertensive medication. Finally, for associations significant in Model 2, a post-hoc analysis was carried out in subjects with and without diabetes using the remaining variables in Model 2. Lastly, to test for linearity between remaining variables with significant associations in Model 2 and independent variables, quartile analyses were carried out.
All analyses were carried out using SPSS 25.0 (IBM, Chicago, IL, USA). A nominal two-sided p-value of less than 0.05 was considered statistically significant.
# Results
## Study characteristics
Characteristics of the study population are presented in [fig_ref] Table 1: Characteristics of the study population AHT antihypertensive, BMI body mass index, DBP... [/fig_ref]. HO individuals were older than NHO. Furthermore, lower levels of total cholesterol and LDL-C, as well as lower systolic and diastolic blood pressures were seen in HO when compared with NHO. However, the use of both lipid-and blood pressure lowering drugs was significantly higher in the HO group. No difference between the two groups were seen in FPG levels, prevalent diabetes, BMI, or waist circumference.
## Biomarker analyses
Of 92 analyzed unadjusted associations between biomarkers and HO, increased levels of two proteins were significant at an FDR < 0.05: Galectin-4 (Gal-4) and insulin-like growth factor-binding protein 1 (IGFBP-1) (Additional file 3: [fig_ref] Table 3: Post-hoc analysis comparing levels of Gal-4 in obese subjects with or without... [/fig_ref]. When these two proteins were included in logistic regression analyses adjusted for age and sex, Gal-4 remained significant (OR 1.76; CI 95% 1.23-2.51; p = 0.002) whereas IGFBP-1 did not (OR 1.24; CI95% 0.97-1.58; p = 0.087). Each 1 SD increase in Galectin-4 (Gal-4) levels was associated with a higher probability of being HO in the fully adjusted logistic regression model (OR 1.72; CI95% 1.16-2.54; p = 0.007) [fig_ref] Table 2: Logistic regression models displaying associations of Galectin-4 levels and probability of being... [/fig_ref]. When further excluding external trauma (n = 38) as a determinant of being HO, the positive association for Gal-4 remained significant (p = 0.024). An interaction analysis was performed, showing no significant interaction between Gal-4 and prevalent diabetes (p = 0.16). However, given the known correlation between these two variables, [bib_ref] Using a targeted proteomics chip to explore pathophysiological pathways for incident diabetes-the..., Molvin [/bib_ref] [bib_ref] In search of causal pathways in diabetes: a study using proteomics and..., Beijer [/bib_ref] a post-hoc stratified analysis was carried out and revealed that the association between Gal-4 and HO was only present among patients with diabetes [fig_ref] Table 3: Post-hoc analysis comparing levels of Gal-4 in obese subjects with or without... [/fig_ref]. To elucidate if the association between Gal-4 and the probability of being HO was linear, we carried out additional quartile analyses. In Model 2, p for trend was 0.009, and further analyses revealed that the risk of being HO was found to be strongest in the upper quartile (Additional file 4: . Finally, we explored how diabetes prevalence and glucose levels differed across quartiles of Gal-4 levels. The highest proportion of subjects with diabetes was found in the upper quartile (Q4) of Gal-4 (65.9%), compared to 27.9% in the lowest quartile of Gal-4 (p for difference between groups = 9.6 × 10 − 9 ). Similarly, glucose levels were higher in the upper quartile (Q4) of Gal-4 (p for difference between Q1 and Q4 = 6.1 × 10 − 7 ) as compared with Q1. .
# Discussion
By using a newly adopted definition of metabolic health in obesity, based on history of hospitalization for somatic disorders up until late mid-life, [bib_ref] Metabolically healthy obesity (MHO) in the Malmo diet cancer study -epidemiology and..., Korduner [/bib_ref] [bib_ref] Proteomic and metabolomic characterization of metabolically healthy obesity: a descriptive study from..., Korduner [/bib_ref] [bib_ref] Antibodies against phosphorylcholine in hospitalized versus non-hospitalized obese subjects, Jujic [/bib_ref] [bib_ref] Metabolically healthy obesity (MHO)-new research directions for personalised medicine in cardiovascular prevention, Nilsson [/bib_ref] we found that increased levels of Gal-4 were independently associated with a higher probability of having been hospitalized in a cohort of middle-aged and older obese subjects. Descriptive data at baseline examination did not reveal any differences in neither BMI nor waist circumference between HO and NHO, suggesting a similar fat distribution. However, plasma total cholesterol, LDL-C and blood pressure were significantly lower among HO, likely because of a higher prevalence of medical treatment with both antihypertensive and lipid-lowering drugs. Finally, the positive association between Gal-4 and the HO phenotype was significant only in subjects with diabetes.
We have previously carried out cross-sectional studies in the Malmö Diet and Cancer Study cohort, where NHO was defined by using a novel approach of a history of non-hospitalization for somatic disorders up until midlife [bib_ref] Metabolically healthy obesity (MHO) in the Malmo diet cancer study -epidemiology and..., Korduner [/bib_ref] [bib_ref] Proteomic and metabolomic characterization of metabolically healthy obesity: a descriptive study from..., Korduner [/bib_ref] [bib_ref] Antibodies against phosphorylcholine in hospitalized versus non-hospitalized obese subjects, Jujic [/bib_ref]. In those studies we found that NHO had a decreased risk of both total mortality and incident CVD compared with HO during a 20-year follow-up period. When comparing NHO with non-obese controls, there were no significant differences in terms of mortality or CVD risk [bib_ref] Metabolically healthy obesity (MHO) in the Malmo diet cancer study -epidemiology and..., Korduner [/bib_ref]. Potential protective factors included a more favorable lipid and glucose profile, downregulation of potentially harmful proteomic biomarkers and a less sedentary lifestyle [bib_ref] Proteomic and metabolomic characterization of metabolically healthy obesity: a descriptive study from..., Korduner [/bib_ref]. Moreover, lower plasma levels of antibodies against anti-phosphorylcholine, which possess anti-inflammatory properties and is coupled with lower CVD risk, were associated with a higher risk of being HO [bib_ref] Antibodies against phosphorylcholine in hospitalized versus non-hospitalized obese subjects, Jujic [/bib_ref]. This is in line with previous research focusing on obesity phenotypes with different cardiometabolic disease risk but with a different terminology, namely metabolically healthy obesity (MHO) [bib_ref] Metabolically healthy obesity (MHO)-new research directions for personalised medicine in cardiovascular prevention, Nilsson [/bib_ref] [bib_ref] Metabolically healthy obesity and cardiovascular events: a systematic review and metaanalysis, Eckel [/bib_ref].
## Metabolically healthy obesity (mho)
The evolving concept of MHO describes obese individuals that through proposed protective mechanisms, such as peripheral body fat distribution, lower grade of chronic inflammation and higher insulin sensitivity, seem to escape metabolic or cardiovascular complications [bib_ref] Metabolically healthy obesity (MHO)-new research directions for personalised medicine in cardiovascular prevention, Nilsson [/bib_ref] [bib_ref] Metabolically healthy obesity and cardiovascular events: a systematic review and metaanalysis, Eckel [/bib_ref] [bib_ref] Heterogeneity in the prevalence of risk factors for cardiovascular disease and type..., Mclaughlin [/bib_ref]. This description could be considered controversial, since increasing evidence suggests that MHO is not a steady state and can transform into metabolically unhealthy obesity over time. Moreover, when compared with metabolically healthy individuals with normal weight, there is a significantly increased risk for incident CVD and metabolic complications linked to MHO [bib_ref] Transition from metabolic healthy to unhealthy phenotypes and association with cardiovascular disease..., Eckel [/bib_ref] [bib_ref] Are people with metabolically healthy obesity really healthy? A prospective cohort study..., Zhou [/bib_ref] [bib_ref] Metabolically healthy obese and incident cardiovascular disease events among 3.5 million men..., Caleyachetty [/bib_ref] [bib_ref] The long-term prognosis of cardiovascular disease and all-cause mortality for metabolically healthy..., Zheng [/bib_ref]. One major concern about the conflicting results lies in the definition of MHO which differs substantially between different studies, but mainly focuses on the absence of risk variables included in the metabolic syndrome [bib_ref] Metabolically healthy obesity: facts and fantasies, Smith [/bib_ref]. There is now an ongoing debate as to whether the term MHO should be avoided and instead be treated as a conceptual model to study mechanisms linking obesity to risk for or protection from cardiometabolic complications [bib_ref] Metabolically healthy obesity, Bluher [/bib_ref].
## Galectin-4
Being part of the galectin family (consisting of 15 small leptin peptides), Gal-4 is expressed almost exclusively in the gastrointestinal tract of healthy individuals, where it plays a role in controlling intestinal inflammation. It reduces proinflammatory cytokine production in the intestinal mucosa, and knockdown of the Gal-4 peptide promotes colorectal cancerogenesis. This suggests that Gal-4 plays a significant role in the pathophysiology of the development of both inflammatory bowel disease and colorectal cancers [bib_ref] The role of galectins as modulators of metabolism and inflammation, Brinchmann [/bib_ref]. However, the physiological role of Gal-4 is multifaceted and further include apical protein trafficking, lipid raft stabilization, intestinal wound healing and bacterial pathogen fighting [bib_ref] Galectin-4 N-terminal domain: binding preferences toward A and B antigens with different..., Quintana [/bib_ref]. Epidemiological data also strongly propose an involvement of Gal-4 in cardiometabolic diseases, suggesting it may be considered as a predictive biomarker for the development of CVD and diabetes [bib_ref] Using a targeted proteomics chip to explore pathophysiological pathways for incident diabetes-the..., Molvin [/bib_ref]. Still, the causal pathway is poorly understood [bib_ref] Proteomic exploration of common pathophysiological pathways in diabetes and cardiovascular disease, Molvin [/bib_ref] [bib_ref] In search of causal pathways in diabetes: a study using proteomics and..., Beijer [/bib_ref]. One theory might lie at the cellular level, where Gal-4 is part of the apical protein transport from the Golgi-apparatus to the apical cell membrane of the enterocyte, including the well-known protease dipeptidyl peptidase-4 (DPP-4) [bib_ref] Galectin-4 and sulfatides in apical membrane trafficking in enterocyte-like cells, Delacour [/bib_ref]. In mice, DPP-4 seems to be misguided and accumulates intracellularly when Gal-4 is depleted [bib_ref] Galectin-4 and sulfatides in apical membrane trafficking in enterocyte-like cells, Delacour [/bib_ref]. DPP-4 plays a major role in promoting cardiometabolic disease by cleaving and thus inactivating glucose-dependent insulinotropic polypeptide and glucagon-like peptide 1 (GLP-1), i.e., two of our most common incretins [bib_ref] DPP4 in cardiometabolic disease: recent insights from the laboratory and clinical trials..., Zhong [/bib_ref]. Modern anti-diabetic drugs such as DPP4-inhibitors and GLP-1 agonists are incretin-based and part of the standard treatment of DM2 as secondline drugs in most patients [bib_ref] The incretin system: glucagon-like peptide-1 receptor agonists and dipeptidyl peptidase-4 inhibitors in..., Drucker [/bib_ref]. Incretins are involved in appetite control and delaying gastric emptying actions that are dependent on GLP-1 receptor activation within the central nervous system, thus having the potential to regulate body weight [bib_ref] The incretin system ABCs in obesity and diabetes -novel therapeutic strategies for..., Joao [/bib_ref]. Furthermore, another study of women with gestational diabetes found an overexpression of Gal-4 in the placental syncytiotrophoblast cells, compared to healthy controls [bib_ref] Overexpression of galectin-4 in placentas of women with gestational diabetes, Schrader [/bib_ref]. Thus, one proposed explanation for our main finding may be Gal-4's involvement in the development of diabetes, which also has been suggested in a previous publication with a similar approach of proteomic exploration. To elucidate this, we carried out a post-hoc analysis, suggesting that elevation in Gal-4 was associated with higher probability of being HO only in those with prevalent diabetes. Gal-4 has a potential inflammatory role in the intestinal mucosa. Previous studies have linked obesity and diabetes to altered composition of the gut microbiota [bib_ref] Gut microbiota, obesity and diabetes, Patterson [/bib_ref] [bib_ref] Connection between BMI-related plasma metabolite profile and gut microbiota, Ottosson [/bib_ref]. Changes in gut microbiota, i.e., through an unhealthy diet, lead to damage of the intestinal barrier, promote leakage and thus endotoxemia through higher levels of lipopolysaccharides systemically, which in turn stimulates the development of low-grade systemic inflammation associated with the negative impact of both obesity and metabolic disorders [bib_ref] Gut microbiota, obesity and diabetes, Patterson [/bib_ref]. Therefore, Gal-4 might, at least in theory, aggravate the pathological processes induced by the obese-diabetic microbiota.
## Study strengths and limitations
By using a definition of individuals with obesity with a more favorable metabolic health as not having been hospitalized for somatic disease up until late midlife, we were able get an objectively defined and more stable phenotype which could serve as an alternative to the conventional way of defining metabolic health within the population with obesity, commonly called MHO. Previous definitions focus on the absence of criteria for the metabolic syndrome, which could shift intra-individually during repeated measurements at different occasions. Moreover, by renaming metabolic health in obesity as non-hospitalized versus hospitalized individuals with obesity instead of MHO, we avoid the perception of certain phenotypes of obesity labeled as healthy.
There are limitations to this study. Its cross-sectional nature precludes any conclusions about causality. However, the study subjects come from a well-characterized, retrospective cohort with excellent national, and wellvalidated, register data on hospitalization, which is why it was possible to apply our approach to define NHO and HO. This study only covers individual data collected at one regional center. A multicenter study to replicate the findings would be preferable, but to reduce false positive findings, the use of FDR analysis was carried out. Furthermore, because our subjects were of European descent, these findings might not be generalizable to other populations. Similarly, the population selection based on glucometabolic disturbances could raise concerns of how well this cohort represents the general population. However, when compared with similar cohorts, the incidence rate of diabetes was proportionate [bib_ref] Cadmium exposure and incidence of diabetes mellitus-results from the Malmo Diet and..., Borne [/bib_ref] [bib_ref] High incidence of diabetes mellitus among a middle-aged population in Iran: a..., Ebrahimi [/bib_ref]. The Olink CVD III panel is partially restricted to proteins associated with CVD and inflammation, and an extended analysis including biomarkers related to diabetes and/or metabolism would most likely add information about the pathophysiology in HO. Lastly, another limitation of this study was that subjects with a non-hospitalization status prior to baseline could still suffer from cardiometabolic disturbances, since no pre-defined diagnoses of hospitalization were decided upon, and many individuals could be treated for chronic illnesses within a primary health care unit. On the other hand, these conditions could have been milder or counterbalanced by protective mechanisms in the affected subjects leading to a status of nonhospitalization in our analyses.
# Conclusions
In obese subjects during late mid-life, increased Galectin-4 levels were associated with a higher probability of being an individual with a history of HO. This association was only significant in subjects with diabetes, implying a role for Galectin-4 in diabetes and its complications.
[fig] Figure 1: Flow-chart of the MPP-RES cohort stratified for individuals with and without obesity, as well as history of hospitalization for somatic disorders in subjects with obesity, respectively [/fig]
[table] Table 1: Characteristics of the study population AHT antihypertensive, BMI body mass index, DBP diastolic blood pressure, HDL-C high density lipoprotein concentration, HO hospitalized subjects with obesity, LDL-C low density lipoprotein concentration, NHO non hospitalized subjects with obesity. Bold values denote statistical significance at the p<0.05 [/table]
[table] Table 2: Logistic regression models displaying associations of Galectin-4 levels and probability of being HO Values are odds ratios (OR) and 95% confidence intervals. Bold values denote statistical significance at the p<0.05 BMI body mass index, FPG fasting plasma glucose, HO hospitalized subjects with obesity HO (n = 407) vs. NHO (n = 110) [/table]
[table] Table 3: Post-hoc analysis comparing levels of Gal-4 in obese subjects with or without prevalent diabetes Values are odds ratios (OR) and 95% confidence intervals. Bold values denote statistical significance at the p<0.05 BMI body mass index, FPG fasting plasma glucose, HO hospitalized subjects with obesity, NHO non hospitalized subjects with obesity [/table]
|
Medulloblastoma group 3 and 4 tumors comprise a clinically and biologically significant expression continuum reflecting human cerebellar development
## In brief
The childhood brain tumor medulloblastoma is classified into multiple DNA methylation-based subtypes. Using RNA-seq, Williamson et al. show that group 3 and group 4 tumors manifest as intermediates on a transcriptomic continuum. Position on the continuum is associated with molecular pathology and disease course. The continuum mirrors early cerebellar development, implicating cells of origin.
# Introduction
The division of medulloblastoma (MB) into molecular subgroups has defined the past decade of MB research, making it all but impossible to interpret future findings except through the prism of these fundamental biological subdivisions. MB was first divided into subgroups on the basis of profiling by expression array [bib_ref] Integrative genomic analysis of medulloblastoma identifies a molecular subgroup that drives poor..., Cho [/bib_ref] [bib_ref] Beta-catenin status in paediatric medulloblastomas: correlation of immunohistochemical expression with mutational status,..., Fattet [/bib_ref] [bib_ref] Medulloblastoma comprises four distinct molecular variants, Northcott [/bib_ref] [bib_ref] Genomics identifies medulloblastoma subgroups that are enriched for specific genetic alterations, Thompson [/bib_ref] and, subsequently, DNA methylation array [bib_ref] DNA methylation profiling of medulloblastoma allows robust subclassification and improved outcome prediction..., Schwalbe [/bib_ref]. The current consensus is that there exist four major MB subgroups (MB SHH , MB WNT , MB Grp3 , MB Grp4 ), each with unique clinico-biological characteristics ; MB WNT and MB SHH are named after characteristic disruptions in the WNT (CTNNB1 mutation [bib_ref] Wnt/Wingless pathway activation and chromosome 6 loss characterize a distinct molecular sub-group..., Clifford [/bib_ref] [bib_ref] Beta-catenin status predicts a favorable outcome in childhood medulloblastoma: the United Kingdom..., Ellison [/bib_ref] and SHH (PTCH, SUFU, SMO mutation, or GLI2 amplification ) pathways, respectively. MB WNT denotes an almost entirely curable disease [bib_ref] Beta-catenin status predicts a favorable outcome in childhood medulloblastoma: the United Kingdom..., Ellison [/bib_ref] , and MB SHH occur more frequently in infants . The remaining two subgroups, group 3 (MB Grp3 ) and group 4 (MB Grp4 ), do not exhibit subgroup-defining mutations [bib_ref] The whole-genome landscape of medulloblastoma subtypes, Northcott [/bib_ref] but nonetheless possess distinct clinico-biological characteristics; MB Grp3 patients have a greater incidence of ''high-risk'' features such as LCA (large-cell/anaplastic) histology and MYC amplification [bib_ref] Medulloblastomics: the end of the beginning, Northcott [/bib_ref] [bib_ref] MYC family amplification and clinical risk-factors interact to predict an extremely poor..., Ryan [/bib_ref] [bib_ref] Molecular subgroups of medulloblastoma: the current consensus, Taylor [/bib_ref]. MB Grp4 tumors more frequently demonstrate isochromosome 17q (i17q) . Some overlap in mutational spectrum, DNA methylation, and expression characteristics between MB Grp3 and MB Grp4 has often been noted, and these are considered more closely related molecularly to one another than to MB SHH and MB WNT , leading them to be considered as a non-WNT/non-SHH group in the latest World Health Organization (WHO) classification [bib_ref] The 2021 WHO classification of tumors of the central nervous system: a..., Louis [/bib_ref]. The advent of routine MB molecular subgrouping has enabled the current generation of molecularly driven trials (e.g., NCT02066220, NCT01878617, NCT02724579, NCT01125800) [bib_ref] Phase I and phase II sonidegib and vismodegib clinical trials for the..., Li [/bib_ref] [bib_ref] Vismodegib exerts targeted efficacy against recurrent sonic hedgehog-subgroup medulloblastoma: results from phase..., Robinson [/bib_ref] , which exploit MB WNT /MB SHH biology to stratify treatments or direct biological therapeutics.
Further elaborations of the consensus subgroups were published, based primarily upon methylomic definitions [bib_ref] Intertumoral heterogeneity within medulloblastoma subgroups, Cavalli [/bib_ref] [bib_ref] The whole-genome landscape of medulloblastoma subtypes, Northcott [/bib_ref] [bib_ref] Molecular subgroups for clinical classification and outcome prediction in childhood medulloblastoma: a..., Schwalbe [/bib_ref]. These were followed by a second consensus study that defined 8 subtypes within MB Grp3 /MB Grp4 , I-VIII, a number of which comprised a mix of MB Grp3 and MB Grp4 tumors . Furthermore, MB SHH can be further divided into subtypes broadly associated with age at diagnosis [bib_ref] Molecular subgroups for clinical classification and outcome prediction in childhood medulloblastoma: a..., Schwalbe [/bib_ref]. The fact that certain MB Grp3 /MB Grp4 subtypes (e.g., I, V) overlap between MB Grp3 and MB Grp4 further supports a relationship between the two subtypes. A recent study of MB used single-cell RNA sequencing (scRNA-seq) analysis of 4,873 individual cells from 17 MB Grp3 /MB Grp4 patients to define two transcriptional meta-programs representing a continuum of neuronal cellular differentiation states. This was mirrored in the transcriptional differences between bulk MB Grp3 and MB Grp4 patients and concluded that MB Grp3/Grp4 contain cells along a common continuum of neuronal differentiation, providing further rationale to support this relationship.
Based on murine modeling, expression, and imaging studies [bib_ref] Subtypes of medulloblastoma have distinct developmental origins, Gibson [/bib_ref] , MB WNT and MB SHH are believed to derive from two spatially distinct developmental origins in the early hindbrain, lower rhombic lip (RL)/dorsal brainstem, and upper RL/early cerebellum, respectively. The developmental origins of MB Grp3 and MB Grp4 were investigated in a study mapping subgroup-specific super-enhancer elements, suggesting deep cerebellar nuclei residing in the nuclear transitory zone as the cell of origin for MB [bib_ref] Active medulloblastoma enhancers reveal subgroup-specific cellular origins, Lin [/bib_ref]. More recently, two studies that compared bulk and single-cell transcriptomic (scRNA-seq) MB profiles with developing murine cerebellar scRNA-seq reference datasets described MB Grp3 and MB Grp4 as most closely resembling Nestin + stem cells and unipolar brush cells (UBCs), respectively, highlighting putative cells of origin . It is notable that the conclusions of each of these studies rely principally upon cross-species comparisons with murine as opposed to human developmental references. Human RL development is more complex and prolonged than that of the mouse, possessing unique features not shared with any other vertebrates [bib_ref] Spatiotemporal expansion of primary progenitor zones in the developing human cerebellum, Haldipur [/bib_ref].
Here, we characterize the transcriptomic landscape of 331 primary MB, with clinico-pathological annotation, DNA methylation, and copy-number profiles, and we catalog subgroup-specific isoforms and RNA-editing events. We show that, despite the discrete methylomic subdivisions of the MB Grp3 /MB Grp4 methylation subtypes I-VIII, these tumors manifest transcriptionally on a bipolar continuum between MB Grp3 and MB Grp4 archetypes. Moreover, the position of an individual tumor on this continuum is predictive of methylation subtype, prognosis, specific copynumber and mutational alterations, and activation of key molecular pathways and regulatory events. By using human scRNA-seq fetal cerebellar reference data, we show that this continuum mirrors and recapitulates the major developmental trajectories within early human cerebellar development, allowing us to map the interplay between key oncogenic events and putative cells of origin for each MB subtype.
# Results
MB shows a continuum of expression between MB Grp3 and MB Grp4 RNA-seq ($90 million paired-end reads) was performed on 331 snap-frozen primary samples from patients with a diagnosis of MB . Transformed gene-level read counts were subject to consensus non-negative matrix factorization (NMF) clustering with resampling to determine the most stable number of clusters and metagenes (i.e., major biological effects described by multiple genes and summarized as a single score). As expected, a four-metagene/four-cluster solution was optimally stable, reflecting the four major consensus subgroups as currently understood [fig_ref] Figure 1: Group 3/group 4 medulloblastoma [/fig_ref]. Approximately 3% (10/331) of samples were defined as non-classifiable (i.e., low probability of classification). Approximately 4% (13/331) samples could only be classified as indeterminate MB Grp3 /MB Grp4 (i.e., confidently classifiable as either MB Grp3 or MB Grp4 but not specific as to which). The distribution of clinico-biological features was consistent with previously described features of the consensus MB subgroups (Figures 1A and S1A); for instance, chromosome 6 loss in 83% (24/29) of MB WNT .
The two metagenes that described MB Grp3 and MB Grp4 samples were notably gradated and overlapping in an anticorrelative manner [fig_ref] Figure 1: Group 3/group 4 medulloblastoma [/fig_ref] , implying that, contrary to some previous descriptions using expression microarrays [bib_ref] Intertumoral heterogeneity within medulloblastoma subgroups, Cavalli [/bib_ref] , MB Grp3 and MB Grp4 are not distinct transcriptional entities but rather exist as a continuum between two transcriptional polarities that we refer to here as G3 and G4. To describe this continuum, we created a continuous score (G3/G4 score) scaled between 0 and 1 to reflect the proportionate amount of G3/G4 metagene expression in each MB Grp3 /MB Grp4 (i.e., all non-WNT/non-SHH tumors) whereby a score of ''0'' indicates a 100% G4 tumor and ''1'' indicates a 100% G3 [fig_ref] Figure 1: Group 3/group 4 medulloblastoma [/fig_ref]. This was applied to the 223 samples classified as MB Grp3 , MB Grp4 ,
[formula] or intermediate MB Grp3 /MB Grp4 . [/formula]
We regard these results as showing that no individuals fall into discrete transcriptional subtypes with respect to the G3/G4 continuum, but for convenient comparison, we subdivided the expression continuum (G3/G4 score) into five purely notional quantiles: highG4 (0-0.2, n = 69/223 [31%]), lowG4 (0.2-0.4, n = 60/223 [27%]), , lowG3 (0.6-0.8, n = 22/223 [10%]), and highG3 (0.8-1 G3/G4 score, n = 33/223 [15%]). All of the samples with >0.5 G3/G4 score were classified as MB Grp3. Notably, 15/20 (75%) MB Grp3 /MB Grp4 samples, which showed disagreement in classification between RNA-seq and DNA methylation array, were classified as indeterminate MB Grp3 /MB Grp4 by RNA-seq [fig_ref] Figure 1: Group 3/group 4 medulloblastoma [/fig_ref]. Examining the MB Grp3 /MB Grp4 subtype (I-VIII) calls by t-distributed stochastic neighbor embedding (t-SNE) [fig_ref] Figure 1: Group 3/group 4 medulloblastoma [/fig_ref] shows clustering by subtype, suggesting that each methylation subtype imparts distinct secondary expression characteristics beyond the primary G3/G4 continuum metagene. Regardless, the MB Grp3 /MB Grp4 subtypes may be broadly ordered upon the G3/G4 continuum in partially overlapping domains from most group 4-like to most group 3-like (VIII, VI, VII, V, I, III, IV, II, respectively) [fig_ref] Figure 1: Group 3/group 4 medulloblastoma [/fig_ref].
Specific clinico-biological features were significantly nonrandomly distributed across the G3/G4 continuum [fig_ref] Figure 2: Clinico-pathology, subtype, and survival are related to an individual's position on the... [/fig_ref]. For instance, the distribution of patients with LCA pathology along the continuum is significantly different from those without LCA pathology (D = 0.339, p = 0.046, n = 158); there appears to be more LCA patients toward the G3 end of the continuum. The distribution along the continuum of patients with certain large (arm level/chromosomal) copy-number alterations are significantly differently distributed compared to those without. Most notably, patients with i17q (D = 0.402, p < 0.001, n = 201) and chromosome 8 gain (D = 0.69, p < 0.001, n = 201) are more frequent toward the G4 and G3 poles, respectively [fig_ref] Figure 1: Group 3/group 4 medulloblastoma [/fig_ref]. Mutations are not frequent in MB Grp3 /MB Grp4 [bib_ref] The whole-genome landscape of medulloblastoma subtypes, Northcott [/bib_ref] ; however, non-synonymous mutations of ZMYM3 and KDM6A are significantly non-randomly distributed with respect to the continuum (each p < 0.01) [fig_ref] Figure 2: Clinico-pathology, subtype, and survival are related to an individual's position on the... [/fig_ref].
We examined the relationship between the G3/G4 score and prognosis. Again, we divided the G3/G4 score into notional quantiles for the purposes of visualization/description showing a progressively poorer 5-year overall survival (OS) across the continuum: Log rank (test for trend) Z = À2.97, p = 0.003, n = 191, highG3 = 46%, lowG3 = 57%, G3.5 = 71%, lowG4 = 81%, and highG4 = 76% [fig_ref] Figure 2: Clinico-pathology, subtype, and survival are related to an individual's position on the... [/fig_ref]. Most important, Cox regression indicates that a continuous G3/G4 score is highly significant (relative risk showing an increase in RR of death of 4.7 times greater for a patient with a G3/ G4 score of 0 compared to a score of 1.
To assess any independent prognostic significance, we used multivariable Cox regression analysis of progression-free survival, including highG3 status alongside other risk factors (MYC amplification, LCA histology, and metastatic disease). The analysis showed that the highG3 status-chosen over a continuous variable in this instance as it overlaps most with other risk factors-retains significance (RR = 2.4, p = 0.014, n = 135), indicating that the G3/G4 score possesses significant independent prognostic power that is distinct from its association with other ''high-risk'' disease features [fig_ref] Figure 2: Clinico-pathology, subtype, and survival are related to an individual's position on the... [/fig_ref].
A G3/G4 continuum score can be reverse-engineered from DNA methylation profiles to validate clinicopathological associations A series of sample cohorts of MB Grp3 /MB Grp4 with DNA methylation profiles have previously been published by our group and others [bib_ref] Intertumoral heterogeneity within medulloblastoma subgroups, Cavalli [/bib_ref] [bib_ref] The whole-genome landscape of medulloblastoma subtypes, Northcott [/bib_ref] [bib_ref] Molecular subgroups for clinical classification and outcome prediction in childhood medulloblastoma: a..., Schwalbe [/bib_ref] [bib_ref] Second-generation molecular subgrouping of medulloblastoma: an international meta-analysis of Group 3 and..., Sharma [/bib_ref]. To these we added 166 profiles to produce a large cohort (n = 1,670) better powered to validate and further expand the findings we made using transcriptomic datasets. We therefore explored the possibility of reverse-engineering a G3/G4 score from DNA methylation data. Using the same method as used for expression was impossible, given that the constrained range (i.e., 0-1 [fully unmethylated]-[fully methylated]) and bimodal distribution of CpG methylation does not lend itself straightforwardly to a continuous score . Unlike expression, which tends to follow a log-linear association with G3/G4 score, methylation follows a sigmoidal distribution from hypo-to hypermethylation or vice versa. The inflection point along the G3/G4 continuum at which these CpGs ''switch'' from one state to the other varies by CpG . We trained a classifier using a training cohort of MB Grp3 /MB Grp4 samples for which we possessed both RNA-seq and DNA methylation profiles (n = 192). Pre-selecting 400 cross-validated CpG features that distinguish between each of the G3/G4 categorical states, we used these to train a random forest classifier to accurately predict (root-mean-square error [RMSE] = 0.036) a G3/G4 score from DNA methylation data alone .
Using this larger MB Grp3 /MB Grp4 methylation cohort, we were able to demonstrate significant differences in distribution along the continuum for patients with infant status (<3 years), metastases, LCA, and MYC amplification (each progressively more frequent toward the G3 pole), and mutations of PRDM6, KDM6A, KMT2C, and ZMYM3 (progressively more frequent toward the G4 pole) compared to patients who lack those features (each p < 0.001; . Likewise, chromosomal gains of 1q, 5, 6, 8, and 16q (each p < 0.001) were progressively more frequent toward the G3 pole, and i17q (p < 0.001) was progressively more frequent toward the G4 pole . These findings thus validated our findings from the initial RNA-seq cohort.
The larger cohort size allowed us to also explore the relationship between the G3/G4 continuum and the MB Grp3 /MB Grp4 subtypes (I-VIII) as well as their previously reported clinico-pathological/mutational characteristics . The MB Grp3 /MB Grp4 subtypes as predicted from DNA methylation data once again occupy discrete but partly overlapping domains within the G3/G4 continuum, broadly ordered, as per the RNA-seq-only cohort, from most archetypally MB Grp4 to MB Grp3 -VIII, VI, VII, V, I, III, IV, II, respectively [fig_ref] Figure 2: Clinico-pathology, subtype, and survival are related to an individual's position on the... [/fig_ref]. We next asked whether the variation in the distribution of clinicopathological features and mutation previously described as being characteristic of MB Grp3 /MB Grp4 subtypes (I-VIII) were attributable to their position on the G3/G4 continuum, the MB Grp3 /MB Grp4 subtype (I-VIII), or, indeed, both. Certain frequent clinicopathological features and copynumber changes (e.g., metastatic disease, MYC amplification, LCA histology, i17q, loss of chromosome 8, gain of chromosome 5) are significantly non-randomly distributed with respect to G3/ G4 continuum, even within individual subtypes . For example, 100% (11/11) of subtype III with MYC amplifications are highG3 compared to 59% (69/117) without MYC amplification. The presence of i17q as the only major chromosomal alteration is a highly characteristic change in subtype VIII, but when considering only MB subtype, VIII is still significantly enriched at the highG4 end of the continuum (D = 0.162, p = 0.014).
The relationship between G3/G4 score and risk of death is significant and striking, allowing us to validate the findings of our RNA-seq cohort with greater confidence. Again, for the purposes of visualization/description, we divided patient G3/G4 scores into notional quantiles: Patients older than 3 years log rank (test for trend) Z = À4.89, p < 0.0001, n = 589, highG3 = 49%, lowG3 = 59%, G3.5 = 64%, lowG4 = 77%, and highG4 = 83% [fig_ref] Figure 2: Clinico-pathology, subtype, and survival are related to an individual's position on the... [/fig_ref]. A similar result is found in patients of all ages: Log rank (test for trend) Z = À5.49, p < 0.0001, n = 654 [fig_ref] Figure 5: Distribution of single cells along the group 3/group 4 continuum is limited... [/fig_ref]. Most important, G3/G4 score is efficiently modeled as a continuous variable using Cox proportional hazards. Again, patients older than 3 years shows a 33 increased risk of death from one end of the continuum to the other (RR = 3, n = 589, p < 0.001). We also note that MB Grp3 /MB Grp4 subtypes (I-VIII) are significantly associated with OS (n = 524, p < 0.001) [fig_ref] Figure 5: Distribution of single cells along the group 3/group 4 continuum is limited... [/fig_ref].
The G3/G4 continuum is associated with differential regulation of oncogenic/developmental pathways The expression of 590 genes is significantly correlated with the G3/G4 score in our RNA-seq cohort (p < 0.01, log 2 fold change >10, n = 223), increasing/decreasing log linearly across the continuum. Most notably, MYC expression correlates significantly with the G3/G4 score (rho = 0.73, p < 0.001, n = 223)-approximately 463 greater from the G4 end of the continuum to the G3 . Performing gene set enrichment analysis (GSEA), we observed that transcriptional targets of MYC were also significantly upregulated (NES = 3.37, p = 0.007) . Single-sample GSEA (ssGSEA) analysis (Hä nzelmann was used to represent the activa-tion/repression of pathways/signatures for each individual and found several oncogenic pathways that were progressively activated or repressed in a manner significantly correlated (each p < 0.001) with the G3/G4 continuum, including MYC, cell cycle, mammalian target of rapamycin (mTOR), transforming growth factor b (TGF-b) (activated at the G3 pole), and NOTCH (activated at the G4 pole) . In addition, a broad pattern of progressive neuronal differentiation at the G4 pole and photoreceptor (CRX/NRL) characteristics at the G3 pole of the G3/G4 continuum were observed.
We examined differentially methylated regions (DMRs) within previously identified MB Grp3 /MB Grp4 specific enhancer loci [bib_ref] Active medulloblastoma enhancers reveal subgroup-specific cellular origins, Lin [/bib_ref] , identifying 45 that also overlapped with gene promoters; each ''switched'' from hypomethylated to hypermethylated or vice versa at specific points along the G3/G4 continuum. The expression of 33/45 of these genes is significantly correlated with the G3/G4 continuum (p < 0.01). This switching appears progressive, with certain MB Grp3 /MB Grp4 enhancer loci ''switching'' earlier and others later. For instance, the enhancer/DMR loci overlapping with the promoters of MB lineage development/differentiation genes LHX1, NEUROD2, LMX1A, and HLX on average ''switch'' at points 0.23, 0.49, 0.56, 0.87, respectively, on the G3/G4 continuum . We note also that the expression of each of these genes is significantly correlated with the G3/G4 continuum and DMR methylation (each p < 0.01). If we presuppose a model by which the G3/G4 continuum reflects interruption of early developmental cell fate at different points in different patients, then this observed switching is consistent with a developmental identity controlled by cumulative changes in underlying epigenetic architecture (i.e., patterns of methylation and/or enhancer usage) throughout a transition from an MB Grp3 to a MB Grp4 cell state.
The G3/G4 continuum is associated with posttranscriptional regulation of isoform expression and RNA editing To explore the clinico-biological significance of differentially expressed transcriptional isoforms across subgroups, Kallisto was used to estimate their abundance. Taking transcripts per million (TPM) >10 as indicative of a moderate-tohighly expressed isoform, it is notable that the diversity of isoforms being expressed across subgroups was significantly greater in MB Grp4 than MB Grp3 (p < 0.001, F = 9.877) [fig_ref] Figure 6: The group 3/group 4 continuum is mirrored in early human cerebellar development [/fig_ref].
A total of 153 genes were identified whose expression overall is invariant but for which the expression of specific isoforms correlates significantly with G3/G4 score . For instance, the overall expression of general transcription factor IIi (GTF2I) is ubiquitous, but a progressive isoform switch corresponding to the balance between b/d (GTF2I-215/GTF2I-218) and a/g (GTF2I-221/GTF2I-212) isoforms correlates significantly to G3/G4 score . These isoform switches are known to alter protein stability [bib_ref] Novel). Splice variants in the 5''UTR of Gtf2i expressed in the rat..., Shirai [/bib_ref] and subcellular localization [bib_ref] Role of splice variants of Gtf2i, a transcription factor localizing at postsynaptic..., Shirai [/bib_ref].
A total of 4,668,508 established RNA editing sites were profiled using the QEdit/Reditools pipeline . We observed significant differences in overall A-I editing level. The Overall Editing Index (OEI, i.e., the total number of reads with G at all known editing positions over the number of all reads covering the positions) differs significantly with respect to subgroup (F = 9.761, n = 223, p < 0.001). Post hoc testing showed RNA editing events in MB Grp4 to be significantly more numerous than in MB Grp3 and MB SHH (each p < 0.01) [fig_ref] Figure 6: The group 3/group 4 continuum is mirrored in early human cerebellar development [/fig_ref]. Analysis of 5,174 non-synonymous RNA editing sites showed 32 significantly differentially edited with respect to the G3/G4 continuum (p < 0.05; , the majority of which were more highly edited in MB Grp3 . One such RNA editing site is AZIN1 chr8:103841636T>C, known to result in a S367G substitution that causes conformational changes, cytoplasmic-to-nuclear translocation, and gain of function, increasing tumor potential in hepatocellular carcinoma [bib_ref] Recoding RNA editing of AZIN1 predisposes to hepatocellular carcinoma, Chen [/bib_ref] , non-small cell lung cancer [bib_ref] RNA editing of AZIN1 induces the malignant progression of non-small-cell lung cancers, Hu [/bib_ref] , colorectal cancer , and gastric cancer . It is also notable that ADAR1 and ADAR2 expression are both correlated with G3/G4 score (rho = 0.54, p < 0.001 and rho = 0.33, p < 0.001, n = 223, respectively), although expression was higher in MB Grp3 , which may speak to a context-dependent effect on specific loci.
Intratumoral cellular heterogeneity with respect to the G3/G4 continuum is apparent but constrained by subtype We projected our MB Grp3 /MB Grp4 metagenes onto a MB Grp3 / MB Grp4 scRNA-seq dataset comprising 4,256 cells from 15 individuals (5xSubtype-II, 2xSubtype-III, 1xSubtype-I, 2xSubtype-V, 4xSubtype-VIII) previously published by [bib_ref] Resolving medulloblastoma cellular architecture by single-cell genomics, Hovestadt [/bib_ref]. The approach used to derive these metagenes is very similar methodologically to the way Hovestadt et al. derived their metaprograms (e.g., use of NMF, projection between bulk and single cell) and both indicate a continuum of scores at both the bulk and single-cell level. We projected our bulk metagenes (describing group 3/group 4 transcriptional variability in 223 bulk tumor profiles) onto scRNA-seq data. In contrast, Hovestadt et al. projected their scRNA-seq metagenes (describing neuronal cellular differentiation and calculated from 17 MB Grp3/Grp4 ) onto bulk expression microarray samples. Our approach allowed us to impose a limit and scale between the extremities of tumor MB Grp3 and MB Grp4 transcriptional states, and in so doing, place each cell within a given sample on a common scale with our bulk tumors, allowing us to align cells with key tumor features such as subtype. MB Grp3 individuals were described by Hovestadt et al. as being dominated by cells with an undifferentiated progenitor-like expression program and MB Grp4 dominated by a differentiated neuronal-like program; to some extent our MB Grp3 and MB Grp4 metagenes appear to equate with the meta-programs described by and it is quite possible that both are describing similar phenomena. Of the 100 genes selected as the top genes by and 8/100 (metaprogram C-differentiated; e.g. KCNA1, ABLIM1, SPOCK2) would have been selected in the equivalent top 100 from our analysis. Notably, were invariant with respect to the G3/G4 continuum.
By placing bulk and scRNA-seq on a common scale, we show that the distribution of G3/G4 scores at the single-cell level indicates a certain amount of intratumoral cellular variation [fig_ref] Figure 5: Distribution of single cells along the group 3/group 4 continuum is limited... [/fig_ref] , but that the majority of cells fall within the same G3/G4 range observed in the equivalent subtype bulk RNA-seq profiles [fig_ref] Figure 5: Distribution of single cells along the group 3/group 4 continuum is limited... [/fig_ref]. For example, among MB subtype VIII individuals, 78% (667/853) of cells fall within the G3/G4 score 0-0.25 range, as per the equivalent subtype VIII bulk profiles [fig_ref] Figure 5: Distribution of single cells along the group 3/group 4 continuum is limited... [/fig_ref]. We should note that different bulk MB Grp3/Grp4 subtypes and their respective scRNA-seq populations occupy either a broader or narrower space on the G3/G4 continuum depending on the subtype; subtype V, for instance, is comparatively broad. In short, the phenomenon of a G3/G4 continuum observed in bulk RNAseq analysis is produced by populations of individual cells, which themselves display continuous G3/G4 expression characteristics. These are constrained to occupy a discrete part of the G3/G4 continuum as dictated by their MB Grp3 /MB Grp4 (I-VIII) subtype. by mouse modeling [bib_ref] Subtypes of medulloblastoma have distinct developmental origins, Gibson [/bib_ref] [bib_ref] Active medulloblastoma enhancers reveal subgroup-specific cellular origins, Lin [/bib_ref] and, more recently, by comparison with reference to mouse fetal cerebellum scRNA-seq datasets, which suggest a UBC origin for MB Grp4 . Such comparisons in embryonal tumors are predicated on the idea that partial transformation in an early prenatal cell interrupts development/differentiation, resulting in a proportion of the expression characteristics of the tumor-initiating cell being retained.
Here, we avoid any cross-species comparisons by using instead a human fetal cerebellum scRNA-seq reference set (69,174 cerebellar cells 9-21 post-conception weeks [PCWs]). We reconstructed a pseudotemporal cellular trajectory within a broadly defined RL lineage (12,243 cells, comprising RL precursors, excitatory cerebellar nuclei [eCN]/UBC, GC precursors , and GC neurons subdivided into four clusters [GN]) [fig_ref] Figure 6: The group 3/group 4 continuum is mirrored in early human cerebellar development [/fig_ref]. We projected our four subgroup metagenes onto these cerebellar cells, identifying those cells that showed the highest expression of each metagene. As an alternative analysis, we also performed canonical correlation analysis (CCA) and achieved comparable results (see description in STAR Methods). These cells occupy distinct branches of our lineage. High MB WNT metagene-expressing cells, as expected, occupy a discrete subset of the RL precursors [fig_ref] Figure 6: The group 3/group 4 continuum is mirrored in early human cerebellar development [/fig_ref]. High MB Grp3 /MB Grp4 metagene-expressing cells occupy a distinct eCN/UBC branch beginning with RL precursors (highly expressing MB Grp3 metagenes) and transitioning midway to eCN/UBC cells highly expressing the MB Grp4 metagene [fig_ref] Figure 6: The group 3/group 4 continuum is mirrored in early human cerebellar development [/fig_ref]. This cell trajectory in effect mirrors the G3/G4 continuum. This can be demonstrated formally by calculating a projected per-cell G3/G4 score, revealing a smooth transition from a MB Grp3 -like to a MB Grp4 -like expression state [fig_ref] Figure 6: The group 3/group 4 continuum is mirrored in early human cerebellar development [/fig_ref]. More straightforwardly, this is demonstrated by observing the significant change in expression with respect to pseudotime of those G3/G4 continuum-associated genes whose expression is sufficiently high to be consistently detectable within the relatively low-depth scRNA-seq data (each p < 0.01; [fig_ref] Figure 7: Key molecular characteristics of MB can be aligned to human fetal cerebellar... [/fig_ref].
Cells that express the MB SHH metagene most highly, as expected, occupy a GC developmental branch beginning with GCPs and extending partly into the earliest GN cell types [fig_ref] Figure 6: The group 3/group 4 continuum is mirrored in early human cerebellar development [/fig_ref]. Two metagenes representing MB SHH-Infant (primarily patients younger than 4 years) and MB SHH-Child (primarily patients older than 4 years), as described in previous studies [bib_ref] Molecular subgroups for clinical classification and outcome prediction in childhood medulloblastoma: a..., Schwalbe [/bib_ref] , were also projected onto the cells in this branch. This indicated a switch midway through the GC pseudotemporal lineage from a predominantly MB SHH-Infant metagene to a predominantly MB SHH-Child metagene expression; this coincided approximately with the first transition from GCPs to GNs [fig_ref] Figure 6: The group 3/group 4 continuum is mirrored in early human cerebellar development [/fig_ref]. Again, where the expression of individual genes that distinguish infant MB SHH from childhood MB SHH were sufficiently detectable within the scRNA-seq profiles, they were significantly associated with pseudotime (each p < 0.01; [fig_ref] Figure 7: Key molecular characteristics of MB can be aligned to human fetal cerebellar... [/fig_ref]. Thus, by aligning the oncogenic G3/G4 scale with the pseudotemporal scale, we were able to order and align tumorigenic events to specific points within fetal cerebellar developmental lineages [fig_ref] Figure 7: Key molecular characteristics of MB can be aligned to human fetal cerebellar... [/fig_ref]. MYC amplification, for instance, tends to coincide with the earlier RL pseudotemporal space, as opposed to KDM6A mutation, which occupies the later, more differentiated eCN/UBC space. Likewise for aneuploidies, the gain of chromosome 8 coincides with the earlier RL developmental space and i17q (as the sole copy-number alteration) with the later eCN/UBC cell types.
We note that as with the pseudotemporal transition from MB Grp3 to MB Grp4 or MB SHH-Infant to MB SHH-Child , there is also a literal temporal transition. The cerebellar cells most closely associated with the archetypal MB Grp3 are predominant at 11 PCW (and possibly before). By 18 PCW, those most closely associated with the archetypal MB Grp4 predominate. This persists until at least 20 PCW. On the RL to GN branch, the cells most closely associated with MB SHH-Infant are predominant at PCW 11 and reduced by PCW 20, at which point MB SHH-Child -associated cells predominate [fig_ref] Figure 7: Key molecular characteristics of MB can be aligned to human fetal cerebellar... [/fig_ref]. We should temper this observation by saying that the uniformity of sampling at each of these time points is uncertain.
This temporal staging from early to late forms of MB Grp3 / MB Grp4 is also mirrored in the average age of onset of disease. The distribution of age at diagnosis of each MB Grp3 /MB Grp4 (I-VIII) subtype closely parallels the distribution across the G3/ G4 continuum [fig_ref] Figure 7: Key molecular characteristics of MB can be aligned to human fetal cerebellar... [/fig_ref] , and there is a significant correlation between G3/G4 score and age at diagnosis [fig_ref] Figure 7: Key molecular characteristics of MB can be aligned to human fetal cerebellar... [/fig_ref].
# Discussion
Here, we show that, in regard to their transcriptomes, the primary intertumoral variation in MB Grp3 /MB Grp4 patients is continuous, in contrast to the discrete nature of the methylation MB Grp3 /MB Grp4 subtypes (I-VIII) [bib_ref] Intertumoral heterogeneity within medulloblastoma subgroups, Cavalli [/bib_ref] [bib_ref] The whole-genome landscape of medulloblastoma subtypes, Northcott [/bib_ref] [bib_ref] DNA methylation profiling of medulloblastoma allows robust subclassification and improved outcome prediction..., Schwalbe [/bib_ref] [bib_ref] Second-generation molecular subgrouping of medulloblastoma: an international meta-analysis of Group 3 and..., Sharma [/bib_ref]. This is not in itself contradictory, as we show that the MB Grp3 /MB Grp4 methylation subtypes are ordered along the G3/G4 continuum in discrete but partially overlapping domains [fig_ref] Figure 1: Group 3/group 4 medulloblastoma [/fig_ref]. Furthermore, as has been demonstrated previously [bib_ref] Intertumoral heterogeneity within medulloblastoma subgroups, Cavalli [/bib_ref] [bib_ref] Second-generation molecular subgrouping of medulloblastoma: an international meta-analysis of Group 3 and..., Sharma [/bib_ref] , the methylation subtypes are reflected to some extent in their expression profiles [fig_ref] Figure 1: Group 3/group 4 medulloblastoma [/fig_ref]. Nonetheless, these are shown here to be secondary expression characteristics subordinate to the overarching primary expression characteristic that is the G3/G4 continuum.
The position of an individual MB Grp3 /MB Grp4 tumor upon the continuum is significantly different in individuals with and without certain mutations, copy-number aberrations, clinicopathology, and histopathology. This is to be expected, as many of these have been shown to be non-randomly associated with MB Grp3 /MB Grp4 subtypes . That both methylation subtype and the expression continuum are related to key tumor characteristics and, indeed, to one another is clear. The question remains as to what extent the intertumoral variation in such characteristics may be better explained by position upon the continuum than by methylation subtype. For at least some of these characteristics, those that are frequent and not specific to single subtypes (e.g., MYC amplification, LCA, i17q, gain of chromosome 5, loss of chromosome 8), it seems that they are more relatable to position on the continuum .
The most striking association is between the G3/G4 continuum and risk of death, at least during the first 5 years post-diagnosis. Risk increases continuously with the G3/G4 continuum [fig_ref] Figure 2: Clinico-pathology, subtype, and survival are related to an individual's position on the... [/fig_ref] , the documented phenomenon of late (>5 years post-diagnosis) relapse in subtype VIII notwithstanding. We regard this study as a description of an extremely close and therefore important relationship between biology and clinical course rather than as an advocation for its use as a clinical biomarker. Those judgments should be made using prospective clinical trials, and the cohort used here, while sizable and carefully reviewed, is a retrospective cohort with all of the limitations and caveats that implies. Nevertheless, we note that when it comes to incorporating molecular data into risk stratification schemes, the use of a single G3/G4 risk score for all MB Grp3 /MB Grp4 patients has a certain pragmatic logic over atomizing a rare cancer into 8 separate subtypes.
Pathway analysis of the G3/G4 continuum shows a concomitant activation of oncogenic processes (e.g., MYC, MTOR, TP53) as tumors become more MB Grp3 -like, which itself suggests a more aggressive phenotype. The influence of the G3/G4 continuum also extends to post-transcriptional regulation (i.e., isoform usage and RNA editing). Here, we describe log-linear relationships showing the primacy of the continuum in multiple aspects of MB Grp3/Grp4 transcriptional biology. A close relationship with cell differentiation (e.g., CRX/NRL, neuronal differentiation) is also evident and consistent with previous descriptions of MB Grp3 /MB Grp4 biology cell identity and differentiation [bib_ref] Neuronal differentiation and cell-cycle programs mediate response to BET-bromodomain inhibition in MYC-driven..., Bandopadhayay [/bib_ref] [bib_ref] NRL and CRX define photoreceptor identity and reveal subgroup-specific dependencies in medulloblastoma, Garancher [/bib_ref]. This is further reflected in the progressive switches in methylation status that we observe within MB Grp3 /MB Grp4 specific enhancers [bib_ref] Active medulloblastoma enhancers reveal subgroup-specific cellular origins, Lin [/bib_ref].
We show here that the MB Grp3 /MB Grp4 continuum is produced by individual cells that themselves exist in the same expression continuum as the bulk tumors. In part, this was observed by [bib_ref] Resolving medulloblastoma cellular architecture by single-cell genomics, Hovestadt [/bib_ref] in their original analysis of their pooled MB Grp3 /MB Grp4 scRNA-seq data. They described two metagenes diverging according to MYC expression and described bulk tumors as composed of cells of either a predominately differentiated, undifferentiated, or intermediate type, which themselves represent a continuum of neuronal differentiation Article ll OPEN ACCESS . We have expanded this by fitting individual cells onto the same metagene scale used to define the bulk tumor transcriptome, thereby defining more precisely the range of transcriptional intratumoral heterogeneity within MB Grp3 /MB Grp4 tumors and showing that it appears to be confined to certain limits prescribed by the MB Grp3 /MB Grp4 subtype. This in turn is consistent with the finding that MB sampled from different areas of the tumor or at diagnosis and relapse rarely alter subgroup [bib_ref] Clinical outcomes and patient-matched molecular composition of relapsed medulloblastoma, Kumar [/bib_ref] [bib_ref] Recurrence patterns across medulloblastoma subgroups: an integrated clinical and molecular analysis, Ramaswamy [/bib_ref].
Unlike previous studies that attempted to define cells of origin, we used a human rather than a mouse scRNA-seq reference set for comparison. The use of a human atlas is significant because human RL persists longer through cerebellar development than the mouse and has unique cytoarchitectural features not shared with any other vertebrates [fig_ref] Figure 1: Group 3/group 4 medulloblastoma [/fig_ref]. Mouse RL is a Methylation subtypes (I-VIII), mutations, and copy-number changes are marked by box and whisker. Dot represents median distribution; thick line represents the interquartile range; and the thinner lines correspond to range. Dotted horizontal lines denote where the range extends up to a G3/G4 score of 0 and 1 (i.e., matching the ne plus ultra pseudotime after which G3/G4 score is unchanged and exact relationship must be extrapolated). Dotted vertical lines denote the boundaries between highG4, lowG4, G3.5, lowG3, and highG3 (these categories are arbitrary divisions of the continuum for the purposes of visualization and comparison and do not represent ''real'' subgroups). transient, proliferative stem cell zone present between embryonic day (E) E12.5 and E17.5, whereas human RL begins as a progenitor niche and is later compartmentalized into ventricular and subventricular zones, forming a human-specific progenitor pool within the posterior lobule, which persists until birth [bib_ref] Spatiotemporal expansion of primary progenitor zones in the developing human cerebellum, Haldipur [/bib_ref]. We show that the MB Grp3 /MB Grp4 continuum is paralleled by a fetal cerebellar lineage that begins with an RL progenitor and ends with eCN/UBC. Aligning oncogenic features to windows within developmental pseudotemporal space suggests that cellular development/differentiation may be interrupted by oncogenic features at (or at least before) a certain point in the developmental trajectory. More speculatively, this may suggest a certain developmental pseudotemporal window of opportunity for specific oncogenic events to provoke MB of a given subtype. How or if this occurs would need to be modeled and tested through further functional experimentation. Nevertheless, we suggest that such future modeling efforts would be best directed to the appropriate window within the developmental trajectory, and we provide here a map to do so. We also demonstrate a putative relationship between earlier/later cell types and the age of onset of the disease. Importantly, we were able to identify a developmental niche for each of the four main MB subgroups including a separate space for MB SHH-Child and MB SHH-Infant . Each of these is contained within a branch of the same early cerebellar lineage explicitly unifying each of the four subgroups to a common developmental antecedent, something not reported in previous studies. For instance, [bib_ref] Resolving medulloblastoma cellular architecture by single-cell genomics, Hovestadt [/bib_ref] were unable to identify a significant matching reference cell type for MB Grp3 and MB did not analyze MB WNT and note a prosaic resemblance of MB Grp3 to Nestin + early neural stem-like cells.
In conclusion, our findings point to the following important insights. First, that group 3/group 4 MB and their methylation subtypes exist transcriptionally upon a continuum and that this is mirrored entirely by an equivalent continuum of transcriptional cell types in early human fetal cerebellar development. Second, that by using a human scRNA-seq reference, all four MB subtypes can be linked to a common developmental antecedent within the RL lineage. Third, that transcriptional intratumoral heterogeneity is limited to certain domains within the continuum as dictated by subtype. Finally, that the continuum is linked with almost every aspect of group 3/group 4 molecular biology and clinico-pathology. We anticipate this to have implications for the future treatment and modeling of the disease-most pressingly, a need to match cell type with specific timing of mutations to develop faithful models.
## Limitations of the study
We wish to highlight the following, which we regard as some of the constraints and limitations of our study. In basing our conclusions upon a human developmental atlas, we note that we were selective, albeit based on prior knowledge, in the subset of cell types we considered to be potential candidate cells of originfiguratively, by assigning them to what we broadly described as the RL lineage, and literally, by the physical process of cell extraction and the points in early human development for which sampling was possible (PCWs 9-21). MB WNT in particular is thought to originate in the dorsal brainstem, and it may be that certain alternative cells of origin were excluded or curtailed on that basis. Nevertheless, previous studies follow a similar logic to our own and the coherent picture of the relationships between the subgroups would seem to bear out our choices. In addition, while we have aligned certain oncogenic features with specific developmental windows by virtue of their transcriptional resemblance, further functional experimentation will be required to determine if and how these oncogenic features provoke tumorigenesis specifically in these cell types.
We demonstrated a strong association between position on the G3/G4 continuum and risk of death. To what extent it may be effective and desirable to incorporate this into future clinical risk stratifications requires a more in-depth study, ideally as part of a prospective clinical trial. We have also touched upon the association between isoform expression or RNA editing and position on the G3/G4 continuum. We did this to demonstrate the primacy of the G3/G4 continuum in determining transcriptional biology; however, our description is by no means exhaustive and many important facets of MB RNA functional biology remain to be explored by future functional studies beyond the scope of the limited descriptions we have initiated here.
Finally, while we have demonstrated that the G3/G4 scores for individual cells appear to fall within a range on the continuum defined by the bulk tumors of the equivalent MB Grp3 /MB Grp4 subtypes (I-VIII), we should note that this was done with a relatively small number (n = 15) of individuals and that not all of the subtypes are covered equally. Further scRNA-seq analysis of individual MB Grp3 /MB Grp4 tumors should be undertaken to confirm the generalizability of this observation.
# Star+methods
Detailed methods are provided in the online version of this paper and include the following: were preprocess_cds, align_cds, reduce_dimension, cluster_cells, learn_graph, order_cells and plot_cells to visualize by UMAP. To rule out the possibility that an association between MB Grp3 and RL was simply an artifact of higher cellular proliferation we estimated the cell cycle phase using Seurat. Whilst there is a higher proportion of cycling cells in RL compared to eCN/UBC the same can also be said of GCP. This speaks against a default matching of MB Grp3 metagenes to any actively cycling cells. We also tried regressing out the effect of the cell cycle using the ''CC.difference'' (Seurat, R/Cran) method and reprojected our metagenes. This had little effect on the projection, as did removing all genes with ''cell cycle'' ontology. Top genes driving association with projected MB Grp3 and MB Grp4 metagenes in the developmental setting include ASIC2, GRIK1, KCNQ3, ANK3, ANKS1B, GRIA2; none of which are classic oncogenes, The GO terms significantly enriched (DAVID/EASE) are ''cell junction'', ''postsynaptic membrane'', ''integral component of plasma membrane'' and ''cell division'' (each Benjamini p < 0.01)
[formula] d KEY RESOURCES [/formula]
The relevant branches for MB Grp3 /MB Grp4 and MB SHH were divided as indicated [fig_ref] Figure 7: Key molecular characteristics of MB can be aligned to human fetal cerebellar... [/fig_ref] and the relationship between pseudotime and G3/G4 score/metagene was defined using a loess curve function. This enabled developmental and oncogenic events to be mapped onto a common scale [fig_ref] Figure 7: Key molecular characteristics of MB can be aligned to human fetal cerebellar... [/fig_ref]. Genes whose expression varied significantly according to pseudotime were detected using Moran's test statistic as implemented by monocle v3. For analysis of the differences between MB SHH-Infant and MB SHH-Child, a further metagene calculated using NMF rank = 2 only on MB SHH (67/331 samples) was additionally projected onto the single cells in the same manner as the other metagenes. For calculating empirical density, the density function was used (R/Bioconductor) except where weighted two-dimensional estimation was needed in which case the kde2d.weighted function from the package ggtern (R/Bioconductor) was used. Weights were calculated as the number of cells at a given sampling point (9-21PCW) as a proportion of the total number of cells sampled.
## Quantification and statistical analysis
Data analysis and visualization was carried out in R 3.5.3 except for the analysis of fetal cerebellar scRNA-seq which was performed using R 4.0.2. CRAN and Bioconductor packages used are given in the key resources table. To test significant association with time to death/progression, a log-rank test (test for trend as implemented by survMisc (R/Cran)) or Cox-regression was used. OS was used when assessing the basic relationship between G3/G4 score/subtype with risk of death. This was to maximise the number of data points (more OS than PFS data was available). When assessing use as an independent biomarker PFS was preferred as standard for the field as patients who relapse, almost without fail, go on to relapse.
A Kolmogorov-Smirnoff test was used to compare distributions across the G3/G4 continuum of patients with or without specific clinico-pathological mutational and copy number features. Where significant this indicates that patients with or without a given feature are significantly likely to be drawn from different G3/G4 score distributions. The implication being that with respect to a given feature patients are non-randomly distributed across the G3/G4 continuum. Where gene expression/pathway associations with G3/ G4 score are assessed, these are assessed using Pearson's correlation coefficient . The test statistics and significant p-values (p < 0.05) are stated in the text and figures and were adjusted for multiple hypothesis testing using Benjamini-Hochberg for high-dimensional analyses. Where values of n are given, these generally pertain to number of samples/individual patients except where otherwise indicated. Boxplots, where used, show dispersion as per standard i.e. (center line = median, box = interquartile range, whisker = range minus outliers).
Data were excluded where samples were clearly indicated to be duplicated across multiple related datasets. Additional exclusions were carried out for samples where methylation array detection p value did not reach significance threshold in at least 90% of the array. Methylation samples were excluded from the analysis if not confirmed as medulloblastoma by MNP2.0. In our analysis of the scRNA-seq dataset GSE119926 we excluded patients SJ970 and SJ723 due to the relatively few available cells.
[fig] Figure 1: Group 3/group 4 medulloblastoma (MB) form a transcriptional continuum (A) Heatmap showing 4 consensus NMF metagenes calculated for n = 331 MB and grouped by subgroup. MB Grp3 /MB Grp4 individuals are ordered by G3/G4 score. Annotation shows subgroup as determined by RNA-seq (expression subgroup), subgroup as determined by methylation (methylation subgroup), and methylation MB Grp3 /MB Grp4 subtype (I-VIII) as per Sharma et al. (2019) defined using Molecular Neuropathology version 2.0 (MNPv2) classifier (Capper et al., 2018) (Grp3/4 subtype). All of the other characteristics are indicated to be present or not by dark gray shading according to the following scheme: Infant, age at diagnosis younger than 3 years; Adult, age at diagnosis older than 16 years; DN, desmoplastic/nodular; LCA, large-cell/anaplastic; STR, subtotal resection; DOD, dead of disease. Side annotation (top left) shows a heatmap of chi-square residuals indicating subgroup enrichment and significance where relevant. The line plot (bottom) shows the G3/G4 score. (B) t-SNE plot showing MB Grp3 /MB Grp4 samples shaded by subgroup (top) and methylation MB Grp3 /MB Grp4 subtype (I-VIII) (bottom). Points where subtype (I-VIII) could not be determined confidently are not shown. (C) Violin plot showing G3/G4 score by MB Grp3 /MB Grp4 subtype (I-VIII).4 Cell Reports 40, [/fig]
[fig] Figure 2: Clinico-pathology, subtype, and survival are related to an individual's position on the group 3/group 4 continuum (A) Rug plot showing distribution of clinico-pathological features with respect to G3/G4 score. Summary counts are given according to the divisions of highG4, lowG4, G3.5, lowG3, and highG3 (these categories are arbitrary divisions of the continuum for the purposes of visualization and comparison and do not represent ''real'' subgroups) and reflected by the red line plots. The presence of a feature is indicated by a bold tick mark, the color of which indicates MB Grp3 /MB Grp4 methylation subtype (I-VIII). Adjusted p values for a Kolmogorov-Smirnoff statistic (D) are shown to denote non-random distribution of features with respect to G3/G4 score. Mismatch, mismatch between methylation and expression call; Infant, age at diagnosis younger than 3 years; M+, metastatic; DOD, dead of disease; LCA, large-cell/anaplastic; PRDM6, PRDM6 rearrangement.(B) Kaplan-Meier plot showing significant differences (Log-Rank test for trend) in MB Grp3 /MB Grp4 overall survival by G3/G4 continuum position. (C) Forest plot showing a multivariate Cox model fitted to progression-free survival and containing the independently significant variables highG3, MYC amplification, LCA, and M+. (D) Violin plot showing G3/G4 score (derived from methylation) by MB Grp3 /MB Grp4 (I-VIII) subtype. (E) Kaplan-Meier plot showing significant differences (Log-Rank test for trend) in MB Grp3 /MB Grp4 overall survival in patients aged older than 3 years by G3/G4 score (as derived from methylation values); n = 589. [/fig]
[fig] Figure 3, Figure 4: Position on the group 3/group 4 continuum corresponds linearly to oncogenic pathway activation and methylation of lineagespecific enhancers (A) Scatterplot showing significant correlation (p < 0.001) between MYC expression and G3/G4 score. Log-linear line of best fit is shown. Dotted lines divide into highG4, lowG4, G3.5, lowG3, and highG3 (these categories are arbitrary divisions of the continuum for the purposes of visualization and comparison and do not represent ''real'' subgroups), and log2 fold changes for each category relative to highG4 are shown. Error bars represent standard error of mean.(B) GSEA enrichment plot showing significant enrichment of MYC target genes. Genes were ranked by correlation with G3/G4 score. (C) Heatmap of ssGSEA results showing level of pathway enrichment for 223 MB Grp3 /MB Grp4 individuals ordered by G3/G4 score. MsigDB pathways are curated into pathways (see STAR Methods). (D) Lollipop plot showing mean beta fold change for DMRs within MB Grp3 /MB Grp4 specific enhancers/super-enhancers. The position on the x axis reflects the average point on the continuum at which the methylation level switches from hypo-to hypermethylation. (E) Plot showing an MB Grp3 /MB Grp4 -specific enhancer within the MB Grp3 -specific gene, LMX1A, which overlaps with a differentially methylated region significantly associated with the G3/G4 continuum. The mean beta value per G3/G4 category (highG4, lowG4, G3.5, lowG3, highG3) and MB Grp3 /MB Grp4 subtype (I-VIII) are shown by line and the 95% confidence interval (CI) by shaded area.8 Cell Reports 40, Position on the group 3/group 4 continuum is linearly associated with isoform usage and non-synonymous RNA editing events (A) Heatmap showing expression of top significantly differentially expressed isoforms of genes whose overall expression is otherwise not significantly differentially expressed with respect to G3/G4 score. (B) Schematic showing exon structure of 4 GTF2I isoforms significantly differentially expressed with respect to G3/G4 score (left) and scatterplot showing expression of these GTF2I isoforms versus G3/G4 score; line represents fitted log-linear model NB: GTF2I is not significantly differentially expressed at the gene level. [/fig]
[fig] Figure 5: Distribution of single cells along the group 3/group 4 continuum is limited according to DNA methylation subtype (A) Violin plot showing per-cell G3/G4 score (derived from projection onto scRNA-seq data) for 15 MB Grp3 /MB Grp4 patients aggregated by subtype. (B) Ridge plot showing distribution of per-cell G3/G4 score (derived from projection onto scRNA-seq data) for each of 15 MB Grp3 /MB Grp4 patients shown alongside the G3/G4 score distribution of equivalent subtype bulk tumors. n = x refers to number of individuals for bulk tumors and number of cells for the scRNA-seq data. Vertical black lines indicate from left to right the fifth percentile, median, and 95th percentile. Dotted vertical lines denote the boundaries between highG4, lowG4, G3.5, lowG3, and highG3 (these categories are arbitrary divisions of the continuum for the purposes of visualization and comparison and do not represent ''real'' subgroups).Cell Reports 40, 111162, August 2, 2022 11 [/fig]
[fig] Figure 6: The group 3/group 4 continuum is mirrored in early human cerebellar development(A) Uniform manifold approximation and projection (UMAP) plot of scRNA-seq profiles showing 12,243 cells of the RL lineage arranged according to developmental trajectory, which is indicated by the black line. Color denotes cell type as determined by graph-based clustering; RL, rhombic lip precursors; GCP, granule cell precursors; GN-I, GN-II, GN-III, GN-IV, 4 granule neuron cell types; eCN/UBC, excitatory cerebellar neurons/unipolar brush cells. (B) UMAP plot of the RL lineage with those cells within the top decile of metagene expression marked with the following colors: MB Grp4 , green; MB Grp3 , yellow; MB SHH , red; MB WNT , blue. (C) Scatterplot showing per-cell scaled metagene expression along the RL to eCN/UBC branch. Fitted sigmoid curves are shown, with SD indicated as dashed lines. The gray line represents a sigmoid curve fitted to per-cell G3/G4 score as a function of pseudotime. (D) Scatterplot showing per-cell scaled metagene expression along the GCP to GN branch. Fitted curves are shown with SD shown as dashed lines. Curves are scaled to be constrained to a range of 0 and 1, to be coherent with bulk analysis. For this reason, by definition, some individual cells lie outside the 0 and 1 range. [/fig]
[fig] Figure 7: Key molecular characteristics of MB can be aligned to human fetal cerebellar developmental niches (A) Schema showing the RL to eCN/UBC developmental branch, the relationship between pseudotime and G3/G4 score, and the staging of key tumor characteristics. From top to bottom: a violin plot showing pseudotime distribution of cells by time of sampling; color transition red to purple marks the point along the developmental trajectory at which cells are defined as eCN/UBC. A fitted sigmoid curve showing the relationship between pseudotime and G3/G4 score. Tumor characteristics are transformed from the G3/G4 scale to the pseudotime scale and marked at the appropriate points. Color bars represent subgroups. [/fig]
[table] TABLE d RESOURCE: AVAILABILITY B Lead contact B Materials availability B Data and code availability d EXPERIMENTAL MODEL AND SUBJECT DETAILS B Human tissue samples d METHOD DETAILS B Patient samples and study cohort B RNA-seq analysis d DNA METHYLATION ANALYSIS d SCRNA-SEQ ANALYSIS d QUANTIFICATION AND STATISTICAL ANALYSIS SUPPLEMENTAL INFORMATION Supplemental information can be found online at https://doi.org/10.1016/j. celrep.2022.111162. This work was supported by Cancer Research UK (C8464/A13457 and C8464/ A23391), the Tom Graham Trust/CCLG, LoveOliver, Star for Harris and the INSTINCT network, co-funded by The Brain Tumor Charity, Great Ormond Street Children's Charity, and Children with Cancer UK (grant no. 16/193). T.S.J. is grateful for additional funding from the Olivia Hodson Cancer Fund, Cancer Research UK, and the National Institute of Health Research. All of the research conducted at Great Ormond Street Hospital NHS Foundation Trust and UCL Great Ormond Street Institute of Child Health is made possible by the NIHR Great Ormond Street Hospital Biomedical Research Center. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health. [/table]
|
Gardnerella vaginalis causing pulmonary infection in a young adult: A novel case
A B S T R A C TGardnerella vaginalis is an anaerobic, gram-variable bacterium primarily found in vaginal microflora of women. Previous reports of G. vaginalis cultured in men are few and have primarily been limited to the gastrointestinal and genitourinary tract. 2−4 Few reports of G. vaginalis causing severe infections have been reported in the literature, including septicemia 7 and two cases of perinephric abscess.8,9There has been one previously reported case of G. vaginalis causing pulmonary complications that occurred in a male alcohol abuser. In our case review, we aim to demonstrate an unusual source of a pulmonary infection and highlight the importance of proper microbial isolation to guide treatment. Our patient is a young male who presented following multiple gunshot wounds including one to his head causing an intracranial hemorrhage, hydrocephalus, and a dural sinus thrombosis. His hospital course was complicated by a decline in neurological status treated with a craniotomy and external drain placement and multiple pulmonary infections. During his fever work-ups, he found to have G. vaginalis on mini-bronchoalveolar lavage and was subsequently treated with metronidazole. After treating his G. vaginalis pneumonia and other infectious sources, namely Haemaphilus influenzae and coagulase-negative staphylococcus pneumonias, his fevers and leukocytosis resolved and he was successfully discharged to a rehabilitation facility for neurologic recovery. To our knowledge, this is the second reported case of G. vaginalis isolated from a pulmonary culture and the first in a previously healthy, immunocompetent young male outside of the urinary tract.
# Introduction
Gardnerella vaginalis is an anaerobic, gram-variable bacterium primarily found in vaginal microflora of women. Previous reports of G. vaginalis cultured in men are few and have primarily been limited to the gastrointestinal and genitourinary tract [bib_ref] The urinary microbiota of men and women and its changes in women..., Gottschick [/bib_ref] [bib_ref] Gardnerella vaginalis and Mollicute detection in rectal swabs from men who have..., Cox [/bib_ref] [bib_ref] Colonization of Gardnerella vaginalis in semen of infertile men: prevalence, influence on..., Andrade-Rocha [/bib_ref]. Here, we present a case of Gardnerella vaginalis isolated from the lungs of a young man after experiencing a gunshot wound to the head. There has only been one prior case of G. vaginalis identified in a pulmonary culture, in which it was found in a 41-year-old alcoholic male [bib_ref] Gardnerella vaginalis bacteremia from pulmonary abscess in a male alcohol abuser, Legrand [/bib_ref]. This is the first case of Gardnerella vaginalis cultured from the lungs of a young adult.
## Case
A 20-year-old, previously healthy, African American male presented to the emergency department following three gunshot wounds to the left shoulder and one to the occiput. The shoulder wounds were superficial and did not require treatment. He had a GCS of 15 on arrival but a head CT revealed a closed skull base fracture with subdural hemorrhage and a CTA was suggestive of left transverse sinus thrombosis. His head injury was initially managed conservatively, without invasive monitoring, and he was started on a low dose heparin drip for his sinus thrombosis following multiple, stable head CT scans.
On hospital day three, his GCS declined, and a head CT revealed a new posterior fossa intracerebral hemorrhage and hydrocephalus. His heparin was discontinued and reversed with protamine. Subsequently, the patient went to the OR for a suboccipital craniectomy and placement of an external ventricular drain with the neurosurgery team. His postoperative course was complicated by stridor, anterior neck swelling, blood tinged sputum, fevers, and leukocytosis. Four days after surgery, a chest X-ray revealed bilateral lower lobe infiltrates and a new finding of pneumomediastinum. This was further visualized on a CT scan of his chest and an otolaryngologist performed direct laryngoscopy, bronchoscopy, and esophagoscopy and found no evidence of tracheal or esophageal perforation. A mini bronchoalveolar lavage (mini-BAL) was performed, whereby a telescoping catheter was introduced into the lower respiratory tract without the use of a bronchoscope and saline irrigation was used to sample the patient's respiratory secretions for culture. Empiric broad-spectrum antimicrobial therapy with intravenous Vancomycin and Piperacillin/ Tazobactam was initiated. The mini-BAL culture showed rare gramnegative rods.
By hospital day 11, the patient went on to develop fevers and leukocytosis, a second mini-BAL was performed, whose culture grew coagulase-negative staph, so the patient was started on ampicillin-sulbactam for four days. added empirically. Despite these regimens, he had continued fevers and subsequent CXR's showed new right-sided infiltrate prompting another mini-BAL and empiric Vancomycin and Cefepime treatment on hospital day 15. Another chest CT revealed a 2.5 cm diameter cavity in the right lower lobe with an air fluid level, consistent with an abscess. His third mini-BAL on hospital day 15 revealed Gardnerella vaginalis (> 10k cfu) growing in the culture obtained from his lavage fluid, so metronidazole was added to his regimen. Another mini-BAL was performed on hospital day 19 for lack of improvement, and suspected pneumonia, which was positive for Haemophilus influenzae. Antibiotics were switched to ceftriaxone and metronidazole, and his fever and leukocytosis resolved after a few days of treatment. After completion of his antibiotic course, the patient showed marked improvement and was discharged to a rehabilitation center where he has continued to improve.
# Discussion
Although G. vaginalis has been described as a component of normal vaginal flora, it is implicated to play a role in the pathogenesis of a number of diseases, primarily in the pathogenesis of Bacterial vaginosis in females [bib_ref] The aetiology of bacterial vaginosis, Turovskiy [/bib_ref]. G. vaginalis is coated with fimbriae that are thought to be responsible for the attachment of the bacterium to the vaginal epithelium as well as to other cells such as red blood cells [bib_ref] The aetiology of bacterial vaginosis, Turovskiy [/bib_ref]. Studies show that 90% of bacteria in the biofilm of BV patients was comprised of G. vaginalis. [bib_ref] The biofilm in bacterial vaginosis: implications for epidemiology, diagnosis and treatment, Verstraelen [/bib_ref] This biofilm development allows for further adherence of the bacterium to the epithelium and may play a part in causing systemic disease.
Male carriage of G. vaginalis is much less prevalent than female, however the organism is commonly detected in urethral and rectal samples. The rates of asymptomatic male urethral colonization has been estimated at 11.4% [bib_ref] Male carriage of Gardnerella vaginalis, Dawson [/bib_ref] , but rectal colonization results are conflicting, with reported rates ranging from 0 to 83.2% [bib_ref] Gardnerella vaginalis and Mollicute detection in rectal swabs from men who have..., Cox [/bib_ref] [bib_ref] Male carriage of Gardnerella vaginalis, Dawson [/bib_ref]. Of note, the 83.2% rate was obtained by qPCR assay rather than culture in a sample of men who have sex with men. Few reports of G. vaginalis causing severe infections have been reported in the literature, including septicemia [bib_ref] An unusual case of Gardnerella vaginalis septicaemia, Wilson [/bib_ref] and two cases of perinephric abscess [bib_ref] Multiple abscesses caused by Gardnerella vaginalis in an immunocompetent man, Calvert [/bib_ref] [bib_ref] Gardnerella vaginalis perinephric abscess in a transplanted kidney, Finkelhor [/bib_ref]. There has been one previously reported case of G. vaginalis causing pulmonary abscess that occurred in a male alcohol abuser [bib_ref] Gardnerella vaginalis bacteremia from pulmonary abscess in a male alcohol abuser, Legrand [/bib_ref].
In our patient, G. vaginalis was isolated from bronchoalveolar lavage culture, and was presumed to be contributing to the patient's pneumonia as well as potentially to his pulmonary abscess. It is likely that the pathogen was able to attach to the lung epithelium via its fimbriae similarly to how it attaches to vaginal epithelium. However, it is unclear how the organism was transmitted to the lung, although it is possible that it spread through direct contact. After identifying this pathogen, metronidazole was added to his antibiotic regimen of vancomycin and cefepime. For unclear reasons, the patient did not show improvement in the first 4 days after metronidazole initiation, but he did quickly improve after day 19 when vancomycin and cefepime were switched to ceftriaxone. Although he was found to have co-infection with H. influenzae, this pathogen was sensitive to the cefepime the patient had been receiving.
# Conclusion
To our knowledge, this is the second reported case of G. vaginalis isolated from a pulmonary culture and the first in a previously healthy, immunocompetent young male outside of the urinary tract. Our case demonstrates an unusual source of pulmonary infection and highlights the importance of proper microbial isolation to guide treatment. |
Central vestibular syndrome in a red fox (Vulpes vulpes) with presumptive right caudal cerebral artery ischemic infarct and prevalent midbrain involvement
A wild young male red fox (Vulpes vulpes) was found in the mountainous hinterland of Rome (Italy) with a heavily depressed mental status and unresponsive to the surrounding environment. Neurological examination revealed depression, left circling, right head tilt, ventromedial positional strabismus and decreased postural reactions on the left side. Neurological abnormalities were suggestive of central vestibular syndrome. Two consecutive MRIs performed with 30 days interval were compatible with lacunar ischemic infarct in the territory of right caudal cerebral artery and its collateral branches. The lesion epicentre was in the right periaqueductal portion of the rostral mesencephalic tegmentum. Neuroanatomical and neurophysiological correlation between lesion localization and clinical presentation are discussed.
# Introduction
Cerebral ischemic stroke is a sudden interruption of arterial blood flow in a limited area of the brain caused by vascular obstruction, impaired vasodilation or increased blood viscosity leading to neuronal injury and parenchymal necrosis [bib_ref] Clinical and topographic magnetic resonance characteristics of suspected brain infarction in 40..., Garosi [/bib_ref] [bib_ref] Hypothyroid-associated central vestibular disease in 10 dogs: 1999-2005, Higgins [/bib_ref] [bib_ref] Vascular encephalopathies in dogs: incidence, risk factors, pathophysiology, and clinical signs, Hillock [/bib_ref] [bib_ref] Ischaemic and haemorrhagic stroke in the dog, Wessmann [/bib_ref]. Depending on the size of the involved vessel, cerebral infarcts are distinct in territorial infarcts, associated with disease of superficial, large diameter blood vessels and lacunar infarcts, deriving from disease of small, intraparenchymal, penetrating arteries [bib_ref] Clinical and topographic magnetic resonance characteristics of suspected brain infarction in 40..., Garosi [/bib_ref]. Clinical signs of focal ischemic encephalopathy are variable and ultimately related to the involved brain area (telencephalon; thalamus or midbrain; cerebellum; brainstem) [bib_ref] Vascular encephalopathies in dogs: incidence, risk factors, pathophysiology, and clinical signs, Hillock [/bib_ref]. Although a large percentage, more or less 40%, of ischemic strokes have an unknown etiology, several underlying causes have been recognized in dogs and cats including, hypertension, endocrine, kidney, heart, metastatic diseases, parasitic thromboembolism [bib_ref] Cerebrovascular disease in dogs and cats, Garosi [/bib_ref] and Evans' syndrome. Focal ischemic encephalopathy is frequently diagnosed in companion animals and, in the last decade, it has been more commonly recognized likely because of both increased awareness of it as a potential neurologic disorder and increased availability of magnetic resonance imaging (MRI) and computed tomography [bib_ref] Vascular encephalopathies in the dog and cat, Dewey [/bib_ref] [bib_ref] Vascular encephalopathies in dogs: incidence, risk factors, pathophysiology, and clinical signs, Hillock [/bib_ref]. However, despite the large amount of medical data regarding ischemic stroke in dogs, description of such condition in wild canids is lacking in veterinary literature. In this paper, the authors describe the clinical signs, MRI findings and follow up of a presumptive focal ischemic encephalopathy in a wild red fox (Vulpes vulpes) in Southern Italy.
Case details A wild young male red fox (Vulpes vulpes) weighting 10 kg was found in the mountainous hinterland of Rome (Italy) with a heavily depressed mental status and unresponsive to the surrounding environment. The animal was able to stand, with pronounced right head tilt, showing no aggressiveness nor fear towards humans. The fox received a single dose of dexamethasone and amoxicillin-clavulanic acid by the first examiner veterinarian and three days later the animal was referred to the Pingry Veterinary Hospital of Bari with an improved reactivity towards the surrounding environment. On physical examination, no abnormalities were observed. Neurological examination revealed depression, circling to the left, right sided head tilt and decreased postural reactions on the left side. Ventromedial positional strabismus was the most reliable abnormality detectable on cranial nerves examination [fig_ref] Figure 1: Neurological examination [/fig_ref]. The menace response was questionable on both eyes. Neurological abnormalities suggested a multifocal encephalic neuroanatomic localization with right forebrain and central vestibular system involvement. Because of the lack of a reliable clinical history, creating an appropriate differential diagnosis list was not possible. Brain MRI was performed using a 0.25Tesla permanent magnet (ESAOTE VET-MR GRANDE, Esaote, Genoa, Italy) with the fox under general anesthesia. MRI sequences used included a Fast SE T2-W acquired in sagittal and transverse plane, a fluid attenuated inversion recovery (FLAIR) image, and a SE T1-W acquired in transverse plane before and after intravenous administration of paramagnetic contrast medium (Magnegita, gadopentetate dimeglumine 500mmol/mL, insight agents; 0.15mmol/kg BW). T2W and FLAIR images showed a sharply hyperintense, well demarcated lesion at the ventromedial surface of the right temporal lobe with focal involvement of the ventrolateral portion of mesencephalic tegmentum and adjacent caudo-ventrolateral portion of the right thalamus (Figs. 2, 3). In the temporal lobe signal changes involved both gray and white matter with major involvement of cerebral cortex of the parahippocampal gyrus and ventral portion of right hippocampus. The lesion appeared isointense on T1-W images with mild and irregular enhancement after contrast medium administration [fig_ref] Figure 2: First MRI of the brain [/fig_ref]. No mass effect was evident. The distribution of the lesion matched the territory of the right caudal cerebral artery and its paramedian branches with possible involvement of caudal perforating arteries arising from basilar bifurcation. These findings were primarily suggestive of vascular ischemic lesion while inflammatory conditions were considered less likely. Protein levels (14 mg/dl; reference interval: < 30 mg/dl) and cell count (3 cell/µl; reference interval: 0-3 cell/µl) of a CSF sample collected from the cerebellomedullary cistern were apparently normal. Fecal flotation test was positive for ascarids eggs and two consecutive Baermann tests were negative for strongyles and Crenosoma vulpis larvae. contrastenhanced T1-weighted MRI images obtained at the level of the rostral midbrain. T2W ad FLAIR images show a sharply hyperintense well demarcated lesion affecting the ventrolateral portion of mesencephalic tegmentum (arrows) and adjacent caudo-ventro-lateral portion of the right thalamus (not shown). The lesion extends to the medial surface of the right temporal lobe with involvement of cerebral cortex of the parahippocampal gyrus and ventral portion of right hippocampus (arrowheads).The lesion appears isointense on T1-W images with faint and irregular enhancement after contrast medium administration. No mass effect is evident. Complete blood (cell) count (CBC), and biochemical profile were also carried out. Hemato-biochemical analysis apparently revealed erythrocytosis, microcytosis, hypochromic red blood cells and mild neutropenia -hemato-biochemical results were compared with the reference values of the island fox and dogs, and with mean values available for Vulpes velox). Within 30 days of hospitalization the fox showed progressive clinical improvement without medical therapy. A second neurological examination revealed normalization of vestibular signs and left side postural reactions with residual mild left circling.
A MRI of the brain was repeated using the same sequence protocol. The second imaging examination showed significant reduction in size of the primary lesions and fluid replacement of T2 and FLAIR hyperintense mesencephalic areas previously detected [fig_ref] Figure 4: Repeat MRI of the brain acquired 30 days after the first examination [/fig_ref]. Based on MRI pattern of distribution and evolution of the lesions, spontaneous improvement of neurological signs and CSF analysis, a presumptive diagnosis of ischemic infarct in the territory of right caudal cerebral artery and its collateral branches was made. Involvement of right caudal perforating arteries arising from basilar bifurcation was also considered.
# Discussion
Canidae is one of the most studied mammalian groups. From comparative studies on their neuroanatomy it has been clarified that the external cerebrum morphology of the modern Canidae is extremely uniform and, except for small differences in the shape and size of frontal gyri (sigmoid and proreal gyrus), characterized by lack of important differences between the genera [bib_ref] Outlines of canid and felid brain evolution, Radinsky [/bib_ref] [bib_ref] Evolution of the canid brain, Radinsky [/bib_ref] [bib_ref] The evolutionary history of dog brains, Radinsky [/bib_ref] [bib_ref] External brain anatomy in relation to phylogeny of Caninae (Carnivora: Canidae), Lyras [/bib_ref]. Such uniformity can also be observed from the comparison of cross-sectional anatomy of the red fox and dog brain as revealed by previous MRI studies in both species [bib_ref] Magnetic Resonance Imaging and cross-sectional anatomy of the brain of the red..., Kassab [/bib_ref] [bib_ref] Clinical anatomy of the canine brain using magnetic resonance imaging, Leigh [/bib_ref]. As reported in dogs [bib_ref] Clinical and topographic magnetic resonance characteristics of suspected brain infarction in 40..., Garosi [/bib_ref] , the imaging findings (shape and distribution of the lesions, absence of mass effect, signal intensity and evolution of the lesions) in the fox herein examined were suggestive of ischemic infarct in the territory of the right caudal cerebral artery (CCA). In dogs, the CCAs arise from the caudal communicant arteries of the cerebral arterial circle (circle of Willis) and run caudo-laterally supplying blood to the caudal and medial surface of each telencephalic hemisphere.
Along their course the CCAs give rise to small branches to the ventro-medial portion of temporal lobes, to the caudo-lateral part of the thalamus and to the lateral mesencephalon (perforating arteries) [bib_ref] Clinical and topographic magnetic resonance characteristics of suspected brain infarction in 40..., Garosi [/bib_ref] [bib_ref] Vascular encephalopathies in dogs: incidence, risk factors, pathophysiology, and clinical signs, Hillock [/bib_ref]. The same territorial distribution of the CCAs have been proven in detailed anatomical studies on the vascular brain anatomy in the red fox and pampas fox (Pseudalopex gumnocercus). In particular, specific collateral CCA branches to the piriform lobe, parahippocampal gyrus, thalamus and mesencephalon have been described in such species [bib_ref] Macro-Anatomical Investigation of the Cerebral Arterial Circle (Circle of Willis) in Red..., Ozudogru [/bib_ref]. The distribution of the lesions observed on MRI in the ventrolateral mesencephalon, ventrolateral caudal thalamus, piriform lobe and para-hippocampal gyrus confirmed the course of the collateral CCA branches as also described in foxes [bib_ref] Macro-Anatomical Investigation of the Cerebral Arterial Circle (Circle of Willis) in Red..., Ozudogru [/bib_ref] and matched the most commonly affected regions reported in dogs with CCA infarction [bib_ref] Clinical and topographic magnetic resonance characteristics of suspected brain infarction in 40..., Garosi [/bib_ref]. However, for thalamic and midbrain lesions, concomitant involvement of caudal perforating arteries arising from basilar bifurcation could not be ruled out [bib_ref] Clinical and topographic magnetic resonance characteristics of suspected brain infarction in 40..., Garosi [/bib_ref]. In this fox, the presence of vestibular signs resulted apparently unsolved and speculative. In cats, unilateral experimentally-induced mesencephalic lesions, resulted in lateral tilt of the head toward the opposite side, while bilateral lesions induced dorsiflexion of the head . Anatomical basis associated with these abnormal head posture involve dysfunction of the interstitial nucleus of Cajal (INC) (in the rostral midbrain adjacent to the periaqueductal gray matter), interstitiospinal fibres (that run in the medial longitudinal fasciculus) [bib_ref] Sites of termination of interstitiospinal fibers in the cat. An experimental study..., Nyberg-Hansen [/bib_ref] and their control on the rostral cervical muscles . In particular, unilateral induced lesion in INC or in its descending fibres up to the caudal region of mesencephalon, cause an increase in activity of the major ipsilateral dorsal neck muscles and in the contralateral obliquus capitis caudalis resulting in controlateral head tilt [bib_ref] Interstitialvestibular interaction in the control of head posture, Fukushima [/bib_ref] [bib_ref] Ipsilateral pallidal control of the sternocleidomastoid muscle of cats: Relationship to the..., Kavaklis [/bib_ref]. Bilateral lesions induce activation of dorsal neck muscles producing dorsiflexion of the head [bib_ref] Rigidity and dorsiflexion of the neck in progressive supranuclear palsy and the..., Fukushima-Kudo [/bib_ref]. In dogs, mesencephalic dysfunction has been related with different abnormal neck and head posture [bib_ref] Clinical and topographic magnetic resonance characteristics of suspected brain infarction in 40..., Garosi [/bib_ref] [bib_ref] 202 Infarction in a Dog with Evans' Syndrome. Case Reports in Veterinary..., Ricciardi [/bib_ref] [bib_ref] Dorsal midbrain syndrome associated with persistent neck extension: clinical and diagnostic imaging..., Canal [/bib_ref]. In dogs with ventrolateral thalamic infractions, the lesions were associated with mesencephalic involvement resulting in controlateral or ipsilateral http://www.openveterinaryjournal.com M. head tilt as prevailing vestibular sing [bib_ref] 202 Infarction in a Dog with Evans' Syndrome. Case Reports in Veterinary..., Ricciardi [/bib_ref]. Moreover, in two dogs with intracranial expansive lesions exerting bilateral compression of dorsal mesencephalon, permanent neck extension (retrocollis) has been observed and attributed to bilateral INC dysfunction [bib_ref] Dorsal midbrain syndrome associated with persistent neck extension: clinical and diagnostic imaging..., Canal [/bib_ref]. Interestingly, neurological signs detected in this fox perfectly reflected those reported for ventrolateral thalamic infractions with mesencephalic involvement in dogs [bib_ref] Clinical and topographic magnetic resonance characteristics of suspected brain infarction in 40..., Garosi [/bib_ref] [bib_ref] 202 Infarction in a Dog with Evans' Syndrome. Case Reports in Veterinary..., Ricciardi [/bib_ref]. In this fox, as well as in previously reported canine cases, while compulsive circling and controlateral proprioceptive deficit were well explained by prosencephalic lesions, the neuroanatomic explanation of the concomitant ipsilateral vestibular signs is consistent with unilateral involvement of nucleus of Cajal in the rostral mesencephalon. Thalamic dysfunction is also suspected to cause vestibular signs in dogs, especially after acute lesions. The pathways for conscious balance perception involving a relay from a thalamic nucleus seems implicated in the vestibular thalamic syndrome [bib_ref] The vestibular cortex: its locations, functions, and disorders, Brandt [/bib_ref]. In attempt to identify possible underling risk factors for brain infarction, hemato-biochemical analysis were performed. Unfortunately, normal reference values for red foxes (Vulpes vulpes) are not available in veterinary medical literature. To the author's knowledge hemato-biochemical reference intervals for wild foxes are available only for the island fox (urocyon littoralis) [bib_ref] Biochemical and hematologic reference intervals for the endangered island fox (Urocyon littoralis), Inoue [/bib_ref]. In addition hematologic and serum chemistry mean values are available for the species Vulpes velox. From the analysis of hematobiochemical abnormalities detected in our fox using the reference values of the island fox and dogs, and by the comparison with mean values available for Vulpes velox species, erythrocytosis associated with microcytosis and hypochromic red blood cells and mild neutropenia were apparently detectable . However, taking into account the possible variability existing among these different species, such abnormal findings were considered of doubtful interpretation. Thus, in this fox the possible cause of brain infarction remains unclear due to the lack of further diagnostic evaluations (blood pressure measurement, thoracic and abdominal imaging evaluation, urinalysis, endocrine tests) and the equivocal relevance of hematobiochemical abnormalities. Finally, this case is the first report of a presumptive thalamic and midbrain infarction in a fox showing multifocal encephalic syndrome with predominant vestibular dysfunction. This case suggests not only an anatomic but also a neurophysiologic analogy between dogs, cats and foxes.
[fig] Figure 1: Neurological examination. (A,B,C): Major neurological examination included depression, left circling with right head tilt. (D): Ventromedial positional strabismus on the right eye. (E,F): decreased postural reactions on the left side. Neurological abnormalities suggested a multifocal encephalic neuroanatomic localization with right forebrain and central vestibular system involvement. [/fig]
[fig] Figure 2: First MRI of the brain. (A): midsagittal and (B): right parasagittal T2-weighted MRI images. (C): Transverse T2weighted, (D): T1-weighted, (E): FLAIR and (F): [/fig]
[fig] Figure 3: First MRI of the brain. (A): Transverse T2-weighted and (B): FLAIR MRI images obtained at the level of the caudal thalamus. The mesencephalic lesion extended cranially with focal involvement of the caudo-ventro-lateral portion of the right thalamus (arrows). In the temporal lobe signal changes involved both gray and white matter of the parahippocampal gyrus. (C): Oblique-transverse FLAIR image at level of cross reference in image (D), showing in detail the mesencephalic and telencephalic lesion extension (thin arrow). [/fig]
[fig] Figure 4: Repeat MRI of the brain acquired 30 days after the first examination. (A): Transverse T2-weighted, (B): T1weighted, (C): FLAIR and (D): contrast-enhanced T1weighted MRI images obtained at the level of the rostral midbrain. There is significant reduction in size of the primary lesions and fluid replacement of T2 and FLAIR hyperintense mesencephalic areas previously detected (arrows). (E): Transverse T2-weighted and (F): FLAIR MRI images obtained at the level of the caudal thalamus showing normal parenchymal signal intensity of the thalamic area previously involved by the lesion. Slight residual hyperintensity is observed at the level of parahippocampal gyrus (arrowheads) with fluid signal at the level of the ventral portion of right hippocampus (thin arrows). [/fig]
|
Cell cycle related proteins in hyperplasia of usual type in breast specimens of patients with and without breast cancer
Background: Hyperplasia of usual type (HUT) is a common proliferative lesion associated with a slight elevated risk for subsequent development of breast cancer. Cell cycle-related proteins would be helpful to determine the putative role of these markers in the process of mammary carcinogenesis. The aim of this study was to analyze the expression of cell cycle related proteins in HUT of breast specimens of patients with and without breast cancer, and compare this expression with areas of invasive carcinomas.Results:Immunohistochemical evaluation was performed using antibodies against cell cycle related proteins ER, PR, p53, p21, p63, and Ki-67 in hyperplasia of usual type (HUT) in specimens of aesthetic reduction mammaplasty (ARM), in specimens of mammaplasty contralateral to breast cancer (MCC), and in specimens of invasive mammary carcinomas (IMC) presenting HUT in the adjacent parenchyma. The results showed that the immunoexpression of ER, PR, p21, p53, p63, and KI-67 was similar in HUT from the three different groups. The p63 expression in myoepithelial cells showed discontinuous pattern in the majority of HUT, different from continuous expression in normal lobules. Nuclear expression of p53 and p21 was frequently higher expressed in IMC and very rare in HUT. We also found cytoplasmic expression of p21 in benign hyperplastic lesions and in neoplastic cells of IMC.Conclusion:Our data failed to demonstrate different expression of cell cycle related proteins in HUT from patients with and without breast cancer. However, we found discontinuous expression of p63 in myoepithelial cells around HUT adjacent to carcinomas and cytoplasmic expression of p21 in epithelial cells of hyperplastic foci. Further studies are needed to determine how these subgroups relate to molecular abnormalities and cancer risk.
# Background
The hypothetical multistep model of carcinogenesis indicates that breast cancer develops via a series of intermediate hyperplastic lesions through in situ to invasive carcinomas, with the risk of developing carcinoma increasing at each stage [bib_ref] Prognostic and predictive factors in breast cancer by immunohistochemical analysis, Allred [/bib_ref] [bib_ref] Prognostic significance of TP53 alterations in breast carcinoma, Andersen [/bib_ref] [bib_ref] Apoptosis inhibitory activity of cytoplasmic p21(Cip1/WAF1) in monocytic differentiation, Asada [/bib_ref]. Epidemiological studies demonstrated an increased risk of developing breast cancer associated with proliferative breast lesions. Hyperplasia of usual type (HUT) is a common proliferative lesion associated with a slightly elevated risk for subsequent development of breast cancer (relative risk = 1.6, augmented to 2.1 with positive family history) [bib_ref] Prognostic and predictive factors in breast cancer by immunohistochemical analysis, Allred [/bib_ref] [bib_ref] Prognostic significance of TP53 alterations in breast carcinoma, Andersen [/bib_ref] [bib_ref] Clinical relevance of immunohistochemical expression of p53-targeted gene products mdm-2, p21 and..., Bankfalvi [/bib_ref]. HUT is not necessarily a direct precursor of invasive breast carcinoma but may identify individuals whose breast tissue has acquired a molecular alteration that can facilitate the eventual development of this disease [bib_ref] p21WAF1 immunohistochemical expression in breast carcinoma: correlations with clinicopathological data, oestrogen receptor..., Barbareschi [/bib_ref].
The defective function of regulatory cell cycle elements, like estrogen and progesterone receptors, Ki-67, p53, p21 WAF1 and p63 leads toward increased proliferation and, in addition, expansion of genome damaged cells. Cell cycle-related markers would be helpful to determine the putative role of these markers in the process of mammary carcinogenesis [bib_ref] Doglioni C: p63, a p53 homologue, is a selective nuclear marker of..., Barbareschi [/bib_ref].
Many studies evaluated cell cycle-related proteins in invasive breast carcinomas, but there are few studies evaluating these proteins in HUT [bib_ref] Increased level of p21 in human ovarian tumors is associated with increased..., Barboule [/bib_ref] [bib_ref] Prognostic value of p21(WAF1) and p53 expression in breast carcinoma: an immunohistochemical..., Caffo [/bib_ref] [bib_ref] Dissociation between steroid receptor expression and cell proliferation in the human breast, Clarke [/bib_ref]. Some molecular alterations may already be present in the earliest stages of breast cancer development. Detection of these alterations may be important for understanding the pathogenesis and also for risk assessment of premalignant breast lesions.
Incidental cancers or precursor lesions are rare in specimens of cosmetic mammaplasty compared to reduction mammaplasty specimens performed for symmetry of contralateral breast in women with breast cancer undergoing mastectomy or conservative surgery. Our hypothesis is that the expression of cell cycle related proteins would be different in HUT lesions from women at higher risk of breast cancer or with breast cancer, compared to those HUT from women without breast cancer. The aim of this study is to analyze the expression pattern of cell cycle related proteins ER, PR, p53, p21 WAF1 , p63, and Ki-67 in hyperplasia of usual type (HUT) of breast specimens of patients with and without breast cancer, and compare this expression with neoplastic cells of invasive carcinomas.
# Results
The age of patients submitted to ARM ranged from 30 to 67 years (mean 43.9 years; SD = ± 7.4 years), of patients with IMC ranged from 30 to 86 years (mean 55. 7; SD = ± 13.1 years), and age of patients submitted to MCC ranged from 30 to 75 years (mean 51.6; SD = ± 12.4 years). Patients were divided according to the menopausal status into pre-menopausal patients (≤ 50 years) and post-menopausal patients (> 50 years). The mean age of patients with IMC and submitted to MCC was significantly greaterthan patients submitted to ARM (p < 0.005). There was no statistically significant difference between mean age of patients from IMC and MCC groups.
The histologic review showed that HUT was associated with other benign breast lesions in the majority of the cases. Histologic findings were varied in ARM, MCC, and IMC specimens. Usually, the strongest ER staining was noted at the periphery of the hyperplastic foci . The majority of the epithelial cells of HUT in all specimens showed positive staining for ER, PR, and Ki-67 (Figure 1; [fig_ref] Table 3: Positivity for ER, PR, p53, p21, p63, [/fig_ref]. The ER immunostaining was localized in the nuclei and showed some variability in intensity even in individual lesions of the same case.
The p63 expression was detected in the majority of the myoepithelial cell nuclei in normal lobules and in HUT. p63-positive cells around HUT foci occurred as a discontinuous layer in 38.1% in ARM, in 73.3% in MCC, and in 64.7% of myoepithelial cells surrounding HUT adjacent to IMC and 2D). The p63 expression was continuous in myoepithelial cells of normal lobules and ducts. There was no difference in the percentage of positive cells for ER, PR, p21 WAF1 , p53, p63, and KI-67 in HUT of ARM, MCC and IMC (p > 0, 05). The mean percentage of ER+, PR+, Ki-67+ in epithelial cells and, p63+ in myoepithelial cells of HUT from all groups was significantly higher than positivity in neoplastic cell of IMC [fig_ref] Table 3: Positivity for ER, PR, p53, p21, p63, [/fig_ref]. The p21 expression in IMC was predominantly nuclear (55.9%). Cytoplasmatic staining was seen in neoplastic cells in 23.5% of cases . Nuclear staining was detected in cells of HUT-MCC in 2 cases (5.9%; and 6 cases showed cytoplasmatic staining in hyperplastic cells .
# Discussion
The aim of this study was to determine alterations in the expression of proteins involved in proliferation and cell cycle in HUT cells of patients with and without invasive breast cancer. Our analysis showed no difference in the cell cycle related proteins immunoexpression in HUT from the three different groups, in spite of age and menopausal status. Similar results were obtained by [bib_ref] Breast cancer risk associated with estrogen receptor expression in epithelial hyperplasia lacking..., Gobbi [/bib_ref] [7] evaluating ER expression in usual hyperplasia without atypia of patients who developed breast cancer compared with patients who did not.
In our study, we found that ER and PR immunoexpression was significantly higher in HUT cells than in neoplastic cells of IMC specimens. Even in all 16 cases of ER-negative tumors the epithelial cells of HUT were positive for ER. Our results are in agreement with other investigators who found higher levels of ER expression in benign breast epithelium of patients who developed breast cancer compared to controls [bib_ref] Increased level of p21 in human ovarian tumors is associated with increased..., Barboule [/bib_ref] [bib_ref] Dissociation between steroid receptor expression and cell proliferation in the human breast, Clarke [/bib_ref] [bib_ref] Transforming growth factor beta induces the cyclin-dependent kinase inhibitor p21 through a..., Datto [/bib_ref] [bib_ref] p63 expression profiles in human normal and tumor tissues, Como [/bib_ref]. The presence of positive ER staining in normal lobules increases the breast cancer risk and the likelihood of progression to cancer [bib_ref] Increased level of p21 in human ovarian tumors is associated with increased..., Barboule [/bib_ref]. It occurs through the increase of the rate of cell proliferation by both recruiting non-cycling cells into the cell cycle and by shortening the overall cell cycle time due to a reduction in the length of G1 phase [bib_ref] p63 expression profiles in human normal and tumor tissues, Como [/bib_ref].
Previous comparison between normal and precancerous breast biopsies has shown that ER expression is relatively low in normal epithelium and slightly increased in HUT [bib_ref] Risk factors for breast cancer in women with proliferative breast disease, Dupont [/bib_ref]. Recent studies indicate that HUT is a heterogeneous entity containing subgroups identified according to the criterion of ER-α (+) proliferating cells and this fact could explain the different biologic behavior of HUT. In our study, the ER positive cells were more often found at the periphery of hyperplastic lesions. Similar pattern of ER positivity was previously described by [bib_ref] Breast cancer risk associated with estrogen receptor expression in epithelial hyperplasia lacking..., Gobbi [/bib_ref] [bib_ref] Increased level of p21 in human ovarian tumors is associated with increased..., Barboule [/bib_ref]. It is possible that the ER+ epithelial cells at the periphery of HUT represent the most proliferative group of cells, in spite of low positivity for Ki-67 in sequential sections of the same lesion. The estrogen exposure may stimulate a clonal proliferation of some ER+ cells or may increase the chance of spontaneous mutations [bib_ref] Dissociation between steroid receptor expression and cell proliferation in the human breast, Clarke [/bib_ref].
In normal breast, there is a negative association between expression of ER and Ki-67, indicating either that ER+ cells are non-dividing or that the receptor is down-regulated as cells enter cycle [bib_ref] Risk factors for breast cancer in women with proliferative breast disease, Dupont [/bib_ref] [bib_ref] Prognostic factors in breast cancer. College of American Pathologists Consensus Statement, Fitzgibbons [/bib_ref]. This important correlation breaks down in many ER+ cancers, where the receptor is often detected in proliferating cells [bib_ref] Prognostic factors in breast cancer. College of American Pathologists Consensus Statement, Fitzgibbons [/bib_ref]. However, cells co-expressing ER-α and Ki-67 have been found in precan-cerous lesions and correlate positively with the level of risk of developing breast cancer [bib_ref] p63 expression profiles in human normal and tumor tissues, Como [/bib_ref]. In our series, we found higher ER and Ki-67 immunoexpression in HUT areas compared to the immunopositivity for these markers in adjacent normal lobules. Our results are similar to those described by [bib_ref] Multistep progression from an oestrogen-dependent growth towards an autonomous growth in breast..., Schmitt [/bib_ref] [bib_ref] p63 expression profiles in human normal and tumor tissues, Como [/bib_ref] who found the existence of a positive correlation between ER status and proliferation in hyperplasic epithelium and a progressive inversion of this relationship in lesions evolving towards malignancy. The observation of higher rates of proliferation in ER positive benign proliferative breast lesions fits with the concept of an initial hormone-dependent status in breast carcinogenesis [bib_ref] p63 expression profiles in human normal and tumor tissues, Como [/bib_ref]. In addition, HUT with higher expression of ER and Ki-67 could represent a subset of hyperplastic lesions with increased risk of subsequent breast cancer development.
Some previous studies suggest that at least some HUTs are clonal [bib_ref] Dissociation between steroid receptor expression and cell proliferation in the human breast, Clarke [/bib_ref] [bib_ref] Breast cancer risk associated with estrogen receptor expression in epithelial hyperplasia lacking..., Gobbi [/bib_ref]. Nevertheless, it remains unclear whether the dysregulation of ER has arisen prior to clonal expansion of the HUT, since cells with apparently abnormal regulation of ER during cell division are scattered randomly throughout the HUT in a non-contiguous pattern in some cases. The variable number of ER+ cells might indicate that in HUT the dysregulation of ER expression is incomplete and may be absent in some lesions, or apparent under certain conditions.
We detected p53 and p21 positivity in neoplastic cells of IMC, especially in high grade carcinomas. Mutations in tumor suppressor gene TP53, which mediates G1 arrest and apoptosis leads to an increased half-life and accumulation of the p53 protein [bib_ref] (waf) correlates with DNA replication but not with prognosis in invasive breast..., Gohring [/bib_ref]. A way to investigate the functional status of TP53 is to evaluate some of its downstream effectors such as p21 gene whose product acts by blocking cyclin-dependent kinases [bib_ref] Immunohistochemical staining for transforming growth factor beta 1 associates with disease progression..., Gorsch [/bib_ref]. In our study, we found a positive association between expression of p21 [bib_ref] Subgroups of non-atypical hyperplasia of breast defined by proliferation of oestrogen receptor-positive..., Iqbal [/bib_ref].
## And ki-67 in hut areas of reduction aesthetic mammaplasty specimens (hut-arm), hut of mammaplasty contralateral to breast cancer (hut-mcc); hut adjacent to invasive mammary carcinoma (hut-imc), and in invasive mammary carcinoma (imc)
## Antibody
## Hut-arm n (%) hut-mcc n (%) hut-imc n (%) imc n (%)
High expression of p21 would result in decreased cell proliferation subsequent to inhibition of cyclin/CDK activity [bib_ref] Estrogen receptor expression in benign breast epithelium and breast cancer risk, Khan [/bib_ref]. However, in our study p21 expression in neoplastic cells was related to higher proliferative index. Our results are in agreement with the theory of p53 independent pathways of p21 regulation in breast cancer [bib_ref] Molecular and biologic markers of premalignant lesions of human breast, Krishnamurthy [/bib_ref] [bib_ref] Atypical ductal hyperplasia of the breast: clonal proliferation with loss of heterozygosity..., Lakhani [/bib_ref]. High amounts of p21 in high proliferating cells may reflect an unsuccessful effort to halt proliferation. This may result from the presence of other cell cycle regulatory pathways, which bypass the p21 mediated cell cycle block, such as c-Myc or B-myb [bib_ref] Immunohistochemical staining for transforming growth factor beta 1 associates with disease progression..., Gorsch [/bib_ref] or due to mutant non-functional forms of p21 which posses prolonged half lives [bib_ref] Prediction of BRCA1 status in patients with breast cancer using estrogen receptor..., Lakhani [/bib_ref]. In addition, p21 expression can indeed be up-regulated by epidermal growth factor receptor and transforming growth factor β1 [bib_ref] Immunohistochemical staining for transforming growth factor beta 1 associates with disease progression..., Gorsch [/bib_ref] which are associated with higher tumor grade and disease progression in breast carcinoma [bib_ref] A prospective study of benign breast disease and the risk of breast..., London [/bib_ref] [bib_ref] Risk of breast cancer associated with atypical hyperplasia of lobular and ductal..., Marshall [/bib_ref].
In our study, cytoplasmic expression of p21 was found in 17.6% of benign hyperplastic cells and in 23.5% of neoplastic cells of IMC. Previous studies have reported exclusive nuclear localization of p21 in neoplastic cells of breast carcinomas [bib_ref] Subgroups of non-atypical hyperplasia of breast defined by proliferation of oestrogen receptor-positive..., Iqbal [/bib_ref] [bib_ref] Prediction of BRCA1 status in patients with breast cancer using estrogen receptor..., Lakhani [/bib_ref] , and in epithelial cells of HUT [bib_ref] Prognostic value of p21(WAF1) and p53 expression in breast carcinoma: an immunohistochemical..., Caffo [/bib_ref] However, other authors reported p21 immunoexpression in the cytoplasm of breast and ovarian tumors and it was considered critical for promoting cell transformation [bib_ref] Determinação da fraçãode proliferação celular no câncer de mama pela marcação imunoistoquímica..., Melo [/bib_ref] [bib_ref] Expression of proliferation and apoptosis-related proteins in usual ductal hyperplasia of the..., Mommers [/bib_ref]. There is no other data in current literature concerning cytoplasmatic p21 expression in HUT similar to our findings. It remains unclear how the elevated cytoplasmic p21 expression might contribute to tumorigenesis. One possibility is that p21 is sequestered away from the nucleus thereby preventing it from binding to nuclear cyclin/CDK complexes, thus allowing sufficient cyclin/ CDK activity for cell cycle progression [bib_ref] Adhesion modulation by antiadhesive molecules of the extracellular matrix, Orend [/bib_ref]. Alternately, relocalization of p21 to the cytoplasm may target cytoplasmic molecules such as apoptosis signal-regulating kinase 1 (ASK1) thereby promoting cell survival [bib_ref] Kosma VM: p21WAF1 expression in invasive breast cancer and its association with..., Pellikainen [/bib_ref].
In our study, p63 was exclusively expressed in myoepithelial cells of normal breast lobules and ducts, partially expressed around the HUT cells, and rarely expressed in invasive breast carcinoma. We observed that p63 staining was discontinuous in 38.1% in HUT-ARM, in 73.3% in HUT-MCC, and in 64.7% in HUT-IMC. The discontinuous p63 expression pattern in HUT was different from continuous expression in normal lobules and could suggest that there is loss of p63 expression in the progression to invasive carcinoma. Our data is similar to the findings of [bib_ref] p63 expression in normal, hyperplastic and malignant breast tissues, Wang [/bib_ref] [bib_ref] Prognostic value of cell cycle regulator molecules in surgically resected stage I..., Peters [/bib_ref] that demonstrate non-continuous expression of p63 in usual ductal hyperplasia. P63 expression has been useful to differentiate DCIS from microinvasive and invasive carcinomas based on lack of myoepithelial cells in invasive tumors without continuous distribution [bib_ref] Taking advantage of basic research: p63 is a reliable myoepithelial and stem..., Reis-Filho [/bib_ref] [bib_ref] p21WAF1/Cip1 is associated with cyclin D1CCND1 expression and tubular differentiation but is..., Rey [/bib_ref]. Although p63 is the most specific marker for myoepithelial cells, limitations exist because discontinuous myoepithelial layer seen in benign lesions, such as in our study may potentially cause diagnostic problems in clinical practice [bib_ref] Multistep progression from an oestrogen-dependent growth towards an autonomous growth in breast..., Schmitt [/bib_ref].
Although our data and genetic studies failed to demonstrate molecular changes in HUT, that are present in columnar cell lesions, ADH cells and in neoplastic cells of DCIS and IMC [bib_ref] Estrogen receptor-positive proliferating cells in the normal and precancerous breast, Shoker [/bib_ref] the argument that HUT may be an early precursor is still supported by consistent data from epidemiological studies [bib_ref] Prognostic and predictive factors in breast cancer by immunohistochemical analysis, Allred [/bib_ref] [bib_ref] Prognostic significance of TP53 alterations in breast carcinoma, Andersen [/bib_ref] [bib_ref] Apoptosis inhibitory activity of cytoplasmic p21(Cip1/WAF1) in monocytic differentiation, Asada [/bib_ref] [bib_ref] Clinical relevance of immunohistochemical expression of p53-targeted gene products mdm-2, p21 and..., Bankfalvi [/bib_ref] [bib_ref] Molecular evolution of breast cancer, Simpson [/bib_ref].
# Conclusion
Our findings and previously published data [bib_ref] Estrogen receptor-positive proliferating cells in the normal and precancerous breast, Shoker [/bib_ref] [bib_ref] Molecular evolution of breast cancer, Simpson [/bib_ref] demonstrate that the imunoprofile of HUT is different from other accepted precursor lesions, since they are composed of a mixed population of cells types with variable proportions of cell-cycle related protein expression, and some alterations could be present in the latest stages of breast cancer development.
# Methods
We selected slides and formalin-fixed, paraffin-embedded blocks from 83 female mammary specimens examined in the Breast Pathology Laboratory of Hospital das Clínicas of Federal University of Minas Gerais received from 1996 to 2004. The specimens selected were 34 specimens of aesthetic reduction mammaplasty (ARM), 15 specimens of mammaplasty contralateral to breast cancer (MCC), and 34 specimens of invasive mammary carcinomas (IMC) presenting HUT in the adjacent parenchyma. The aesthetic reduction mammaplasty was indicated only for cosmetic reasons or for back pain related to hypertrophic breast. There was no clinic or mammography alteration in the breasts of these patients. The mammaplasties contralateral to breast cancer were indicated in order to obtain an aesthetic balance and equilibrium related to the contralateral lumpectomy or mastectomy indicated because of breast cancer. Clinical and pathological data were obtained from the Breast Pathology Laboratory and hospital files. Clinical features evaluated were age, and menopausal status. Slides were reviewed by two observers and criteria used to classify HUT were those from Page & Anderson (1987)and the terminology adopted by the WHO classification [bib_ref] Edgerton SM: p(21WAF1/ CIP1) expression in breast cancers: associations with p53 and..., Thor [/bib_ref]. We performed immunostainings using monoclonal antibodies (summarized in [fig_ref] Table 1: Primary antibodies, dilutions, and sources of antibodies used in immunohistochemical study [/fig_ref] and the streptavidin-biotin method (Biogenex, USA) with previous heat-induced epitope retrieval. Immunoreactivity for ER, PR, p53, p21 WAF1 , p63, and Ki-67 was evaluated in HUT of ARM specimens (HUT-ARM), in HUT of MCC specimens (HUT-MCC), in HUT adjacent to invasive mammary carcinomas (HUT-IMC), and in neoplastic cells of IMC specimens.
Only nuclear staining was considered in the evaluation of ER, PR, p53, p63, and KI-67. For p21 WAF1 , both nuclear and cytoplasmatic staining were considered positive [bib_ref] Benign breast disease and breast cancer risk: morphology and beyond, Schnitt [/bib_ref].
We also evaluated ER expression in normal lobules of ARM and MCC specimens, and in adjacent normal lobules of IMC. Two to four lobular units adjacent to HUT areas were evaluated in each case.
Cases were classified as ER, PR, and p53 positive when more than 10% of cells exhibited positive nuclear staining [bib_ref] Increased level of p21 in human ovarian tumors is associated with increased..., Barboule [/bib_ref] [bib_ref] p63 expression in normal, hyperplastic and malignant breast tissues, Wang [/bib_ref] [bib_ref] Cytoplasmic p21WAF1/CIP1 expression is correlated with HER-2/neu in breast cancer and is..., Winters [/bib_ref]. The Ki-67 labeling index was obtained by the percentage of neoplastic and HUT cells showing nuclear staining. The tumors and HUT were grouped in three categories: < 10%, low proliferative index; 10-25%, intermediate proliferative index; and > 25%, high proliferative index [bib_ref] On the shoulders of giants: p63, p73 and the rise of p53, Yang [/bib_ref]. We considered p63 positive cases when at least 10% of myoepithelial cells exhibited positive nuclear staining [bib_ref] p21WAF1/Cip1 is associated with cyclin D1CCND1 expression and tubular differentiation but is..., Rey [/bib_ref]. Cases were classified as p21 WAF1 positive when more than 2% of cells exhibited positive nuclear staining [bib_ref] Benign breast disease and breast cancer risk: morphology and beyond, Schnitt [/bib_ref].
## Abbreviations
# Authors' contributions
LSAT: obtained the samples, carried out the histopathological and immunohistochemical analysis and wrote the first draft of manuscript. GFSR: carried out the histology and immunohistochemistry. HG: conceived and designed the study, confirmed the histopathological and immunohistochemical analysis and provided expert input for writing and supervised the study.
All authors have read and approved the final manuscript.
[table] Table 1: Primary antibodies, dilutions, and sources of antibodies used in immunohistochemical study [/table]
[table] Table 3: Positivity for ER, PR, p53, p21, p63, [/table]
|
Spontaneous Rupture in a Non-Laboring Uterus at 20 Weeks: A Case Report
Objective:Unusual clinical course Background:Uterine rupture is uncommon but when it happens, it can cause significant morbidity and mortality to both mother and fetus. Incidence reportedly is higher in scarred than in unscarred uteri. Most cases occur in laboring women in their third trimester with a previous history of uterine surgery, such as caesarean delivery or myomectomy. We present a case of spontaneous uterine rupture in a non-laboring uterus in the mid-trimester of pregnancy. Case Report:The patient presented with threatened miscarriage at 17 weeks' gestation and ultrasound findings were that raised suspicion of a morbidly adherent placenta. Her history was significant for two previous cesarean deliveries more than 5 years ago followed by two spontaneous complete miscarriages in the first trimester. The patient was managed conservatively until 20 weeks' gestation, when she presented with acute abdomen with hypotensive shock. Her hemoglobin dropped to a level such that she required blood transfusion. An emergency exploratory laparotomy was performed, which revealed a 5-cm rupture in the lower part of the anterior wall of the uterus, out of which there was extrusion of part of the placenta. Given the patient's massive bleeding, the decision was made to proceed with subtotal hysterectomy. Histopathology of the specimen confirmed the diagnosis of placenta percreta.Conclusions:Identification of uterine scarring with morbidly adherent placenta is crucial because even in early pregnancy, it can lead to uterine rupture. Furthermore, failure to recognize and promptly manage uterine rupture may prove fatal.
# Background
Although uncommon, uterine rupture is a serious obstetrics complication that can cause significant maternal and fetal morbidity and mortality. A population-based cohort study in the Netherlands found that incidence of uterine rupture in scarred uteri was 5.1 per 10 000 births [bib_ref] Uterine rupture in The Netherlands: A nationwide population-based cohort study, Zwart [/bib_ref]. A prospective study in India revealed that the overall incidence of uterine rupture was 0.35%, and there, women without and with scarred uteri have 1.5 and 1.7 times higher risks for rupture, respectively, than women in other developed countries [bib_ref] Uterine eupture: Still a harsh reality, Singh [/bib_ref]. Uterine rupture can cause massive hemorrhage and extensive damage to the uterus, leading to hysterectomy when the uterus is beyond repair. It can lead to significant disruption of blood, and hence, oxygen supply to the fetus, which can cause temporary or permanent hypoxic injury to the fetus, and in worst scenario, fetal death. Most of the cases of uterine rupture described in previous literatures have involved women with a history of previous uterine surgery, such as cesarean delivery or myomectomy, usually occurred in the third trimester during labor. Uterine rupture without onset of labor is an unusual presentation, as persistent uterine contraction is what typically leads to scar dehiscence and the subsequent rupture. We present a case of spontaneous uterine rupture in a nonlaboring uterus in the mid-trimester of pregnancy.
## Case report
A 31-year-old woman in her fifth pregnancy presented at 17 weeks' gestation with painless per vaginal bleeding. She had experienced intermittent staining per vagina since early pregnancy, which was diagnosed as threatened miscarriage. The patient had two previous caesarean deliveries both for major placenta previa more than 5 years prior to this pregnancy, and two uncomplicated first-trimester spontaneous miscarriages. Examination revealed a soft abdomen with an 18-week size, non-irritable uterus. Transabdominal ultrasonography showed a viable fetus with fetal growth parameters equivalent to 18 weeks' gestation. The placenta was low-lying and there was no distinct plane of demarcation between the placenta and the uterine wall, with some placental lakes noted with high probability of adherent placenta. The patient was counselled about the need for regular ultrasound for surveillance and magnetic resonance imaging (MRI) in the third trimester was planned.
At 20 weeks' gestation, the woman presented to the district hospital with sudden onset of severe lower abdominal pain associated with vomiting. She had no vaginal bleeding but she was hemodynamically unstable and needed prompt fluid resuscitation. Her hemoglobin level dropped to 8.3 g/dL and she was transfused with 1 pint of packed red blood cells (PRBCs) and immediately transferred to the tertiary hospital. Upon arrival, the patient was in severe pain, pale, tachycardic, and hypotensive. Her abdomen was soft with a tender uterus of approximately 20-week size. Ultrasonography revealed a viable 20-week fetus with anterior placenta, which was still lowlying, and presence of placental lakes posterior to the bladder. There was free fluid seen at Morrison's pouch. Repeat blood testing showed that the patient's hemoglobin level had dropped further, to 6.6 g/dL.
A diagnosis of intra-abdominal hemorrhage likely due to uterine rupture was made, therefore, emergency exploratory laparotomy was arranged. Intraoperatively, there was a 5-cm rupture in the lower part of the anterior uterine wall and out of which part of the placenta was extruding, with ongoing active bleeding. The decision was made to perform a subtotal hysterectomy with the fetus in situ [fig_ref] Figure 1: Post subtotal hysterectomy specimen showing uterine rupture with intact amniotic sac and... [/fig_ref]. The patient received 7 pints of PRBCs and one cycle of a regimen for disseminated intravascular coagulopathy (DIVC). She recovered well postoperatively and was discharged on Day 5.
Histopathology of the specimen showed placental tissue adherent to the uterine wall with an intervening layer of decidua. There was evidence of chorionic villi infiltrating the myometrium with no layer of intervening decidua in some focal areas. The lower uterine wall and endocervix was thinned out and showed areas of chorionic villi perforating the muscular wall, with focal area of necrosis and hematoma on the placenta. These findings were consistent with placenta percreta.
The patient was followed up in the outpatient clinic 2 months following discharge and reported no complications. She fully understood the diagnosis and its implication.
# Discussion
Uterine rupture during pregnancy is rare and can occur in women with native, unscarred uteri or in uteri with surgical scars from previous surgery. Incidence of uterine rupture in scarred uteri is higher than in unscarred uteri, as supported by studies from other countries such as the Netherlands and India [bib_ref] Uterine rupture in The Netherlands: A nationwide population-based cohort study, Zwart [/bib_ref] [bib_ref] Uterine eupture: Still a harsh reality, Singh [/bib_ref]. In unscarred uteri, the incidence of uterine rupture was found to be 1 in 2770 deliveries [bib_ref] Unscarred uterine rupture and subsequent pregnancy outcome -a tertiary centre experience, Peker [/bib_ref]. A Swedish population-based cohort study showed that women who had previous uterine scar via cesarean delivery were at increased risk of uterine rupture in their second deliveries, compared to those who had previous vaginal deliveries [bib_ref] Risk factors for uterine rupture and neonatal consequences of uterine rupture: A..., Kaczmarczyk [/bib_ref]. Other risk factors include previous myomectomy scar, particularly one that has breached the uterine cavity. The risk is 0.6% to 0.75% in pregnancies after myomectomy [bib_ref] Uterine rupture in pregnancies following myomectomy: A multicenter case series, Kim [/bib_ref]. On the other hand, uterine rupture in unscarred uteri typically occurs in the second stage of labor due to mismanaged labor, injudicious use of oxytocin, obstructed labor, or use of instrumental delivery [bib_ref] Unscarred uterine rupture: A retrospective analysis, Vernekar [/bib_ref].
Although uterine rupture is commonly associated with the laboring uterus, it has also been reported in the non-laboring uterus. Uterine rupture in the non-laboring uterus without external trauma is rare but cornual pregnancy and multiparity are two reported risk factors [bib_ref] Rupture of the pregnant uterus: A review, Schrinsky [/bib_ref]. A case has been reported of a grand multipara with spontaneous rupture at 30 weeks in a scarred uterus [bib_ref] Preterm spontaneous uterine rupture in a non-labouring grand multipara: A case report, Albrecht [/bib_ref]. Similarly, a case of spontaneous uterine rupture in a grand multipara with one cesarean delivery followed by four vaginal deliveries also has been reported [bib_ref] Extrusion of fetus into the abdominal cavity following complete rupture of uterus:..., Segal [/bib_ref]. In our case, the patient only had two pregnancies beyond 24 weeks, which should have made her less predisposed to uterine rupture. Both those pregnancies were delivered via cesarean. However, in a prospective, multicenter, observational study, no significant difference was found in the rate of uterine rupture between patients with a single cesarean delivery versus those with multiple such deliveries [bib_ref] Risk of uterine rupture with a trial of labor in women with..., Landon [/bib_ref].
Uterine rupture usually occurs in the third trimester. In the first and second trimester, it is rare and the diagnosis is often made intraoperatively [bib_ref] Uterine rupture in first or second trimester of pregnancy after in-vitro fertilization..., Arbab [/bib_ref]. Most patients present with abdominal pain, vaginal bleeding, and hypotension [bib_ref] Uterine rupture during pregnancy and delivery: Risk factors, symptoms and maternal and..., Andonovova [/bib_ref]. In this case, the patient presented at 17 weeks' gestation, 3 weeks before the event with vaginal bleeding that was trivial and required only conservative measures. There was a suspicion of morbidly adherent low-lying placenta, but the incidence of uterine rupture due to placenta percreta is 1 in 5000 pregnant women, which is rare [bib_ref] Uterine rupture in pregnancy reviewed, Gardeil [/bib_ref].
Recent evidence indicates that ultrasonography at 12 to 16 weeks' gestation can accurately predict morbidly adherent placenta [bib_ref] Screening for morbidly adherent placenta in early pregnancy, Panaiotova [/bib_ref]. The patient in this case had an ultrasound suspicious for morbidly adherent placenta when she first presented at 17 weeks' gestation. There was loss of uterine wall-bladder demarcation and presence of placental lakes. Other ultrasound features suggestive morbidly adherent placenta include nonvisible cesarean scar, thin retroplacental myometrium, bladder wall interruption, and presence of intra-placental lacunar spaces. Other than that, three-dimensional power Doppler can demonstrate presence of retroplacental arterial-trophoblastic blood flow and irregular placental vascularization [bib_ref] Screening for morbidly adherent placenta in early pregnancy, Panaiotova [/bib_ref]. There is a suggested scoring module for prediction of intrapartum morbidly adherent placenta and maternal morbidity which assesses these features: placenta lacunae, retroplacental echo lucent space, retroplacental myometrium thickness, hyperechoic uterine-bladder interface, and vascularity of subplacental, uterine serosa-bladder wall, intra-placental and bladder wall [bib_ref] An ultrasound scoring model for the prediction of intrapartum morbidly adherent placenta..., El-Haieg [/bib_ref]. MRI plays a vital role in evaluation of inconclusive cases by ultrasonography. The sensitivity and specificity of MRI in evaluating the invasion topography in placenta percreta ranges from 87.5% to 100%. For prediction of parametrial, bladder, and cervical invasion, MRI has 100% specificity [bib_ref] Placenta percreta evaluated by MRI: Correlation with maternal morbidity, Chen [/bib_ref]. Therefore, MRI has an important role in confirmation of morbidly adherent placenta to enable precise mapping of placental abnormalities and aid multidisciplinary planning and management [bib_ref] MRI evaluation of the placenta from normal variants to abnormalities of implantation..., Zaghal [/bib_ref]. There are limited reports with regard to use of MRI in mid-trimester pregnancy for diagnosing this condition. However, a morbidly adherent placenta during mid-trimester can be diagnosed by an MRI finding of low-lying, inhomogeneous placenta, dysplastic vascular hypertrophy, ill-defined placental bands, and an overall impression of some areas of increta with no overt evidence of percreta [bib_ref] Conservative management of abnormally invasive placenta previa after midtrimester foetal demise, Macgibbon [/bib_ref].
Uterine rupture caused by placenta percreta can be more lifethreatening than that caused by a previous scar because placenta percreta-induced uterine rupture exhibits more vascularization than the site of previous scar-induced rupture [bib_ref] Clinical risk factors for placenta previaplacenta accreta, Miller [/bib_ref].
Morbidly adherent placenta, like increta and percreta, typically develops due to dehiscence of a previous uterine scar, thus facilitating better access by cells from the trophoblast column to the large outer myometrial vessels [bib_ref] Pathophysiology of placenta creta: The role of decidua and extravillous trophoblast, Tantbirojn [/bib_ref]. This explains the intraoperative finding in the patient in this case. That is, she had massive bleeding from the percreta-induced uterine rupture site, which led to the need for hysterectomy.
A retrospective analysis reported that subtotal hysterectomy is the most common surgical intervention for management of uterine rupture (73.6%) [bib_ref] Uterine rupture: A retrospective analysis of causes, complications and management outcomes at..., Kidantou [/bib_ref]. The procedure was performed in this case and it has been recommended as the best choice of surgical intervention for uterine rupture. With it, hemostatic control is faster, blood loss and need for blood transfusions reduced, there are fewer perioperative complications, and operating time is reduced [bib_ref] FIGO consensus guidelines on placenta accreta spectrum disorders: Nonconservative surgical management, Allen [/bib_ref]. However, subtotal hysterectomy has not been shown to provide protection against urinary tract injury [bib_ref] Approaches to reduce urinary tract injury during management of placenta accreta, increta,..., Tam [/bib_ref]. In uterine rupture due to underlying morbidly adherent placenta that extends into the cervix, total hysterectomy is superior to subtotal hysterectomy particularly to ensure better hemostatic control [bib_ref] FIGO consensus guidelines on placenta accreta spectrum disorders: Nonconservative surgical management, Allen [/bib_ref]. Nonetheless, subtotal hysterectomy was effective in the case presented here because there was no cervical involvement.
# Conclusions
The current trend of rising cesarean delivery globally inevitably had led to an increased incidence of both placenta previa and morbidly adherent placenta. Thus, it is crucial for clinicians to screen all pregnancies with previous uterine scar for morbidly adherent placenta during routine second-trimester ultrasonography. Early identification allows the provider time to prepare and plan for the high-risk nature of the delivery. Mid-trimester uterine rupture is a rare event; however, a high index of suspicion is crucial to allow early recognition and timely management in these cases, particularly with an underlying morbidly adherent placenta, so as to prevent a catastrophic and fatal outcome.
[fig] Figure 1: Post subtotal hysterectomy specimen showing uterine rupture with intact amniotic sac and fetus in situ. [/fig]
[fig] Figure 2: Stillbirth protruding out after iatrogenic rupture of the amniotic sac. [/fig]
|
What impact could DMPA use have had in South Africa and how might its continued use affect the future of the HIV epidemic?
Introduction: Some studies suggest that use of the injectable contraceptive depot medroxyprogesterone acetate (DMPA) may increase susceptibility to HIV infection. We aim to determine the influence that such an association could have had on the HIV epidemic in South Africa. Methods: We simulate the heterosexual adult HIV epidemic in South Africa using a compartmental model stratified by age, behavioural risk group, sex, male circumcision status and contraceptive use. We model two possible scenarios: (1) The "With Effect" scenario assumes that DMPA increases susceptibility to HIV infection by 1.20-fold (95% confidence interval 1.06 to 1.36) based on a combination of the results of a recent randomised controlled trial (ECHO trial) and a number of observational studies. (2) The "No Effect" scenario assumes that DMPA has no effect on HIV acquisition risk. We calculate the difference in HIV-related outcomes between the With Effect and No Effect scenarios to determine the potential impact that DMPA use could have had on the HIV epidemic.
# | introduction
Injectable hormonal contraception (IHC) is used by over 50 million women worldwide, the majority of whom are using depot medroxyprogesterone acetate (DMPA). It is particularly popular in southern and eastern Africa, for example in South Africa 17.7% of married and sexually active unmarried women use DMPA . However, some evidence suggests that there is a link between DMPA use and increased risk of acquiring HIV.
The available evidence consists of observational studies that have re-analysed data that was initially collected for another purpose and so most studies contain some methodological limitations. A recent meta-analysis including 14 studies with the fewest methodological limitations suggests that DMPA may increase the risk of HIV acquisition by a factor of 1.4 (95% confidence interval 1.23 to 1.59). Conversely, collective evidence suggests that oral contraceptive pills, injectable norethisterone enanthate (NET-EN) and implants (for which limited data is available) do not affect HIV acquisition risk.
A number of animal and ex-vivo studies have investigated possible biological mechanisms that would explain this association. There is evidence to suggest that the progestin medroxyprogesterone acetate (MPA) has been shown to inhibit parts of the immune system, and furthermore MPA may increase the frequency of target cells for HIV in the genital tract. These immunomodulatory effects could allow the virus to more easily infect target cells and avoid detection and removal by the immune system. MPA has also been shown to make the female reproductive tract more permeable, potentially allowing HIV to move through the lining of the reproductive tract more easily.
The World Health Organisation has updated its medical eligibility criteria for injectable contraceptives, including both DMPA and NET-EN, in response to this evidence from a category 1which states that there should be "no restrictions on the use of the contraceptive method"to a category 2which generally indicates that "the advantages of using the contraceptive method generally outweigh the theoretical or proven risks, " and advises that women at risk of HIV infection wishing to use injectable contraceptives should be counselled as to its potential side effects.
Based on the uncertainty surrounding the observational evidence, the Evidence for Contraceptive Options and HIV Outcomes (ECHO) trial, randomized 7829 HIV-negative women to either DMPA, a copper intrauterine device (IUD) or a levonorgestrel (LNG) implant and followed them for up to 18 months throughout the trial, testing for HIV seroconversion every three months. The trial results have recently been released and showed no significant association between use of any of the contraceptive methods and HIV acquisition risk.
The results of the ECHO trial reduce the likelihood that DMPA use increases HIV risk, particularly for a "large" increase in risk of 1.5-foldthe effect size that the trial was powered to detect. However, it does not conclusively disprove that a causal association exists. It is important to consider all available evidence, and so we have updated our understanding of the likelihood that a causal association exists and the effect size of such an association. An increase in risk by a factor of between 1 and 1.5 may still be a substantial increase in risk, particularly in areas of high HIV incidence.
To understand the impact that DMPA may have had on the HIV epidemic so far, and the potential future impact, it is important to consider the context of a complex epidemic including men and women of different ages and sexual behaviours. The impact would be greatest in those countries that have had historically high rates of DMPA use, as well as high HIV prevalence. DMPA has been available in South Africa since 1963, and has been widely used for more than 30 years. Furthermore, South Africa has the largest number of people living with HIV of any country worldwide, with adult prevalence of more than 15% since 2002. Therefore, a possible effect of DMPA use on HIV risk, even if the effect size is lower than the 1.5-fold increase that the ECHO trial was powered to detect, could have particularly large implications in South Africa. To estimate the possible effects of such a causal association, we perform a mathematical modelling analysis comparing scenarios in which DMPA increases HIV susceptibility to scenarios in which there is no effect of DMPA use on HIV risk.
# | methods
We use a deterministic compartmental mathematical model of heterosexual HIV transmission in South Africa. The population is stratified by age, behavioural risk, sex, male circumcision status and contraceptive use. The population age structure is based on previous model projections of the South African population, based on single year age groups. The HIV-infected population is divided into nine compartments based on stage of infection, CD4 count and treatment status. HIV-infected individuals in the model are able to begin antiretroviral therapy (ART) at different stages of infection. The population is stratified into three behavioural risk groups, which have different rates of partner change, frequencies of sex acts and condom use. Rates of HIV transmission are dependent on the risk group of each partner, the treatment status of the HIV positive partner, the circumcision status of the HIV negative male partner, the infection stage of the HIV positive partner, and the contraceptive use of the female partner. For a full description of the HIV transmission within the model, see the Supporting Information. We calibrate our model using data on age and sex-specific HIV prevalence, total HIV incidence and incidence in high risk women (see Supporting Information). A number of parameters were fitted including those related to the per sex act probability of HIV transmission, rates of mixing between different behavioural risk groups, the size of these risk groups and the start time of the epidemic.
Injectable contraceptives have been widely used in South Africa for the majority of the HIV epidemic, the 1987 to 1988 DHS reported that 19.6% of married women were using injectable methods. Estimates of contraceptive use between different time points are often incompatible due to differences in methodologies between surveys; here we calibrate contraceptive use in the model to data from the 2016 Demographic and Health Survey (DHS) [2] and assume constant contraceptive total prevalence (including the prevalence of DMPA use) and age-specific prevalence in South Africa over the period of observation (1980 to 2037). We believe that these assumptions are acceptable since injectables have been widely used throughout periods of high HIV prevalence in South Africa. We also believe that a substantially different pattern of age-specific prevalence is unlikely.
We assume that scale-up of male circumcision continues such that 60% of men aged 15 to 49 are circumcised by 2037 and ART scale-up continues such that the 90-90-90 targets are reached by 2032. We ran the model 20,000 times and used a filtration method to select 100 acceptable epidemic fits. We then perform 1000 pairwise runs of the model, sampling from the set of fitted epidemic parameters each time. For each sampled parameter set, the model is run twice; in one run we assume that DMPA increases HIV susceptibility by a factor randomly sampled from a log-normal hazard ratio (HR) distribution ("With Effect" scenario) and in the other we assume a hazard ratio of 1 and hence no effect of DMPA use on HIV risk ("No Effect" scenario). Due to the uncertainty surrounding the true association between DMPA and HIV risk, in the With Effect scenario we use a HR distribution with a median of 1.20 (2.5th to 97.5th percentiles 1.06 to 1.36) This distribution is based on a combination of the recent ECHO trial results, which report a HRs for DMPA use of 1.04 and 1.23 compared with the copper IUD and LNG implant respectively, and the most recent meta-analysis of observational studies which found a mean HR of 1.40 for DMPA use. In using this distribution we acknowledge that the ECHO trial suggests that there is no substantial causal association between DMPA use and HIV risk. However, we also incorporate the body of observational evidence that suggests that such an association does exist. The ECHO trial was powered to detect a relative increase in risk of 1.5and therefore the HR distribution represents a plausible range of values that may not have been detected by the ECHO trial.
For each parameter set we estimate two different epidemic curves and hence we are able to estimate the impact of DMPA use on the HIV epidemic in South Africa, under the assumption given by the current body of evidence. In this way, our credible range represents uncertainty due to the epidemic fit and the effect of DMPA use on the risk of HIV acquisition.
Full details of the parameter values used, model specification and calibration can be found in the Supporting Information.
3 | RESULTSshows that for simulations in which DMPA increases HIV susceptibility, the epidemic is larger relative to the No Effect scenario. We observe a median excess 430,000 (90% of model runs 160,000 to 960,000) infections from 1980 to 2017, which represents 4.3% (1.6% to 9.6%) of the total HIV infections in this period. Of these, we predict that on average, 275,000 occur in women and 155,000 occur in men. The fraction of excess infections that occur in men is 36% (90% of model runs 25% to 48%). The extra infections caused by the potential increase in HIV susceptibility would necessitate a median estimate of 640,000 (190,000 to 1,660,000) extra years of ART from 1980 to 2017.
Furthermore, we see an increase in AIDS deaths in the With Effect scenarioand estimate that a causal association between DMPA and HIV risk would have caused a median 230,000 additional deaths from AIDS (90,000 to 470,000) to date. This represents 6.9% (2.6% to 15.2%) of the AIDS deaths in this period.
Projecting forward, from 2018 to 2037 we predict an excess 130,000 (90% of model runs 50,000 to 270,000) infections between 2018 and 2037 as well as 60,000 (20,000 to 150,000) additional deaths. As a result of the excess HIV infections an extra 2,870,000 (890,000 to 7,270,000) years of ART would be required from 2018 to 2037.
# | discussion
The use of DMPA may have had a profound impact in exacerbating the spread of HIV in South Africa. Its use could be responsible for 4.3% of infections to date. The impact would most heavily affect the female users of DMPA themselves, but additional infections would also occur in male partners of DMPA users and in the wider population.
The major limitation of this analysis is the uncertainty on the true association between DMPA use and HIV susceptibility. We have therefore sampled HRs from a distribution, to incorporate the significant body of observational evidence, suggesting a mean HR of 1.4, and the ECHO trial, which produced HRs of 1.04 for DMPA compared to the copper IUD and 1.23 for DMPA compared with the LNG implant. Whilst the ECHO trial provides the best evidence of the available studies, the trial was powered to detect a hazard ratio of 1.5, and so it is still possible that DMPA does increase HIV risk, but this was not detected by the trial. This becomes more likely for HRs, such as 1.2 (the median of the HR used in this modelling analysis) which would still represent a substantial increase in risk in high-incidence areas. We therefore believe that the HR distribution used in this study represents credible uncertainty surrounding the possible causal association of DMPA use on HIV acquisition risk.
There is also some uncertainty regarding scale up of contraception. DMPA was introduced in South Africa in the 1960sand, in a 1977 to 1978 national survey injectable contraceptives (including DMPA) were reported to be used by 19.6% of women married or in a union. Due to incompatibility between different contraceptive prevalence estimates, we assume constant contraceptive use throughout the HIV epidemic. This is a limitation of our analysis, but since DMPA has been widely used throughout the HIV epidemic in South Africa, changes in contraceptive use are unlikely to be large, and so we would not expect a large impact on the results of our analysis.
Our model assumes 17.7% of women use DMPA, based on estimates of contraceptive prevalence from 2016 [2]. If DMPA use has increased during the course of the HIV epidemic, and there continues to be increased use in the future, then our results will overestimate the past impact of a possible effect of DMPA use on HIV risk, but will underestimate future impact. The contraceptive age-profile of the model is based on national survey estimates, DMPA use is highest for women in their 20s and 30s and lower for women under 20 and over 40 . If DMPA use was in fact most commonly used by women in their 40s, for whom HIV incidence is lower, then the impact of the putative effect of DMPA on the HIV epidemic would be lower. However, we believe it is unlikely that such a substantially different age-profile of DMPA use could occur.
While we have focussed on South Africa, this research has implications for many other countries. The impact of a potential causal association would vary between countries depending on a number of factors including, the pattern of DMPA use in the country, the size and timing of the HIV epidemic and the scale up of interventions. For example, in Zambia, there is a large HIV epidemic, with 11.5% of the adult population living with HIV in 2017. DMPA use is also high, with the most recent Demographic and Health Survey reporting that 17.9% of married women used the method in 2013 to 2014. However, DMPA use has largely been expanded in the 21st century, particularly since 2007. Therefore there would be a shorter period of overlap between DMPA use and the HIV epidemic in Zambia; the number of new HIV infections peaked in the mid-1990s and AIDS deaths had peaked by 2003, at which point between 4.5% to 8.5% of married women were using DMPA. Since much of the scale up of DMPA in Zambia happened after the peak of the HIV epidemic, we would expect a potential causal association between DMPA use and HIV risk to contribute a smaller fraction of HIV-related outcomes in Zambia, when compared with South Africa, where DMPA had been used at high levels since the late 1980s. We would predict that a causal association between DMPA use and HIV risk, if it exists, to be responsible for a larger fraction of HIV-related outcomes in countries where DMPA has been used extensively throughout the HIV epidemic. In countries across southern and eastern Africa, DMPA accounts for a large fraction of the contraceptive method mix and many of these countries have also had high HIV prevalence. It is likely that any effects of DMPA on HIV risk would be setting specific and depend on many factors, including the temporal overlap of HIV incidence and patterns of DMPA in particular age groups.
The use of a highly stratified mathematical model has a number of advantages. We account for age-specific differences in HIV prevalence as well as contraceptive differences by age. One limitation is that the model does not capture the formation and dissolution of partnerships and so may introduce bias by not accounting for specific sexual behaviour patterns in individuals.
# | conclusions
Discussion as to the relative benefits, and potential risks, of DMPA use are ongoing. The World Health Organization has made several updates to its guidelines for injectable contraceptive use for those at high risk of HIV. We aim to aid discussion by providing estimates of the impact that DMPA use may have had on the HIV epidemic as well as projections for future impact if use is continued and a real risk exists. Our results suggest that such an association between DMPA use and HIV risk could have increased the number of HIV infections by 430,000 (90% of model runs 160,000 to 960,000) which could have caused an additional 230,000 (90,000 to 470,000) AIDS deaths in South Africa between 1980 and 2017. We estimate that on average, 36% (25% to 48%) of infections occurred in men, showing that the impact of an association between DMPA use and HIV risk would affect not only women using DMPA, but also their partners and the wider population.
In South Africa, many women have limited access to safe healthcare facilities, and hence the maternal mortality ratio is approximately 8 to 10 times higher than those of high income countries. Hence a reduction in DMPA use, without transition to another equally effective method of contraception, could lead to an increase in unintended pregnancies, causing a range of negative consequences including maternal morbidity and mortality. DMPA remains a valuable contraceptive for women throughout the world, allowing women to control their fertility and avoid potential negative consequences, both health and socioeconomic, of unintended pregnancy. These results reiterate the need for governments and family planning service providers to establish broad contraceptive method mix, to allow women to safely manage both their HIV prevention and family planning needs.
## F u n d i n g
The work reported in this article was supported by a grant from the United States Agency for International Development (USAID) through an award subcontracted through Abt Associates. We acknowledge joint Centre funding from the UK Medical Research Council and Department for International Development. Funding was also received from Bill & Melinda Gates Foundation.
## Supporting information
Additional information may be found under the Supporting Information tab for this article.. Natural history of HIV infection and ART initiation as represented in the model. . The proportion of adult men that are circumcised with respect to time. The level of circumcision in the model was calibrated to data reported in a nationally representative surveyBeacroft L et al. . The percentage of HIV positive adults (15 to 49) receiving antiretroviral therapy in South Africa. Model data is compared to estimates of the percentage of HIV positive adults on ART in South Africa . . Total contraceptive prevalence calibration. Total contraceptive prevalence among 15 to 49 year old women is calibrated to nationally representative survey data . The percentage of HIV positive adults (15 to 49) receiving antiretroviral therapy in South Africa. Model data is compared to estimates of the percentage of HIV positive adults on ART in South Africa . . Age-specific contraceptive prevalence calibration. Contraceptive prevalence among different age groups are calibrated to nationally representative survey data . . Population pyramids for South Africa for . Model population structure is compared to annual age-structured population size model estimates produced by the Actuarial Society of South Africa. . Population size with respect to time. The total population of the model was calibrated to previous estimates from a demographic model of the South African population. . HIV incidence in 15 to 49 year olds. HIV incidence in adults was calibrated to incidence data from a nationally representative survey as well as incidence estimates produced by a mathematical model calibrated to prevalence data. Blue dotted lines represent 10th, 50th and 90th percentiles of model variation. . Prevalence in 15 to 49 year olds. Adult HIV prevalence is calibrated to nationally representative survey data from South Africa as well as UNAIDS prevalence estimates . Blue dotted lines represent 10th , 50th and 90th percentiles of model variation.. Male HIV prevalence. HIV prevalence in the model was calibrated to sex-specific prevalence data. Blue dotted lines represent 10th, 50th and 90th percentiles of model variation.. Female HIV prevalence. HIV prevalence in the model was calibrated to sex-specific prevalence estimates. Blue dotted lines represent 10th, 50th and 90th percentiles of model variation.. Male HIV incidence. HIV incidence in the model was calibrated to sex-specific incidence estimates produced by a mathematical model calibrated to nationally representative prevalence data. Blue dotted lines represent 10th, 50th and 90th percentiles of model variation.. Female HIV incidence. HIV incidence in the model was calibrated to sex-specific incidence estimates produced by a mathematical model calibrated to nationally representative prevalence data. Blue dotted lines represent 10th, 50th and 90th percentiles of model variation.. Natural history of infection parameters . Behavioural parameters and values . Factor increments in transmission probability per sex act with respect to baseline transmission probability (b 0 ) . Contraceptive efficacy and continuation rates for methods used in the model. |
Exploration of the relationship between biogas production and microbial community under high salinity conditions
High salinity frequently causes inhibition and even failure in anaerobic digestion. To explore the impact of increasing NaCl concentrations on biogas production, and reveal the microbial community variations in response to high salinity stress, the Illumina high-throughput sequencing technology was employed. The results showed that a NaCl concentration of 20 g/L (H group) exhibited a similar level of VFAs and specific CO 2 production rate with that in the blank group, thus indicating that the bacterial activity in acidogenesis might not be inhibited. However, the methanogenic activity in the H group was significantly affected compared with that in the blank group, causing a 42.2% decrease in CH 4 production, a 37.12% reduction in the specific CH 4 generation rate and a lower pH value. Illumina sequencing revealed that microbial communities between the blank and H groups were significantly different. Bacteroides, Clostridium and BA021 uncultured were the dominant species in the blank group while some halotolerant genera, such as Thermovirga, Soehngenia and Actinomyces, dominated and complemented the hydrolytic and acidogenetic abilities in the H group. Additionally, the most abundant archaeal species included Methanosaeta, Methanolinea, Methanospirillum and Methanoculleus in both groups, but hydrogenotrophic methanogens showed a lower resistance to high salinity than aceticlastic methanogens.An increasing amount of food waste (FW) is produced by the sorting, cooking, peeling and dining processes 1 . According to the statistics, more than 2000 ton of FW is generated every day in Beijing 2 , and the high organics, salinity and water contents of FW have caused serious environmental problems in modern societies 3 . Within different possible treatment routes, anaerobic digestion (AD) of FW into biogas is proven to be an effective solution for FW treatment 4 . Although the composition of FW is highly variable depending on the collection sources, it usually contains high salinity. Oh et al. reported that the NaCl-amended FW contained 10 to 35 g/L NaCl, while the non-washed FW contained 11.6 g/L NaCl 5 . Dai et al. collected FW from cafeterias in Shanghai with NaCl concentration of 8.0 g/L 6 . Wang et al. found that the NaCl concentration from FW anaerobic digestate could reach 13.8 g/L 7 .This high salinity could cause cell osmotic stress imbalance, resulting in plasmolysis and/or loss of activity of cells8,9, which would further cause inhibition and even failure of the AD process. For instance, Nagai et al. found that the utilization of FW from soy sauce was difficult because of its high salinity of 10% (w/w) despite its highly nutritious biomass 10 . Lee et al. studied the effect of salinity on biogas production from food waste leachate, and found that 0.5~2 g/L NaCl would increase the methane yield while 5 and 10 g/L NaCl resulted in a reduction of methane yield by 36% and 41%, respectively 11 . Rinzema et al. found that the formation of methane from acetate would be inhibited by 10, 50 and 100% respectively at Na + concentrations of 5, 10 and 14 g/L 12 .Research on how high salinity would affect the biogas production has been drawing increasing attention. It was reported that a Na + concentration ranging from 2 to 10 g/L would moderately inhibit the methanogenic activity, while a concentration exceeding 10 g/L would strongly inhibit methanogenesis 11, 13 . Lefebvre et al. reported that methanogenesis started to be affected at a NaCl concentration of 5 g/L while acidogenesis was severely affected Published: xx xx xxxx OPEN www.nature.com/scientificreports/ 2 Scientific RepoRts | 7: 1149 |
only at NaCl concentration exceeding 20 g/L 8 . However, detailed and intensive analysis is still scarce. Exploring the influence of high salinity on biogas production in the AD process is therefore of great significance.
Generally, the AD process includes four steps, namely, hydrolysis, acidogenesis, acetogenesis, and methanogenesis [bib_ref] Peracetic acid oxidation as an alternative pre-treatment for the anaerobic digestion of..., Appels [/bib_ref]. The first three steps are mediated by bacterial populations in which organic matter is converted to volatile fatty acids (VFAs) and further digested into acetate, H 2 and CO 2 . The last step is performed by the archaeal group that produces methane using the acidogenic products [bib_ref] Peracetic acid oxidation as an alternative pre-treatment for the anaerobic digestion of..., Appels [/bib_ref] [bib_ref] Qualitative and quantitative assessment of microbial community in batch anaerobic digestion of..., Shin [/bib_ref]. The microorganisms in the four steps affect the production of methane in different ways. It is therefore important to comprehensively understand the microbial behavior to fundamentally improve the efficiency of the AD process. Although some researchers have investigated the influence of the high salinity inhibition on microbial community structure [bib_ref] Impact of increasing NaCl concentrations on the performance and community composition of..., Lefebvre [/bib_ref] [bib_ref] Effect of elevated salt concentrations on the aerobic granular sludge process: linking..., Bassin [/bib_ref] , few findings provide an entire and in-depth analysis by correlation of biogas production with microbial community structure.
The microbial community of biogas production is commonly determined via construction of 16S rRNA clone libraries and Denaturing Gradient Gel Electrophoresis (DGGE) [bib_ref] Quantitative and qualitative transitions of methanogen community structure during the batch anaerobic..., Lee [/bib_ref]. However, it is too difficult to investigate a complex microbial community by DGGE due to the limited information and low resolution [bib_ref] A comprehensive microbial insight into two-stage anaerobic digestion of food waste-recycling wastewater, Shin [/bib_ref]. The analysis of the microbial community using a clone library is moreover very tedious and expensive [bib_ref] A pilot scale two-stage anaerobic digester treating food waste leachate (FWL): Performance..., Kim [/bib_ref]. With the ongoing development of sequencing technology, the high-throughput sequencing shows a high efficiency in microbial community structure identification. Especially, Illumina sequencing has been widely used for its low cost and high sequence depth merits [bib_ref] Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms, Caporaso [/bib_ref].
In this study, the culture-independent Illumina high-throughput sequencing technology was employed and the microbial community changes between the blank and high salinity groups (H group) were systematically investigated and compared. The objectives were: to evaluate the influence of increasing NaCl concentrations on the biogas production; to reveal the bacterial and archaeal communities diversity and structure; and to explore the relationship between microbial community structure and process performance under high salinity conditions.
# Results
Effect of increasing NaCl concentrations on biogas production. Prior to this study, the effects of different substrate concentrations [bib_ref] Kinetics of phenol biodegradation in high salt solutions, Peyton [/bib_ref] [bib_ref] Sodium inhibition of acetoclastic methanogens in granular sludge from a UASB reactor, Rinzema [/bib_ref] g/L starch) on the biogas production were firstly investigated. The results showed that the methane yields were 305.91, 236.90 and 245.71 mL/g-VS, respectively . Therefore, 6 g/L starch was chosen as the concentration for the subsequent experiment.
To investigate the influence of high salinity conditions on the biogas production, a blank group (0 g/L NaCl) and three supplement groups (L, M, and H group) with increasing NaCl concentrations (5, 10 and 20 g/L NaCl respectively) were assessed. The effects of increasing NaCl concentrations on characteristics of different groups were shown in [fig_ref] Figure 1: Cumulative methane productions at different concentrations of NaCl [/fig_ref]. CH 4 production. Generally, methane production rates of each group were higher in the early stage of fermentation (0~92 h), but gradually decreased thereafter [fig_ref] Figure 1: Cumulative methane productions at different concentrations of NaCl [/fig_ref]. The final cumulative CH 4 productions for blank, L, and M group were 2249.0 ± 131.3 mL, 2167.3 ± 127.6 mL and 2137.2 ± 138.1 mL respectively, which were nearly identical with a slight decrease at increasing NaCl concentration. Notably, the supplement of 5 g/L NaCl (L group) could even promote the biogas production before 264 h, although an overall higher CH 4 production was obtained at the end of fermentation in the blank group. The methane production in the H group was significantly reduced to 1301 ± 125.7 mL compared with other three groups (all P < 0.05), indicating that NaCl inhibition clearly occurred in the H group. Besides, a distinct lag phase (0~48 h) could also be observed in the H group.
pH. As can be seen in [fig_ref] Figure 2: pH variation at different concentrations of NaCl [/fig_ref] , the addition of NaCl differently influenced the pH value. The pH value in both the blank and L groups decreased sharply within 24 h. In contrast, this decreasing trend in the M and H groups was delayed to 48 h and 60 h respectively. After the hydrolysis/acidogenesis steps, the pH in all groups recovered due to the utilization of volatile fatty acids (VFAs) by methanogens for biogas production [fig_ref] Figure 1: Cumulative methane productions at different concentrations of NaCl [/fig_ref]. The pH value in all groups decreased with an increasing NaCl concentration, with the final pH of 7.60, 7.56, 7.44 and 7.32 for the blank, L, M and H groups respectively, indicating that an overloading of NaCl could lead to a higher VFAs accumulation and a lower pH value in the AD process. CO 2 production. The cumulative CO 2 productions at different concentrations of NaCl were shown in [fig_ref] Figure 3: Cumulative CO 2 productions at different concentrations of NaCl [/fig_ref]. Overall, the CO 2 production in all groups was rapidly accumulated within 48 h, where substrates were hydrolyzed and utilized by bacteria. In this step, the L group exhibited the highest CO 2 production rate, followed by the M group and the H group, while the blank group showed the lowest CO 2 accumulation rate. However, the CO 2 production rates in the L, M and blank groups were slowed down afterwards, while the H group maintained a high CO 2 accumulation rate. As a result, the final cumulative CO 2 productions for blank, L, M and H group were 1228.7 ± 41.7 mL, 1554.2 ± 55.6 mL, 1576.6 ± 62.4 mL and 1658.7 ± 54.9 mL respectively.
Overview of the 16S rRNA high-throughput sequencing results. For bacterial communities analysis, 42,855~53,163 clean reads were obtained for each sample with an average length of 412 bp after removing low quality sequences and chimeras. The sequence number of each sample was normalized and 491~530 operational taxonomic units (OTUs) were generated. Analysis of archaeal communities resulted in 44,098~57,074 sequences for each sample with an average length of 382 bp. Richness and diversity of the bacterial and archaeal communities, indicated by OUTs, Chao 1 value and Shannon index, were calculated and listed in .
For the bacterial community, the OUTs and Chao 1 values in both the blank and H groups at 480 h showed no significant difference (P > 0.05), but both were higher than those of the original sludge at 0 h (P < 0.05). The Shannon index in the blank group significantly increased from 4.82 ± 0.02 to 5.23 ± 0.16 during the AD process, while it decreased to 4.30 ± 0.21 in the H group (P < 0.05), demonstrating that the high concentration of NaCl could significantly reduce the sludge diversity. Towards the archaeal community, the richness of both groups was reduced, as reflected by a reduced number of OTUs, Chao 1 . Moreover, the blank group showed a much higher Shannon index compared with that of the H group at 480 h (3.84 ± 0.37 vs. 3.01 ± 0.25, P < 0.05). This also revealed that the high salinity negatively affected the diversity of the archaeal community. Meanwhile, three significant tests, i.e. Adonis, ANOSIM (analysis of similarity) and MRPP (multi-response permutation procedure), were carried out based on Jaccard distances to analyze the differences of microbial communities of different samples . A P value less than 0.05 meant a significant difference. The difference analysis indicated that the microbial communities of different samples were significantly different (all P < 0.05). [fig_ref] Figure 4: Taxonomic composition of the microbial community at phyla level in each sample [/fig_ref]. Of the total sequences, less than 2% were not classified at any phylum level. In general, the phylum Euryarchaeota, which includes all known methanogens, significantly increased from 5.29% (0 h) to 48.77% (blank group) and 18.26% (H group) after 480-hour operation in an AD process. For bacterial communities, Bacteroidetes (34.59%), Firmicutes (25.82%) and Proteobacteria (20.41%) were the three dominant phyla in the original sludge (0 h). However, a considerable decrease could be observed in both the blank and H groups in abundance of Bacteroidetes (8.76% and 8.66%) and Proteobacteria (4.33% and 4.45%) during the AD process. Firmicutes significantly decreased to 11.29% in the blank group while it remained relatively stable (24.41%) in the H group. In addition, the abundance of Synergistetes (9.36% and 14.92%) and OP9 (10.11% and 9.52%) also increased in both the blank and H groups. Comparing the bacterial community changes between the two groups in the late phase, Firmicutes, Thermotogae, Actinobacteria and Chloroflexi were more dominant in the H group while Bacteroidetes, Proteobacteria and OP9 showed no distinct change under high concentrations of NaCl.
## Bacterial community variations at genus level under different inhibitory conditions.
To further reveal the bacteria shift under high salinity inhibition, the bacterial communities of the blank and H groups at genus level were analyzed and the phylogenetic tree showing the phylogenetic identities of different genera was illustrated in . At the genus level, Marinilabiaceae uncultured bacterium (35.90%) and Pseudomonas (17.60%) were the two primary genera in the original sludge, but were hardly detected after the AD process . Other major genera included Bacteroides (2.75%), Paludibacter (3.25%), Clostridium (2.50%), Blautia (2.70%), Soehngenia (6.80%), BA021 uncultured (3.45%) and Thermovirga (3.15%). However, bacterial communities and major genera greatly shifted in the blank and H groups after 480-hour of AD. In the blank group, Bacteroides (9.80%), BA021 uncultured (9.60%) and Clostridium (8.80%) increased to be the three foremost genera, followed by Thermovirga (5.20%), Tissierella (2.70%), Soehngenia (2.50%) and Kosmotoga (2.20%). In the H group, Thermovirga, BA021 uncultured and Soehngenia were noted to be the most three prevalent genera, with relative abundance of 13.20%, 8.80% and 8.50%, respectively. Actinomyces (4.30%), Clostridium (4.50%), Tissierella (5.20%) and Kosmotoga (6.00%) increased to some extent compared with the blank group. In addition, the percentage of Bacteroides decreased to 2.10% while other genera affiliated to the same phylum such as Marinilabiaceae uncultured bacterium (1.40%) and Porphyromonas (2.10%) increased .
## Archaeal community variations at genus level under different inhibitory conditions. the rela-
tive abundances of the archaeal community at genus level were shown in . There were only four major genera, Methanosaeta, Methanolinea, Methanospirillum, and Methanoculleus, that were predominant in both groups where these genera accounted for more than 99.5% of all archaea. The genus Metanosaeta was extraordinary . Effective DNA sequences, richness and alpha diversity from bacterial and archaeal communities analysis. a Chao 1 richness estimator: a higher number represents higher richness. b Shannon index (H): a higher value indicates more diversity. dominant in the original sludge (91.32%), but significantly reduced to 69.82% in the blank group during the AD process. However, this genus showed more dominant in the H group with sequence percentage of 84.59%. Comparatively, the genus Methanolinea, which was the second primary genera in the original sludge (6.71%), . Percentages of the major genera (A) in each sample and neighbor-joining phylogenetic tree (B) of different genera. The major means sequence percentage is above 1% in any sample. 0 h stands for the samples that were collected immediately after inoculation. B-480 h and H-480 h stand for the samples at 480 h of the blank and H group respectively. A more red intense color corresponds with higher percentages, and deeper green colors indicate the lower percentages. For the phylogenetic tree construction, the neighbor-joining method was used and sequences were aligned using Clustal X 1.8 and MEGA 5.1. The bar represents 5% sequence divergence.
increased continuously reaching 25.12% and 13.99% in the blank and the H group respectively. In addition, Methanospirillum and Methanoculleus also showed a modest increase in the blank group with a percentage of 2.74% and 1.46% while they were hardly detected in the H group. It was worth noting that all the major hydrogenotrophic methanogens showed a significantly decrease in response to high salinity stress while the aceticlastic methanogens increased.
# Discussion
Some literature reported that low salinity concentrations were beneficial for the growth methanogens with value of 350 mg Na + /L (~0.8 g/L NaCl), while 8~13 g/L NaCl would cause moderate inhibition and values over 20 g/L NaCl would lead to severe impairment [bib_ref] Principles and potential of the anaerobic digestion of waste-activated sludge, Appels [/bib_ref] [bib_ref] Inhibition of anaerobic digestion process: a review, Chen [/bib_ref] [bib_ref] Anaerobic treatment of seafood processing waste waters in an industrial anaerobic pilot..., Omil [/bib_ref]. Consistent with these studies, the cumulative CH 4 productions in [fig_ref] Figure 1: Cumulative methane productions at different concentrations of NaCl [/fig_ref] showed no obvious difference among the blank, L and M groups (all P > 0.05). Moreover, the L group even showed higher production rate before 264 h. The H group however exhibited a strong inhibition with a methane production of 1301 ± 125.7 mL, which was 42.2% lower than that of the blank group. Two distinct phases, acidogenesis and methanogenesis, could be observed in the pH variation [fig_ref] Figure 2: pH variation at different concentrations of NaCl [/fig_ref]. During the first phase, i.e. acidogenesis, organic matter was converted to VFAs and CO 2 , thus resulting in an increasing VFA concentration and decreasing pH. Notably, the acidogenesis phase lasted 30 hours in the blank and L groups, while it was extended to 48 and 60 hours in the M and H groups respectively. Ren et al. reported that it required some time before microorganisms (mostly bacteria in this phase) became dominant and displayed their functions [bib_ref] Microbial community structure in an integrated A/O reactor treating diluted livestock wastewater..., Ren [/bib_ref]. In the present study, the higher the NaCl concentration was, the longer the time it required. It should be noted that the pH levels in the four groups did not vary significantly with a minimum value ranging from 6.45 to 6.65 during this phase, despite being extended at high salt concentrations (10-20 g/L NaCl). In addition, VFA accumulation as shown in [fig_ref] Figure 1: Cumulative methane productions at different concentrations of NaCl [/fig_ref] also supported this conclusion, where the VFA concentrations in the M and H groups were even higher than those of the blank and L groups. However, this higher VFA accumulation should be attributed to the lower VFA assimilation efficiency of methanogens which was inhibited by a high salt concentration. Therefore, it could be concluded that the increasing NaCl concentrations from 0 g/L to 20 g/L might not have a great impact on the substrate consumption by bacteria in the acidogenesis phase. [fig_ref] Figure 3: Cumulative CO 2 productions at different concentrations of NaCl [/fig_ref] moreover showed that the supplement of additional NaCl (L, M and H groups) did not inhibit but even improve the CO 2 production compared with that of the blank group in acidogenesis, which might also support this conclusion. During the second phase, i.e. methanogenesis, the VFAs generated during the acidogenesis were used for methane production, thus resulting in a decreasing VFA concentration and increasing pH. However, pH levels varied significantly at high salt concentrations [fig_ref] Figure 2: pH variation at different concentrations of NaCl [/fig_ref]. The final pH values were only 7.44 and 7.32 in the M and H groups, compared with values of 7.56 and 7.60 in the L and blank groups, indicating that the methanogens might be inhibited, especially in the H group.
To further quantify the NaCl inhibition on the acidogenic and methanogenic activity, the specific CO 2 production rate of bacteria obtained in acidogenesis and the specific CH 4 generation rate of archaea calculated in methanogenesis were compared and shown in [fig_ref] Figure 6: Comparison of specific CH 4 and CO 2 generation rates in different... [/fig_ref]. The highest specific CO 2 (59.47 ± 1.90 mL CO 2 g −1 VSS day −1 ) . Taxonomic composition of archaeal community at genus level of blank and H groups. and CH 4 production rates (47.39 ± 0.86 mL CH 4 g −1 VSS day −1 ) were obtained at 5 g/L NaCl in the L group, while NaCl concentration exceeding 5 g/L would decrease both the specific CO 2 and CH 4 production rates, suggesting that the biogas production could be enhanced by supplying an appropriate salt concentration (5 g/L NaCl in this case). More importantly, although it could lead to a reduction in the specific CO 2 production rate when the NaCl concentration exceeded 5 g/L, the specific CO 2 production rate in the H group was even higher than that in the blank group (50.11 ± 1.61 vs. [bib_ref] Field-based evidence for copper contamination induced changes of antibiotic resistance in agricultural..., Hu [/bib_ref].62 ± 1.90 mL CO 2 g −1 VSS day −1 , P < 0.05). This result further proved the above conclusion that the increasing NaCl concentrations did not negatively affect the degrading capability of bacteria in acidogenesis. Instead, it could enhance the acidogenic effect even when NaCl concentration reached 20 g/L. Unlike acidogenesis, the specific CH 4 generation rate in methanogenesis showed that a significant inhibition occurred when the NaCl concentrations increased from 5 to 20 g/L, which caused a decrease of 37.12% of the specific CH 4 generation rate (from 47.39 ± 0.86 to 29.80 ± 1.48 mL CH 4 g −1 VSS day −1 , P < 0.05). The results suggested that methanogenesis, rather than acidogenesis, was strongly affected at 20 g/L NaCl in the H group.
The above results showed that a high NaCl concentration could differently affect bacteria and archaea. In order to provide a detailed insight into how the microbial community changes the biogas production under high salinity inhibitory conditions, the samples of the blank and H groups at different periods (0 h and 480 h) were selected for further investigation by Illumina high-throughput analysis. illustrated that the bacterial richness of both the blank and H groups showed no difference (P > 0.05), while the diversity of the H group was less than the blank group. The result indicated that the high salinity only reduced the bacterial diversity rather than their richness during the AD process. However, the archaeal richness and diversity of the H group was significantly lower than that of the blank group, suggesting that both the richness and diversity of archaea were inhibited by the high NaCl concentration. These results were consistent with the above conclusion that methanogenesis showed more inhibition than acidogenesis in the H group.
Dominant microbial communities of the two groups were notably different [fig_ref] Figure 4: Taxonomic composition of the microbial community at phyla level in each sample [/fig_ref]. As commonly found in various AD processes, phyla Bacteroidetes (34.59%), Firmicutes (25.82%) and Proteobacteria (20.41%) dominated in the original sludge (0 h). These bacteria were responsible for biomass degradation and digestion [bib_ref] Presence and role of anaerobic hydrolytic microbes in conversion of lignocellulosic biomass..., Azman [/bib_ref] [bib_ref] Characterization of microbial community structure during continuous anaerobic digestion of straw and..., Sun [/bib_ref]. After 480 h of AD, Euryarchaeota exhibited an extremely high abundance (48.78%) while the relative abundance of bacteria such as Firmicutes, Bacteroidetes, and Proteobacteria significantly decreased due to the shift from acidogenesis to methanogenesis. However, the high concentration of NaCl as supplement changed the phyla abundance. Euryarchaeota significantly decreased from 48.78% to 18.26%, suggesting that the methanogenesis was severely inhibited. Interestingly, the phyla Firmicutes, Synergistetes, Thermotogae and Actinobacteria, which were important bacteria for substrate hydrolysis in acidogenesis, even increased in comparation with the blank group [fig_ref] Figure 4: Taxonomic composition of the microbial community at phyla level in each sample [/fig_ref]. This also further indicated that methanogenesis rather than acidogenesis was significantly inhibited in the H group. illustrated more details about the change of the blank and H group at genus level. For the blank groups, the phylum Bacteroidetes, which was the primary dominant bacteria (34.59%) in the original sludge, severely decreased to 8.76% in the AD process. It could be attributed to the significant decrease of Marinilabiaceae uncultured bacterium from 35.9% to 0.6% . Although it was reported that the family of Marinilabiaceae was obligatory anaerobic and saccharolytic bacteria [bib_ref] Genome sequence of Anaerophaga sp. strain HS1, a novel, moderately thermophilic, strictly..., Gao [/bib_ref] [bib_ref] Mangroviflexus xiamenensis gen. nov., sp. nov., a member of the family Marinilabiliaceae..., Zhao [/bib_ref] , these bacteria might not be suitable for the AD process and disappeared. However, Bacteroides, which was affiliated to the same phylum, was the major in the blank group with a proportion of 9.80%. It was reported that these species could secrete different hydrolyzing enzymes such as cellulase, amylase, protease and lipase 29 , indicating its importance for the depolymerisation of organic matter in the acidogenesis phase. Clostridium was also found dominant in the blank group (8.80%). These species were widely distributed in various anaerobic systems and played an important role in the acidogenesis/acetogenesis stage [bib_ref] Evaluation of A Novel Split-Feeding Anaerobic/Oxic Baffled Reactor (A/OBR) For Foodwaste Anaerobic..., Wang [/bib_ref] [bib_ref] Clostridium phytofermentans sp. nov., a cellulolytic mesophile from forest soil, Warnick [/bib_ref]. Our previous work demonstrated that these acid-generating species such as Bacteroides sp. and Clostridium sp. could reduce the pH value and further affect the bacterial community, leading to an increase of the adaptive species and a decrease of the non-adaptive species [bib_ref] Semi-continuous anaerobic digestion for biogas production: influence of ammonium acetate supplement and..., Su [/bib_ref].
The microbial community in the H group also shifted and a greater evenness of species was observed . For example, Bacteroides decreased from 9.8% to 2.1%, while other genera such as Actinomyces (4.30%), N09 (2.60%), and Porphyromonas (2.10%) increased and displayed similar hydrolytic abilities [bib_ref] Aspergillus oryzae S-03 Produces Gingipain Inhibitors as a Virulence Factor for Porphyromonas..., Danshiitsoodol [/bib_ref] [bib_ref] Spatial succession and metabolic properties of functional microbial communities in an anaerobic..., Peng [/bib_ref]. The percentage of Clostridium decreased from 8.80% to 4.50%, whereas two major genera, Tissierella and Soehngenia, which belong to the same family Tissierellaceae, were found to be more abundant than Clostridium in the H group within this phylum . These two genera were also reported to be functionally important in acidogenesis [bib_ref] Multilocus analysis reveals diversity in the genus Tissierella: Description of Tissierella carlieri..., Alauzet [/bib_ref] [bib_ref] Soehngenia saccharolytica gen. nov., sp. nov. and Clostridium amygdalinum sp. nov., two..., Parshina [/bib_ref]. Thermovirga was observed to be the most prevalent genera in the H group at the late phase, with relative abundance of 13.2%. Thermovirga was reported to have a high tolerance to high salinity [bib_ref] Thermovirga lienii gen. nov., sp. nov., a novel moderately thermophilic, anaerobic, amino-aciddegrading..., Dahle [/bib_ref] and dispalyed a high hydrolysis ability 37 . As described above, the H group was not inhibited in acidogenesis, in which VFAs accumulation and CO 2 production were similar to the blank group [fig_ref] Figure 3: Cumulative CO 2 productions at different concentrations of NaCl [/fig_ref] and [fig_ref] Figure 1: Cumulative methane productions at different concentrations of NaCl [/fig_ref]. This further revealed that the microbial community changed in high salinity condition to maintain the hydrolysis capacity.
It should be noticed that both the blank and H groups harbored a significantly high percentage of BA021 uncultured bacterium (9.6% and 8.8%, respectively), which was assigned to candidate phylum OP9 (now termed as phylum Atribacteria). With the help of culture-independent technologies, these bacteria were frequently detected in anaerobic digesters, petroleum reservoirs, and deep marine sediment 38 , indicating that these bacteria might be tolerant to a high concentration of NaCl. Moreover, a recent study reported that the phylum OP9 was abundant in high-methane conditions, and might play a key role in regulating both the production of methane and the diversity of methane producers [bib_ref] Abundant Atribacteria in deep marine sediment from the Adélie Basin, Carr [/bib_ref].
Taken together, although a significantly bacterial community shift could be observed between the H group and the blank group and many bacteria were inhibited under high salinity conditions, a variety of halotolerant bacteria that exhibited similar hydrolytic and acidogenetic abilities adapted to the high salt concentrations and even became dominant. Therefore, in line with the above conclusion, the increasing NaCl concentrations might not have a great impact on the substrate degradation. As previously stated, the decrease in methane production might largely relate to the inhibition of the archaeal community.
Unlike the bacterial community, the archaeal compositions in the both groups were quite simple with only four major genera representing more than 99.5% of all archaea . Generally, the relative abundance of phylum Euryarchaeota decreasing from 48.77% (blank group) to 18.26% (H group) suggested the significant inhibition in response to high salinity stress [fig_ref] Figure 4: Taxonomic composition of the microbial community at phyla level in each sample [/fig_ref]. Detailed archaeal community analysis at genus level showed that the aceticlastic methanogen Methanosaeta was the most important dominant species in the both groups, which was consistent with our previous result [bib_ref] Semi-continuous anaerobic digestion for biogas production: influence of ammonium acetate supplement and..., Su [/bib_ref]. Methanosaeta was reported to be the predominant methane producer on earth [bib_ref] Methanosaeta, the forgotten methanogen?, Smith [/bib_ref] and showed a high affinity and low minimum threshold for acetate [bib_ref] Effect of temperature on carbon and electron flow and on the archaeal..., Fey [/bib_ref] [bib_ref] Methanogenesis from acetate: a comparison of the acetate metabolism in Methanothrix soehngenii..., Jetten [/bib_ref]. These species could consume acetate for CH 4 production, resulting in an increase of pH and decrease of VFAs [fig_ref] Figure 2: pH variation at different concentrations of NaCl [/fig_ref] and [fig_ref] Figure 1: Cumulative methane productions at different concentrations of NaCl [/fig_ref]. It was worth noting that the relative abundance of Methanosaeta increased from 69.82% to 84.59% with elevated NaCl concentrations , indicating that these species might not be affected significantly in the H group. reported that the methanogen population would shift from acetoclastic methanogens to hydrogenotrophic methanogens with the release of VFAs [bib_ref] A pilot scale two-stage anaerobic digester treating food waste leachate (FWL): Performance..., Kim [/bib_ref]. Consistently in the blank group, the hydrogenotrophic methanogens such as Methanolinea, Methanospirillum and Methanoculleus significantly increased with the relative abundance of 25.12%, 2.74% and 1.46% respectively . These hydrogenotrophic genera were capable of producing CH 4 through the reduction of CO 2 with H 2 and could use formate and alcohols as alternative electron donors [bib_ref] Reviewing the anaerobic digestion of food waste for biogas production, Zhang [/bib_ref] [bib_ref] A pyrosequencing-based metagenomic study of methane-producing microbial community in solid-state biogas reactor, Li [/bib_ref]. However, it seemed that the shift from acetoclastic methanogens to hydrogenotrophic methanogens was significantly inhibited in the H group due to the high concentration of NaCl. The major hydrogenotrophic genus Methanolinea decreased dramatically to 13.99% while the other two genera were hardly detected . According to our experiment data in [fig_ref] Figure 3: Cumulative CO 2 productions at different concentrations of NaCl [/fig_ref] , the cumulative CO 2 production of the H group was 34.99% higher than that of the blank group, whereas this increase should be largely attributed to the inhibition of hydrogenotrophic methanogens.
To verify the above conclusion, the final CH 4 and CO 2 concentrations of the blank and H groups were further analyzed and shown in [fig_ref] Figure 2: pH variation at different concentrations of NaCl [/fig_ref]. As mentioned above, hydrogenotrophic methanogens could utilize H 2 and CO 2 as substrate to produce CH . Therefore, a higher CO 2 and lower CH 4 concentration would be expected if hydrogenotrophic methanogens were inhibited. As expected, a significant decrease in CH 4 concentration (from 61.08 ± 1.47 to 55.57 ± 0.43, P < 0.05) and increase in CO 2 concentration (11.79 ± 0.66 vs. 14.21 ± 0.18, P < 0.05) could be observed in the H group than in the blank group. Hence, the high salinity might be less toxic to aceticlastic methanogens than to hydrogenotrophic methanogens.
In summary, the impact of increasing NaCl concentrations on the biogas production was systematically investigated and the Illumina high-throughput sequencing method was employed to reveal the microbial community variations in response to high salinity stress. The increasing NaCl concentrations from 0 g/L to 20 g/L did not negatively affect the activity of bacteria during acidogenesis but even promoted it, while the methanogenic activity in methanogenesis was strongly inhibited. Microbial communities between the blank and H groups were significantly different. Bacteroides, Clostridium and BA021 uncultured were the major genera in the blank group. Nevertheless, these species were inhibited to some extent in the H group while some halotolerant bacteria, such as Thermovirga, Soehngenia and Actinomyces, were dominant and displayed similar hydrolytic and/or acidogenetic abilities. The archaeal community compositions were quite simple and were more easily affected by high salinity. Hydrogenotrophic methanogens showed a lower resistance to high salinity than aceticlastic methanogens.
# Materials and methods
Inoculums and anaerobic digestion. The seed sludge was obtained from a 100 m 3 food waste anaerobic digester in the Changping district, Beijing, and had been acclimated over one year for FW treatment. The seed sludge was kept for 5~8 days at 35 °C to consume the residual organic matter before inoculation.
The batch anaerobic digestion was operated using 1 L glass digesters with total solids (TS) of 5 wt% for anaerobic digestion at mesophilic temperature (35 ± 1 °C). To avoid the influence of the high fluctuating compositions of FW on the AD process, starch and peptone were used instead of FW. The inoculum and substrates were fully mixed before being added to the digesters. After inoculation, each digester was flushed for 10 min (300 mL/min) with argon gas to provide anaerobic condition. Prior to this study, the medium was optimized to be 6 g/L starch and 1.22 g/L peptone (C/N = 15). Four different groups were examined in triplicates, where NaCl was supplemented in increasing concentrations of 0 g/L (blank group), 5 g/L (L group), 10 g/L (M group) and 20 g/L (H group), respectively. All the digesters were operated under identical conditions. All the reagents used in this study were analytically pure and were purchased from Xilong Scientific Co., Ltd, Beijing, China.
DNA extraction and pyrosequencing. Sludge samples of the original and final stages of the blank and H groups were collected and stored at −80 °C before use. The total genomic DNA was extracted using Soil DNA extraction kit (OMEGA, Georgia, United States). The DNA concentration was determined by Nano drop (NanoDrop Technologies, Inc, Wilmington, United States). Prepared genomic DNA samples were sent to Novegene (Beijing, China) for shotgun library construction using an Illumina Hiseq2500 platform. The raw sequences were joined and treated according to Hu et al. [bib_ref] Field-based evidence for copper contamination induced changes of antibiotic resistance in agricultural..., Hu [/bib_ref]. Quantitative Insights Into Microbial Ecology (QIIME) 1.7.0 was used to sift the raw reads, the resulting high quality sequences were clustered into operational taxonomic units (OTUs) at the 97% sequence similarity threshold by Uclust clustering. The phylogenetic tree was constructed by MEGA version 5.1, using the neighbor-joining method.
Analytical methods. The total solid (TS), volatile solid (VS) and volatile suspended solid (VSS) were measured according to the standard methods. The pH was monitored by an ion meter (MP 523 pH/ISE meter, San-Xin Instrumentation, Inc, Shanghai, China). Biogas was collected by water displacement method. Biogas and VFAs concentrations were examined by gas chromatography (GC-2014C, Shimadzu, Kyoto, Japan) using a thermal conductivity detector (TCD) and hydrogen flame ionization detector (FID) respectively, as detailed in a previous study [bib_ref] The anaerobic co-digestion of food waste and cattle manure, Zhang [/bib_ref].
Scientific RepoRts | 7: 1149 | DOI:10.1038/s41598-017-01298-y Statistical analysis. Cluster analysis was performed to evaluate the overall differences in microbial community structure based on Jaccard distances. Dissimilarity tests (i.e. MRPP, ANOSIM and Adonis) were performed to determine the significance of differences between the microbial community compositions in R v.3.3.2 using the 'vegan' package [bib_ref] From immunosuppression to immunomodulation: current principles and future strategies, Team [/bib_ref]. Data were analyzed using SigmaStat 3.5. The one-way ANOVA were used to test the significance of differences between groups, and P < 0.05 was considered as significant.
[fig] Figure 1: Cumulative methane productions at different concentrations of NaCl. Blank group (0 g/L), L group (5 g/L), M group (10 g/L) and H group (20 g/L). The experiments were operated in 1 L digesters at mesophilic temperature of 35 ± 1 °C. [/fig]
[fig] Figure 2: pH variation at different concentrations of NaCl. [/fig]
[fig] Figure 3: Cumulative CO 2 productions at different concentrations of NaCl. Scientific RepoRts | 7: 1149 | DOI:10.1038/s41598-017-01298-y Microbial community variations at phyla level under different inhibitory conditions. The taxonomic compositions of the microbial community at phyla level were shown in [/fig]
[fig] Figure 4: Taxonomic composition of the microbial community at phyla level in each sample. The sequence percentage is above 1% in at least one sample. [/fig]
[fig] Figure 6: Comparison of specific CH 4 and CO 2 generation rates in different NaCl concentrations. [/fig]
|
Performance of Dexcom G5 and FreeStyle Libre sensors tested simultaneously in people with type 1 or 2 diabetes and advanced chronic kidney disease
BACKGROUNDAdvanced chronic kidney disease (CKD) is a common complication for people with type 1 and 2 diabetes and can often lead to glucose instability. Continuous glucose monitoring (CGM) helps users monitor and stabilize their glucose levels. To date, CGM and intermittent scanning CGM are only approved for people with diabetes but not for those with advanced CKD.AIMTo compare the performance of Dexcom G5 and FreeStyle Libre sensors in adults with type 1 or 2 diabetes and advanced CKD.METHODSThis was a non-randomized clinical trial that took place in two outpatient clinics in western Sweden. All patients with type 1 or 2 diabetes and an estimated glomerular filtration rate (eGFR) of < 30 mL/min per 1.73 m 2 were invited to participate. Forty patients (full analysis set = 33) carried the DexcomG5sensor for 7 d and FreeStyle Libre sensor for 14 d simultaneously. For referencing capillary Ólafsdóttir AF et al. Dexcom G5 and FreeStyle Libre simultaneous-testing WJCC https://www.wjgnet.com 7795 August 6, 2022 Volume 10 Issue 22blood glucose (SMBG) was measured with a high accuracy glucose meter (HemoCue ® ) during the study period. At the end of the study, all patients were asked to answer a questionnaire on their experience using the sensors.RESULTSThe mean age was 64.1 (range 41-77) years, hemoglobin A1c was 7.0% [standard deviation (SD) 3.2], and diabetes duration was 28.5 (SD 14.7) years. A total of 27.5% of the study population was on hemodialysis and 22.5% on peritoneal dialysis. The mean absolute relative difference for Dexcom G5 vs SMBG was significantly lower than that for FreeStyle Libre vs SMBG [15.2% (SD 12.2) vs 20.9% (SD 8.6)], with a mean difference of 5.72 [95% confidence interval (CI): 2.11-9.32; P = 0.0036]. The mean absolute difference was also significantly lower for Dexcom G5 than for FreeStyle Libre, 1.21 mmol/L (SD 0.78) and 1.76 mmol/L (SD 0.78), with a mean diffrenec of 0.55 (95%CI: 0.27-0.83; P = 0.0004).The mean difference (MD) was -0.107 mmol/L and -1.10 mmol/L (P = 0.0002), respectively. In all, 66% of FreeStyle Libre values were in the no risk zone on the surveillance error grid compared to 82% of Dexcom G5 values.CONCLUSIONDexcom G5 produces more accurate sensor values than FreeStyle Libre in people with diabetes and advanced CKD and is likely safe to be used by those with advanced CKD.Core Tip: This study bridges a needed gap within the diabetes device area for people with diabetes and advanced chronic kidney disease and was done in a home setting for analyses as close to real life as possible. The study found that Dexcom G5 showed greater accuracy both in relation to the mean absolute relative difference and on a surveillance error grid, but participants rated their user experience for FreeStyle Libre higher but rated no difference in feeling safe. Citation: Ólafsdóttir AF, Andelin M, Saeed A, Sofizadeh S, Hamoodi H, Jansson PA, Lind M. Performance of Dexcom G5 and FreeStyle Libre sensors tested simultaneously in people with type 1 or 2 diabetes and advanced chronic kidney disease. World J Clin Cases 2022; 10(22): 7794-7807
# Introduction
For people with diabetes, good glycemic control is essential to avoid problems due to diabetes complications [bib_ref] The effect of intensive treatment of diabetes on the development and progression..., Nathan [/bib_ref]. To reach recommended glucose levels, it is important to monitor glucose levels and for several years, self-measurement of blood glucose (SMBG) with capillary measurements has been the best way to do this [bib_ref] The effect of intensive treatment of diabetes on the development and progression..., Nathan [/bib_ref]. Over the last decades, continuous glucose monitoring (CGM) and intermittent glucose monitoring (isCGM) has become more common within diabetes management and for many, has replaced the multiple capillary tests. Both systems are made up of a small sensor that is inserted under the skin where it measures glucose levels in the interstitial fluid. CGM measures glucose levels continuously and every 5 min sends a glucose value to a handheld receiver or mobile telephone. It sends alarms for high and low glucose levels. The isCGM collects data, and when the user scans the sensor with a handheld receiver or a mobile phone, it sends the glucose levels to the receiver.
Within the diabetes field, there are many discussions regarding who should be given CGM and isCGM. To date, CGM and isCGM are only approved for people with diabetes but not with chronic kidney disease (CKD)and mainly recommended for those with type 1 diabetes and who have problems with recurrent hypoglycemia [bib_ref] The effect of intensive treatment of diabetes on the development and progression..., Nathan [/bib_ref].
Advanced CKD is a common complication in people with type 1 and 2 diabetes. It is estimated that 20%-40% of people with diabetes will develop diabetic kidney disease, and it is the leading cause of end-stage renal failure [bib_ref] Renal Complications and Duration of Diabetes: An International Comparison in Persons with..., Dena [/bib_ref]. A recent study showed that up to 5.1% of people with type 1 diabetes in Germany and Austria had an estimated glomerular filtration rate (eGFR) below 30 mL/min, and for Sweden and United States, the corresponding figures were 1.5% and 2.1%. Advanced CKD increases the risk of hypoglycemia and great glycemic variation, and therefore it can be helpful to monitor blood glucose with a CGM or isCGM . There are very few studies available on the accuracy of CGMs or isCGMs for people with advanced CKD . Two of the most common systems are Dexcom and FreeStyle Libre. Neither of these systems are approved for people in dialysis.
The aim of this study was to compare the performance of Dexcom G5 and FreeStyle Libre in adults with type 1 or 2 diabetes with CKD and an eGFR < 30 mL/min/1.73 m 2 , including patients on maintenance dialysis.
# Materials and methods
This study took place at NU Hospital Group and Sahlgrenska University Hospital, Sweden. It was a non-randomized, non-blinded clinical study over a 14 d period to compare the performance of FreeStyle Libre 1 and Dexcom G5 for people with diabetes and advanced CKD in an at-home situation. The protocol was approved by the regional ethics review board of Gothenburg, Sweden.
## Study procedures
All participants provided written informed consent before the study began. The inclusion criteria were: type 1 or type 2 diabetes, between 18-years-old and 80-years-old, and eGFR < 30 mL/min per 1.73 m 2 for people undergoing or not undergoing dialysis. The exclusion criteria were pregnancy, patients with severe cognitive dysfunction or other diseases that makes glucose monitoring difficult, continuous use of paracetamol, history of allergic reaction to chlorhexidine or alcohol antiseptic solution, abnormal skin at the anticipated glucose sensor attachment sites, and eGFR ≥ 30 mL/min per 1.73 m 2 .
After obtaining written and informed consent, a diabetes nurse inserted two different sensors in accordance with instructions from the manufacturer. Dexcom G5 was inserted in the abdomen and FreeStyle Libre on the upper arm. Participants were instructed on how they should use each monitor and instructed how to calibrate the Dexcom G5. Calibrations were done using the HemoCue ® DM RD 201 (Ängelholm, Sweden). All HemoCue meters were calibrated before being assigned to participants using the absolute isotope dilution gas chromatography/mass spectrometry measurement system . The total measurement error/reproducibility imprecision of HemoCue is less than 6.5% . Earlier studies using HemoCue showed a strong correlation between capillary and venous HemoCue concentrations, and capillary concentrations were considered to be a suitable reference . All participants were instructed by a diabetes nurse on how to use the HemoCue meter. Participants were instructed to simultaneously document their blood glucose measured by HemoCue and the value of the FreeStyle Libre and Dexcom G5 in a diary a minimum of three times per day. Participants were instructed to calibrate their Dexcom G5 twice daily in accordance with the manufacturer's instructions and to do so after recording its value in the diary. Participants on maintenance dialysis (peritoneal dialysis or hemodialysis) were also asked to register the start and finish of each session in their diary. After 7 d, Dexcom G5 was removed by the participants but they continued to record results from the FreeStyle Libre and HemoCue. After the 14 d period, participants returned the meters to the site. The study personnel downloaded data from the meters using the Glooko-Diasend system. HemoCue measurements were manually validated by personnel going through each value and comparing to the diary. When each sensor was finished, participants rated their experience on a 10-item visual analogue scale. Similar questionnaires have been used in earlier studies [bib_ref] A Clinical Trial of the Accuracy and Treatment Experience of the Flash..., Ólafsdóttir [/bib_ref] [bib_ref] Comparative Accuracy Analysis of a Real-time and an Intermittent-Scanning Continuous Glucose Monitoring..., Link [/bib_ref].
## Predefined endpoints
All endpoints were predefined and registered on ClinicalTrials.gov. The primary endpoint was the difference of mean absolute relative difference (MARD) between Dexcom G5 and FreeStyle Libre using HemoCue (capillary glucose meter) as a reference. Secondary endpoints were the difference in mean absolute difference (MAD) between the Dexcom G5 and FreeStyle Libre sensors, the difference in mean difference (MD) between the Dexcom G5 and FreeStyle Libre sensors, and the correlation between the different systems. Predefined subgroup analyses for glucose ranges below 3.9 mmol/L, between 3.9 and 10 mmol/L, and above 10 mmol/L as well as for those without dialysis and undergoing dialysis.
## Independence of the study
The manufacturers of FreeStyle Libre and Dexcom G5 were not involved in the design, performance, data analysis, or publication of the article. No support was received from the manufacturers.
# Statistical analysis
After sample size analysis, 40 patients were included in the study (see supplement). All main analyses between Dexcom G5 and FreeStyle Libre were performed with paired analyses. All statistical analyses were predefined in the statistical analysis plan before database lock. All participants having at least 10 matched time points, with evaluable blood glucose values from both sensors and HemoCue (reference capillary value) during the whole study period, were included in the Full Analysis Set (FAS). All matching time points were used. For paired analysis regarding continuous variables, Fisher's nonparametric permutation test for paired observations was used and for dichotomous and ordered categorical variables sign test was used. For comparison between dialysis subjects and subjects not in dialysis, Fisher's non-parametric permutation test was used for continuous variables.
The primary variable was MARD, which is the mean absolute relative difference between the estimated sensor glucose value of FreeStyle Libre or Dexcom G5 and blood glucose measured with HemoCue. For each individual mean of following differences from each time point was evaluated for both sensors: |(sensor і -HemoCue i )|/HemoCue i . The secondary variables were MAD and MD.
MAD is the mean absolute difference between estimated sensor glucose value of FreeStyle Libre or Dexcom G5 and blood glucose measured with HemoCue. For each individual mean of following differences from each time point was evaluated for both sensors: |sensor і -HemoCue i |. MD is the mean difference between estimated sensor glucose value of FreeStyle Libre or Dexcom G5 and blood glucose measured with HemoCue. For each individual mean of following differences from each time point was evaluated: (sensor і -HemoCue i ), where i = time-point during the analyzed days in the study.
The MD between Dexcom G5 and FreeStyle Libre with 95% confidence intervals (CIs) was calculated based on Fisher's non-parametric permutation test for paired observations for continuous variables. All analyses for different glucose ranges were based on HemoCue values within respective range.
To study the covariation between Dexcom G5/FreeStyle Libre and HemoCue Pearson correlation coefficient between each of the devices and HemoCue was calculated for each subject. These correlations were also analyzed both for Dexcom G5 and FreeStyle Libre with Fisher's non-parametric permutation test one sample test.
Agreement between each of the devices and HemoCue were analyzed with Bland-Altman' methods. The main result was the limit of agreement. If one got a value measured with one of the sensors, you can calculate an interval where 95% of the HemoCue values would have been. The distributions of the difference between each of the sensor and HemoCue was also given together with Intraclass correlation coefficient (ICC), Bland-Altman plots, and scatterplots.
All significance tests were two-sided and conducted at the 5% significance level. All statistical analyses were performed with SAS System Version 9.4 (Cary, NC, United States).
## Post-hoc analyses
The surveillance error grid graph for Dexcom G5/FreeStyle Libre vs HemoCue was calculated by using https://www.diabetestechnology.org/seg/. The proportion of sensor values within 15%, 20%, and 30% of reference values HemoCue for blood glucose > 100 mg/dL (5.6 mmol/L) or within 15, 20, and 30 mg/dL (0.8, 1.1, 1.7 mmol/L) of reference values for blood glucose ≤ 100 mg/dL (5.6 mmol/L), respectively, was calculated (%15/15, %20/20, %30/30). MARD FreeStyle Libre vs HemoCue the first week was compared with the second week with the same requirements as main study with Fisher's nonparametric permutation test for paired observations.
# Results
The study included 40 participants with type 1 and 2 diabetes and advanced CKD; 33 (FAS) met the criteria for data analysis and at least 10 time points with evaluable values from both systems and the HemoCue within 5 min during the whole study period (June 2016-March 2019). Of the 7 patients who were not included in FAS, 2 chose not to participate after starting the study and 5 did not meet the criteria for data analysis described above; that is, they did not have 10 matched time points for both sensors. Mean hemoglobin A1c (HbA1c) was 7.0%, 25.6% were women, mean age was 64.1 (range 41-77), and 50% were on dialysis. Additional baseline characteristics are shown in [fig_ref] Table 1: Baseline characteristicsFor categorical variables n [/fig_ref].
## Accuracy evaluations
The MARD analyzed for all participants for Dexcom G5 was significantly lower than that for FreeStyle Libre vs SMBG (15.2% [SD 12.2] vs 20.9% [SD 8.6]), respectively, with mean difference of 5.72 (95%CI: 2.11-9.32; P = 0.0036). The MAD was also significantly lower for Dexcom G5 than for FreeStyle Libre, 1.21 mmol/L (SD 0.78) and 1.76 mmol/L (SD 0.78), with a mean difference of 0.55 (95%CI: 0.27-0.83; P = 0.0004). There was also a significant difference between the MD of the systems. There was a systematic MD between FreeStyle Libre and HemoCue of -1.10 mmol/L (95%CI: -1.55 to -0.66 mmol/L; P < 0.0001) but no systematic MD between Dexcom G5 and HemoCue -0.107 (95%CI: -0.439 to 0.225; P = 0.052).
We found that for glucose values that were in range (3.9-10.0 mmol/L) and above range (> 10 mmol/L), there was a significantly lower MARD, MAD, and MD for Dexcom G5 than for FreeStyle Libre. For glucose values in range, the MARD was 14.8% (SD 10.6) for Dexcom G5 and 22.6% (SD 8.9) for the FreeStyle Libre, with a mean difference of 7.83 (95%CI: 4.32-11.33; P < 0.0001). The MARD for hyperglycemic values were 12.3% (SD 11.6) and 16.6% (SD 11.1), respectively, with a mean difference of 4.22 (95%CI: 1.06-7.39; P = 0.010). There were few values below range (< 3.9 mmol/L), 14 values from 9 individuals.
## Subgroup analysis: people needing and not needing dialysis
Subgroup analyses for MARD, MAD, and MD were done for people requiring and not requiring dialysis. The MARD for FreeStyle Libre for people in dialysis was 19.3% (SD 7.4) compared to 22.5% (SD 9.5) for those not in dialysis (P = 0.29). The corresponding values for Dexcom G5 were 15.5% (SD 14.8) and 15.0% (SD 9.6), respectively (P = 0.91). For people not in dialysis, there was a significant difference between the sensors MARD and MAD (P = 0.0033 and P = 0.0057, respectively). For people in dialysis, there was a significant difference between the systems MAD (P = 0.035), whereas a numerical difference was found between the sensors MARD, although not statistically significant. Further subgroup analysis with people on peritoneal dialyses showed numerically lower MARD and MAD for Dexcom G5 compared to FreeStyle Libre as in the total population, and there was a significant systematic difference between FreeStyle Libre and HemoCue -1.58 (P = 0.01). There were 7 people on hemodialysis and Dexcom G5 showed a numerically lower MARD and MAD compared to FreeStyle Libre in this subgroup, but the differences were less. [fig_ref] Table 1: Baseline characteristicsFor categorical variables n [/fig_ref]. This could clearly be seen on the Bland-Altman plot in [fig_ref] Figure 1: Bland-Altman plot, scatterplot [/fig_ref] and Supplementary [fig_ref] Figure 1: Bland-Altman plot, scatterplot [/fig_ref].
## Correlation between the systems
## Patient experience
After using the systems, participants evaluated their experience. Participants were significantly more positive towards FreeStyle Libre than Dexcom G5 in all factors except feeling safe, for which there was no significance between the two systems. FreeStyle Libre scored 7.94 of 10 and Dexcom G5 scored 7.19 of 10 (P = 0.32;.
# Post hoc analysis
For Dexcom G5, %20/20 = 79.6, which indicates that 79.6% of the values above 5.6 mmol/L were within 20% of the reference instrument and within 1.11 mmol/L (20 mg/dL) for values below 5.6 mmol/L.
# Discussion
Dexcom G5 showed greater overall accuracy than FreeStyle Libre. Dexcom G5 also showed greater accuracy for glucose values within range (3.9-10 mmol/L) and above range (> 10 mmol/L). Furthermore, in a subgroup analysis, Dexcom G5 showed greater accuracy for people not in dialysis. However, for people in dialysis, Dexcom G5 had a numerically lower MARD and a significantly lower MAD compared with FreeStyle Libre. On the surveillance error grid, Dexcom G5 had 82% of values within the no risk zone compared to 66% for FreeStyle Libre. Glucose values from both sensors correlated well with the reference instrument, HemoCue. FreeStyle Libre showed a greater systematic deviation than Dexcom G5. Participants rated their user experience of FreeStyle Libre higher after a 2 wk period than Dexcom G5 but did not experience a difference in safety. Earlier studies with similar methodology and the same reference instrument showed that the FreeStyle Libre had a MARD of 13.2% and an earlier Dexcom sensor (Dexcom 4G) had a MARD of 13.8% when tested in people with type 1 diabetes [bib_ref] A Clinical Trial of the Accuracy and Treatment Experience of the Flash..., Ólafsdóttir [/bib_ref] [bib_ref] Comparative Accuracy Analysis of a Real-time and an Intermittent-Scanning Continuous Glucose Monitoring..., Link [/bib_ref]. A recent study analyzed how well FreeStyle Libre correlates with capillary measurements (Medisafe ® Fit) during hemodialysis in people with type 2 August 6, 2022
Volume 10 Issue 22 diabetes, and showed that the FreeStyle Libre had a MARD between 13% and 22% depending on the glycemic range and that it showed a 18.4 mg/dL (1.0 mmol/L) lower value than the capillary reference instrument. The same study found that the Medtronic iPro Enlite sensor had a MARD between 5% and 30% depending on the glycemic value and showed a 4.7 mg/dL (0.3 mmol/L) lower value than the reference instrument[10]. It was previously shown that the FreeStyle Libre deviates systematically by - 0.5 mmol/L in people with type 1 diabetes using HemoCue capillary measurements as a reference [bib_ref] Comparative Accuracy Analysis of a Real-time and an Intermittent-Scanning Continuous Glucose Monitoring..., Link [/bib_ref]. The Dexcom G5 was found to have a MARD of 7.1%-15.7% when tested in people with type 1 diabetes and using a Yellow Spring Instrument as a reference [bib_ref] Glycemic pattern in diabetic patients on hemodialysis: continuous glucose monitoring (CGM) analysis, Gai [/bib_ref]. People with advanced CKD more frequently experience glycemic excursions [bib_ref] Comparative Accuracy Analysis of a Real-time and an Intermittent-Scanning Continuous Glucose Monitoring..., Link [/bib_ref]. During hemodialysis, there is an increased risk for hypoglycemia, whereas patients with peritoneal dialysis have an increased hyperglycemia risk [bib_ref] Continuous glucose monitoring system in 72-hour glucose profile assessment in patients with..., Skubala [/bib_ref]. It is therefore important that this group of patients receives all possible help to monitor their glucose levels and to increase their possibility of better glycemic control. It is possible to speculate if these increased glucose excursions can possibly be the cause to the lower accuracy of these sensors for people with advanced CKD. This study found that the accuracy of FreeStyle Libre and Dexcom G5 while being used by people with advanced CKD is similar to the accuracy of earlier sensors which were used as glucose indicators and not for insulin dosing decisions [bib_ref] A Clinical Trial of the Accuracy and Treatment Experience of the Flash..., Ólafsdóttir [/bib_ref] [bib_ref] Comparative Accuracy Analysis of a Real-time and an Intermittent-Scanning Continuous Glucose Monitoring..., Link [/bib_ref]. An earlier study has found that when people on dialysis used CGM it led to more frequent treatment changes and better glycemic control .
This study showed that even people undergoing peritoneal dialysis, which can have high glucose fluctuations, had a MARD which is similar to previous systems. The peritoneal dialysis fluids did not seem to affect the MARD.
FreeStyle Libre had a higher MARD and MAD than Dexcom G5 and there was a greater percentage of values within the safe zone for Dexcom G5. This can partly be explained by the fact that the FreeStyle Libre showed a systematic deviation of -1.1 mmol/L. It is important that users of the system are aware of the systems tendency of reporting lower glucose values. This systematic deviation is not only evident when the sensor is used by people with advanced CKD although it seems to be greater for this patient group [bib_ref] Comparative Accuracy Analysis of a Real-time and an Intermittent-Scanning Continuous Glucose Monitoring..., Link [/bib_ref]. The surveillance error grid showed that only 66% of FreeStyle Libre values were in the no risk zone whilst 82% of Dexcom G5 values were within the no risk zone.
Participants rated the user experience of the FreeStyle Libre significantly higher than for the Dexcom G5. They found the system easier to use and easier to interpret the data on the receiver. The sensor was more comfortable, and it was less painful to insert. There was a greater interest to use the system in their daily life. This might be different with Dexcom's latest sensors which do not require calibration by the user. It is important to note that the users did not experience any difference of safety when using the system.
The strength of this study is that it was done independently from the manufacturers of this study. The study was done in a real-life environment as patients used the sensors in their daily life. All analyses were predefined. The limitations of this study were the short duration the participants used the sensors, and the evaluation of the user experience might change if the users become more comfortable and confident in the use of the sensors, and the questionnaire used is not validated. For certain subgroup analysis the number of participants or values obtained was low, therefore these analyses have to be interpreted with caution. It should be noted that Dexcom G5 was calibrated with the same capillary method as the reference system, and it cannot be excluded that more novel generations of Dexcom sensors which do not need calibrations may have a greater systematic deviation from HemoCue. Neither Dexcom G5 nor FreeStyle Libre are approved to be used by people with advanced chronic kidney disease. Another limitation is that the most novel sensors often used today were not evaluated. However, these data must be viewed in the light that CGM accuracy data are overall lacking in people with Diabetes and advanced CKD and data are therefore urgently needed.
# Conclusion
In conclusion, this study supports that Dexcom G5 has a similar accuracy in people with diabetes and advanced CKD as in people with diabetes without advanced CKD. The FreeStyle Libre system showed similar correlations between sensor value and blood glucose values as Dexcom, but a lower number of values in the no risk zone indicating that greater caution should be taken to use it in the current population. The FreeStyle Libre showed a systematic deviation at least partly explaining the lower accuracy.
## Article highlights
# Research background
People with diabetes and advanced chronic kidney disease (CKD) often have fluctuating blood glucose levels and today no blood glucose sensors are approved to be used in this patient group.
## Research motivation
It is of great importance to give the best possible care to all people with diabetes. This is a patient group with difficult complications due to their diabetes and need all the help they can get.
## Research objectives
The objective of this study was to see if the sensors FreeStyle Libre and Dexcom G5 were accurate when used by people with advanced CKD.
# Research methods
This was a non-randomized clinical study. The results were evaluated by using mean absolute relative difference as a main analysis. Mean absolute difference and mean difference was also calculated. A surveillance error grid was even used for accuracy evaluations.
# Research results
The main analysis found that the Dexcom G5 had a mean absolute relative difference of 15.2% while it was 20.9% for the FreeStyle Libre. There was no significant difference if the patients were on maintenance dialysis or not. There was no significant difference between those with type 1 or 2 diabetes. The surveillance error grid showed that Dexcom G5 had 82% of its values within the safe zone while FreeStyle Libre had 66% within the safe zone.
# Research conclusions
The study concludes that the Dexcom G5 produces more accurate values than the FreeStyle Libre.
## Research perspectives
This study is a great start for evaluating how we can use glucose sensors for this patient group, but further studies have to be done with more novel glucose sensors.
[fig] Figure 1: Bland-Altman plot, scatterplot. A: All individual measurements FreeStyle Libre vs HemoCue; B: All individual measurements Dexcom G5 vs HemoCue. Thick dotted line represents the mean difference. [/fig]
[fig] Figure 2: Surveillance error grid. A: FreeStyle Libre vs HemoCue -66.3% of values fall within the dark green area; B: Dexcom G5 vs HemoCue -82% of values fall within the dark green area. [/fig]
[table] Table 1: Baseline characteristicsFor categorical variables n (%) is presented. For continuous variables mean (SD)/median (min; max)/n = is presented. FAS: Full analysis set; MDI: Multiple daily injection. [/table]
|
The change in high-sensitivity troponin-T as a risk factor for significant coronary stenosis in patients with acute coronary syndrome
Background/Aims: High-sensitivity cardiac troponin (hs-TnT) assays detect very low levels of cardiac troponin. This study examined the interval change between initial and subsequent hs-TnT levels and evaluated its ability to predict significant coronary stenosis.Methods:The study analyzed 163 patients who presented with acute coronary syndrome (ACS) and underwent coronary angiography (CAG) between April 2014 and May 2018. The 0 and 3-hour hs-TnT were checked. The patients were subdivided into positive (n = 32) and negative (n = 131) interval change groups. The presence of significant coronary artery stenosis on CAG in the two groups was compared.Results:The positive interval change group was older and had higher 0 and 3-hour hs-TnT and blood glucose levels than the negative interval change group. Significant coronary stenosis was more common in the positive interval change group than in the negative interval change group (68.8% vs. 23.7%, p = 0.001). However, vasospasm was more common in the negative interval change group (6.3% vs. 31.3%, p = 0.003). The positive interval change group had higher rates of bifurcation lesions and received more percutaneous coronary intervention. In multivariate analysis, age, interval change of serial hs-TnT and diabetes mellitus were independent predictors of significant coronary artery stenosis.Conclusions:This study identified a relationship between the serial change in cardiac biomarkers and the presence of significant coronary stenosis in patients with ACS. Serial hs-TnT change was associated with real angiographic stenosis in patients with ACS.
# Introduction
In South Korea, the incidence of acute coronary syndrome (ACS), including acute myocardial infarction (AMI), is increasing gradually and becoming a serious socioeconomic problem. In the emergency setting, an early diagnosis of myocardial infarction (MI) is advantageous for patients and the physicians treating them . Therefore, several rapid diagnosis methods have been devised, and cardiac biomarkers are considered the most important . Cardiac troponin is the biomarker of choice for diagnosing myocardial necrosis, as it is www.kjim.org https://doi.org/10.3904/kjim.2020.090 considered the most sensitive and specific biomarker of myocardial injury. High-sensitivity troponin (hs-Tn) has emerged for the assessment of acute chest pain and the improved sensitivity allows the identification of practically any amount of myocardial necrosis. The marked evolution of these cardiac biomarker assays has reduced the time interval to the second Tn assessment. If hs-Tn is available, a rapid 'rule-out' protocol (3-hour hs-Tn) is recommended. Furthermore, Reichlin et al.introduced a 1-hour 'rule-out' algorithm as a faster diagnostic tool than the 3-hour 'ruleout' protocol.
However, no studies have examined a rapid 'rule-out' algorithm using high-sensitivity troponin T (hs-TnT) in Korean patients with ACS. In addition, there is little information on the prediction of significant coronary artery stenosis using hs-TnT. We hypothesized that if the change in myocardial enzymes is significant, there would be a high likelihood that significant coronary artery stenosis is present in patients with chest pain and suspected non-ST-segment elevation ACS (NSTE-ACS).
# Methods
## Study population
From April 2014 to May 2018, 484 patients presented to our tertiary center with acute chest pain suspected of being ACS, which developed within 6 hours. The initial hs-TnT (or 0-hour hs-TnT) was tested routinely. To estimate the interval change in cardiac enzymes, hs-TnT was re-checked 3 hours after presentation and defined as the 3-hour hs-TnT. The study flowchart is shown in. The following patients were excluded from the study: (1) those with duration of chest pain > 6 hours; (2) those with possible 'rule-in' MI; (3) those with medical conditions that could affect Tn levels (e.g., systemic infection, acute cerebrovascular accident, tachyarrhythmia, renal disease, heart failure, or other acute medical conditions rather than ACS); and (4) those who previously underwent an interventional cardiac procedure. Patients with a 0-hour hs-TnT ≥ 0.1 ng/mL or who presented with electrocardiogram findings suggesting ST-segment elevation were excluded because they were identified as a 'rule-in' MI. Patients with a history of coronary artery disease, previous percutaneous coronary intervention (PCI), or previous coronary artery bypass graft were also excluded. The patients who required cardiopulmonary resuscitation during admission, were in cardiogenic shock, or had chronic heart failure were excluded because these conditions may affect the initial Tn levels.
Patients who did not undergo coronary angiography (CAG) were excluded. Ultimately, the study enrolled 163 patients. These were subdivided into hs-cTnT positive (n = 32) and negative (n = 131) interval change groups. This study was a prospective, non-randomized, observational study. The study protocol was approved by the Chonnam National University Hospital Institutional Review Board (IRB No. CNUH-2014-124). Informed consent was obtained from all study participants.
## Study definitions and endpoint
Blood samples were collected in serum-separating tubes. Samples were obtained from 484 patients immediately after emergency room visits. The hs-TnT level was measured by Roche e411 (Roche Diagnostics, Rotkreuz, Switzerland). An interval change in serial hs-TnT was defined as follows: (1) 3-hour hs-TnT ≥ 0.014 ng/mL and more than 50% greater than 0-hour hs-TnT in patients with a 0-hour hs-TnT < 0.014 ng/mL; or (2) 3-hour hs-TnT ≥ 0.014 ng/mL and more than 20% greater than 0-hour hs-TnT in patients with a 0-hour hs-TnT ≥ 0.014 ng/ mL. The current cutoff value for the hs-TnT assay is 0.014 ng/mL. Although the 2015 European Society of Cardiology (ESC) guidelines recommend invasive management for patients with acute chest pain if 3-hour hs-TnT is greater than 0.014 ng/mL, we defined positive or negative interval change groups according to an older guideline that subdivided patients with 3-hour hs-TnT ≥ 0.014 ng/mL into two groups. The clinical manifestations of NSTE-ACS reference the 2015 ESC guidelines, as follows: (1) prolonged (over 20 minutes) angina at rest; (2) new-onset class II or III angina of the Canadian Cardiovascular Society (CCS) classification; (3) recent destabilization of previously stable angina with at least CCS class III angina (crescendo angina); and (4) post-MI angina. The term non-ST-segment elevation myocardial infarction (NSTEMI) was used when there was evidence of myocardial necrosis (defined as an elevated troponin value with at least one value above the 99th percentile upper reference limit) and the aforementioned clinical features. In this study, hs-TnT was used as a cardiac biomarker to determine whether the subjects had unstable angina or MI. In this study, significant coronary artery stenosis was confirmed by CAG and defined as follows: (1) ≥ 50% diameter stenosis identified in the left main coronary artery at CAG; or (2) ≥ 70% diameter stenosis identified in coronary arteries other than the left main coronary artery at CAG. Multi-vessel disease was determined when there were significant coronary stenoses in two or more epicardial coronary arteries. A bifurcation lesion was defined as narrowing of a coronary artery adjacent to or involving the origin of a significant side branch. Vasospasm was defined as a > 90% reduction in diameter on the ergonovine provocation test or a spontaneous, dynamic > 90% decrease in diameter observed during CAG. The medical and social histories were obtained using a questionnaire. The conventional cardiac risk factors included were hypertension, diabetes, dyslipidemia, stroke, and smoking status. Body data such as height (m) and weight (kg) were measured by experienced nurses and body mass index (BMI, kg/m 2 ) was calculated. Blood was tested in all subjects on admission, including white blood cell count, hemoglobin level, platelet count, serum glucose, serum creatinine, and hs-TnT. The hs-TnT was measured at presentation and 3 hours later. All study subjects underwent echocardiography and CAG. The left ventricular ejection fraction was considered an indicator of ventricular systolic function on transthoracic echocardiography. Patients who had an interval change underwent CAG between admission and hospital discharge, while patients who were negative for an interval change underwent CAG within 1 to 2 months of admission via a hospital visit. The coronary arteries were assessed quantitatively and qualitatively using CAG. If significant coronary artery stenosis was found at CAG, PCI was performed simultaneously.
# Statistical analysis
All statistical analyzes were performed using SPSS for Windows version 25.0 (IBM Co., Armonk, NY, USA). Continuous variables are expressed as the mean ± standard deviation. Discrete (categorical) variables are expressed as percentages and counts. Student's t test was used to analyze the difference between two independent groups of continuous variables. In the analysis of categorical variables, the difference between the observed www.kjim.org https://doi.org/10.3904/kjim.2020.090 and expected frequencies was confirmed by Pearson's chi-square test. To demonstrate the impact of the interval change in hs-TnT on the presence of significant coronary artery stenosis, the odds ratio (OR) and 95% confidence interval (CI) were calculated using univariable and multivariable logistic regression analysis. The ORs were adjusted according to the variables that had statistical significance (p < 0.05) in the univariable analysis in the multivariable model.compares the baseline clinical characteristics. There was no significant difference between the two groups in age (60.09 ± 10.40 years vs. 57.79 ± 11.74 years, p = 0.312), proportion ≥ 65 years old (34.4% vs. 28.2%, p = 0.495), or proportion of males (71.9% vs. 65.6%, p = 0.502). The positive interval change group had higher 0-hour hs-TnT level (0.020 ± 0.014 ng/mL vs. 0.011 ± 0.008 ng/mL, p = 0.001), proportion of 0-hour hs-TnT ≥ 0.014 ng/mL (56.3% vs. 16.0%, p = 0.001), and 3-hour hs-TnT level (0.150 ± 0.206 ng/mL vs. 0.010 ± 0.007 ng/mL, p = 0.050) than the negative interval change group. The laboratory tests did not differ significantly, except for the serum glucose level (151.90 ± 57.92 ng/mL vs. 115.07 ± 29.68 mg/dL, p = 0.031). There was no difference in conventional cardiac risk factors such as hypertension, diabetes mellitus, and dyslipidemia in both groups.summarizes the angiographic findings of both groups. There was a marked difference in the presence
# Results
## Baseline characteristics
## Coronary angiographic characteristics
## Independent predictors of significant coronary artery stenosis
The univariable and multivariable analyses of predictors of significant coronary stenosis are described in, respectively. In the univariable analyses, age, Values are presented as number (%) or mean ± SD. TIMI, thrombolysis in myocardial infarction; PCI, percutaneous coronary intervention. 0-hour hs-TnT, 3-hour hs-TnT, interval change of serial hs-TnT, serum glucose level and history of diabetes mellitus were associated with significant coronary stenosis. Multivariable analysis was performed on these significant variables, and age (adjusted OR, 1.054; 95% CI, 1.013 to 1.097; p = 0.024), interval change of serial hs-TnT (adjusted OR, 10.196; 95% CI, 3.571 to 29.111; p = 0.001), and history of diabetes mellitus (adjusted OR ,3.390; 95% CI, 1.216 to 9.451; p = 0.013) were identified as independent predictors of the presence of significant coronary artery stenosis.
# Discussion
This study examined the association between significant coronary stenosis and the interval change of serial cardiac biomarkers. A higher proportion of significant stenosis was observed in the positive interval change group. In the positive group, the 0 and 3-hour hs-TnT were higher than in the negative group. As there was a higher rate of significant coronary stenosis in the positive group, the proportion undergoing PCI was also higher in that group. Multivariable analysis of the risk factors for significant coronary stenosis demonstrated that the interval change in serial cardiac biomarkers was the most important predictor of severe artery coronary stenosis angiographically. The advent of the hs-TnT assay has led to the development of a method for the rapid assessment of 'rule-out' and 'rule-in' AMI. Many studies have examined hs-TnT worldwide. Recent guidelines mention the early 'rule-out' method for AMI. This method is based on the hs-TnT assay and sampling at 0 and 1 hour. This algorithm includes two conditions to rule out MI: (1) hs-Tn is a continuous variable and the probability of MI increases with hs-Tn; and (2) early absolute changes in the level within 1 hour can be interpreted as surrogates of absolute changes over 3 or 6 hours and add incremental diagnostic value to the assessment of cardiac troponin at admission. This 'rule-out' algorithm should always be integrated with a detailed clinical assessment. It is also mandatory to evaluate the 12-lead electrocardiogram and repeat blood sampling including Tn for all cases of ongoing or recurrent angina.
We hypothesized that cardiac troponin, especially hs-TnT, can predict the presence of angiographic significant coronary stenosis. In many cases of MI, actual significant coronary stenosis may not be present. CAG is needed to determine whether stenosis exists. All NSTE-ACS cases that have a rise or fall in Tn compatible with MI are classified as high-risk patients, and should undergo CAG within 24 hours, according to the 2015 ESC guidelines. All STEMI cases should undergo primary PCI within 12 hours, according to the 2017 ESC guidelines. Indeed, the greater the likelihood of having significant coronary stenosis is, the greater the need for early revascularization.
A few studies describe the relationship between cardiac biomarkers and significant coronary stenosis. von Jeinsen et al.noted that cardiac troponin-I is useful for diagnosing significant obstructive coronary artery disease in the hemodynamically stable setting of suspected AMI with special electrocardiographic findings, such as bundle branch block or a wide QRS complex. Sanchis et al.measured hs-TnT at the time of presentation and after 6 hours in patients presenting with non-STE acute chest pain and elevated hs-TnT, estimated the maximum hs-cTnT and interval change in hs-TnT, and optimized the cut-off values. They noted that low maximum and low interval changes in hs-TnT were associated with angiographically proven no-stenosis.
In this study, we selected 3-hour hs-TnT as the reference value for a change in cardiac biomarkers. The 2015 ESC guidelines recommend applying the 0/3-hour 'rule-out' algorithm of NSTE-ACS using the hs-TnT assay. In accordance with prior studies regarding the hs-TnT assay, we set the upper normal limit of 3-hour hs-cTnT as 0.014 ng/mL. We included NSTEMI with hs-TnT of at least 0.14 ng/mL and at most 0.1 ng/ mL. In such patients, it may be difficult to predict the presence of significant coronary stenosis, so these may be considered 'gray zone' patients clinically. In these 'gray zone' cases, we found that the interval change in hs-TnT predicts some coronary stenotic lesions before CAG. This supports the evidence for performing CAG. As expected, both 0 and 3-hour hs-TnT were significantly higher in the positive interval change group. Interestingly, the proportion of patients with a significant vasospasm was statistically higher in the negative group. In other words, if ACS is suspected, but the serial changes in cardiac markers are not significant, vasospastic angina may contribute to the clinical situation. Vasospastic angina was diagnosed when spontaneous spasm or ergonovine-induced spasm occurred. If no spontaneous spasm occurred, the ergonovine provocation test was only performed in patients with clinical suspicion of vasospastic angina. However, the prevalence of vasospasm in the negative hs-TnT group was relatively high (6.3% vs. 31.3%, p = 0.003). Therefore, routine vasospasm provocation test can be considered in patients with moderately elevated hs-TnT. The 1-hour 'rule-in' algorithm was introduced in the 2015 ESC guidelines. However, it was not available in our institution at the time of this study. Our study had several limitations. First, there is no expert consensus on how to determine the interval change in cardiac biomarkers. Many studies define the interval change as the arithmetic difference between the 0-hour cardiac 0 and subsequent estimated Tn level. However, we determined the interval change using a slightly different method. Second, the symptom duration might have affected study outcomes. Patients with long symptom duration likely had pre-conditioned myocardium and rich collateral flows in coronary artery circulation, and these aspects might have influenced the study results. Although patients in the current study had symptom durations of less than 6 hours, there was no exact information regarding symptom duration in the case report form. Third, there was some diversity among the underlying clinical characteristics. Fourth, the study was conducted in a single medical center with a small sample size. Therefore, caution is needed when interpreting the results. Finally, this study did not include follow-up information on the clinical prognosis of the subjects. Therefore, large randomized trials should examine the clinical outcomes in both groups.
In conclusion, the change in serial hs-TnT was associated with real angiographic severe stenosis in patients with ACS. A multicenter study with a large number of patient samples is needed. |
Preferred Provider Organizations and Physician Fees
Preferred provider organizations (PPOs) represent a form of managed care in which providers agree to accept discounted fees in exchange for the expectation that their patient volume will increase or at least be maintained. Managed care plans that rely on discounted fee-for-service (FFS) payments have increased from about 10 plans in 1981 to over 700 plans in 1994. In this study, we document levels of discounts achieved by two large national insurers and discuss how the size of the discount varies by type of service and how the discounted rates relate to Medicare fees. Our results show that, despite achieving large discounts (approximately 10-20 percent) relative to their indemnity plans, the two nationwide PPOs studied here pay at rates substantially above Medicare levels.
# Introduction
One of the most significant changes in the health care delivery system in the past decade has been the movement away from traditional FFS insurance to managed care arrangements. This has been largely driven by payers' demands for cost control and a growing provider surplus. Payers are becoming more discriminating consumers, searching for low-cost providers and using market power to negotiate discounted prices. Spurred by a need to maintain patient volume, physicians are increasingly accepting discounted payments.
Among the broad range of managed care arrangements that are evolving, PPOs rep-
The research presented in this article was funded by the Health Care Financing Administration (HCFA) under Contract Number 500-92-0024. The authors are with The Urban Institute. The conclusions and opinions expressed are solely those of the authors and do not necessarily reflect the views of The Urban Institute or HCFA. resent one of the fastest growing alternatives to FFS or traditional health insurance. In PPOs, an insurer or other third-party payer contracts with selected physicians, hospitals, and other providers to deliver services at discounted rates. Since financial incentives are provided for enrollees to use these preferred providers, providers are willing to accept discounted fees in exchange for the expectation that their patient volume will increase or, at least, be maintained. PPOs are an attractive type of managed care arrangement because they offer something for everyone. The payer hopes to control costs through discounted fees and utilization management controls. Enrollees save money by choosing a contracted provider, although they are free to go to other providers. Providers can potentially increase their patient loads by agreeing to discounted fees and forms of utilization review, while often avoiding capitated payments.
Contracting with a panel of providers for health care services, which occurs in PPOs, is not entirely new. In fact, the Blue Cross and Blue Shield insurance plans' participating provider program rests on a variant of PPO contracting. However, only recently has the notion of selective contracting been used extensively as a costcontainment device. The recent growth in PPOs has been remarkable. Over a decade ago, PPOs played almost no role in the health insurance market. According to one source, in 1981 fewer than 10 PPOs had contracts to serve enrollees [bib_ref] Employee Benefit Research Institute: Sources of Health Insurance and Characteristics of the..., Barger [/bib_ref]. By 1987, that number increased to over 100 plans . In 1994, there were over 700 plans. Similarly, the number of people enrolled in PPOs has also increased from an estimated 10.4 percent of individuals with private insurance in 1988 to . Physician participation in PPOs mirrors these enrollment trends-increasing from 45 percent in 1988 to . Overall, physicians receive about one-fifth of their revenue from PPO arrangements.
According to a recent study of 30 PPO plans, the predominant payment method for providers was FFS 1 [bib_ref] Employee Benefit Research Institute: Sources of Health Insurance and Characteristics of the..., Barger [/bib_ref]. In fact, none of the PPOs surveyed used capitation as a basic form of physician reimbursement. Although this study investigated provider payment methods, it did not address the level of discount payers are able to obtain from panels of preferred providers relative to what they pay for indemnity claims. The larger the size of the PPO discount the payer achieves, the greater the payer's ability to control health plan costs.
PPO discounting is not merely a privatesector issue. It may also play an important role in payment decisions among public payers. Historically, policymakers have been concerned that relatively low public fees in comparison with private payers could create access problems for beneficiaries. Their concerns, for example, have led to annual tracking of the relationship between Medicare and private fees by the Physician Payment Review Commission (PPRC). Their findings suggest that Medicare fees have been, on average, between 30 and 40 percent below private fees for most of the 1990s . However, it has been argued that, by relying on very little data from managed care plans, PPRC's results are misleading and "in many instances, Medicare has started paying more to doctors than private payers do" [bib_ref] The Medicare Relative Value Scale and Private Payers: The Potential Impact on..., Miller [/bib_ref]. This conclusion, though, appears to be based on fee data from a limited number of services and geographic areas.
If Medicare fees were, in fact, becoming more generous relative to private payers, then this could allow public payers to save money by lowering their fees without a great risk of impeding access. Alternatively, if generous private fees had been providing cross-subsidies for public (and uninsured) patients, discounting could pressure public payers to raise their fees under certain circumstances. For example, lower private payments brought about by market forces could threaten some essential providers' financial viability, causing public payers to respond by providing additional sources of revenue (e.g., a higher Medicare bonus in personnel shortage areas). Although it is not possible to predict how private discounting will affect public payers' decisions, it is important to understand how the growth in privately-discounted fees may be changing the relationship between public and private payments.
In this article, we measure PPO discounts achieved by two large private payers in 1993. For the two payers, we consider these discounts from a national perspective and explore how discounts vary across types of physician services. In addition, we contrast the discounted PPO rates with Medicare fees during the study year. Although these private payers may not be representative of all private payers, they are large and national in scope and, as such, provide a reasonable basis for making comparisons with Medicare. In the section that follows, we describe the data available to this study and outline our methods. We then present our findings and conclude with a discussion of the potential implications of these results.
# Data and methods
## Data sources
This analysis relies on data from Medicare's National Claims History System and two private-sector third-party payers. These private-sector payers include two large national insurers, one covering 4.5 million lives. 2 Both payers operate PPO and indemnity plans in all 50 States and the District of Columbia and provided data from all geographic areas. Due to data use agreements, neither payer providing data can be identified by name. Therefore, the data sources are referred to as Payer 1 and Payer 2. We recognize that data from two payers are not generalizable to all private payers. However, these payers were willing to cooperate and provide data that is rarely available publicly; we contacted other payers who were not receptive to the idea. Moreover, the insurer covering 4.5 million lives may also apply the same payment rules to many other health plans across the country. Therefore, the payments derived from this payer are likely to reflect a larger share of the private market than is represented in the claims available for us to analyze. It would require arbitrary assumptions, however, to estimate the market share of either payer in this study. [fig_ref] Table 1: Characteristics of Private Payer Data Sources [/fig_ref] provides a summary of the private payer data used in the analysis. The claims experience represented in Payer 1 reflects 3 months of data from 1993 while Payer 2 represents the entire year's data. Because we are primarily interested in estimating average payment rates and discounts, there is a sufficient amount of data in 3 months of claims experience to produce reliable estimates. Although Payer 1 data represents a much shorter time period than Payer 2, the number of indemnity 2 The other insurer would not provide data on the number of covered lives. services reported in these data bases are comparable. PPO claims in each data set are defined by whether the service was rendered by a PPO provider and not by the health plan in which the patient is enrolled. Claims for patients enrolled in PPO plans who receive services by non-PPO providers are considered indemnity services and classified accordingly. There are twice as many PPO services reported in the Payer 2 data. In total, there are over $950 million in payments for Payer 1 and $1,100 million in payments for Payer 2.
We reviewed and edited each source of data in order to develop analytic files which consisted of claims for physician services and clinical laboratory services only. This editing involved omitting claims with invalid Current Procedural Terminology (CPT) codes, claims for medical supplies, durable medical equipment, and ambulance services, and claims for oncology, dialysis, and anesthesia services. In each data set, about 10-15 percent of the total charges were dropped as a result of this process. It was possible to identify claims for surgical assistants and the professional component of radiology and other services. This is important for expressing payment rates in terms of relative value units (RVUs) (discussed later).
In addition, claims with apparently erroneous payment data in the private and Medicare sources are screened out. The objective was to remove those claims that seemed to have very high or very low average payments and, specifically, eliminate claims for partial payments (i.e., payments for surgical assistants) that are not adequately identified in the data. The approach we used eliminates all claims that were more than three times or less than one-third of the mean payment for the service. 3 In order to avoid screening out disproportionate numbers of claims from high-or low-cost areas, we first adjust for differences in payments using HCFA's Geographic Practice Cost Index (GPCI). Overall, outlier claims accounted for only about 2 percent of payments. Also, we identified some Medicare services with very low or very high payments per RVU relative to the 1993 fee schedule conversion factors. These services tended to be unusual services within the Medicare program (e.g., newborn care, obstetrical care, and antigen therapy). To address this issue, all national average Medicare payments per RVU that were less than onethird or greater than three times the fee schedule conversion factors were excluded from the computations.
# Methods
The size of the PPO discount for each payer in this study can be measured as the ratio of the average payment for services paid through the payer's PPO plans to the average payment for services paid through the indemnity plan. Analogously, the payment differential between Medicare and the PPOs equals the ratio of the average Medicare payment to the average PPO payment. To compute these ratios nationally, two issues must be addressed. First, how can payments for different services (e.g., office visits and surgeries) be expressed in terms of consistently-defined units of volume so that they can be combined into summary measures of PPO discounts or Medicare-to-PPO differentials? Second, if our goal is to focus solely on the price discount, how do we control for differences in the mix of services between PPO, indemnity, and Medicare claims when computing the price ratios?
The first issue can be addressed directly by expressing payments for individual services in terms of the number of RVUs contained in the service. RVUs, as defined by the Medicare Relative Value Scale, can be thought of as basic units of service volume that have the same meaning in all physician services. 4 For example, if an office visit contains one RVU and is paid at a rate of $45, its payment per RVU would be $45. If an arthroscopy, on the other hand, contains 25 RVUs and is paid at a rate of $1,500, its payment per RVU would be $60. By expressing these payments in terms of RVUs, we can combine the payment rates for different services into a weighted average payment rate, using the distribution of RVUs across services as the weights. 5 These weighted average payments per RVU can be computed separately for all PPO and indemnity claims, as well as for groups of services within these payment categories, allowing for fairly easy computation of PPO discounts. The aver-4 For additional information on the process used to assign RVUs to physician services, see [bib_ref] The Medicare Relative Value Scale and Private Payers: The Potential Impact on..., Miller [/bib_ref]. The reader may note that the payment per RVU for an individual service, computed as described here, is simply the conversion factor that would need to be applied to that service's RVUs to yield the current payment per service. age payment per RVU can be computed similarly for Medicare. The issue of service mix differences between PPO, indemnity, and Medicare claims can also be dealt with in a straightforward fashion. The solution relates to the weights used in computing the average payment per RVU when aggregating services. Our basic choice is to use the PPO shares of RVUs, the indemnity shares of RVUs, or the Medicare shares in all instances. Since our goal is to measure the size of the discount the PPO receives relative to the indemnity part of the payer's business, it makes sense to weight all prices by the distribution of indemnity service RVUs. In this way, the discount we compute will tell us how much lower indemnity payments might have been if their prices paid were at PPO levels. Therefore, PPO and indemnity average payments per RVU and the resulting discounts are computed as if the indemnity service mix prevailed within the PPO claims. Once the indemnity service mix is selected as the basis for computing the PPO discount, it makes sense to be consistent and use it in deriving the Medicare-to-PPO differential. This is done separately for each PPO, since the indemnity service In 1993, the average Medicare payment per RVU was not simply the published conversion factor for the Medicare fee schedule because Medicare was only in the second year of the 5-year transition to a system based fully on the Medicare Relative Value Scale. The discount we derive can be viewed as a Laspeyres price index, where the indemnity claims and prices represent the base and the PPO prices the comparison. mix varies by payer. 8 When indemnity weights are applied to Medicare payments per RVU, the resulting average payment per RVU can be thought of as the level of payment indemnity plans would have offered if they paid at Medicare rates. The differential then reflects the potential additional savings (beyond the PPO discount) that might have accrued to indemnity plans had they adopted Medicare's rates. The GPCI adjustment used in the data screening (described previously) was performed in order to remove potential distortions resulting from a disproportionate number of claims coming from either highor low-cost areas. For similar reasons, the GPCI-adjusted payment rates were used in computing national average payments per RVU for all payers. [fig_ref] Table 2: Average PPO Payments per RVU for Broad Categories of Physician Services [/fig_ref] presents weighted average payment rates per RVU for each payer's PPO and indemnity plan for All Services and four broad type of service categories. Recall that the weights used in both the PPO and indemnity calculations represent the service mix of indemnity claims. First, the results show sizeable differences in indem- The resulting differences in the average Medicare payment per RVU across payers is small. Nevertheless, using the payer-specific service mix is appropriate in order for the PPO discounts and Medicare-to-PPO differentials to reflect purely price information. Some of these potential savings to indemnity plans could be offset by volume responses that could occur in response to lower fees.
# Results
HEALTH CARE FINANCING REVIEW/Spring 1996/Volume 17, Number 3 nity payment rates across the two payers. For Payer 1, the average payment per RVU for all indemnity claims is $58.66, 17.2 percent higher than the average indemnity payment per RVU of $50.05 for Payer 2. In all services categories indemnity payments for Payer 1 exceed those for Payer 2 (although, for imaging the difference is negligible). Second, as a result of the discounts each payer is able to achieve, PPO payment levels for Payer 1 and Payer 2 are fairly close to each other, with average weighted payments per RVU for All Services of about $46 and $44, respectively. For these two payers, this new form of managed care seems to result in a reduction in rate differentials across payers. Data from a broader set of payers would be required before more general conclusions can be reached.
Taken together, the indemnity and PPO payment rates show that Payer 1 is able to achieve a higher and more uniform discount across the broad service categories shown than Payer 2. For example, for All Services, the discount achieved by Payer 1 is nearly twice the size of the discount achieved by Payer 2 (20.8 percent versus 11.5 percent). Further, the payment discount for Payer 1 varies only slightly by broad type of service category (17.3 percent for imaging services to 22.1 percent for evaluation and management services), while for Payer 2 the level of the discount ranges from 8.0 and 16.0 percent. It may be that, because Payer 1 is paying substantially more for services among its indemnity claims, it has more room to negotiate larger discounts. However, other factors, such as differences in the two payers market shares may also be playing an important role in determining these discounts. [fig_ref] Table 3: Average PPO and Indemnity Payments per RVU for Detailed Categories of Physician... [/fig_ref] shows the size of the differentials between PPO payments and indemnity payments for each payer by more detailed type of service categories. To some extent, the size of the discount for each payer varies within each of the broad service categories. These varying discounts are most pronounced in the Evaluation and Management (E&M) service category. They range from 8.7 percent for emergency visits to 26.5 percent for office visits for Payer 1 and 2.1 percent for nursing home visits to 11.7 percent for hospital visits for Payer 2.
The E&M category also has the greatest difference in discounts between Payer 1 and Payer 2. For example, in the office visit category, Payer 1 is able to achieve a 26.5 percent discount while the discount achieved by Payer 2 is only 8.6 percent In fact, the PPO discounts for Payer 1 are so large among E&M services that, despite having higher indemnity rates than Payer 2 in all but one of the E&M categories, Payer 1 actually has lower PPO rates than Payer 2 in five of the six service groups. For example, the average indemnity payment per RVU for office visits is about $42 for Payer 1 and $41 for Payer 2, but drops to $31 and $37, respectively, among PPO claims. The only E&M service category where the opposite finding is true is emergency room visits, where discounts are comparable and average PPO rates for Payer 1 are about 23 percent higher than those for Payer 2.
Payer 1 is also able to negotiate lower PPO rates than Payer 2 in the imaging category. Similar to the E&M category, Payer 1 starts with higher indemnity payments than Payer 2 in three out of the four imaging categories, but achieves larger discounts. For instance, the average indemnity payment rate for echographies is about $65 for Payer 1 and $63 for Payer 2 but drops to $50 and $55 for Payers 1 and 2, respectively, as a result of the discounting.
Despite the sizeable PPO discounts achieved by both payers, [fig_ref] Table 4: Average Medicare and PPO Payments per RVU for Broad Categories of Physician... [/fig_ref] shows that PPO rates are still well above those paid by Medicare in 1993. Across all services, on average, 1993 Medicare fees were 35.2 percent and 33.2 percent below the PPO rates of Payer 1 and Payer 2, respectively. These differentials are consistent with those reported by . Not surprisingly, given that the Medicare fee schedule increased payments for E&M services relative to Procedures, the differential between Medicare and PPO rates was lowest for E&M services (15.3 and 24.2 percent) and highest for Procedures. With the exception of E&M services, the Medicare-to-PPO differential is similar for both payers across service categories. For E&M, payments made by Payer 1 are closer to Medicare than are those made by Payer 2. The results in [fig_ref] Table 4: Average Medicare and PPO Payments per RVU for Broad Categories of Physician... [/fig_ref] provide a general sense of the differences between Medicare and PPO payments. However, because the computations are based on average payments per RVU, they do not show how payments for specific services vary across payers. While it would be impractical to include an exhaustive listing of Medicare and PPO prices, we have included national average payments per service for a selected set of services in [fig_ref] Table 5: Average Medicare and PPO Payments per Service for Selected Services [/fig_ref]. Our goal is to offer an alternative, and potentially more tangible, basis for comparing Medicare prices with PPO prices. Within each type of service group, the services shown were selected from among those accounting for the largest shares of spending among the indemnity side of private payers. 10 10 Generally, the Medicare payment rates are very close to those that would have been in effect had the Medicare fee schedule been fully phased-in in 1993. Differences seem to be due to the fact that in 1993, Medicare was still in the midst of its 5-year transition to payments based fully on relative values. In 1993, the three E&M payments shown are slightly below their fee schedule amounts. As the transition proceeds, it is likely that this differential between Medicare and these PPOs will be reduced. On the other hand, for a procedure such as a coronary angioplasty (CPT 92982), the fully phased-in fee schedule amount would be below the $1,181 average payment shown in [fig_ref] Table 5: Average Medicare and PPO Payments per Service for Selected Services [/fig_ref]. In 1993, its fully phased-in fee schedule amount would have been $875. Therefore, the transition could result in a potentially greater differential for this service in comparison with the two private payers used in this study. Of course, if these payers' PPO payments for angioplasties are reduced at the same rate, or a greater rate, than is reflected in the Medicare fee schedule, the Medicare payment may not lose ground to these PPOs. The payments shown in [fig_ref] Table 5: Average Medicare and PPO Payments per Service for Selected Services [/fig_ref] highlight large variations in the size of the Medicare-to-PPO differential across specific services for these payers. The size of the differential ranges from a low of 6.3 percent (established office visit for Payer 1 PPO) to a high of 57.8 percent (inguinal hernia repair for Payer 1 PPO). Within imaging services, the differentials for a two-view chest X ray (CPT 71020) and a CAT Scan of the Head (CPT 70470) are very close to the overall differential of approximately 40 percent shown in [fig_ref] Table 4: Average Medicare and PPO Payments per RVU for Broad Categories of Physician... [/fig_ref]. However, for the other two imaging services reported in [fig_ref] Table 5: Average Medicare and PPO Payments per Service for Selected Services [/fig_ref] , the differentials are smaller. In fact, Medicare's payment for a Magnetic Resonance Image (MRI) of the Brain (CPT 70553) is only 11.7 percent below the average payment rate of Payer 2's PPO.
# Discussion
If PPOs are going to thrive as a form of managed care that is able to control private spending, then they must be able to pay providers at rates well below those in traditional insurance plans. Our results show that the two payers for which we have data were able to establish heavily discounted rates within their PPOs. This suggests that, unless volume responses completely offset these discounts, purchasers covered by these payers' rules should see PPOs as an effective means of lowering their health care spending. However, the payer who was paying at the more generous level in its traditional insurance plans was able to negotiate larger PPO discounts than the low-rate payer. Though preliminary, these findings suggest that, for any given payer, the extent of the reduction in spending may be a function of their initial level of prices.
A potentially important implication of the variation in the size of PPO discounts is that the extent of heterogeneity in private payment rates across payers for physician services could gradually diminish. Based on the two payers in this study, we see a 17-percent differential in indemnity payment rates reduced to 5 percent in the PPO market. However, the process of PPO discounting did not seem to undue historical differences in payments across types of services that the Medicare fee schedule was designed to address. Despite some variations in the discounts, for both payers, indemnity and PPO payment rates for procedures, imaging services and tests were considerably higher than rates for E&M services. If the expectation is that through the discounting process fees for E&M services will increase relative to those for other services, as Medicare accomplished through the adoption of its Relative Value Scale (Levy and Borowitz, 1992), we find little evidence to suggest that has been accomplished by either of the two PPOs analyzed here. These changes taking place in private sector fees can have important implications for future developments in the policies of public payers. Decisionmakers have always been concerned that reductions in public rates could make their beneficiaries less attractive to providers and create access barriers. As PPOs grow and their lower fees become more typical of the private market, public payers may have hoped that they would have had the opportunity to lower their rates without risking serious access problems for its beneficiaries. Of course, this opportunity may not have materialized if the lower private rates reduced cross-subsidies to public payers and created demands for higher public fees. However, based on the two payers studied here, there is little reason to conclude that the gap between public and private rates is disappearing.
Based on these findings, two conclusions revelant to Medicare policy follow:
- First, simply enrolling Medicare beneficiaries in private PPOs paying at the rates reflected by the two payers in this study would not necessarily result in lower spending, on average, as a result of lower prices. If the argument that "the Medicare program could be rescued if only the Government would adopt some of the cost controls that employers have imposed on their workers under the banner of 'managed care'" is true and these two payers' prices are reasonably repre-sentative, then savings would have to come from lower rates of service utilization. Whether or not utilization controls would be acceptable to beneficiaries would be an issue policymakers would have to confront. - Second, despite arguments to the contrary [bib_ref] The Medicare Relative Value Scale and Private Payers: The Potential Impact on..., Miller [/bib_ref] , there is no evidence from this study or earlier work by PPRC to suggest that national average Medicare fees are generous or are much closer to private payer fees than they have been historically. 11 Therefore, the view that significant program savings can be achieved by reducing Medicare fees without access concerns being an issue may be overly optimistic. Although we acknowledge that these results are based on 1993 data and that the rapidly changing market may have already led to lower private fees than those observed here, a great deal of ground would have had to have been closed in order to put Medicare fees near those of the average private payer. If Medicare reduces its fees and access is not adversely affected, as has occurred at various times over the past decade, it is more likely to be due to providers' reliance on Medicare revenues than on the fact that Medicare is now the generous payer in the physician services market
[table] Table 1: Characteristics of Private Payer Data Sources: 1993 [/table]
[table] Table 2: Average PPO Payments per RVU for Broad Categories of Physician Services: 1993NOTE: PPO is preferred provider organization. SOURCE: Urban Institute analysis of 1993 claims from two large private payers. [/table]
[table] Table 3: Average PPO and Indemnity Payments per RVU for Detailed Categories of Physician Services: 1993 [/table]
[table] Table 4: Average Medicare and PPO Payments per RVU for Broad Categories of Physician Services by Private Payer: 1993 [/table]
[table] Table 5: Average Medicare and PPO Payments per Service for Selected Services: 1993 CPT Code and Service Description Office or Other Outpatient Evaluation of an Established Patient, 15 Minutes Subsequent Hospital Evaluation of a Patient, 25 Minutes Office Consultation with a New or Established Patient, 60 Minutes Contrast CAT Scan of the Head Magnetic Resonance Image of the Brain, Without Contrast Material Chest X Ray, Two Views Echography Exam of a Pregnant Uuterus Cardiovascular Stress Test Evaluation of Wheezing Sensory Nerve Conduction Study Includes performing procedure and interpretation of imaging results. Includes performing procedure and interpretation of test results. NOTE: PPO is preferred provider organization. SOURCE: Urban Institute analysis of 1993 claims from the Medicare National Claims History System and two large private payers. [/table]
[bib_ref] The Medicare Relative Value Scale and Private Payers: The Potential Impact on..., Miller [/bib_ref] |
The Complimentary Role of Methoxy-Isobutyl-Isonitrile and Hand-Held Gamma Probe in Adamantinoma
Adamantinoma is a rare locally aggressive osteolytic tumor that is found 90% of the time in the diaphysis of the tibia with the remaining lesions found in the fibula and long tubular bones. A case of adamantinoma of the tibia is presented. The added value of nuclear medicine investigations in the workup of this patient is described. A three-phase whole body 99m Tc-methylene diphosphonate bone and a whole body 99m Tc-methoxy-isobutyl-isonitrile scans were complimentary in the demarcation of viable bone tumor and the assessment of the remainder of the bone and soft tissue to exclude other sites. Intra-operative assistance with a hand-held gamma probe, guided the biopsy of the most metabolically active tumor tissue. Histology revealed a biphasic tumor composed of epithelial and fibrous components, in keeping with an adamantinoma.
## Case report
## Interesting image
A 28-year-old male patient presented with 3 months history of pain and swelling over the anterior aspect of the left tibia.
Planar X-rays revealed at least one section of confluent lytic areas with an associated pathological fracture, surrounded by cortical thickening and demarcated superiorly by an area of irregular sclerosis [ -c].
Magnetic resonance imaging demonstrated a lobulated lesion involving the midshaft of the left tibia, measuring ± 9 cm in total length. Cartilaginous features in the proximal aspect of the lesion were confirmed by areas of irregular hyperintensity. An irregular hyperintense component in relation to the lytic area and the pathological fracture was observed suggesting extraosseous extension of tumor tissue causing overlying displacement of soft tissue. The radiological features were that of an aggressive Adamantinoma.
The patient was referred to nuclear medicine to further evaluate the extent of the primary lesion and to determine other sites of possible metastases. 99m Tc-methylene diphosphonate ( 99m Tc-MDP) whole body blood pool and 3 h delayed whole body [ ] bone scan were done. There was a localized area of increased vascularity and inhomogeneous concentration of radiopharmaceutical in the left proximal tibia. Intense localized osteoblastic activity was seen in the superolateral and anterior aspect of the lesion corresponding to the fracture site, a relatively photopenic defect was seen medial and adjacent to this area with a rim of low-grade increased osteoblastic activity. The remainder of the bone scan was normal. A 99m Tc-methoxy-isobutyl-isonitrile ( 99m Tc-MIBI) whole body scan [ ] was performed 2 days after the bone scan to determine the extent of the metabolically active component of the tumor. It showed avid uptake corresponding to the photopenic area on the bone scan. The rest of the scan was normal.
Directly after the MIBI study the patient was transferred to theater for intraoperative localization, biopsy of the metabolically active tumor tissue and excision of the lesion. This was aided by a hand-held gamma probe (World of Medicine Gamma Finder) [ [fig_ref] Figure 3: Largest tissue fragment taken [/fig_ref] ]. Multiple fragments of tissue were taken for histology; the largest was 17 × 15 × 5 mm. Counts of more than 20% of background activity were considered as significant. The patient underwent en block excision of the proximal tibia with preservation of the growth plate. The defect was filled by use of a bone transplant technique and an external fixation was applied. A week post total excision of the tumor a repeat MIBI whole body scan was performed to evaluate for residual active tumor. This scan showed no abnormal uptake [ [fig_ref] 5: Moretti JL, Hauet N, Caglar M, Rebillard O, Burak Z [/fig_ref] ]. It was confirmed on histology that the tumor was removed with a clear margin. Postoperatively, the patient did well and long-term follow-up has been planned.
# Discussion
Adamantinoma is a rare locally aggressive osteolytic tumor, found 90% of the time in the diaphysis of the tibia with remaining lesions found in the fibula and long [bib_ref] Adamantinoma of long bones: Clinical, pathological and ultrastructural features, Pieterse [/bib_ref] [bib_ref] Adamantinoma: A clinicopathological review and update, Jain [/bib_ref] The accurate assessment of musculoskeletal tumors may be unreliable due to pathological fractures, hemorrhage, calcification, and inflammation on a bone scan. In 1990, Delmon-Moingeon et al., described Sestamibi as an in vivo tumor imaging agent. [bib_ref] Uptake of the cation hexakis (2-methoxyisobutylisonitrile)-technetium-99m by human carcinoma cell lines in..., Delmon-Moingeon [/bib_ref] Better delineation of tumor outlines and cellular activity is an advantage of MIBI scintigraphy, which is helpful in the evaluation of musculoskeletal tumors. No relationship could be found between tumor MIBI uptake and 99m Tc-MDP osteoblastic activity in musculoskeletal tumor's. [bib_ref] Correlation of 99mTc-sestamibi uptake with blood-pool and osseous phase 99mTc-MDP uptake in..., Ozcan [/bib_ref] Although the exact uptake mechanisms of MIBI into the myocardial and tumor cells are not well-understood, it is postulated to be related to blood flow, blood residence time, and the cellular uptake due to passive influx of the lipophilic cation, driven by the plasma and mitochondrial membrane potentials generated in living cells. Elevated potentials are directly related to metabolic state. MIBI is physiologically taken up by the salivary glands, thyroid, heart, liver, spleen, and skeletal muscle. There is physiological hepatobiliary and renal clearance. MIBI has the advantage of being readily available, easy to prepare and cost-effective.
# Conclusion
As demonstrated in this case report, molecular imaging techniques can provide complementary information to anatomical imaging techniques in staging, therapeutic monitoring and surveillance as well as additional applications that include the localization of a malignant lesion using a hand held gamma probe. The best test will be defined by local cost, availability, and expertise within a given modality, in addition to patient specific circumstances.
[fig] Figure 1, Figure 2: (a) Planar X-ray of left tibia, (b) Magnetic resonance imaging transaxial cut of proximal left tibia, (a) Whole body blood pool and whole body delayed images, (b) Whole body bone scan versus whole body methoxy-isobutylisonitrile scan b a tubular bones. Adamantinoma typically ranges from 3 to 15 cm in size, and metastases occur in approximately 15-20% of patients. Diagnosis is histological. The tumor is known to have a dominant lytic component. This tumor is insensitive to radiation and chemotherapy. [/fig]
[fig] 5: Moretti JL, Hauet N, Caglar M, Rebillard O, Burak Z. To use MIBI or not to use MIBI? That is the question when assessing tumour cells. Eur J Nucl Med Mol Imaging 2005;32:836-42. [/fig]
[fig] Figure 3: Largest tissue fragment taken [/fig]
[fig] Figure 4: Histology revealed biphasic tumor [/fig]
|
National Trends in Main Causes of Hospitalization: A Multi-Cohort Register Study of the Finnish Working-Age Population, 1976–2010
Background: The health transition theory argues that societal changes produce proportional changes in causes of disability and death. The aim of this study was to identify long-term changes in main causes of hospitalization in working-age population within a nation that has experienced considerable societal change.Methodology: National trends in all-cause hospitalization and hospitalizations for the five main diagnostic categories were investigated in the data obtained from the Finnish Hospital Discharge Register. The seven-cohort sample covered the period from 1976 to 2010 and consisted of 3,769,356 randomly selected Finnish residents, each cohort representing 25% sample of population aged 18 to 64 years.Principal Findings: Over the period of 35 years, the risk of hospitalization for cardiovascular diseases and respiratory diseases decreased. Hospitalization for musculoskeletal diseases increased whereas mental and behavioral hospitalizations slightly decreased. The risk of cancer hospitalization decreased marginally in men, whereas in women an upward trend was observed.Conclusions/Significance: A considerable health transition related to hospitalizations and a shift in the utilization of health care services of working-age men and women took place in Finland between 1976 and 2010.
# Introduction
Change in population-wide disease patterns is a major public health issue. Knowledge of the disease trends and changing burden of disease is essential for estimating the impact of primary prevention, defining public health priorities, and predicting future health care needs [bib_ref] The disabling effect of diseases: a study on trends in diseases, activity..., Hoeymans [/bib_ref]. Information on sex-specific patterns is important as different strategies may be needed for women and men. Previous studies examining sex-specific trends in hospitalization have typically focused on specific diseases or diagnostic categories, including diabetes [bib_ref] Age and sex differences in hospitalizations associated with diabetes, Lee [/bib_ref] , stroke [bib_ref] Ageand sex-specific trends in fatal incidence and hospitalized incidence of stroke in..., Lewsey [/bib_ref] , heart failure [bib_ref] Decreasing one-year mortality and hospitalization rates for heart failure in Sweden: Data..., Schaufelberger [/bib_ref] , myocardial infarction [bib_ref] National gender-specific trends in myocardial infarction hospitalization rates among patients aged 35..., Towfighi [/bib_ref] [bib_ref] Recent trends in hospitalization for acute myocardial infarction, Wang [/bib_ref] [bib_ref] 25 year trends in first time hospitalisation for acute myocardial infarction, subsequent..., Schmidt [/bib_ref] , acute coronary syndromes [bib_ref] Age-and sex-specific trends in the incidence of hospitalized acute coronary syndromes in..., Nedkoff [/bib_ref] , coronary heart disease [bib_ref] Discordant age and sex-specific trends in the incidence of a first coronary..., Briffa [/bib_ref] , peripheral artery disease [bib_ref] Sexspecific time trends in first admission to hospital for peripheral artery disease..., Inglis [/bib_ref] , asthma [bib_ref] Recent trends in asthma hospitalization and mortality in the United States, Getahun [/bib_ref] , chronic obstructive pulmonary disease [bib_ref] All Danish first-time COPD hospitalisations 2002-2008: incidence, outcome, patients, and care, Lykkegaard [/bib_ref] , and carpal tunnel syndrome [bib_ref] Incidence rates of in-hospital carpal tunnel syndrome in the general population and..., Mattioli [/bib_ref]. However, research on sex-specific national trends in all-cause hospitalization and hospitalization for main diagnostic categories is rare and information on long-term trends is lacking. In Germany, Nowossadeck [bib_ref] Population aging and hospitalization for chronic disease in Germany, Nowossadeck [/bib_ref] analyzed changing rates of hospitalization for individual diagnoses between 2000 and 2009 and found increased rates for congestive heart failure and diseases for spine and back; whereas the hospitalization rates for ischemic heart disease, cerebrovascular diseases and certain cancers decreased. The underlying trends were mainly similar for women and men. However, there were also notable differences; lung cancer hospitalizations decreased in men but sharply increased in women [bib_ref] Population aging and hospitalization for chronic disease in Germany, Nowossadeck [/bib_ref].
The individual research findings on shifts in the disease burden based on long follow-ups can be interpreted from the point of view of the health transition theory. This theory suggests that economic, social, cultural and political changes are likely to produce proportional changes in causes of disability and death [bib_ref] Elements for a theory of the health transition, Frenk [/bib_ref] [bib_ref] The health transition: trends and prospects, Mesle [/bib_ref]. It is surprising that -to the best of our best knowledge -there are no previous studies on national hospitalization trends examining longterm proportional changes in disease burden across main diagnostic groups. The present study was set out to investigate the health transition and changing burden of disease in workingage population in Finland; a country which experienced dramatic societal change between the mid-1970s and 2010. In the workingage population the societal shift was reflected in changes in the occupational structure (from agricultural and industrial occupations to service and knowledge-sector occupations) and job content (from physically demanding and chemically hazardous jobs to mentally, socially and cognitively intensive tasks), overall growth of highly-skilled work force, the emergence of late modern work organisations based on the use of information technologies and mobile networks, and considerable changes in leisure time activities [bib_ref] Economic Growth and Structural Change: A Century and a Half of Catching-up, Hjerppe [/bib_ref] [bib_ref] Työolojen kolme vuosikymmentä. Työolotutkimusten tuloksia 1977-2008 [Three decades of working conditions. Results..., Lehto [/bib_ref]. As an indicator of severe health problems and the use of medical treatment we analysed the transition in all-cause hospitalizations and hospitalizations for the main diagnostic categories from 1976 to 2010 in seven representative cohorts of working-age men and women. Hospitalization was measured by the number of discharges rather than by the total number of inpatient days.
# Materials and methods
## Study population
The total study period was divided into seven five-year time periods . These seven independent cohorts were constructed by randomly selecting 25% of the 18 to 64-year-old Finns. In total, the study population included 3,769,356 working-age adults (52% men). Data on age and sex were obtained from a population database maintained by Statistics Finland in which every Finnish resident is registered. The dates of death, where applicable, were obtained from the National Death Register kept by Statistics Finland.
## Hospitalization data
Hospitalization data were obtained from the Finnish Hospital Discharge Register that is maintained by the National Institute for Health and Welfare. The data consist of information on all cases of inpatient medical treatment in Finnish public sector hospitals. A recent systematic review showed that completeness and accuracy in this register varies from satisfactory to very good [bib_ref] Quality of the Finnish Hospital Discharge Register: a systematic review, Sund [/bib_ref]. For each individual the diagnosis data were linked to Statistics Finland records by using national identification numbers. Outcomes were all-cause hospitalization and hospitalizations for the five main diagnostic categories: Diseases of the circulatory system (International Classification of Diseases, Eighth (ICD-8) and Ninth Revision (ICD-9) codes 390-459, International Classification of Diseases, Tenth Revision (ICD-10 codes I00-I99); Mental and behavioral disorders (ICD-8 and ICD-9 codes 291-319; ICD-10 codes F04-F99); Diseases of the musculoskeletal system and connective tissue (ICD-8 and ICD-9 codes 710-739; ICD-10 codes M00-M99); Diseases of the respiratory systems (ICD-8 and ICD-9 codes 460-519; ICD-10 codes J00-J99); and Neoplasms (ICD-8 and ICD-9 codes 140-208; ICD-10 codes C00-C97)).
Since the vast majority of pregnancy-related hospitalizations relate to healthy deliveries and therefore do not indicate morbidity, we excluded pregnancy-related causes (ICD8-9 630-679 and ICD10 000-099) from the analyses.
The proportions of the hospitalizations for the five main categories of all first hospitalizations were as follows: musculoskeletal disorders 20%, cardiovascular diseases 15%, respiratory diseases 10%, mental and behavioral disorders 6%, and cancer 4%.
We followed-up the hospitalization data over a five year period in each of the seven cohorts. The follow-up began on 1st January at the beginning of each cohort and ended on the day the participant was hospitalized or died. For the rest of the participants, the follow-up period ended five years after it began, on 31st December.
All records/information was anonymized and de-identified prior to linkage and analysis. Ethical approval for the study was received from the Finnish Institute of Occupational Health.
## Statistical analyses
Age-standardized incidence rates for all-cause and the five main diagnostic categories of hospitalization were calculated separately for men and women and are expressed as the annual number of cases per 10,000 persons in each five-year cohort. For every individual within each cohort, only the first hospitalization for the given diagnostic category was included and the individual was then removed from the risk population (as in the case of death). We did not exclude individuals' subsequent hospitalizations for other diagnostic categories. To detect time trends we applied Cox regression model, which produced age-adjusted sex-stratified proportional hazard ratios (HR) with 95% confidence intervals (95% CI) for all-cause hospitalization and hospitalization for the five main diagnostic categories for each five-year cohort between 1981 and 2010 in relation to the earliest cohort . Due to the random selection method the same individual could appear in several cohorts, but such cases could not be identified and in the analyses the cohorts were assumed to be independent.
All analyses were performed using the SAS 9.2 (SAS Institute, Cary, NC, USA) software.
# Results
Altogether 1,453,190 hospitalizations in men and 1,918,936 hospitalizations in women were recorded in the study cohorts between 1976 and 2010. The mean follow-up time per cohort was 4.97 years. [fig_ref] Figure 1: The total number of hospitalizations by cohort in men and women and... [/fig_ref] presents the total number of hospitalizations by cohort in men and women and the mean age and the mean follow-up time in each cohort. Figures 2-7 present the age-standardized incidence rates per 10,000 individuals for all-cause hospitalization and the five main diagnostic categories in the seven cohorts. Compared to women, men had higher age-standardized rates of hospitalization for cardiovascular diseases and mental and behavioral disorders, whereas age-standardized incidence rates of musculoskeletal disorders and cancer were higher in women. Up until 2001-2005 men had considerably higher levels of hospitalization for respiratory diseases, but the gap has been narrowing and in 2006-2010 the difference was very small.
In men, the age-standardized incidence rate decreased in all main causes of hospitalization apart from musculoskeletal disorders where it rose from 87 in 1976-1980 to 161 in 1996-2000 and remained stable until 2006-2010 (overall increase 85%). In women the age-standardized incidence rate of hospitalizations for musculoskeletal diagnoses increased from 112 in 1976-2010 to 174 in 2006-2010 (overall increase 55%). During the same period, the incidence rate of hospitalization for cardiovascular diseases decreased from 165 to 95 (overall decrease 42%) in men and from 165 to 66 (60%) in women. In a similar way, the incidence rate for respiratory diseases decreased, from 82 to 57 (30%) in men and from 66 to 51 (23%) in women. The incidence rate of hospitalizations for mental and behavioral disorders decreased from 45 to 30 (50%) in men and from 29 to 23 (21%) in women. In men the incidence rate for cancer-related hospitalizations decreased slightly from 31 to 28 (10%), whereas in women it increased from 29 to 44 in (52%). In men, cardiovascular diseases was the most common diagnostic category at the start of the study period, but was replaced by musculoskeletal disorders in 1996-2000. In women, musculoskeletal diagnoses overtook cardiovascular diagnoses as the largest diagnostic category already in 1986-1990. [fig_ref] Table 1: Age-adjusted Proportional Hazard Ratios [/fig_ref]
# Discussion
The present study examined incidence rates and proportional risks for all-cause hospitalization and for the five main diagnostic categories in seven representative consecutive cohorts of Finnish working-age men and women between 1976 and 2010. The findings show that musculoskeletal disorders have increased their proportion of the total of all five diagnostic categories that were included in this study whilst the proportion of cardiovascular diseases has decreased. The observed overall trends were very similar in men and women with the exception of cancer: these hospitalizations marginally decreased in men but increased in women. The findings provide new evidence which specifies the health transition theory. It seems that the general transition occurs in hospitalizations when the nation moves towards the late modern societal order but the pattern of the changes is partly different for women and men. The general secular shifts and sex-specific findings are probably related to notable changes in lifestyles and in living and working conditions but at the same time they also are likely to be connected to developments in health care, health promotion, and health policies.
## Cardiovascular disease
Our findings demonstrate a declining incidence of hospitalization for cardiovascular diseases; the decrease being a slightly larger in women. Studies from other national samples show that ageadjusted coronary heart disease and cerebrovascular disease hospitalization rates have significantly decreased [bib_ref] Changes in cardiovascular hospitalization and comorbidity of heart failure in the United..., Liu [/bib_ref]. The role of widespread primary prevention and decreases in conventional risk factors have been highlighted as important contributors to these reductions; including decreases in smoking prevalence and population wide levels of systolic blood pressure as well as favorable lipid effects [bib_ref] Decline in incident coronary heart disease: why are the rates falling?, Luepker [/bib_ref] [bib_ref] Explaining the decrease in U.S. deaths from coronary disease, Ford [/bib_ref]. The widespread and increasing use of antihypertensive drugs as well as statins and other lipid-lowering drugs, has probably also contributed to these trends [bib_ref] National, regional, and global trends in systolic blood pressure since 1980: systematic..., Danaei [/bib_ref] [bib_ref] National, regional, and global trends in serum total cholesterol since 1980: systematic..., Farzadfar [/bib_ref]. However, it has been predicted that increasing prevalence of overweight and obesity may be compromising favorable trends in the risk factors in the future [bib_ref] Explaining the decrease in U.S. deaths from coronary disease, Ford [/bib_ref].
Moreover, it is well known that trends in cardiovascular diseases may be set decades before the conditions become manifest [bib_ref] Secular trends in cardiovascular risk factors: an age-period cohort analysis of 698,954..., Ulmer [/bib_ref] , and that early-life poor socio-economic conditions are associated with a higher cardiovascular risk later in life [bib_ref] Systematic review of the influence of childhood socioeconomic circumstances on risk for..., Galobardes [/bib_ref] ; an important part of this is their potential effect on development of conventional risk factors [bib_ref] Social determinants and the decline of cardiovascular diseases: understanding the links, Harper [/bib_ref]. It is therefore plausible that population-wide improvements in childhood socio-economic environments have contributed to declining cardiovascular risk. In addition, decreases in cardiovascular risk factors relate to an upward shift in educational attainment; a substantially smaller proportion of the population than before is exposed to the risk associated with low education [bib_ref] Social determinants and the decline of cardiovascular diseases: understanding the links, Harper [/bib_ref]. The contribution of educational shift has been shown to be somewhat stronger for women than for men. In addition, changes in policies, such as introduction of comprehensive smoke-free policy, may have helped to change social norms [bib_ref] Smoke-free policies and the social acceptability of smoking in Uruguay and Mexico:..., Thrasher [/bib_ref].
## Musculoskeletal disorders
Our results show a clear increase in hospitalizations for musculoskeletal disorders. Indeed, musculoskeletal disorders have become the most common reason for claiming sickness allowance in Finland. Hospitalization due to musculoskeletal disorders has also increased in Germany [bib_ref] Population aging and hospitalization for chronic disease in Germany, Nowossadeck [/bib_ref]. In the US health care utilisation for spine conditions, the prevalence of chronic impairing low back pain, and back and neck-related health expenditure have increased [bib_ref] The rising prevalence of chronic low back pain, Freburger [/bib_ref] [bib_ref] Expenditures and health status among adults with back and neck problems, Martin [/bib_ref]. In Finland it has been noted that there have been improvements in diagnostic and therapeutic methods that reveal an increasing range of musculoskeletal disorders.
Increasing rates of musculoskeletal hospitalization may be attributable to changes in the nature of working life: the shift from manual types of jobs towards jobs involving more sedentary tasks such as more office-based and computer-based jobs. It is now acknowledged that for a large number of employees, lack of physical activity -both at work and on leisure time -is a major risk factor for musculoskeletal ill health [bib_ref] Increased physical work loads in modern worka necessity for better health and..., Straker [/bib_ref]. At the same time, obesity, which is a risk factor for particularly certain musculoskeletal conditions [bib_ref] Musculoskeletal disorders associated with obesity: a biomechanical perspective, Wearing [/bib_ref] , has rapidly increased.
Work-related psychosocial stress has also been linked to hospitalizations for musculoskeletal disorders [bib_ref] The natural history and risk factors of musculoskeletal conditions resulting in disability..., Lincoln [/bib_ref]. Because of the globalisation of work, widespread information technology and a growing pressure to increase productivity and more stressful psychosocial environments and working patterns are now affecting increasing proportions of both the male and female workforce.
## Mental and behavioral disorders
The risk of psychiatric hospitalization decreased slightly. Mental disorders as a reason for sickness absence and disability pension have increased in recent decades in Finland. As in most high income countries, mental health service in Finland has been dramatically transformed by an increase in ambulatory treatment and care. However, despite decrease in the total number of hospital beds, the annual number of psychiatric in-patients has remained largely unchanged over time; this is due to the decrease in average length of psychiatric hospital stay [bib_ref] Outcomes of Nordic mental health systems: life expectancy of patients with mental..., Wahlbeck [/bib_ref].
## Respiratory diseases
Our data show that the risk of hospitalization for respiratory diseases has decreased. In a similar way, in Danish working population, the number of hospitalizations for chronic lower respiratory diseases has been reduced over time [bib_ref] Time trend in hospitalised chronic lower respiratory diseases among Danish building and..., Tüchsen [/bib_ref].
The decrease in hospitalizations both for cardiovascular and respiratory diseases can be partially explained by widespread smoke-free legislation, introduced in Finland from 1990s onwards. A recent meta-analysis showed that smoke-free legislation is associated for a lower risk of hospitalization for cardiac, cerebrovascular and respiratory diseases [bib_ref] Association between smoke-free legislation and hospitalizations for cardiac, cerebrovascular, and respiratory diseases:..., Tan [/bib_ref].
In addition, considerable regulatory measures have been taken to prevent the adverse health outcomes of hazardous substances, for example all asbestos usage has been banned and lead-exposed workers are being biologically monitored [bib_ref] Epidemiologic estimate of the proportion of fatalities related to occupational factors in..., Nurminen [/bib_ref]. Occupational inhalation exposure to most chemical agents has decreased in Finland since 1970s and chemical exposures and related disease burden are expected to further decrease in the future [bib_ref] Mä kinen I (2013) Trends of Occupational Exposure to Chemical Agents in..., Kauppinen [/bib_ref].
## Cancer
Our results showed that the risk of cancer hospitalization decreased marginally in men, whereas in women an upward trend in risk can be observed. An earlier study in the Finnish general population similarly showed that in men there have been only minor changes in the age-standardized total cancer incidence rates, whereas in women the total cancer morbidity has slightly increased since the 1950s [bib_ref] Helsinki: National Public Health Institute KTL; National Research and Development Centre for..., Teppo [/bib_ref].
A number of reasons may have contributed to increasing cancer rates in working-age women. First, the risk of breast cancer has continuously increased. The introduction of organised mammographic screening programme affects the breast cancer incidence rate in a population as the diagnosis is advanced in time [bib_ref] The influence of mammographic screening on national trends in breast cancer incidence, Møller [/bib_ref]. In Finland national mammographic screening was introduced in 1987; in that year the incidence of breast cancer increased by about one-tenth. An analysis of secular trends which corrected for the influence of screening showed that in Finland the breast cancer rates increased by 13% per 5-year period [bib_ref] The influence of mammographic screening on national trends in breast cancer incidence, Møller [/bib_ref]. Another important contributing factor for the increased breast cancer risk relates to the increased use of postmenopausal hormone therapy (HT). A study examining the HT use and breast cancer risk in Nordic countries in 1995 showed that in three other Nordic countries (Sweden, Norway and Iceland) that had had a significant drop in HT use, the increasing trends of breast cancer incidences either experienced a down-turn or the increase declined whereas in Finland, with only a small decline in HT, this was less evident [bib_ref] Postmenopausal hormone drugs and breast and colon cancer: Nordic countries, Hemminki [/bib_ref].
Second, strong and consistent increases in endometrial cancer have been reported over the last 50 years; the biggest increase has been seen in postmenopausal women but in the last decade the rates have also increased in younger women.
Third, in Finland smoking has systematically decreased in men already over several decades whereas in women smoking has started to decline only in recent years. As a consequence, the incidence of lung cancer has constantly increased in women and is projected to continue to increase.
Finally, obesity has strongly increased over the last three decades and it is well established that excess body adiposity is a risk factor for cancer development [bib_ref] Incident cancer burden attributable to excess body mass index in 30 European..., Renehan [/bib_ref]. It has also been shown that both in relative and absolute terms, obesity-related cancer is a greater problem for women than men; and endometrial, breast and colorectal cancers have been identified as priorities for research and public health measures [bib_ref] Incident cancer burden attributable to excess body mass index in 30 European..., Renehan [/bib_ref].
# Strengths and limitations
A major strength of our study was that the hospitalization data were derived from a national hospital discharge register with proven good accuracy and coverage [bib_ref] Quality of the Finnish Hospital Discharge Register: a systematic review, Sund [/bib_ref]. There was no loss to follow-up and the changes in International Classification of Diseases were recoded to correspond to the most recent classification. As far as we are aware, this was the first study to examine long-term national trends in main causes of hospitalization among working-age men and women.
Despite these strengths, it is also important to acknowledge potential limitations. First, we used hospitalization as a proxy for underlying morbidity. Hospitalization is considered to be a reliable indicator of the use of medical treatment [bib_ref] An evaluation of hospital discharge records as a tool for serious work..., Alamgir [/bib_ref]. However, the longterm shifts in the use of hospital services and proportional share of diagnoses related to hospitalizations within a population are also likely to reflect changes in treatments and options of health care. Moreover, over time changes in coding practices and diagnostic criteria may have affected hospitalization trends for some causes and proportional shares of main diagnostic categories.
Second, referral bias and differential access to hospital services might have affected the results [bib_ref] Referral bias among health workers in studies using hospitalization as a proxy..., Tüchsen [/bib_ref]. Self-referral and referral by general practitioners may be influenced by a number of factors such as geography, employment status, age, education, and economic status [bib_ref] Referral bias among health workers in studies using hospitalization as a proxy..., Tüchsen [/bib_ref]. There are significant differences in health care service provision between local authorities and significant National Trends in Causes of Hospitalization PLOS ONE | www.plosone.org socioeconomic and regional inequities in the provision of and access to health care services in Finland [bib_ref] Keskimä ki I (2013) Socioeconomic differences in mortality amenable to health care..., Mccallum [/bib_ref]. Third, we considered only the first hospitalization for any of the five diagnostic categories and excluded any subsequent hospitalizations for the same category.
# Conclusions
The present study demonstrates that there have been significant secular shifts in main causes of hospitalization and consequently the content of disease burden requiring hospital treatment has considerably changed among the Finnish working-age population between 1976 and 2010. This study provides new evidence of the health transition in a nation that has moved towards the late modern spheres of life within a short period of time. Generally the main trend both in men and women was away from cardiovascular and respiratory diseases and towards musculoskeletal disorders. Interestingly, cancer treatments in hospitals decreased marginally in men but clearly increased in women. The exact reasons for the observed trends are not known. However, it is likely that two types of factors -health interventions and socio-economic and lifestyle changes -have affected these trends. Health interventions include better control of many cardiovascular risk factors and advances in medical treatments. Socio-economic and lifestyle changes beyond the health sector include changes in working life and in the occupational distribution of the labor force, higher educational levels, increased social mobility, increased use of digital technology, more sedentary work and lifestyles, and increasing obesity rates. The present findings can have significant policy implications. Awareness of sex-specific long-term temporal changes in main causes of hospitalizations can help to identify public health priorities and guide health care planning.
[fig] Figure 1: The total number of hospitalizations by cohort in men and women and the mean age and the mean follow-up time in each cohort, Finland 1976-2010. doi:10.1371/journal.pone.0112314.g001 [/fig]
[fig] Figure 2: Age-standardized incidence rates per 10,000 individuals for all-cause hospitalizations in men and women, Finland 1976-2010. doi:10.1371/journal.pone.0112314.g002 [/fig]
[fig] Figure 3: Age-standardized incidence rates per 10,000 individuals for hospitalizations for cardiovascular disease in men and women, Finland 1976-2010. doi:10.1371/journal.pone.0112314.g003 [/fig]
[fig] Figure 4: Age-standardized incidence rates per 10,000 individuals for hospitalizations for musculoskeletal disorders in men and women, Finland 1976-2010. doi:10.1371/journal.pone.0112314.g004 [/fig]
[fig] Figure 5: Age-standardized incidence rates per 10,000 individuals for hospitalizations for mental and behavioral disorders in men and women, Finland 1976-2010. doi:10.1371/journal.pone.0112314.g005 [/fig]
[fig] Figure 6: Age-standardized incidence rates per 10,000 individuals for hospitalizations for respiratory diseases in men and women, Finland 1976-2010. doi:10.1371/journal.pone.0112314.g006 National Trends in Causes of Hospitalization PLOS ONE | www.plosone.org [/fig]
[fig] Figure 7: Age-standardized incidence rates per 10,000 individuals for hospitalizations for cancers in men and women, Finland 1976-2010. doi:10.1371/journal.pone.0112314.g007 National Trends in Causes of Hospitalization PLOS ONE | www.plosone.org [/fig]
[table] Table 1: Age-adjusted Proportional Hazard Ratios (HR) and their 95% Confidence Intervals for Hospitalization in Relation to the First Cohort in Seven Consecutive Cohorts, Finland, 1976-2010. doi:10.1371/journal.pone.0112314.t001 [/table]
|
The influence of cardiac output on propofol and fentanyl pharmacokinetics and pharmacodynamics in patients undergoing abdominal aortic surgery
Cardiac output (CO) is expected to affect elimination and distribution of highly extracted and perfusion rate-limited drugs. This work was undertaken to quantify the effect of CO measured by the pulse pressure method on pharmacokinetics and pharmacodynamics of propofol and fentanyl administrated during total intravenous anesthesia (TIVA). The data were obtained from 22 ASA III patients undergoing abdominal aortic surgery. Propofol was administered via target-controlled infusion system (Diprifusor) and fentanyl was administered at a dose of 2-3 lg/kg each time analgesia appeared to be inadequate. Hemodynamic measurements as well as bispectral index were monitored and recorded throughout the surgery. Data analysis was performed by using a non-linear mixed-effect population modeling (NONMEM 7.4 software). Three compartment models that incorporated blood flows as parameters were used to describe propofol and fentanyl pharmacokinetics. The delay of the anesthetic effect, with respect to plasma concentrations, was described using a biophase (effect) compartment. The bispectral index was linked to the propofol and fentanyl effect site concentrations through a synergistic E max model. An empirical linear model was used to describe CO changes observed during the surgery. Cardiac output was identified as an important predictor of propofol and fentanyl pharmacokinetics. Consequently, it affected the depth of anesthesia and the recovery time after propofol-fentanyl TIVA infusion cessation. The model predicted (not observed) CO values correlated best with measured responses. Patients' age was identified as a covariate affecting the rate of CO changes during the anesthesia leading to age-related difference in individual patient's responses to both drugs.
# Introduction
Providing an adequate level of anaesthesia is challenging and requires careful monitoring and dose titration. Especially in the light of recent studies which have suggested Electronic supplementary material The online version of this article (https://doi.org/10.1007/s10928-020-09712-1) contains supplementary material, which is available to authorized users. that ''too deep'' anesthesia may increase long-term postoperative mortality in cardiac surgery patients [bib_ref] Pick up the pieces depth of anesthesia and long-term mortality, Kalkman [/bib_ref] [bib_ref] Depth of anesthesia measured by bispectral index and postoperative mortality: a meta-analysis..., Liu [/bib_ref]. Therefore understanding the pharmacokinetics and pharmacodynamics (PK/PD) of drugs used in anesthesia is crucial, especially for rarely studied groups of patients or conditions.
Propofol is a short-acting hypnotic widely used for induction and maintenance of general anesthesia as well as for postoperative sedation in patients undergoing abdominal aortic surgery [bib_ref] Changes in drug plasma concentrations of an extensively bound and highly extracted..., Hiraoka [/bib_ref] [bib_ref] Pilot study on the influence of liver blood flow and cardiac output..., Peeters [/bib_ref] [bib_ref] A physiological model of induction of anaesthesia with propofol in sheep. 1...., Upton [/bib_ref]. Different pharmacokinetic (PK) models of propofol have been presented for healthy patients, critically ill patients, as well as for animals [bib_ref] A physiological model of induction of anaesthesia with propofol in sheep. 1...., Upton [/bib_ref] [bib_ref] A two-compartment effect site model describes the bispectral index after different rates..., Bjornsson [/bib_ref] [bib_ref] Population pharmacokinetics of propofol: a multicenter study, Schuttler [/bib_ref] [bib_ref] Advances in propofol pharmacokinetics and pharmacodynamics, Shafer [/bib_ref] [bib_ref] Assessing circadian rhythms in propofol PK and PD during prolonged infusion in..., Bienert [/bib_ref] [bib_ref] Population pharmacokinetic and pharmacodynamic modeling of propofol for long-term sedation in critically..., Knibbe [/bib_ref] [bib_ref] Disease severity is a major determinant for the pharmacodynamics of propofol in..., Peeters [/bib_ref] [bib_ref] Human physiologically based pharmacokinetic model for propofol, Levitt [/bib_ref] [bib_ref] A physiologically based, recirculatory model of the kinetics and dynamics of propofol..., Upton [/bib_ref] [bib_ref] A model of the kinetics and dynamics of induction of anaesthesia in..., Upton [/bib_ref] [bib_ref] Pharmacokinetic model driven infusion of propofol in children, Marsh [/bib_ref] [bib_ref] The influence of method of administration and covariates on the pharmacokinetics of..., Schnider [/bib_ref]. The 3-compartmental models published by Marsh [bib_ref] Pharmacokinetic model driven infusion of propofol in children, Marsh [/bib_ref] and Schnider [bib_ref] The influence of method of administration and covariates on the pharmacokinetics of..., Schnider [/bib_ref] are incorporated into the targetcontrolled infusion system (TCI) and serve as a guide for propofol administration. These models were developed based on healthy adults without any additional drug coadministrated. However, propofol is usually combined with an opioid drug to ensure adequate analgesia during total intravenous anesthesia. As a consequence various PK/PD interactions can occur [bib_ref] Pharmacokinetics and pharmacodynamics of propofol during propofol-alfentanil and propofol-remifentanil total intravenous anaesthesia..., Bienert [/bib_ref] [bib_ref] Mixed-effects modeling of the influence of alfentanil on propofol pharmacokinetics, Mertens [/bib_ref] [bib_ref] Hemodynamic effects of propofol: data from over 25,000 patients, Hug [/bib_ref]. For molecules with high hepatic extraction ratio, such as propofol and fentanyl, it is expected that changes in cardiac output (CO) influence their elimination clearance by affecting liver blood flow. Also, the distribution rate of drugs exhibiting perfusion rate-limited distribution, such as propofol and fentanyl, is expected to be affected by tissue blood flow [bib_ref] A physiologically based, recirculatory model of the kinetics and dynamics of propofol..., Upton [/bib_ref] [bib_ref] Cardiac output is a determinant of the initial concentrations of propofol after..., Upton [/bib_ref] [bib_ref] Influence of cardiac output on the pharmacokinetics of sufentanil in anesthetized pigs, Birkholz [/bib_ref] [bib_ref] Aortic crossclamping and reperfusion in pigs reduces microvascular oxygenation by altered systemic..., Siegemund [/bib_ref]. This mechanism was confirmed for several drugs used in anesthesia [bib_ref] Cardiac output is a determinant of the initial concentrations of propofol after..., Upton [/bib_ref] [bib_ref] Influence of cardiac output on the pharmacokinetics of sufentanil in anesthetized pigs, Birkholz [/bib_ref] [bib_ref] Pharmacokinetics and pharmacodynamics of propofol and fentanyl in patients undergoing abdominal aortic..., Wiczling [/bib_ref] and for some of them the impact of cardiac output values on the adequate dosing scheme of anesthetic drugs was confirmed under experimental conditions [bib_ref] Influence of cardiac output on the pharmacokinetics of sufentanil in anesthetized pigs, Birkholz [/bib_ref]. However, there is still very little clinical data to support dose-adjustments based on CO measurements. Upton et al [bib_ref] Cardiac output is a determinant of the initial concentrations of propofol after..., Upton [/bib_ref] have shown that after a short infusion of propofol, its initial concentrations are determined by cardiac output. On the other hand, in a pilot study by Peeters et al. [bib_ref] Pilot study on the influence of liver blood flow and cardiac output..., Peeters [/bib_ref] no significant relationship between measured CO and propofol clearance in ICU patients was observed. Therefore, it is still an open question as to whether CO may really by useful to predict the concentrations of propofol or fentanyl. Cardiovascular surgery ensures specific conditions which may be useful to study the influence of CO on propofol/fentanyl PK/PD, as it is characterized by changes in CO [bib_ref] Perioperative use of propofol for cardiac surgery, Jain [/bib_ref] [bib_ref] Vascular surgery critical care. perioperative cardiac optimization to improve survival, Venkataraman [/bib_ref] [bib_ref] Intraoperative management of aortic aneurysm surgery, Shine [/bib_ref] [bib_ref] Effects of hypothermia on the disposition of morphine, midazolam, fentanyl, and propofol..., Bjelland [/bib_ref]. Today, the minimally invasive methods of the cardiac output monitoring have become more popular in the clinical practice [bib_ref] Cardiac output monitoring: an integrative perspective, Alhashemi [/bib_ref] especially during cardiovascular surgeries, thus continuous CO monitoring might be useful in understanding PK/PD of propofol and fentanyl and consequently in guiding adequate drugs' dosing.
The aim of this work was to build a PK/PD model for propofol-fentanyl TIVA in patients undergoing vascular surgeries. The model included CO as a dependent variable and included a relationship between CO and the PK parameters of both drugs. Further we examined whether CO measurements by minimally invasive pulse pressure method could be useful for clinical decision regarding the propofol and fentanyl dosing.
# Methods
## Patients
After the approval from the local Research Ethics Committee and written informed consent, 22 patients undergoing major aortic surgery, classified as ASA III according to the American Society of Anesthesiologists (ASA) physical status classification system, were enrolled in the study. We analyzed data collected in two studies in which propofol-opioid TIVA was used during major aortic surgery. In both of them CO was continuously measured during anesthesia and the same surgical and anesthetic procedures were applied as well. The only difference was related to more frequent measurement of CO, as well as propofol and fentanyl concentrations in the second study. The data from Study 1 were obtained from a previous publication published by our group [bib_ref] Perioperative use of propofol for cardiac surgery, Jain [/bib_ref] , nevertheless due to sparse data, we were unable to assess the influence of CO on the PK/PD of propofol and fentanyl. As clinical conditions of the studies were identical, we pooled all the data to better characterize propofol and fentanyl pharmacokinetics in these patients.
The exclusion criteria in both studies were: previous cardiac surgery, ejection fraction \ 40%, valvular heart disease and myocardial infarction within 3 months prior to surgery, significant renal (serum creatinine [ 1.5 mg/dL check units) or hepatic dysfunction (aspartate and alanine transaminase [ 50% above normal level), cerebrovascular and central nervous system diseases, history of drug or alcohol abuse, morbid obesity and hearing disorders. No sedative or opioid drugs were administered before the induction of anesthesia. All surgeries were performed under propofol-fentanyl TIVA. The target-controlled infusion (TCI) system DiprifusorÒ (Astra Zeneca, UK) was used to administer propofol. In the operating room, intravenous and arterial lines were inserted under local anesthesia, and standard monitors were applied (ECG, SpO2). Hemodynamic measurements were carried out with a FloTrac/Vigileo TM System (Edwards, USA). The system consists of a sensor unit (FloTrac) and a stand-alone monitor (Vigileo). It is a pulse wave analysis technique allowing continuous cardiac output measurement. This system uses the arterial pressure waveform to measure the CO detected through a proprietary transducer (FloTrac) attached to a standard arterial line connected to the Vigileo monitor [bib_ref] Cardiac output monitoring: an integrative perspective, Alhashemi [/bib_ref] [bib_ref] Pulse waveform hemodynamic monitoring devices: recent advances and the place in goal-directed..., Hendy [/bib_ref] After the surgery, the patients were mechanically ventilated in the intensive care unit (ICU) until full recovery. Simultaneously propofol infusion was maintained until extubation. Pancuronium 0.1 mg/kg was injected to facilitate intubation and then administered as required. This study was initiated with a bolus injection of fentanyl (1.5 lg/kg). The propofol infusion started 5 min later and was maintained with intermittent injections of fentanyl (2-3 lg/kg) administered whenever inadequate analgesia was assessed throughout the surgery, ie. whenever episodes of tachycardia/hypertension in response to surgical stimuli were noted. The bispectral index (BIS; A-2000, Aspect Medical System, Newton, MA) was used to measure the depth of anesthesia and propofol dosage was adjusted to maintain the BIS level between 40 and 60. The BIS uses highly processed electroencephalographic (EEG) signals, acquired from a single self-adhesive forehead sensor, to measure the depth of sedation and hypnosis which is expressed on a unitless scale ranging from 0 to 100 (0, coma or absence of brain electrical activity; 0-40, deep hypnotic state; 40-60, general anesthesia; 60-90, deep to light sedation; and 90-100, awake). The BIS is a complex parameter composed of a combination of time domain, frequency domain and high order spectral subparameters. It is a unique quantitative electroencephalogram parameter (QEEG) which integrates several disparate descriptors of the EEG into a single variable based on a large volume of clinical data, to synthesize a combination that correlates behavioral assessments of sedation and hypnosis yet insensitive to the specific anesthetic agents chosen [bib_ref] Depth of anesthesia measured by bispectral index and postoperative mortality: a meta-analysis..., Liu [/bib_ref]. During the surgery, crystalloid and colloid fluids were infused according to the following protocol: continuous infusion of crystalloid at a rate of [10 9 body weight (kg)] ml/h, interventional colloid infusion to preserve normovolemia (stroke volume variation (SVV) \ 12) compensatory to the volume of blood loss. Arterial blood samples (3.5 mL) for plasma propofol and fentanyl concentration measurements were drawn before propofol infusion i.e. 1, 3, 5, 10, 15, 30 min after the beginning of the infusion, then every 30 min until the end of anesthesia and also after 1, 3, 5, 10, 15, 30, 60 min after the termination of propofol infusion for Study 1 and 1, 3, 5, 10, 15, 30, 60 90, 120, 240 min after the termination of propofol infusion for Study 2. The blood samples were transferred into heparinized tubes and centrifuged immediately after collection. Plasma was divided into two equal volumes. Half was stored at 4°C (propofol analysis) [bib_ref] Improved method for the determination of propofol in blood by high-performance liquid..., Plummer [/bib_ref] and another half in -70°C (fentanyl analysis). The BIS values as well as hemodynamic parameters were recorded continuously throughout the study.
# Analytical method
The propofol concentration in the plasma was measured within 8 weeks by means of high-performance liquid chromatography with fluorescence detection [bib_ref] Improved method for the determination of propofol in blood by high-performance liquid..., Plummer [/bib_ref]. The limit of quantification was estimated at 10 ng/ml. The withinday coefficients of variation were less than 10%. The fentanyl samples were measured by a validated highpressure liquid chromatography (Waters 2695 Separation Module, Milford, USA) coupled with a triple quadrupole mass spectrometer, equipped with an electrospray ionization source (ESI?) (Waters Quattro Micro, Milford, USA). The mass spectrometer operated in the multiple-ion monitoring (MRM) mode. Fentanyl and internal standard (IS) were monitored by means of the fragment ions at 387.1? 238.0 and 532.0? 219.1, respectively. The column used was a Thermo BDS Hypersil C18 100 9 2. Fentanyl and terconazole (IS) were extracted using a single-step liquid-liquid extraction (LLE) with a mixture of ethyl acetate and hexane. The lower limit of quantification was 0.05 ng/ml for fentanyl using a 0.250 ml sample volume, with a bias of 4.6% and RSD of 5.4%. The calibration curves were linear (r2 C 0.990) over the working range of 0.05-50.0 ng/ml, using 1/9 2 as a weighting factor. Quality control samples at three concentration levels (LQC 0.2 ng/ml, RSD = 9.9%; MQC 1.50 ng/ ml, RSD = 9.7%; HQC 15.0 ng/ml, RSD = 9.4%) were used for validation purposes of the analytical run.
## Model
The population nonlinear mixed-effect modelling was done using NONMEMÒ (version 7.4 ICON Development, Ellicott City, MD, USA)) and the gfortran compiler. NONMEM runs were executed using Wings for NON-MEM (WFN743, https://wfn.sourceforge.net). The FOCE estimation method with the interaction option in NON-MEM was applied. The minimum value of the NONMEM objective function (OFV), typical goodness of fit diagnostic plots, and evaluation of the precision of the PK/PD parameter and variability estimates were used to discriminate between various models during the model-building process. The NONMEM data processing, simulations, and plots were carried out using MatlabÒ Software version 7.0 (The MathWorks, Inc., Natick, MA, USA). The model predictive performance was assessed by means of Visual Predictive Checks (VPC). The VPC calculation was based on 1000 datasets simulated with the final parameter estimates. Different dosing regimens and variable infusion length required the use of prediction corrected VPC (pcVPC) [bib_ref] Prediction-corrected visual predictive checks for diagnosing nonlinear mixed-effects models, Bergstrand [/bib_ref]. The pcVPCs were created by correcting the observed and simulated values for the average population prediction in the time-bin divided by population predictions for each observed and simulated value. In this study the 10th, 50th and 90th percentile were used to summarize the data and VPC prediction. The pcVPC enables a comparison of the confidence intervals obtained from prediction with the observed data over time. If the corresponding percentile from the observed data falls outside the 95% confidence interval derived from predictions, it indicates the model misspecification. Since the PK/PD data deviated from nominal times to some extent, binning across time was used. A nonparametric bootstrap was performed to evaluate the uncertainty of final model parameters. Individual patients were randomly sampled with replacement from original dataset to form 300 new data sets with the same number of patients as original dataset. Each new dataset was fitted to the final model and all model parameters were estimated. The bootstrap empirical parameter distributions were summarized as a median with 90% (5th-95th percentile) confidence intervals.
A schematic representation of the proposed PK/PD model is given in . A three-compartment model was used to describe the PK of both propofol and fentanyl. It was parametrized using systemic clearance, distribution clearance and volumes of distributions. The delay of the anaesthetic effect, with respect to plasma concentrations, was described by an effect compartment. The bispectral index (BIS) was linked to the propofol and fentanyl effectsite concentrations (C e,P and C e,F ) through the following E max model [bib_ref] Propofol and fentanyl act additively for induction of anesthesia, Ben-Shlomo [/bib_ref] [bib_ref] Pharmacodynamic interaction between propofol and remifentanil regarding hypnosis, tolerance of laryngoscopy, bispectral..., Bouillon [/bib_ref] [bib_ref] Response surface model for anesthetic drug interactions, Minto [/bib_ref] :
[formula] BIS ¼ BIS 0 1 À E max C e;P [/formula]
where C e50,P and C e50,F denote the concentrations of propofol or fentanyl in the biophase compartment that produce half-maximal decrease in the BIS response, BIS 0 denotes the baseline BIS score (fully awake), E max is the maximal effect fixed to 1 in this work (BIS value of zero at sufficiently high concentrations of propofol or fentanyl), c is a Hill coefficient also fixed to 1, and a is a first order interaction term (a = 0 suggests additivity, a = 0 suggests nonadditivity). The additive and nonadditve model for drug interactions was explored during the model building process resulting in an estimate of the a parameter not significantly different from 0, suggesting that the interaction between propofol and fentanyl beyond additivity was not supported by the data. CO was fitted to an empirical linear equation based on the visual inspection of the data: The proposed PK/PD model of propofol, fentanyl, CO and BIS
[formula] COðtÞ ¼ CO 0 þ a CO t if t 300 min CO 0 þ a CO 300 if t [ 300 minð2Þ [/formula]
where CO 0 is a baseline CO and a CO is a linear rate of change of CO during the surgery. Due to the lack of data, the model assumed that CO was constant after 300 min. This is a very crude assumption and the model should not be extrapolated beyond 300 min, as CO likely returned to the baseline values. The CO was a priori assumed to affect the distribution and elimination clearances of both drugs. A proportional relationship was assumed (here presented for propofol clearance only):
[formula] CL P;i ðtÞ ¼ h CLp COðtÞ 6:5 expðg CLP Þ ð 3Þ [/formula]
Inter-individual variability (IIV) for all PK/PD parameters was modelled assuming log-normal distribution:
[formula] P i ¼ h P expðg P;i Þ ð 4Þ [/formula]
where P i is the set of PK/PD parameters for i th individual, h P is the population estimate of PK/PD parameters, g P,i is a random effect for a particular parameter with mean 0 and variance x P 2 . Further, any j th observation of propofol and fentanyl concentration, BIS and CO values for the i th individual, C P,obs,ij , C F,obs,ij , BIS obs,ij , and CO obs,ij measured at time t j , were defined by the following equations:
[formula] C P;obs;ij ¼ C P ðP i ; t j Þð1 þ e P;ij Þ C F;obs;ij ¼ C F ðP i ; t j Þð1 þ e F;ij Þ BIS obs;ij ¼ BISðP i ; t j Þ þ e BIS;ij CO obs;ij ¼ COðP i ; t j Þ þ e CO;ijð5Þ [/formula]
where C p , C F , BIS and CO denote the basic structural population model. P i are pharmacokinetic parameters for the i th individual, and e Pij , e Fij , e BIS,ij , e CO,ij represent the proportional or additive residual intra-individual random error. We assumed that e was symmetrically distributed around a mean of 0, with variance denoted by r 2 .
# Covariate analysis
Initially, the base model described above was compared to the model with no relationship between CO and PK parameters. For comparison, also the observed values of CO were regressed with PK parameters. Under this scenario, the missing covariates for an individual were obtained by carrying forward the last measured value. Further the classical covariate search was performed by plotting individual (post-hoc) estimates of the PK/PD parameters against covariates (weight, age) to identify their potential effects. Categorical covariates (i.e. study type)
were included into the model based on indicator variables. The covariates were added based on biological plausibility and clinical relevance. Also the statistical significance was calculated based on the difference in the minimum of the NONMEM OFV obtained for the two hierarchical models (likelihood ratio). This statistic is approximately v2 distributed and when the difference in OFV between two nested models is estimated near to 3.84 for one degree of freedom, it corresponds to p \ 0.05.
## Model simulations
The final PK/PD model with estimated fixed-and randomeffect parameters was used to simulate the concentrations of propofol and fentanyl, the BIS index and CO for an exemplary dosing schemes and patients. Context-sensitive detrimental-time (CSDT) was used to assess the influence of propofol and fentanyl administration on the time required for a decline in the effect compartment concentration upon infusion cessation in relation to patients' age [bib_ref] Context-sensitive halftime in multicompartment pharmacokinetic models for intravenous anesthetic drugs, Hughes [/bib_ref]. The CSDT is the time necessary for a certain decline in virtual effect-site concentration (and consequently increase in BIS) after termination of a continuous infusion of a given duration. This virtual effect-site concentration represents the sum of normalized effect-site concentrations of propofol and fentanyl, assuming an additive interaction between the drugs (C P,e /Ce 50,P ? C F,e /C e50,F ). The context-sensitive decrement times were simulated based on typical parameter estimates of the final PK/PD model assuming dosing scheme that leads to constant concentrations of propofol and fentanyl (equivalently certain BIS values).
# Results
The data were collected from 22 patients with the demographic characteristic presented in the [fig_ref] Table 1: Demographic characterization of patients [/fig_ref] and Tables 1S. The raw observations are presented in [fig_ref] Figure 2: The individual propofol PK, fentanyl PK, BIS and CO time profiles [/fig_ref].
## Pk/pd model
In agreement with the literature the disposition of propofol and fentanyl were described by three compartment mamillary model. The CO was assumed to be proportionally related to the distribution and elimination clearances of propofol and fentanyl. It is consistent with the assumption that both drugs are high-extraction drugs with perfusion limited distribution. The effect compartment and the additive model (Eq. 1 with a fixed to 0) was able to describe the pharmacodynamics response (BIS). The propofol-fentanyl interaction beyond additivity was not supported by the data, likely due to the limitations of the experimental design (e.g. lack of sufficient concentration range of both drugs). Different models regarding CO effects were tested during model building process as summarized in. A substantial improvement in model fit was noted when model predicted CO was included into the PK/PD model (DOFV = 83.349, df = 0). The inclusion of observed CO improved model fits slightly when compared to the model without CO effects on propofol and fentanyl PK (DOFV = 5.713, df = 0). Thus, there is a slight benefit of using a model with measured CO as a covariate. Further the addition of patients' age and study number was found to improve the model fits to the data.
The typical goodness-of-fit plots of the final PK/PD model are presented in . The individual predictions for propofol and fentanyl concentrations as well as CO and BIS values were organized around the line of identity. Also weighted residuals showed that the model was reasonably unbiased with respect to the data. The pcVPC plots for the PK/PD measurements are presented in . No major misspecifications were noted for the propofol PK, fentanyl PK, BIS values and CO values indicating agreement between the observation and model prediction. The pcVPC confirms that the model has sufficient predictive performance and can be used to simulate different clinical scenarios if one agrees to all model assumptions. The predicted and measured responses vs. time profiles for each individual are presented in Figs. 1S-4S. [fig_ref] Table 3: The parameter estimates of the final PK/PD model of propofol and fentanyl [/fig_ref] shows the final parameter estimates along with the inter-and intra-individual variability. Majority of parameters were estimated with low (lower than 50%) coefficients of variation (CV). The bootstrap confidence intervals show higher uncertainty, especially for betweensubject variability parameters. However it can be expected given the overall complexity of the model and a small number of patients included in the study. presents the relationship between the CO rate of change during the surgery and two covariates included into the final model: study and patients' age. There was a significant and consistent decrease in the rate of change of CO during the surgery (a CO ) with patients' age estimated at about 3.23% per year of age. The difference in the observed CO values between two studies was also noted. Patients in Study II had CO rate of change decreased by about 50.9% in comparison to patients in Study I. presents the dependence of the CSDT on patients' age for a propofol-fentanyl infusion that led to propofol and fentanyl biophase concentrations of 3.0 mg/l and 1.5 ng/ml for 200 min. These concentrations correspond to the BIS values of about 35. The CSDT corresponds to time needed to decrease the effect-site concertation (and consequently increase BIS) by a given percent. The CSDT increases with age of the patients. Also some betweenstudy difference was noted. As an example, BIS value of 61 (corresponding to 60% in was achieved in 41 min (Study I) and 48 min (Study II) for a 50 years of age patient, and 53 min (Study I) and 57 min (Study II) for a 80 years old patient after infusion cessation that was kept at BIS value of 35 for 200 min. The time required to achieve 80% decrease was 129 min (Study 1) or 183 min (Study 2) for a 50 years of age patient and 206 min (Study 1) and 221 min (Study 2) for a 80 years of age patient. presents the simulation using the final model of propofol and fentanyl concentrations i.e. BIS values and CO values after intravenous infusion of both drugs in relation to 50, 65 and 80 years old patient (patients are shown for Study II) for the artificial dosing protocol. The patients' age related differences can be observed in this plot. The CO increased most rapidly for the youngest subjects. This difference influences the recovery from anesthesia. Basically, the higher CO values in younger patients, the lower the propofol and fentanyl concentrations and higher the BIS values, especially at the end of surgery. shows the influence of different fentanyl doses on the responses measured in a 65-year-old patient. For the same propofol dosing scheme, the BIS values showed small dose-dependent variations around each fentanyl dose. Also different rates of BIS increase after infusion cessation were predicted.
## Simulations
# Discussion
To our best knowledge, this is the first study assessing the influence of CO on the PK/PD of propofol and fentanyl during TIVA in patients undergoing abdominal aortic surgery. To assess the usefulness of CO measurement for propofol and fentanyl dosing, we proposed a PK/PD model that took blood flow into account and used BIS as an efficacy response.
## Model structure and parameters
The influence of fentanyl on the recovery from propofolfentanyl TIVA has already been demonstrated by our group [bib_ref] Pharmacokinetics and pharmacodynamics of propofol and fentanyl in patients undergoing abdominal aortic..., Wiczling [/bib_ref] however without the inclusion of CO effects. In the present study, a three compartment model was used to describe the concentration-time profiles of both propofol and fentanyl. It clearly removes bias in model parameters introduced by using a two compartment model. As an example, the elimination clearance of propofol was estimated at 1.54 L/min, whereas in our earlier studies at 2.64 L/min and 2.22 L/min [bib_ref] Pharmacokinetics and pharmacodynamics of propofol and fentanyl in patients undergoing abdominal aortic..., Wiczling [/bib_ref] [bib_ref] Pharmacokinetics and pharmacodynamics of propofol in patients undergoing abdominal aortic surgery, Wiczling [/bib_ref]. Eleveld et al. in study based on 21 previously published data sets [bib_ref] A general purpose pharmacokinetic model for propofol, Eleveld [/bib_ref] used a three-compartment PK model for propofol PK with the elimination clearance estimated at 1.53 L/min, which is very close to our typical value of CL (1.54 L/min). Considering the distribution of propofol, a membrane barrier has been postulated to separate the slow distribution compartment [bib_ref] A physiologically based, recirculatory model of the kinetics and dynamics of propofol..., Upton [/bib_ref] , assuming flow dependent distribution to the well-perfused tissues and membrane permeability limited distribution to the second tissue compartment. In our study, for propofol the sum of peripheral compartment clearence (Q 1,P ? Q 2,P ) was smaller compared to fentanyl (2.50 L/min vs 7.26 L/min) and smaller than the CO value (6.5 L/min) therefore some contribution of permeability limited kinetics (not dependent on CO) might be expected for propofol.
The typical values of C e50 of propofol (2.25 mg/L) and fentanyl (8.77 ng/ml) are consistent with the literature data regarding PD of both drugs with the narcotic EEG effect as a PD response. For propofol C e50 , for BIS as a PD effect was estimated at 2.71-3.44 mg /L, whereas for fentanyl the C e50 with EEG power spectrum analysis is equal to 7-10 ng/ml [bib_ref] Updates of the Clinical Pharmacology of Opioids with Special Attention to Long-Acting..., Lotsch [/bib_ref].
## The influence of patients' age
The patients' age was shown to be a significant covariate associated with the rate of change of CO during the surgery. The same phenomenon was noted by Heilbrunn and Allbritten [bib_ref] Pharmacokinetic drug interaction between propofol and remifentanil?, Ludbrook [/bib_ref] who examined the cardiac output changes during and after surgical procedures. However, in our study, the time-related elevation in CO was smaller for older patients. This can be explained by the decrease in the cardiac functional reserve occurring as an effect of aging [bib_ref] Cardiac Output During and Following Surgical Operations, Heilbrunn [/bib_ref] [bib_ref] Cardiovascular physiology in the older adults, Dai [/bib_ref]. In our study, CO changes directly influenced propofol and fentanyl disposition and thus also affected the BIS values. However, due to the age-related blunting of the CO increase at the end of the surgery, older patients achieved higher propofol and fentanyl concentrations resulting in lower BIS values (Figs. 5 and 5s) which further affected the recovery profile. This is in agreement with the theoretical paradigm related to age related changes in PK, which points to the diminished liver blood flow being responsible for the slower elimination rate of highly extracted drugs [bib_ref] Cardiac anesthesia and surgery in geriatric patients: epidemiology, current surgical outcomes, and..., Castillo [/bib_ref] [bib_ref] Pharmacokinetics in older person, Cusack [/bib_ref]. However, this effect was not large within the range of infusion durations under study. As presented in , the difference between 50 and 80 years of age patients in the 60% and 80% decrement times of propofol-fentanyl infusion is estimated at about 10 or 25 min, respectively. This effect could be much more significant for infusions lasting more than 4 h, due to fentanyl accumulation. However, due to a small impact of The prediction-corrected VPC plots for the final PK/PD. The VPC plots show the simulation-based 90% confidence intervals around the 10th, 50th, and 90th percentiles of the PK data in the form of blue (50th) and gray (10th and 90th) areas. The corresponding percentiles from the prediction corrected observed data are plotted in black color Journal of Pharmacokinetics and Pharmacodynamics (2020) 47:583-596 591 fentanyl on BIS values when compared to propofol, the differences in recovery of consciousness would be less visible [bib_ref] Effects of Fentanyl, Alfentanil, Remifentanil and Sufentanil on Loss of Consciousness and..., Lysakowski [/bib_ref]. Another issue is the recovery of spontaneous breathing which is known to be affected by opioids. Unfortunately, this aspect cannot be assessed based on this data only.
## Model simulations and clinical significance
The clinical consequences of the results of this study are presented in Figs. 5S and 6S which show some examples of simulations performed based on the final PK/PD model. At the age of 50, higher CO values were achieved at the end of surgery when compared to the age of 65 and 80. This resulted in lower propofol and fentanyl concentrations, higher BIS values and as a consequence faster recovery . The differences were most visible at the recovery period which is consistent with the physiologically-based, recirculatory model of the kinetics and dynamics of propofol in a man and developed by Upton and Ludbrook [bib_ref] A physiologically based, recirculatory model of the kinetics and dynamics of propofol..., Upton [/bib_ref]. In this model, similar to our result, the changes in CO had only minor effect on the time of loss of consciousness but largely affected the time to recovery.
The proposed PK/PD model might be useful in optimizing the dosing of propofol and fentanyl during TIVA in patients undergoing abdominal aortic surgery. The model suggests that the predicted CO values could be more precise in adjusting the propofol and fentanyl doses than CO values directly measured throughout the surgery. The less significant impact of the measured CO on PK of studied drugs can be related to the fact that (a) the measured CO only approximates organ blood flow and (b) it is a measured variable that is stochastically related to the true value of liver and tissue blood flow of each subject. The inclusion of CO into the model led to a small decrease (less than about 20%) in inter-individual variability of PK parameters as presented in. It indicates rather limited clinical applicability of the final model in predicting drug dosing in comparison to a simple model (without CO effect on PK/ PD parameters). Please note that in the final model the variability in dose-rate leading to similar steady-state concentration across the subjects is a combination of interindividual variability in PK and CO parameters. The latter source of variability can be decreased by conditioning propofol and fentanyl dosing decisions on the previously measured CO in an individual subject. In principle, it could The elimination and distribution clearance of highlyextracted and perfusion limited drugs is flow-dependent. The hepato-splanchnic blood flow reaches 25-30% of the CO [bib_ref] Splanchnic and total body oxygen consumption differences in septic and injuried patients, Dahn [/bib_ref]. However, some changes in the distribution of CO during cardiovascular surgeries cannot be excluded due to the different adaptation processes [bib_ref] Aortic crossclamping and reperfusion in pigs reduces microvascular oxygenation by altered systemic..., Siegemund [/bib_ref]. Peeters et al. [bib_ref] Pilot study on the influence of liver blood flow and cardiac output..., Peeters [/bib_ref] in a pilot study, examined the influence of the CO and liver blood flow on the clearance of propofol in five critically ill patients. They noted that liver blood flow is a more predictive indicator of propofol clearance than measured CO in the studied population. It is in agreement with our study, as the model predicted CO is a better surrogate of liver blood flow than measured CO. Similarly, the animal studies confirm the significance of CO in predicting propofol concentrations during constant infusion [bib_ref] Influence of haemorrhage on the pseudo-steadystate remifentanil concentration in a swine model:..., Tassonyikurita [/bib_ref] [bib_ref] Influence of cardiac output on the pseudo-steady state remifentanil and propofol concentrations..., Kurita [/bib_ref] [bib_ref] Influence of cardiac output on plasma propofol concentrations during constant infusion in..., Kurita [/bib_ref]. In all these studies the plasma remifentanil and propofol concentrations were influenced by CO during continuous infusions, with concentrations decreasing with increased CO and increasing with decreased CO. In our study CO changes occurred as a result of inter-and intra-patient variability connected with patients' individual characteristics and clinical scenario, whereas in the animal studies by Kurita et al. [bib_ref] Influence of haemorrhage on the pseudo-steadystate remifentanil concentration in a swine model:..., Tassonyikurita [/bib_ref] [bib_ref] Influence of cardiac output on the pseudo-steady state remifentanil and propofol concentrations..., Kurita [/bib_ref] [bib_ref] Influence of cardiac output on plasma propofol concentrations during constant infusion in..., Kurita [/bib_ref] , the animals were divided into groups based on the CO maintained throughout the study. Also, it is consistent with the physiologically-based models of propofol or fentanyl [bib_ref] A physiologically based, recirculatory model of the kinetics and dynamics of propofol..., Upton [/bib_ref] [bib_ref] A physiologically-based recirculatory meta-model for nasal fentanyl in man, Upton [/bib_ref] which show CO as a main determinant of drug clearances. Another clinical implication of our study is related to the interaction between propofol and fentanyl. The interaction between propofol and opioids was studied extensively in the literature where the differences in propofol hypnosis were observed depending on the type of opioid used [bib_ref] Mixed-effects modeling of the influence of alfentanil on propofol pharmacokinetics, Mertens [/bib_ref] [bib_ref] Influence of demographic factors, basic blood test parameters and opioid type on..., Bienert [/bib_ref] [bib_ref] Remifentanil dose/electroencephalogram bispectral response during combined propofol/regional anesthesia, Koitabashi [/bib_ref]. One of the proposed reasons of such differences is the pharmacokinetic hypothesis related to the opioid-driven changes in CO. In our work we were unable to show the effects of either propofol or fentanyl on CO, instead we used an empirical relationship describing the surgery-related increase in CO.
In summary, we illustrated the widely recognized theoretical paradigm of the relationship between CO and the clearance of highly extracted drugs and distribution clearance of perfusion-rate limited drugs, under real clinical scenarios. For that purpose a PK/PD model was built to describe the relationship between propofol and fentanyl dosing, the measured CO and BIS values in patients undergoing abdominal aortic surgery. The uncertainty and the indirectness of the measured CO to liver and tissue blood flow required using a model predicted CO values as a predictor. The measured CO values were shown to be of rather limited usefulness for propofol and fentanyl doseadjustment in patients undergoing abdominal aortic surgery. The patients' age was identified as a covariate for the observed CO changes during anesthesia. Thus, patients' age can be associated with different PK profiles, depth of anesthesia and recovery profiles.
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://creativecommons. org/licenses/by/4.0/. Context-sensitive effectsite decrement times (CSDT) for propofol-fentanyl infusions showing the time for decreasing the effect-site concentrations of a given percentage (20-80%) from the maintained effect-site concentration after propofol/ fentanyl infusion cessations for subjects of different age. The solid and dashed lines corresponds to study I and study II predictions. The propofol and fentanyl biophase concentrations were kept at 3.0 mg/l and 1.5 ng/ml respectively for 200 min, which corresponds to the BIS values of about 35. CSDT of 80%, 70%, …, 20% corresponds to achieving a BIS values of 77, 68, 61, 55, [bib_ref] Influence of cardiac output on plasma propofol concentrations during constant infusion in..., Kurita [/bib_ref] [bib_ref] Effects of Fentanyl, Alfentanil, Remifentanil and Sufentanil on Loss of Consciousness and..., Lysakowski [/bib_ref] [bib_ref] Cardiovascular physiology in the older adults, Dai [/bib_ref]
[fig] 1: mm 3 lm (Thermo Scientific, Waltham, USA). The mobile phase was: formate buffer pH4.0 [A] and acetonitrile [B] (J.T. Baker, Avantor, the Netherlands). The flow rate was 0.2 ml/min, isocratic separation was applied -the mobile phase was used as follows: 70% [B] and 30% [A]. [/fig]
[fig] Figure 2: The individual propofol PK, fentanyl PK, BIS and CO time profiles. The dots represent raw measurements. They are connected with [/fig]
[table] Table 3: The parameter estimates of the final PK/PD model of propofol and fentanyl. CO , L/min 1.41 (6.0), 1.42 [1.28-1.54] lead to more precise dosing for an individual subject. However, based on model simulations, even an exact knowledge of CO changes in an individual subjects, would lead to a decrease in drug-dosing variability of at most 20%. It makes CO measurements of rather limited usefulness in guiding propofol and fentanyl dosing, despite clear casual effect of CO on clearances of both drugs. [/table]
|
Generation and Characterization of Alloantigen-Specific Regulatory T Cells For Clinical Transplant Tolerance
Donor-specific CD4 + CD127 − CD25 + FOXP3 + regulatory T cells (AgTregs) have the potential to induce clinical transplant tolerance; however, their expansion ex vivo remains challenging. We optimized a novel expansion protocol to stimulate donor-specific Tregs using soluble 4-trimer CD40 ligand (CD40L)activated donor B cells that expressed mature antigen-presenting cell markers. This avoided the use of CD40L-expressing stimulator cells that might otherwise result in potential cellular contamination. Purified allogeneic "recipient" CD4 + CD25 + Tregs were stimulated on days 0 and 7 with expanded "donor" B cells in the presence of IL-2, TGFβ and sirolimus (SRL). Tregs were further amplified by polyclonal stimulation with anti-CD3/CD28 beads on day 14 without SRL, and harvested on day 21, with extrapolated fold expansion into the thousands. The expanded AgTregs maintained expression of classical Treg markers including demethylation of the Treg-specific demethylated region (CNS2) and also displayed constricted TcR repertoire. We observed AgTregs more potently inhibited MLR than polyclonally expanded Tregs and generated new Tregs in autologous responder cells (a measure of infectious tolerance). Thus, an optimized and more clinically applicable protocol for the expansion of donor-specific Tregs has been developed.Organ and tissue transplantation currently rely on nonspecific immunosuppressive agents (IS) given life-long to prevent graft rejection. Although the introduction of new immunosuppressive agents, in particular calcineurin inhibitors (CNI), has resulted in a dramatic reduction in acute rejection rates following renal transplantation, these drugs have failed to prevent chronic allograft dysfunction (CAD). Moreover, these IS are associated with significant morbidity, particularly to the kidneys, and increased rates of infections, malignancy, diabetes, and hypertension. Therefore, since organs were first transplanted in humans, the elusive goal has been to establish donor-specific immunological tolerance, a state where a donated organ is accepted as "self, " eliminating the need for IS. To date the best clinical success at achieving tolerance has been through the combined use of hematopoietic stem cells (HSC) and solid organ transplantation 1-3 . However, the long term safety, efficacy, and broad applicability of approaches using therapeutic transfer of donor HSC is not yet known.Regulatory CD4 + CD25 + FOXP3 + T cells (Tregs) have been shown to be elevated in tolerant human transplant recipients, and polyclonally expanded Tregs have been used to delay allograft rejection in experimental animals 4,5 including humanized mice 6 , despite contrary results recently obtained in a heart transplant model in non-human primates 7 . As such, a number of centers, including our own, are beginning to explore the use of autologous regulatory T cells clinically to induce transplant tolerance8,9.Polyclonally expanded Tregs have also been shown to have possible benefit in human GVHD 10-12 and diabetes 13 . At our center, we have perfected large scale isolation and polyclonal expansion of good manufacturing 1
practice (GMP)-grade autologous Tregs and successfully completed a phase 1 safety trial in living donor kidney transplant patients with no study-related adverse events [bib_ref] Interim Results of a Phase 1 Trial of Treg Adoptive Cell Transfer..., Leventhal [/bib_ref]. Additionally, these patients developed increased numbers of circulating Tregs post-infusion. Based on this, we are poised to pursue a phase II efficacy trial.
Compared to polyclonally-expanded Tregs, donor-specific Tregs have the potential advantage of being more potent and specifically targeted to control alloimmune responses [bib_ref] Clinical Grade Manufacturing of Human Alloantigen-Reactive Regulatory T Cells for Use in..., Putnam [/bib_ref]. Therefore, a number of groups have developed Treg expansion protocols with restricted T cell receptor (TCR) repertoire recognizing only donor antigens and have shown such Tregs to delay graft rejection in rodent models of solid organ transplantation [bib_ref] Clinical Grade Manufacturing of Human Alloantigen-Reactive Regulatory T Cells for Use in..., Putnam [/bib_ref] [bib_ref] Conferring indirect allospecificity on CD4+CD25+ Tregs by TCR gene transfer favors transplantation..., Tsang [/bib_ref] [bib_ref] Human regulatory T cells with alloantigen specificity are more potent inhibitors of..., Sagoo [/bib_ref] [bib_ref] Specificity of CD4+CD25+ regulatory T cell function in alloimmunity, Sanchez-Fueyo [/bib_ref]. In this manuscript we have tested the hypothesis whether we can generate and expand in culture potent antigen-specific Tregs that can potentially be utilized clinically for tolerance induction.
A critical reagent to generate these alloantigen-specific Tregs is the donor antigen-presenting cells (APC). The source of these APC has been peripheral blood mononuclear cells (PBMC) alone [bib_ref] Ex vivo generation of human alloantigen-specific regulatory T cells from CD4(pos)CD25(high) T..., Peters [/bib_ref] or, in combination with FACS sorting [bib_ref] CD27/CFSE-based ex vivo selection of highly suppressive alloantigen-specific human regulatory T cells, Koenen [/bib_ref] , dendritic cells [bib_ref] Expansion of FOXP3high regulatory T cells by human dendritic cells (DCs) in..., Banerjee [/bib_ref] , and B cells [bib_ref] Clinical Grade Manufacturing of Human Alloantigen-Reactive Regulatory T Cells for Use in..., Putnam [/bib_ref] [bib_ref] Efficient generation of human alloantigen-specific CD4+ regulatory T cells from naive precursors..., Tu [/bib_ref]. Of these methods, activated B cells have been used most often. However, current published protocols rely on CD40L-expressing feeder cells for B cell activation and expansion, which have raised issues about their suitability for use in patients. Recently developed 4-trimer soluble CD40L circumvents the need for CD40L-expressing feeder cells to produce the adequate activation and expansion of B cells for use as APCs.
This study examines the effect of a 4-trimer soluble form of CD40L (UltraCD40L) on the expansion and activation of B cells, demonstrates generation of donor-specific Tregs using these activated B cells as APCs, and reports the ability of these Tregs to inhibit "recipient" anti-"donor" responses and preferentially induce the generation of new Tregs from "recipient" naïve CD4 + cells. Thus, an optimized and more clinically applicable protocol for the expansion of Tregs has been developed. It should be stressed that this report is not to provide a side-by-side comparison against other protocols being utilized elsewhere, but rather to describe the development of a Treg expansion protocol that does not require complicated starting cell purifications or CD40L-expressing feeder cells.
# Results
The cellular reactants. In the proposed transplant situation, recipient Tregs are expanded against the organ or tissue donor and infused into the recipient after the transplant. Logistically, this can be achieved by pre-transplant non-mobilized leukophoresis of the recipient and cryopreservation of the product. Simultaneously, donor cells from either peripheral blood of a living donor or spleen cells of a deceased donor are obtained and the B cells are expanded and then cryopreserved, if required. Twenty-one days before the expected Treg infusion, the recipient leukophoresis product is thawed and the CD4 + CD25 + cells are isolated and stimulated with expanded and irradiated donor B cells. Towards this, as a proof of principle, the following novel ex vivo expansion protocol was developed and optimized using peripheral blood mononuclear cells (PBMC) from two normal volunteers, henceforth denoting the individual from whom Tregs were isolated as "recipient" and the allogeneic B cells donor as "donor".
## Generation of activated b cells from "donor" pbmc in culture. previous reports have demonstrated
that B cells can be expanded using activation by cells expressing membrane CD40L plus IL-4 along with T-cell inhibition by CsA in culture [bib_ref] Efficient generation of human alloantigen-specific CD4+ regulatory T cells from naive precursors..., Tu [/bib_ref]. In the current study, a similar system was developed that used a cell-free source of CD40L, namely UltraCD40L (a 4-trimer form of CD40L produced as a fusion between the body of surfactant protein D (SPD) as a scaffold and the extracellular domain of CD40L, also referred to as SPD-CD40L). The 4-trimer structure of UltraCD40L allows for better clustering of the CD40 molecules on B cells and thus increased expansion [bib_ref] Multimeric soluble CD40 ligand and GITR ligand as adjuvants for human immunodeficiency..., Stone [/bib_ref] [bib_ref] Design of CD40 agonists and their use in growing B cells for..., Kornbluth [/bib_ref]. The UltraCD40L was produced by transfected CHO cells cultured in RPMI-1640 medium supplemented with 10% normal AB − serum. The concentration of UltraCD40L in the culture supernatant ranges from 2.5-6 μg/ml and was used at 20% vol/vol for B cell expansion.
"Donor" PBMC were cultured under stimulation with UltraCD40L and 40 ng/mL IL-4 in the presence of 400ng/ml CsA to prevent the growth of T cells. We monitored the number of total cells and CD19 + cells in the culture to evaluate B cell growth. There was ~10-fold increase in the absolute number of cells after 14 days in culture when PBMC was the starting population, as opposed to a ~100-fold expansion if the initiating culture was CD19 + purified by immunomagnetic beads [fig_ref] Figure 1: Robust expansion of "donor" B cells to be used as stimulators using... [/fig_ref]. The phenotypic nature of the expanded cells were monitored by 5-color flow cytometric analyses performed on a Beckman-Coulter FC500 flow cytometer as previously described [bib_ref] Interim Results of a Phase 1 Trial of Treg Adoptive Cell Transfer..., Leventhal [/bib_ref] [bib_ref] The human "Treg MLR": immune monitoring for FOXP3+T regulatory cell generation, Levitsky [/bib_ref] [bib_ref] Nonchimeric HLA-Identical Renal Transplant Tolerance: Regulatory Immunophenotypic/Genomic Biomarkers, Leventhal [/bib_ref]. This growth in the PBMC cultures was due to the amplification of the number and proportion of CD19 + B cells [fig_ref] Figure 1: Robust expansion of "donor" B cells to be used as stimulators using... [/fig_ref]. The B cell cultures also showed an increased percentage of cells that expressed CD86 and CD80 [fig_ref] Figure 1: Robust expansion of "donor" B cells to be used as stimulators using... [/fig_ref] with the majority expressing HLA-DR throughout the culture. Thus, these expanded "donor" B cells demonstrated the phenotype of mature APCs. A detailed monitoring with more frequent flow cytometric analysis revealed that the maximal cell growth with optimum costimulatory molecule expression occurred on day 11-12, and therefore, subsequent B cell cultures were harvested at that period (data not shown). Since we obtained equally pure and activated B cells from cultures initiated from total PBMCs or purified CD19 + cells, for ease of clinical translation all further B cells expansions were performed from total PBMCs. The expanded B cells could be utilized either freshly or after cryopreservation with equivalent stimulatory capabilities (data not shown).
## Expansion of purified tregs with stimulation by activated b cells.
Next, we determined if purified CD4 + CD25 + cells could be expanded using these activated B cells. The Tregs were isolated from "recipient" PBMC using Miltenyi Treg isolation kit II and were then cultured in 24-well culture plates with irradiated (3,000 cGy) "donor" B cells that had been expanded and activated in culture as above at a ratio of 1:1. Culture medium consisted of RPMI-1640 supplemented with 15% human AB serum, 1,000 IU/mL IL-2 and 1ng/ml TGF-β. 100 nM sirolimus was added to cultures except for the final 7 days. In terms of absolute number of cells, the Tregs proliferated ~20-fold during 28 days in culture, with robust proliferation occurring after SRL was removed from culture on day 21 [fig_ref] Figure 2: Figure 2 [/fig_ref]. However, since <5% of the purified CD4 + CD25 + are expected to be alloreactive and even lower percentages could be specific to the particular stimulator, this would represent an expansion from 400 to >1,000 fold.
Phenotypically, the purified Tregs on day 0 prior to culture, were >95% CD4 + and >85% FOXP3 + . A minor population of conventional T cells staining CD127 dim was present in the purified product. However, it gradually disappeared with the final product becoming virtually devoid of CD127 + cells [fig_ref] Figure 2: Figure 2 [/fig_ref]. Because of the flow cytometric configuration available in the laboratory at the time, the CD25 + population could not be assessed on day 0 for the starting product, as the selection antibody sterically hindered the detection antibody. However, there was a progressive enrichment with >90% being CD25 + FOXP3 + cells at the end of the culture period [fig_ref] Figure 2: Figure 2 [/fig_ref]. Thus, >80% of the cells were CD4 + CD127 − CD25 + FOXP3 + Tregs in the final product [fig_ref] Figure 2: Figure 2 [/fig_ref]. Henceforth, these are designated as antigen-specific Tregs (AgTregs).
To test if any of the "donor" irradiated B cells still remained as a contaminant in the final Treg product, two approaches were utilized: (i) Expanded B cells were irradiated at varying cGrey and cultured in 15% HAB medium; no viable B cells could be detected beyond 1 week of the culture at any of the irradiation doses [fig_ref] Figure 1: Robust expansion of "donor" B cells to be used as stimulators using... [/fig_ref]. This indicated that no B cell contamination was to be expected in the expanded Tregs. (ii) Expanded Treg product was monitored for any residual B cells; no contamination with CD20 + B cells was observed. Similarly, no CD14 + monocytes, CD56 + NK cells or CD8 + T cells were also detected. A representative analysis on a 21 day culture is shown in [fig_ref] Figure 2: Figure 2 [/fig_ref]. Expansion of AgTregs (n = 6). 1 × 10 6 CD4 + CD25 + CD127 low cells were purified from healthy volunteers and stimulated with equal number of "donor" irradiated B cells that were expanded as in [fig_ref] Figure 1: Robust expansion of "donor" B cells to be used as stimulators using... [/fig_ref] and had one HLA-DR in common with the Tregs. Rapamycin at 100ng/mL was also included during the first 21 days. Tregs were restimulated the "donor" B cells at weekly intervals. (A) Tregs expanded ~20 fold during the 28 days in culture with the rapid expansion occurring after the removal of SRL on day 21. (B) Representative data of the flow cytometric scheme and analysis that demonstrated the gradual disappearance of residual CD127 dim cells with culture progression (top row), and robust expression of CD25 and FOXP3, the hallmark of Tregs (bottom row). [Note: Because of the flow cytometric configuration available in the laboratory, the CD25 + could not be accurately assessed on day 0 due to steric hindrance of the detection antibody by the selection antibodybeads.]. (C) Residual contamination by non-Treg subsets of cells were also monitored. No contamination with CD8 + T cells, CD14 + monocytes, CD56 + NK cells or CD20 + B cells was observed. Expanded AgTregs demonstrate potent "donor"-specific MLR inhibition. To examine the functional potency and specificity of expanded Tregs, MLR inhibition assays were performed. "Recipient" responder PBMC autologous to the Tregs were stimulated with irradiated PBMC from either the "donor" stimulator used in the generation of Tregs or a third party irrelevant individual in presence of either expanded AgTregs or irradiated "recipient" cells as modulator controls. The back-response CPM given by the AgTregs or control modulators at the tested doses [as assessed in (Rx + Dx + AgTregs) combinations] were subtracted from the (R + Dx + AgTregs) experimental cultures to obtain the delta CPM (Δ CPM), and the percentage of inhibition was calculated. [fig_ref] Figure 3: Specificity and Potency of AgTregs [/fig_ref] shows the results from a representative experiment as Δ CPM obtained in the cultures with AgTreg or control modulators at indicated ratios with "recipient" PBMC responders. When the percentage of inhibition by each ratio of AgTregs versus the control modulator in each experiment was calculated, a dose dependent and potent inhibition was observed [fig_ref] Figure 3: Specificity and Potency of AgTregs [/fig_ref]. This inhibition was significantly higher in the MLR against the specific donor stimulator than against irrelevant third party stimulator, thus demonstrating the antigen specificity of the expanded AgTregs.
Cultured antigen-specific Tregs generate new Tregs by infectious tolerance. One potential advantageous characteristic for maintaining a lasting immune tolerance by infused Tregs is that they can cause the generation of new Tregs in the recipients which is termed "infectious tolerance" [bib_ref] Infectious" transplantation tolerance, Qin [/bib_ref]. The Treg-MLR assay developed in the laboratory is an in vitro correlate of this, and we employed it to assess the ability of expanded AgTregs in generating new Tregs from naïve autologous responder cells. We utilized total PBMCs and not Treg depleted PBMCs as responder cells because, clinically, patients that receive Treg infusion will have native Tregs present and therefore the overall effect/ability of infused Tregs to generate new Tregs within recipients is better assessed with total PBMCs. Responder PBMC labelled with CFSE were stimulated with either PKH26-labelled and donor irradiated PBMC or third party irrelevant irradiated PBMC in the presence of PKH26-labelled modulator cells (AgTregs or additional responder PBMC as controls) at varying concentrations. The percentages of CD4 + CD127 − CD25 High FOXP3 + cells that were generated in CFSE diluted proliferating responder cells were determined by flow cytometry on day 7 and analyzed after gating out all PKH + modulators and residual stimulators as well as CD127-PE positive effector cells [fig_ref] Figure 4: Infectious Generation of new Tregs in responders by expanded AgTregs [/fig_ref].
When autologous PBMC served as the modulator cells for either "donor"-specific (R + Dx + Rx) or third party indifferent (R + Ix + Rx) reactions, the percentage of CD4 + CD127 − CD25 High FOXP3 + Tregs in the proliferating responders remained constant at around 20% at all modulator doses tested [fig_ref] Figure 4: Infectious Generation of new Tregs in responders by expanded AgTregs [/fig_ref]. However, there was a The CD4 cells that were negative for PKH26 and CD127-PE and then those that diluted the CFSE were sequentially gated and analyzed for CD25 and FOXP3 expressions. Thus, the cells of interest were CD4 + CD127 − CD25 + FOXP3 + Tregs in the CFSE diluted proliferating responders. A representative experiment with "donor" and third party stimulators at 1:50 modulator: T responder ratio is shown. (B) The percentages of CD4 + CD127 − CD25 + FOXP3 + Tregs obtained with the AgTreg modulators were divided by those obtained with the Rx control modulators to calculate the fold change. Thus, in the example in [fig_ref] Figure 4: Infectious Generation of new Tregs in responders by expanded AgTregs [/fig_ref] two-fold (200%) increase in the proportion CD4 + CD127 − CD25 High FOXP3 + cells among proliferating responders when AgTregs were added as modulator cells into the "donor"-specific (R + Dx + AgTreg) cultures at 1:10 and 1:50 Treg: T-responder ratios [fig_ref] Figure 4: Infectious Generation of new Tregs in responders by expanded AgTregs [/fig_ref]. In contrast, no such increase in the Tregs newly generated in the responders was observed in the third party specific (R + Ix + AgTreg) cultures [fig_ref] Figure 4: Infectious Generation of new Tregs in responders by expanded AgTregs [/fig_ref]. These results demonstrate that AgTregs are capable of infectiously generating new Tregs in autologous naïve responder cells.
Further optimization of antigen-specific Treg expansion protocol for the GMP. We recognize that although our AgTregs are highly potent and specific, an expansion protocol with duration of a month will be costly and time consuming for clinical applications. We addressed this concern by asking if amplified expansion without the loss of antigen specificity could be obtained by polyclonal restimulation after two rounds of antigen specific stimulations. Purified CD4 + CD25 + cells were stimulated "donor"-specifically on days 0 and 7, and then polyclonally expanded with MACS GMP ExpAct Treg beads at 1:1 ratio on day 14 without SRL. The cultures were harvested on day 21 and compared against the standardized protocol described above. We also included another comparator wherein the third stimulation on day 14 was also antigen-specific and in the absence of SRL. When analyzed on day 21, all the three cultures [fig_ref] Figure 5: Optimization of AgTreg Expansion with Polyclonal Re-stimulation [/fig_ref] showed equivalent proportion of cells that were CD4 + and CD127 − as well as CD25 + and FOXP3 + [fig_ref] Figure 5: Optimization of AgTreg Expansion with Polyclonal Re-stimulation [/fig_ref] , proving polyclonal restimulation does not negatively affect Treg phenotype. However, there was a two-fold increase in the absolute number of Tregs when polyclonal restimulation was used as compared to antigen-specific stimulation throughout, both in the absence of SRL during the final 7-day culture period [fig_ref] Figure 5: Optimization of AgTreg Expansion with Polyclonal Re-stimulation [/fig_ref]. When assayed functionally in MLR inhibition assays on day 21, all three cultures demonstrated equivalently potent inhibition against the specific "donor" stimulator used in expanding the Tregs [fig_ref] Figure 5: Optimization of AgTreg Expansion with Polyclonal Re-stimulation [/fig_ref]. However, there were subtle differences in the inhibition against the third party irrelevant stimulated MLRs [fig_ref] Figure 5: Optimization of AgTreg Expansion with Polyclonal Re-stimulation [/fig_ref] , bottom dashed lines); the cultures from which SRL was withdrawn on day 14 (middle and right) had higher inhibition at 1:10 and 1:50 Treg: responder ratios, when compared to the culture with SRL present throughout the 21 day culture (left). Importantly, the inhibition was "donor"-specific in all cultures at higher and more clinically relevant Treg: responder ratios (>1:250). These results demonstrate that amplified expansion without the loss of antigen specificity can be obtained by polyclonal restimulation on day 14 subsequent to antigen specific stimulations on days 0 and 7.
Comparison of antigen-specific versus polyclonally expanded Tregs. We recently developed the technology for the expansion of Tregs polyclonally and completed a phase I safety trial in kidney transplant recipients with these polyclonally expanded autologous regulatory T cells [bib_ref] Interim Results of a Phase 1 Trial of Treg Adoptive Cell Transfer..., Leventhal [/bib_ref]. Therefore, it was of interest to make a direct comparison of the characteristics of polyclonal versus AgTregs. We also purified the Tregs by first depleting non-CD4 cells and then positively selecting CD25 + cells using the reagents and CliniMACS system that would be used for clinical expansion. Tregs were expanded from the same individuals using the optimized AgTreg and polyclonal Treg expansions, i.e. (i) antigen-specific stimulation on days 0 and 7 followed by restimulation on day 14 with anti-CD3/CD28 beads, and (ii) polyclonal stimulation with anti-CD3/CD28 beads on days 0 and 7. Both the cultures were in the absence of SRL from day 14 onwards and harvested on day 21. As expected, the fold expansion was highest with polyclonal stimulations [fig_ref] Figure 6: Superiority of Antigen Specific Tregs versus polyclonal Tregs [/fig_ref]. Phenotypically, both cultures demonstrated equivalent characteristics and similar to those shown in Figs 2 and 4. However, when tested as modulators in MLRs, the polyclonally expanded Tregs displayed less inhibition of MLR compared to AgTregs which inhibited the responses in an antigen-specific manner, especially at the higher Treg: T-responder ratios [fig_ref] Figure 6: Superiority of Antigen Specific Tregs versus polyclonal Tregs [/fig_ref]. For instance, ~75% inhibition was demonstrated by AgTregs versus ~35% by polyclonally expanded Tregs at Treg: T responder ratio of 1:250. In another series of experiments in which head-to-head comparisons were not made, polyclonally expanded Tregs showed lower dose-dependent inhibition also (n = 9; not shown). These results indicated that antigen-specific Tregs are more potent and specifically targeted.
Expanded Tregs demonstrated stable FOXP3 demethylation profile. The demethylation of a conserved region in the first intron of the Foxp3 gene Treg-specific demethylated region (TSDR) has been demonstrated as the most reliable criterion for identification of Tregs [bib_ref] DNA demethylation in the human FOXP3 locus discriminates regulatory T cells from..., Baron [/bib_ref] [bib_ref] Intragraft regulatory T cells in protocol biopsies retain foxp3 demethylation and are..., Bestard [/bib_ref]. Therefore, the methylation / demethylation status of the expanded Tregs was compared against those of a number of cell subsets by bisulfite conversion, and pyrosequencing of isolated DNA through the use of a commercially available vendor services (EpigenDx). As can be seen in [fig_ref] Figure 6: Superiority of Antigen Specific Tregs versus polyclonal Tregs [/fig_ref] and D, freshly isolated CD25 − effector T cells, the negative controls had >90% cells methylated and the starting population of CD25 + cells (Day 0) had ~50% cells methylated. The polyclonally expanded Tregs retained similar methylation / demethylation status throughout the culture. Similarly, the AgTregs also demonstrated equivalent TSDR despite the analyzed samples were the total Treg culture, without depletion of residual contaminating stimulator DNA (this was done as the total cellular product would be infused into the patient). Please also note that no correction for X-linked inactivation [bib_ref] Ex Vivo-Expanded but Not In Vitro-Induced Human Regulatory T Cells Are Candidates..., Rossetti [/bib_ref] was made when the results were plotted. These results showed that the expanded Tregs had stable TSDR.
## Expanded agtregs have restricted but sufficiently broad tcr repertoire usage. a diverse t cell
repertoire has been found to be essential to maintain an adequate level of recognition and response to environmental antigens [bib_ref] The many important facets of T-cell repertoire diversity, Nikolich-Zugich [/bib_ref] and this is true even for regulatory T cells controlling responses to the wide range of antigens [bib_ref] Continuous requirement for the TCR in regulatory T cell function, Levine [/bib_ref]. Since alloreactivity encountered in a transplant situation has an array of antigens acting as immune targets, a wide indicated AgTreg: T responder ratios for each experiment. The data are shown as mean ± SD fold change from n = 5 experiments. *p < 0.05. There was an amplification of Tregs that were newly generated in the proliferating fraction of autologous responders by the expanded AgTregs in a "donor" specific manner. range of clonal diversity is expected. Thus, a TcR repertoire analysis (curtesy of ArcherDx, Boulder, CO), revealed that there was sufficient breadth, albeit with significant shrinkage in the TCR diversity in the AgTregs cultured for 21 days from the Day 0 CD25 + starting population of cells [fig_ref] Figure 6: Superiority of Antigen Specific Tregs versus polyclonal Tregs [/fig_ref]. This was in contrast to an expansion in clonal diversity in the 21-day polyclonally expanded Tregs (i.e. a more polyclonal repertoire); this was likely due to uncovering of low frequency clones within the Treg product that were below the detection level in the starting cells and not by the generation of new clones as expansion was an ex vivo procedure. However, when the top clonotypes that expanded in culture were assessed, the AgTregs had more clonotypes in common with the starting cells than the polyclonal Tregs [fig_ref] Figure 6: Superiority of Antigen Specific Tregs versus polyclonal Tregs [/fig_ref] thus indicating that the more abundant clones in the CD25 + product were the ones that expanded in AgTreg cultures. Overall, Treg receptor diversity was maintained post expansion in the polyclonal product with the expected shrinkage in the AgTregs.
# Discussion
This study describes the development of the technology for effective isolation and expansion of allospecific regulatory T cells through the use of "donor" B cells that have been expanded by a novel clinically applicable method. Several reports 14,21 describe the challenges in the generation of large numbers of antigen-specific Tregs for clinical use. The first step in the process is to produce large numbers of "donor" APC to stimulate and expand donor antigen-reactive Tregs. Although Tu et al. first reported roughly equal effectiveness of soluble CD40L compared with CD40L-transfected feeder cells expressing membrane CD40L to generate activated B cells, most protocols currently rely on the latter which may produce cellular contamination and limit clinical use. Using newly developed technology where 4 trimers of CD40L are linked together (UltraCD40L), large quantities of activated B cells can be expanded without feeder cells, thus making this approach more clinically applicable. This is one of the features that distinguishes this study from others, i.e. the use of cell-free, soluble 4-trimer CD40L (Ultra-CD40L) to expand "donor" B cells. Our simplified B cell expansion protocol used in these experiments has thereby produced similar expansion and activation [fig_ref] Figure 1: Robust expansion of "donor" B cells to be used as stimulators using... [/fig_ref] of B cells from PBMC compared to previous reports [bib_ref] Clinical Grade Manufacturing of Human Alloantigen-Reactive Regulatory T Cells for Use in..., Putnam [/bib_ref] [bib_ref] Efficient generation of human alloantigen-specific CD4+ regulatory T cells from naive precursors..., Tu [/bib_ref]. Whereas most protocols begin by isolating B cells from PBMC we found this step unnecessary and achieved equally robust expansion when starting from unfractionated PBMC. Under our culture conditions using CsA to suppress T cell growth along with B cell stimulation by UltraCD40L and IL-4, we produced highly immunogenic mature APC expressing CD86, CD80 and HLA-DR. This protocol thereby produced >200 million B cells starting with 20 mL of peripheral blood, i.e., ~100-fold expansion, indicating that clinical-scale production would be possible assuming 2-3 billion stimulator B cells would yield an effective number of antigen-specific Tregs.
The AgTregs produced in this study also confirm reports by others that activated B cells can be used to select ("donor") antigen-specific Tregs [bib_ref] Clinical Grade Manufacturing of Human Alloantigen-Reactive Regulatory T Cells for Use in..., Putnam [/bib_ref] [bib_ref] Efficient generation of human alloantigen-specific CD4+ regulatory T cells from naive precursors..., Tu [/bib_ref]. In our final optimized protocol now poised to be transferred to our clinical GMP facility, the Tregs were stimulated "donor"-specifically on days 0 and 7 in presence of SRL and then polyclonally with anti-CD3/CD28 on day 14 without SRL, and the cells harvested on day 21. Another distinguishing feature of our protocol is that the starting population of cells is prepared using a less cumbersome method than that reported by Putnam et al. [bib_ref] Clinical Grade Manufacturing of Human Alloantigen-Reactive Regulatory T Cells for Use in..., Putnam [/bib_ref] who used flow-sorting of the starting Treg cells (and CD40L-expressing K562 leukemia cells as stimulators). However, the amplification of Tregs is specifically favored in our cultures by the use of 100ng/ml SRL to prevent the expansion of effector T cells [bib_ref] Rapamycin promotes expansion of functional CD4+CD25+FOXP3+ regulatory T cells of both healthy..., Battaglia [/bib_ref] [bib_ref] Costimulatory blockade with mTor inhibition abrogates effector T-cell responses allowing regulatory T-cell..., Bestard [/bib_ref] [bib_ref] Rapamycin inhibits differentiation of Th17 cells and promotes generation of FoxP3+T regulatory..., Kopf [/bib_ref] [bib_ref] Allospecific Regulatory Effects of sirolimus and Tacrolimus in the Human Mixed Lymphocyte..., Levitsky [/bib_ref] [bib_ref] Immunoregulatory Effects of Everolimus on In Vitro Alloimmune Responses, Levitsky [/bib_ref] [bib_ref] Ex vivo-expanded but not in vitro-induced human regulatory T cells are candidates..., Rossetti [/bib_ref] [bib_ref] Differential impact of mammalian target of rapamycin inhibition on CD4+CD25+Foxp3+ regulatory T..., Zeiser [/bib_ref] , giving ~60 fold expansion in the total number of Tregs by this protocol. Since < 5% of the purified CD4 + CD25 + are expected to be alloreactive and even lower percentages could be specific to the particular "donor", this would represent a fold expansion into the thousands.
One of the critical questions that is relevant to immunomodulation therapy with Tregs is whether sufficient numbers of cells can be obtained through the expansion protocol. In a typical human there are ~1,500 lymphocytes per µl of peripheral blood and there are ~5 × 10 6 µl (5 liters) of blood, and therefore ~7.5 × 10 9 total lymphocytes. In our renal transplant clinical protocol, we reduce the peripheral lymphocyte count to ~150 per µl (~7.5 × 10 8 total) or lower by the use of lymphodepleting agents such as Alemtuzumab. Under these circumstances, infusion of 10 7 -10 8 ex vivo expanded donor-specific Tregs is easily achievable, i.e., ~1:2 ratio of Treg to T responder in recipient peripheral blood after Alemtuzumab has been metabolized. Of course, this calculation does not take into account the recipient's other lymphoid compartments. However, since the Tregs are highly potent and specific immunoregulators (Figs 3 and 5) and can infectiously induce the generation of new Tregs, even lower proportions may possibly be tolerogenic.
The stability and potency of the AgTregs are very important considerations in Treg expansion protocols. Since our expanded Tregs demonstrate stable TSDR, it may be assumed that they may not show plasticity by turning into Th17 inflammatory cells [bib_ref] Instability of the transcription factor Foxp3 leads to the generation of pathogenic..., Zhou [/bib_ref] [bib_ref] Disease Manifestation and Inflammatory Activity as Modulators of Th17/Treg Balance and RORC/FoxP3..., Almanzar [/bib_ref]. Similarly, the expanded AgTregs also demonstrated a restricted yet sufficiently broad TcR repertoire expected of alloreactive regulatory T cells. Further, the Tregs produced in this study appears to be potent specific MLR inhibitors [fig_ref] Figure 6: Superiority of Antigen Specific Tregs versus polyclonal Tregs [/fig_ref]. Please note that the lowest ratio of Treg: T responder used here was 1:10 as opposed to 1:1 by others and the highest 1:1,250 by us vs 1: 125 by others 14 as we had reasoned that infusion(s) of AgTregs would achieve only higher ratios in vivo, and therefore, only equivalent ratios would provide clinically meaningful results. The higher potency of our Tregs, whether allo-specific or polyclonal, is possibly due to the use of SRL in the initial expansion culture period.
Another critical factor that determines successful in vivo tolerance induction is the maintenance of amplified Treg proportion and functions in the recipient. This can be achieved through prolonged survival of the infused Tregs, as observed by in the peripheral circulation in humans [bib_ref] Polyclonal Treg Adoptive Therapy for Control of Subclinical Kidney Transplant Inflammation (TASK..., Chandran [/bib_ref] , and by Zhang, et al. in the circulation and secondary lymphoid tissues of non-human primates [bib_ref] Sequential Monitoring and Stability of Ex Vivo-Expanded Autologous and Nonautologous Regulatory T..., Zhang [/bib_ref]. Alternate and relevant mechanisms conferring lasting clinical tolerance may include better homing and functioning 44 and infectious tolerance described by Waldman and co-workers as well as others [bib_ref] Mechanisms of peripheral tolerance and suppression induced by monoclonal antibodies to CD4..., Cobbold [/bib_ref] [bib_ref] Infectious tolerance: human CD25(+) regulatory T cells convey suppressor activity to conventional..., Jonuleit [/bib_ref] [bib_ref] Human CD4(+)CD25(+) regulatory, contact-dependent T cells induce interleukin 10-producing, contact-independent type 1-like..., Dieckmann [/bib_ref]. To assess this potential of AgTregs to infectiously generate new Tregs from naïve cells in the recipient, we have utilized the "Treg-MLR, " an in vitro correlate of the in vivo phenomenon [bib_ref] The human "Treg MLR": immune monitoring for FOXP3+T regulatory cell generation, Levitsky [/bib_ref]. Using this assay, we have demonstrated that proliferating autologous PBMC preferentially develop new Tregs with "donor"-specific but not third party stimulation [fig_ref] Figure 4: Infectious Generation of new Tregs in responders by expanded AgTregs [/fig_ref]. Previously, we had shown that the CD4 + cells that were newly generated in the Treg-MLR would in turn inhibited fresh MLRs again in an antigen-specific manner demonstrating their own specific inhibitory capabilities [bib_ref] Favorable effects of alemtuzumab on allospecific regulatory T-cell generation, Levitsky [/bib_ref]. Similar infectious tolerance through Treg development has been indirectly observed in transplant recipients after infusion of Tregs or donor hematopoietic stem cells [bib_ref] Chimerism and Tolerance Without GVHD or Engraftment Syndrome in HLA-Mismatched Combined Kidney..., Leventhal [/bib_ref] [bib_ref] Long-Term Results in Recipients of Combined HLA-Mismatched Kidney and Bone Marrow Transplantation..., Kawai [/bib_ref] [bib_ref] Chimerism, Graft Survival, and Withdrawal of Immunosuppressive Drugs in HLA Matched and..., Scandling [/bib_ref] [bib_ref] Efficient generation of human alloantigen-specific CD4+ regulatory T cells from naive precursors..., Tu [/bib_ref] [bib_ref] The human "Treg MLR": immune monitoring for FOXP3+T regulatory cell generation, Levitsky [/bib_ref] [bib_ref] Nonchimeric HLA-Identical Renal Transplant Tolerance: Regulatory Immunophenotypic/Genomic Biomarkers, Leventhal [/bib_ref] [bib_ref] Tolerance Induction in HLA Disparate Living Donor Kidney Transplantation by Donor Stem..., Leventhal [/bib_ref] [bib_ref] Immune Reconstitution/Immunocompetence in Recipients of Kidney Plus Hematopoietic Stem/Facilitating Cell Transplants, Leventhal [/bib_ref] [bib_ref] Genomic biomarkers correlate with HLA-identical renal transplant tolerance, Leventhal [/bib_ref] [bib_ref] Mechanisms of donor-specific tolerance in recipients of haploidentical combined bone marrow/kidney transplantation, Andreola [/bib_ref]. The expanded AgTregs thus have tremendous clinical implications particularly in the induction and maintenance of transplant tolerance where current IS has failed long-term.
To summarize, this study has examined the effect of a cell-free, 4-trimer form of soluble CD40L (UltraCD40L) on the expansion and activation of B cells which in turn function as effective stimulators of antigen-specific regulatory T cells. Some of the salient and distinguishing features of the present study are: (i) the use of soluble 4-trimer CD40L to expand "donor" B cells instead of CD40L-expressing feeder cells; (ii) simplified purification of the initial "recipient" Treg cell to be expanded; (iii) stimulation of the Tregs with expanded "donor" B cells on days 0 and 7, then with anti-CD3/CD28-beads on day 14 and product harvest on day 21; and finally (iv) use of SRL in the initial stimulation phase (days 0-14) to prevent the expansion of effector T cells and favor the amplification of Tregs. We are in the process of transferring the technology to our GMP facility, so that a phase I clinical trial can be initiated in the near future. Thus, we have verified our hypothesis that we can generate and expand in culture potent antigen-specific Tregs that can potentially be utilized clinically for tolerance induction.
# Methods
Subjects. The subjects of the study were normal laboratory volunteers. Peripheral blood samples were obtained from them under a protocol following written informed consent approved and supervised by a Northwestern University Institutional Review Board. All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2008. The healthy volunteers were HLA-typed by the Northwestern HLA laboratory using molecular methods (reverse sequence specific oligonucleotide probe hybridization). In a typical experiment B cells were expanded using UltraCD40L and these "donor" cells were used to stimulate and expand purified CD4 + CD127 -CD25 + cells (Tregs) from an allogeneic individual designated as "recipient" using the following protocols. Generation of B cells from "donor" PBMC. Peripheral blood mononuclear cells (PBMC) were isolated using Ficoll-Hypaque density centrifugation and cultured in RPMI media (Invitrogen, Carlsbad, CA) containing 15% heat-inactivated human AB serum (Sigma, St. Louis, MO), HEPES buffer (Invitrogen), Penicillin/ streptomycin/glutamine (Invitrogen) with the addition of 40 ng/mL IL-4 (R&D Systems), 20% soluble 4-trimer CD40L (SPD-CD40L fusion protein, UltraCD40L; Multimeric Biotherapeutics, La Jolla, CA), and 400ng/ml cyclosporine-A (Novartis, NJ) at 37 °C with 5% CO 2 . Cells were cultured in T75 flasks (Corning, NY), counted every 3-4 days, and maintained at a cell density of 0.5-1 × 10 6 cells/mL. After 14 days in culture, the expanded B cells were phenotypically characterized by flow cytometry and were irradiated prior to being used as stimulators in the expansion of Tregs. Henceforth the volunteer from whom B cells were expanded and used as stimulators is denoted as "donor".
Treg isolation and expansion. PBMC were isolated from healthy human donors (mis)matched at one HLA-DR locus to the corresponding B cell donors. Initially, CD4 + CD127 − CD25 + Tregs were isolated using Miltenyi Treg isolation kit II (Miltenyi Biotec, Auburn, CA) by first depletion of CD8, CD19, CD123 and CD127 positive cells followed by positive selection for CD25 + cells. Subsequently at the final optimized stages, Tregs were isolated by first depletion of CD8 and CD19 positive cells and then positive selection for CD25 + cells using GMP grade reagents. Isolated Tregs were then cultured in 24-well culture plates (Corning) for 21 or 28 days with irradiated (3,000 rads) B cells that had been expanded and activated in culture as above at a ratio of 1:1. Culture medium consisted of RPMI-1640, 15% human AB serum, penicillin/streptomycin/L-glutamine, HEPES, 1,000 IU/mL IL-2 (R&D Systems), and 1ng/ml TGF-β (R&D Systems). 100 nM sirolimus (SRL; Pfizer, NY) was added to cultures except for the final 7 days. Initially, the cultures were restimulated on days 7, 14, and 21 with the irradiated B cells at a 1:1 ratio. In the final optimized protocol, the Tregs were stimulated with irradiated "donor" B cells at a 1:1 ratio on days 0 and 7 in presence of SRL and restimulated polyclonally with MACS GMP ExpAct Treg beads (Miltenyi Biotec) at 1:1 ratio on day 14 in the absence of SRL; the cells were harvested on day 21. For polyclonal expansion of Tregs isolated CD4 + CD25 + cells were stimulated with MACS GMP ExpAct Treg beads at 1:4 ratio on days 0 and 1:1 on day 7 in the presence of SRL and on day 14 in absence of SRL; the cells were harvested on day 21. Tregs were cultured at a cell density of 1 × 10 6 cells/mL with flow cytometric phenotyping performed at weekly intervals. Henceforth, the volunteer from whom Tregs are expanded is denoted as "recipient".
Flow Cytometry. For "donor" B cell expansion experiments, flow cytometry was performed on cultured PBMC on indicated days using antibodies against CD19-PC7, CD80-PE, CD86-PC5, HLA-DR-ECD (all from Beckman-Coulter, Miami, FL). To phenotype Tregs, antibodies against CD4-FITC, CD127-PE, CD3-ECD, CD25-PC7 (all from Beckman-Coulter) and FOXP3-PC5 (eBioscience, San Diego, CA) were used on days 0, 14 and 21, (also on day 28 in longer-term cultures). All detection was performed on a Beckman-Coulter FC500 flow cytometer as previously described [bib_ref] Interim Results of a Phase 1 Trial of Treg Adoptive Cell Transfer..., Leventhal [/bib_ref] [bib_ref] The human "Treg MLR": immune monitoring for FOXP3+T regulatory cell generation, Levitsky [/bib_ref] [bib_ref] Nonchimeric HLA-Identical Renal Transplant Tolerance: Regulatory Immunophenotypic/Genomic Biomarkers, Leventhal [/bib_ref]. The gating strategy used for the analyses including that for the negative controls is shown in Supplemental [fig_ref] Figure 1: Robust expansion of "donor" B cells to be used as stimulators using... [/fig_ref].
## Mixed lymphocyte reaction suppression assay. freshly isolated responder pbmc autologous to the
Tregs were stimulated with freshly isolated "donor"-specific irradiated PBMC from the B cell donor or an irrelevant third party at a ratio of 1:1 in U-bottom 96-well plates in triplicate. Tregs or irradiated R-PBMC (control) were added at indicated modulator: responder ratios at the initiation of the suppression assays. Even though the "none" control is used widely, the true control should be where the same number of modulator control cells are added to the culture but without observing any inhibitory effect. If conventional T cells from the same donor were to be expanded in the same way as Tregs, they would generate cytotoxic effector T cells that would then cause the lysis of the stimulator cells of the inhibition experiment 53 resulting in either non-responsiveness or partial response. Therefore, it was found that the best control would be equivalent number of either irradiated or non-irradiated responder PBMC added as the modulator control [bib_ref] Modulatory effects of human donor bone marrow cells on allogeneic cellular immune..., Mathew [/bib_ref]. The cells were irradiated at 3,000 rads. After 7 days, 3 H-thymidine was added to the cultures during the final 16-20 hours and incorporation of 3 H-thymidine was used to measure proliferation. The back-response CPM given by the Tregs or control modulators at the tested doses were also assessed in separate cultures (Rx + Dx/Ix + AgTregs/Rx) and were subtracted from (R + Dx/ Ix + AgTregs/Rx) experimental cultures to obtain the delta CPM (Δ CPM). The mixed lymphocyte reaction (MLR) inhibition was quantified as:
[formula] = − Δ Δ × % inhibition 1 [/formula]
CPM in presence of Tregs CPM in presence of Rx controls 100
Treg-MLR (Infectious tolerance assay). The ability of expanded Tregs to generate new Tregs infectiously from naïve responder cells was measured using the "Treg-MLR" as described previously [bib_ref] The human "Treg MLR": immune monitoring for FOXP3+T regulatory cell generation, Levitsky [/bib_ref]. Briefly, MLR cultures were set up with "recipient" CFSE-labelled responders stimulated with irradiated (3,000 rads) and PKH26-labelled donor-specific or allo-irrelevant PBMC. PKH26-labelled modulator cells composed of either Tregs or R-PBMC treated in an equivalent manner to serve as controls were then added at modulator: responder ratios of 1:10, 1:50, and 1:250. On day 7, flow cytometry was performed on the cultured cells after labelling with CD127-PE, CD4-ECD, CD25-PC7 (all from Beckman-Coulter), and FOXP3-PC5 (eBioscience). PKH26-labelled modulators and any surviving stimulators as well as CD127-PE + responder cells were gated out, and CD4 + cells that proliferated (as determined by reduced CFSE expression) were analyzed for CD25 and FOXP3 expressions. Thus, the percentage of CD4 + CD127 − CD25 High FOXP3 + cells that were newly generated in the proliferating responder cells was determined. Methylation Analysis. Cell pellets from whole Treg cultures without purification or depletion of residual stimulator B cells were sent to EpigenDx (Hopkinton, MA) for DNA extraction, bisulfite conversion, and pyrosequencing. Percent methylation was calculated as (% methylated cytosine)/(% methylated cytosine + unmethylated cytosine). All Treg cultures were from female donors; however, no correction for X-linked inactivation [bib_ref] Ex Vivo-Expanded but Not In Vitro-Induced Human Regulatory T Cells Are Candidates..., Rossetti [/bib_ref] was made when the results were plotted.
TCR repertoire analysis. Cell pellets of 0.5 × 10 6 from 21-day expanded Tregs cultures (from 3 different AgTregs and 3 different polyclonal expansions) and the Day 0 CD4 + CD25 + starting Treg culture (in duplicate) were sent to ArcherDx (Boulder, CO) for survey level TCRβγ Immunoverse Analysis. Normal PBMC was used as a control to ensure that the largest number of T cell clones were identified in the assay. The RNA were extracted with Qiagen RNeasy and quantified via Qubit. Libraries were prepared with sample input normalized by volume (20uL eluate) and the TCRβγ Immunoverse analyses were performed. The data including the number of clones, number of clonotypes and the diversity were analyzed using algorithms developed by ArcherDx.
Statistical Analysis. Paired Student T-tests and Wilcoxon signed rank tests for parametric and nonparametric calculations respectively were used. P values of ≤0.05 were considered statistically significant and indicated by * in the Figures. If not otherwise mentioned, the data are shown as mean ± SD.
Data Availability. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
[fig] Figure 1: Robust expansion of "donor" B cells to be used as stimulators using ultraCD40L (n = 8). Peripheral blood was obtained from healthy volunteers and PBMC were isolated using Ficoll centrifugation. Either PBMC or CD19 + B cells purified using Miltenyi microbeads were cultured for 14 days in B cell culture media containing 20% ultraCD40L and 40ng/mL IL-4. Cyclosporine at 400ng/mL concentration was also included with PBMC cultures to prevent the growth of T cells. Cells were counted and flow cytometric analysis were performed on indicated days.(A) Absolute number of cells in the culture (mean ± SD); (B) Percentage of CD19 + B cells in the cultures where the starting cells were total PBMC; (C) A representative flow cytometry analysis showing the gating strategy for the classic B cell marker (CD19) and activation markers (CD80, CD86, and HLA-DR) in CD19 + gated cells. (D) The percentages CD19 + cells that expressed the activation markers on indicated days. Please note the robust expansion of B cells displaying high expression of HLA-DR, CD80 and CD86, hallmark of effective antigen presenting cells. (E) The expanded B cells that were to be used as stimulator cells were irradiated at indicated centi-Grey and cultured in 15% Human AB serum supplemented RPMI-1640 medium (15% HAB Medium). No viable B cells could be observed beyond 1 week of the culture, indicating that no B cell contamination was to be expected in the final expanded Treg product (of Fig. 2). SCIeNTIFIC REpoRTs | (2018) 8:1136 | DOI:10.1038/s41598-018-19621-6 [/fig]
[fig] Figure 2: Figure 2. Expansion of AgTregs (n = 6). 1 × 10 6 CD4 + CD25 + CD127 low cells were purified from healthy volunteers and stimulated with equal number of "donor" irradiated B cells that were expanded as in Fig. 1 and had one HLA-DR in common with the Tregs. Rapamycin at 100ng/mL was also included during the first 21 days. Tregs were restimulated the "donor" B cells at weekly intervals. (A) Tregs expanded ~20 fold during the 28 days in culture with the rapid expansion occurring after the removal of SRL on day 21. (B) Representative data of the flow cytometric scheme and analysis that demonstrated the gradual disappearance of residual CD127 dim cells with culture progression (top row), and robust expression of CD25 and FOXP3, the hallmark of Tregs (bottom row). [Note: Because of the flow cytometric configuration available in the laboratory, the CD25 + could not be accurately assessed on day 0 due to steric hindrance of the detection antibody by the selection antibodybeads.]. (C) Residual contamination by non-Treg subsets of cells were also monitored. No contamination with CD8 + T cells, CD14 + monocytes, CD56 + NK cells or CD20 + B cells was observed. [/fig]
[fig] Figure 3: Specificity and Potency of AgTregs. On Day 28, AgTregs were harvested and were used as modulators in mixed lymphocyte reaction of autologous PBMC stimulated with irradiated PBMCs from either the donor used for expanding the Tregs or an irrelevant third party. Additional responder PBMC was used as control modulators. Thus the combinations were [Recipient responder PBMC + donor irradiated stimulator + control = (R + Dx + Rx)], [Recipient responder PBMC + donor irradiated stimulator + AgTregs = (R + Dx + AgTregs)], [Recipient responder PBMC + irrelevant irradiated stimulator + control = (R + Ix + Rx)] and [Recipient responder PBMC + irrelevant irradiated stimulator + AgTregs = (R + Ix + AgTregs)]. After 7 days, a standard thymidine incorporation assay was performed. (A) Representative experiment showing the delta counts per minute (ΔCPM ± SD) values with the various modulators at indicated modulator: T responder ratios. (B) Data are shown as mean ± SD percentage of suppression that were calculated for each individual experiment (n = 9) using the formula shown in the text. Tregs demonstrated potent MLR inhibition against the specific donor stimulator with minimal inhibition against the third party, thus showing the antigen specificity of the inhibition. *p < 0.05; **p < 0.01. [/fig]
[fig] Figure 4: Infectious Generation of new Tregs in responders by expanded AgTregs. Responder PBMCs (R) were labeled with CFSE; and stimulator Dx and Ix as well as the modulators (Rx and Tregs) were labeled with PKH26 prior to assay performance. The responders were cultured with equal number of irradiated stimulators in presence of indicated ratios of the modulators. After 7 days flow cytometric analysis was performed with monoclonal antibodies CD127-PE, CD4-ECD, CD25-PC7 and FOXP3-PC5. (A) Flow cytometric gating strategy: [/fig]
[fig] Figure 5: Optimization of AgTreg Expansion with Polyclonal Re-stimulation. To improve expansion potential, the standard expansion protocol was modified with antigen specific B cell stimulation on days 0 and 7 followed by polyclonal stimulation with anti-CD3/CD28-beads on day 14. (A) Flow chart showing the method used to develop the technology (standard) vs. the method modified for transfer to GMP (optimized). (B) Fold expansion (Mean ± SD) with the optimized method of anti-CD3/CD28-bead restimulation and removal of SRL on day 14 (green line with triangular markers), with antigen-specific restimulation and removal of SRL on day 14 (purple line with square markers) and standard method (blue line with round markers). n = 7. (C) Representative experiment showing the phenotypic profile in terms of CD127, CD25 and FOXP3 expressions in CD4 + gated cells. (D) Comparison of "donor"-specific vs non-specific suppression by AgTregs expanded using the three methods and MLR inhibition performed as described in Fig. 3. n = 6, *p < 0.05 **p < 0.01. The method optimized for transfer to the GMP facility demonstrated superior expansion without the loss of antigen specificity. SCIeNTIFIC REpoRTs | (2018) 8:1136 | DOI:10.1038/s41598-018-19621-6 [/fig]
[fig] Figure 6: Superiority of Antigen Specific Tregs versus polyclonal Tregs. Tregs were expanded for 21 days from the same individual by (i) antigen-specific stimulations on days 0 and 7 plus polyclonal stimulations with anti-CD3/CD28-beads on day 14 (triangle markers), and (ii) polyclonal stimulations with anti-CD3/ CD28-beads on days 0, 7 and 14 (diamond markers). SRL was withdrawn from the cultures on day 14. (A) Fold expansion with the two methods of Treg expansions (Mean ± SD; n = 7). (B) Comparison of "donor"-specific vs non-specific suppression by Tregs expanded using the two methods and MLR inhibition performed as described in Fig. 3. n = 7 (please note that for the dilution 1: 1,250 the n = 4 only). The dotted line represents 50% inhibition. *p < 0.05. (C) Heat map of methylation analysis of sub-fractions of cells from a representative experiment. (D) Methylation status of freshly isolated effector cells and Tregs (Day 0) as well as Tregs expanded either antigen-specifically or polyclonally on days 14 and 21. Cells assayed without purification of Tregs and the results were plotted without correction for X-linked inactivation 38 , even though all Tregs were generated from females (n = 3). (E) Mean ± SD clonal diversity observed in 21 day cultures of AgTregs (n = 3) and polyclonally expanded Tregs (n = 3) versus Day 0 CD25 + starting population of cells (in duplicate). (F) The top clonotypes shared by 21 day cultures of AgTregs and polyclonally expanded Tregs (n = 3) versus Day 0 CD25 + starting population of cells. P = * < 0.05, ** < 0.01, *** < 0.001. As expected, the fold expansion by "donor"-specific stimulations was lower. The AgTregs demonstrated higher and specific inhibition at all Treg: T responder ratios. The Tregs demonstrated stable demethylated state of the FOXP3 gene. The AgTregs showed a slight shrinkage in the clonal diversity (as to be expected) but had higher top clonotypes in common with the starting cells than polyclonally expanded Tregs. SCIeNTIFIC REpoRTs | (2018) 8:1136 | DOI:10.1038/s41598-018-19621-6 [/fig]
[fig] SCIeNTIFIC: REpoRTs | (2018) 8:1136 | DOI:10.1038/s41598-018-19621-6 [/fig]
|
What can cohort studies in the dog tell us?
This paper addresses the use of cohort studies in canine medicine to date and highlights the benefits of wider use of such studies in the future. Uniquely amongst observational studies, cohort studies offer the investigator an opportunity to assess the temporal relationship between hypothesised risk factors and diseases. In human medicine cohort studies were initially used to investigate specific exposures but there has been a movement in recent years to more broadly assess the impact of complex lifestyles on morbidity and mortality. Such studies do not focus on narrow prior hypotheses but rather generate new theories about the impact of environmental and genetic risk factors on disease. Unfortunately cohort studies are expensive both in terms of initial investment and on-going costs. There is inevitably a delay between set up and the reporting of meaningful results. Expense and time constraints are likely why this study design has been used sparingly in the field of canine health studies. Despite their rather limited numbers, canine cohort studies have made a valuable contribution to the understanding of dog health, in areas such as the dynamics of infectious disease. Individual exposures such as neutering and dietary restriction have also been directly investigated. More recently, following the trend in human health, large cohort studies have been set up to assess the wider impact of dog lifestyle on their health. Such studies have the potential to develop and test hypotheses and stimulate new theories regarding the maintenance of life-long health in canine populations.Lay summaryCohort studies involve the repeated collection of information through time. Investigators are able to assess whether exposure to a particular risk factor (these could be environmental, diet, lifestyle, genetic etc) is associated with, and followed by, clinical outcomes in individuals. This paper highlights the benefits of wider use of such studies in the future. Thus researchers collect information (e.g. clinical data and records), rather than DNA samples to use in genetic studies. Information is longitudinally collected at regular time intervals, and allows investigators to assess whether exposure to particular risk factors (these could be environmental, diet, lifestyle, genetic, etc) are associated with, and followed by, clinical outcomes in individuals. This paper highlights the benefits of wider use of such studies in the future.Uniquely amongst observational studies, cohort studies offer researchers an opportunity to assess the time relationship between suggested risk factors and diseases. In human medicine, cohort studies were originally used to investigate specific exposures but there has been a recent move to more broadly assessing the impact of complex lifestyles on life and death. Such studies can generate new theories about the impact of environmental and genetic risk factors on disease. However, cohort studies are expensive and there is, inevitably, a delay between setting them up and the final reporting of meaningful results. Expense and time constraints are why this type of study has been used sparingly in canine health studies. Despite limited numbers, canine cohort studies have made a valuable contribution to the understanding of dog health, in areas such as infectious disease. Individual exposures such as neutering and dietary restriction have also been investigated. More recently, large cohort studies have been set up to assess the impact of dog lifestyle on their health. Such studies have the potential to develop and test hypotheses (ideas) and stimulate new theories regarding the maintenance of life-long health in canine populations.
# Introduction
Understanding the factors relating to disease in a population is important for anticipating and dealing with health care needs. The health of populations can be studied in a number of ways. Beyond descriptive approaches, analytical studies can be split into experimental and observational investigations. Dohoo et al. [bib_ref] Introduction to Observational Studies, Dohoo [/bib_ref] distinguished observational studies from experimental studies, where investigators control the allocation of subjects to study groups, by suggesting that in observational studies, investigators "try not to influence the natural course of events for the study subjects".
Epidemiologists traditional divide observational studies into case-control, cross-sectional or cohort study designs [bib_ref] Introduction to Observational Studies, Dohoo [/bib_ref]. The advantages and disadvantages of each of these study types, particularly with regard to susceptibility to bias, are fully described in [fig_ref] Table 1: The advantages and disadvantages of different observational study types [/fig_ref]. In brief however, case-control studies are particularly useful for rare diseases but lack an ability to clarify temporal relationships between events and exposures. Cross-sectional studies can be performed at a single time point and allow investigators to seek associations between potential risk factors and outcomes, but again do not allow the assessment of temporal dependencies. Cohort studies, where individuals are tracked through time, solve this problem as investigators can assess whether risk factor exposures are followed by outcomes in individuals. This element of time dependency is crucial to infer causation between risk factors and disease, and to understand transmission dynamics of infectious diseases. Further, cohort studies lend themselves to analysis of the effect of long-term exposure to a risk factor or treatment and, with targeted recruitment, are ideally suited to examine the effect of rare risk factors. Unfortunately cohort studies necessarily involve a large investment of time and finances, both to set up and maintain, and have therefore been used sparingly in the field of canine disease. In this review we will discuss studies found in broader medical literature before describing the types of canine cohort studies reported to date. The techniques found in human medicine may be applied to canine epidemiology with immense potential for health advances.
## Review
## The benefits of cohort studies: comparative examples
One of the most widely renowned cohort studies in human medicine is the Framingham Heart Study. Researchers recruited a group of over 5,000 women and men aged between 30 and 62 years old living in Framingham, Massachusetts in 1948. The cohort were evaluated every two years regarding their medical status and lifestyles, including physical examinations and collection of biological samples for laboratory testing. The study identified many of the major cardiovascular disease risk factors which we take for granted today, such as high blood pressure, high blood cholesterol, smoking, obesity and diabetes [bib_ref] The Framingham Heart Study and the epidemiology of cardiovascular disease: a historical..., Mahmood [/bib_ref]. The analysis of the Framingham cohort has resulted in over 2,000 peer reviewed publications, and aptly demonstrates how the detailed, repeated evaluation of modestly sized cohort groups can result in the identification of risk factors for disease which have global significance. Another early cohort study of human health was undertaken in the UK in 1951. The aim was to address concerns about an observed association between smoking and lung cancer. To examine the question of causality, the study was designed to determine whether it was possible to predict someone's risk of developing lung cancer from their smoking habits earlier in life [bib_ref] Tobacco: a medical history, Doll [/bib_ref]. Over 40,000 doctors were recruited, which was over two thirds of the doctors on the British Medical Register at the time. The study went on to investigate the impact of smoking on diseases beyond lung cancer, including vascular disease and other neoplasias [bib_ref] Mortality in relation to smoking: 50 years' observations on male British doctors, Doll [/bib_ref]. Ultimately the cohort was so valuable that the members were followed for their lifetime and the last questionnaire was sent out some 50 years later.
Two more recent studies which have illustrated the power of large scale cohort studies are the Avon Longitudinal Study of Parents and Children (ALSPAC) [bib_ref] Children of the nineties. A longitudinal study of pregnancy and childhood based..., Golding [/bib_ref] and the Italian NINFEA cohort [bib_ref] Feasibility of recruiting a birth cohort through the Internet: the experience of..., Richiardi [/bib_ref]. Both are birth cohorts, initially designed without specific hypotheses in mind. Instead they set out to collect information on a variety of exposures to broadly investigate pregnancy and the early life of children. In the case of ALSPAC, investigations went back even earlier, with assessment of antenatal risks, such as the impact of maternal drinking prior to conception and in early pregnancy on birth weight [bib_ref] The effect of maternal drinking before conception and in early pregnancy on..., Passaro [/bib_ref].
The ALSPAC study team faced great difficulty obtaining funding in the initial years of the project [bib_ref] History of Avon Longitudinal Study of Parents and Children (ALSPAC), c, Overy [/bib_ref]. As time passed and significant risk factors started to be found and reported, it became more widely recognised that the cohort was an incredible resource that should be maintained in the long-term. This open-ended investigative approach resulted in the identification of a range of phenotypes and influencing factors that could not have been predicted by the investigators at the start of the study. The costs of recruiting the cohort would have been wasted if contact with members were lost before these discoveries could be made.
Analyses of the ALSPAC cohort did not stop with exploration of early-life influences. As the costs of collecting, archiving and analysing DNA reduced it became possible to add genetic data to the wealth of phenotypic data and explore the interaction of genotype with other variables. Over more than 20 years the ALSPAC team moved from having a relationship with pregnant mothers to having a relationship with the children from those pregnancies. These children have grown to start their own families and the next generation are also being recruited into the study. A wealth of discoveries guiding national public health policy have been made during the study. These include understanding the influence of sleeping position on the risk of cot death [bib_ref] Cot deaths and sleep position campaigns, Golding [/bib_ref] [bib_ref] Does the supine sleeping position have Any adverse effects on the child?:..., Hunt [/bib_ref] and the benefits of eating oily fish on children's mental development [bib_ref] Fish intake during pregnancy and early cognitive development of offspring, Daniels [/bib_ref] , both of which have directly led to the development of guidelines for best practice.
Between 1996 and 2001, the Million Women Study recruited women over 50 in the UK. Recruitment through breast cancer screening centres built-in a reliable method of ascertaining the primary outcome of interestthe incidence of breast cancer. Environmental influences were captured in a lifestyle questionnaire that was completed at recruitment and periodically thereafter. Information regarding other disease events such as incidence of fractures was also collected via the followup questionnaires [bib_ref] Fracture incidence in relation to the pattern of use of hormone therapy..., Banks [/bib_ref].
The main finding from the Million Women Study regarding the impact of Hormone Replacement Therapy (HRT) on the incidence of breast cancerremains controversial. An increased incidence of breast cancer was found in the women taking HRT but it has subsequently been argued that these women were more likely to be tested for breast cancer, resulting in increased diagnoses. The women involved were not randomly assigned to receive HRT so the potential for confounding cannot be ignored. Nevertheless, the study built on results from earlier cross-sectional studies and it had enormous power to detect associations. As the women were followed with time, causal inference is possible. At the very least the results of the many publications about the cohort will influence the direction of future randomised controlled trials to try and definitively determine causal relationships.
The value of cohorts has been recognised and data collected previously are increasingly the foundation for further analysis. For example, a team from Edinburgh University took advantage of historic data collection to develop a cohort of people with results that span over 80 years. In Scotland 95% of children born in 1921 and 1936 were given an intelligence test at the age of 11. The team recruited a subset of the survivors from these tested cohorts some 60-70 years later to investigate their cognitive function [bib_ref] Cohort profile: the Lothian Birth Cohorts of 1921 and 1936, Deary [/bib_ref] and the environmental and genetic influences upon them. Their continuing assessment of cognitive function has led to the discovery of an association between carrying the APOE E4 allele (also associated with Alzheimer's disease) and non-pathological cognitive decline [bib_ref] APOE E4 status predicts age-related cognitive decline in the ninth decade: longitudinal..., Schiepers [/bib_ref]. The cohort is a unique resource for the investigation of the effects of aging on cognition and it continues as participants enter their tenth decade.
While the benefits of cohort studies are well understood [fig_ref] Table 1: The advantages and disadvantages of different observational study types [/fig_ref] , the extended time to finding results and relatively high costs are undeniable. In part to address these costs, the US Department of Defense started to move cohort studies into the internet age when they set up the Millennium Cohort Study of US military personnel [bib_ref] The millennium cohort study: a 21-year prospective cohort study of 140,000 military..., Gray [/bib_ref]. Current and ex-military personnel were recruited and offered the chance to answer the questionnaire by post or online. The financial savings associated with participants replying online were such that they offered a $5 incentive and still estimated their savings per online response at $50 compared to those responding by post [bib_ref] When epidemiology meets the internet: Web-based surveys in the millennium cohort study, Smith [/bib_ref]. As internet access has increased, epidemiological studies have gradually made greater use of the technology. The NINFEA cohort is based entirely online . Whilst the costs of setting up and maintaining functional and appealing web portals are not insignificant, studies are now possible that would not have been feasible if based on face-to-face, telephone or postal questionnaires. Building on this experience of human studies, canine cohort studies that would have been inconceivable are now financially viable and the potential to exploit this avenue of research is immense.
## Canine cohort studies
Despite the extensive number of findings uncovered by human cohort studies, the design has not been widely used in canine research in the past. As discussed, the cost and time burdens can be prohibitively high. A number of canine cohort studies have been reported and in each case attempts have been made to overcome the associated financial burden. The different strategies used are discussed below and their merits summarised in [fig_ref] Table 2: The advantages and disadvantages of different cohort study types [/fig_ref].
# Retrospective methods
Retrospective cohort studies involve looking back at individuals after the events of interest have occurred (for example disease incidence, death or pregnancy) and the follow-up period has ended. These studies can be undertaken on a large scale with relatively little lead time or up-front costs by using pre-existing databases such as those maintained by insurance companies and groups of secondary veterinary hospitals or primary clinics.
Insurance databases in particular are an extremely valuable resource and are discussed in detail by O'Neill et al. [bib_ref] Approaches to Canine health surveillance, O'neill [/bib_ref]. There is a long tradition of pet insurance in Sweden and Agria insure approximately 40% of Swedish dogs [bib_ref] Age patterns of disease and death in insured Swedish dogs, cats and..., Bonnett [/bib_ref]. Their database provides a powerful measure of events in the Swedish pet canine population [bib_ref] Validation of computerized Swedish dog and cat insurance data against veterinary practice..., Egenvall [/bib_ref]. Such large electronic resources offer the chance to study incidence rates and survival time from diagnosis for specific diseases, such as mammary tumours [bib_ref] H: Incidence of and survival after mammary tumors in a population of..., Egenvall [/bib_ref]. However there is no requirement for private companies to make their data available. When using insurance databases, there is likely to be bias relating to the non-random socioeconomic status of owners who insure their pets and to specific insurance policy exclusions such as preexisting conditions and age limits. In addition, in countries where dog insurance rates are low, the resource would be even less representative of the population as a whole.
Veterinary medical databases provide an alternative resource of information on the health of populations [bib_ref] Approaches to Canine health surveillance, O'neill [/bib_ref]. They have the advantage of that they can be linked to ancillary resources (such as radiographic archives and biological samples). However, the plethora of recording systems, and lack of agreement of diagnostic criteria for the definition of specific diseases, makes them cumbersome to use and extracting and extrapolating data is difficult. With modern textural mining tools there is scope to revisit this area [bib_ref] Approaches to Canine health surveillance, O'neill [/bib_ref] but the challenge of collating records from diverse recording systems remains. Further, when these databases rely solely on groups of specialist hospitals, there is the risk of referral bias as demonstrated by Bartlett et al. [bib_ref] Disease surveillance and referral bias in the veterinary medical database, Bartlett [/bib_ref].
Risk factor studies using both insurance and veterinary medical databases are also limited by the type of data collected. In both cases, the data refer to phenotype of the dog but not their wider environment. Postcode (location) data have been used to assess the spatial distribution of atopic dermatitis [bib_ref] The spatial distribution of atopic dermatitis cases in a population of insured..., Nødtvedt [/bib_ref] but the impact of the dogs' lifestyles is not available from such records. For example Glickman et al. [bib_ref] Association between chronic azotemic kidney disease and the severity of periodontal disease..., Glickman [/bib_ref] were able to investigate a link between severity of periodontal disease in dogs and subsequent chronic azotemic kidney disease (kidney disease causing high levels of blood urea and creatinine) because both diseases were recorded in clinical records, but environmental risk factors like diet could not be considered. This is a major limitation of such databases; otherwise their data on multiple disease outcomes, covering large numbers of dogs from different breeds, would be unparalleled in terms of potential for use in investigations.
# Prospective methods
Prospective studies are set up before the outcome of interest occurs and allow investigators to pre-select study subjects and specifically determine which data they wish to collect.
# Prospective methods: time-limited
Limiting the time at risk has been used to minimise the costs of studies where pre-existing data are not available. This also helps reduce bias through loss to follow up. A wealth of investigations have utilised this methodology, such as those investigating the spread of Leishmaniasis and other vector borne diseases in dog cohorts. Studies investigating disease incidence [bib_ref] Epidemiology of canine leishmaniasis: a comparative serological study of dogs and foxes..., Courtenay [/bib_ref] [bib_ref] Peridomestic risk factors for canine leishmaniasis in urban dwellings: new findings from..., Moreira [/bib_ref] [bib_ref] Cohort study on canine emigration and Leishmania infection in an endemic area..., Paranhos-Silva [/bib_ref] , detection methods [bib_ref] Longitudinal study on the detection of canine Leishmania infections by conjunctival swab..., Gramiccia [/bib_ref] [bib_ref] Incidence and time course of leishmania infantum infections examined by parasitological, serologic,..., Oliva [/bib_ref] [bib_ref] Diagnosis of canine vector-borne diseases in young dogs: a longitudinal study, Otranto [/bib_ref] [bib_ref] IgG subclass responses in a longitudinal study of canine visceral leishmaniasis, Quinnell [/bib_ref] [bib_ref] Evaluation of rK39 rapid diagnostic tests for canine visceral leishmaniasis: longitudinal study..., Quinnell [/bib_ref] and the impact of a culling regime [bib_ref] Assessment of an optimized dog-culling program in the dynamics of canine Leishmania..., Moreira [/bib_ref] have all used this approach. Cohort methodology was necessary in each study
## Prospective methods: single factor
If time is not constrained, then the focus or numbers of dogs in a study may be narrowed. Perhaps the best example of this comes from a study of dietary restriction using a small group of Labrador Retrievers (48 dogs) in an experimental setting. This controlled trial has yielded an array of findings on the effect of dietary restriction on mortality [bib_ref] Effects of diet restriction on life span and age-related changes in dogs, Kealy [/bib_ref] [bib_ref] Influence of lifetime food restriction on causes, time, and predictors of death..., Lawler [/bib_ref] , immune function [bib_ref] Modulation of canine immunosenescence by life-long caloric restriction, Greeley [/bib_ref] , and developmental joint disease [bib_ref] A longitudinal study of the influence of lifetime food restriction on development..., Huck [/bib_ref] [bib_ref] Effects of limited food consumption on the incidence of hip dysplasia in..., Kealy [/bib_ref] [bib_ref] Five-year longitudinal study on limited food consumption and development of osteoarthritis in..., Kealy [/bib_ref] [bib_ref] Evaluation of the effect of limited food consumption on radiographic evidence of..., Kealy [/bib_ref] [bib_ref] Use of the caudolateral curvilinear osteophyte as an early marker for future..., Powers [/bib_ref] [bib_ref] The effects of lifetime food restriction on the development of osteoarthritis in..., Runge [/bib_ref] [bib_ref] Lifelong diet restriction and radiographic evidence of osteoarthritis of the hip joint..., Smith [/bib_ref] [bib_ref] Chronology of hip dysplasia development in a cohort of 48 Labrador retrievers..., Smith [/bib_ref] [bib_ref] Evaluation of a circumferential femoral head osteophyte as an early indicator of..., Szabo [/bib_ref]. The time-span and depth of this trial (including blood sampling and radiography at regular intervals) made it prohibitively expensive to perform on a larger scale but data on specific aspects, such as the life-long progression of osteoarthritis, could only be collected by following a cohort longitudinally in this manner. Dobson et al. [bib_ref] Mortality in a cohort of flatcoated retrievers in the UK, Dobson [/bib_ref] undertook a study with a similarly narrow focus but were able to recruit dogs from the normal pet population in the UK. Following 174 flat-coated retrievers for up to 10 years they investigated the impact of neoplasia on mortality in that breed. Costs were also minimised in this case by contacting recruited owners just once per year for a health update and asking them to proactively contact the investigators if their dog fell ill. The study demonstrated that over 40% of the dogs died as a result of neoplasia, reducing their lifespan by three years compared to those that died from other causes.
Recruiting a large enough cohort to give the required power for an investigation and retaining that cohort to minimise bias are both key to the success of populationbased cohort studies. Thrusfield et al. [bib_ref] Acquired urinary incontinence in bitches: its incidence and relationship to neutering practices, Thrusfield [/bib_ref] studied a cohort of bitches for up to five years in an attempt to assess the impact of neutering on urinary incontinence. The onus for recruiting and maintaining the cohort was placed on volunteering veterinary surgeons. Perhaps because of this responsibility, some difficulty was encountered recruiting veterinarians to participate; whilst 233 initially agreed, only 16 went on to return data (a 7% response rate). The authors made every effort to minimise bias through randomisation techniques but the potential impact of selection bias on the study should not be overlooked.
Each veterinarian was asked to recruit 40 female puppies from their practices. Should these bitches subsequently become incontinent then they were no longer followed, whilst, by design, the remaining (continent) cohort were to be followed for five years. The veterinarians received letters encouraging them to continue with the study at one and three years, and a request to contact the involved owners to check that their dogs were not incontinent after five years. The authors cite slow initial recruitment as the main reason why only 504 dogs from an original 809 were followed for the full five years. They do not directly address how many of the remaining 305 dogs were lost to follow-up (only 22 developed incontinence), but the potential impact of retention bias on their results cannot be ignored. Nevertheless, by focussing on a single phenotype and spreading the responsibility for dealing with recruited animals amongst a number of veterinarians, it was possible to follow enough dogs to determine that neutered bitches had a risk of urinary incontinence that was nearly eight-fold that of intact bitches.
## Prospective methods: hypothesis generation
Beyond studies that focus on one disease or one exposure, there has been a movement in canine epidemiology toward the broader studies undertaken in human medicine such as the example of ALSPAC mentioned above [bib_ref] History of Avon Longitudinal Study of Parents and Children (ALSPAC), c, Overy [/bib_ref]. These studies do not necessarily aim to test a single hypothesis but rather gather data to identify new areas of investigation. In canine medicine, questionnaires have been developed that cover a wide range of potential exposures and disease outcomes and they are directed at breeders, owners and veterinarians. These studies have the disadvantage of relying on non-standardised data inputs where each animal is assessed by a different person with disparate (or no) training. However the studies are able to recruit more participants, and their subjects are more representative of dog lifestyle in the wider population than those followed under controlled conditions.
A 10-year cohort study of pedigree Boxers in the Netherlands recruited over 90% of the litters born in 14 months of 1994-5, initially comprising 2629 puppies. The study used diary-format records and face-to-face assessment with the breeders but moved on to sixmonthly questionnaires with owners. Pre-weaning mortality [bib_ref] Investigation of mortality and pathological changes in a 14-month birth cohort of..., Nielen [/bib_ref] [bib_ref] Evaluation of genetic, common-litter, and within-litter effects on preweaning mortality in a..., Van Der Beek [/bib_ref] and post-weaning mortality [bib_ref] Life expectancy in a birth cohort of Boxers followed up from weaning..., Van Hagen [/bib_ref] were assessed and, due to the large numbers of dogs involved in the study, all with pedigree information, the investigators were able to make heritability estimates for phenotypes [bib_ref] Heritability estimations for diseases, coat color, body weight, and height in a..., Nielen [/bib_ref] and common diseases such as cryptorchidism (failure of one or both testes to descend to the scrotum), cranial cruciate disease (degeneration of the cranial cruciate ligament) and epilepsy [bib_ref] Genetic and epidemiological investigation of a birth cohort of boxers, Nielen [/bib_ref] (a neurological disease characterised by the development of seizures) and hip dysplasia (a developmental malformation of the hip joints) [bib_ref] Incidence, risk factors, and heritability estimates of hind limb lameness caused by..., Van Hagen [/bib_ref].
Similarly a group in Norway followed a cohort of 700 dogs from four large breeds. Again they gave questionnaires to breeders and owners but they also involved the dogs' veterinarians. To date they have published studies on the prevalence and risk factors of neonatal mortality [bib_ref] Canine neonatal mortality in four large breeds, Indrebø [/bib_ref] , the effect of weight and growth rates on the development of hip dysplasiaand the incidence and risk factors associated with vomiting and diarrhoea [bib_ref] A longitudinal study on diarrhoea and vomiting in young dogs of four..., Saevik [/bib_ref].
Relatively newly created is the Dogslife Project, which is focussed on the owners of Kennel Club registered Labrador Retrievers in the UK [bib_ref] Dogslife: a webbased longitudinal study of Labrador Retriever health in the UK, Clements [/bib_ref]. It is limiting costs by utilising a website-based questionnaire and has recruited over 4,200 dogs in three and a half years. As a prospective study, it was possible to specifically tailor the questionnaire to address areas of interest. Data collection includes detail regarding phenotype and lifestyle which will be examined with reference to dog health. Like the studies in Norway and the Netherlands, the Dogslife Project is an attempt to develop a large-scale cohort of dogs with thoroughly documented history, similar to those cohorts found in human medicine.
## The future of canine cohort research
With the relative dearth of cohort studies in canines to date, there is scope to address new questions in the future. For example, the cohort of Dutch Boxer dogs discussed earlier were reported to have a pre-weaning mortality rate over 20% [bib_ref] Investigation of mortality and pathological changes in a 14-month birth cohort of..., Nielen [/bib_ref]. Such a loss is a clear welfare problem for dogs and more detailed studies of potential risk factors could have a great impact. Indrebø et al. [bib_ref] Canine neonatal mortality in four large breeds, Indrebø [/bib_ref] , Nielen et al. [bib_ref] Investigation of mortality and pathological changes in a 14-month birth cohort of..., Nielen [/bib_ref] and van der Beek et al. [bib_ref] Evaluation of genetic, common-litter, and within-litter effects on preweaning mortality in a..., Van Der Beek [/bib_ref] each address early mortality through cohort studies but their findings focus largely on factors from birth onward. Van der Beek et al. [bib_ref] Evaluation of genetic, common-litter, and within-litter effects on preweaning mortality in a..., Van Der Beek [/bib_ref] included an analysis of inbreeding coefficients but found that genetic effects in general had less effect than environmental effects at puppy and litter level. Relatively short cohort studies including the lifestyle of the dam prior to birth may shed new light on risk factors associated with both stillbirths and early mortality, minimising distress in owners who are currently unable to prevent early losses.
Cohort studies such as those undertaken in human medicine could play a vital role if veterinarians are to be able to offer advice to owners on minimising the risks of developing disease and injury. Beyond death in very early life, morbidity and mortality in dogs in developed nations reflects the epidemiological shift in morbidity and mortality in human medicine from infectious diseases to noncommunicable diseases. This shift is increasingly relevant in canine health as vaccination, antibiotics and better veterinary care ensure that more dogs in developed nations live to suffer from developmental diseases and diseases of aging. Bonnett et al. [bib_ref] Mortality in insured Swedish dogs: rates and causes of death in various..., Bonnett [/bib_ref] demonstrated that whilst the highest mortality rate in dogs over six weeks of age in Sweden was trauma (typically car accidents), the next highest rate was due to tumours, followed by locomotor problems. Cohort studies of canine lifestyle have the power to investigate the risk factors associated with developing these noncommunicable diseases, facilitated by the release of a draft canine genome sequence [bib_ref] The dog genome: survey sequencing and comparative analysis, Kirkness [/bib_ref] [bib_ref] Genome sequence, comparative analysis and haplotype structure of the domestic dog, Lindblad-Toh [/bib_ref] and the increasing access to high density genotyping and eventually low cost whole genome sequencing. Since the dog has a shorter lifespan than humans, associations between genetic variation and disease that are also relevant to human aging are likely to be revealed.
Human medicine is again ahead of the veterinary field with regard to incorporating biological data in cohort studies. UK Biobank that have recruited 500,000 people between 40-69 years of age. The investigative team collect blood, saliva and urine samples, phenotypic data and the agreement of all participants to have their health status followed. The collection of genetic information in particular adds a new element to the traditional cohort study, and with such a large cohort the potential power to detect risk factors involving genetic-environmental interactions is enormous. Projects on such a scale are currently financially prohibitive in dogs, but projects on a smaller scale such as Dogslifehave collected buccal swabs for DNA extraction from a subset of their cohort enabling comparisons of genotype with phenotype. Should such sampling be repeated throughout the lives of the dogs, investigations of genetic and epigenetic changes throughout that lifetime and comparison with concurrent phenotypic data would potentially permit investigators to seek environmental factors associated with genetic variation, disease and aging. The merging of lifestyle and whole genome data should increasingly reveal associations between genotype and environment in the dog and ultimately in humans as well.
# Conclusions
Cohort studies have already yielded results in the field of canine health. With the advent of large databases and internet technology the costs of such studies are being reduced to the point whereby large-scale studies are possible in canine populations. The potential to identify risk factors and inform an evidenced-based medicine approach to preventative health measures in dogs mean that cohort studies can have a great impact on dog health and welfare. Given how long it takes to achieve results from prospective studies, the time to start is now.
Abbreviations ALSPAC: Avon longitudinal study of parents and children; HRT: Hormone replacement therapy.
[table] Table 1: The advantages and disadvantages of different observational study types [/table]
[table] Table 2: The advantages and disadvantages of different cohort study types [/table]
|
Intron Retention in the 5′UTR of the Novel ZIF2 Transporter Enhances Translation to Promote Zinc Tolerance in Arabidopsis
## Figure s4
Figure S4 ZIF2 expression in the Arabidopsis zif2-1 mutant. RT-PCR analysis of ZIF2 expression in 14-d old wild-type (Col-0) and mutant (zif2-1) seedlings. The location of the F1', R1, F2, R2, F3 and R3 primers used is shown in . Expression of the UBQ10 gene was used as a loading control. Results are representative of three independent experiments.
## Zif2
## Ubq10
Col-0 zif2-1
## F1'/r1
F3/R3 F2/R2 |
MPER-specific antibodies induce gp120 shedding and irreversibly neutralize HIV-1
Interference with virus entry is known to be the principle mechanism of HIV neutralization by antibodies, including 2F5 and 4E10, which bind to the membrane-proximal external region (MPER) of the gp41 envelope protein. However, to date, the precise molecular events underlying neutralization by MPER-specific antibodies remain incompletely understood. In this study, we investigated the capacity of these antibodies to irrevocably sterilize HIV virions. Long-term effects of antibodies on virions can differ, rendering neutralization either reversible or irreversible. MPER-specific antibodies irreversibly neutralize virions, and this capacity is associated with induction of gp120 shedding. Both processes have similar thermodynamic properties and slow kinetics requiring several hours. Antibodies directed to the CD4 binding site, V3 loop, and the MPER can induce gp120 shedding, and shedding activity is detected with high frequency in plasma from patients infected with divergent genetic HIV-1 subtypes. Importantly, as we show in this study, induction of gp120 shedding is closely associated with MPER antibody inhibition, constituting either a primary event leading to virion neutralization or representing an immediate consequence thereof, and thus needs to be factored into the mechanistic processes underlying their activity.
## Br ief definitive repor t
Neutralization of HIV by antibodies is generally attributed to antibody occupancy of the envelope trimers and interference with viral attachment to host cell receptors or entry, but the precise underlying molecular mechanisms leading to neutralization by most neutralizing antibodies identified to date await clarification . A considerable effort in the development of HIV vaccines has been directed toward eliciting neutralizing antibody responses that mimic activities of the membrane-proximal external region (MPER)specific mAbs 2F5 and 4E10. Information on the structural composition of the MPER has broadened in recent years, and studies of epitope accessibility have highlighted the possibility that neutralization by MPER-specific antibodies may require the recognition of their epitopes in the context of membrane lipids [bib_ref] Structural basis of enhanced binding of extended and helically constrained peptide epitopes..., Cardoso [/bib_ref] [bib_ref] The membrane-proximal external region of the human immunodeficiency virus type 1 envelope:..., Montero [/bib_ref] [bib_ref] Role of HIV membrane in neutralization by two broadly neutralizing antibodies, Alam [/bib_ref] and may depend on the formation of the prehairpin intermediate state during viral entry [bib_ref] A fusion-intermediate state of HIV-1 gp41 targeted by broadly neutralizing antibodies, Frey [/bib_ref] [bib_ref] Distinct conformational states of HIV-1 gp41 are recognized by neutralizing and non-neutralizing..., Frey [/bib_ref] [bib_ref] Role of HIV membrane in neutralization by two broadly neutralizing antibodies, Alam [/bib_ref]. Whether these antibodies can act on free virus in the absence of target cells and receptor engagement remains uncertain, and thus the precise window of action of these antibodies, their modes of epitope recognition, and a detailed mechanism of neutralization still await further definition.
Our understanding of the biochemical activity and biological function of neutralizing antibodies is shaped by studies that have assessed the initial interaction between antibodies and HIV and their capacity to block virus entry in relatively short-term experimental settings but have left the fate of neutralized virions activity was probed by preincubating serial dilutions of mAbs b12, 2G12, 2F5, and 4E10 with 11 different R5 and X4 pseudoviruses (Tables S1 and S2) for 1 h or overnight (o/n; 16-18 h) at 37°C before infection of TZM-bl target cells. The 50% inhibitory concentration (IC 50 ) for each mAb/treatment condition is depicted. IC 50 values were derived from sigmoid dose-response curve fits of pooled data of two to three independent experiments. The number of viruses not neutralized (NN) at the highest mAb concentration probed is indicated on top of each antibody panel. (B) Decay of HIV upon long-term exposure to 37°C. A fixed virus input (JR-FL enveloped-pseudotyped Br ief Definitive Repor t induced gp120 shedding upon prolonged contact with the virus, rendering neutralization irreversible. The kinetic and thermodynamic requirements of the shedding process were virtually identical to those of neutralization, identifying gp120 shedding as a key process associated with HIV neutralization by MPER antibodies.
## Results and discussion kinetics of hiv neutralization
To study the effects of long-term interaction of HIV with neutralizing antibodies, we compared short-(1 h) and longterm (overnight; 16-18 h) inhibitory activity of the neutralizing antibodies b12 (gp120 CD4 binding site [CD4bs] specific; [bib_ref] Recombinant human Fab fragments neutralize human type 1 immunodeficiency virus in vitro, Barbas [/bib_ref] , 2G12 (gp120 carbohydrate specific; [bib_ref] Human monoclonal antibody 2G12 defines a distinctive neutralization epitope on the gp120..., Trkola [/bib_ref] , and the gp41 MPER-specific antibodies 2F5 [bib_ref] A conserved neutralizing epitope on gp41 of human immunodeficiency virus type 1, Muster [/bib_ref] and 4E10 [bib_ref] Broadly neutralizing antibodies targeted to the membrane-proximal external region of human immunodeficiency..., Zwick [/bib_ref] against a panel of envelope-pseudotyped luciferase reporter viruses [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref]. Although long-term coincubation of virus with mAb 2G12 showed no or only modest increases in neutralization, prolonged coincubation markedly improved the neutralizing capacity of 2F5, 4E10, and b12. As noted previously [bib_ref] Comprehensive cross-clade neutralization analysis of a panel of anti-human immunodeficiency virus type..., Binley [/bib_ref] , neutralization activity of MPER mAbs, in particular, increased by more than one order of magnitude upon extended interaction with several of the tested isolates. We ascertained that these increases in inhibitory activity upon prolonged interaction with the virus were a consequence of the antibodies' action and not a result of spontaneous decay of virions. Although HIV infectivity decreases upon prolonged incubation at 37°C, largely unexplored. HIV neutralization is commonly assessed under conditions that do not allow discrimination between actions of the neutralizing antibody on the free virus and actions after receptor engagement, as antibody is usually present throughout all steps, including preincubation and infection. In this study, we specifically investigated the actions of MPER mAbs on virions in the absence of target cell interactions to determine whether and to what extent this type of antibody can contribute to clearance of virus particles in vivo.
Whether neutralization causes an irreversible deactivation of HIV or whether virions can regain activity after antibody dissociation may also significantly impact the in vivo efficacy of a given antibody. In vivo virions can persist for extended periods of time before encountering appropriate target cells or being cleared by phagocytes. The sequestration of HIV by DCs is well documented in this respect. Trapped by DCs, the virus can remain infectious for several days and efficiently be transferred to CD4 + T cells [bib_ref] Immunodeficiency virus uptake, turnover, and 2-phase transfer in human dendritic cells, Turville [/bib_ref] [bib_ref] HIV traffics through a specialized, surface-accessible intracellular compartment during transinfection of T..., Yu [/bib_ref]. Should neutralization be reversible, trafficking of antibody-opsonized virus to intracellular compartments or anatomical sites with lower antibody concentration could potentially lead to antibody dissociation and reconstitute the virus's infectivity. Achieving irreversible neutralization is thus clearly desirable for both natural and vaccine-elicited immune responses.
In this study, we aimed to obtain insight into the longterm effect of broadly neutralizing antibodies on cell-free HIV particles and their capacity to irreversibly inactivate the virus. Most notably, the MPER-specific antibodies potently virus; 3,000 TCID 50 /well in culture medium; no mAbs) was incubated for the indicated time periods at 37°C before infection of TZM-bl target cells, and virus infectivity was monitored by luciferase reporter gene production 72 h after infection. Relative light units (RLU) measured per well are depicted. Values are means of 15 replicas, and error bars indicate SEM. One out of five independent experiments is shown. (C) mAb neutralization is constant over a wide range of viral input. Serial dilutions of JR-FL envelope-pseudotyped virus were pretreated for 1 h at 37°C with a fixed concentration of each mAb at concentrations that typically yield neutralization below 70% (0.02 µg/ml mAb b12, 1 µg/ml 2G12, 1.5 µg/ml 2F5, and 7 µg/ml 4E10). In parallel, for each virus dilution, a mock treatment (incubation without mAb) was performed. The viral infectivity of each virus dilution (relative light units of mock-treated sample, measured 72 h after infection) is plotted on the x axis. The y axis indicates the neutralization each antibody achieved at the respective virus input. Error bars indicate SEM. (D) Enhanced mAb neutralization over prolonged incubation is not caused by spontaneous virus decay. In experiments depicted by red symbols, virus input and mAb concentration were kept constant, and only incubation time was varied. In experiments depicted by black symbols, incubation time (1 h) and mAb concentration were kept constant, and only virus input was varied. mAbs were probed at concentrations that typically yield neutralization below 70% (0.02 µg/ml mAb b12, 1 µg/ml 2G12, 1.5 µg/ml 2F5, and 7 µg/ml 4E10). For the red symbols, JR-FL pseudo virus was pretreated with the respective mAbs for the indicated times at 37°C. In parallel, for each time point, a mock control (incubation in medium without mAb) was performed. The viral infectivity measured for the mock treatment is plotted on the x axis for each time point (relative light units, measured 72 h after TZM-bl cell infection). The y axis indicates the neutralization each antibody achieved at after the respective preincubation time. For the black symbols, JR-FL pseudo virus was titrated and preincubated for 1 h at 37°C with the respective mAbs or a mock control (incubation in medium without mAb). The viral infectivity of the mock-treated samples is plotted on the x axis for the respective virus input. The y axis indicates the neutralization activity each antibody achieved at the respective virus input. Data points are means of 20 replicates. Error bars indicate SEM. One of two independent experiments is shown. (E) Increases in inhibitory activity upon prolonged interaction with the virus are a consequence of the antibodies' action. Envelope-pseudotyped JR-FL was incubated with serial dilutions of b12, 2G12, 2F5, and 4E10 or left untreated for either 1 or 18 h. Alternatively, virus was left untreated for 17 h before being treated for 1 h with neutralizing antibodies. Pretreated virus was then added to TZM-bl cells, and virus infectivity was measured by luciferase reporter gene production 72 h after infection. The 50% inhibitory concentration (IC 50 ) for each mAb/treatment condition is depicted. IC 50 values were derived from sigmoid dose-response curve fits of pooled data of three independent experiments. (F) Time of addition experiments were conducted to probe the kinetics of mAb neutralization. Antibody concentrations were chosen to yield neutralization of 70% after 20 h of preincubation to allow for the monitoring of increases in neutralization activity over time. 0.03 µg/ml b12, 12 µg/ml 2G12, 2.5 µg/ml 2F5, 8.5 µg/ml 4E10, and 0.1 µg/ml CD4-IgG2 were incubated with JR-FL pseudovirus for the indicated time periods at 37°C. Percent neutralization of virus infectivity on TZM-bl cells was calculated in reference to the respective mock-treated virus control of each time point. One of six independent experiments is shown. Data points depict means and SEM of duplicate measurements. a reversible event. To probe reversibility, we generated JR-FL pseudovirus that coexpresses mouse CD4, allowing rapid separation of virions and unbound antibodies using magnetic beads [fig_ref] Figure 2: Reversibility of HIV neutralization [/fig_ref]. We subjected virions to neutralizing antibodies, followed by the opportunity for antibody dissociation and removal of unbound antibody (treatment A). Two sets of controls determined the extent of neutralization by antibody-treated virus without dissociation (treatment B) and the maximal inhibitory activity without removal of antibody (treatment C). Most interestingly, neutralization by the MPER mAbs proved to be irreversible. Removal of excess antibody (treatment B) led to only a small reduction in 2F5 and 4E10 neutralization activity (24% and 22%, respectively), likely accounting for ongoing neutralization when mAbs were not removed. Importantly, MPER mAb-opsonized virus remained stably neutralized over the 11-h dissociation period (treatment A). Virtually complete irreversibility of neutralization was observed for mAb b12, as the virus remained neutralized throughout the observation period. In contrast, 2G12 immediately lost the majority of its neutralization activity once excess antibody was removed (treatment B), and the virus was able to regain its infectivity almost completely (treatment A), indicating that neutralization by 2G12 is reversible and steered by a high off rate in binding to the HIV envelope.
## Neutralizing antibodies induce gp120 shedding
Having established that certain neutralizing antibodies irrevocably inactivate HIV, we investigated plausible mechanisms of this process. To probe whether neutralizing antibodies induce conformational changes in the envelope trimer that irrevocably perturb functionality, we assessed the capacity of opsonized JR-FL pseudovirus to bind to 3D6, an antibody specific for the immunodominant (cluster I) region in gp41 inaccessible on intact envelope trimers [bib_ref] Epitope exposure on functional, oligomeric HIV-1 gp41 molecules, Sattentau [/bib_ref] [bib_ref] Antigenic properties of the human immunodeficiency virus transmembrane glycoprotein during cell-cell fusion, Finnegan [/bib_ref]. Soluble CD4 (sCD4) treatment is known to induce conformational changes within gp120 [bib_ref] Structure of an HIV gp120 envelope glycoprotein in complex with the CD4..., Kwong [/bib_ref] and to cause shedding of gp120 [bib_ref] Dissociation of gp120 from HIV-1 virions induced by soluble CD4, Moore [/bib_ref] , both of which lead to exposure of otherwise hidden domains in gp41 [bib_ref] Epitope exposure on functional, oligomeric HIV-1 gp41 molecules, Sattentau [/bib_ref] [bib_ref] Antigenic properties of the human immunodeficiency virus transmembrane glycoprotein during cell-cell fusion, Finnegan [/bib_ref]. In line with this, we observed a strong increase in 3D6 binding (7.5-fold) to virions upon sCD4 treatment . Most strikingly, treatment with mAbs 2F5 or 4E10 also led to a pronounced exposure of cluster I epitopes on virions, whereas the gp41targeting inhibitor T-20 had no such effect, suggesting that MPER mAbs either induced gp120 shedding or caused an extensive structural reorganization of gp120 and gp41 molecules within the trimer.
To directly probe for gp120 shedding, JR-FL pseudovirus was first treated with neutralizing antibodies. After bead immobilization (via coexpressed mouse CD4), unbound antibody and potentially shed gp120 was removed, and virus-associated gp120 content was quantified . Comparing mock-treated with CD4-IgG2 (tetrameric CD4 Ig)-and neutralizing mAb-treated JR-FL pseudovirus, we observed remarkable potency of the MPER mAbs to induce gp120 shedding upon long-term interaction with the virus. Likewise, antibody b12 induced gp120 this loss is moderate (<1 log within 24 h; [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref]. To study HIV neutralization by mAbs in vitro, antibodies are commonly supplied in stoichiometric excess over the viral target, allowing robust assessment of neutralization activity over a wide range of virus input. This was also true in our assay setup, in which we measured the virtually identical neutralization capacity of b12, 2G12, 2F5, and 4E10 over a >2-log range of virus inputs [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref]. Assessment of mAb-independent decrease in virus infectivity and the enhanced mAb neutralization activity during long-term incubation in parallel confirmed that the moderate loss observed in infectivity does not impact the assessment of neutralization activity in our experimental setup [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref]. Increases in the neutralization capacity of 2F5, 4E10, and b12 upon prolonged incubation are thus caused by the antibodies' action. However, under the same conditions, long-term treatment by 2G12 induced no equivalent improvement in neutralization. To formally exclude the influence of virus decay, we performed a series of further control experiments [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref] in which virus was subjected to a 1-(condition 1) and 18-h (condition 2) preincubation with mAbs at 37°C. Condition 3 controlled for viral decay during the 18-h incubation by incubating the virus, in absence of mAbs, for 17 h at 37°C, followed by a 1-h mAb treatment. The latter confirmed that viral decay had no general influence on neutralization activity, as conditions 1 and 3 resulted in identical inhibitory activity.
We next defined the specific kinetics of mAb binding to virions [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref] , A and B) and neutralization [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref] in time of addition experiments. To monitor increases in neutralization activity over time, fixed antibody doses that typically yield 70% neutralization of JR-FL pseudovirus only after 20-h preincubation were chosen. 2G12 rapidly bound to virions and instantly neutralized, but this activity only moderately improved upon sustained treatment, suggesting that easy access to its well exposed glycan epitope fosters a comparatively fast and efficient binding [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref] and neutralization reaction. In contrast, both binding and neutralization by b12 and the MPER mAbs gradually and substantially increased over time. The latter clarifies a longstanding debate as it confirms that 2F5 and 4E10 can access the MPER on virions before receptor engagement. Yet, in agreement with previous studies, the slow kinetics of binding and neutralization (>7 h; [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref] suggest that access to the MPER epitope is highly restricted [bib_ref] Redox-triggered infection by disulfide-shackled human immunodeficiency virus type 1 pseudovirions, Binley [/bib_ref] [bib_ref] A fusion-intermediate state of HIV-1 gp41 targeted by broadly neutralizing antibodies, Frey [/bib_ref]. Ineffective binding could be a consequence of steric interference of antibody binding to the MPER epitope, which is thought to be partially buried in the viral membrane. In fact, efficient MPER mAb binding may depend on a prior association of the mAbs with the viral membrane [bib_ref] Structural basis of enhanced binding of extended and helically constrained peptide epitopes..., Cardoso [/bib_ref] [bib_ref] Role of HIV membrane in neutralization by two broadly neutralizing antibodies, Alam [/bib_ref] [bib_ref] Crystal structure of HIV-1 gp41 including both fusion peptide and membrane proximal..., Buzon [/bib_ref] [bib_ref] Distinct conformational states of HIV-1 gp41 are recognized by neutralizing and non-neutralizing..., Frey [/bib_ref]. Alternatively, envelope conformations that expose the MPER may only infrequently be formed.
## Is neutralization by antibodies reversible?
One specific purpose of studying long-term effects of antibody action was to determine whether neutralization of HIV is Br ief Definitive Repor t [bib_ref] Dissociation of gp120 from HIV-1 virions induced by soluble CD4, Moore [/bib_ref] [bib_ref] Mechanisms of human immunodeficiency virus Type 1 (HIV-1) neutralization: irreversible inactivation of..., Mcdougal [/bib_ref] [bib_ref] Neutralizing antibodies to human immunodeficiency virus type-1 gp120 induce envelope glycoprotein subunit..., Poignard [/bib_ref]. Although sCD4 has been long known to induce dissociation of gp120 [bib_ref] Dissociation of gp120 from HIV-1 virions induced by soluble CD4, Moore [/bib_ref] , to what extent shedding contributes to sCD4 inhibitory activity has never been fully unraveled as these processes can differ in concentration and temperature dependence [bib_ref] Mechanisms of human immunodeficiency virus Type 1 (HIV-1) neutralization: irreversible inactivation of..., Mcdougal [/bib_ref] [bib_ref] Envelope glycoprotein incorporation, not shedding of surface envelope glycoprotein (gp120/SU), is the..., Chertova [/bib_ref]. Importantly, sCD4 mediates weaker shedding and neutralization of PBMCderived viruses than T cell line-adapted strains, likely accounting for the failure of recombinant CD4-based therapies [bib_ref] High concentrations of recombinant soluble CD4 are required to neutralize primary human..., Daar [/bib_ref] [bib_ref] Virions of primary human immunodeficiency virus type 1 isolates resistant to soluble..., Moore [/bib_ref] [bib_ref] Two mechanisms of soluble CD4 (sCD4)-mediated inhibition of human immunodeficiency virus type..., Orloff [/bib_ref]. Previous investigations into shedding as a potential mechanism for antibody neutralization of HIV yielded inconclusive results [bib_ref] Dissociation of gp120 from HIV-1 virions induced by soluble CD4, Moore [/bib_ref] [bib_ref] Mechanisms of human immunodeficiency virus Type 1 (HIV-1) neutralization: irreversible inactivation of..., Mcdougal [/bib_ref] [bib_ref] Neutralizing antibodies to human immunodeficiency virus type-1 gp120 induce envelope glycoprotein subunit..., Poignard [/bib_ref]. MPER-specific antibodies had not been investigated, and neutralizing CD4bs mAbs (including b12) were reported to lack shedding activity [bib_ref] Mechanisms of human immunodeficiency virus Type 1 (HIV-1) neutralization: irreversible inactivation of..., Mcdougal [/bib_ref] [bib_ref] Neutralizing antibodies to human immunodeficiency virus type-1 gp120 induce envelope glycoprotein subunit..., Poignard [/bib_ref]. Although several mAbs directed to other domains in gp120 were found to partially dissociate the envelope of T cell line-adapted strains and a direct association between antibody neutralization and gp120 shedding was inferred to, a causal link between these modes of action was not formally experimentally established [bib_ref] Neutralizing antibodies to human immunodeficiency virus type-1 gp120 induce envelope glycoprotein subunit..., Poignard [/bib_ref]. Thus, the relevance of shedding as a mechanism of HIV neutralization by antibodies has remained to some extent uncertain. The potential in vivo significance of the shedding process was equally unclear as activity against primary viruses was not confirmed [bib_ref] Mechanisms of human immunodeficiency virus Type 1 (HIV-1) neutralization: irreversible inactivation of..., Mcdougal [/bib_ref] [bib_ref] Neutralizing antibodies to human immunodeficiency virus type-1 gp120 induce envelope glycoprotein subunit..., Poignard [/bib_ref].
shedding, whereas 2G12 had no such effect. Of note, in contrast to CD4-IgG2, which induced rapid shedding, short-term incubation with the probed neutralizing mAbs had only minimal influence on gp120 dissociation . The latter may account for the difficulty of earlier studies to identify induction of gp120 shedding by CD4bs antibodies, including b12 Antibody concentrations were chosen to yield neutralization activities of 90% after 11 h of preincubation to allow for monitoring decreases in neutralization activity after the dissociation step. JR-FL neutralization by 0.1 µg/ml b12, 15 µg/ml 2G12, 11.4 µg/ml 2F5, and 45 µg/ml 4E10 was probed under the indicated treatment conditions: after antibody dissociation (treatment A), neutralization without dissociation (treatment B), and without removal of antibody (treatment C). Means and SEM of triplicates measurements of one of three independent experiments are shown. Abs, antibodies.
period and compared these activities with those in a conventional assay format, in which antibody was present throughout . S1, C and D). The results were strikingly clear; neither 2F5 nor 4E10 were capable of neutralizing JR-FL when only present during the 1-h preincubation period. They only neutralized when either was allowed to interact with the cell free virus for prolonged periods or when present during both preincubation and the entire infection period. These data confirm that neutralization of free virions by MPER antibodies in the absence of receptor interactions requires prolonged interaction of the mAbs with the HIV particle and thus follows the reactivity pattern observed for MPER-induced gp120 shedding.
Our finding that MPER mAbs induce gp120 shedding is of obvious interest not only as it provides a long sought for insight into their mechanism of action but also as this process leads to irreversible neutralization of HIV. We confirmed our Considering that MPER antibodies even at high concentration (100 µg/ml) did not induce shedding after short-term (1 h) preincubation , whereas after 1-h preincubation, potent inhibition by the same mAbs can be seen in a conventional neutralization assay [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref] , it was prudent to explore whether shedding is a mere artifact of the mAb's interaction with the virus or is intimately connected with the neutralization process. In conventional neutralization assays, antibody is present during the preincubation period and also during the entire infection process. Thus, for MPER antibodies, for which an action after receptor engagement has been postulated [bib_ref] Structural basis of enhanced binding of extended and helically constrained peptide epitopes..., Cardoso [/bib_ref] [bib_ref] A fusion-intermediate state of HIV-1 gp41 targeted by broadly neutralizing antibodies, Frey [/bib_ref] [bib_ref] Distinct conformational states of HIV-1 gp41 are recognized by neutralizing and non-neutralizing..., Frey [/bib_ref] [bib_ref] Role of HIV membrane in neutralization by two broadly neutralizing antibodies, Alam [/bib_ref] , experimental conditions have to be chosen that allow discrimination between actions on the free virus and actions that occur in the context of virus-host cell interactions. We thus performed neutralization experiments in which virus was only exposed to antibody during the preincubation . MPER-specific antibodies induce gp120 shedding. (A) To assess exposure of gp41 epitopes upon neutralizing antibody treatment of virions, the capacity of antibody-treated JR-FL pseudovirus to bind to the gp41 cluster I mAb 3D6 was assessed. Virus was pretreated for 18 h with 10 µg/ml sCD4, 100 µg/ml 2F5, 100 µg/ml 4E10, or 10 µg/ml of the fusion inhibitor T-20 or left untreated (medium control). Virus was then captured onto 3D6coated ELISA plates and quantified by p24 ELISA. Error bars indicate SEM. One of three independent experiments is shown. (B) To assess gp120 shedding, JR-FL virus coexpressing mouse CD4 was pretreated with 10 µg/ml CD4-IgG2, 10 µg/ml b12, 100 µg/ml 2G12, 100 µg/ml 2F5, and 100 µg/ml 4E10 or left untreated for either 1 h or overnight (18-20 h). Virus was immobilized on anti-CD4-coated magnetic beads, and gp120 was quantified by ELISA. Percent gp120 shedding induced by the respective conditions in relation to mock-treated virus was determined. Data were normalized to p24 content (ELISA) to ensure that identical numbers of virions were assessed under each condition. Mean and SEM of three independent experiments are depicted. (C) Neutralization activity against free virus was determined as depicted in the flow chart shown in [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref]. JR-FL pseudovirus was pretreated with serial dilutions of CD4-IgG2, b12, 2G12, 2F5, and 4E10 for either 1 or 16 h, with a maximal mAb concentration identical to the shedding assay shown in B. Virus was then bead immobilized, washed, and used to infect TZM-bl cells. In comparison, the entire preincubation mix without washing was transferred onto TZM-bl cells. In the latter case, 1-h preincubation corresponds to the conditions in a conventional neutralization assay, and 16-h preincubation corresponds to the long-term incubation depicted in [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref]. Percent neutralization induced by the respective conditions in relation to mock-treated virus was determined. Mean and SEM of two independent experiments are depicted.
shedding from the probed R5, R5X4, and X4 isolates, demonstrating that shedding of primary viruses is possible and may thus occur in vivo.
## Is gp120 shedding a cause or a byproduct of antibody neutralization?
To probe whether gp120 shedding is mechanistically linked with the neutralization process or whether these two activities are independent, separable, or occur consecutively, we first defined the kinetics of the antibody-induced gp120 shedding process in detailed time of addition experiments [fig_ref] Figure 5: Kinetics of neutralizing antibody-induced gp120 shedding [/fig_ref]. The loss of gp120 triggered by the probed mAbs proved to be strikingly different over the 33 h of monitoring. results by alternative experimental approaches: studying gp120 shedding from virions using size-exclusion chromatography [fig_ref] Figure 4: MPER antibodies induce gp120 shedding from virions and HIV envelope-expressing cells [/fig_ref] and shedding from envelope-expressing cells by Western blot and flow cytometry [fig_ref] Figure 4: MPER antibodies induce gp120 shedding from virions and HIV envelope-expressing cells [/fig_ref] , B and C; and [fig_ref] Figure 2: Reversibility of HIV neutralization [/fig_ref]. The same pattern of reactivity as in the bead-immobilized virion shedding assay was observed in these experiments: MPER mAbs and CD4bs-directed agents induced envelope dissociation, whereas 2G12 did not. To obtain insights into a potential physiological relevance of the shedding process, we verified our initial results (obtained with pseudovirus expressing the primary virus envelope JR-FL) with replicationcompetent, PBMC-derived viruses [fig_ref] Figure 4: MPER antibodies induce gp120 shedding from virions and HIV envelope-expressing cells [/fig_ref]. As observed for pseudovirus, mAbs 2F5 and 4E10 induced potent gp120 To analyze gp120 shedding by sizeexclusion chromatography, JR-FL pseudovirus was treated with 25 µg/ml of CD4-IgG2, b12, 2G12, 2F5, 4E10, or total human IgG (huIgG) for 20-25 h. Samples were then separated on a Sephacryl S-1000 column, and collected fractions were analyzed for virion content and viral infectivity (relative light units [RLU] luciferase reporter production; relative light units are depicted in units of 3 × 10 3 ). One of three independent experiments is shown. (B) Shedding of gp120 from envelope-expressing cells was detected by Western blot analysis. 293-T cells expressing CT JR-FL gp160 were treated with 10 µg/ml CD4-IgG2, 50 µg/ml 2G12, 50 µg/ml 2F5, or 50 µg/ml human IgG as control for the indicated time periods to allow for shedding. After cell lysis, total protein was separated by SDS-PAGE and analyzed by Western blotting for gp120 content. gp120 content was quantified by densitometric analysis, and shedding induced by neutralizing mAbs was expressed in relation to human IgG control. One of three independent experiments is shown. (C) 293-T cells that were mock transfected or transfected with CT JR-FL gp160 were treated with 10 µg/ml CD4-IgG2, 50 µg/ml 2F5, or 50 µg/ml human IgG as control for 20 h to allow for shedding. Cell surface-associated gp120 was detected by flow cytometry upon staining with biotinylated 2G12 and streptavidin-APC. One of three independent experiments is shown. (D) Induction of gp120 shedding from PBMC-derived, replication-competent HIV isolates was analyzed using PBMC-derived viruses JR-FL, DH123, and NL4-3. Coreceptor usage of viruses is indicated. Viruses were incubated with 10 µg/ml CD4-IgG2, 100 µg/ml 2F5, and 100 µg/ml 4E10 or left untreated for 24 h (mock control). The extent of gp120 shedding in mAb-treated samples was analyzed as described in on virus infectivity upon short incubation [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref]. Neutralization activity therefore depended on interaction with the virus during the preincubation period and thus allowed direct assessment of the impact of shedding on the antibodies' neutralization activity. Strikingly, the timing of JR-FL neutralization and shedding was virtually identical for the two MPER antibodies, indicating that the process of shedding is causally linked to their mechanism of neutralization. Although we cannot formally rule out that MPER binding itself confers neutralization in the absence of shedding, the kinetics of the processes were so closely related that, if these events were indeed functionally separable, shedding must occur very rapidly after JR-FL neutralization by MPER antibodies. In contrast to MPER antibodies, no time-dependent association between shedding and neutralization of JR-FL pseudovirus was detected for b12. Although b12 induced shedding of 50% of JR-FL gp120 at concentrations >0.1 µg/ml [fig_ref] Figure 5: Kinetics of neutralizing antibody-induced gp120 shedding [/fig_ref] , lower concentrations (0.03 µg/ml; , which are perfectly capable of inducing neutralization, failed to induce potent gp120 shedding. The latter observation strongly suggests that the primary mechanism of b12 neutralization is independent of shedding. Likewise, as shown above [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref] , 2G12 neutralized without inducing gp120 shedding.
We next assessed the thermodynamic requirements of neutralization and shedding activity using mAb concentrations that yield 70% neutralization after 20-h CD4-IgG2 provoked instant and almost complete shedding. 2F5 and 4E10 reacted more rapidly than b12 and caused almost complete gp120 shedding upon long-term treatment, indicating that the antibodies must be able to release gp120 from both functional (trimeric) and nonfunctional envelope species (monomers and dimers) present on viral particles [bib_ref] Nature of nonfunctional envelope proteins on the surface of human immunodeficiency virus..., Moore [/bib_ref]. Intriguingly, even after long-term incubation, b12 only achieved shedding of 58% of JR-FL gp120. The dose dependency of the b12 shedding process followed a similar pattern. gp120 shedding continuously increased in a dose-dependent manner for 2F5 and 4E10, whereas shedding by b12 plateaued at 50% and could not be augmented by increasing doses [fig_ref] Figure 5: Kinetics of neutralizing antibody-induced gp120 shedding [/fig_ref]. Determining precisely which functional properties of b12 shape its shedding activity will require further investigation. For instance, partial gp120 depletion could result if mAb binding prompts shedding of neighboring units in a trimer (or dimer) but not of the unit the mAb is bound to. In general, differences in stoichiometric requirements, steric constraints, and mAb-induced conformational changes [bib_ref] HIV-1 evades antibody-mediated neutralization through conformational masking of receptor-binding sites, Kwong [/bib_ref] could be envisioned to cause the observed variable patterns of shedding activity among antibodies.
To explore whether the shedding process is an integral component of neutralization by individual antibodies, we monitored the kinetics of neutralization and shedding in parallel . Each mAb was probed at a fixed antibody concentration that typically yields 70% neutralization of JR-FL pseudovirus after 20-h preincubation of mAb and virus at 37°C but has (with the exception of 2G12) no notable effect (10 µg/ml), b12 (10 µg/ml), 2G12 (100 µg/ml), 2F5 (100 µg/ml), and 4E10 (100 µg/ml) or left untreated (medium control) for the indicated time periods at 37°C, and gp120 shedding was assessed as described in B. The rate of gp120 loss was fitted according to the formula L + (100 L) × exp(d × time). L denotes the maximal level of gp120 loss (lowest level of gp120) inflicted by the respective antibody. The rate of gp120 loss is characterized by a constant d (units per hour). T1/2 max denotes the time until half maximal gp120 loss was reached. An overlay of five independent experiments is depicted. (B) The dose dependency of the shedding process was assessed by treating JR-FL pseudovirus with increasing concentrations of b12, 2F5, and 4E10 or by leaving the pseudovirus untreated (medium control) for 22 h at 37°C. Shedding was analyzed as described in B. Data are expressed in relation to the medium control. Pooled data from three independent experiments are depicted. Curves depict sigmoid dose-dependent fits with variable slope and constraining upper limits at 100%. The dotted curves indicate the respective 95% confidence intervals.
## Br ief definitive repor t
Neutralizing antibodies induce gp120 shedding of genetically diverse viruses Held together by labile, noncovalent intersubunit interactions between gp41 and gp120 [bib_ref] Topological layers in the HIV-1 gp120 inner domain regulate gp41 interaction and..., Finzi [/bib_ref] [bib_ref] Structure of HIV-1 gp120 with gp41-interactive region reveals layered envelope architecture and..., Pancera [/bib_ref] , the HIV envelope trimer is characterized by a high conformational diversity particularly within gp120, which is central to the envelope's function in virus entry and immune evasion [bib_ref] HIV-1 evades antibody-mediated neutralization through conformational masking of receptor-binding sites, Kwong [/bib_ref] [bib_ref] Structural basis of immune evasion at the site of CD4 attachment on..., Chen [/bib_ref]. Notably, neutralization-sensitive HIV strains have been postulated to present envelope conformations, which, although optimized for receptor interaction, allow easier access for neutralizing antibodies and potentially destabilize the trimer [bib_ref] A global neutralization resistance phenotype of human immunodeficiency virus type 1 is..., Park [/bib_ref] [bib_ref] The prolonged culture of human immunodeficiency virus type 1 in primary lymphocytes..., Pugach [/bib_ref]. These structural characteristics are also likely to steer the sensitivity of a given virus envelope to antibody-induced shedding. To address this, we first screened a panel of gp120-and gp41-specific antibodies for their shedding activity against JR-FL pseudovirus A and Tables S2-S4). With the exception of 2G12, all mAbs that neutralized the virus also induced potent gp120 shedding. Particularly notable were the CD4bs-specific antibody 1F7 [bib_ref] Molecular characterization of five neutralizing anti-HIV type 1 antibodies: identification of nonconventional..., Kunert [/bib_ref] and the V3 loop antibody 447-52D [bib_ref] Neutralization of diverse human immunodeficiency virus type 1 variants by an anti-V3..., Gorny [/bib_ref] , which achieved 96.3% and 87.6% shedding, respectively. Interestingly, the CD4bs antibody b6, which is known to bind predominantly nonfunctional envelope forms [bib_ref] Heterogeneity of envelope molecules expressed on primary human immunodeficiency virus type 1..., Poignard [/bib_ref] and lacks neutralization activity against JR-FL, induced partial shedding (41.7%), demonstrating that both neutralizing and nonneutralizing antibodies can participate in this process, presumably by acting against envelope monomers or dimers on the virion surface. preincubation at 37°C. CD4-IgG2 potently neutralized virions and induced gp120 shedding even at low temperatures. 2G12 displayed modest neutralization activity at 19.4°C, which increased with temperature. Confirming our previous observations [fig_ref] Figure 4: MPER antibodies induce gp120 shedding from virions and HIV envelope-expressing cells [/fig_ref] and Figs. 5 A and 6 A), no notable shedding of JR-FL gp120 was induced by 2G12 at any of the probed temperatures. Antibodies b12, 2F5, and 4E10 all required temperatures >28°C to neutralize. Strikingly, MPER mAb neutralization and gp120 shedding activity followed an identical temperature profile. Together with the observed close time-dependent synchronization of their activities, these data provide strong evidence that with the majority of isolates, MPER neutralization and shedding are indeed mechanistically linked. The thermodynamic processes observed for b12 were more complex: although the chosen dose (0.03 µg/ml) induced potent neutralization at 37°C, which decreased in a temperaturedependent manner, the same conditions induced only comparatively little shedding, particularly at lower temperatures. Although the pattern was similar (more shedding with higher temperature), neutralization by b12 was not obligatorily linked to shedding. Likewise, the high potency of b12 to irreversibly neutralize HIV appears not to depend entirely on the mAb's shedding capacity, as complete irreversibility is achieved at antibody doses which only induce partial shedding [fig_ref] Figure 2: Reversibility of HIV neutralization [/fig_ref]. The latter suggests that the degree of conformational fixation of gp120 inflicted upon b12 binding [bib_ref] Structural definition of a conserved neutralization epitope on HIV-1 gp120, Zhou [/bib_ref] alone mediates irreversible inactivation.
## Figure 6. shedding and neutralization activity of mper mabs follow identical kinetics and thermodynamic requirements. (a)
To compare the kinetics of shedding and neutralization, neutralization on TZM-bl cells and shedding activity (as described in of JR-FL pseudovirus upon treatment with mAbs b12 (0.03 µg/ml), 2G12 (12 µg/ml), 2F5 (2.5 µg/ml), and 4E10 (8.5 µg/ml) was assessed at the indicated times at 37°C. Chosen concentrations correspond to 70% neutralization after 20-h incubation at 37°C. One of two independent experiments is depicted. (B) To compare thermodynamic requirements of shedding and neutralization, neutralization on TZM-bl cells and shedding activity (as described in of JR-FL pseudovirus upon treatment with mAbs b12 (0.03 µg/ml), 2G12 (12 µg/ml), 2F5 (2.5 µg/ml), and 4E10 (8.5 µg/ml) for 20 h at the indicated temperature were assessed. One of three independent experiments is depicted. Chosen concentrations correspond to 70% neutralization after 20-h incubation at 37°C.
shedding in 11 of the 14 probed viruses. A comparatively low level of shedding (<30%) occurred with three viruses, WITO, TRO, and PVO.4, of which the latter two were not neutralized by 2F5. The pattern for 4E10 was similar: neutralization was detected against all 14 viruses, against 13 of which shedding activity was high. The only comparatively sheddinginsensitive virus (<30% shedding by 4E10), WITO, was nevertheless neutralized by 4E10. It is to be expected that both shedding efficacy and dependency of the MPER neutralization process on shedding induction vary among divergent virus strains, as the induction of gp120 shedding by sCD4 also varies among virus strains [bib_ref] Virions of primary human immunodeficiency virus type 1 isolates resistant to soluble..., Moore [/bib_ref] [bib_ref] Mechanisms of human immunodeficiency virus Type 1 (HIV-1) neutralization: irreversible inactivation of..., Mcdougal [/bib_ref] [bib_ref] Envelope glycoprotein incorporation, not shedding of surface envelope glycoprotein (gp120/SU), is the..., Chertova [/bib_ref]. Besides epitope accessibility and antibody affinity, the conformational flexibility of gp120 and the stability of the noncovalent gp120/gp41 association of a specific virus strain will influence shedding efficacy [bib_ref] Virions of primary human immunodeficiency virus type 1 isolates resistant to soluble..., Moore [/bib_ref] [bib_ref] Topological layers in the HIV-1 gp120 inner domain regulate gp41 interaction and..., Finzi [/bib_ref] [bib_ref] Structure of HIV-1 gp120 with gp41-interactive region reveals layered envelope architecture and..., Pancera [/bib_ref]. Thus, envelope trimers that can undergo the required To explore how ubiquitously neutralizing antibodies cause shedding across divergent strains, we assessed the shedding and neutralizing capacity of seven mAbs against a panel of 14 pseudoviruses encompassing relatively easy to neutralize tier 1 and more resistant tier 2 and 3 isolates and Tables S2-S4). The results confirmed our previous observations with the virus strain JR-FL . mAbs directed to the CD4bs, V3 loop, and the MPER possessed potent shedding activity against the majority of isolates. 2G12, which was ineffective in inducing shedding from JR-FL envelope-bearing virions, induced 30-60% shedding with 5 of the 14 probed viruses. Although clearly not a potent shedding inducer, 2G12 can thus not be considered incapable of inducing shedding. mAb b12 promoted shedding of >30% in 8 of 14 viruses but had only marginal effects against the other 6 isolates, 3 of which were not neutralized by b12. The CD4i (CD4 induced)-specific mAb 17b was largely ineffective in both, inducing neutralization and shedding. 2F5 induced potent . Antibodies with gp120 shedding activity are commonly elicited in HIV infection. (A) Assessment of gp120 shedding activity induced by HIV envelope-specific mAbs. 20 µg/ml of mAbs directed to diverse epitopes in gp120 and gp41 were assessed for shedding activity (as described in after 20-h treatment of JR-FL pseudovirus at 37°C. Red bars denote mAbs that neutralize JR-FL, and gray bars denote nonneutralizing mAbs. Mean and SEM of three independent experiments are depicted. (B) Shedding activity elicited in HIV infection. 84 plasma samples derived from patients with chronic infection with the indicated HIV subtypes were probed for shedding (as described in , and neutralization activity was measured using TZM-bl cells against JR-FL (left) and NL4-3 pseudovirus (right) under identical preincubation conditions (20-h preincubation at 37°C). Correlation coefficient r (Pearson) and p-values are depicted. Data are means of two independent experiments. Dotted lines correspond to spontaneous gp120 shedding induced by healthy donor plasma (mean of three donors). (C) Shedding induced by a wide spectrum of mAbs with divergent specificities. Shedding (as described in and neutralizing activity measured using TZM-bl cells of the indicated mAbs against 14 divergent viruses from tier 1-3 were determined under identical preincubation conditions (10 µg/ml mAb and 20-h preincubation at 37°C). Correlation coefficient r (Pearson) and p-values are depicted. Significance was assessed after correcting for multiple testing (Bonferroni test). Data are means of two to four independent experiments. Error bars indicate SEM. n.s., not significant. account for its high in vivo efficacy, particularly in established infection [bib_ref] Broadly neutralizing human anti-HIV antibody 2G12 is effective in protection against mucosal..., Hessell [/bib_ref].
In HIV vaccine development, a considerable effort has been directed toward eliciting neutralizing antibody responses that match the activities of the MPER-specific antibodies 2F5 and 4E10, yet the molecular basis of their neutralization remained unresolved. Recently, protection by MPER-directed antibodies was suggested not to involve antibody-dependent cellular cytotoxicity, emphasizing the fact that neutralization rather than effector functions dominate the in vivo activity of these mAbs [bib_ref] Broadly neutralizing monoclonal antibodies 2F5 and 4E10 directed against the human immunodeficiency..., Hessell [/bib_ref]. Our current findings underline the importance of the neutralization process to the MPER mAb antiviral activity. Here, we provide compelling evidence that neutralization by MPER mAbs involves induction of gp120 shedding, rendering virus inhibition irreversible. In our study, we explored the interaction of the antibodies with HIV before receptor engagement to obtain information explicitly on the virion-directed activity of the mAbs. Notably, as we show here, virion neutralization by MPER mAbs is a comparatively slow process that requires several hours. In the light of our findings, it will be interesting to verify the postulated action of MPER mAbs after receptor engagement and investigate the contribution of shedding in this process [bib_ref] Redox-triggered infection by disulfide-shackled human immunodeficiency virus type 1 pseudovirions, Binley [/bib_ref]. MPER antibodies have been shown to preferentially bind to the prehairpin intermediate that forms upon receptor engagement [bib_ref] Structural basis of enhanced binding of extended and helically constrained peptide epitopes..., Cardoso [/bib_ref] [bib_ref] A fusion-intermediate state of HIV-1 gp41 targeted by broadly neutralizing antibodies, Frey [/bib_ref] [bib_ref] Distinct conformational states of HIV-1 gp41 are recognized by neutralizing and non-neutralizing..., Frey [/bib_ref] [bib_ref] Role of HIV membrane in neutralization by two broadly neutralizing antibodies, Alam [/bib_ref]. Whether these confirmations are also adopted spontaneously in the absence of receptor engagement or upon initial, low affinity interaction of MPER mAbs with the viral envelope remains to be determined. Be it upon receptor engagement or mAb interaction, gp120 shedding can be envisioned to be closely associated with the formation of prehairpin intermediate.
Notably, our data only allow us to affirm a causal link between JR-FL virus neutralization and gp120 shedding by MPER antibodies within the temporal resolution of our assay systems. From this we can conclude that these actions are highly synchronized and follow identical kinetic, thermodynamic, and stoichiometric requirements. Yet, this does not exclude that MPER mAb binding alone would suffice to neutralize the virus by blocking the refolding process of gp41 and that shedding occurs as an immediate consequence of mAb binding. Whether a primary event leading to neutralization or a consequence of neutralization, our data clearly indicate that induction of gp120 shedding is closely associated with MPER antibody inhibition of the majority of HIV isolates and needs to be factored into the mechanistic processes underlying HIV neutralization by MPER mAbs.
How can antibody binding principally lead to gp120 dissociation from the envelope trimer? The HIV trimer association is relatively labile as it must support substantial conformational changes and dissociation of the gp120-gp41 heterodimers during interaction with cellular receptors and fusion. In analogy to the entry process, conformational changes resulting from antibody binding or conformational fixation by neutralizing antibodies can be envisioned to similarly disrupt conformational changes and tolerate neutralizing mAb binding without jeopardizing subunit association may be less prone to shed gp120 upon interaction with neutralizing antibodies. However, overall, we found that shedding activity is prominently associated with neutralization activity. Of the probed neutralizing antibodies, MPER mAbs yielded the highest and broadest shedding activity overall. Particularly notable again were mAbs 447-52D and 1F7, which both induced potent shedding that correlated with their neutralization activity.
To obtain information on the frequency at which shedding-inducing antibodies are elicited in HIV infection, we analyzed shedding and long-term neutralization activity in 84 plasma samples derived from patients with chronic infection of subtypes A, B, C, CRF01_AE, and CFR02_AG and . Cross-neutralizing activity against JR-FL was low in most plasma samples (reciprocal neutralization titer <100 in 38 of 84 samples) but was frequently associated with induction of shedding against the same virus. The low plasma neutralization titers against JR-FL did not allow firm conclusions on a mechanistic link between JR-FL directed neutralization and shedding activity to be drawn. In contrast, NL4-3 was neutralized by a high proportion of isolates at relatively high titers (reciprocal neutralization titer >1,000 in 60 of 84 samples). Shedding was induced by plasma of chronic patients across all subtypes and, in accordance with neutralization activity, was higher against NL4-3 than JR-FL. Notably, the neutralization and shedding activity directed toward NL4-3 were tightly correlated, further underlining a potential mechanistic association of these activities.
# Conclusions
Although the ability of neutralizing antibodies to protect against HIV infection has been demonstrated in vivo [bib_ref] Protection of macaques against vaginal transmission of a pathogenic HIV-1/SIV chimeric virus..., Mascola [/bib_ref] [bib_ref] Delay of HIV-1 rebound after cessation of antiretroviral therapy through passive transfer..., Trkola [/bib_ref] [bib_ref] Effective, low-titer antibody protection against low-dose repeated mucosal SHIV challenge in macaques, Hessell [/bib_ref] , whether one or more actions, direct virus neutralization, induction of phagocytosis, inhibition of transfer (e.g., via DCs) to target cells, or killing of infected cells via antibody-dependent cellular cytotoxicity, are key to block transmission is currently not known [bib_ref] Humoral immunity to HIV-1: neutralization and beyond, Huber [/bib_ref]. Both gp120 shedding and antibody-induced conformational changes can lead to irreversible inactivation of HIV, which may positively influence in vivo efficacy of antibody neutralization [bib_ref] Mechanisms of human immunodeficiency virus Type 1 (HIV-1) neutralization: irreversible inactivation of..., Mcdougal [/bib_ref] [bib_ref] Neutralizing antibodies to human immunodeficiency virus type-1 gp120 induce envelope glycoprotein subunit..., Poignard [/bib_ref]. The impact of antibodies mediating reversible neutralization may depend on rapid clearance of neutralized virions by phagocytes as influences of the milieu (e.g., migration to anatomical sites with lower antibody concentration; engulfment by DCs) may lead to antibody dissociation and reconstitute the virus's infectivity. Neutralizing antibodies that irrevocably sterilize virions before target cells are encountered would in these scenarios be of clear benefit. However, failure to irreversibly neutralize does not necessarily render an antibody ineffective in vivo. Although reversible, 2G12 neutralization occurs, as we show in this study, more rapidly than neutralization with b12, 2F5, and 4E10, and its immediate action may indeed study, a variety of neutralizing antibodies induce shedding and thus irreversibly neutralize HIV. Thus, elicitation of antibodies that irreversibly neutralize should be within reach, and it will certainly be of interest to assess the recently identified potently neutralizing antibodies PG9, PG16, HJ16, and VRC01 for their capacity to irreversibly neutralize HIV [bib_ref] Analysis of memory B cell responses and isolation of novel monoclonal antibodies..., Corti [/bib_ref] [bib_ref] Human anti-HIV-neutralizing antibodies frequently target a conserved epitope essential for viral fitness, Pietzsch [/bib_ref] [bib_ref] Rational design of envelope identifies broadly neutralizing human monoclonal antibodies to HIV-1, Wu [/bib_ref]. A combination of antibodies that act rapidly and neutralize irreversibly may prove to be advantageous. Intriguingly, 2G12, which lacks the capacity to irreversibly neutralize but instead rapidly binds to and neutralizes HIV, effectively blocks the virus in vivo, as passive immunization trials with combinations of 2G12, 2F5, and 4E10 confirmed [bib_ref] Adjunctive passive immunotherapy in human immunodeficiency virus type 1-infected individuals treated with..., Mehandru [/bib_ref]. Although only 2G12 was assumed to be effective in these trials based on escape virus formation, a matter for speculation remains whether in vivo efficacy may have in part resulted from a rapid but reversible inhibition by 2G12 followed by sustained, irreversible inhibition by the MPER antibodies.
Although commonly low titered and often subneutralizing in vitro, neutralizing antibody concentrations in mucosal fluids nevertheless protected against SHIV transmission in challenge studies [bib_ref] Protection of macaques against vaginal transmission of a pathogenic HIV-1/SIV chimeric virus..., Mascola [/bib_ref] [bib_ref] Broadly neutralizing human anti-HIV antibody 2G12 is effective in protection against mucosal..., Hessell [/bib_ref]. Our data provide a possible mechanistic basis for these observations as we demonstrate that upon long-term contact, efficient (and in most cases irreversible) inactivation of HIV occurs at comparatively low antibody concentrations. Although the precise delineation of the molecular processes that induce gp120 shedding upon MPER mAb binding will require further investigation, our current observations reveal important aspects of these antibodies' mode of action and lay a foundation for the design of vaccines and compounds that induce gp120 shedding and irreversible inactivation in a similar manner. . Tetrameric CD4-Ig (CD4-IgG2; [bib_ref] Expression and characterization of CD4-IgG2, a novel heterotetramer that neutralizes primary HIV..., Allaway [/bib_ref] and sCD4 [bib_ref] A soluble form of CD4 (T4) protein inhibits AIDS virus infection, Deen [/bib_ref] were provided by W. Olson (Progenics Pharmaceuticals, Tarrytown, NY); T-20 [bib_ref] Potent suppression of HIV-1 replication in humans by T-20, a peptide inhibitor..., Kilby [/bib_ref] was provided by Roche. Human IgG was purchased from Sigma-Aldrich. Plasma samples were obtained from individuals with chronic HIV-1 infection (>6 mo infected) enrolled in (a) the , (b) the Zurich Primary HIV infection study [bib_ref] Complement lysis activity in autologous plasma is associated with lower viral loads..., Huber [/bib_ref] , and (c) the Swiss Spanish Treatment Interruption trial. All experiments were approved by the local ethics committee of the University Hospital Zurich, and written informed consent was obtained from all individuals. Detailed patient demographics are listed in .
# Materials and methods
## Reagents
Cells. 293-T and TZM-bl cells were obtained from the American Type Culture Collection and the National Institutes of Health AIDS repository, respectively, and cultured as described previously [bib_ref] Divergent effects of cell environment on HIV entry inhibitor activity, Rusert [/bib_ref].
Virus preparation. Envelope-pseudotyped virus was prepared as previously described [bib_ref] Divergent effects of cell environment on HIV entry inhibitor activity, Rusert [/bib_ref]. For generation of envelope-pseudotyped virus the fragile inter-and intrasubunit associations. It can be speculated that the trimer association may depend on a certain degree of flexibility of all three subunits to allow synchronized movements. Once one unit's conformation is arrested, for example by mAb binding to the MPER domain and the viral membrane, trimer association may be jeopardized. Alternatively, envelope conformations that expose the MPER and enable 2F5 and 4E10 binding may induce instability of the trimer. The strength of the subunit interaction as well as the degree by which conformational changes are inflicted will likely determine whether and to what extent shedding is associated with the neutralization process of a given virus antibody combination. Related observations have been made in other virus infections in which envelope or capsid conformation changes need to occur before or during neutralization [bib_ref] Structural dynamics, an intrinsic property of viral capsids, Witz [/bib_ref] [bib_ref] Further evidence that papillomavirus capsids exist in two distinct conformations, Selinka [/bib_ref] [bib_ref] Binding of a neutralizing antibody to dengue virus alters the arrangement of..., Lok [/bib_ref].
A principle question remains: is induction of shedding by neutralizing antibodies of physiological relevance? Clearly it is no prerequisite of neutralization as 2G12 and b12, which both can neutralize without inducing shedding, have proven in vivo efficacy [bib_ref] Fc receptor but not complement binding is important in antibody protection against..., Hessell [/bib_ref] [bib_ref] Broadly neutralizing human anti-HIV antibody 2G12 is effective in protection against mucosal..., Hessell [/bib_ref] [bib_ref] Adjunctive passive immunotherapy in human immunodeficiency virus type 1-infected individuals treated with..., Mehandru [/bib_ref] Likewise, the in vitro and in vivo action of sCD4 has been suggested to be dominated by receptor binding site occupancy rather than by induction of gp120 shedding [bib_ref] Dissociation of gp120 from HIV-1 virions induced by soluble CD4, Moore [/bib_ref] [bib_ref] Virions of primary human immunodeficiency virus type 1 isolates resistant to soluble..., Moore [/bib_ref] [bib_ref] Differential loss of envelope glycoprotein gp120 from virions of human immunodeficiency virus..., Mckeating [/bib_ref] [bib_ref] Two mechanisms of soluble CD4 (sCD4)-mediated inhibition of human immunodeficiency virus type..., Orloff [/bib_ref] [bib_ref] Mechanisms of human immunodeficiency virus Type 1 (HIV-1) neutralization: irreversible inactivation of..., Mcdougal [/bib_ref] [bib_ref] Soluble CD4 and CD4-mimetic compounds inhibit HIV-1 infection by induction of a..., Haim [/bib_ref]. Yet, we would argue that the capacity of neutralizing antibodies to induce shedding or otherwise irreversibly neutralize HIV could be expected to substantially impact their in vivo activity. Virus transmission across mucosal surfaces is an inefficient process, and only few founder viruses establish the new infection [bib_ref] Genetic identity, biological phenotype, and evolutionary pathways of transmitted/founder viruses in acute..., Salazar-Gonzalez [/bib_ref]. Mucus and epithelia trap virus, physically restricting transmission, and the limited number of transmitted virions depend on locating appropriate target cells [bib_ref] Targeting early infection to prevent HIV-1 mucosal transmission, Haase [/bib_ref]. These processes require time during which virions remain vulnerable to antibody attack. During established infection, the bulk of HIV particles in the periphery are rapidly turned over with an estimated half-life of plasma virus of 28-110 min. If this turnover is to be influenced by neutralizing antibodies, it requires a more rapid action than MPER antibodies appear capable of. Yet, not all virions are in the periphery and cleared. HIV is known to be trapped in various tissue compartments (e.g., by immature and mature DCs, follicular DCs, and B cells) where virions can be retained in an infectious state for prolonged time periods [bib_ref] B cells of HIV-1-infected patients bind virions through CD21-complement interactions and transmit..., Moir [/bib_ref] [bib_ref] Immunodeficiency virus uptake, turnover, and 2-phase transfer in human dendritic cells, Turville [/bib_ref] [bib_ref] Complement dependent trapping of infectious HIV in human lymphoid tissues, Bánki [/bib_ref] [bib_ref] Role for CD21 in the establishment of an extracellular HIV reservoir in..., Ho [/bib_ref] [bib_ref] HIV traffics through a specialized, surface-accessible intracellular compartment during transinfection of T..., Yu [/bib_ref]. Although encompassing only a small fraction of the circulating virions, these trapped HIV particles are thought to contribute substantially toward dissemination and the formation of a viral reservoir in the infected individual. In this setting, antibodies that irreversibly neutralize the virus could indeed be of clear benefit, and antibody activities that require prolonged interaction with the virus may come in play.
Should vaccine-elicited antibody responses be specifically tailored to induce irreversible neutralization? As we show in this Br ief Definitive Repor t Detection of gp120 shedding from bead immobilized virus by ELISA. (a) Pseudovirus expressing mouse CD4 on the viral surface was incubated at the indicated time/temperature conditions in the presence or absence of neutralizing antibodies. Virus was then immobilized onto magnetic beads coated with rat anti-mouse CD4 antibodies (Invitrogen) and washed with TBS/2% BSA using a magnetic 96-well plate (OZ Biosciences) to separate virus from unbound antibodies and shed gp120. Finally, the magnetic pellets containing washed virus were lysed in TBS containing 1% Empigen and gp120, and the p24 content of the lysate was quantified by ELISA. The percentage of gp120 shedding induced by the respective conditions is dependent on the experiment, expressed either in relation to the original virus stock (decay experiments; Figs. 5 A and 6 A) or in relation to mock-treated virus controls (all other results). Mock-treated virus controls correspond to 0% shedding and 100% gp120 content. (b) Analysis of gp120 from PBMCderived, replication-competent virus was essentially identical. The only difference in this set up was that virus was immobilized on magnetic anti-CD44 beads (Miltenyi Biotec) for 1 h at 4°C and washed on µMACS columns. Purified virus was eluted in TBS/1% Empigen and gp120, and the p24 content of the lysate was quantified by ELISA.
Detection of gp120 shedding by gel filtration. Size-exclusion chromatography was performed as described previously [bib_ref] Dissociation of gp120 from HIV-1 virions induced by soluble CD4, Moore [/bib_ref]. In brief, JR-FL-envelope-pseudotyped virions were concentrated by ultracentrifugation over a 32% sucrose cushion in a centrifuge (Centrikon Ultra, Rotor TST 28.38; Kontron Instruments) at 28,000 rpm for 2 h. Aliquots of the concentrated virus (100 ng p24) were treated with 25 µg/ml CD4-IgG2, 25 µg/ml b12, 25 µg/ml 2G12, 25 µg/ml 2F5, and 25 µg/ml 4E10 or 25 µg/ml human IgG for 20-25 h. Samples were then subjected to size-exclusion chromatography using a (10 ml, 16/50) Sephacryl S-1000 column (GE Healthcare) on a purifier system (AEKTA; GE Healthcare). 200-µl fractions were collected and analyzed for p24 and gp120 content by ELISA, and infectivity was probed on TZM-bl reporter cells.
Detection of gp120 shedding by Western blot analysis. 293T cells were transfected in 24-well plates with CT JR-FL gp160, a cytosolic tail truncated gp160 construct known to yield higher envelope trimer expression [bib_ref] Redox-triggered infection by disulfide-shackled human immunodeficiency virus type 1 pseudovirions, Binley [/bib_ref] , and pCMV-rev [bib_ref] Identification of a cis-acting element in human immunodeficiency virus type 2 (HIV-2)..., Lewis [/bib_ref] using jetPEI (Polyplus Transfection) according to the manufacturer's instruction. As mock control, empty pcDNA3.1 vector was cotransfected together with pCMV-rev under identical conditions. 20 or 1 h before harvest, the cells were treated with 50 µg/ml total human IgG, 10 µg/ml CD4-IgG2, 50 µg/ml 2G12, or 50 µg/ml 2F5. Cells were then washed to remove shed gp120 in PBS/10 mM EDTA and lysed with 100 µl of 50 mM Tris, 150 mM NaCl, SDS 0.1% (wt/ vol), 0.5% (wt/vol) Na deoxycholate, and 1% Triton X-100, pH 7.4 (RIPA buffer). The cell lysates were analyzed on a NuPAGE 4-12% gradient gel (Invitrogen) and subsequently blotted onto a Hybond C membrane (GE Healthcare) and analyzed for gp120 content using goat anti-gp120 antibody (D7324; Aalto) and horseradish peroxidase-conjugated rabbit anti-goat IgG and detection with ECL (GE Healthcare). Imaging analysis was performed using the LAS4000 imaging system (Fujifilm), and densitometry was performed using Multi Gauge version 3.0 software (Fujifilm). All data were normalized to uncleaved (intracellular) gp160 of the respective sample to correct for potential differences in transfection efficiency.
Detection of gp120 shedding by FACS. 293-T cells expressing CT JR-FL gp160 and controls were prepared as described for the analysis of gp120 shedding by Western blot. 20 h before harvest, the culture medium was exchanged for fresh medium containing protein transport inhibitor Brefeldin A (BioLegend) at 1:1,000 to limit incorporation of new trimer molecules in the cell membrane during the following shedding analysis. Cells were then incubated for 1 or 20 h with total human 50 µg/ml IgG, 10 µg/ml CD4-IgG2, or 50 µg/ml 2F5, after which cells were detached from plates with PBS/10 mM EDTA (Invitrogen) and washed twice with PBS, 1% FCS, and 2 mM EDTA. The level of residual surface gp120 was determined by flow cytometry upon staining with biotinylated 2G12 and detection by streptavidin-APC (BioLegend).
with mouse CD4 on the viral surface, 30 µg of mouse CD4 plasmid (mCD4-pCMV-SPORT6, IRAVp968B1245D; RZPD Deutsches Ressourcenzentrum für Genomforschung GmbH) was cotransfected with 7.5 µg of envelope expression plasmid, 22.5 µg of the envelope-deficient HIV-1 backbone vector pNL-luc-AM, and 120 µg polyethylenimine. Stocks of replication-competent virus were prepared as described previously .
Neutralization assay using envelope-pseudotyped virus. The neutralization activity of mAbs and patient plasma was evaluated on TZM-bl cells essentially as described previously [bib_ref] Divergent effects of cell environment on HIV entry inhibitor activity, Rusert [/bib_ref]. Virus input was chosen to yield comparable reporter gene production (10,000-20,000 relative light units) in untreated control samples across divergent viruses. The antibody concentrations causing 50% reduction in viral infectivity (IC 50 ) were calculated in Prism (GraphPad Software, Inc.) by sigmoid dose-response curve fit through the pooled data of two to five independent assays. If the appropriate degree of inhibition was not achieved at the highest or lowest drug concentration, a greater than or less than value was recorded. gp120 and p24 ELISA. gp120 and p24 antigen were quantified by ELISA as described previously [bib_ref] Divergent effects of cell environment on HIV entry inhibitor activity, Rusert [/bib_ref].
Probing reversibility of neutralization. See [fig_ref] Figure 2: Reversibility of HIV neutralization [/fig_ref] for flow chart. Antimouse CD4 magnetic beads were incubated with envelope-pseudotyped virus expressing mouse CD4 on the viral surface for 30 min at 4°C. For treatment A, the bead-virus complexes were bound to µMACS columns (Miltenyi Biotec), washed with DME/10% FCS, and subsequently eluted according to the manufacturer's protocol. Virus-bead complexes were then treated for 11 h at 37°C with neutralizing antibodies and unbound mAbs and afterward removed via column separation as described for the previous step. Antibodyopsonized virus was then incubated in medium for 11 h at 37°C to allow for antibody dissociation. After a final wash step to remove all unbound antibodies and shed gp120, viral infectivity was assessed on TZM-bl cells. To normalize for viral input, p24 content in the virus preparations was quantified by ELISA. Controls to account for neutralization without dissociation were conducted for each antibody (treatment B). For these, virus was treated identically as described for treatment A, with the exception that antibody treatment was performed in the second 11-h incubation step. Time periods of virus incubation and antibody exposure were thus identical in treatment A and B, only the order differed. The maximal inhibitory activity without removal of antibody was probed in treatment C, in which incubation was performed as described for treatment B but without the final removal of the unbound antibody. Percent inhibition was calculated in reference to a mocktreated control (100% virus infectivity).
Assessing neutralization activity against free virus. See [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref] for flow chart. JR-FL-pseudotyped virus was incubated with neutralizing mAbs for the indicated time period. To separate virions and unbound mAbs, anti-CD44 (BioLegend)-coupled magnetic beads (Dynabeads MyOne streptavidin; Invitrogen) were added to the preincubation mix and allowed to react with virions for 30 min at 4°C. Virus-bead complexes were then separated, washed, and transferred to TZM-bl cells to assess viral infectivity. Percent inhibition was calculated in reference to a mock-treated control (100% virus infectivity).
Virion binding assay. ELISA plates (Costar) were coated with 2 µg/ml goat anti-human IgG (SouthernBiotech) in NaHCO 3 , pH 8.5, and blocked with TBS containing 2% BSA. mAb 3D6 was diluted to 10 µg/ml in TBS/2% BSA and captured on coated plates. Subsequently, antibody-covered plates were blocked with human IgG diluted to 10 µg/ml in TBS/2% BSA. JR-FL-pseudotyped virus was pretreated for 18 h at 37°C with 10 µg/ml sCD4, 100 µg/ml 2F5, 100 µg/ml 4E10, or 10 µg/ml of the fusion inhibitor T-20 or left untreated (medium control). Virus was then added to the antibodycovered plates and incubated for 12 h at 37°C. Plates were washed with TBS/2% BSA to remove unbound virus, and bound virus was lysed in TBS containing 1% Empigen. Finally, p24 content was determined by ELISA.
Viability stain propidium iodide (Invitrogen) was added to all samples 10 min before acquisition, and dead cells were excluded from analysis. Samples were acquired on a CyAn ADP analyzer (Beckman Counter) and analyzed using FlowJo software (Tree Star, Inc.). Data analysis. Statistical analysis and fitting were performed using Prism version 5.0 with the exception of data depicted in [fig_ref] Figure 5: Kinetics of neutralizing antibody-induced gp120 shedding [/fig_ref] where fitting was performed using the R language of statistical computing [bib_ref] R: A language and environment for statistical computing, Development Core [/bib_ref].
Online supplemental material. [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref] shows virus-antibody interaction kinetics. [fig_ref] Figure 2: Reversibility of HIV neutralization [/fig_ref] depicts gp120 staining on transfected cells after 1-h treatment with CD4-IgG2 and 2F5, demonstrating that short-term treatment with 2F5 does not release gp120. provides IC 50 concentrations for all virus mAb combinations probed in [fig_ref] Figure 1: Kinetics and reversibility of HIV neutralization [/fig_ref] A. lists characteristics of pseudovirions used. lists the characteristics and origin of antibodies used. , included as an Excel file, provides shedding and neutralization data depicted in C. , included as an Excel file, provides shedding and neutralization data depicted in B. Online supplemental material is available at http://www.jem.org/cgi/content/full/jem.20101907/DC1.
[fig] Figure 1: Kinetics and reversibility of HIV neutralization. (A) Time dependence of neutralization [/fig]
[fig] Figure 2: Reversibility of HIV neutralization. (A and B) Reversibility of neutralization was probed as depicted in the flow chart. [/fig]
[fig] Figure 4: MPER antibodies induce gp120 shedding from virions and HIV envelope-expressing cells. (A) [/fig]
[fig] Figure 5: Kinetics of neutralizing antibody-induced gp120 shedding. (A) Shedding kinetics were determined over a 33-h time period. JR-FL pseudovirus was treated with fixed concentrations of CD4-IgG2 [/fig]
|
Giant magnetoresistance, three-dimensional Fermi surface and origin of resistivity plateau in YSb semimetal
Very strong magnetoresistance and a resistivity plateau impeding low temperature divergence due to insulating bulk are hallmarks of topological insulators and are also present in topological semimetals where the plateau is induced by magnetic field, when time-reversal symmetry (protecting surface states in topological insulators) is broken. Similar features were observed in a simple rock-salt-structure LaSb, leading to a suggestion of the possible non-trivial topology of 2D states in this compound. We show that its sister compound YSb is also characterized by giant magnetoresistance exceeding one thousand percent and low-temperature plateau of resistivity. We thus performed in-depth analysis of YSb Fermi surface by band calculations, magnetoresistance, and Shubnikov-de Haas effect measurements, which reveals only three-dimensional Fermi sheets. Kohler scaling applied to magnetoresistance data accounts very well for its low-temperature upturn behavior. The field-angle-dependent magnetoresistance demonstrates a 3D-scaling yielding effective mass anisotropy perfectly agreeing with electronic structure and quantum oscillations analysis, thus providing further support for 3D-Fermi surface scenario of magnetotransport, without necessity of invoking topologically non-trivial 2D states. We discuss data implying that analogous field-induced properties of LaSb can also be well understood in the framework of 3D multiband model. Yttrium monoantimonide has mainly been used as a non-magnetic reference or as a 'solvent' in monoantimonides of f-electron-elements solid solutions with anomalous physical properties such as dense Kondo behavior and complex magnetic ground-states 1 . Within that context it has been characterized as a metal by low-temperature specific heat measurements 2 . Later Hayashi et al. have shown that the first-order phase transition from the NaCl-type to a CsCl-type crystal structure occurs in YSb at 26 GPa 3 . That discovery induced numerous calculations of electronic structure of the compound, among them those by Tütüncü, Bagci and Srivastava, who directly compared electronic structure of YSb with that of LaSb 4 . Results of those calculations were very similar for both compounds, revealing low densities of states at Fermi level and characteristic anti-crossings leading to band inversion at X-points of the Brillouin zone. LaSb has a simple NaCl-type structure without broken inversion symmetry, perfect linear band crossing or perfect electron-hole symmetry, yet it exhibits the exotic magnetotransport properties of complex-structure semimetals like TaAs, NbP (Weyl semimetals) 5,6 , Cd 3 As 2 (Dirac semimetal) 7 and WTe 2 (resonant compensated semimetal)8,9 . Recently Tafti et al. discovered in LaSb field-induced resistivity plateau at low temperatures up to ≈ 15 K, ultrahigh mobility of carriers in the plateau region, quantum oscillations, and magnetoresistance (MR) of nearly one million percent at 9 T 10 . Their calculations, including spin-orbit coupling (SOC) effect, suggested that LaSb is a topological insulator with a 10 meV gap open near the X-point of the Brillouin zone. They also observed specific angular dependence of frequencies of quantum oscillations and ascribed them to two-dimensional Fermi surface (FS) possibly formed of topologically nontrivial states, and thus proposed LaSb as a model system for understanding the consequences of breaking time-reversal symmetry in topological semimetals 10 .However, such angular dependence has already been observed in LaSb by de Haas-Van Alphen measurements and well explained by the presence of elongated pockets of 3D-Fermi surface 11,12 .
Scientific RepoRts | 6:38691 | DOI: [bib_ref] Resistivity plateau and extreme magnetoresistance in LaSb, Tafti [/bib_ref].1038/srep38691
Tafti et al. also invoked the opening of insulating gap (i.e. metal-insulator transition) as a source of the field-induced resistivity plateau in LaSb 10 but it should be noted that in the case of WTe 2 the existence of a magnetic-field-driven metal-insulator transition has been excluded by means of Kohler scaling analysis of magnetoresistance [bib_ref] Origin of the turn-on temperature behavior in Wte 2, Wang [/bib_ref].
Motivated by these ambiguities in the interpretation of LaSb properties we decided to carry out a comprehensive characterization of magnetotransport properties of a sister compound YSb. We found that YSb displays physical properties in many aspects very similar to those of LaSb. Our results are in accord with those of other groups that appeared during preparation of our article [bib_ref] Magnetotransport of single crystalline YSb, Ghimire [/bib_ref]. The interpretation proposed by follows that presented by for LaSb [bib_ref] Resistivity plateau and extreme magnetoresistance in LaSb, Tafti [/bib_ref] , implying the role of field-induced metal-insulator transition in YSb.
However, our analysis of magnetoresistance and Shubnikov-de Haas (SdH) effect provides strong support for 3D-Fermi surface scenario of magnetotransport, without necessity of invoking topologically non-trivial 2D states or metal-insulator transition in YSb.
# Results
Magnetoresistance and the origin of its plateau. Electrical resistivity (ρ) was measured on two samples (denoted as #1 and #2) and its dependence on temperature in zero field is plotted in. Shape of ρ(T) curves is typical for a metal. When measured in different applied fields ρ(T) exhibits universal plateau at temperatures 2-15 K, as shown for sample #1 in. Temperature range of this plateau is very similar to that reported for LaSb [bib_ref] Resistivity plateau and extreme magnetoresistance in LaSb, Tafti [/bib_ref].
Magnetoresistance, MR ≡ [ρ(B) − ρ(B = 0)]/ρ(B = 0), is plotted versus magnetic field, B, inand (b), for samples #1 and #2, respectively. Following the approach of Wang et al. [bib_ref] Origin of the turn-on temperature behavior in Wte 2, Wang [/bib_ref] we performed Kohler scaling analysis of magnetoresistance to test for the existence of a magnetic-field-driven metal-insulator transition in YSb.and (d) show very good Kohler scaling of our data, MR ∝ (B/ρ 0 ) m , yielding at 2.5 K exponents m = 1.64 and 1.74, respectively, very close for both samples, despite significant difference of their MR values. Efficiency of this scaling indicates that resistivity plateau is due purely to the magnetoresistance, but not to a field-induced metal-insulator transition.
In order to elucidate the dimensionality of FS we measured magnetoresistance of the sample #2 in fields applied at different angles to its surface,. Here θ = 0° denotes the field perpendicular to the sample surface and the current direction, whereas θ = 90° means that the field is parallel to the current. As shown in the inset to, ρ at strongest field of 9 T follows θ ∝cos . This is typical behavior for materials without magnetic anisotropy, but the change of ρ expressed as anisotropic magnetoresistance, AMR ≡ [ρ(90°) − ρ(0°)]/ρ(0°), has an outstanding − 80% value.
Moreover, when field strength is scaled by a factor ε θ dependent on mass anisotropy and θ-angle, all ρ(T) data ofcollapse on single curve, as shown in. Inset ofshows that values of ε θ plotted against field angle θ can be perfectly fitted with ε θ = (cos 2 θ + γ −2 sin 2 θ) 1/2 function, shown with red line. Such scaling has initially been proposed for anisotropic superconductors [bib_ref] From isotropic to anisotropic superconductors: A scaling approach, Blatter [/bib_ref] , and recently used to interpret MR behavior of WTe 2 based on its 3D electronic nature [bib_ref] Temperature-Dependent Three-Dimensional Anisotropy of the Magnetoresistance in WTe 2, Thoutam [/bib_ref]. Parameter γ represents effective mass anisotropy of carriers mostly contributing to the magnetoresistance. We ascribe this behavior to a strongly anisotropic sheet of 3D-FS revealed by SdH effect data, as shown below.
Hall effect. Hall resistivity of sample #1h (cut from the same single crystal as #1) measured at several temperatures between 2.2 K and 300 K is shown in. Nonlinear ρ xy (B) indicates that at least two types of charge carriers are responsible for the Hall effect observed in YSb. The ρ xy (B) curves for temperatures from 2.2-15 K range are almost identical, which points to nearly constant carrier concentrations and mobilities, and coincides with the plateau of ρ xx (T) observed in the same range of T. Changes of sign of ρ xy (B) observed at T ≤ 100 K, indicate conducting bands of both electrons and holes, at higher temperatures ρ xy (B) is positive in the whole range of magnetic field (0 < B ≤ 9T). Thus, the Hall contributions of holes and electrons nearly compensate, but both depend on temperature in different manner. Clear quantum oscillations are observed in ρ xy in temperature range 2.2-15 K (cf.and (b)).
Since ρ xy ≪ ρ xx , the off-diagonal component of conductivity tensor σ ρ ρ ρ = − + /( ) xy xy xx xy 2 2 should be used for multiple-band analysis of Hall data. In this case simple Drude model can be used: summing up conductivities of individual bands, with n i and μ i denoting respectively concentration and mobility of carriers from the i-th band. Asshows, for data collected at 300 K accounting for two bands yielded a good fit. On the other hand, fitting with two bands was insufficient for 2.2 K data, but addition of a contribution of another band with small concentration of more mobile holes brought a very satisfactory fit. Inset: resistivity at 2.5 K and in 9 T versus field rotation angle; blue line represents ρ θ ∝ cos dependence. (b) Data of (a) replotted with B scaled by angle-dependent factor ε θ . Inset: angle dependence of ε θ ; red line represents fit with ε θ = (cos 2 θ + γ −2 sin 2 θ) 1/2 function yielding mass anisotropy γ = 3.4. The fit to data collected at 300 K yielded: the concentrations: n e = 1.34 × 10 18 cm −3 , n h = 1.94 × 10 19 cm −3 , and mobilities μ e = 1.8 × 10 3 cm 2 /(Vs), μ h = 8.2 × 10 2 cm 2 /(Vs), so YSb has similar concentration of carriers but with significantly lower mobility than LaSb. This seems to be the main reason for its significantly smaller magnetoresistance, as it was well demonstrated for WTe . Overall, Hall effect results are in perfect agreement with characteristics of Fermi surface presented in the next section.
[formula] ∑ σ µ µ = + B eB n B ( ) 1 ( ) ,(1) [/formula]
Fermi surface analysis: electronic structure calculations and Shubnikov-de Haas effect. We performed the electronic structure calculations for YSb using a full potential all-electron local orbital code (FPLO) within GGA approximation. [fig_ref] Figure 4: Relativistic energy band structure for YSb calculated using FPLO approach with GGA... [/fig_ref] shows obtained energy band structure for YSb, with a few bands crossing Fermi level. Near X-point there is an anti-crossing present with a gap of ≈ 0.8 meV, similar to those reported for lanthanum monopnictides, which led to the proposal of 2D topologically non-trivial states at the origin of extraordinary behavior of their magnetoresistance [bib_ref] Resistivity plateau and extreme magnetoresistance in LaSb, Tafti [/bib_ref]. Fermi surface resulting from our calculations is presented in [fig_ref] Figure 5: Fermi surface of YSb from electronic band calculations [/fig_ref] : two electron sheets centered at X-points and three nested hole sheets centered at Γ -point. Our calculations are in good agreement with those of ref. 14.
Shubnikov-de Haas oscillations of resistivity are discernible for YSb at temperatures up to 15 K and in fields above 6 T, as seen in Figs 1(a), 1(b) and 2(a). Since ρ xy ≪ ρ xx , we may safely assume that the conductivity σ ρ − xx xx 1 and analyze directly oscillations of ρ xx . Points in [fig_ref] Figure 6 FFT: Shubnikov-de Haas oscillations of resistivity at T = 2 [/fig_ref] represent resistivity of sample #2 measured at 2.5 K, after subtraction of smooth background, plotted versus inverse field 1/B (for 7 < B < 9T). Complex shape of Δ ρ xx (1/B) dependence indicates that observed SdH oscillation has several components. Indeed fast Fourier transform (FFT) analysis reveals clearly six well separated frequencies. We chose to fit Δ ρ xx (1/B) with the multi-frequency Lifshitz-Kosevich function [bib_ref] On the Theory of the Shubnikov-De Haas Effect, Lifshitz [/bib_ref] [bib_ref] The Shubnikov-de Haas Effect: A Powerful Tool for Characterizing Semiconductors, Seiler [/bib_ref] (Eq. 2) because, as the maxima of total Δ ρ xx do not correspond to the maxima of particular components with different frequencies, it is inadequate to determine phases in a multicomponent SdH oscillation using the so-called Landau-level fan diagram (plot of the values of 1/B N corresponding to the N-th maximum in Δ ρ xx versus N). The fit including six components was of very good quality, as shown by blue line in [fig_ref] Figure 6 FFT: Shubnikov-de Haas oscillations of resistivity at T = 2 [/fig_ref]. Obtained parameters are collected in . All frequencies converged almost exactly to f i FFT values obtained from FFT analysis. We ascribe oscillations with frequencies of 720 and 1072 T to second and third harmonic of the strongest one with f i = 360 T. The phases ϕ i of fundamental oscillations resulting from the fit are close to Onsager phase factor of 1/2 expected for free electrons. Thus all components of SdH oscillation can be assigned to 3D-FS sheets predicted by band calculations (shown above in [fig_ref] Figure 5: Fermi surface of YSb from electronic band calculations [/fig_ref] and no Berry phase of π was observed, which could reveal topologically non-trivial charge carriers.
[formula] ∑ ρ π ϕ ∆ = − − − a B c B b B f B 1/ exp( / ) sinh( / ) cos 2 / 1 8 ,(2) [/formula]
We performed FFT analysis for all data sets presented in, which allowed us to observe angular behavior of frequencies corresponding to all extreme cross-sections of Fermi sheets and compare them to those derived from our band structure calculations, as shown in. Three of observed frequencies were clearly changing Green lines indicate cyclotron orbits (extreme cross-sections of FS-sheets α, β, δ and α 1 ), for which we observed SdH oscillations in fields applied at θ = 0° (cf.and . T at θ = 0°), and its harmonics, 2α and 3α. These frequencies are plotted versus θ in.
It became apparent from the shape of FS obtained from band calculations (cf. [fig_ref] Figure 5: Fermi surface of YSb from electronic band calculations [/fig_ref] that angular behavior of α f FFT follows a cross-section area of a prolate ellipsoid (which well approximates the shape of electron sheet centered at X-point shown in 1/2 . This holds for a two-axial ellipsoid described by the equation: (x/k x ) 2 + (y/k x ) 2 + (z/k z ) 2 = 1, with r = k z /k x . Band structure shown in [fig_ref] Figure 4: Relativistic energy band structure for YSb calculated using FPLO approach with GGA... [/fig_ref] yields r ≈ 3.6 (with k x estimated as average size of α-sheet of FS along X-U and X-W lines, and k z as its size along Γ -X line).
After rotation by θ = 90° the α frequency meets the one denoted as α 1 , initially (i.e. at θ = 0°) corresponding to the largest cross-section of the same ellipsoidal FS sheet. Two other observed frequencies, β and δ do not change notably with θ, as expected for almost isotropic hole Fermi sheets centered at Γ -point.
We plotted S(θ) (for r = 3.6) inas solid lines. The θ ∝ − cos 1 dependence, expected for two-dimensional FS, behaves similarly and is shown for comparison with dashed lines. The θ ∝ − cos 1 dependence was used by Tafti et al. [bib_ref] Resistivity plateau and extreme magnetoresistance in LaSb, Tafti [/bib_ref] as a hint of possible topologically-nontrivial states in LaSb. However, Fermi surface of that compound has already been well characterized by band calculations and angle-dependent de Haas-van Alphen measurements [bib_ref] Fermi surface of LaSb and LaBi, Hasegawa [/bib_ref] , revealing FS very similar to the one we found in YSb, namely consisting of one ellipsoidal electron sheet centered at X-point and two isotropic hole pockets centered at Γ -point. assigned angular dependence of the principal de Haas-van Alphen frequency (identical to the SdH frequency in ref. [bib_ref] Resistivity plateau and extreme magnetoresistance in LaSb, Tafti [/bib_ref] to the cross-section of the ellipsoidal sheet S(θ) in accord with our interpretation of S(θ) behavior for YSb. This underscores the similarities between these two compounds and implies that there is no need to invoke topologically non-trivial states to explain exotic magnetotransport properties neither in YSb nor in LaSb (contrary to ref. [bib_ref] Resistivity plateau and extreme magnetoresistance in LaSb, Tafti [/bib_ref].
# Discussion and conclusions
YSb is another material displaying giant magnetoresistance (1100% in 9 T), three orders of magnitude smaller than that of sister compound LaSb 10 , thus it cannot be termed 'extreme magnetoresistance' (XMR). This is due mainly to its lower carrier mobility and weaker electron-hole compensation revealed by our Hall effect measurement.
Kohler scaling analogous to that shown inand (d) has recently been used to explain the remarkable up-turn behavior of MR in WTe 2 without the field-induced metal-insulator transition or significant contribution of an electronic structure change 13 . The same authors have shown that perfect carrier compensation leads to: α (circles) and 2α (squares), as well as 3α (triangles), versus sample rotation angle θ. Dashed lines represent θ ∝ − f cos 1 , solid line corresponds to f = f 0 (sin 2 θ + r 2 cos 2 θ) −1/2 with r = 3.6. This line is redrawn for θ + π/2, reflecting the symmetry of cubic YSb lattice. exponent m = 2 in this scaling. Kohler scaling for YSb yielded for our samples the exponents m = 1.64 and 1.74, which seems related to weaker carrier compensation than nearly perfect one in WTe 2 , where m = 1.92 13 . Thus, analogously to WTe 2 , Kohler scaling indicates that the field-induced metal-insulator transition is unnecessary to explain up-turn and low-temperature plateau of resistivity in YSb. The origin of the up-turn is a combination of magnetoresistance with the low-temperature resistivity plateau present already at zero field. Given the similarity of YSb and LaSb the same may also be true for the latter compound.
[formula] i = α 2α β 3α δ α1 = f f (T) [/formula]
Comparing results of SdH measurements with those of electronic structure calculations we obtained comprehensive description of the Fermi surface of YSb. Presence of both electron and hole sheets of similar volumes provides partial charge compensation responsible for its strong magnetoresistance. Band structures of YSb and LaSb are very similar. All Fermi sheets in YSb but the smallest one centered at Γ -point have their counterparts in LaSb 10-12 . Our analysis of angular behavior of SdH frequencies in YSb indicates it is related to the three-dimensional FS, in line with findings for LaSb, but not connected to possible non-trivial topology of electronic structure analogous to that suggested by Tafti et al. [bib_ref] Resistivity plateau and extreme magnetoresistance in LaSb, Tafti [/bib_ref] Angular behavior of MR can also be perfectly explained by anisotropy of 3D-FS. When field strength is scaled by the angle-dependent factor ε θ , all data ofcollapse on single curve. The effective mass anisotropy factor γ = 3.4, obtained from the fit of ε θ (θ) with the expression θ γ θ + − (cos sin ) 2 2 2 1/2 , is in excellent agreement with k z /k x = 3.6 we estimated for α-sheet of FS. This is not surprising, since the mass anisotropy directly reflects the shape of FS, but it shows that angular behavior of MR in YSb is mainly governed by anisotropic form of α-sheet of FS. That sheet corresponds to the electron band, all other FS-sheets contain holes and are nearly isotropic. The effective mass and mobility of α-sheet electrons change significantly with field angle, which strongly modifies the magnetoresistance.
It has been proposed that the magnetic field induces the reconstruction of the FS in a Dirac semimetal by breaking the time reversal invariance [bib_ref] Dirac Semimetal in Three Dimensions, Young [/bib_ref] [bib_ref] Dirac semimetal and topological phase transitions in A 3 Bi (A =..., Wang [/bib_ref] [bib_ref] Topological nodal semimetals, Burkov [/bib_ref] [bib_ref] Discovery of a three-dimensional topological Dirac semimetal, Na 3 Bi, Liu [/bib_ref]. Assisted by the high mobility of carriers such reconstruction has been suggested to induce very large MR observed in Cd 3 As 2 and NbSb . We also observe features in the electronic structure of YSb, buried under the Fermi level, which may possibly allow the magnetic field to transform this compound into Dirac semimetal. A small gap between inverted bands near the X-point (cf. inset to [fig_ref] Figure 4: Relativistic energy band structure for YSb calculated using FPLO approach with GGA... [/fig_ref] might result in topologically non-trivial states. The effect of FS reconstruction could be similar to temperature-induced Lifshitz transition in WTe 2 [bib_ref] Temperature-Induced Lifshitz Transition in WTe 2, Wu [/bib_ref] , whereas its mechanism might be related, for example, to that of Lifshitz transition driven by magnetic field in CeIrIn . Very recently Dirac states have been observed by angle-resolved-photoemission spectroscopy in NbSb, a compound with bulk electronic structure very similar to that of YSb 31 , however topologically protected states were not detected in YSb by this method.
Although a small contribution of topologically non-trivial 2D states cannot be completely excluded our analysis of magnetoresistance and Shubnikov-de Haas effect provides strong support for 3D-Fermi surface scenario of magnetotransport properties in YSb. Analogous field-induced properties of LaSb can most probably be also described in the framework of 3D multiband model.
# Methods
Measurements were performed using a Physical Property Measurement System (Quantum Design) on two samples cut from one single crystal and labeled as #1 and #1 h, and a sample cut from another single crystal and labeled #2. All samples had shapes of rectangular cuboid with all edges along 〈 1 0 0〉 crystallographic directions. Their sizes were: 0.56 × 0.25 × 0.12 mm 3 , 0.4 × 0.47 × 0.13 mm 3 and 0.41 × 0.32 × 0.09 mm 3 , for samples #1, #1 h and #2, respectively. The electric current was always flowing along [1 0 0] crystallographic direction. Single crystals were grown from Sb flux and their NaCl-type crystal structure was confirmed by powder X-ray diffraction carried out using an X'pert Pro (PANanalytical) diffractometer with Cu-Kα radiation. No other phases were detected and lattice parameter of 6.163 Å was determined, reasonably close to literature value 6.155 Å 3 . Electronic structure calculations were carried out using FPLO-9.00-34 code within generalized gradient approximation (GGA) method [bib_ref] Full-potential nonorthogonal local-orbital minimum-basis band-structure scheme, Koepernik [/bib_ref]. The full-relativistic Dirac equation was solved self-consistently, treating exactly all relativistic effects, including the spin-orbit interaction without any approximations. The Perdew-Burke-Ernzerhof exchange-correlation potential [bib_ref] Generalized gradient approximation made simple, Perdew [/bib_ref] was applied and the energies were converged on a dense k mesh with 24 3 points. The convergence was set to both the density (10 −6 in code specific units) and the total energy (10 −8 Hartree). For the Fermi surface a 64 3 mesh was used to ensure accurate determination of the Fermi level.
[fig] Figure 1: (a) and (b) Magnetoresistance of samples #1 and #2, respectively, versus strength of applied magnetic field at different temperatures. (c) and (d) Kohler scaling of magnetoresistance, MR ∝ (B/ρ 0 ) m fitted to 2.5 K data yields m = 1.64 and 1.74, for samples #1 and #2, respectively. [/fig]
[fig] Figure 2: (a) Resistivity of YSb (sample #2) at 2.5 K versus strength of magnetic field applied at different angles θ. [/fig]
[fig] Figure 3: (a) Hall resistivity of YSb (sample #1h) versus magnetic field recorded at several temperatures. (b) Magnetic field dependent Hall conductivity at T = 2.2 K (left axis) and 300 K (right axis). Lines represent the fits with multiple-band model (Eq. 1). For LaSb Tafti et al. 10 estimated uncompensated carrier concentration, n, and the Hall mobility m H , using the relations n = 1/eR H (0) (with R H (0) being the zero-temperature limit of R H (T)) and μ H = R H (0)/ρ 0 . They obtained n ≈ 10 20 cm −3 and μ H ≈ 10 5 cm 2 /(Vs).For our YSb sample the fit with with multiple-band model (Eq. 1), shown inFig. 3(b), yielded at T = 2.2 K the concentrations: n e = 1.52 × 10 20 cm −3 , n h = 1.16 × 10 20 cm −3 , and mobilities μ e = 2.7 × 10 3 cm 2 /(Vs), μ h = 1.9 × 10 3 cm 2 /(Vs). Third band necessary for that fit consists of more mobile holes [n = 3.4 × 10 19 cm −3 , μ = 7.9 × 10 3 cm 2 /(Vs)]. These results are consistent with band calculations presented below. [/fig]
[fig] Figure 4: Relativistic energy band structure for YSb calculated using FPLO approach with GGA approximation. Bands crossing Fermi level are marked in colors: electron band α in red, hole bands: β and δ in blue, ζ (not observed in SdH oscillations) in green. Vicinity of X-point with an anti-crossing and opened gap is shown as blow-up in inset. [/fig]
[fig] Figure 5: Fermi surface of YSb from electronic band calculations. [/fig]
[fig] Figure 6 FFT: Shubnikov-de Haas oscillations of resistivity at T = 2.5 K. Blue solid line represents the fit with multiple Lifshitz-Kosevich function (Eq. 2). Scientific RepoRts | 6:38691 | DOI: 10.1038/srep38691 upon rotation of the magnetic field: the principal, labeled as α ( [/fig]
[fig] Figure 7: (a) FFT-frequency spectrum of ρ xx (1/B) oscillations of sample #2 measured at different temperatures. Diamond symbols indicate normalized amplitudes obtained from fitting with multiple Lifshitz-Kosevich function shown in Fig. 6. Black solid line (for the sake of clarity offset by 5 units) represents the spectrum of ρ xy (1/B) oscillations of sample #1h at 2.5 K. (b) Frequencies of two strongest FFT components of SdH oscillations shown in [/fig]
[fig] Table 1: Parameters obtained from fit of the multiple Lifshitz-Kosevich function (Eq. 2) to data-points shown in Fig. 6. Frequencies derived by FFT analysis of SdH oscillations of ρ xy are shown for comparison. [/fig]
|
An evaluation of activity tolerance, patient-reported outcomes and satisfaction with the effectiveness of pulmonary daoyin on patients with chronic obstructive pulmonary disease
# Introduction
Chronic obstructive pulmonary disease (COPD) is the forth leading cause of death worldwide, and its prevalence and consequent mortality is expected to increase in the coming decades.In China, COPD is the third leading cause of death from Global Burden of Disease Study 2013, [bib_ref] Cause-specific mortality for 240 causes in China during 1990-2013: a systematic subnational..., Zhou [/bib_ref] and an estimated 65 million people will die of COPD between 2003 and 2033. [bib_ref] Effects of smoking and solid-fuel use on COPD, lung cancer, and tuberculosis..., Lin [/bib_ref] COPD is characterized by a progressive deterioration of debilitating symptoms and increasingly frequent exacerbations. Its common symptoms are chronic cough, abnormal sputum, and breathlessness. Initially, patients lose their exercise tolerance due to airflow limitation, and the condition gradually worsens. The goal of treatment in COPD is to reduce symptoms, frequency and severity of exacerbations, and improve exercise tolerance and health status.Pulmonary rehabilitation (PR) has been defined as an "evidence-based, multidisciplinary and comprehensive intervention for patients with chronic respiratory disease who are symptomatic and often have decreased daily life activities." [bib_ref] European Respiratory Society statement on pulmonary rehabilitation, Nici [/bib_ref] PR is a cornerstone of management for patients with COPD, in whom it decreases respiratory rate (by prolonging expiration). The benefits of PR have been extensively reported in COPD, with the assumption that the recommendations are applicable to subjects with other lung diseases. [bib_ref] Pulmonary rehabilitation for chronic obstructive pulmonary disease, Lacasse [/bib_ref] In 2006, PR was included in the Global Initiative for Chronic Obstructive Lung Disease (GOLD) for the therapy of stable COPD by the American Thoracic Society (ATS) for the first time, and was the only recommended non-pharmacological treatment. [bib_ref] European Respiratory Society statement on pulmonary rehabilitation, Nici [/bib_ref] The positive effects of PR programs on functional outcome parameters such as dyspnea, exercise capacity, and healthrelated quality of life have been investigated and proven mostly in patients with COPD, [bib_ref] Pulmonary rehabilitation for chronic obstructive pulmonary disease, Mccarthy [/bib_ref] [bib_ref] Benefits of pulmonary rehabilitation in patients with COPD and normal exercise capacity, Lan [/bib_ref] [bib_ref] Practical recommendations for exercise training in patients with COPD, Gloeckl [/bib_ref] and are recommended in the recent treatment guideline.Traditional Chinese exercise therapy has a long history. Through a combination of mind, breathing and limb movement, it improves body organs function, regulates blood, dredges the meridians, gains the balance between body and mind, and achieves disease prevention, health care and longevity. The commonly used methods are Daoyin, Tai Chi and Qigong. [bib_ref] Effectiveness of t'ai chi and qigong on chronic obstructive pulmonary disease: a..., Ding [/bib_ref] Daoyin is a form of exercise that is easy to learn and requires no specific training equipment. It should be considered a potential substitute for the PR. Although the effects of daoyin for chronic respiratory disease have been reported in some studies, data supporting it in patients with COPD are scarce and the effects are largely unknown. Therefore, pulmonary daoyin (PD) has been established on the basis of the daoyin skills and theory of Traditional Chinese Medicine (TCM) with the characteristics of COPD. Thus, the aim of our study was to evaluate the benefit of PD program in patients with COPD in terms of activity tolerance, patient-reported outcomes and satisfaction with the effectiveness.
## Pulmonary daoyin
Daoyin is a health preservation and therapy method of ancient China that combines specially designed movements of the limbs and trunk and controlled breathing exercises.It has the effects of dredging the meridians, promoting the circulation of Qi and blood, cultivating primordial Qi, strengthening the body resistance to eliminate pathogenic factors, stretching the tendons and activating the collaterals, and keeping Yin and Yang in balance. [bib_ref] The mechanism and clinical application of daoyin method, Yu [/bib_ref] Daoyin has some advantages and characteristics in prevention and treatment of diseases, and has irreplaceable advantages in preventive treatment of diseases and the application of kinesiotherapy compared with modern medicine. [bib_ref] The mechanism and clinical application of daoyin method, Yu [/bib_ref] Daoyin is Chinese traditional exercise that includes Tai Chi, Qigong and Baduanjin (eight-section brocades). Studies have shown the benefit of Tai Chi and Qigong in improving lung function, physical performance, activity tolerance level and quality of life in COPD patients. 14-16 PD, a TCM PR technology, was established on the basis of the daoyin skills and theory of TCM with the characteristics of COPD. It is a gentle meditative technique that includes a series of physical movements, breathing exercises, and mind regulation. It is based on the principle of integrating and harmonizing one's mind, breath, posture, and movement. Its effectiveness lies in the element of special breathing and respiratory muscle training, which are important aspects of respiratory management. [bib_ref] Functional and psychosocial effects of PD on patients with COPD in China:..., Yu [/bib_ref] Thus the effects of PD on COPD patients are worthy of further investigation.
# Methods
## Study design
The protocol of this study has been published in the Journal of Integrative Medicine in 2013. [bib_ref] Functional and psychosocial effects of PD on patients with COPD in China:..., Yu [/bib_ref] This study utilized a multicenter, cluster randomized controlled clinical trial. Subjects were randomly assigned to one of the two groups, namely PD and control group. The random number was generated by SAS 9.2 software. The trial was registered in the ClinicalTrials.gov (NCT01482000) on 29 November 2011, and was conducted from November 2011 (when the first patient was enrolled) to April 2013 (when the last patient completed). All patients signed the informed consent before inclusion, and ethical approval was obtained from the Ethical Research Committee of the First Affiliated Hospital of Henan University of Traditional Chinese Medicine. The batch number is 2011HL-034.
## Sample size
According to previous results of a Tai Chi and COPD study, [bib_ref] Tai chi Qigong improves lung functions and activity tolerance in COPD clients:..., Chan [/bib_ref] forced expiratory volume in 1 second (FEV 1 ) was 0. compared with the control group (0.85±0.35). The allowable error (δ) value was 0.11, and the SD value was 0.37, with a power of 0.10 at a 5% significant level (two-sided), and 193 subjects per group were required. In order to cover the potential attrition rate of 20%, 464 subjects (232 per group) were targeted.
## Selection of subjects
Subjects clinically diagnosed with COPD according to the Global Strategy for the Diagnosis, Management, and Prevention of COPD, and the Chinese Treatment Guidelines of COPDwere eligible for inclusion in this study. Exclusion criteria can be viewed in the published study protocol. [bib_ref] Functional and psychosocial effects of PD on patients with COPD in China:..., Yu [/bib_ref] Subjects were recruited from the local communities in cities of the 11 research centers.
## Intervention protocol
All subjects received patient education. Patients in the PD group were taught the PD technique and continued with their usual therapy; in the control group, patients continued with their usual therapy. The protocol of this study has been published, and more about the patient's health education and Pulmonary Daoyin actions can be found in the published research program. [bib_ref] Functional and psychosocial effects of PD on patients with COPD in China:..., Yu [/bib_ref] Subjects in the PD group completed a 3-month PD program, which consisted of two sessions per day at least 5 days each week. Along with PD pictures, a DVD was also given to each subject to facilitate daily self-practice. A diary was also provided to each subject for recording the frequency of their self-practice sessions.
## Measurement
The 6-min walking distance (6MWD) test, COPD patientreported outcomes (COPD-PRO) and Effectiveness Satisfaction Questionnaire for COPD (ESQ-COPD) were used because of their simplicity and sensitivity when utilized for health evaluation purposes and their response to the treatment. The 6MWD protocol following ATS guidelines was used.The COPD-PRO 21 and ESQ-COPD 22,23 was developed and validated by our team according to the highest standards and international development specification. Data collection was performed at baseline and 3-months.
# Data analysis
Data analyses were conducted using SAS 9.2 software. Descriptive statistics were used to define the demographic characteristics of the sample. The paired-sampled t-test or independent-sampled t-test were used to examine the outcome measures. To account for differences between the two groups at baseline, comparisons between the PD and control groups were conducted using an analysis-of-variance model including the baseline values as covariates (analysis of covariance). A P-value of 0.05 (two-sided) was taken as the level of significance. To preserve the value of randomization, an intention-to-treat (ITT) analysis was applied. Data of last observations were carried forward for withdrawals.
# Results
## Demographic data
In total, 464 subjects were randomly assigned to each of the following groups: the PD group (n=232) and control group (n=232). A total of 429 patients fully completed the study. Therefore, the per-protocol analysis set (PPS) population was 429 with 213 in the PD group and 216 in the control group. The full analysis set (FAS) population was 461 with 232 in the PD group and 229 in the control group. Patient enrollment and completion values for the study are shown in [fig_ref] Figure 1: The consort flowchart [/fig_ref].
There was no significant difference in gender, age, the course of disease, lung function, and GOLD classification between the two groups (FAS, PPS: P.0.05).
Demographic characteristics by group allocation are shown in [fig_ref] Table 1: Baseline characteristics of the patientsNotes [/fig_ref]. The conventional Western medication of the two groups before randomization and research period are shown in [fig_ref] Table 2: Baseline of original respiratory medications Abbreviations [/fig_ref].
## Comparison of 6mwd
There was no significant difference in the mean value of 6MWD between the two groups before treatment (FAS: P=0.899; PPS: P=0.960). After 3 months, the mean value of 6MWD in all the groups was significantly higher than before treatment (P,0.001; [fig_ref] Figure 2: Comparison of the results of 6-minute walking distance [/fig_ref]. At 3 months, the mean value of 6MWD was significantly higher in the PD group compared with the control group (FAS: P=0.049; PPS: P=0.041). The program appeared to have significant effect on the measures of the difference of 6MWD between the two groups, in change from baseline [fig_ref] Table 4: Change in 6MWD, COPD-PrO and eSQ-COPD baseline values 3 months after enrollment,... [/fig_ref].
## Comparison of copd-pro
There was no significant difference in the COPD-PRO scores of the two groups before treatment (FAS, PPS: P.0.05). After 3 months, the COPD-PRO scores in all groups were significantly lower than before treatment (P,0.001;
## Gold 4
Inhaled corticosteroid and long-acting beta 2 -agonist or long-acting anticholinergic Abbreviations: GOlD, Global Initiative for Obstructive lung Disease; PD, pulmonary daoyin.
group (FAS, PPS: P,0.05). The program appeared to have significant effect on measures of the difference in clinical symptoms, effectiveness satisfaction, health satisfaction and the total score between the two groups in change from baseline [fig_ref] Table 4: Change in 6MWD, COPD-PrO and eSQ-COPD baseline values 3 months after enrollment,... [/fig_ref].
## Comparison of esq-copd
There was no significant difference in the ESQ-COPD scores of the two groups before treatment (FAS, PPS: P.0.05). After 3 months, the ESQ-COPD scores in all groups were significantly higher than before treatment (P,0.001; [fig_ref] Figure 4: Comparison of the results of eSQ-COPD [/fig_ref]. At 3 months, the PD group had significantly higher ESQ-COPD scores compared with those of the control group (FAS, PPS: P,0.05). The program appeared to have significant effect on measures of the difference of capacity for life and work, clinical symptoms, effect of therapy, convenience of therapy, whole effect domain and the total score between the two groups in change from baseline [fig_ref] Table 4: Change in 6MWD, COPD-PrO and eSQ-COPD baseline values 3 months after enrollment,... [/fig_ref].
# Discussion
Patients with COPD frequently complain of dyspnea and exercise limitation and become trapped in a vicious cycle of inactivity, initiated by breathlessness. [bib_ref] Role of exercise in testing and in therapy of COPD, Divo [/bib_ref] [bib_ref] Exercise assessment and training in pulmonary rehabilitation for patients with COPD, Singh [/bib_ref] Exercise training, the important part of PR, has been shown to improve dyspnea and health status and decrease health care use. [bib_ref] Role of exercise in testing and in therapy of COPD, Divo [/bib_ref] In addition to smoking cessation, PR is the most important method in the management of COPD. [bib_ref] Assessment of pulmonary rehabilitation efficacy in chronic obstructive pulmonary disease patients using..., Candemir [/bib_ref] Although the effectiveness of daoyin in COPD patients has been reported, most studies were not randomized controlled trials. In addition, sample sizes were small, outcome parameters were not accurate and there was no specific rehabilitation technique for COPD. Therefore, a randomized controlled trial was submit your manuscript | www.dovepress.com
## Dovepress
## Dovepress
## 2338
Zhang et al conducted and demonstrated the efficacy of PD in patients with COPD. Our study suggests that the PD program has positive effects on exercise capacity in COPD patients.
The results of our study show that the patient-reported outcomes were reduced, while satisfaction with the effectiveness of the patients significantly increased after the 3-month-long PD program. This exercise program can be recommended as an effective alternative training modality in PR programs. Decreased exercise capacity is one of the main symptoms of COPD patients. Numerous studies have confirmed reduced exercise tolerance of COPD patients with increased airway resistance, ineffective ventilation, hyperinflation and increased elastic load to breathing, gas exchange abnormalities and a mechanical disadvantage of the respiratory muscles; 27,28 moreover, reduced exercise capacity further reduces patients' quality of life. [bib_ref] Effects of Qigong on respiratory function and quality of life in patients..., Zhu [/bib_ref] In addition, skeletal muscle dysfunction has been reported as an important contributor to exercise limitation in COPD. Therefore, assessment of exercise capacity helps evaluate motor function, quality of life, and prognosis in these patients.
The 6MWD mainly reflects the exercise capacity of patients for comprehensive evaluation of moderate-tosevere disease of systemic functional status. It is a valid indicator for evaluating exercise capacity in patients with clinically stable COPD. [bib_ref] Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE) study investigators...., Spruit [/bib_ref] [bib_ref] Interpretation of treatment changes in 6 minute walk distance in patients with..., Puhan [/bib_ref] The 6MWD is also considered useful in the determination of the point at which patients should be listed for rehabilitation programs. [bib_ref] Six-minute walk distance in patients with chronic obstructive pulmonary disease: which reference..., Andrianopoulos [/bib_ref] The analysis of 6MWD in this study showed a statistically significant increase in the PD group compared with the control group. The PD program has been established on the basis of the daoyin skills and theory of TCM with the characteristics of COPD. The improvement of exercise capacity is considered to be associated with physical movements and breathing exercises, thereby improving ventilation and exercising skeletal muscle. PD is a form of exercise that requires minimal equipment and no specific training facility, and should be considered a potential substitute for the PR that is currently prescribed.
Symptoms in COPD patients may impair exercise capacity and quality of life. Therefore, an important aim of the assessment and treatment programs in COPD patients is symptom management. In the present study, PR has good advantages in improving clinical symptoms and enhancing the quality of life. The COPD-PRO has inherent correlation with the evaluation of efficacy of Chinese medicine based on clinical symptoms. The COPD-PRO as a valid indicator for evaluating PR can solve the problem of a single evaluation. According to the highest standards and procedures of international scales, the COPD-PRO was developed and validated by our study group. [bib_ref] Development and validation of a patient reported outcome instrument for chronic obstructive..., Li [/bib_ref] The COPD-PRO contains 17 items in three domains: amelioration of clinical symptoms, satisfaction of health condition, and satisfaction of treatment effect. The COPD-PRO has good validity, reliability and responsiveness. The COPD-PRO can provide patients' responses to the treatments and then evaluate the effect of treatment in a standardized way. Using the COPD-PRO, our results showed that after 3 months of PR, the improvement of effectiveness satisfaction was 13.55% in the PD group and 9.51% in the control group, and there was more improvement of the effectiveness satisfaction scores of the COPD-PRO in the PD group than that in the control group. Dyspnea, cough, and phlegm have been shown to be the main COPD symptoms. The PD program improves the quality of life by reducing symptoms of COPD based on the principle of integrating and harmonizing one's mind, breath, posture, and movement. Effectiveness satisfaction is defined as how patients evaluate the process of taking the current treatment and their response to the treatment. [bib_ref] Satisfaction with medication: an overview of conceptual, methodologic, and regulatory issues, Shikiar [/bib_ref] The evaluation of effectiveness satisfaction, although inherently subjective, can reflect the patient's unique perspective and perceptions on the process, efficiency, and outcomes of the medical care and treatment. [bib_ref] Treatment satisfaction instruments for different purposes during a product's lifecycle: keeping the..., Rofail [/bib_ref] The ESQ-COPD was developed and examined by our group according to the highest standards and procedures of international scales, and contains 18 items in five domains: capacity for life and work (five items), clinical symptoms (five items), effect of therapy (four items), convenience of therapy (three items), and whole effect domain (one item). [bib_ref] Study on the effectiveness satisfaction questionnaire of COPD for chronic obstructive pulmonary..., Li [/bib_ref] The ESQ-COPD proved to be a reliable and structurally valid instrument through evaluation, [bib_ref] The development of the Effectiveness Satisfaction Questionnaire of COPD for chronic obstructive..., Li [/bib_ref] and has been used to evaluate the effect of traditional Chinese medicine on satisfaction in COPD patients. [bib_ref] An evaluation of self-efficacy and satisfaction with the effectiveness of Bu-Fei Yi-Shen..., Li [/bib_ref] Using the ESQ-COPD, our results showed that after 3 months of PR, the improvement of effectiveness satisfaction was 13.60% in the PD group and 7.82% in the control group, and there was more improvement of the effectiveness satisfaction scores of the ESQ-COPD in the PD group than in the control group. The PD program improves the COPD patient's efficacy satisfaction by improving exercise tolerance and reducing clinical symptoms. Despite the fact that much of the evidence pertaining to the physiological benefit of exercise is based on conventional physical exercise, such as Tai Chi, Baduanjin (eight-section brocades) and Qigong, this study confirms that PD program has a good clinical efficacy. "PD is a gentle meditative technique that applies a series of physical movements, breathing exercises, and mind regulation. It is based on the principle of integrating and harmonizing one's mind, breath, posture, and movement. Its effectiveness lies in the element of special breathing and respiratory muscle training which are important aspects of COPD management." [bib_ref] Functional and psychosocial effects of PD on patients with COPD in China:..., Yu [/bib_ref] Thus PD may be a suitable exercise for COPD patients in the community.
# Conclusion
This study confirms that the 3-month PD program has beneficial effects on activity capacity and satisfaction of effectiveness of COPD patients. The benefits of practicing PD on COPD clients should be further investigated with a longer follow-up period in order to detect further improvements in physiological and psychosocial status.
[fig] Figure 3: At 3 months, the PD group had significantly lower COPD-PRO scores compared with those of the control International Journal of COPD 2017:12 submit your manuscript | www.dovepress. [/fig]
[fig] Figure 1: The consort flowchart: to track participants through the randomized controlled trial. Abbreviation: PD, pulmonary daoyin. [/fig]
[fig] Figure 2: Comparison of the results of 6-minute walking distance. Abbreviations: FaS, full analysis set; PPS, per-protocol analysis set. [/fig]
[fig] Figure 4: Comparison of the results of eSQ-COPD. Note: The specific changes and comparisons of the results for ESQ-COPD in capacity for life and work domain (A), clinical symptoms domain (B), effect of therapy domain (C), convenience of therapy domain (D), whole effect domain (E) and total score (F) at baseline and third month between two groups. Abbreviations: COPD, chronic obstructive pulmonary disease; eSQ-COPD, effectiveness Satisfaction Questionnaire for COPD; FaS, full analysis set; PPS, per-protocol analysis set. [/fig]
[table] Table 1: Baseline characteristics of the patientsNotes: Data presented as mean ± standard deviation or n. a The course of disease was calculated in months. b The BMI is the weight in kilograms divided by the square of the height in meters. c exacerbations during the 12 months before screening were self-reported. d Clinical data are from visit 1 (the screening visit). e Severity grades of lung function were determined by guidelines of COPD. Abbreviations: BMI, body mass index; FeV 1 , forced expiratory volume in 1 second; FVC, forced vital capacity; GOlD, Global Initiative for Obstructive lung Disease. [/table]
[table] Table 2: Baseline of original respiratory medications Abbreviations: GOlD, Global Initiative for Obstructive lung Disease; PD, pulmonary daoyin. [/table]
[table] Table 3: Comparison of the usual therapy in both groups [/table]
[table] Table 4: Change in 6MWD, COPD-PrO and eSQ-COPD baseline values 3 months after enrollment, by intervention status [/table]
|
Mammography Screening in a Large Health System Following the U.S. Preventive Services Task Force Recommendations and the Affordable Care Act
BackgroundPractice recommendations for mammography screening were issued by the U.S. Preventive Services Task Force in 2009 and expansion of insurance coverage was provided under the Patient Protection and Affordable Care Act soon thereafter, yet the influence of these changes on screening practices in the United States is not known.MethodsTo determine changes in mammography screening and their associations with new practice recommendations and the Affordable Care Act, we examined patient-level data from 249,803 screening mammograms from January 1, 2008 through December 31, 2012 in a large community-based health system in the northwestern United States. Associations were determined by an intervention analysis of time-series data method.ResultsAmong women screened, 64% were age 50-74 years; 84% self-identified as white race; 62% had commercial insurance; and 70% were seen in facilities located in metropolitan areas. Practice recommendations were associated with decreased screening volumes among women age <40 (-37.4 mammograms/month; -39.4% change; P<0.001), 40-49 (-106.0 mammograms/month; -11.2% change; P<0.001), and 75 (-54.7 mammograms/ month; -10.0% change; P<0.001), but not women age 50-74. Implementation of the Affordable Care Act was associated with increased screening among women age 50-74 (+184.3 mammograms/month; +7.2% change; P=0.001), but not women <40 or 75; increases for age 40-49 were of borderline statistical significance (+56.9 mammograms/month; +6% change; P=0.06). Practice recommendations were also associated with decreased PLOS ONE | screening for women with commercial insurance, while the Affordable Care Act was associated with increased screening for women with Medicare, Medicaid, or other noncommercial sources of payment.ConclusionsMammography screening volumes in a large community health system decreased among women age <50 and 75 in association with new U.S. Preventive Services Task Force practice recommendations, while insurance coverage changes under the Affordable Care Act were associated with increased screening volumes among women age 50-74.Influences on Mammography Screening PLOS ONE |
# Introduction
In November 2009, the U.S. Preventive Services Task Force (USPSTF) issued new age-based recommendations for mammography screening for women without high risks for breast cancerbased on research of its benefits and harms [bib_ref] Screening for breast cancer: an update for the U.S. Preventive Services Task..., Nelson [/bib_ref]. For all women age 50-74 years, they recommended routine mammography every two years. For women age 40-49, they recommended selective screening depending on risk factors for breast cancer and personal values regarding the trade-offs between benefits and harms. The USPSTF acknowledged that research among women age 75 and older was too limited to support a recommendation and screening would depend on individual considerations. The new recommendations represented a major change from the previous practice of screening every one to two years beginning at age 40that continues to be supported by several organizations in the United States [bib_ref] Breast cancer screening with imaging: recommendations from the Society of Breast Imaging..., Lee [/bib_ref].
Four months later, the Patient Protection and Affordable Care Act was passed by the U.S. Congress mandating insurance coverage for mammography screening with no co-pay or deductible fees. Coverage includes yearly mammography for women age 40 and older beginning in the next plan year after October 2010 for health plans created after passage of the law. Although Medicare, Medicaid, and all but one state had existing regulations regarding coverage of mammography screening, these varied widely.
How these changes influenced breast cancer screening practices among women in the United States is unclear. Studies of the effects of the new USPSTF recommendations have mixed results. Surveys of women providing self-reported mammography information indicated no statistically significant changes in screening rates [bib_ref] Mammography rates after the 2009 U.S. Preventive Services Task Force breast cancer..., Howard [/bib_ref] [bib_ref] Trends in mammography screening rates after publication of the 2009 U.S. Preventive..., Pace [/bib_ref] , whereas data from a statewide registry of screening mammography [bib_ref] Registry-based study of trends in breast cancer screening mammography before and after..., Sprague [/bib_ref] and from insurance claims [bib_ref] Mammography rates 3 years after the 2009 US Preventive Services Task Force..., Wharam [/bib_ref] showed significant decreases in screening. These discrepancies may be due to the methodological limitations of self-reported survey data that have been shown to overestimate actual mammography rates [bib_ref] Bias associated with self-report of prior screening mammography, Cronin [/bib_ref] [bib_ref] Utilization of screening mammography in New Hampshire: a population-based assessment, Carney [/bib_ref] , although other contributing factors may also exist. A study of women in their forties indicated a modest decline in screening rates initially after the new guidelines were released, followed by an increase in rates two years later [bib_ref] Preventive Services Task Force guidelines on screening mammography rates on women in..., Wang [/bib_ref]. However, this study did not consider the subsequent effects of the Affordable Care Act on mammography screening. Understanding the influence of both of these changes is important to consumers, policy makers, and health systems responsible for delivering effective screening services.
The objective of this study is to consider the influences of both the new USPSTF recommendations and the implementation of the Affordable Care Act on changes in mammography screening by examining patient-level mammography data across five years from women receiving care in a large community-based health system.
# Methods
This study was approved by the Providence Institutional Review Board and Providence Privacy Board. Informed consent was not required because the data were analyzed anonymously. This study is an intervention analysis of patient-level time-series data obtained from January 1, 2008 through December 31, 2012. Monthly mammography screening volume is the main outcome measure.
## Data sources
The study was based at a large not-for-profit community health system serving urban and rural populations in Oregon and southwest Washington (Providence Health & Services, Oregon Region). This is an integrated health system of eight community hospitals and affiliated outpatient facilities that provides comprehensive care for breast cancer and related conditions including screening, diagnosis, treatment, and survivorship care. Since Providence is an open-access health system, patients are covered by a variety of private and public health insurance plans or receive unsponsored or charity care, and closely match the demographic and socio-economic profiles of their communities. Physician practices are similarly varied, and include private as well as health system employed groups. While the health system provided guidance about mammography screening on their website that incorporated much of the language and rationale of the USPSTF recommendation, it issued no specific requirements or mandates.
Data for this analysis were obtained from women receiving mammography screening within the health system as part of their usual health care. The health system uses a master patient identifier for each patient accessing its extensive clinical network of hospitals and clinics. The master patient identifier and other unique patient identifiers are used to track patients across multiple encounters and over time. A breast care specific data mart developed by the health system integrates patient-level data from various internal sources, including administrative databases, electronic medical records, imaging data, and pathology data from the laboratory information system [bib_ref] Actualizing personalized healthcare for women through connected data systems: breast cancer screening..., Nelson [/bib_ref] [bib_ref] Development of an electronic breast pathology database in a community health system, Nelson [/bib_ref]. Data are subsequently linked based on matching algorithms that group data for individual patients creating disease-specific data tables accessed by customized interactive queries. Data security and patient confidentiality are protected by existing health system safeguards and procedures.
All screening mammograms (Current Procedural Terminology codes G0202/77052) obtained within the health system from January 1, 2008 through December 31, 2012 were imported to the data mart. Mammograms obtained for purposes other than screening, such as diagnostic or unilateral views, were not included. Each patient provided data for only one screening mammogram per year. The study dates were selected to capture mammography volume before, during, and after the release of the USPSTF screening recommendations in November 2009 and implementation of coverage for mammography screening under the Affordable Care Act in January 2011. Although coverage was implemented by different health plans at different times, January 1, 2011 represents the point in which the majority of plans at the health system had changed their coverage policies for mammography screening because it was the start of a new insurance year.
Although our primary patient variable was age at the time of mammography, we were also interested in how other factors influenced screening changes, including insurance type, race and ethnicity (self-reported by patients), and location of mammography. Thus, each patient's medical record number, date of birth, date of mammography, race and ethnicity, insurance type, and facility location were also imported to the data mart.
# Statistical analysis
Age was categorized according to age-based screening recommendations (<40, 40-49, 50-74, and 75 years); insurance type by commercial, Medicare, Medicaid, and other sources of payment; and facility location by whether it was located in a metropolitan or non-metropolitan area. Comparisons of age, insurance type, and facility location by year; and age by insurance type and facility location were determined by the Chi-square test.
Monthly mammography volumes across the health system were treated as time series data that could be affected by the two specified interventions: the announcement of new USPSTF screening recommendations and the implementation of insurance coverage under the Affordable Care Act. We used this approach because it provides a method to assess the effect of interventions over time despite the influence of multiple other effects that are not related to the primary interventions, such as background noise and seasonal changes, and in the absence of a fixed patient denominator.
Using an intervention analysis for time series data method [bib_ref] Intervention analysis with applications to economic and environmental problems, Box [/bib_ref] , we modeled monthly mammography volume by simultaneously incorporating two additive parts. First, the underlying intrinsic fluctuation (i.e. fluctuations attributed to background noise and seasonal changes) was modeled using the integrated autoregressive moving average (ARIMA) regression approach. Secondly, the effects of interventions (i.e., new recommendations and insurance coverage) on the mean monthly mammography volume were modeled as simple step changes. For this study, the interventions were presented as step functions at the time of the start of their implementation (December 2009 for USPSTF recommendations and January 2011 for the Affordable Care Act), and their effects were assumed to be simple step changes in the screening process. Mean monthly mammography volumes were obtained from the model for the effect of no intervention to establish baseline measures and for the effect of change due to each intervention. The analysis was repeated stratified by age (<40, 40-49, 50-74, and 75 years) and insurance categories (commercial, Medicare, Medicaid, and other sources of payment). To account for the possibility that changes occurred in a graded rather than step fashion, we modified the modeling process with a ramping up effect to see if results differed. All analyses were performed using R software (version 3.1.0); the intervention analysis used the TSA package.
# Results
A total of 249,803 screening mammograms were obtained from January 1, 2008 through December 31, 2012. The majority of mammograms were from women between the ages of 50-74 years (64.1%); and 21.8% were 40-49, 12.5% were 75 and older, and 1.7% were younger than 40 [fig_ref] Table 1: Screening Mammograms by Age, n [/fig_ref].
At baseline, the types of insurance coverage varied across age groups (P<0.001). Most women age 74 and younger were covered by commercial insurance (61.9%), while nearly all women age 75 and older were covered by Medicare (98.6%). Overall, 2.3% of women were covered by Medicaid and 3.2% by other sources including self-pay and charity care. When adjusted for age, the proportion of women covered by commercial insurance decreased over time (65.1% in 2008 to 60.3% in 2012, P<0.001), while proportions covered by Medicare, Medicaid, or other sources increased.
The majority of women in all age groups obtained their mammography in health system facilities located in a metropolitan compared with non-metropolitan area (70.2% versus 29.8%, P<0.001). This distribution did not change over time and was not further considered in the statistical models. The majority of women were white (84%), followed by a category including combined, other, and unknown racial heritage (8%), Asian (6%), and African American (2%).
The completeness of data on race and ethnicity varied between facility locations, rendering it unreliable for statistical models.
## Associations of the uspstf recommendations and the affordable care act on screening volume
Mammography screening volume (mean number of mammograms/month) displayed yearly cyclic changes. These cycles have been observed over preceding years and are related to outreach activities during breast cancer awareness month each October as well as general increases in health care utilization at the end of each insurance year. Overall, the USPSTF recommendations were associated with a decrease in volume that was not statistically significant (-138.9 mammograms/month, P = 0.10), while implementation of the Affordable Care Act was associated with increased volume (+232.7 mammograms/month, P = 0.006) [fig_ref] Table 2: Associations of Screening Recommendations and Affordable Care Act Implementation on Volume of... [/fig_ref].
The effect of the USPSTF recommendations on screening volume varied by age and insurance type. Volumes decreased from baseline for women younger than 40 (-37.4 mammograms/month, P<0.001), 40-49 (-106.0 mammograms/month, P<0.001) [fig_ref] Fig 1: Screening Mammography Changes for Women <40 and 40-49 [/fig_ref] , and 75 and older (-54.7 mammograms/month, P<0.001), but did not change for women 50-74 [fig_ref] Fig 2: Screening Mammography Changes for Women 50-74 and 75 [/fig_ref]. When stratified by insurance type, volume decreased from baseline for commercial insurance (-125.5 mammograms/month, P = 0.02), but not for other insurance types.
Implementation of the Affordable Care Act was associated with increased screening volume for women 50-74 (+184.3 mammograms/month, P = 0.001), and among women with Medicare (+86.2 mammograms/month, P<0.001), Medicaid (+20.5 mammograms/month, P = 0.002), and other noncommercial sources of payment (+43.3 mammograms/month, P = 0.003) compared to baseline [fig_ref] Table 2: Associations of Screening Recommendations and Affordable Care Act Implementation on Volume of... [/fig_ref]. Changes were not statistically significant in other age groups or for commercial insurance.
Results were similar when we modified the modeling process with a ramping up effect to account for gradual rather than stepped interventions.
# Discussion and conclusions
Mammography screening in a community health system increased among women age 50-74 and decreased for other ages over five years in association with new practice recommendations and insurance coverage changes. While the USPSTF recommendations were specifically associated with decreased screening among women younger than 50 and 75 and older, but not women age 50-74, the Affordable Care Act was associated with increased screening among women age 50-74, but not women in the other age groups. These results are consistent with current practice standards and quality performance measures, such as revised HEDIS (Healthcare Effectiveness Data and Information Set) measures, that target screening for women age 50-74. The USPSTF recommendations were associated with decreased screening among women in age groups for whom recommendations either did not apply (younger than 40); were changed from universal to selective (40-49); or the evidence of effectiveness was lacking (75 and older). The recommendation may have discouraged screening among women younger than 40 who were never appropriately included in the screening pool unless they were at high-risk for breast cancer (e.g., genetic mutation, previous high risk breast lesion). Although many women continued screening after the new recommendations were issued, our study was unable to determine the rationale and decision making process for screening in these women. Implementation of the Affordable Care Act was associated with increased screening among women age 50-74, reinforcing screening efforts in this population. It is possible that economic barriers to screening, whether real or perceived, may have been diminished by insurance coverage expansion. Screening in other age groups was not affected by the Affordable Care Act.
The effects of the USPSTF recommendations and Affordable Care Act on women with different types of insurance were mixed. The USPSTF recommendations were associated with decreased screening among women with commercial insurance, while the Affordable Care Act was associated with increased screening for women with Medicare, Medicaid, or other noncommercial sources of payment. Since age and insurance type are significantly related (P<0.001), their relationships with screening were confounded. Our results are inconsistent with other published studies of the impact of the USPSTF recommendations [bib_ref] Mammography rates after the 2009 U.S. Preventive Services Task Force breast cancer..., Howard [/bib_ref] [bib_ref] Trends in mammography screening rates after publication of the 2009 U.S. Preventive..., Pace [/bib_ref] [bib_ref] Registry-based study of trends in breast cancer screening mammography before and after..., Sprague [/bib_ref] [bib_ref] Mammography rates 3 years after the 2009 US Preventive Services Task Force..., Wharam [/bib_ref]. Two smaller studies based on self-reported information indicated no statistically significant differences in mammography rates over time for all age groups studied [bib_ref] Mammography rates after the 2009 U.S. Preventive Services Task Force breast cancer..., Howard [/bib_ref] [bib_ref] Trends in mammography screening rates after publication of the 2009 U.S. Preventive..., Pace [/bib_ref]. These include a study of 29,857 women responding to Medical Expenditure Panel Surveys from 2006 to 2010 [bib_ref] Mammography rates after the 2009 U.S. Preventive Services Task Force breast cancer..., Howard [/bib_ref] , and a study of 27,829 women age 40 and older from the National Health Interview Survey from 2008 to 2011 [bib_ref] Trends in mammography screening rates after publication of the 2009 U.S. Preventive..., Pace [/bib_ref]. In contrast, two large studies indicated statistically significant declines in mammography [bib_ref] Registry-based study of trends in breast cancer screening mammography before and after..., Sprague [/bib_ref] [bib_ref] Mammography rates 3 years after the 2009 US Preventive Services Task Force..., Wharam [/bib_ref]. These include a retrospective study of a statewide mammography registry of 150,000 women age 40 and older from 2009 to 2011 [bib_ref] Registry-based study of trends in breast cancer screening mammography before and after..., Sprague [/bib_ref] , and an analysis of 5.5 million insured women age 40-64 from 2005 to 2013 [bib_ref] Mammography rates 3 years after the 2009 US Preventive Services Task Force..., Wharam [/bib_ref]. None of these studies consider other influences on screening rates besides the USPSTF recommendations. They also differ by using self-reported or registry data, comparing screening rates from one time period with another, using shorter or dissimilar time periods, or enrolling varying numbers of participants.
A recently published time-series analysis of several million privately insured women obtaining screening mammography between 2006 and 2011 reported a decline in screening rates two months after release of the USPSTF recommendation among women age 40-49, but not women age 50-64 [bib_ref] Preventive Services Task Force guidelines on screening mammography rates on women in..., Wang [/bib_ref]. However, this decline was not sustained after two years, contrasting with our results that indicated a sustained decline in screening for age 40-49. Results of both studies showed no significant declines for women age 50-64 for whom screening recommendations did not change. Differences in results for women age 40-49 could be related to methodological differences between the studies, particularly the inclusion of only privately insured women compared with all types of payers; inclusion of data from different time periods; use of interrupted time-series analysis versus intervention analysis for time-series data; including screening and diagnostic mammograms compared with screening mammograms only; and assumptions of expected baseline screening rates. Importantly, our study also considered the concurrent effect of implementation of the Affordable Care Act that exhibited its own effect on screening.
The strengths of this study include its use of patient-level data from a large health system that closely represents the diversity of patients, providers, and practices across a broad geographic region, improving its applicability. In addition, we used highly reliable data from actual screening mammography, rather than self-reported accounts, from nearly 250,000 encounters over a sufficiently long period of time to appropriately evaluate changes and interventions. We used a highly accurate method of identifying and integrating patient data across the health system assuring completeness and accuracy.
This study may be limited by its use of data for women actually screened without considering the pool of candidates that was not screened, as would be available in a population-based registry or closed health system. Patients enter and leave the health system and it is possible that specific subgroups preferentially migrated, such as those of a certain age or insurance type. However, major demographic shifts of patients obtaining care at the health system for other health services have not been observed. Patient migration could potentially occur in closed systems and registries over time as well. Also, we used an analytic approach that was appropriate for our data and was not dependent on screening rates. Health systems often require pragmatic approaches to analyzing and interpreting data based on electronic health system data sources (i.e., big data) such as we used in this study.
The analysis was based on several assumptions. Our approach assumed that other than the specified interventions, all other factors that might affect the screening volume (the underlying intrinsic fluctuation of the process) stayed the same over the study period. It is possible that other factors could have had important influences. Also, we assigned a specific time point for implementation of the Affordable Care Act, when there was actually a phasing in period. However, our rationale for using January 2011 as the starting point was based on the implementation of insurance plans in our health system and the simplicity and ease of interpretation of the model. We assumed that intervention effects were step changes, even though changes in medical practice are often gradual. To test this assumption, we modified the modeling process with a ramping up effect and the results did not indicate differences in the model. Although these findings support our assumptions, they could also reflect the limited number of time points in the time series rather than the absence of effect. In addition, our assumption that the interventions only affected the mean mammography volume may not be correct if the intervention also changed the intrinsic fluctuation (e.g., by altering the variance of the background noise). Due to the small number of time points in the data, the validity of this assumption could not be meaningfully assessed.
Future studies of screening changes in other health systems could provide additional information about the impact of changes in practice recommendations and insurance coverage, as well as other factors. Results of these studies would be particularly useful to health systems providing screening services.
In conclusion, mammography screening volumes in a large community health system decreased among women younger than 50 and 75 and older in association with new USPSTF practice recommendations, while insurance coverage changes under the Affordable Care Act were associated with increased screening volumes among women age 50-74. These changes align with current practice standards and quality performance measures that target screening for women age 50-74.
[fig] Fig 1: Screening Mammography Changes for Women <40 and 40-49. The mean number of screening mammograms per month performed in the health system from 2008 through 2012. Arrows indicate the times of new screening recommendations and implementation of the Affordable Care Act. Recommendations were associated with decreased screening for women <50, while the Affordable Care Act had no statistically significant associations. doi:10.1371/journal.pone.0131903.g001 [/fig]
[fig] Fig 2: Screening Mammography Changes for Women 50-74 and 75. The mean number of screening mammograms per month performed in the health system from 2008 through 2012. Arrows indicate the times of new screening recommendations and implementation of the Affordable Care Act. New recommendations were not associated with changes for women age 50-74, but were associated with decreased screening for 75, while the Affordable Care Act was associated with increased screening among women age 50-74 and no changes for 75. doi:10.1371/journal.pone.0131903.g002 [/fig]
[table] Table 1: Screening Mammograms by Age, n (row %). doi:10.1371/journal.pone.0131903.t001 [/table]
[table] Table 2: Associations of Screening Recommendations and Affordable Care Act Implementation on Volume of Screening Mammography. [/table]
|
Reexpression of LSAMP inhibits tumor growth in a preclinical osteosarcoma model
Background: Osteosarcomas are the most common primary malignant tumors of bone, showing complex chromosomal rearrangements with multiple gains and losses. A frequent deletion within the chromosomal region 3q13.31 has been identified by us and others, and is mainly reported to be present in osteosarcomas. The purpose of the study was to further characterize the frequency and the extent of the deletion in an extended panel of osteosarcoma samples, and the expression level of the affected genes within the region. We have identified LSAMP as the target gene for the deletion, and have studied the functional implications of LSAMP-reexpression. Methods: LSAMP copy number, expression level and protein level were investigated by quantitative PCR and western blotting in an osteosarcoma panel. The expression of LSAMP was restored in an osteosarcoma cell line, and differences in proliferation rate, tumor formation, gene expression, migration rate, differentiation capabilities, cell cycle distribution and apoptosis were investigated by metabolic dyes, tumor formation in vivo, gene expression profiling, time-lapse photography, differentiation techniques and flow cytometry, respectively.Results:We found reduced copy number of LSAMP in 45/76 osteosarcoma samples, reduced expression level in 25/42 samples and protein expression in 9/42 samples. By restoring the expression of LSAMP in a cell line with a homozygous deletion of the gene, the proliferation rate in vitro was significantly reduced and tumor growth in vivo was significantly delayed. In response to reexpression of LSAMP, mRNA expression profiling revealed consistent upregulation of the genes hairy and enhancer of split 1 (HES1), cancer/testis antigen 2 (CTAG2) and kruppel-like factor 10 (KLF10).Conclusions:The high frequency and the specificity of the deletion indicate that it is important for the development of osteosarcomas. The deletion targets the tumor suppressor LSAMP, and based on the functional evidence, the tumor suppressor function of LSAMP is most likely exerted by reducing the proliferation rate of the tumor cells, possibly by indirectly upregulating one or more of the genes HES1, CTAG2 or KLF10. To our knowledge, this study describes novel functions of LSAMP, a first step to understanding the functional role of this specific deletion in osteosarcomas.
# Background
Osteosarcomas are the most common primary malignant tumors of bone. They are highly aggressive with poor prognosis [bib_ref] The genetics of osteosarcoma, Martin [/bib_ref] and occur most frequently in children and adolescents [bib_ref] Bone cancers, Dorfman [/bib_ref]. The efficacy of the current treatments has reached a plateau, and the need of increased biological understanding is crucial to improve treatment options and thus the life of patients.
At the genomic level, osteosarcomas show complex chromosomal rearrangements with multiple gains and losses [bib_ref] Focus on sarcomas, Mackall [/bib_ref] [bib_ref] Mechanisms of sarcoma development, Helman [/bib_ref]. Array comparative genomic hybridization (aCGH) has been used extensively to analyze DNA copy number changes at a higher resolution, identifying recurrent chromosomal alterations [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref] [bib_ref] Evaluation of high-resolution microarray platforms for genomic profiling of bone tumours, Kresse [/bib_ref] [bib_ref] Genome-wide array comparative genomic hybridization analysis reveals distinct amplifications in osteosarcoma, Man [/bib_ref] [bib_ref] Genomic imbalances associated with methotrexate resistance in human osteosarcoma cell lines detected..., Hattinger [/bib_ref] [bib_ref] Gene amplifications in osteosarcoma-CGH microarray analysis, Atiye [/bib_ref] [bib_ref] Genomic signatures of chromosomal instability and osteosarcoma progression detected by high resolution..., Selvarajah [/bib_ref]. We have previously identified a novel, frequent deletion in 3q13.31 in osteosarcomas [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref]. Of the genes located within the deleted region, three have been proposed to be involved in cancer biology: the protein-coding gene limbic system-associated membrane protein (LSAMP) and the two non-coding RNAs LSAMP RNA antisense 3 (LSAMP-AS3) (also known as LOC285194 or TUSC7) and LSAMP RNA antisense 4 (LSAMP-AS4) (also known as BC040587) [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref] [bib_ref] Expression analysis of genes associated with human osteosarcoma tumors shows correlation of..., Sadikovic [/bib_ref] [bib_ref] Recurrent focal copy-number changes and loss of heterozygosity implicate two noncoding RNAs..., Pasic [/bib_ref] [bib_ref] Identification of chromosomal aberrations associated with disease progression and a novel 3q13.31..., Yen [/bib_ref] [bib_ref] The IgLON family in epithelial ovarian cancer: expression profiles and clinicopathologic correlates, Ntougkos [/bib_ref] [bib_ref] LSAMP and norei: novel genes in hereditary and sporadic clear cell renal..., Kanayama [/bib_ref] [bib_ref] The t(1;3) breakpointspanning involved in clear cell renal cell genes LSAMP and..., Chen [/bib_ref] [bib_ref] LncRNA loc285194 is a p53-regulated tumor suppressor, Liu [/bib_ref] [bib_ref] Low expression of LOC285194 is associated with poor prognosis in colorectal cancer, Qi [/bib_ref] [bib_ref] High-resolution genomic profiling of adult and pediatric core-binding factor acute myeloid leukemia..., Kuhn [/bib_ref]. LSAMP has previously been reported to be a candidate tumor suppressor gene in clear cell renal cell carcinoma and epithelial ovarian cancer [bib_ref] The IgLON family in epithelial ovarian cancer: expression profiles and clinicopathologic correlates, Ntougkos [/bib_ref] [bib_ref] LSAMP and norei: novel genes in hereditary and sporadic clear cell renal..., Kanayama [/bib_ref] [bib_ref] The t(1;3) breakpointspanning involved in clear cell renal cell genes LSAMP and..., Chen [/bib_ref] , and subsequently also in osteosarcomas [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref] [bib_ref] Expression analysis of genes associated with human osteosarcoma tumors shows correlation of..., Sadikovic [/bib_ref] [bib_ref] Recurrent focal copy-number changes and loss of heterozygosity implicate two noncoding RNAs..., Pasic [/bib_ref] [bib_ref] Identification of chromosomal aberrations associated with disease progression and a novel 3q13.31..., Yen [/bib_ref]. All three of these genes have also been proposed to act in conjunction as tumor suppressors in osteosarcomas [bib_ref] Recurrent focal copy-number changes and loss of heterozygosity implicate two noncoding RNAs..., Pasic [/bib_ref].
In this study, the frequency and extent of the deletion and the aberrations of LSAMP were further investigated. To study the potential importance of LSAMP in osteosarcoma biology, we have examined the functional implications of LSAMP-reexpression in an osteosarcoma cell line with a homozygous deletion of the gene.
# Results
The deletion in 3q13. [bib_ref] Notch signaling contributes to the pathogenesis of human osteosarcomas, Engin [/bib_ref]
## Targets lsamp
In order to precisely define the deletion in 3q13.31, highresolution DNA copy number data obtained using Affymetrix Genome-Wide Human SNP Array 6.0 on a total of 76 osteosarcoma samples (32 clinical samples, 25 xenograft samples and 19 cell lines) [bib_ref] Integrative analysis reveals relationships of genetic and epigenetic alterations in osteosarcoma, Kresse [/bib_ref] [bib_ref] AM: Identification of osteosarcoma driver genes by integrative analysis of copy number..., Kuijjer [/bib_ref] and were investigated. We determined the minimal recurrent deletion to be from chr3:116,560,000-116,577,000 and present in 59% (45/76) of the samples , with a similar distribution across the different sample types (56% of the clinical samples , 64% of the xenograft samples and 58% of the cell lines . No differences were observed among the different osteosarcoma subtypes investigated, although the majority of the samples were of osteoblastic subtype (subtype information in Additional file 1: . The high frequency suggests that loss of 3q13.31 is important for development of osteosarcoma, and that the region may harbor tumor suppressor gene(s).
The number of annotated genes within the deleted region differs between the two databases ENSEMBL and RefSeq, with the newly published ENCODE/GENCODE data (version 17) supporting the ENSEMBL annotation . Within the deleted region (chr3:116,000,000-117,500,000), there are two genes commonly annotated by both databases; LSAMP and LSAMP-AS3 (or TUSC7) . To investigate whether loss of other genes besides LSAMP could be important, we performed gene Chromosome map and frequency plot of the observed deletions in 3q13.31 in osteosarcoma samples (n = 76). The shaded gray area (chr3:116,269,000-116,896,000) corresponds to a frequency of ≥ 30% of the samples. Below the frequency plot are the genes within 3q13.31 annotated by the databases ENSEMBL (red) and RefSeq (blue), and supporting annotation by ENCODE/GENCODE data (green). expression analysis of LSAMP RNA antisense 1 (LSAMP-AS1), LSAMP-AS3 and LSAMP-AS4 in a panel of 5 osteosarcoma clinical samples, 13 xenograft samples, 19 cell lines and 14 control samples (n = 51). Expression of LSAMP-AS1 was detected in 30/51 samples, with a similar level between the osteosarcoma samples and the control samples (Additional file 2: . Furthermore, since LSAMP-AS1 is located in the flanking region of the deletion , it was excluded from further experiments. Low expression of LSAMP-AS3 was detected in 5/46 samples, independent of the DNA copy number status, but not in any of the control samples (Additional file 2: . Expression of LSAMP-AS4 was not detected in any of the samples, cancer nor control (0/42) (Additional file 2: . In comparison, expression of LSAMP was detected, although in variable amounts, in 43/49 of these samples, including all the control samples [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref] and . These results indicate that the deletion in 3q13.31 is not inactivating any of the genes LSAMP-AS1, LSAMP-AS3 or LSAMP-AS4, but rather that the expression level, or lack thereof, is a normal state for both non-cancerous and cancerous cells. Thus, LSAMP is most likely the target gene for the deletion.
## Aberrations of lsamp
Aberrations of LSAMP were investigated at the copy number, expression and protein level [fig_ref] Figure 2: LSAMP aberrations and patient survival [/fig_ref] The correlation between LSAMP copy number, expression level and protein level in osteosarcoma samples (n = 42). The samples are sorted according to their copy number (reduced, normal or gain). The relative expression level was measured by qRT-PCR using two probes to cover the length of the gene (located in exon junction 1-2 and 6-7). No detectable expression level is indicated with an asterisk (*). The corresponding LSAMP protein level was determined using western blot, with β-actin as loading control. B: Kaplan-Meier plot showing overall survival for patients with loss (n = 6) and normal/gain (n = 12) of LSAMP copy number. C: Kaplan-Meier plot showing overall survival for patients with low (n = 11) and high (n = 7) expression of LSAMP, compared to the average expression of two normal bone samples. of all the cell lines, except CAL 72 and G-292, have been determined previously using aCGH and qRT-PCR, respectively [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref]. The copy number and expression levels of the remaining samples were determined using TaqMan DNA Copy Number Assay and qRT-PCR, respectively. In total, 16/42 (38%) of the samples had reduced copy number, 16/42 (38%) had normal copy number and 10/ 42 (24%) had increased copy number. There were no differences between sample types or osteosarcoma subtypes. Of the 16 samples with loss of copy number, 11 samples had no or lower expression of LSAMP compared to the average expression level of two normal bone samples, detected by at least one of the probes [fig_ref] Figure 2: LSAMP aberrations and patient survival [/fig_ref]. Of the 16 samples with normal copy number, 9 samples had lower expression of LSAMP detected by at least one of the probes. Of the 10 samples with increased copy number, 5 samples had lower expression of LSAMP detected by at least one of the probes, indicating that copy number aberrations might not be the only mechanism regulating the expression level of LSAMP. In total, 25/42 (60%) samples showed reduced expression level compared to the normal bone samples.
The protein level was investigated by western blotting [fig_ref] Figure 2: LSAMP aberrations and patient survival [/fig_ref]. Of the samples with loss of copy number, 2/ 16 had detectable levels of the LSAMP protein, shown by a band of approximately 62 kDa, corresponding to the size reported by others (60-68 kDa) [bib_ref] Initial biochemical-characterization of limbic system associated membrane-protein (lamp), Levitt [/bib_ref] [bib_ref] Proteomic analysis of membrane microdomain-associated proteins in the dorsolateral prefrontal cortex in..., Behan [/bib_ref]. Of the samples with normal copy number, 3/16 had detectable levels of the LSAMP protein, whereas of the samples with gain of copy number, the protein was detected in 4/ 10 samples. In total, the protein was detected in one clinical sample and eight cell lines. There was no clear correlation between mRNA level and protein level, as some samples with relatively high mRNA level had undetectable protein levels.
We have previously shown an association between low expression of LSAMP and poor survival [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref]. Of the samples investigated in this study (n = 18, of which 10 xenografts and 8 clinical samples), although not statistically significant (p = 0.083, Mantel-Cox test), there was a trend towards poorer survival in patients with loss of LSAMP copy number [fig_ref] Figure 2: LSAMP aberrations and patient survival [/fig_ref]. There was no association between the expression of LSAMP and overall survival [fig_ref] Figure 2: LSAMP aberrations and patient survival [/fig_ref] (p = 0.486).
## Restoring the expression of lsamp
The cell line IOR/OS14 was chosen to ectopically reexpress LSAMP as it has a homozygous deletion of the gene [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref]. In total, 23 clones transfected with LSAMP ORF were assayed for levels of ectopic LSAMP mRNA and protein, and compared to two control clones (backbone vector). All 23 clones had detectable levels of the LSAMP mRNA, but only 12 showed detectable though variable protein levels [fig_ref] Figure 3: LSAMP reexpression [/fig_ref]. The clones were categorized to whether they had undetectable, low, medium or high levels of the LSAMP protein [fig_ref] Figure 3: LSAMP reexpression [/fig_ref]. Low protein levels were most comparable to the endogenous protein levels found in osteosarcoma samples investigated [fig_ref] Figure 2: LSAMP aberrations and patient survival [/fig_ref] and the non-cancerous mesenchymal cell line HEPM [fig_ref] Figure 3: LSAMP reexpression [/fig_ref]. By immunofluorescence confocal microscopy, the ectopically expressed LSAMP protein was shown to be localized to the cell membrane [fig_ref] Figure 3: LSAMP reexpression [/fig_ref] , which is consistent with LSAMP being a membrane protein [bib_ref] A unique membrane-protein is expressed on early developing limbic system axons and..., Horton [/bib_ref].
## Ectopic expression of lsamp delays tumor formation in vivo
The clones with low levels of the LSAMP protein (LSAMP #7, #9 and #21) were chosen for functional characterization and compared to cells without LSAMP-expression (Ctrl #1, #2 and non-transfected cells). The proliferation rate of the clones with the LSAMP protein was significantly reduced 15-20% compared to the cells without LSAMP-expression [fig_ref] Figure 4: Proliferation rate and in vivo tumor growth [/fig_ref] (p = 0.004, Mann-Whitney test).
With exception of non-transfected cells, the same clones were used for investigating the in vivo tumorforming ability of the cells. Each clone was injected into both flanks of six mice, giving a total of 12 potential tumor sites, with the exception of LSAMP #9, which was injected into three mice and thus had a total of six potential tumor sites. Time until tumor appearance is shown by a Kaplan-Meier plot in [fig_ref] Figure 4: Proliferation rate and in vivo tumor growth [/fig_ref] , showing that cells with the LSAMP protein have a significant delayed tumor formation compared to the clones without LSAMP-expression (p = 0.002, Mantel-Cox test).
In addition, other cancer phenotypes were investigated. Neither cell cycle distribution (Additional file 3: [fig_ref] Figure 2: LSAMP aberrations and patient survival [/fig_ref] nor apoptosis (Additional file 4: [fig_ref] Figure 3: LSAMP reexpression [/fig_ref] identified by flow cytometry, or the migration rate using time-lapse photography in the IncuCyte (Additional file 5: [fig_ref] Figure 4: Proliferation rate and in vivo tumor growth [/fig_ref] , were shown to be affected by the expression of LSAMP. As the parental cell line IOR/OS14 has been shown to successfully differentiate towards the adipogenic and osteogenic lineage [bib_ref] Functional characterization of osteosarcoma cell lines provides representative models to study the..., Mohseny [/bib_ref] , differences in differentiation capabilities were investigated. The degree of differentiation was not affected by the expression of LSAMP (data not shown).
Ectopic expression of LSAMP upregulates HES1, CTAG2 and KLF10
To identify possible mechanisms involved in tumor suppression by LSAMP, changes in gene expression in response to LSAMP reexpression were investigated by mRNA expression profiling. Seven clones with different levels of the LSAMP protein (LSAMP #7, #9 #11, #16, #17, #18 and #21) were compared to the two control clones (Ctrl #1 and #2). The analysis revealed that compared to the average expression of the two control clones, three genes, in addition to LSAMP, were differentially expressed in all seven clones [fig_ref] Figure 5: Differential gene expression [/fig_ref]. These genes were hairy and enhancer of split 1 (HES1), cancer/testis antigen 2 (CTAG2) and kruppel-like factor 10 (KLF10), which were all overexpressed compared to the controls, and the upregulation was validated by qRT-PCR [fig_ref] Figure 5: Differential gene expression [/fig_ref]. In addition, one clone (LSAMP #1) expressing LSAMP mRNA, but without detectable levels of the protein [fig_ref] Figure 2: LSAMP aberrations and patient survival [/fig_ref] , was included. Interestingly, this clone had similar levels to the two control clones, supporting that the presence of the LSAMP protein is the cause for the induction of these genes.
Of the 42 samples tested for LSAMP protein [fig_ref] Figure 2: LSAMP aberrations and patient survival [/fig_ref] , 30 had available mRNA expression profiling data [bib_ref] Integrative analysis reveals relationships of genetic and epigenetic alterations in osteosarcoma, Kresse [/bib_ref] and . The expression levels of the genes LSAMP, HES1, CTAG2 and KLF10 were compared between samples with detectable (n = 7) and undetectable (n = 23) levels of the LSAMP protein. For LSAMP, the expression was higher in the samples with LSAMP protein (Additional file 6: [fig_ref] Figure 5: Differential gene expression [/fig_ref] A, p = 0.026, Mann-Whitney test). For HES1, CTAG2 and KLF10, no differences in the median expression levels between the two groups were observed (Additional file 6: [fig_ref] Figure 5: Differential gene expression [/fig_ref].
# Discussion
Osteosarcomas are cytogenetically complex malignancies, with a vast number of DNA copy number aberrations. A recurrent deletion within 3q13.31 has been The expression level was measured using qRT-PCR, using two probes (located in exon junction 1-2 and 6-7). The corresponding LSAMP protein level was determined using western blot, with β-actin as loading control. The clones are sorted according to whether they have undetectable, low, medium or high protein levels. B: Comparison of the endogenous levels of the LSAMP protein in the non-cancerous cell line HEPM and seven LSAMP-expressing clones with increasing amount of protein. C: Subcellular location of the ectopically expressed LSAMP protein in LSAMP-expressing cells (LSAMP #16) and control cells (Ctrl #1) determined using immunofluorescence confocal microscopy. Red color represents stain for anti-LSAMP-antibody, blue represents staining of the nuclei (DAPI).
identified [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref] [bib_ref] Recurrent focal copy-number changes and loss of heterozygosity implicate two noncoding RNAs..., Pasic [/bib_ref] [bib_ref] Identification of chromosomal aberrations associated with disease progression and a novel 3q13.31..., Yen [/bib_ref] [bib_ref] Genomic alterations and allelic imbalances are strong prognostic predictors in osteosarcoma, Smida [/bib_ref]. This deletion has so far only been described in one other cancer type [bib_ref] High-resolution genomic profiling of adult and pediatric core-binding factor acute myeloid leukemia..., Kuhn [/bib_ref] , indicating that it is highly specific to osteosarcomas. This specificity argues that the deletion is not due to a fragile site and strongly suggests involvement of this region in osteosarcoma development. Thus, the deleted region is likely to contain sequences preventing or retarding osteosarcoma development or progression.
The simplest interpretation would be that the region harbors one or more tumor suppressor genes. Concordantly, three genes located within the deleted region have been proposed to have tumor suppressive functions in osteosarcomas: LSAMP [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref] [bib_ref] Expression analysis of genes associated with human osteosarcoma tumors shows correlation of..., Sadikovic [/bib_ref] [bib_ref] Recurrent focal copy-number changes and loss of heterozygosity implicate two noncoding RNAs..., Pasic [/bib_ref] [bib_ref] Identification of chromosomal aberrations associated with disease progression and a novel 3q13.31..., Yen [/bib_ref] , LSAMP-AS3 [bib_ref] Recurrent focal copy-number changes and loss of heterozygosity implicate two noncoding RNAs..., Pasic [/bib_ref] and LSAMP-AS4 [bib_ref] Recurrent focal copy-number changes and loss of heterozygosity implicate two noncoding RNAs..., Pasic [/bib_ref]. However, our results showed either no or low expression levels of LSAMP-AS3 and LSAMP-AS4 in all the samples investigated, including the control samples. The control samples originated from osteoblastic cultures, bone, brain and smooth muscle. In addition, we included the Universal Human Reference RNA, which is a pool of RNA from ten different human cell lines representing various cancer types, excluding osteosarcoma. As neither LSAMP-AS3 nor LSAMP-AS4 were expressed in any of these control samples, it is reasonable to argue that the low or lack of expression in the osteosarcoma samples is not due to inactivation, but a normal state for both noncancerous and cancerous cells. The other genes located within the deleted region (only annotated by ENSEMBL, supported by GENCODE/ENCODE) have been annotated only recently, and was thus not investigated in this study. However, stranded total RNA-sequencing data of 10 osteosarcoma cells lines and the non-tumorigenic cell line iMSC#3 (both undifferentiated and with osteogenic differentiation), showed no detectable expression corresponding to any of these genes (Meza-Zepeda et al., unpublished). Even though we cannot exclude that other genes within the deleted region are important for the development of osteosarcomas, several lines of evidence indicate that LSAMP is a tumor suppressor gene and the target for the deletion in 3q13.31. DNA methylation of the promoter region has been described, indicating that LSAMP is epigenetically inactivated in cancer [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref] [bib_ref] The t(1;3) breakpointspanning involved in clear cell renal cell genes LSAMP and..., Chen [/bib_ref] , and absent or low expression of LSAMP has been reported to be a frequent characteristic of osteosarcomas [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref] [bib_ref] Expression analysis of genes associated with human osteosarcoma tumors shows correlation of..., Sadikovic [/bib_ref] [bib_ref] Recurrent focal copy-number changes and loss of heterozygosity implicate two noncoding RNAs..., Pasic [/bib_ref] [bib_ref] Identification of chromosomal aberrations associated with disease progression and a novel 3q13.31..., Yen [/bib_ref]. We have previously shown that expression of LSAMP is associated with poor survival in a larger panel of osteosarcoma patients [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref] , and it has also been shown in patients with epithelial ovarian cancer [bib_ref] The IgLON family in epithelial ovarian cancer: expression profiles and clinicopathologic correlates, Ntougkos [/bib_ref]. In this study, we saw a trend towards poorer survival in patients with loss of LSAMP copy number, although not statistically significant, possibly due to the limited sample size (n = 18). Overexpression of LSAMP was shown to inhibit proliferation in a renal cell carcinoma cell line [bib_ref] The t(1;3) breakpointspanning involved in clear cell renal cell genes LSAMP and..., Chen [/bib_ref] , whereas depletion of LSAMP promoted cell proliferation in osteoblasts [bib_ref] Recurrent focal copy-number changes and loss of heterozygosity implicate two noncoding RNAs..., Pasic [/bib_ref]. These observations are consistent with the function of LSAMP to be a tumor suppressor gene. Our results add to these evidences as we showed reduced proliferation rate in vitro and inhibited tumor growth in vivo when the expression of LSAMP was restored.
By western blotting, we investigated whether the LSAMP transcript was translated, which to our knowledge has not been done in tumor samples. However, we did not find a clear correlation between the mRNA level of LSAMP and the protein level in our samples. Interestingly, one study has shown that in osteosarcomas with high mRNA levels of LSAMP, a premature termination codon was detected [bib_ref] Identification of chromosomal aberrations associated with disease progression and a novel 3q13.31..., Yen [/bib_ref]. It is tempting to speculate that this is the cause for the apparent lack of translation in some of the samples. However, the premature termination codon was not found in any of 19 osteosarcoma cell lines based on available RNA-sequencing data (Meza-Zepeda et al., unpublished), indicating that this is not a frequent event of regulation in osteosarcomas. Taking into account that 11/23 clones transfected with LSAMP ORF failed to translate the transcripts to detectable levels, there could also be post-transcriptional regulation of LSAMP, even though these transcripts are lacking features of the endogenous mRNA. Also, as only the clones with detectable levels of the LSAMP protein had upregulated HES1, CTAG2 and KLF10, our results indicate that the LSAMP protein is indirectly affecting the upregulation of these genes. However, it is also possible that the cells upregulated one or more of these genes to be able to grow in the presence of the LSAMP protein. On the other hand, there were no differential expression of these genes between the investigated samples with (n = 7) and without (n = 23) detectable levels of the LSAMP protein, based on mRNA expression profiling data [bib_ref] Integrative analysis reveals relationships of genetic and epigenetic alterations in osteosarcoma, Kresse [/bib_ref] , . This discrepancy could be explained by the relatively small size of the cohort with LSAMP protein. However, LSAMP is one of four IgLONs [bib_ref] The limbic system-associated membrane-protein is an Ig superfamily member that mediates selective..., Pimenta [/bib_ref] , which are cell adhesion molecules that function as dimers, referred to as Diglons [bib_ref] Diglons are heterodimeric proteins composed of IgLON subunits, and diglon-CO inhibits neurite..., Reed [/bib_ref]. LSAMP has been shown to only function as heterodimers with either Opioid-binding cell adhesion molecule (OBCAM) or Neurotrimin (NTM) [bib_ref] Diglons are heterodimeric proteins composed of IgLON subunits, and diglon-CO inhibits neurite..., Reed [/bib_ref]. If the dimerization partner of LSAMP is not present, it would most likely not function properly. OBCAM was not expressed in any of the investigated samples (n = 7 with LSAMP protein, n = 23 without LSAMP protein), but NTM was expressed at different levels. Furthermore, the non-transfected parental cell line IOR/OS14 had among the highest expression levels of NTM and ranking as number 11 out of the 30 samples (data not shown). Also, the majority of the cohort with LSAMP protein had low expression of NTM. Thus, it is possible that we do not find an upregulation of HES1, CTAG2 and/or KLF10 in these samples due to the lack of expression of a LSAMP dimerization partner.
HES1, CTAG2 and KLF10 have all been shown to play a role in cancer biology. HES1 has been suggested to have both oncogenic and tumor suppressive functions [bib_ref] Notch signaling contributes to the pathogenesis of human osteosarcomas, Engin [/bib_ref] [bib_ref] Potential role of Notch1 signaling pathway in laryngeal squamous cell carcinoma cell..., Jiao [/bib_ref] , whereas KLF10 has been suggested to be a tumor suppressor gene [bib_ref] Tissue, cell type, and breast cancer stage-specific expression of a TGF-beta inducible..., Subramaniam [/bib_ref] [bib_ref] Role of TIEG1 in biological processes and disease states, Subramaniam [/bib_ref]. CTAG2 is mainly expressed in the reproductive organs, in addition to a variety of cancers [bib_ref] LAGE-1, a new gene with tumor specificity, Lethe [/bib_ref] [bib_ref] Expression of multiple cancer-testis antigen genes in gastrointestinal and breast carcinomas, Mashino [/bib_ref] [bib_ref] Expression profile of cancer-testis genes in 121 human colorectal cancer tissue and..., Li [/bib_ref]. To our knowledge, the function of CTAG2 is not known, but both HES1 and KLF10 have been shown to be involved in bone biology. HES1 has been shown to bind to bone-specific promoters together with the runt-related transcription factor 2 (RUNX2) and the retinoblastoma protein (RB) [bib_ref] HES1 cooperates with pRb to activate RUNX2-dependent transcription, Lee [/bib_ref]. KLF10 has been shown to have an impact on the proliferation of osteoblasts, osteoclasts and osteosarcoma cells [bib_ref] Functional role of KLF10 in multiple disease processes, Subramaniam [/bib_ref] [bib_ref] TGF-beta inducible early gene 1 regulates osteoclast differentiation and survival by mediating..., Cicek [/bib_ref] , which could be the cause for the observed reduced proliferation rate. In a recent study, LSAMP clustered together with pro-apoptotic genes when gene expression changes of osteosarcoma cell lines were analyzed after induction of apoptosis [bib_ref] Combinatorial treatment of DNA and chromatin-modifying drugs cause cell death in human..., Thayanithy [/bib_ref]. Furthermore, depletion of LSAMP in osteoblasts had an effect on the expression of the proapoptotic genes BCL2 and BimEL [bib_ref] Recurrent focal copy-number changes and loss of heterozygosity implicate two noncoding RNAs..., Pasic [/bib_ref]. Although apoptosis could be a possible mechanism for LSAMP to suppress or delay tumors formation, we did not find evidence of apoptosis in our clones with LSAMPreexpression. This is consistent with another study where LSAMP was overexpressed in a clear cell renal cell carcinoma cell line without any evidence of apoptotic cells [bib_ref] The t(1;3) breakpointspanning involved in clear cell renal cell genes LSAMP and..., Chen [/bib_ref].
The high frequency of the deletion in 3q13.31 and the specificity for osteosarcomas strongly suggest a functional role for this region and that it harbors a tumor suppressor gene important for the development of osteosarcomas.
Our results indicate that among the genes investigated in 3q13.31, LSAMP is the target for the deletion. The function of HES1 and KLF10 in bone biology also implies a function for these genes in osteosarcoma development.
Furthermore, the reduced proliferation rate in vitro and inhibited tumor growth in vivo is further pointing to a tumor suppressor function of LSAMP.
# Conclusions
We have identified a frequent deletion in osteosarcomas and shown LSAMP to be the target gene within the deletion. We believe that LSAMP is a tumor suppressor gene in osteosarcomas and that LSAMP suppress tumors by reducing the proliferation rate of cancer cells, possibly through upregulation of one or more of the genes HES1, CTAG2 and KLF10.
# Materials and methods
## Samples
Clinical data for all osteosarcoma and control samples are listed in Additional file 1: .
Human osteosarcoma clinical samples (n = 39) were analyzed, of which 13 were primary or metastatic samples collected at the Norwegian Radium Hospital, Oslo, Norway and 26 were grown subcutaneously in immunodeficient mice as xenografts (suffix x), obtained either from the Norwegian Radium Hospital (n = 15) or the Department of Pathology, University of Valencia, Spain (n = 11). All tumors were diagnosed according to the current World Health Organization classification. The informed consent used and the collection of samples were approved by the Ethical Committee of Southern Norway, Project S-06133 or the Institutional Ethical Committee of the University of Valencia. The samples were collected immediately after surgery, snap frozen in liquid nitrogen and stored at −80°C. The establishment and passing of xenografts were in accordance with national and institutional animal care guidelines.
In addition, osteosarcoma cell lines (n = 21) were analyzed. These were: The cells were grown in RPMI 1640 or DMEM (both Lonza) supplemented with 10% FBS (Fisher Scientific) and GlutaMAX (Life Technologies), at 37°C in a humidified atmosphere with 5% CO 2 . All cell lines were tested for mycoplasma and found negative, and authenticated as previously described [bib_ref] Modulation of the osteosarcoma expression phenotype by microRNAs, Namlos [/bib_ref].
Control samples (n = 15) were also analyzed. Normal long bone samples were purchased from Capital Biosciences (n = 2) or obtained from amputations of cancer patients at the Norwegian Radium Hospital (n = 4) or University College London (n = 1). The normal bone was collected as distant as possible from the tumor site, and SNP array confirmed normal DNA copy number. Primary osteoblast cultures (n = 2) isolated from human calvaria of different donors were purchased from Scien-Cell Research Laboratories. The non-tumorigenic cell lines (n = 3) HEPM, hFOB (both ATCC) and iMSC#3 were included; the latter being an immortalized human bone marrow-derived mesenchymal stromal cell line established in our laboratory . Commercial normal RNA samples (n = 3) were used, one from brain tissue (Life Technologies), one from smooth muscle (Clonetech Laboratories) and Universal Human Reference RNA (Agilent Technologies).
## Dna copy number and expression level
DNA copy number was determined either by highresolution aCGH or the real-time PCR based assay Taq-Man DNA Copy Number Assay (Life Technologies), as previously described [bib_ref] Integrative analysis reveals relationships of genetic and epigenetic alterations in osteosarcoma, Kresse [/bib_ref]. aCGH was performed using the Affymetrix Genome-Wide Human SNP Array 6.0 (Affymetrix) and DNA copy number analysis was performed using the Nexus software (BioDiscovery), as previously described [bib_ref] Integrative analysis reveals relationships of genetic and epigenetic alterations in osteosarcoma, Kresse [/bib_ref]. Expression level was determined using TaqMan Gene Expression Assays (Life Technologies) as previously described [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref] , and is hereafter referred to as qRT-PCR (quantitative real-time reverse transcription PCR). The expression level of the clinical samples and cell lines was analyzed as previously described [bib_ref] LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by..., Kresse [/bib_ref] , and the expression level of the clones was analyzed using the 2 -ΔΔCt -method [bib_ref] Analysis of relative gene expression data using real-time quantitative PCR and the..., Livak [/bib_ref] , with TATA box binding protein (TBP) as an endogenous reference. The assays used and their respective ID number are listed in Additional file 7: .
## Western blotting
Total protein lysate was run on a 4-12% Bis-Tris NuPAGE precast gel (Life Technologies) and transferred onto a PVDF membrane (Millipore). The antibodies used and their respective conditions are listed in Additional file 8: . The proteins were visualized using SuperSignal West Duration Substrate (Thermo Scientific).
## Vector construction and transfection
The LSAMP expression vector was constructed using Gateway Technology (Life Technologies), recombining a vector containing LSAMP open reading frame (ORF) (ID number OCAAo5051A0349D, imaGenes) with pT-REx-DEST30 (Life Technologies). The cell line IOR/OS14 was stably transfected with either the expression vector (named LSAMP ORF) or the backbone vector using Lipofectamine 2000 (Life Technologies). Selection was performed using 450 μg/ml Geneticin (Life Technologies) for 14 days, after which the concentration was reduced to 225 μg/ml.
## Immunofluorescence confocal microscopy
The anti-LSAMP antibody was a kind gift from Dr. Aurea F. Pimenta, Vanderbilt University, Nashville, USA. The cells were grown on coverslips and fixed in 10% formalin solution (Sigma-Aldrich), rinsed in PBS and blocked in 5% FBS in PBS for 30 min before incubation with the antibody in 1:100 dilution in 5% FBS for 1 h at RT. After incubation, the cells were washed 3 × 5 min in 5% FBS and incubated in a 1:200 dilution of anti-mouse-IgG/Cy3 (Jackson ImmunoResearch Laboratories) for 30 min at RT. The cells were then washed 3 × 5 min in 5% FBS and rinsed in dH 2 O. The nuclei were stained using ProLong Gold Antifade Reagent with DAPI (Life Technologies). The fluorescence was visualized using a Zeiss LSM 510 confocal microscope (Zeiss) and pictures were taken of thin single plane sections.
## Proliferation rate
Proliferation rate was measured using the CellTiter 96 AQueous One Solution Cell Proliferation Assay (MTS) (Promega). Cells stably transfected with LSAMP ORF or the backbone vector were seeded in quadruplicates in a 96-well plate with 10,000 cells per well in 100 μl medium. The cell viability was measured after 96 h.
## In vivo tumorigenicity
Animal experiments were performed according to protocols approved by the National Animal Research Authority in compliance with the European Convention of the Protection of Vertebrates Used for Scientific Purposes (approval ID 1499 and 3275, http://www.fdu.no/). The experiments were performed as previously described [bib_ref] Functional characterisation of osteosarcoma cell lines and identification of mRNAs and miRNAs..., Lauvrak [/bib_ref].
## Apoptosis and cell cycle distribution
Apoptosis and cell cycle distribution were investigated using flow cytometry. For apoptosis, APO-BRDU (TUNEL) assay (Life Technologies) was performed, according to the manufacturer's instructions. For investigation of cell cycle distribution, 2*10 6 cells were harvested and resuspended in 200 μl ice-cold PBS and added to 4 ml ice-cold ethanol and incubated on ice for 45 min. Then, 6 ml of ice-cold staining buffer (SB: 0.5% BSA in PBS) was added, and the cells were centrifuged at 300 × g, at 4°C for 5 min. The pellet was resuspended in 1 ml SB and the centrifugation repeated. The cells were resuspended in 300 μl SB containing 2 μg/ml Hoechst 33342 (Sigma-Aldrich). For both assays, the LSR II UV Laser (BD Bioscience) was used, and the data was analyzed using FlowJo v8.8.7 software (Tree Star).
## Migration rate
The migration assay was performed using the IncuCyte system (Essen Bioscience).
## Differentiation
Adipogenic differentiation and Oil red O staining were performed as previously described [bib_ref] Adipocyte differentiation of human bone marrow-derived stromal cells is modulated by MicroRNA-155,..., Skarn [/bib_ref] , except that cells were seeded at a density of 3,000 cells/cm 2 .
For the osteogenic differentiation, the cells were seeded at a density of 4,500 cells/cm 2 . Osteogenesis was initiated with osteogenic induction medium containing 10 nM Dexamethasone, 3.5 mM β-Glycerolphosphate and 66.7 μM Ascorbic acid 2-phosphate (all from Sigma-Aldrich). The osteogenic induction medium was replaced every third day. To estimate the degree of differentiation, the wells were washed with PBS, fixed in ice-cold 70% ethanol for 1 h at 4°C, washed with ddH 2 O and subsequently stained with 0.4% Alizarin Red S solution (w/v, pH 4.2; Sigma-Aldrich) for 10 min at RT. The staining solution was removed by washing the cells 5 × in ddH 2 O, followed by a 15 min wash in PBS. The cells were dehydrated with 70% ethanol, followed by absolute ethanol and air-dried. This procedure was performed after 0, 14, 21 and 28 days of differentiation, respectively. mRNA expression profiling RNA was isolated using the miRNeasy Mini Kit (QIA-GEN GmbH). The RNA integrity was evaluated using the Agilent 2100 Bioanalyzer and the RNA nano 6000 kit (Agilent Technologies). For each sample, 500 ng of total RNA was used to make biotin-labeled and amplified cRNA with the Illumina TotalPrep Amplification Kit (Life Technologies). cRNA was hybridized to Illumina's HumanHT-12 v4 Expression BeadChip as previously described [bib_ref] Tumor-infiltrating macrophages are associated with metastasis suppression in high-grade osteosarcoma: a rationale..., Buddingh [/bib_ref]. Expression values were annotated using the file HumanHT-12_V4_O_R2_15002873_B.bgx (Illumina). The expression data was quantile normalized [bib_ref] A comparison of normalization methods for high density oligonucleotide array data based..., Bolstad [/bib_ref] in GenomeStudio Gene Expression module v1.9 (Illumina) and log 2 -transformed, and a rank product analysis [bib_ref] Rank products: a simple, yet powerful, new method to detect differentially regulated..., Breitling [/bib_ref] was performed in J-Express [bib_ref] J-express: exploring gene expression data using java, Dysvik [/bib_ref] using a qvalue < 0.05 to identify significant changes of gene expression. The dataset has been deposited in the GEO data repository (www.ncbi.nlm.nih.gov/geo/, accession number GSE52089).
# Statistical analysis
Statistical analyses were performed using SPSS version 20. A p-value < 0.05 was regarded as statistically significant.
## Additional files
Additional file 1: . Clinical data for osteosarcoma samples and control samples. Additional file 5: [fig_ref] Figure 4: Proliferation rate and in vivo tumor growth [/fig_ref]. Migration rate. A representative figure showing the migration rate investigated by time-lapse photography using the IncuCyte. The migration of two clones with low levels of the LSAMP protein (#9 and #21), two control clones (#1 and #2) and non-transfected cells were monitored as relative wound density over time (h). The experiment was performed twice. Error bars represent standard deviations of the technical replicates (n = 6). . Overview TaqMan assays.
Additional file 8: . Overview antibodies used for western blotting.
[fig] Figure 2: LSAMP aberrations and patient survival. A: [/fig]
[fig] Figure 3: LSAMP reexpression. A: LSAMP expression and protein level in clones with ectopic expression of LSAMP. [/fig]
[fig] Figure 5: Differential gene expression: A) Venn diagram showing the number of differentially expressed genes in the clones with low (LSAMP #7, #9 and #21), medium (LSAMP #11 and #17) and high (LSAMP #16 and #18) LSAMP protein level compared to the average of the two control clones (Ctrl #1 and #2). B: Validation of the mRNA expression profiling results for the genes HES1, CTAG2 and KLF10, determined by qRT-PCR. Fold change is compared to the average of the two control clones (Ctrl #1 and #2). [/fig]
[fig] Figure 4: Proliferation rate and in vivo tumor growth. A: Relative proliferation rate of LSAMP-expressing cells (LSAMP #7, #9 and #21) compared to cells without LSAMP-expression (Ctrl #1 and #2 and non-transfected cells). The experiment was performed twice. The midline is the median observation, and the whiskers represent the total spread of the observations. The difference was tested statistically significant with a Mann-Whitney test (p = 0.004). B: In vivo tumorigenicity measured as time until tumor appearance, represented by a Kaplan-Meier plot. The difference between the LSAMP-expressing cells (LSAMP #7, #9 and #21) and the cells without LSAMP-expression (Ctrl #1 and #2) was tested statistically significant by a Mantel-Cox test (p = 0.002). [/fig]
[fig] Additional file 2: Figure S1. Expression level of other genes in 3q13.31. The expression of LSAMP-AS1, LSAMP-AS3 and LSAMP-AS4 was investigated by qRT-PCR. The different expression levels are shown as relative percent to an endogenous reference gene (TBP) within the same sample. UHR: Universal Human Reference RNA, OB: Osteoblast.Additional file 3: Figure S2. Cell cycle distribution. A representative figure showing cell cycle distribution investigated by flow cytometry. A and B: Two clones with low levels of the LSAMP protein (A: #7 and B: #21), C and D: two control clones (C: #1 and D: #2) and E: nontransfected cells were included in the analysis. The experiment was performed twice. Additional file 4: Figure S3. Apoptosis. A representative figure showing apoptosis investigated by flow cytometry. Included in the analysis were A: Negative control cells, B: Positive control cells, C-E: Three clones with low levels of the LSAMP protein (C: #7, D: #9 and E: #21), F and G: two control clones (F: #1 and G: #2) and H: non-transfected cells. The experiment was performed thrice. [/fig]
[fig] Additional file 6, Figure S5: Expression levels of LSAMP, HES1, CTAG2 and KLF10 in samples with and without LSAMP protein. Shown are the expression levels of A: LSAMP, B: HES1, C: CTAG2 and D: KLF10 in samples with detectable (n = 7) and undetectable levels (n = 23) of the LSAMP protein. The expression level of CTAG2 was detected by two probes in the bead array (probe ID ILMN 1787578 and ILMN 1715347), and shown in C is the median expression level of the two probes. The expression level of KLF10 was detected by three probes (probe ID ILMN 1720080, ILMN 1659122 and ILMN 167594), and shown in D is the median expression level of the three probes.Additional file 7: [/fig]
|
Off-target effects of tribendimidine, tribendimidine plus ivermectin, tribendimidine plus oxantel-pamoate, and albendazole plus oxantel-pamoate on the human gut microbiota
A B S T R A C TSoil-transmitted helminths infect 1.5 billion people worldwide. Treatment with anthelminthics is the key intervention but interactions between anthelminthic agents and the gut microbiota have not yet been studied. In this study, the effects of four anthelminthic drugs and combinations (tribendimidine, tribendimidine plus ivermectin, tribendimidine plus oxantel-pamoate, and albendazole plus oxantel-pamoate) on the gut microbiota were assessed. From each hookworm infected adolescent, one stool sample was collected prior to treatment, 24 h post-treatment and 3 weeks post-treatment, and a total of 144 stool samples were analyzed. The gut bacterial composition was analyzed using 16S rRNA gene sequencing. Tribendimidine given alone or together with oxantel-pamoate, and the combination of albendazole and oxantel pamoate were not associated with any major changes in the taxonomic composition of the gut microbiota in this population, at both the short-term posttreatment (24 h) and long-term post-treatment (3 weeks) periods. A high abundance of the bacterial phylum Bacteroidetes was observed following administration of tribendimidine plus ivermectin 24 h after treatment, due predominantly to difference in abundance of the families Prevotellaceae and Candidatus homeothermaceae. This effect is transient and disappears three weeks after treatment. Higher abundance of Bacteroidetes predicts an increase in metabolic pathways involved in the synthesis of B vitamins. This study highlights a strong relationship between tribendimidine and ivermectin administration and the gut microbiota and additional studies assessing the functional aspects as well as potential health-associated outcomes of these interactions are required.
# Introduction
Recent estimates suggest that the worldwide prevalence of soiltransmitted helminths (STHs), including infection with Ascaris lumbricoides, Trichuria trichuris and hookworm, is 1.5 billion people [bib_ref] Rescuing the bottom billion through control of neglected tropical diseases, Hotez [/bib_ref] [bib_ref] Soil-transmitted helminth infections, Jourdan [/bib_ref]. The total estimated burden due to STHs is 3.4 million disability-adjusted life years . The symptoms associated with STH infection are not specific and are usually more severe and debilitating in school-aged children and in the elderly. Children who are chronically infected can display malnutrition and developmental delay while elderly people infected with STHs often display reduced work-related productivity [bib_ref] Soil-transmitted helminth infections: ascariasis, trichuriasis, and hookworm, Bethony [/bib_ref].
Four drugs (albendazole, mebendazole, pyrantel pamoate, levamisole) are on the World Health Organization's list of essential medicines for the treatment of STH infections . The two benzimidazoles (albendazole and mebendazole) are most commonly used drugs in preventive chemotherapy programs [bib_ref] Efficacy of current drugs against soil-transmitted helminth infections: systematic review and meta-analysis, Keiser [/bib_ref] [bib_ref] Schistosomiasis and soil-transmitted helminths among an adult population in a war affected..., Abou-Zeid [/bib_ref]. Ivermectin has a broad spectrum of activity against different parasites ranging from nematodes https://doi.org/10.1016/j.ijpddr.2018.07.001 Received 12 April 2018; Received in revised form 27 June 2018; Accepted 2 July 2018 such as Strongyloides stercoralis or A. lumbricoides to filarial parasites, and arthropods and even affects the Anopheline vectors of malaria [bib_ref] A comparative trial of a single-dose ivermectin versus three days of albendazole..., Marti [/bib_ref] [bib_ref] A randomized, double-blind, multicenter clinical trial on the efficacy of ivermectin against..., Wen [/bib_ref] [bib_ref] Ivermectin and malaria control, Ōmura [/bib_ref]. Oxantel has high activity against T. trichuris [bib_ref] Clinical trial with a new anti-Trichuris drug, trans-1, 4, 5, 6 tetrahydro-2-(3-hydroxystyryl)-l-methyl..., Zaman [/bib_ref] [bib_ref] Global anthelmintic chemotherapy programs: learning from history, Horton [/bib_ref] [bib_ref] Oxantel pamoate-albendazole for Trichuris trichiura infection, Speich [/bib_ref]. Finally, tribendimidine has been shown to have a similar spectrum to that of albendazole, and could be an alternative to the latter in case of emergence of benzimidazole resistance [bib_ref] Advances with the Chinese anthelminthic drug tribendimidine in clinical trials and laboratory..., Xiao [/bib_ref]. There is growing consensus to use these drugs in combination to increase efficacy and decrease the risk of drug resistance.
Recent advancements in high-throughput sequencing technologies enable the characterization of the human microbiome [bib_ref] The human microbiome project: exploring the microbial part of ourselves in a..., Turnbaugh [/bib_ref] [bib_ref] Enterotypes of the human gut microbiome, Arumugam [/bib_ref] , which was not possible at the time those drugs were introduced. Thus, the interactions of these anthelminthic drugs and drug combinations with the human gut microbiota composed of a variety of microorganisms, including other eukaryotic parasites, protozoa, viruses, fungi and most importantly bacteria [bib_ref] Gut microorganisms, mammalian metabolism and personalized health care, Nicholson [/bib_ref] [bib_ref] Diversity, stability and resilience of the human gut microbiota, Lozupone [/bib_ref] [bib_ref] Metagenomic diagnostics for the simultaneous detection of multiple pathogens in human stool..., Schneeberger [/bib_ref] , can now be studied. Drug-microbiota interactions can modulate the bioavailability and activity of drugs and hence modulate their efficacy [bib_ref] Gut microorganisms, mammalian metabolism and personalized health care, Nicholson [/bib_ref] [bib_ref] Pharmacometabonomic identification of a significant host-microbiome metabolic interaction affecting human drug metabolism, Clayton [/bib_ref] [bib_ref] The role of gut microbiota in drug response, Wilson [/bib_ref] [bib_ref] Xenobiotics shape the physiology and gene expression of the active human gut..., Maurice [/bib_ref]. STHs share the same environmental niche as the gut bacterial microbiota, but potential interactions between the gut microbiota and anthelminthic agents have not been assessed to date. The aim of this study was to identify potential interactions between these treatments and the non-target microbiota.
# Methods
## Sample collection and ethics statement
Samples were collected in the framework of a randomized, controlled, single blind, non-inferiority trial in the Agboville district in Côte d'Ivoire. Hookworm positive adolescents (age 15 to 18), confirmed by quadruplicate Kato-Katz thick smears, were randomly assigned to four treatment arms, including tribendimidine (400 mg), tribendimidine (400 mg) plus ivermectin (200 μg/kg), tribendimidine (400 mg) plus oxantel pamoate (25 mg/kg) and albendazole (400 mg) plus oxantel pamoate (25 mg/kg). Details about the study are presented elsewhere [bib_ref] Efficacy and safety of tribendimidine, tribendimidine plus ivermectin, tribendimidine plus oxantel pamoate,..., Moser [/bib_ref]. From each adolescent, one stool sample was collected prior to treatment, 24 h post-treatment and 3 weeks post-treatment, and a total of 144 stool samples were analyzed [fig_ref] Table 1: Summary of volunteers investigated in this study [/fig_ref].
Ethical approval was obtained from the Comité National d'Ethique et de la Recherche in Côte d'Ivoire (083/MSHP/CNER-kp) and the Ethics Committee of North-western and Central Switzerland (EKNZ UBE-15/35). The trial was registered with ISRCTN (number 14373201).
## Sample collection
For DNA isolation, 150-250 mg of stool sample was diluted in 500 μl of AVL buffer (Qiagen, Darmstadt, Germany) and subsequently homogenized using a soil-grinding SK38 kit on the Precellys 24 system (Bertin Technologies, Saint-Quentin, France). Homogenized samples were further extracted on a Magna Pure 96 system (Roche, Basel, Switzerland) using the DNA and Viral NA Large Volume kit (Roche, Basel, Switzerland) according to the manufacturers' protocol.
## 16s amplicon pcr
2.5 μl of isolated DNA was used to perform amplification of the V3-V4 region using the following primer pair:
Forward primer = 5′-TCGTCGGCAGCGTCAGATGTGTATAAGAGA-CAGCCTACGGGNGGCWGCAG Reverse primer = 5′-GTCTCGTGGGCTCGGAGATGTGTATAAGAGA-CAGGACTACHVGGGTATCTAATCC The polymerase chain reaction (PCR) was performed in 25 μl reaction volumes using the 2X KAPA HiFi HotStar ReadyMix (KAPA Biosystems; Boston, MA, USA). The thermocycler was set as follows: 95°C for 3 min, 25 cycles of 95°C (30 s), 55°C (30 s) and 72°C (30 s), one additional step at 72°C for 5 min and finally set on hold indefinitely at 4°C. The quality of the amplified product was controlled visually on a 1% agarose gel. The amplicons were purified using an AMPure XP beads (Beckman-Coulter; Fullerton, CA, USA) protocol.
## Sequencing
We used a Nextera XT Index kit (Illumina, San Diego, CA, United States of America) to perform the index PCR and generate barcoded pools of 96 samples. The amplification reaction was conducted on a thermocycler under the following condition: 95°C for 3 min, 20 cycles of 95°C for 30 s, 55°C for 30 s, 72°C for 30 s, one additional step at 72°C for 5 min and a final step at 4°C until further processing. Amplicons were cleaned with AMPure XP beads according to Illuminas' library preparation guide. The quality of the product was assessed using a 1% agarose gel and the quantification was performed using a Qubit Fluorimeter (Life Technologies, Carlsbad, CA, United States of America) and the corresponding High-sensitivity DNA assays (Life Technologies, USA). The 96 amplicon samples were pooled together in an equimolar way and loaded on a cartridge on the Illumina Miseq sequencing platform (Illumina, USA). We used V3 reagents (2*300bp) (Illumina, USA) for this experiment in 2*300 base pair mode.
## Data processing and statistical analysis
Raw Fastq files were filtered using a quality score (Phred) above Q20. Filtered reads were fed in the QIIME pipeline [bib_ref] QIIME allows analysis of high-throughput community sequencing data, Caporaso [/bib_ref] v 1.9.1, which was configured for standard OTU picking (closed reference, Greengenes database v.13_8). Shannon diversity index was computed with QIIME using a rarefaction depth of 19500 sequences and principal component analysis (PCA) plots were generated with the STAMP statistical analysis package [bib_ref] STAMP: statistical analysis of taxonomic and functional profiles, Parks [/bib_ref] using Euclidean distance matrices. Taxonomic profiles were further analyzed with the PICRUSt suite , in order to predict the functional content of the gut microbiota based on the abundance of 16S rRNA genes. To compare groups (treatment arms), we performed an analysis of variance using the STAMP software [bib_ref] STAMP: statistical analysis of taxonomic and functional profiles, Parks [/bib_ref] , both for taxonomical and functional profiles. Benjamini-Hochberg correction [bib_ref] Controlling the false discovery rate -a practical and powerful approach to multiple..., Benjamini [/bib_ref] was applied to the uncorrected statistics, to control for multiple testing bias. Statistical significance is reached with a corrected p-value (q-value) below 0.1.
# Results
## Overall comparison of the taxonomic composition between treatment arms
Pre-processing, including filtering and denoising of the sequence datasets resulted in the analysis of a total of 2,311,784 high quality reads for the 48 pre-treatment samples (median = 55,000), 2,261,185 reads for the 48 samples collected 24 h after treatment (median = 56,826), and 2,580,596 filtered reads for the 48 samples collected in the follow-up period of 3 weeks (median = 59,889). We compared the overall microbiota diversity of all samples, grouped by treatment arm, using a principal component analysis and the Shannon diversity index, as shown in [fig_ref] Figure 1: Overall diversity comparison of taxonomic composition at the phylum level, stratified by... [/fig_ref]. Complete taxonomic profiles stratified by treatment arm are shown in [fig_ref] Figure 1: Overall diversity comparison of taxonomic composition at the phylum level, stratified by... [/fig_ref].
Treatment groups were not associated with significant differences in taxonomic composition or Shannon diversity at any time-point. Although two samples (P-3 and P-8) from treatment arm 3 (tribendimidine plus oxantel pamoate) differed significantly from the rest of the population before treatment, there was no observable difference in gut bacterial composition of these two patients both at 24 h post-treatment and 3 weeks post treatment.
At 24 h after drug administration, the principal component analysis showed a homogeneous population and the SDI across all treatment arms was not significantly different. However, there was a noticeable spread (∼0.35-0.2) of samples belonging to treatment arm 2 over component 1 (x-axis), while samples from this group all clustered tightly before treatment (within coordinates −0.1 and ∼0.1 on both axis).
Finally, the observation of a generally homogeneous bacterial composition of the gut microbiota at the phylum level in the studied population was confirmed in the samples collected 3 weeks after administration of the treatments. According to the PCA, the sample collected from patient 39 differed from the rest of the population, at this sampling time. The SDI for this sample was low (SDI = 1.72) compared to the rest of the group.
## Differences in taxonomic composition after administration of the four treatments
In order to assess the potential effect of each drug/drug combination tested in this study on the gut microbiota, we compared abundance means of each bacterial taxon between groups, at different taxonomic levels (phylum and family). At baseline, there was no phylum which was differentially abundant between the four treatment arms. 24 h after treatment, there was a significantly higher abundance (q-value < 0.1) of species from the Bacteroidetes phylum in samples from treatment arm 2. The relative abundance of Bacteroidetes was greater in treatment arm 2 than all other groups 24 h after treatment, but not at baseline nor at 3 weeks.
In order to quantify the degree of variation of Bacteroidetes over time, we compared those from pre-to 24 h post-treatment samples and those from 24 h post-to 3 weeks post-treatment samples, as shown in [fig_ref] Figure 2: Box plot showing the abundance variation of the phylum Bacteroidetes [/fig_ref].
In order to further characterize potential drug-bacteria interactions, we screened lower taxonomic levels and found two families, which follow a similar abundance pattern (A) to that of Bacteroidetes (A TA2 > A TA1 = A TA3 = A TA4) , namely S24_7 (recently renamed to Candidatus homothermaceaeand Prevotellaceae. However, for both families, the significance for inter-group difference was not reached (q-value > 0.1).
## Prediction of metabolic pathways
In order to highlight possible interactions, we compared the predicted abundance of metabolic pathways obtained with PICRUSt between the four treatment arms at each sampling time .
In samples collected at baseline, the abundance of sequences related to biotin metabolism, folate, and N-glycan biosynthesis was the same across all treatment groups. The situation changed 24 h after treatment when these metabolic pathways became significantly more abundant in the tribendimidine-ivermectin treatment arm (q-values = 0.047, 0.016, and 0.038, for biotin metabolism, folate biosynthesis, and N-glycan biosynthesis, respectively). The abundance pattern was the same for each of the three pathways, the highest abundance being associated to treatment with tribendimidine in combination with ivermectin (treatment arm 2) and being lower for the three other groups. The situation after 3 weeks was similar to pre-treatment, and no differentially abundant pathway could be identified.
# Discussion
There is a growing interest in associations between the human microbiome and drug therapy, which has been called "pharmacomicrobiomics" [bib_ref] The Human Microbiome Project, personalized medicine and the birth of pharmacomicrobiomics. Curr, Rizkallah [/bib_ref] [bib_ref] Gut flora metabolism of phosphatidylcholine promotes cardiovascular disease, Wang [/bib_ref] [bib_ref] Vaginal bacteria modify HIV tenofovir microbicide efficacy in African women, Klatt [/bib_ref] [bib_ref] Gut microbiome interactions with drug metabolism, efficacy, and toxicity, Wilson [/bib_ref]. This study is the first investigating effects of different anthelminthic drugs and their combinations on the human gut microbiota of adolescents infected with hookworms in a tropical setting.
Several studies have shown that the effects of various parasitic infections on the gut microbiota are relatively modest [bib_ref] Impact of experimental hookworm infection on the human gut microbiota, Cantacessi [/bib_ref] , yet, there is only scarce information about the impact, direct (= drug effect on bacterial composition) or indirect (= change in bacterial composition is due to clearance of the worm), of drugs used to treat parasitic infections on the gut microbiota. Cross-reaction of drugs designed to target eukaryotic parasites between the gut microbiota and various treatments (e.g. anticancerous drugs) has been shown [bib_ref] Gut microbiota modulation of chemotherapy efficacy and toxicity, Alexander [/bib_ref] [bib_ref] Microbiota: a key orchestrator of cancer therapy, Roy [/bib_ref] , and given the fact STHs and gut bacteria share the same environment these effects are possible for anthelminthic drugs as well.
Our analyses indicate that tribendimidine given alone or together with oxantel pamoate, and the combination of albendazole and oxantel pamoate is not associated with any major changes in the taxonomic composition of the gut microbiota in this population, at both the shortterm post-treatment (24 h) and long-term post-treatment (3 weeks) periods. However, treatment with tribendimidine in combination with ivermectin was found to be associated with shifts in the relative abundance of Bacteroidetes. This phylum increased in relative abundance in the first 24 h post-administration and decreased over the 3 weeks post-treatment. Within the Bacteroidetes phylum, two bacteria families, Prevotellaceae and Candidatus Homeothermaceae (former S24_7, accounted for most of the variation observed in the abundance of Bacteroidetes.
We assessed the potential metabolic significance of this taxonomic shift using PICRUSt prediction, which accurately infers metabolic gene content from taxonomic composition [bib_ref] An evaluation of the accuracy and speed of metagenome analysis tools, Lindgreen [/bib_ref]. Of all . Box plot comparing the abundance of predicted metabolic pathways at the different sampling times. TA1 = Tribendimidine; TA2 = Tribendimidine plus ivermectin; TA3 = Tribendimidine plus oxantel pamoate; TA4 = Albendazole plus oxantel pamoate. Panel A shows the abundance of biotin metabolism, folate biosynthesis and N-glycan biosynthesis pathways before treatment, by treatment arm. Panel B shows the abundance of the three metabolic pathways, 24 h after administration of treatment. Panel C highlights the abundance of the three metabolic pathways, 3 weeks after administration of treatment. Significance is achieved with a q-value < 0.1. q-values are shown only for comparisons which are significantly different.
predicted metabolic pathways, three were significantly more abundant in samples from treatment arm 2, namely, biotin metabolism, folate biosynthesis, and N-glycan biosynthesis, and all are involved at some point in the biosynthesis of B vitamins. While the statistical proof of higher abundance of Prevotellaceae and Candidatus Homeothermaceae in patients from treatment arm 2 (tribendimidine plus ivermectin) is somewhat weaker to that of Bacteroidetes, it has been shown that both groups present characteristics highly relevant to each of the three differentially abundant metabolic pathways. Indeed, and while acknowledging that it does not encompass the complete range of their activities in the gut, members from the Prevotellaceae family almost ubiquitously harbor the genes associated with B vitamin biosynthesis [bib_ref] Systematic genome assessment of B-vitamin biosynthesis suggests co-operation among gut microbes, Magnúsdóttir [/bib_ref]. Similarly, one of the purposes of Candidatus Homeothermaceae in the gut of mammals is to regulate the synthesis of B vitamins. Of note, a possible cause for the lack of statistical power for findings observed in our study for these two taxa is the intrinsically weak taxonomic resolution of 16S rRNA gene identification, which frequently does not allow categorizing sequences at lower taxonomic levels, in this case, the family level. High level of biotin, or vitamin B 7 , in rat livers, has been associated with decreased absorption of orally administered vitamin B 12 [bib_ref] The effect of vitamin B(12) and biotin on the metabolism of vitamin..., Puddu [/bib_ref]. For folate (vitamin B 9 ), decreased concentrations in combinations with low B 12 levels have been associated with various symptoms, including anemia [bib_ref] Folate and vitamin B-12 status in relation to anemia, macrocytosis, and cognitive..., Morris [/bib_ref] [bib_ref] Folate-vitamin B-12 interaction in relation to cognitive impairment, anemia, and biochemical indicators..., Selhub [/bib_ref] , a common consequence of ivermectin administration [bib_ref] Efficacy of ivermectin and albendazole alone and in combination for treatment of..., Ndyomugyenyi [/bib_ref]. However, in order to validate these observations and to understand these interactions in detail, additional studies investigating other parameters (e.g. monitoring of B vitamins levels after treatment) and tests with higher taxonomic resolution (e.g. strain identification with shotgun metagenomics) are required.
This study has limitations. Indeed, while the study design in its current form enables the identification of significant differences in the composition of the gut microbiota between different treatment groups, it does not allow distinguishing between direct or indirect effects. Future studies might be considered including treatment of uninfected patients and patients who received a placebo, although this might be challenging from an ethical point of view. However, it is worth highlighting that the observed effects (higher abundance of Bacteroidetes and specific metabolic pathways) are distributed homogeneously across all samples analyzed from the tribendimidine-ivermectin group, except for one sample measured 24 h post-treatment. Additionally, stable abundance of Bacteroidetes and the metabolic pathways is observed homogeneously across all samples from the other treatment groups, except for two samples. Both facts support the hypothesis of a recompositional event triggered by the administration of tribendimidine plus ivermectin and rule out observed effects driven mainly by random outlying samples.
# Conclusion
In conclusion, our study revealed several key findings of antihelminthic drugs/microbiota interactions. Tribendimidine, tribendimidine plus oxantel pamoate and albendazole plus oxantel-pamoate do not cause microbiome-specific effects and hence are unlikely to cause variability in response. On the other hand, treatment with a combination of tribendimidine and ivermectin triggers a re-compositional event of the gut microbiota, which was not observed for the other treatments studied. Moreover, a higher abundance of Bacteroidetes, and to some extent, of bacteria from the Prevotellaceae and Candidatus Homeothermaceae families as well as an overrepresentation of metabolic pathways related to B vitamins biosynthesis, was observed in patients treated with tribendimidine and ivermectin, 24 h after treatment. Importantly, the observed effects are transient as they disappear within three weeks after administration of this anthelminthic treatment.
# Funding information
JK is grateful to the Swiss National Science Foundation for financial support (number 320030_14930/1).
[fig] Figure 1: Overall diversity comparison of taxonomic composition at the phylum level, stratified by treatment arm. TA1 = Tribendimidine; TA2 = Tribendimidine plus ivermectin; TA3 = Tribendimidine plus oxantel pamoate; TA4 = Albendazole plus oxantel pamoate. Panel A: pre-treatment. Panel B: 24 h post-treatment. Panel C: 3 weeks post-treatment. Panel D shows the Shannon diversity index at each time point. h = hours; w = weeks. [/fig]
[fig] Figure 2: Box plot showing the abundance variation of the phylum Bacteroidetes. This plot summarizes variations of abundance of Bacteroidetes between two sampling times (from pre-treatment to 24 h post-treatment and 24 h to 3 weeks post-treatment), by treatment arm. Treatment arm 1 = Tribendimidine; Treatment arm 2 = Tribendimidine plus ivermectin; Treatment arm 3 = Tribendimidine plus oxantel pamoate; Treatment arm 4 = Albendazole plus oxantel pamoate. [/fig]
[table] Table 1: Summary of volunteers investigated in this study. From each adolescent of each of the four treatments, a sample was collected before, 24 h after and 3 weeks after treatment. EPG = Hookworm eggs per gram of stool, FU = follow-up. [/table]
|
A Phase II Study of Sequential Capecitabine Plus Oxaliplatin Followed by Docetaxel Plus Capecitabine in Patients With Unresectable Gastric Adenocarcinoma
Fluorouracil and platinum are considered the standard treatment options for advanced gastric cancer. Docetaxel is also an effective agent and it shows no cross-resistance with fluorouracil and platinum. The combination treatment of docetaxel with fluorouracil and platinum has been explored, but it demonstrated intolerable toxicities. An alternative approach in the first-line treatment of gastric adenocarcinoma may be to use these agents sequentially. We aimed to evaluate the activity and safety profile of sequential chemotherapy with capecitabine plus oxaliplatin, followed by docetaxel plus capecitabine in the first-line treatment of unresectable gastric cancer.We conducted a phase II study of sequential first-line chemotherapy in advanced gastric cancer. Treatment consisted of 6 cycles of capecitabine plus oxaliplatin (capecitabine 1000 mg/m 2 bid on days 1-10 and oxaliplatin 85 mg/m 2 on day 1, every 2 weeks), followed by 4 cycles of docetaxel plus capecitabine (docetaxel 30 mg/m 2 on days 1 and 8, capecitabine 825 mg/m 2 bid on days 1-14, every 3 weeks). The primary end-point was the objective response rate.Fifty-one patients were enrolled: median age, 63 years; male/ female: 37/14. The main grade 3 to 4 toxicities were a decreased absolute neutrophil count (25.4%), diarrhea (9.8%), and hand-foot syndrome (15.7%). The objective response rate was 61.7%. The median progression-free survival and overall survival were 8.6 and 11.0 months, respectively. Six patients (11.8%) received surgery after chemotherapy and 5 are still disease-free.This sequential treatment demonstrated feasibility with a favorable safety profile and produced encouraging results in terms of activity and efficacy.(Medicine 95(3):e2565)Editor: Patrick Wall. ). Novelty & impact statements: a phase II trial which enrolled 51 patients to clarify the safety and efficacy of XELOX followed by TX in unresectable gastric cancer patients. We found it is effective, well-tolerated, convenient, and practical for advanced gastric cancer patients in daily practice Human rights statement and informed consent: all procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and later versions. Informed consent or substitute for it was obtained from all patients for being included in the study.
# Introduction
G astric cancer (GC) is the fourth most common cancer and the third most common cause of cancer-related death worldwide. [bib_ref] Management of gastric cancer in Asia: resource-stratified guidelines, Shen [/bib_ref] Surgery is the only curative modality. However, many patients are initially diagnosed with locally advanced or metastatic disease. The cancer has a high recurrence rate after surgery, especially for advanced disease. [bib_ref] Update on treatment of gastric cancer, Kuo [/bib_ref] For these patients, systemic chemotherapy has been shown to improve quality of life and survival compared with best supportive care. [bib_ref] Chemotherapy for advanced gastric cancer, Wagner [/bib_ref] One meta-analysis study of 11 gastric cancer trials found that combination chemotherapy resulted in better overall survival (OS) compared with single-agent chemotherapy. [bib_ref] Chemotherapy in advanced gastric cancer: a systematic review and meta-analysis based on..., Wagner [/bib_ref] There is still no well-established standard regimen, but doublet chemotherapy, including fluoropyrimidine (5-FU) and platinum, is widely used in worldwide clinical trials. Docetaxel is another effective agent in the treatment of GC. In the phase III TAX-V325 trial, the addition of docetaxel to 5-FU and cisplatin (DCF regimen) improved response rate, time to progression, and survival compared with 5-FU and cisplatin (CF). [bib_ref] Phase III study of docetaxel and cisplatin plus fluorouracil compared with cisplatin..., Van Cutsem [/bib_ref] However, the DCF regimen is associated with increased toxicity, especially myelosuppression and infection, and it is not widely used in clinical practice. Several modifications of the DCF regimen have been developed to increase the tolerability of the regimen while maintaining the same level of activity. [bib_ref] A phase II study of weekly docetaxel and cisplatin plus oral tegafur/uracil..., Li [/bib_ref] [bib_ref] Docetaxel plus oxaliplatin with or without fluorouracil or capecitabine in metastatic or..., Van Cutsem [/bib_ref] [bib_ref] Randomised, noncomparative phase II study of weekly docetaxel with cisplatin and 5-fluorouracil..., Tebbutt [/bib_ref] [bib_ref] Randomized multicenter phase III study of a modified docetaxel and cisplatin plus..., Wang [/bib_ref] An alternative way to include these active agents in the first-line treatment of advanced GC is to use them sequentially. Sequential schedules may maximize the dose-intensity of each single agent and avoid the overlapping toxicity of concomitant agents. Three studies using sequential strategies to treat advanced GC have been reported. [bib_ref] Sequential chemotherapy with cisplatin, leucovorin, and 5-fluorouracil followed by docetaxel in previously..., Catalano [/bib_ref] [bib_ref] Randomized phase II study of sequential docetaxel and irinotecan with 5-fluorouracil/ folinic..., Gubanski [/bib_ref] [bib_ref] Sequential treatment with epirubicin, oxaliplatin and 5FU (EOF) followed by docetaxel, oxaliplatin..., Petrioli [/bib_ref] All studies showed that sequential therapies produced good treatment efficacy with manageable toxicities.
In the REAL2 study, capecitabine, the oral pro-drug form of 5-FU, has already shown the same efficacy as 5-FU, cisplatin or oxaliplatin. 14 Capecitabine is an oral chemotherapy, which is more convenient than continuous 5-FU infusion. Oxaliplatin is an alkylating agent and is a third-generation platinum. Compared with cisplatin, oxaliplatin has shown more favorable safety profiles. In our previous study, we showed that a modified biweekly capecitabine and oxaliplatin (XELOX) regimen is a practical and effective regimen in the treatment of GC. [bib_ref] A multicenter phase II study of biweekly capecitabine in combination with oxaliplatin..., Chao [/bib_ref] [bib_ref] Modified biweekly oxaliplatin and capecitabine for advanced gastric cancer: a retrospective analysis..., Kuo [/bib_ref] Lo et al and Giordano et al also showed that docetaxel and capecitabine (TX) is a well-tolerated, easily administrated regimen in advanced GC. [bib_ref] Docetaxel and capecitabine in patients with metastatic adenocarcinoma of the stomach and..., Giordano [/bib_ref] [bib_ref] A phase II study of weekly docetaxel in combination with capecitabine in..., Lo [/bib_ref] Therefore, we conducted a phase II study to investigate the efficacy and feasibility of the sequential administration of the XELOX regimen followed by the TX regimen in patients with GC.
# Material and methods
## Study design
This trial was a multicenter, open label, single-arm phase II study evaluating sequential chemotherapy with the XELOX regimen followed by the TX regimen (NCT01558011). The primary end-point was the objective response rate (ORR), and the secondary end-points were OS, progression-free survival (PFS), and assessment of toxicity. The study was approved by the Institutional Review Board of each participating center or the competent authority and the Ethics Committee. The study was conducted in full accordance with the International Conference on Harmonization Good Clinical Practice guidelines and the Declaration of Helsinki. All patients provided written informed consent before entering the study.
## Eligibility
Patients were enrolled from the following medical centers: Taipei Veterans General Hospital, Mackay Memorial Hospital, National Health Research Institutes, National Cheng Kung University Hospital, and Tri-Service General Hospital. Patients with pathologically proven unresectable recurrent or metastatic gastric adenocarcinoma were assessed for eligibility. The major inclusion criteria were as follows: at least 1 measurable disease; age > 20 years; an Eastern Cooperative Oncology GROUP (ECOG) performance status (PS) of 0 to 2; adequate bone marrow function (defined by a leukocyte count of 4000 leukocytes/mL, an absolute neutrophil count of 1500 neutrophils/mL, a platelet count of 100,000 platelets/mL, and a serum hemoglobin level of 9 g/dL); adequate renal function (serum creatinine level at least 1.5-fold lower than the reference value); and adequate hepatic function (bilirubin level at least 2-fold lower than the reference value and aspartate aminotransferase and alanine aminotransferase levels at least 2.5-fold lower than the reference value). Prior radiotherapy was permitted if it was not administered to the target lesions evaluated in this study and if it had been completed at least 2 weeks prior to the patient's enrollment into the study. Patients who had completed adjuvant chemotherapy at least 6 months before recruitment were enrolled. Patients with brain metastasis or those who could not take study medication orally were excluded. Patients whose tumor samples revealed overexpression of the HER-2/neu protein (3þ) by immunohistochemical (IHC) staining were also excluded.
## Treatment schedule
Eligible patients received oral capecitabine (Xeloda 1 ; Roche, Basel, Switzerland) -1000 mg/m 2 twice daily on days 1 to 10 every 2 weeks, plus oxaliplatin (Eloxatin 1 , Sanofi-Aventis, Paris, France) 85 mg/m 2 (2 h IV infusion) on day 1 (XELOX regimen) every 2 weeks for 6 cycles [fig_ref] FIGURE 1: The Consolidated Standards of Reporting Trials [/fig_ref]. Patients were allowed to rest for 1 week after XELOX treatment. Then the treatment was shifted to docetaxel (Taxotere 1 , Sanofi-Aventis, Paris, France) 30 mg/m 2 (a 30-min intravenous infusion) on days 1 and 8, plus oral capecitabine 825 mg/m 2 twice daily on days 1 to 14 (TX regimen) every 3 weeks for 4 cycles. After completing all planned regimens, a further regimen was independently decided by the investigator. Prophylactic dexamethasone was prescribed to prevent any potential hypersensitivity reactions to docetaxel. The standard antiemetic prophylaxis of intravenously administered 5-HT3 antagonists was administered before chemotherapy. Granulocyte colonystimulating factor was administered to treat neutropenic events; however, prophylactic granulocyte colony-stimulating factor and prophylactic antibiotics were not administered to patients who had experienced a neutropenic event in the previous cycle.
## Response and toxicity evaluation
The response to therapy was assessed by the radiological evaluation of any measurable lesion every 8 weeks based on Response Evaluation Criteria In Solid Tumors version 1.1 [bib_ref] New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1), Eisenhauer [/bib_ref] using computed tomography; it was determined by an independent response review committee. After discontinuation of the study treatment, patients were followed up every 3 months until disease progression or death. Toxicity was evaluated and recorded according to version 4.0 of the Common Terminology Criteria for Adverse Events of the National Cancer Institute. All of the patients were included in the toxicity assessment. For the toxicity analysis, the data indicating the worst toxicity for each patient from all of the chemotherapy cycles were used.
# Statistical analysis
According to Simon's optimal 2-stage design, [bib_ref] Optimal two-stage designs for phase II clinical trials, Simon [/bib_ref] [bib_ref] Modified biweekly oxaliplatin and capecitabine for advanced gastric cancer: a retrospective analysis..., Kuo [/bib_ref] patients were required for enrollment to test the null hypothesis that the true ORR is 40% versus the alternative hypothesis that the true ORR is at least 60%, at a significance level of 0.05 with a power of 80%. If 8 or more responses were observed among 16 patients in the first stage, the study was continued with 30 additional patients included. As the dropout rate was assumed 10%, the number of patients necessary for recruitment into the study was calculated to be 51.
The perprotocol (PP) population excluded those patients who received treatment of <1 cycle for reasons other than disease progression or death or those who received <50% of the anticipated treatment during the first 6 weeks of the trial. The response and toxicity data were described using simple descriptive statistics. PFS was calculated from the first day of treatment until the first day of documented disease progression or death from any cause. PFS was censored at the date of the last followup visit for the patients who were still alive and had no documented disease progression. OS was calculated from the first day of treatment until the day of death. PFS and OS were estimated using the Kaplan-Meier method.
# Results
## Patient characteristics
From March 2012 to Sep 2014, 51 patients were enrolled in this study. The patient characteristics were shown in [fig_ref] TABLE 1: Patient Characteristics ECOG ¼ Eastern Cooperative Oncology Group [/fig_ref]. The median age was 63 years (range 32-83 years). Thirty-seven patients (72.5%) were men, and most patients (98%) had ECOG PS of 0 or 1. Twenty-eight patients (54.9%) had poorly differentiated adenocarcinoma. Four patients (7.8%) had received adjuvant chemotherapy. Forty-two patients (82.4%) had metastatic disease and the most common metastatic sites were the distant lymph nodes (62.7%), liver (39.2%), and peritoneum (33.3%). Forty-three patients (84.3%) had low-level HER2 protein expression (IHC staining 0 or 1þ) in their tissue samples.
## Treatment delivery
Out of the 51 patients, 4 patients received only 1 cycle of the XELOX regimen and were excluded from analysis by PP population [fig_ref] TABLE 2: Cycles of Treatment and Follow-Up Time [/fig_ref]. Forty-three patients (84.3%) completed 6 cycles of the XELOX regimen, with a median of 6 cycles (range 1-6 cycles). Among these patients, 42 patients (82.4%) then proceeded to receive the TX regimen. A total of 31 patients (60.8%) completed 4 cycles of the TX regimen; the median was 4 cycles (range 0-4 cycles).
## Efficacy
Of the 51 patients, 47 were eligible for response evaluation. Four patients were not available for response evaluation: 1 was intolerant to capecitabine, 1 had disease progression and could not take oral drugs, and 2 patients were excluded at the investigator's discretion. Tumor responses are summarized in [fig_ref] TABLE 3: Best Response Rate According to Response Evaluation Criteria In Solid Tumors [/fig_ref]. During the XELOX regimen period, 25 patients (53.2%) achieved a partial response, 21 patients (44.7%) had stable disease, and 1 patient (2.7%) had disease progression. During the TX period, 8 patients (19.5%) achieved a partial response, 27 patients (65.9%) had stable disease, and 6 patients (14.6%) had disease progression. Overall, 29 patients (61.7%) achieved a partial response and 18 patients (38.3%) had stable disease. Among 8 patients who had response during the TX period, 5 patients had partial response, 2 patients had stable disease, and 1 patient had progressive disease during the XELOX period. The median PFS was 8.6 months (95% confidence interval [CI] 5.6-13.7 months, [fig_ref] FIGURE 2: Kaplan-Meier curves for progression-free survival of 47 advanced gastric cancer patients [/fig_ref] , and the median OS was 11.0 months (95% CI 9.6-14.5 months, [fig_ref] FIGURE 3: Kaplan-Meier curves for overall survival of 47 advanced gastric cancer patients [/fig_ref]. The median follow-up time was 10.1 months. Six patients (11.8%) completed treatment and their tumors became resectable. They underwent surgery and 5 of these patients were still alive with disease-free status during their last follow-up.
## Safety
Safety was assessed in 51 patients; the adverse events are listed in [fig_ref] TABLE 4: Toxicities According to Treatment [/fig_ref]. In the XELOX period, the most common grade 3/4 adverse event was hand-foot syndrome (9.8%). In the TX period, the most common grade 3/4 adverse events were neutropenia (28.5%), leucopenia (14.3%), and hand-foot syndrome (9.5%). Overall, the most common grade 3/4 adverse events were neutropenia (25.4%), leucopenia (11.8%), diarrhea (9.8%), and hand-foot syndrome (15.7%). Other toxicities were usually mild and manageable. Grade 3 toxicities with a frequency of 5% or more included oral mucositis (5.9%), nausea (5.9%), fatigue (5.9%), and thrombocytopenia (5.9%). No grade 3/4 peripheral neuropathy was observed in this study.
# Discussion
The prognosis of advanced gastric cancer is still dismal. For gastric cancer patients with unresectable tumors, systemic chemotherapy is the cornerstone of treatment, and it shows improved survival compared with best supportive care. [bib_ref] Chemotherapy for advanced gastric cancer, Wagner [/bib_ref] Several novel drugs have been developed in recent years. One of the most important drugs is docetaxel, which was recently approved as a first-line treatment on the basis of the TAX-V325 trial. [bib_ref] Phase III study of docetaxel and cisplatin plus fluorouracil compared with cisplatin..., Van Cutsem [/bib_ref] In this pivotal trial, the DCF regimen proved superior to the CF regimen; it showed improved time to progression (5.6 vs 3.7 months, respectively; P < 0.001), response rate (37% vs 25%, respectively; P ¼ 0.01), and OS (9.2 vs 8.6 months, respectively; P ¼ 0.02). However, the regimen is limited in clinical practice due to severe hematologic toxicity. When we consider that the aim of the treatment in advanced gastric cancer patients is generally palliative, the tolerability of the treatment is a very important issue. Therefore, we sequentially administered capecitabine, oxaliplatin, and docetaxel in order to increase the tolerability of this combination. This is the first phase II study investigating the sequential therapy of the XELOX regimen followed by the TX regimen in the treatment of advanced gastric cancer. The 61.7% overall response rate of the evaluable patients, the PFS of 8.6 months and the median survival of 11 months are not inferior to the results of the other phase III clinical trials of the current reference regimens, including DCF, EOX, XP and TS1þ cisplatin. [bib_ref] Phase III study of docetaxel and cisplatin plus fluorouracil compared with cisplatin..., Van Cutsem [/bib_ref] [bib_ref] Randomized phase III study comparing irinotecan combined with 5-fluorouracil and folinic acid..., Dank [/bib_ref] [bib_ref] Capecitabine/cisplatin versus 5-fluorouracil/cisplatin as first-line therapy in patients with advanced gastric cancer:..., Kang [/bib_ref] [bib_ref] Multicenter phase III comparison of cisplatin/S-1 with cisplatin/infusional fluorouracil in advanced gastric..., Ajani [/bib_ref] [bib_ref] Capecitabine and oxaliplatin for advanced esophagogastric cancer, Cunningham [/bib_ref] In these trials, the objective tumor response rate and the median survival ranged from 29% to 47% and from 8.6 to 11.1 months, respectively. The results are also not inferior to the results of docetaxel-based phase II/III studies, [bib_ref] Phase III study of docetaxel and cisplatin plus fluorouracil compared with cisplatin..., Van Cutsem [/bib_ref] [bib_ref] A phase II study of weekly docetaxel and cisplatin plus oral tegafur/uracil..., Li [/bib_ref] [bib_ref] Docetaxel plus oxaliplatin with or without fluorouracil or capecitabine in metastatic or..., Van Cutsem [/bib_ref] [bib_ref] Randomised, noncomparative phase II study of weekly docetaxel with cisplatin and 5-fluorouracil..., Tebbutt [/bib_ref] [bib_ref] Randomized multicenter phase III study of a modified docetaxel and cisplatin plus..., Wang [/bib_ref] in which the objective tumor response rate and median survival ranged from 21% to 54% and from 8.6 to 14.5 months, respectively.
In this study, the ORR in evaluable patients was 61.7%. Six patients whose gastric cancers were initially unresectable became resectable after receiving treatment [fig_ref] TABLE 5: Patients Who Became Operable After Chemotherapy Case Number Initial Stage Best Response... [/fig_ref]. Notably, 1 patient (case number 2) received operation was due to tumor rupture but the pathology showed pathologic complete remission. These patients underwent surgery and 5 of them are still alive with disease-free status. Because of the high ORR and disease control rate, this sequential chemoregimen may be considered for the neoadjuvant setting of gastric cancer. Further studies are warranted.
In the phase III TAX-V325 trial, the DCF regimen had better results than the CF regimen but it was associated with intolerable toxicity; 82% of patients experienced grade 3/4 neutropenia and 14.3% of patients experienced febrile neutropenia. [bib_ref] Phase III study of docetaxel and cisplatin plus fluorouracil compared with cisplatin..., Van Cutsem [/bib_ref] In other docetaxel-based chemoregimens, 5-10 the rate of grade 3/4 neutropenia was lower and ranged from 37.2% to 70%. In our study, the sequential therapy with XELOX followed by TX was well tolerated by most patients. The most common grade 3/4 side effects were neutropenia (28.5%), decreased white blood cell counts (14.2%), and hand-foot syndrome (9.5%). It is worth mentioning that no patient experienced febrile neutropenia or grade 3/4 peripheral sensory neuropathy, the most concerning adverse events associated with docetaxel and oxaliplatin. To the best of our knowledge, the rate of grade 3/4 neutropenia is lowest with docetaxel-based chemotherapy. Few grade 3/4 nonhematologic adverse events were observed; these included mucositis, nausea, vomiting, fatigue, and anorexia. The favorable results may be due to the sequential therapy, which may have avoided the overlapping toxicity of concomitant agents and reduced the exposure doses of each agent. The administration of the sequential therapy with XELOX followed by TX was convenient and practical for the gastric cancer patients. Unlike the traditional DCF regimen which requires intravenous infusion of 5-FU for 5 days, the sequential therapy only required a 2-h infusion of oxaliplatin in the XELOX regimen and a 1-h infusion of docetaxel in the TX regimen. Reducing the length of time required for the intravenous infusion may also decrease the number central line infections. This regimen can easily be administered at outpatient clinics.
[formula] 0(0) 2(5) 0(0) 0(0) 14(27) 2(4) 0(0) Limbs edema 2(4) 0(0) 0(0) 2(5) 0(0) 0(0) 3(6) 0(0) 0(0) Fatigue 21(41) 1(2) 0(0) 24(57) 2(5) 0(0) 30(59) 3(6) 0(0) AST increased 3(6) 0(0) 0(0) 1(2) 0(0) 0(0) 3(6) 0(0) 0(0) ALT increased 3(6) 0(0) 0(0) 2(5) 0(0) 0(0) 5(10) 0(0) 0(0) TB increased 0(0) 0(0) 0(0) 1(2) 0(0) 0(0) 1(2) 0(0) 0(0) Cr increased 0(0) 1(2) 1(2) 1(2) 0(0) 0(0) 1(2) 1(2) 1(2) Anorexia 21(41) 1(2) 0(0) 15(36) 0(0) 0(0) 27(53) 1(2) 0(0) Dizziness 2(4) 0(0) 0(0) 0(0) 0(0) 0(0) 2(4) 0(0) 0(0) Paresthesia 4(8) 0(0) 0(0) 4(10) 0(0) 0(0) 5( [/formula]
In conclusion, the sequential therapy of the XELOX regimen followed by the TX regimen is effective, well-tolerated, convenient, and practical for advanced gastric cancer patients in daily practice.
[fig] FIGURE 1: The Consolidated Standards of Reporting Trials (CONSORT) diagram depicting the trajectory of the trial. CONSORT ¼ Consolidated Standards of Reporting Trials. [/fig]
[fig] FIGURE 2: Kaplan-Meier curves for progression-free survival of 47 advanced gastric cancer patients. [/fig]
[fig] FIGURE 3: Kaplan-Meier curves for overall survival of 47 advanced gastric cancer patients. [/fig]
[table] TABLE 1: Patient Characteristics ECOG ¼ Eastern Cooperative Oncology Group. [/table]
[table] TABLE 2: Cycles of Treatment and Follow-Up Time [/table]
[table] TABLE 3: Best Response Rate According to Response Evaluation Criteria In Solid Tumors (RECIST) [/table]
[table] TABLE 4: Toxicities According to Treatment [/table]
[table] TABLE 5: Patients Who Became Operable After Chemotherapy Case Number Initial Stage Best Response Outcome PR ¼ partial remission; CR ¼ complete remission. [/table]
|
Radical cystectomy (bladder removal) against intravesical BCG immunotherapy for high-risk non-muscle invasive bladder cancer (BRAVO): a protocol for a randomised controlled feasibility study
for the BRAVO study group To cite: Oughton JB, Poad H, Twiddy M, et al. Radical cystectomy (bladder removal) against intravesical BCG immunotherapy for high-risk non-muscle invasive bladder cancer (BRAVO): a protocol for a randomised controlled feasibility study.
Introduction High-risk non-muscle invasive bladder cancer (HRNMIBC) is a heterogeneous disease that can be difficult to predict. While around 25% of cancers progress to invasion and metastases, the remaining majority of tumours remain within the bladder. It is uncertain whether patients with HRNMIBC are better treated with intravesical maintenance BCG (mBCG) immunotherapy or primary radical cystectomy (RC). A definitive randomised controlled trial (RCT) is needed to compare these two different treatments but may be difficult to recruit to and has not been attempted to date. Before undertaking such an RCT, it is important to understand whether such a comparison is possible and how best to achieve it. Methods and analysis BRAVO is a multi-centre, parallelgroup, mixed-methods, individually randomised, controlled, feasibility study for patients with HRNMIBC. Participants will be randomised to receive either mBCG immunotherapy or RC. The primary objective is to assess the feasibility and acceptability of performing the definitive phase III trial via estimation of eligibility and recruitment rates, assessing uptake of allocated treatment and compliance with mBCG, determining quality-of-life questionnaire completion rates and exploring reasons expressed by patients for declining recruitment into the study. We aim to recruit 60 participants from six centres in the UK. Surgical trials with disparate treatment options find recruitment challenging from both the patient and clinician perspective. By building on the experiences of other similar trials through implementing a comprehensive training package aimed at clinicians to address these challenges (qualitative substudy), we hope that we can demonstrate that a phase III trial is feasible. Ethics and dissemination The study has ethical approval (16/YH/0268). Findings will be made available to patients, clinicians, the funders and the National Health Service through traditional publishing and social media. Trial registration number ISRCTN12509361; Pre results.
# Introduction
## Context
Bladder cancer (BC) is a common disease that is one of the most expensive malignancies to manage. [bib_ref] International variations in bladder cancer incidence and mortality, Chavan [/bib_ref] Around 25% of patients present with poorly differentiated, low-stage tumours, termed 'high-risk non-muscle invasive bladder cancer' (HRNMIBC; including tumours with carcinoma in situ, invasion into the lamina propria and intraepithelial spread into the prostatic urethra). The two main treatment options for HRNMIBC are intravesical immunotherapy (using a maintenance regime of intravesical maintenance BCG (mBCG)) and radical cystectomy (RC). The former aims to induce an immune response against the tumour and may reduce the risk of progression to muscle invasion. [bib_ref] An individual patient data meta-analysis of the long-term outcome of randomised studies..., Malmström [/bib_ref] While mBCG avoids bladder removal, it leaves patients at risk of local progression and may impact on quality of life (QoL) through local symptoms and anxiety. RC removes the risk of local disease progression and may have the best oncological outcomes but could be overtreatment for non-progressing tumours.
## Strengths and limitations of this study
► This is an important comparison that has not been attempted before. ► This study will not determine which intervention is the superior treatment; a definitive phase III trial will still be needed. ► Recruitment may be challenging and may not be possible through traditional care pathways.
## Open access
Many patients develop short-term postoperative complications after RC, and others have a reduction in QoL following surgery. To date, RC and mBCG have not been directly compared. Their comparative risks and benefits are unknown, hampering decision making, clinical care and exposing patients to both overtreatment and undertreatment.
## Current knowledge
The natural history of HRNMIBC is unpredictable. Rates of progression to muscle invasion and metastases vary between 25% and 75%, 3 and long-term outcomes suggest around 20%-25% of patients with HRNMIBC may die from BC. mBCG avoids bladder removal, and meta-analyses report potential reductions in progression by 5% at 2.5 years 6 . However, mBCG can be poorly tolerated, its impact on progression is debated 2 and there are manufacturing problems. 7 mBCG involves 27 intravesical instillations and 10 cystoscopies over 3 years. Many (74%) patients report local and systemic toxicity, so only 30% of patients complete mBCG. Furthermore, there are few data to support that mBCG with bladder preservation preserves a good QoL. With regards to oncological outcomes, reports of BCs failing mBCG find upstaging to invasion in 27%-63% of tumours and the cancer-specific survival is worse than for BC with de novo muscle invasion (eg, 37% vs 67%/3 years). [bib_ref] The increasing use of intravesical therapies for stage T1 bladder cancer coincides..., Lambert [/bib_ref] [bib_ref] Prognosis of muscleinvasive bladder cancer: difference between primary and progressive tumours and..., Schrier [/bib_ref] [bib_ref] Cystectomy in patients with high risk superficial bladder tumors who fail intravesical..., Huguet [/bib_ref] [bib_ref] Early versus deferred cystectomy for initial high-risk pT1G3 urothelial carcinoma of the..., Denzinger [/bib_ref] [bib_ref] The optimum timing of radical cystectomy for patients with recurrent high-risk superficial..., Solsona [/bib_ref] RC includes removal of the bladder and adjacent organs and reconstruction of urinary drainage. Many patients develop short-term bowel, respiratory or cardiovascular problems, including up to 20% require intervention. [bib_ref] Defining early morbidity of radical cystectomy for patients with bladder Cancer using..., Shabsigh [/bib_ref] Prospective studies report that recovery of QoL following RC takes 6 months or longer to recover to preoperative levels. [bib_ref] Quality of life in patients with bladder carcinoma after cystectomy: first results..., Hardt [/bib_ref] Recurrence-free survival rates following primary RC for HRNMIBC cancers appear superior to those from mBCG (eg, 79%/10 years). 18
## Surgical rcts
As contemporary data challenge the role of mBCG 2 and lessons have been learnt from large surgical randomised controlled trials (RCTs), [bib_ref] Active monitoring, radical prostatectomy, or radiotherapy for localised prostate cancer: study design..., Lane [/bib_ref] we believe it is time to compare mBCG with RC. This is an important comparison, and this opportunity may be lost as RC for HRNMIBC becomes more popular. [bib_ref] Invasive T1 bladder cancer: indications and rationale for radical cystectomy, Stein [/bib_ref] Importantly, the 2015 NICE BC guidelines selected this comparison as one of the highest-ranked research priorities in the disease.The BRAVO study aims to compare surgical and non-surgical treatments. Trials of similarly disparate treatments in BC have previously failed to recruit (eg, CRUK-SPARE trial). Here we propose the preliminary work necessary to understand if we can undertake a large RCT of mBCG versus RC. Anticipated barriers to recruitment include patient and clinician preferences, BC treatment pathways, a lack of high-quality information and the need for staff training in equipoise and communicating RCT methods. [bib_ref] Improving the recruitment activity of clinicians in randomised controlled trials: a systematic..., Fletcher [/bib_ref] To address these issues, we will develop a tailored staff training package to facilitate informed decision making about participation and to better understand RCT methodology. The development work will be informed by existing knowledge and context-specific evidence derived from interviews with patients and healthcare staff exploring: (a) treatment perceptions, (b) patient pathways to treatment, (c) barriers to participation and (d) training needs of site staff. This qualitative work to develop and deliver the training package is described in a separate protocol (see online supplementary file 1). We will then undertake a feasibility study to assess whether recruitment could be achieved in a definitive trial, embedding a qualitative component to establish patient experience.
## Study aims
Our aims are to assess whether a larger phase III RCT is possible and to acquire sufficient 129 data to aid planning such a trial. Primary outcomes are: ► To assess the number of patients screened and identified as eligible within these six centres. ► To assess recruitment rates (number of patients randomised per month).
Secondary outcomes are: ► To assess acceptance of allocated treatment. ► To assess the rate of compliance with mBCG at 12 months after randomisation and collect reasons for non-compliance. ► To assess the feasibility and optimal frequency of collecting QoL data in patients treated for HRNMIBC. ► To obtain preliminary data on the QoL data of patients treated for HRNMIBC. ► To explore the reasons expressed by patients for declining recruitment into the study.
# Methods and design
Trial design BRAVO is a multi-centre, parallel-group, mixed-methods, individually randomised, controlled feasibility study in patients with HRNMIBC suitable for treatment by either mBCG or RC. Eligible, consenting patients will be randomised (1:1) to receive either mBCG or RC (figure 1). Due to the different treatment modalities in the two arms, it is not feasible to blind patients or clinicians to treatment allocation. Patient reported outcome data will be collected at 3, 6 and 12 months postrandomisation in clinic or by postal questionnaire if the patient is not due to attend a clinic visit. i. pT1 stage. 5. Either re-resection of the bladder (following the initial diagnostic TURBT) within the 3 months prior to randomisation confirming the absence of muscle invasion OR a. the initial diagnostic TURBT biopsy contains muscle, AND b. the radiological and pathological stage assessment are in agreement regarding stage and absence of muscle invasion, AND c. a re-resection is not appropriate in the opinion of the treating clinician AND Open Access d. the initial TURBT is within 3 months prior to randomisation. 6. CT or cross-sectional imaging of the abdomen and pelvis within the year prior to starting treatment. 7. Imaging of the lungs and thorax within 3 months prior to randomisation. 8. Suitable and fit for both mBCG and RC as determined by the treating clinician. 9. Central multidisciplinary team (MDT) pathological review agrees with diagnosis. 10. If female, must be (as documented in patient notes):
## Trial population
a. postmenopausal (no menses for 12 months without an alternative medical cause) or b. surgically sterile (hysterectomy, bilateral salpingectomy or bilateral oophorectomy) or c. using acceptable contraception (which must be continued for 7 days after the last dose of BCG or until RC is carried out). Women of childbearing potential must undergo a pregnancy test before randomisation. d. not breast feeding. The exclusion criteria are: 1. solely non-urothelial or any variant urothelial pathology 2. unable or not willing to give informed consent 3. previous high-risk (high grade or grade 3) non muscle invasive (NMI) or invasive BC 4. any previous treatment with intravesical BCG 5. any previous treatment with pelvic radiotherapy 6. any other malignancy (excluding non-melanomatous skin cancer, low-risk prostate cancer and prior low-risk BC).
Eligibility waivers are not permitted. Prior to entry, patients must be accurately staged (eg, cross-sectional imaging (eg, CT) of the abdomen, pelvis and thorax, or bone scan if indicated, within 3 months prior to randomisation) and judged to be eligible for both treatments (anaesthetic evaluation in those with borderline fitness for RC). After trial entry, women of childbearing age must be proven to be not pregnant (pregnancy test).
## Sample size
The sample size for this feasibility study has been set to give confidence that the recruitment target for the main trial can be met. A formal power calculation is not appropriate as effectiveness is not being evaluated. It is estimated that, per year, over the six centres, there will be approximately 1000 new diagnoses of NMIBC, where 20% are likely to be eligible (200 patients). [bib_ref] Stage, grade and pathological characteristics of bladder cancer in the UK: British..., Boustead [/bib_ref] We would need to show that we are able to randomise approximately 25% of all eligible patients to be confident that the recruitment target for the main trial would be met within 3 years, with an additional nine centres. We therefore plan to recruit 60 patients over an 18-month period in the feasibility study. For the phase III trial, we anticipate either a single primary endpoint (cancer-specific survival) or coprimary endpoints (cancer-specific survival and averaged quality-adjusted life years (QALYs)). We estimate that 506 participants are required to have 80% power to show a superiority HR of 0.626 (based on an improvement in 5-year cancer-specific survival from 70% in the BCG arm to 80% in the RC arm), assuming a 3-year accrual period, 5 years of follow-up and accounting for 5% loss-to-follow-up.
## Setting
Participants will be recruited from six cancer centres (and seven neighbouring district hospitals) within Yorkshire and Northumberland. National Health Service (NHS) demographic data show that Yorkshire and Northumberland have some of the highest rates of BC incidence and some of the lowest rates of survival from this cancer. Recruitment Patients will be identified through MDT meetings and approached once they know their diagnosis of HRNMIBC. This approach may be at any hospital involved in their care and by medical or nursing staff. The team will introduce the trial when treatment options are being discussed, provide the introduction leaflet and ask permission (and contact details) for a research nurse to contact the patient with more information. The number of eligible and screened patients will be recorded. Interested participants will be invited to attend an appointment at the research site and/or receive telephone calls, to be given a full explanation of the BRAVO study. Experience in similar studies suggests that patients can be overwhelmed by information given in clinic and that telephone contact can help and provides another opportunity to support patients. Up to five attempts will be made to contact the participant by telephone, after which it will be assumed they have decided to not participate. Eligible patients can be contacted by post if the immediate care team deem this best. No contact information will be shared outside of the team directly caring for the patient unless consent has been obtained.
## Consent
Informed consent takes place in a face-to-face setting at the research site. Patients will have at least 24 hours to consider participation and will be encouraged to discuss the study with their family and other healthcare professionals. A full verbal explanation of the study, a written Patient Information Sheet (detailing rationale, design and personal implications of trial entry) and informed consent form will be provided. Participants may withdraw at any stage of the trial. Consent will be obtained prior to collection of baseline assessment data and subsequent randomisation.
## Staff training
We recognise the challenge of comparing these two treatment choices and that the patient pathway includes interaction with numerous healthcare providers. To minimise bias and to maintain equipoise, a training package Open Access will be developed from interviews with patients and clinicians and delivered to staff who are likely to care for patients before and during the study. Training will incorporate lectures and role play exercises with simulated patients. A careful explanation of the potential risks and benefits of the two treatment interventions is crucial; such risks will be clearly explained to interested patients in an unbiased and fair way, assisted by written study-specific patient information.
## Randomisation
Patients will be randomised, using a 24-hour centralised telephone or web-based randomisation system, on a 1:1 basis to receive either RC or mBCG. A computer-generated adaptive minimisation algorithm that incorporates a random element will be used to ensure that the treatment groups are balanced (stratified) for: ► age (<75,>=75) ► sex (male, female) ► recruiting cancer centre ► tumour stage (pTa/pTis, pT1) ► presence of carcinoma in situ (yes, no) ► previous low-risk BC (yes, no).
## Intervention: bcg immunotherapy
Maintenance BCG immunotherapy will be administered at either the cancer centre or district general hospital using the SWOG protocol. [bib_ref] Maintenance Bacillus Calmette-Guerin immunotherapy for recurrent TA, T1 and carcinoma in situ..., Lamm [/bib_ref] At least 12 months of BCG treatment are required, and 6 weeks of induction BCG will be followed by 3 doses at 4 and 10 months after diagnosis. Delays and deferrals are common and allowed within this study. BCG induction should include at least 4 (of 6) doses of BCG, and induction should be completed within 10 weeks. The presence of an invasive BC requires the cessation of mBCG and a change in treatment intent. Maintenance BCG may continue in the presence of low-risk NMI and HRNMI BC at the first cystoscopy; thereafter, these are managed as recurrences and require patient discussion. Rigid cystoscopy with bladder biopsy and bladder washings is mandated at the first check. After this, bladder surveillance is performed as per local protocol (flexible or rigid instruments). All cystoscopies will be undertaken or directly supervised (with a visual check) by a consultant urologist who manages HRNMIBC. Fluorescence or narrow band imaging may be used, as per local protocols. Histological review of the bladder biopsies and urinary cells should be performed to determine the presence or absence of BC.
Local and systemic complications are common in mBCG regimens and should be managed as per local protocol. The study will collect data on the frequency of expected BCG toxicities and whether this leads to the cessation of BCG treatment. Cystectomy may be performed within BRAVO for severe BCG-related toxicities, if these warrant such an intervention. Patients undergoing BCG treatment may stop treatment due to disease progression, disease recurrence, serious BCG intolerance or side effects or patient choice. Disease progression: patients who have confirmed progressive disease after any of the check cystoscopies (presence of pT2 tumours, cancer in lymph nodes or metastases) should stop BCG and be offered curative treatment for muscle invasive BC. Disease recurrence is defined as the presence of low-risk NMI or HRNMIBC from the second check cystoscopy onwards. Participants with recurrence should be offered the option of changing treatment, including RC or using second-line intravesical approaches.
Intervention: RC RC should be performed at each cancer centre by teams specialising in this service. Variations in surgical performance and practice produce wide differences in morbidity and mortality from RC. [bib_ref] Impact of hospital and surgeon volume on in-hospital mortality from radical cystectomy:..., Konety [/bib_ref] To mitigate these, surgeons within BRAVO will have individually undertaken at least 10 RCs per year for the last 2 years (or 20 in the last year), have median length of stay rates under 16 days and have a 90-day post-RC mortality rate of less than 10% (collected outcomes from the British Association of Urological Surgeons RC complex dataset). [bib_ref] Stage, grade and pathological characteristics of bladder cancer in the UK: British..., Boustead [/bib_ref] Postoperative complication rates and intraoperative and postoperative transfusion rates will also be taken into consideration. Individual surgeon data will act as surrogate measures for the entire surgical team and require accreditation from the trial management group before entry into BRAVO. Submitted data for surgical accreditation should reflect the practice to be undertaken within this study (eg, open or robotic approaches). Surgery should take place within 8 weeks of randomisation.
Cystectomy should include removal of adjacent organs. In males, this includes the prostate and seminal vesicles. In females, this should include a section of adjacent anterior vaginal wall, the uterus, cervix and fallopian tubes and, if no bladder reconstruction is planned, the urethra. Oophorectomy is optional, as per local practice and individualised for each patient. Pelvic lymphadenectomy is mandated within BRAVO. The template should at least include the regional lymph nodes up to the level of the ureteric crossing of the common iliac vessels. This includes the obturator fossa, the external iliac and internal iliac nodes. A more extended lymphadenectomy is acceptable. Excised lymphatic tissue should be submitted for histological analysis. Perioperative care is to be carried out as per enhanced recovery after surgery protocols. Withdrawal of treatment In line with usual clinical care, cessation or alteration of regimens will be at the discretion of attending clinicians or the participants. All participants who withdraw or are withdrawn from their allocated treatment will still attend for follow-up assessments and complete questionnaires unless unwilling to do so and outcomes will continue to be collected. In the event that a patient withdraws consent prior to randomisation, data collected up to the point of withdrawal will be analysed. Open Access data collection A screening form, to include demographic details and reasons for ineligibility, exclusion or refusal, will be completed for all patients considered for BRAVO. A feedback questionnaire will be used to identify patients who are willing to take part in the qualitative substudy (see online supplementary file 1). Baseline assessments prior to randomisation include QoL scores (EuroQuol-5D (EQ-5D), 37 EORTC QLQ-C30, [bib_ref] The European Organization for Research and Treatment of Cancer QLQ-C30: a quality-of-life..., Aaronson [/bib_ref] EORTC QLQ-BLM30) at trial entry.
Within mBCG, outcomes and compliance data will be collected at each cystoscopy. For RC, patient and operative data will be collected at the time of surgery, as per our national register, [bib_ref] Stage, grade and pathological characteristics of bladder cancer in the UK: British..., Boustead [/bib_ref] and then at each subsequent follow-up visit (3, 6 and 12 months postrandomisation). Follow-up imaging (CT scan) to assess response to treatment will be performed in both arms at 1-year postrandomisation. QoL questionnaires will be collected at 3, 6 and 12 months postrandomisation in face-to-face consultations or by telephone. These include EQ-5D, [bib_ref] EQ-5D: a measure of health status from the EuroQol Group, Rabin [/bib_ref] EORTC QLQ-C30 38 and either EORTC QLQ-BLM30 (for those randomised to RC) or EORTC QLQ-NMIBC24 (for those randomised to BCG). Information will be collected on deaths, complications and toxicities (adverse events) and related and unexpected serious adverse events up to 1 year postrandomisation or 3 months after the last participant is randomised if earlier.
## Statistical analyses
A detailed statistical analysis plan will be written before any analysis is undertaken. All analyses and data summaries will be conducted on the intention-to-treat population. No formal interim analyses are planned, and final analysis will take place when all available data have been received. The analysis will focus on descriptive statistics and CI estimation. Primary analysis will include summaries of the number of patients at each stage of the recruitment pathway (screening, eligibility, consent and randomisation) and assessment of the overall monthly recruitment rate. Secondary analysis will include summaries of acceptance of randomised treatment and mBCG treatment compliance. Participant retention and self-reported QoL outcomes during follow-up, including withdrawal data (timing and reason), will also be summarised overall and by time point. Levels of missing data in QoL outcomes will be assessed. The median cancer-specific survival estimate and its corresponding 95% CI will be calculated to inform the sample size calculation of the phase III trial. As this is to aid the design of a pragmatic phase III trial, all randomised patients will be included in the calculation, regardless of treatment received. Cancer-specific survival will be calculated from the date of randomisation to the date of cancer-specific death. Participants with missing follow-up data or who are alive at the time of the analysis will be censored at the date they were last known to be alive. Overall survival, calculated from the date of randomisation to the date of death, will also be summarised as for cancer-specific survival.
The frequent collection of QoL data within this feasibility study is necessary in order to assess the burden to patients. This will be assessed by monitoring collection compliance rates and will inform the optimal frequency of data collection for the main trial. Averaged QALYs may be a coprimary endpoint for the main trial; as such, determining the optimal frequency of EQ-5D data collection within this feasibility study is crucial.
## Safety
The number of adverse events and related unexpected serious adverse events will be summarised descriptively by arm, grade and body system. The proportion of participants experiencing each toxicity will be summarised by maximum National Cancer Institute's Common Terminology Criteria for Adverse Events grade 39 experienced, overall and by arm. Operative RC complications will be graded using the Clavien Dindo classification. [bib_ref] Classification of surgical complications: a new proposal with evaluation in a cohort..., Dindo [/bib_ref] Criteria for progression to the definitive phase III trial The following guidelines for progression to a definitive phase III trial have been defined: ► The recruitment and follow-up rates must demonstrate that a definitive trial using similar procedures will achieve sufficient power to test the hypothesised difference between treatment arms. ► The sample size calculation for the feasibility study and proposed phase III trial are provided earlier. This assumes that 20% of all new diagnoses of NMIBC would be eligible and approximately 25% of those would be randomised. To proceed to a definitive trial, we need to show that at least 20% of eligible patients can be randomised.
## Qualitative substudy
There are two qualitative studies. The first was undertaken prior to the start of the RCT to identify a priori the barriers to recruitment from the perspectives of patients and staff to inform the development of a bespoke training package for staff 41 (see online supplementary file 1). A second qualitative study is embedded into the RCT trial to understand patients' views and experiences of the treatments and explore patients' acceptability of the study and recruitment processes:
Qualitative substudy objectives 1. To gauge patients' understanding of the study and their views on the recruitment process. 2. To qualitatively explore patient's acceptability of the study to assist in optimisation of recruitment strategies employed for the definitive trial. 3. To explore reasons for participation and nonparticipation of eligible patients. 4. To understand patients' experience of the randomisation process on decision making. 5. To understand why people refuse to participate or do not take up allocated treatment.
## Open access
6. Patient understanding of study materials, that is, do patients understand what will happen if they take part and do they understand what they are being randomised to. 7. Acceptability of study procedures. 8. Acceptability of randomisation.
## Qualitative substudy overview
In order to examine the views and experiences of patients with BC, we will conduct in-depth semistructured interviews with patients approached to take part in the trial. Qualitative findings will help illuminate the acceptability of trial processes and explore barriers to uptake. Recruitment to RCTs with very different treatment arms can be difficult, and recruitment to trials involving surgery is particularly challenging. [bib_ref] What influences recruitment to randomised controlled trials? A review of trials funded..., Mcdonald [/bib_ref] Trials present practical and methodological challenges, including difficulties in recruitment, randomisation and lack of clinical equipoise. [bib_ref] Barriers to recruitment for surgical trials in head and neck oncology: a..., Kaur [/bib_ref] Understanding why patients do or do not participate in trials is important, and clinical trials have recently begun to incorporate a qualitative component to address these issues. These studies have been able to successfully identify aspects of the trial design that hindered recruitment and identify possible solutions. Qualitative substudy design All eligible patients will be asked to complete a questionnaire to gauge their understanding of BRAVO and their views on the recruitment process. We will collect data from patients who decline the study, who consent but refuse allocation and who consent and accept allocation. A short questionnaire will be given to seek patient views on the recruitment process and to ask if participants would be willing to provide detailed feedback by face-toface or telephone interview. A purposive sample of 15 patients will be selected for interview. Written consent will be taken prior to the interview and a flexible topic guide developed in conjunction with patient and public involvement (PPI) representatives, clinical colleagues and informed by the literature used to assist questioning. The topic guide will be devised to ensure that the key issues are covered but do not dictate data collection and will be flexible enough to elicit participants' own experiences and views of the trial as well as issues unanticipated by the interview team. Interviews will be audio-recorded, transcribed and anonymised to protect confidentiality. With their consent, participants may be contacted after the interview to answer questions that may emerge during the analysis or to explore issues that emerged in the interviews in more depth.
Qualitative substudy data analysis Qualitative data will be analysed by the qualitative researcher. Interview transcripts will be checked for accuracy and then managed using NVivo qualitative data analysis software (QSR International, Daresbury, UK), which aids the indexing of qualitative data. Analysis will start during data collection and will inform later data collection; for example, emerging themes may identify new questions to explore in later interviews. The data will be analysed using thematic analysis using an inductive (bottom up) approach to identify and analyse patterns across the dataset using constant comparison methods. Inductive coding will follow using a line-byline coding approach, with codes assigned to segments of data that provide insight into participants' views of the trial. An initial coding frame will be developed from the first interviews and will be modified, if necessary, as the analysis develops. A subset of transcripts will be independently coded by another member of the team and compared with ensure consistency. Any discrepancies will be discussed with the research team and resolved to achieve coding consensus. The data will be examined for negative cases, and the reasons will be explored by comparison with the overall dataset. data monitoring Trial supervision includes a core project team, a trial management group (TMG) and an independent trial steering committee (TSC). For a feasibility study of this nature and duration, a separate data monitoring and ethics committee is not required; rather; TSC adopts a safety monitoring role and will review safety issues if this becomes necessary. Data will be monitored for quality and completeness by the Clinical and Translational Research Unit (CTRU). Missing data (except individual items collected via questionnaires) will be chased until received, confirmed as not available or the trial is at analysis. Any protocol changes will be disseminated by CTRU to the relevant parties.
## Trial organisation and administration
The trial was developed by the BRAVO TMG. The trial is funded by Yorkshire Cancer Research and is sponsored by the Sheffield Teaching Hospitals NHS Trust (Clinical Research Office, Royal Hallamshire Hospital, D Floor, Glossop Road, Sheffield), coordinated by CTRU, University of Leeds, and is registered (ISRCTN12509361). The trial will be conducted in accordance with the principles of Good Clinical Practice (GCP) in clinical trials, as applicable under UK regulations, the NHS Research Governance Framework and through adherence to CTRU standard operating procedures. CTRU/sponsor have systems in place to ensure that serious breaches of GCP or the trial protocol are identified and reported. Ethical approval has been obtained from the National Research Ethics Service Committee Yorkshire & Humber-South Yorkshire (reference no. 16/YH/0268). Any on-site source data verification carried out by CTRU is not independent from sponsor. Sheffield Teaching Hospitals NHS Trust will not be liable for negligent harm caused by the design of the trial. No additional compensation for clinical negligence will be provided for trial participants over that which is available to NHS patients. All identifiable information collected during the course of the study will be kept strictly confidential and not transferred outside Open Access of the research team. Patient name (via consent form), email address and telephone number will be collected when a patient is randomised into the study, but all other data collection forms that are transferred to or from the CTRU will be coded with a study number and will include two patient identifiers, usually the patient's initials and date of birth. Both electronic and paper data will be held in a secure locations with restricted access.
# Discussion
The 2015 NICE BC guidelines identified the comparison between mBCG and RC as one of their highest research priorities.This reflects the importance of this question but does not address how randomisation between two very different treatment options should occur or whether such a comparison is possible. Within this feasibility study, we are attempting to understand, address and develop methodology to allow such a comparison. This will require several key issues to be addressed. First, it is clear from other surgical versus non-surgical treatment trials [bib_ref] Quality improvement report: improving design and conduct of randomised trials by embedding..., Donovan [/bib_ref] that the most important element for RCT recruitment is keeping equipoise when discussing the treatment options by medical and nursing staff. While previous studies used research nurses to keep equipoise, this is not viable across many centres within the current research funding climate. In an attempt to replicate this model, we ran a number of educational days to train relevant medical and nursing staff about the importance of equipoise and to discuss their beliefs about HRNMIBC. All staff had opinions about the efficacy of BCG and QoL with RC, and so it was important to discuss these in an open forum to challenge these views and use evidence to dispel prior beliefs. We proposed a six-stage consultation plan to help staff keep patients at equipoise and so facilitate trial entry and treatment acceptance. [bib_ref] A new simple six-step model to promote recruitment to RCTs was developed..., Realpe [/bib_ref] Within this feasibility study, we will determine if this approach is possible and successful. Second, UK data do not accurately identify the number of patients with HRNMIBC, what proportion of these are suitable for both RC and mBCG and how many of these would accept randomised treatment options. Within this feasibility study, we will establish accurate data about the number of eligible cases across this population and understand what proportion accepts their randomised treatment allocation. We will use these findings to power the phase III comparative study. Finally, there are very few reliable data about QoL with mBCG and none that compare this directly to RC. Within this study, we will produce these data within 60 patients (30 for each arm) and so allow this endpoint to be modelled for the larger phase III study.
## Ethics and dissemination
The study has ethical approval from Research Ethics Service Committee Yorkshire & Humber-South Yorkshire (reference no. 16/YH/0268). The results of the study will be published in peer-reviewed publications and will be presented at relevant national and international conferences. We will work with our patient panel of BC survivors to develop lay reports to disseminate research findings to patient groups and the clinical teams at participating sites.
Availability of data CTRU will control the final trial dataset and any requests for access will be reviewed by TMG and TSC, subject to existing contractual arrangements with the funders. The protocol, sample case report forms and participant information are available on a case-by-case basis as agreed by TMG, on request to the corresponding author.
## Trial status
The trial opened to recruitment in October 2016 using protocol version 2.0 (8 October 2016) and is due to close in March 2018.
The protocol was amended to version 3 in October 2016 to account for additional inclusion and exclusion criteria and updated surgeon accreditation criteria. The protocol was amended to version 4 in November 2016 to further update the inclusion criteria and surgeon accreditation criteria. Both amendments were reviewed and approved by the sponsor and the National Research Ethics Service Committee Yorkshire & Humber-South Yorkshire (reference no. 16/YH/0268). Protocol amendments are disseminated to relevant parties by CTRU. Twitter @BRAVO_RCT, @jamieoughton, @JimCatto
[fig] Figure 1: Study flow diagram. HRNMIBC, high-risk non-muscle invasive bladder cancer; IVU, intravenous urogram; MDT, multidisciplinary team; QoL, quality of life; TURBT, transurethral resection of bladder tumour. on July 7, 2021 by guest. Protected by copyright. http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2017-017913 on 11 August 2017. Downloaded from [/fig]
|
Quality indicators of palliative care for cardiovascular intensive care
Healthcare providers working for cardiovascular intensive care often face challenges and they play an essential role in palliative care and end-of-life care because of the high mortality rates in the cardiac intensive care unit. Unfortunately, there are several barriers to integrating palliative care, cardiovascular care, and intensive care. The main reasons are as follows: cardiovascular disease-specific trajectories differ from cancer, there is uncertainty associated with treatments and diagnoses, aggressive treatments are necessary for symptom relief, and there is ethical dilemma regarding withholding and withdrawal of life-sustaining therapy. Quality indicators that can iterate the minimum requirements of each medical discipline could be used to overcome these barriers and effectively practice palliative care in cardiovascular intensive care. Unfortunately, there are no specific quality indicators for palliative care in cardiovascular intensive care. A few indicators and their domains are useful for understanding current palliative care in cardiovascular intensive care. Among them, several domains, such as symptom palliation, patient-and family-centered decision-making, continuity of care, and support for health care providers that are particularly important in cardiovascular intensive care.Historically, the motivation for using quality indicators is to summarize mechanisms for external accountability and verification, and formative mechanisms for quality improvement. Practically, when using quality indicators, it is necessary to check structural indicators in each healthcare service line, screen palliative care at the first visit, and integrate palliative care teams with other professionals. Finally, we would like to state that quality indicators in cardiovascular intensive care could be useful as an educational tool for practicing palliative care, understanding the minimum requirements, and as a basic structure for future discussions.
# Introduction
Cardiovascular disease is one of the major causes of deaths worldwide. In the past several decades, in addition to the progress of surgical procedures and lessinvasive devices, the birth of cardiac intensive care units (CICUs) has contributed to the outcome improvement in cardiovascular disease patients [bib_ref] Evolved role of the cardiovascular intensive care unit (CICU), Kasaoka [/bib_ref]. The role of CICUs has shifted from care for patients with acute coronary syndrome without complications to more complex patients, including heart failure with extracardiac organ dysfunction, high-risk pulmonary embolism, malignant arrhythmia, acute aortic syndrome, and cardiogenic shock, especially necessary for monitoring [bib_ref] Demographics, care patterns, and outcomes of patients admitted to cardiac intensive care..., Bohula [/bib_ref] [bib_ref] How to manage various arrhythmias and sudden cardiac death in the cardiovascular..., Kobayashi [/bib_ref] [bib_ref] Management of acute aortic dissection and thoracic aortic rupture, Fukui [/bib_ref] [bib_ref] Achieving the earliest possible reperfusion in patients with acute coronary syndrome: a..., Nakashima [/bib_ref] [bib_ref] Management of patients with high-risk pulmonary embolism: a narrative review, Yamamoto [/bib_ref]. Even with the progress of the treatment strategies, the mortality rate in the CICUs is still high compared to other general wards due to the complex nature of background characteristics [bib_ref] Demographics, care patterns, and outcomes of patients admitted to cardiac intensive care..., Bohula [/bib_ref]. Therefore, healthcare providers working in CICUs should be competent in dealing with patients' death and end-of-life care [bib_ref] Understanding physicians' skills at providing end-of-life care perspectives of patients, families, and..., Curtis [/bib_ref]. In the past several decades, the perception and attitude toward death has dramatically changed according to the aging society, both for the patients and the physicians. For example, the concept of "less is more" has been spreading around the world [bib_ref] Less is more" in modern ICU: blessings and traps of treatment limitation, Ricou [/bib_ref]. This concept raises the issue of overtreatment, and it is becoming increasingly important to ensure that the level of care provided matches that of patients, families, and society. Considering this background, palliative care, which usually manages patients' death and end-oflife care, has gained more attention in the field of cardiovascular intensive care [bib_ref] The business case for palliative care: translating research into program development in..., Cassel [/bib_ref].
Several barriers to the integration of palliative care and critical care have long been discussed and guidelines for cardiovascular palliative care have been recently published [bib_ref] The changing role of palliative care in the ICU, Aslakson [/bib_ref] [bib_ref] Consensus document and recommendations on palliative care in heart failure of the..., García Pinilla [/bib_ref] [bib_ref] JCS/JHFS 2021 statement on palliative care in cardiovascular diseases, Anzai [/bib_ref]. However, palliative care in the field of cardiovascular intensive care, CICUs, and the integration of these three specialties: palliative care, cardiovascular care, and intensive care, are still in its infancy . In this article, we focus on palliative care in the field of cardiovascular intensive care and introduce the quality indicators for acute cardiovascular disease, which might be useful for many healthcare providers to implement palliative care in cardiovascular intensive care, as well as for educational purposes.
## The definition of palliative care and several barriers in implementing it in cardiovascular intensive care
The term "palliative care" is popular; however, what "palliative care" indicates is vague [bib_ref] Defining the palliative care patient: a systematic review, Van Mechelen [/bib_ref]. Palliative care is usually defined by the World Health Organization as "An approach that improves the quality of life of patients and their families facing problems associated with life-threatening illness, through prevention and relief of suffering by means of early identification, impeccable assessment, and treatment of pain and other problems: physical, psychological and spiritual". Although many physicians totally agree with this common definition and the general concept of palliative care, we could not differentiate "palliative care" from "medicine itself, which are based on similar concepts [bib_ref] The essence of palliative care is best viewed as the "Problematization, Mizuno [/bib_ref]. Due to broadening of the spectrum of palliative care from cancer to chronic disease, palliative care is largely affected by external social needs. The attitude to tackle patients' unmet needs, in other words "problematization" is the fundamental concept of palliative care [bib_ref] The essence of palliative care is best viewed as the "Problematization, Mizuno [/bib_ref]. Although the term "palliative care" is often mistakenly regarded identical to "end-of-life care, many palliative care specialists often declare that the provision of palliative care should depend on need not prognosis [bib_ref] Palliative care is not same as end of life care, Reid [/bib_ref]. However, this needs-driven palliative care model sometimes faces difficulties, especially due to resource limitation, such as access to palliative care team [bib_ref] Palliative care in intensive care units: why, where, what, who, when, how, Mercadante [/bib_ref]. For example, one-third of providers do consider daily participation in intensive care unit rounds by the palliative care team as an optimal way [bib_ref] Survey on barriers to critical care and palliative care integration, Kyeremanteng [/bib_ref]. We should balance the patients' need and available resources in each hospital and each intensive care unit.
Palliative care has been developed mainly in the field of cancer patients and its coverage has expanded to many non-cancer patients [bib_ref] Palliative care for patients with noncancer illnesses, Harrison [/bib_ref]. In addition to the complexity of palliative care itself, there are several challenges and barriers to its implementation in the field of cardiovascular disease compared to that of cancer. First, the disease trajectory differs between patients with cancer and patients with cardiovascular disease. It is well known that the trajectory of cardiovascular disease is characterized by an overall gradual decline in function with intermittent serious episodes and exacerbations, which is in contrast to that of cancer with a short period of evident decline [bib_ref] Illness trajectories and palliative care, Murray [/bib_ref]. Especially in the phase of serious episodes, such as admission to the intensive care unit, not only patients but also physicians are not sure whether the patients could recover to the status before admission. This uncertainty related to cardiovascular disease makes palliative care strategy difficult to apply [bib_ref] The impact of uncertainty on bereaved family's experiences of care at the..., Robinson [/bib_ref]. A survey of palliative care in the intensive care unit revealed that patients and their families often express unrealistic expectations, which could result in a barrier to integrating palliative care into intensive care [bib_ref] Survey on barriers to critical care and palliative care integration, Kyeremanteng [/bib_ref]. Second, the treatment for baseline conditions differ from that of cancer patients. Symptom palliation is one of the important elements in palliative care, and many symptoms overlap between cancer and cardiovascular disease patients [bib_ref] Are there differences in the prevalence of palliative care-related problems in people..., Moens [/bib_ref]. However, one important aspect of symptom palliation in cardiovascular disease is that guideline-based treatments for cardiovascular diseases, such as vasodilators or inotropes, could be effectively used [bib_ref] Palliative care and hospice in advanced heart failure, Lemond [/bib_ref]. For example, dyspnea is a frequently Scope of cardiovascular intensive care unit. The fields of palliative care, intensive care, and cardiovascular care overlap observed symptom in patients with cardiovascular disease. Because guideline-based optimization of conventional therapies, such as angiotensin-converting enzyme inhibitors, can ameliorate dyspnea, we should refrain from blinded opioid prescription before that [bib_ref] Palliative care and hospice in advanced heart failure, Lemond [/bib_ref]. These treatment differences would result in a more aggressive strategy in the end-of-life phase of cardiovascular disease compared with cancer patients and the exact same palliative care strategy would not be applicable in the field of cardiovascular disease [bib_ref] Differences in aggressive treatments during the actively dying phase in patients with..., Mizuno [/bib_ref]. Of course, refractory physical symptoms should be managed by a palliative care specialist, but this should be done alongside fundamental treatments for heart failure itself [bib_ref] Living with and dying from heart failure: the role of palliative care, Gibbs [/bib_ref]. Finally, cardiovascular disease could be associated with unstable hemodynamics, including cardiopulmonary arrest, which could be restored by life-sustaining therapy including cardiopulmonary resuscitation (CPR) and mechanical circulatory support devices. Life-sustaining therapy can prolong patients' lives; however, it is sometimes not desirable for patients and their families. In addition to these fundamental CPRs, the discussion about withdrawing and withholding these mechanical circulatory devices in the end-of-life phase is quite important but still challenging. For example, venoarterial extracorporeal membrane oxygenation (VA-ECMO) provides temporary oxygenation and perfusion to patients with cardiopulmonary failure [bib_ref] Venoarterial-ECMO in the intensive care unit: from technical aspects to clinical practice, Gall [/bib_ref]. Compared with potential long-term usage of mechanical ventilation, VA-ECMO has no long-term option, owing to the technology itself. For the patient who fails to recover and are not eligible for transplantation or destination therapy with ventricular assist devices, there is only terminal discontinuation. Although continuing life-sustaining therapy for prolonged period itself could be the goal of care for patient's family, treating physicians generally need surrogate consent to withdraw life-sustaining therapy, including VA-ECMO. This can present a dilemma and end-of-life conflicts between healthcare providers and families, leaving physicians wanting a greater degree of professional autonomy [bib_ref] A survey of physicians' attitudes toward decision-making authority for initiating and withdrawing..., Meltzer [/bib_ref]. As destination therapy using ventricular assist device was approved and reimbursed this May 2021 in Japan, these new options could make it more difficult for treating physicians to obtain surrogate consent to withdraw life-sustaining therapy.
## The quality indicators for palliative care field
Considering these difficulties of palliative care for cardiovascular disease, the detailed components of palliative care might need to be clarified for many clinicians to implement it in cardiovascular intensive care effectively. Although some guidelines or statements described components of palliative care narratively, iterations or checklists would be better implemented [bib_ref] Consensus document and recommendations on palliative care in heart failure of the..., García Pinilla [/bib_ref] [bib_ref] palliative care and cardiovascular disease and stroke: a policy statement from the, Braun [/bib_ref]. Quality indicators are one of the major options for iterating minimum requirements for each medical field. Quality of care itself is defined as "the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge". One of the definitions of quality indicator is "quantitative measures that provide information about the effectiveness, safety and/or people-centeredness of care. " Across several definitions of quality indicators, there are three essential components: (1) quality goal, (2) measurement concept, and (3) appraisal concept. Quality indicators are also sometimes classified into several categories [bib_ref] Defining and classifying clinical indicators for quality improvement, Mainz [/bib_ref]. Among these, the most widely used classification of quality indicators in healthcare, Donabedian's Structure-Process-Outcome (SPO) framework proposed by Donabedian, included three levels of class: structure, process, and outcome of care [bib_ref] The quality of care. How can it be assessed?, Donabedian [/bib_ref]. Briefly, the structure is applicable to the environment the instruments for palliative care, such as the presence of a palliative care team and 24/7 access to the palliative care team. Process is the actual medical treatment and care provided. Outcome is usually considered as actual outcome, such as mortality of each target population, but in palliative care context, bereaved family survey was only available form as an indicator of quality. The National Consensus Project and National Quality Forum provided eight major domains to capture palliative care: (1) structure and process of care; (2) physical aspects of care; (3) psychological and psychiatric aspects of care; (4) cultural aspects of care; (5) spiritual, religious, and existential aspects of care; (6) ethical and legal aspects of care; (7) care of the patient at the end of life; (8) social aspects of care.In the field of cancer, several quality indicators for palliative care have already been discussed and updated according to these domains [bib_ref] Quality indicators for palliative care: update of a systematic review, De Roo [/bib_ref] [bib_ref] A framework for assessing quality indicators for cancer care at the end..., Seow [/bib_ref] [bib_ref] Identifying potential indicators of the quality of end-of-life cancer care from administrative..., Earle [/bib_ref]. From the perspective of quality improvement measuring, quality of end-of-life care, palliative care utilization and site of death have been discussed in the palliative care for cancer patients [bib_ref] Transition points for the routine integration of palliative care in patients with..., Collins [/bib_ref]. As for quality of end-of-life care, many aggressive treatments especially during endof-life period were frequently discussed and monitored. For example, proportion receiving chemotherapy in the last 14 days of life and admission to the ICU in the last month of life were frequently monitored. According to the claim data, the proposed appropriate threshold for proportion receiving chemotherapy in the last 14 days of life and admission to the ICU in the last month of life were 10% and 4%, respectively [bib_ref] Evaluating claims-based indicators of the intensity of end-of-life cancer care, Earle [/bib_ref]. Palliative care utilization and site of death have also been frequently discussed especially based on the whether the medical care was consistent with the patient's needs [bib_ref] Improving palliative care in selected settings in England using quality indicators: a..., Iliffe [/bib_ref] [bib_ref] Quality indicators for the evaluation of end-of-life care in Germany-a retrospective cross-sectional..., Van Baal [/bib_ref]. Not only would the indicators themselves but also the strategy and frameworks to make quality indicators be useful for many non-cancer field healthcare providers, especially in CICUs, to understand the current concepts and methods of palliative care. In the following sections, we will introduce detailed examples and lists of quality indicators of palliative care in intensive care and cardiovascular intensive care.
## Quality indicators and domains of palliative care in intensive care
There are only a limited number of studies on quality indicators for palliative care in intensive care units (not limited to cardiovascular disease). Clarke et al. summarized 53 quality indicators and seven domains: (1) patient-and family-centered decision-making, (2) communication, (3) continuity of care, (4) emotional and practical support, (5) symptom management and comfort care, (6) spiritual support, and (7) emotional and organizational support for intensive care unit clinicians [bib_ref] Quality indicators for end-of-life care in the intensive care unit, Clarke [/bib_ref]. These seven domains largely overlap with the eight domains of general quality indicators for palliative care by NCP and NQF as described above. Patient-and family-centered decision-making, continuity of care, and support for health care providers are not covered by NCP and NQF domains and could be considered as a unique point of quality indicators for palliative care in the intensive care unit [bib_ref] The national agenda for quality palliative care: the National Consensus Project and..., Ferrell [/bib_ref]. Metaxa et al. performed a systematic review of palliative intervention in the intensive care unit field and reported that these seven domains were not practical, because many of them overlapped with each other. Half of interventions are categorized into patient-and family-centered decision-making and one-third are categorized into communication within the team and with patients and families. Therefore, they advocated a more pragmatic classification of palliative intervention in the intensive care unit by following the intervention taxonomy framework, which summarizes palliative interventions as follows: (1) communication interventions, (2) ethics consultations, (3) educational interventions, (4) palliative care team involvement, and (5) advance care planning. Using these five new categories, half of them are categorized into palliative care team intervention. Both of these, the five new intervention categories and seven domains by Clarke et al., are not perfect, but would be useful to comprehend the current important aspects of palliative care in cardiovascular intensive care.
Studies on quality indicators of palliative care for cardiovascular intensive care are scarce. We performed an updated and a structured PubMed literature review on quality indicators for palliative care in cardiovascular disease patients, with only two articles related to cardiovascular intensive care (Additional file 1). We contrasted these two studies with NQF eight domains and Clarke's most popular quality indicator of palliative care in intensive care unit, as shown in [bib_ref] Development and practical test of quality indicators for palliative care in patients..., Hamatani [/bib_ref] [bib_ref] Quality indicators of palliative care for acute cardiovascular diseases, Mizuno [/bib_ref]. The indicators for heart failure included appropriate heart failure treatment and care, which is similar to performance measures for heart failure treatment itself [bib_ref] 2020 ACC/AHA clinical performance and quality measures for adults with heart failure:..., Heidenreich [/bib_ref]. These indicators implied that even if we consider palliative care for patients with heart problems, we should not forget baseline treatment as described above. Hamatani et al. also measured these indicators in patients with heart problems from three teaching hospitals, which revealed that three indicators were quite low performance: "intervention by multidisciplinary team, " "opioid therapy for patients with refractory dyspnea, " and "screening for psychological symptoms. " The other palliative care quality indicator for acute cardiovascular disease mainly focuses on palliative care itself. The two major domains were "symptom palliation" and "supporting the decision-making process. " The seven sub-categories included: "presence of palliative care team", "patient-family relationship", "multidisciplinary team approach", "policy of approaching patients", "symptom screening and management", "presence of ethical review board", "collecting and providing information for decision-maker", and "determination of treatment strategy and the sharing of the care team's decision".
## Symptom palliation and support decision-making process
Symptom palliation and support decision-making processes are two major domains of quality indicators of palliative care in cardiovascular intensive care and reflect structural and process indicators in the SPO framework . The symptom palliation domain includes five subdomains. These domains and clinical indicators are based on the concept "total pain, " which was advocated by Cicely Saunders and consisted of physical, psychological, social, and spiritual pain [bib_ref] Embracing Cicely Saunders's concept of total pain, Ong [/bib_ref]. Each clinical indicator embraced total pain concept and indicated that we should effectively and efficiently screen total pain and ameliorate these spectra of pain. Based on these quality indicators, there are three steps to approach the total pain. Although palliative care interventions to improve patients' symptoms in the chronic phase have been evaluated in the past several years, there is no specific randomized control trial to evaluate the impact of palliative care on cardiovascular intensive care. In addition, even in the chronic phase, the impact of palliative care interventions remains inconsistent. Palliative care in heart failure (PAL-HF) was the first randomized, controlled clinical trial reported in 2017, which evaluated the additional palliative care intervention on usual care in patients with heart problems [bib_ref] palliative care in heart failure: the PAL-HF randomized, controlled clinical trial, Rogers [/bib_ref]. [bib_ref] The impact of palliative care on clinical and patient-centred outcomes in patients..., Sahlollbey [/bib_ref]. They also reported that there are only three studies evaluating symptom burden, and for individual symptoms, there was no clear impact of palliative care on anxiety, dyspnea, or pain. Quinn et al. also performed a systematic review and meta-analysis of palliative care interventions in chronic noncancer illness [bib_ref] Association of receipt of palliative care interventions with health care use, quality..., Quinn [/bib_ref]. They reported that palliative care was significantly associated with lower symptom burden translated to the Edmonton Symptom Assessment Scale (standardized mean difference − 1.6; 95% CI − 2.6 to − 0.4). Finally, regarding symptom palliation, we should pay attention not only to patients' symptoms but also to caregivers' symptoms at the same time. Shinada et al. reported that about 15% of caregivers experienced depression and complicated grief after the death of a patient due to acute cardiovascular disease, which was also encompassed by symptom palliation of quality indicators. The support decision-making process is imperative for current palliative care in cardiovascular intensive care. As described above, decision-making in cardiovascular intensive care is more complex and time limited compared to other healthcare settings. There are two important essences of supporting the decision-making process [bib_ref] The experiences and perspectives of family surrogate decision-makers: a systematic review of..., Su [/bib_ref]. First, we should determine the surrogate decisionmaker with whom we can discuss patients' treatment strategy. Ideally, we should discuss this with the patients themselves. However, in cardiovascular intensive care settings, patients often cannot communicate with healthcare providers due to loss of consciousness or cognitive dysfunction [bib_ref] Assessment of decision-making capacity in older adults: an emerging area of practice..., Moye [/bib_ref]. In this situation, we should determine and document who should be approached about the patients' care or carefully discuss with patients and surrogate decision makers through advance care planning [bib_ref] Definition and recommendations for advance care planning: an international consensus supported by..., Rietjens [/bib_ref]. Of note, any decision making should be toward the patient himself. Second, we should acknowledge that many decisions making in cardiovascular intensive care depends not on the best available evidence but on patients and their families' preferences, which is associated with ethical dilemmas. Furthermore, not only patients but also healthcare providers cannot predict patient prognosis, which results in decision-making under uncertainty. Conquering these dilemmas, several ethical topics should be prepared in the field of cardiovascular intensive care, including withholding and withdrawal of the treatment, especially life-sustaining therapy and terminal/palliative sedation. A previous systematic review revealed that most of the recommendations referred to withholding and withdrawal of life-sustaining therapy, and these two were considered as morally equivalent and permissible [bib_ref] Ethical content of expert recommendations for end-of-life decision-making in intensive care units:..., Spoljar [/bib_ref]. Although both withholding and withdrawal of life-sustaining therapy have increased recently, withholding of treatment is perceived as less difficult even if the end result is the same. The difficulty in withdrawing life-sustaining therapy is often explained by a status-quo bias. The decisions to maintain the status quo tend to be regretted less than decisions to change, which could best apply to the status, where the patient has already been intubated and needs to discuss the withdrawal [bib_ref] The status quo bias and decisions to withdraw life-sustaining treatment, Breslin [/bib_ref]. A surrogate decision-maker should be encouraged to share the decision with others that could reduce their responsibility and minimize the risk of being blamed. Finally, continuous deep sedation, which is sometimes referred to as terminal or palliative sedation, is also an important ethical and a sensitive topic. Previous cardiovascular intensive care quality indicators did not include this specific topic; however, it is important for many clinical professionals. Continuous deep sedation could sometimes be confused with euthanasia by non-healthcare professionals, which could result in understanding gaps between healthcare professionals and non-healthcare professionals [bib_ref] Consensus guidelines on analgesia and sedation in dying intensive care unit patients, Hawryluck [/bib_ref]. We should recognize that the intent and documentation of the intent of physicians and other healthcare providers are important after careful discussions.
Finally, as described above, many patients and families frequently do not have any experiences about advance care planning, lack medical background knowledge, and are under time-restricted situation in intensive care unit. To support this, several decision aids and programs support surrogate decision maker have been developed [bib_ref] Effectiveness of an intervention supporting shared decision making for destination therapy left..., Allen [/bib_ref] [bib_ref] Effect of ethics consultations on nonbeneficial life-sustaining treatments in the intensive care..., Schneiderman [/bib_ref]. As the effectiveness of these educational materials and programs are still controversial [bib_ref] A randomized trial of a family-support intervention in intensive care units, White [/bib_ref] , future trials would be necessary to implement patient and family side education for an ideal decision making. More importantly, getting familiar with these ethical topics and noticing the presence of ethical dilemmas could be essential for implementing palliative care for cardiovascular intensive care.
## How to use quality indicators of palliative care for cardiovascular intensive care
Theoretically, there are two major motivations for using quality indicators. The first is for a quality assurance system as a summative mechanism for external accountability and verification. Pay for quality of care, such as pay for performance programs, is categorized into this. The second is a formative mechanism to improve quality. Internal audit and feedback for continuous improvement at the hospital level can be categorized into a formative mechanism. For quality assurance and accountability, high-level precision and advanced statistical techniques are required; otherwise, providers will resist the usage of the quality indicator itself and its potential consequences, such as certification issues. Unfortunately, there is still no sufficient evidence and high-quality randomized control trials to prove the validity and importance of quality indicators in cardiovascular intensive care. Even in the cancer field, many quality indicators of palliative care have not been well accepted for accountability purposes, which implies that many quality metrics are too difficult to monitor or do not directly reflect the quality of care in each department and hospital. Only a few of them, mainly structure indicators, such as specialist allocation, are valid and easily measured.
Considering these usage limitations in palliative care quality indicators for quality assurance, some other insights are needed to use quality indicators [fig_ref] Figure 3: Scheme of development and usage of quality indicators [/fig_ref]. First, it is important to know that the quality indicators are derived not only from the evidence but also from expert opinions, policy priorities, regulations, and ethical standpoints. Quality indicators made in each target disease or field and the domain sorted by the SPO framework are usually useful for understanding the essence of a specific field. After recognizing these backgrounds, as discussed, quality indicators and domains would be useful for learning the minimum requirements for quality improvement. Especially in cardiovascular intensive care, palliative care is yet to be considered a common practice, and could be useful for many beginners to understand the minimum requirements for palliative care in the field of cardiovascular intensive care. Furthermore, after monitoring each quality indicator, structural and behavioral changes are necessary to improve palliative care quality. There are several barriers to implementing palliative care, especially in intensive care units. Design modifications or behavioral scientific approaches will be necessary to modify the structure and care process in the intensive care unit [bib_ref] A behavioral blueprint for improving health care policy, Loewenstein [/bib_ref]. Using behavioral insights and implementation strategies would be helpful to tackle the evidencepractice and quality indicator-practice gap . These processes of making quality indicators and using quality indicators for quality improvement are continuous and repetitive cycles similar to the Plan-Do-Check-Action cycle to create new quality indicators.
Finally, we should acknowledge that there are several limitations in the use of quality indicators and speculations on how to use indicators effectively. As described above, quality indicators of palliative care in cardiovascular intensive care have not been fully evaluated and have not been validated with patient outcomes. The domain and subcategories could also be just hypothetical contrasts, and we could not deny arbitrariness. The following strategies could be realistically implemented to use In addition, quality indicators are educational tools to help people understand the nature of a particular field. Quality indicators are also used to monitor the quality of care in hospitals and departments. Structural/behavioral changes after monitoring and education lead to quality improvement. These quality improvement processes are a continuous and iterative cycle. SPO Structure-Process-Outcome quality indicators of palliative care for cardiovascular intensive care from a clinical perspective. First, we should check the structure indicators, including workflow, policy, and resources about palliative care in each healthcare service line, such as hospitals and departments. Second, palliative care needs and symptoms should be screened at the first encounter and then repeatedly. Finally, we should integrate multidisciplinary palliative care teams and other specialists when necessary, mainly reflected by process indicators.
# Conclusion
Although there are limited numbers and limited usage of quality indicators of palliative care for cardiovascular intensive care compared with other fields, quality indicators could be used as educational tools to implement palliative care and the fundamental structure for future discussion. To implement palliative care for cardiovascular intensive care, we should learn several basic concepts of palliative care through quality indicators and monitor indicators to see, where we stand now. With the increasing demand for palliative care for every patient, further high-quality evidence and valid quality indicators are warranted.
## Supplementary information
The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s40560-022-00607-6.
Additional file 1. Initial search for quality indicators of palliative care in heart disease.
Additional file 2: [fig_ref] Table 1: In addition, details of each quality indicators are shown in Additional file... [/fig_ref]. Comparison of quality indicators for palliative care.
[fig] Figure 3: Scheme of development and usage of quality indicators. Quality indicators are based on expert opinion, evidence, political priorities, regulations, and ethical positions. [/fig]
[table] Table 1: In addition, details of each quality indicators are shown in Additional file 2: Table S1. Hamatani et al. listed 35 quality indicators for palliative care in patients with heart problems and Mizuno et al. made 21 quality indicators for acute cardiovascular disease [/table]
|
Using behavior change communication to lead a comprehensive family planning program: the Nigerian Urban Reproductive Health Initiative
Greater exposure to a comprehensive family planning program in urban Nigeria that emphasized demand generation and communication theory was associated with improved ideation among women (their beliefs, ideas, and feelings about family planning), and more positive ideation was associated with greater contraceptive use, especially among the poor. Improving providers' knowledge, attitudes, and skills was also key. By the end of the observation period, outreach through mobile service delivery contributed nearly one-half of the project clinics' family planning services.ABSTRACTBackground: The Nigerian Urban Reproductive Health Initiative (NURHI), a 6-year comprehensive family planning program (2009)(2010)(2011)(2012)(2013)(2014)(2015) in 4 cities, intentionally applies communication theories to all program elements, not just the demand generation ones, relying mainly on a theory called ideation-the concept that contraceptive use is influenced by people's beliefs, ideas, and feelings and that changing these ideational factors can change people's behavior.Program Description: The project used multiple communication channels to foster dialogue about family planning, increase social approval for it, and improve accurate knowledge about contraceptives. Mobile service delivery was started in the third year to improve access to clinical methods in slums.Methods: Data from representative baseline (2010-11) and midterm (2012) surveys of women of reproductive age in the project cities were analyzed. We also used propensity score matching to create a statistically equivalent control group of women not exposed to project activities, and we examined service delivery data from NURHI-supported clinics (January 2011-May 2013) to determine the contribution of mobile services to total family planning services.Results: Three years into the initiative, analysis of longitudinal data shows that use of modern contraceptives has increased in each city, varying from 2.3 to 15.5 percentage points, and that the observed increases were predicted by exposure to NURHI activities. Of note is that modern method use increased substantially among the poorest wealth quintiles in project cities, on average, by 8.4 percentage points. The more project activities women were exposed to, the greater their contraceptive use. For example, among women not using a modern method at baseline, contraceptive prevalence among those with no exposure by midterm was 19.1% vs. 43.4% among those with high exposure. Project exposure had a positive dose-response relationship with ideation, as did ideation and contraceptive use. By the end of the observation period, mobile services were contributing nearly 50% of total family planning services provided through NURHI-supported clinics. Propensity score matching found that the increase in contraceptive use in the 4 cities attributable to project exposure was 9.9 percentage points. Intention to use family planning in the next 12 months also increased by 7.5 to 10.2 percentage points across the 4 cities.Conclusion: Demand-led family planning programs, in which demand generation is the driving force behind the design rather than the conventional, service delivery-oriented approach, may be more suitable in places where expressed demand for contraceptives is low.
# Introduction
W ith a population of 169 million, Nigeria has some of the poorest measures of reproductive health in Africa, including an estimated maternal mortality ratio of 630 deaths per 100,000 live births and an infant mortality rate of 69 per 1,000 live births.The Government of Nigeria has committed to improving these indicators as part of the Millennium Development Goals (MDGs). For MDG 5 (improve maternal health), a pillar of achievement is increasing the contraceptive prevalence rate (CPR), a core driver of maternal and reproductive health.The Nigerian Urban Reproductive Health Initiative (NURHI), a comprehensive family planning program encompassing supply, demand, and advocacy interventions, aims to increase voluntary use of contraceptives by [bib_ref] Mass media, ideation, and behavior: a longitudinal analysis of contraceptive change in..., Kincaid [/bib_ref] project. MLE has conducted baseline and midterm surveys to measure the impact of NURHI, and a final evaluation will be available in early 2015. For more information on the initiatives in India, Kenya, and Senegal, see the MLE website (www. urbanreproductivehealth.org), which is designed to share the learning from these programs.
The project's activities, which included performance improvement at facilities, training providers in contraceptive provision, and ensuring efficient and effective commodity logistics systems, will be familiar to anyone who has designed and implemented a comprehensive family planning program; what NURHI has done differently than most programs is to use communication methodologies to adapt each activity-even the service delivery ones-and to put serious and sustained effort and resources into demand generation activities.
The NURHI initiative was designed based on the hypothesis that when demand for family planning rises, supply will rise to meet the demand over time. NURHI defines demand for family planning as the desire and ability among women and/or men to take action to plan their families. Our hypothesis does not imply that one can leave the supply side to itself; it simply reframes the often unstated but very real assumption built into some large-scale family planning programs that if you build it, they will come.
Creating demand for family planning was clearly a priority in Nigeria, as just a fraction of women were articulating a desire and need for family planning. For example, in the 2008 Nigeria Demographic and Health Survey,the national CPR for modern methods was 10.5% (with less than 2 percentage points of growth since the 1999 surveyand 39% of women cited opposition to contraceptive use, but 20% had an unmet need and 21% intended to use contraception in the future. While designing the NURHI project, the Bill & Melinda Gates Foundation shared with program designers an overview of its 2008 reproductive health strategy, in which it estimated through its own calculations that ''demand issues comprise 70% of the problem and, therefore, are an even larger driver for achievement of our goals.'' Furthermore, there are adequate sources of short-acting contraceptive methods in parts of the country, including through the nonprofit health sector (the public sector and nongovernmental organizations) and through a robust and entrepreneurial for-profit health care sector that includes patent medicine vendors (who serve as frontline health care providers for a large percentage of Nigerians), pharmacists, and an array of small to large-scale health facilities (general and maternity clinics and hospitals). Interestingly, the majority of family planning users in Nigeria already purchase their contraceptives from the nonclinical private sector.In NURHI's urban sites, the primary issue was thus not a lack of sites that could provide family planning services, but rather that no one was asking for them. This was supported by the baseline survey, which found that under 1% of women in the 4 intervention cities cited cost, distance, or access as a reason for not using family planning.The purpose of this article is to describe the activities designed and implemented by NURHI to meet the project's stated objectives and to illustrate how having a demand lens influenced programming decisions in ways that other family planning programs probably would not have considered. We also present findings about the project's outcomes at midterm, primarily from the MLE evaluation surveys.
## Intervention design and components
In this section, we introduce the theoretical foundation that underpins the NURHI project and describe in some detail each component of the project, including its formative research, demand generation activities, service delivery interventions, and advocacy activities. The research, strategies, and materials used to design and implement NURHI can be found at www.nurhitoolkit.org.
## Theoretical foundation
NURHI's overall design and strategy are driven by the project's theory of change. NURHI understands the barriers to contraceptive use in its intervention cities to be primarily ones of knowledge, attitudes, and social norms, and the causal pathway to improve the CPR is through changes in these factors at each level of society, from the individual up through communities, service sectors, and the policy environment. Communication is the driver of this change at every level, from demand creation at the individual level, to supportive supervision and training in interpersonal communication at the provider level, to advocacy at the policy level.
In developing strategies for demand generation, service delivery interventions, and advocacy, NURHI has made use of a communication theory called ideation. Ideation is the concept that people's actions are influenced strongly by their beliefs, ideas, and feelings (''ideational factors'') and that changing them can change behavior, including contraceptive behavior [fig_ref] FIGURE 1: Ideation Model of CommunicationSource [/fig_ref]. 7 Some of these ideational factors are personal, such as what a person knows about family planning and how they think it will affect them. Others reflect social norms, such as what people believe other people will think of them if they use family planning. The more positive ideational factors a person holds, the greater the likelihood the person will adopt the desired behavior.
While ideation has often been applied to designing demand generation interventions in family planning, we also applied the basic ideas of ideation to designing our service provision activities. We examined service providers' ideas, beliefs, and feelings about family planning, and adapted service delivery interventions to address them. For example, among the ''ideas and feelings'' that service providers hold is the common belief that women should not use family planning if they only have 1 or 2 children or if they are young, beliefs that pose a barrier to quality family planning provision. Ideation also includes knowledge, which, for a service provider, would encompass skills in clinical care.
## Formative research
Reflecting our theory of change, we designed the formative research to explore potential barriers to contraceptive use related to knowledge, attitudes, and social norms, and to pay specific attention to ideational factors in both qualitative and quantitative research. The full set of formative research included a household baseline survey with men and women; focus group discussions with men and women (contraceptive users and non-users) and family planning providers of different cadres; a facility assessment survey; and a family planning social mapping survey in 3 of the project cities.
The NURHI project team worked with the MLE evaluation team to tailor the instruments used in the baseline and midterm surveys to measure indicators important to this communicationinfused program. Questions were tailored to measure ideational factors (partner communication, beliefs and attitudes, correct knowledge, perceptions of peer support, self-efficacy, and perceptions of religious approval), which, taken together, are used as an index predictive of contraceptive use; changing these factors contributes to increased contraceptive use. We also tailored the evaluation to add a baseline survey of men, because although men are not family planning clients for most methods, they are integral to the decision-making process and NURHI needed information about their beliefs, needs, and desires.
Qualitative research with users and nonusers as well as with service providers was used to complement the baseline survey findings. The research with the providers uncovered and described the biases they held against certain types of clients and methods. Focus groups with men and women explored their beliefs, motivations, fears, and perceptions of use and non-use. A number of important findings emerged from these quantitative and qualitative research methods (Box). Taken together, we concluded that a major barrier to contraceptive use in the project cities was fear and bias. NURHI's interpretation of these data is that people approve of family planning as a concept but believe individual methods are risky. Service providers believe it is their role to uphold social norms around family size, marriage, and spousal consent. These are major challenges, but ones that can be addressed using communication approaches-by encouraging people to talk about family planning and helping to make it a normal part of life, by providing accurate information about the safety of contraceptive methods, and by helping providers use their clinical knowledge, rather than their personal values, in the counseling room. NURHI designed the project's demand generation, service delivery, and advocacy interventions to achieve these goals [fig_ref] FIGURE 2: Nigerian Urban Reproductive Health Initiative [/fig_ref].
## Demand generation
NURHI's demand generation strategy for women and men focuses on demedicalizing and demystifying the practice of family planning, including fostering dialogue around family planning-in the home, on the street, at work, in the clinic, in the media; increasing understanding, appreciation, and social approval for planning one's family; improving knowledge and perceptions of family planning methods; and reinforcing existing contraceptive use and reducing discontinuation.The initial strategy included a 3-phase approach: Phase 1 was designed to increase access In urban Nigeria, people approve of family planning as a concept but believe particular methods are risky.
## Box. formative research findings guide program design
The baseline survey provided the following essential information that guided the design of program interventions:
N Contraceptive prevalence was low in the 4 project cities, ranging from 19.6% in Kaduna to 33.3% in Ibadan. The majority of modern contraceptive users were using short-acting methods; for example, in Ibadan only 5.4% of married women used 1 of 3 long-acting or permanent methods available (sterilization, IUDs, implants).
N Most women cited no intention to use family planning in the next year (for example, in Ibadan only 7.5% of non-users intended to use contraceptives in the future). The main reasons women gave for not using contraception related to either being pregnant or wanting to be pregnant (36.7% of women in Abuja fit this profile) or having no/infrequent sex (36.2% in Abuja). This indicated to NURHI that a major challenge was to help people think about the benefits to spacing their children and planning their families.
N Fear of specific methods and misconceptions about their side effects was a major non-fertility related reason for not using contraception (in Kaduna, for example, 13.8% of women said they did not use contraceptives due to fear of side effects). The majority of women and men stated that they approved of family planning as a practice, but they held fearful and negative views of actual available methods.
N Despite high levels of awareness of contraceptives (over 90% of women knew of at least 1 method and where to get it), there was limited knowledge of clinical methods (IUDs, implants, and sterilization). Qualitative research showed that what ''knowledge'' there was of these methods was generally based on myths and misconceptions and contributed to fear of these methods.
N Of women who were using a method, most were using short-acting methods through private-sector pharmacies and drug shops rather than clinical methods from clinics or hospitals with a trained provider. This point dovetails with those above: Women did not know of clinical methods and what they did ''know'' was negative. Furthermore, women may not have seen access as a barrier because they were not trying to access clinical methods, where services may not be readily available.
N Women and men did not report discussing family planning, contraceptives, or their desired number of children.
Spousal discussion is strongly predictive of family planning use, [bib_ref] Spousal communication and contraceptive use in rural Nepal: an event history analysis, Link [/bib_ref] and so lack of discussion is a barrier.
N While women said religious approval was important, and some felt that their religion did not approve of family planning, a majority of women believed that they could use a contraceptive method despite religious disapproval, a surprising finding. Gender preference was also prevalent but not predictive of non-use, another surprising finding that shaped program interventions.
Qualitative research provided in-depth understanding of the barriers to family planning use:
N Focus groups with service providers showed that the providers had biases and myths and misconceptions about family planning that reflected those of the larger city populations. In particular, service providers believed women should have many children and should not use contraceptives to space them until they have already had a large family. They also disliked providing services to young women, unmarried women, and women with few children. The following quotes from in-depth interviews with service providers illustrate these medical barriers 10 :
We do not provide family planning to unmarried young girls because it can make them promiscuous.
-Female, 24 years old, middle income, head of nursing at a private clinic in Kaduna I don't like attending to youth because of their involvement in what they are not due for. Also, I don't like attending to the unmarried people.
-Female, 18-29 years old, owner of patent medicine store in a slum in Ibadan Furthermore, sometimes providers perpetuate biases and myths about contraception to potential family planning users. One 21-year-old married woman from Ibadan with 1 child said, ''The advice given to us in the hospital is that the IUD is risky.'' N A survey of service providers at clinics and hospitals showed that they restricted access to certain methods based on a woman's age, marital status, or parity rather than on medical eligibility. For example, 48% of providers restricted to basic family planning information and to heighten awareness of family planning; Phase 2 meant to deepen understanding, discussion, and exploration around the concept of family planning and about specific methods; and Phase 3 was designed to increase the level and localization of communication efforts but was subsequently rolled in with Phase 2 based on timing issues. From the outset, NURHI's demand side activities have been orchestrated to mutually reinforce one another, in addition to being closely integrated with the service delivery and advocacy objectives. We use multiple communication channels, based on the theory that communication interventions have a synergistic impact, so that hearing or seeing messaging through more than one medium has more impact than hearing or seeing messaging in just one way. Furthermore, the communication activities operate at different levels of the socioecological environment, from the individual up through the community and to the policy environment, with messaging designed to address essential cognitive, emotional, and normative ideational factors. The main NURHI demand generation activities consist of mass media, entertainment-education, social mobilization, and integrated branding with a memorable, colorful puzzle logo and tagline that helps tie all program activities together under one identity. The tagline is ''Know. Talk. Go.'', meaning ''know'' your family planning options, ''talk'' to your partner, and ''go'' for services.
## Mass media
A media campaign featuring the overarching puzzle logo and the ''Know. Talk. Go.'' tagline uses radio and TV spots and print materials (eg, posters, umbrellas, flyers, t-shirts) to get the word out. Some of the scenes or materials illustrate partner communication; others show barbers or hairdressers discussing family planning with their friends in an open and easy way or couples going to clinics for services, allowing NURHI to model healthy, happy family planning users. For example, in one spot, a couple gets the happy news they are expecting a second child soon after stopping their contraceptive method, access to injectables if they felt a woman has not had enough children; 60% restricted access to IUDs if a woman is not married; and 30% restricted access to pills without spousal consent. 11 N Focus groups with women and men found that a major barrier to family planning use was the need for women to obtain their husband's permission to use family planning, but women found it difficult to start a conversation about family planning with their husbands. Both men and women, in general, approved of planning one's family. However, men felt it was the women's responsibility to begin the family planning discussion, and women felt it was the men's responsibility, and so the conversation was not happening. N Women and men described the need to plan a family as a way to ensure one had only the number of children one could ''cater for,'' meaning feed, clothe, house, educate, and love. People saw children as a blessing and a gift but also as a great responsibility; they described this responsibility as the reason for supporting family planning. A 24-year-old married woman from Ibadan with 2 children and middle socioeconomic status explained 12 :
Having too many children is not good. Everyone knows his capacity, and I think it is necessary to limit your childbirth to what your capacity can take you. God will not come down from heaven to help.
N Aside from condoms, people viewed contraceptives as highly medical, requiring medical tests and a perfect fit with one's anatomy. Hormonal and clinical methods were seen as risky-more risky than giving birth to many children. A 30-year-old married man from Ibadan with 1 child and middle socioeconomic status described this fear, which was rooted in misconceptions about contraception 11 :
I will advise her [his wife] not to do it. Family planning is very dangerous to a person's health. Great caution needs to be exercised.
refuting the myth of impaired fertility with contraceptive use.
## Entertainment-education
A 30-minute weekly radio magazine program (a radio show with various magazine elements, such as listener interviews and call-in ''ask the expert'' segments) was also produced and broadcast in each project city. These programs include additional content about contraceptive methods, they address myths and misconceptions, and they model discussion of family planning between spouses and with providers. In the initial plans, NURHI had expected to produce one program that could be translated for each location. However, to fully localize it to the specific city context, ultimately a unique program was designed for each site although the format remained consistent. The second phase of the radio programs integrates a live call-in component, with a quiz and an opportunity to ask questions to an ''onair'' expert. Radio listening groups, formed within the city environment, are convened on a weekly basis by social mobilizers to listen to and discuss the content of the programs, thus deepening the dialogue, reflection, and understanding of family planning.
## Social mobilization
NURHI social mobilizers were chosen not for their expertise in health but for their expertise in talking to people and making connections in their slum communities. In Nigeria, these are hairdressers, barbers, and tailors. Working through professional associations and community-based organizations, NURHI recruited mobilizers from these professions, trained them in family planning, and equipped them with materials, including ''Know. Talk. Go.'' referral cards. In addition to leading the radio listening groups, they talk to their clientele in their shops about family planning, mobilize clients for family planning outreach services, and discuss family planning at key life events, such as graduations and naming ceremonies. They are now widely sought out by community members, as their participation is considered highly prestigious. This focus in the community has been critical to personalizing the agenda, making family planning a socially acceptable topic, and providing a bridge between the community members and the services. The mobilizers are not paid, which is both a strength and an ongoing challenge for retention and commitment. Recognition of their role and contribution to the well-being of others in the community has inspired many of the mobilizers to continue with the work.
## Service delivery
NURHI's service delivery component is based on best practices in service integration and quality improvement, [bib_ref] Elements of success in family planning programming, Richey [/bib_ref] but with the added dimension of treating service providers as an audience for behavior change. The formative research identified key biases among providers related to their attitudes toward family planning, the provision of specific methods, and the women who seek services. Many providers lacked basic family planning knowledge and, in many instances, the technical competency to provide particular services. As NURHI launched and program staff spent time in clinical facilities, it also became apparent that the decrepit family planning facilities (lack of privacy for clients, lack of running water, leaking roofs, dirty floors and walls) were not just an issue of hygiene or safety; they also indicated to providers how little family planning mattered to hospital administrators, which was demotivating to staff.
The issue of decrepit facilities is illustrative of how we approached a supply intervention (renovating the facilities) with a demand lens. NURHI viewed the decrepit facilities as an indicator of the ideas and feelings (the ideation) of stakeholders, policy makers, the larger community, and service providers, specifically that they lacked motivation for and did not value family planning. The solution therefore involved advocacy with local stakeholders, engagement with facility administrators, participation of facility staff in the renovation process (coined as the ''72-Hour Clinic Makeover''), and a launch of a ''new and improved'' family planning facility that built support for the providers in their community. It is important to note that the facility renovations generally entailed a coat of fresh paint, scrubbing, connecting a sink to the hospital's water line, and making sure contraceptive commodities and equipment were on hand-not, in most cases, major costs or construction.
NURHI also applies a demand lens to improving counseling sessions between providers and clients, which we consider to be critical episodes of interpersonal communication. We have ensured that provider training sessions include ample time and priority for interpersonal communication and counseling. In addition, we have made sure that providers have the tools they need to counsel their clients well to provide voluntary, free choice of methods, and we have developed those materials to seamlessly integrate with demand generation outside the clinic walls. For example, counseling materials and job aids were part of the overall NURHI communication approach, with consistent branding, creative A poster produced by the NURHI project for the ''Get It Together'' campaign encourages partners to discuss family planning together.
''72-Hour Clinic Makeovers'' not only improved facility conditions but also engaged facility staff to value family planning.
Using communication to lead a comprehensive family planning program www.ghspjournal.org approach, and messaging so that both clients and providers would associate what happens inside the clinic with the television, radio, and interpersonal communication they were exposed to in the community. Finally, our selection of service sites for the project's clinical interventions was informed by communication approaches by first considering, through the baseline household survey, women's preferences, needs, and behaviors regarding where they access health services. We matched that input with sites with a high volume of clients and also asked women where they spent time in their community, so we had an idea of where we could reach women outside of the clinic. Using this information, we selected service delivery points for clinical quality improvement, almost all of which were public-sector facilities where clinical services were available, plus a broad network of mostly private, nonclinical service delivery sites where providers expressed interest in family planning. We connected all of these providers through a new branded network called the ''Family Planning Providers Network'' to ensure access to the full basket of services in every project city. We also adopted and adapted a clinical service outreach model to fill a gap in clinical services in slum neighborhoods, in which we provide services on advertised days in tents in markets and in small health posts with no regular family planning providers. These outreach visits are linked to our demand generation work, by using social mobilizers to promote the outreach events and make referrals to them.
## Advocacy
In NURHI's view, policy makers and traditional and religious leaders-as well as service providers-are important audiences in need of communication and ''demand generation'' just as much as the general public. The difference is in the kind of information they need and in how they can access that information.
Baseline research showed that hearing a religious leader voice support for family planning was an important ideational factor for women and men in Nigeria.NURHI enlists prominent leaders of multiple faiths to speak publicly and in the media about family planning. The project also developed advocacy kits for each city's policy makers, many of whom are motivated to make progress toward the MDGs, that synthesized baseline data at the city level and highlighted MDGrelated trends and how family planning could impact them. In each city, we also formed advocacy groups that included all interested partners working to improve family planning, and these groups oversaw the development and use of the advocacy kits and took ownership for progress.
While funds for family planning are often allocated in State and Local Government Area budgets, the funds are often not released so that family planning coordinators and facilities can actually use them. Through intensive communication and mentoring, NURHI staff coached government staff in the intricacies of local-level budgeting, requesting processes, and spending decisions, with the result that modest amounts of funding began flowing in many Local Government Areas, where the funds had been previously ''stuck.''
# Methods
Data on the outcomes of the NURHI project come primarily from analysis of the MLE baseline and midterm surveys. The surveys are representative of men and women of reproductive age in each NURHI project city. The same women were interviewed for the baseline and midterm surveys, providing a unique longitudinal sample in which sophisticated analytical techniques could be applied to have greater confidence that results of the project could be attributed to exposure to the interventions. [bib_ref] Demand generation activities and modern contraceptive use in urban areas of four..., Speizer [/bib_ref] women and men was conducted in 2010-2011 and the midterm survey of women in 2012.We also conducted additional analysis of the MLE data using a technique called propensity score matching (PSM). This technique allowed us to estimate the probability (propensity) that a woman will be exposed to the program activities and to create an unexposed control group that is statistically equivalent to those exposed. Using PSM, we estimated what the CPR would have been among the women exposed to the NURHI project had they not been exposed to it. The difference between the CPR of the women exposed to the NURHI project and the estimated CPR of those same women had they not been exposed is considered the treatment effect of the intervention.
Finally, we performed secondary analysis of the MLE data to determine whether there was a positive relationship between communication activities and ideation (factors such as beliefs, spousal discussion, perceived peer behavior, perceived self-efficacy, and personal advocacy), and whether there was a positive relationship between ideation and contraceptive use. Specifically, 32 ideational items were measured across 3 domains: cognitive, emotional, and social interaction. The items included aspects of contraceptive awareness (12 items), myths and rumors about contraceptives (8 items), perceived self-efficacy to take action regarding contraceptive use (7 items), and approval of leaders talking about family planning (2 items), as well as descriptive norms about contraceptive use in one's community, personal advocacy for family planning, and perceived social support for personal use of contraceptives. The resulting scores were then categorized into quintiles denoting women's overall level of ideation: very low (8 items or fewer), low (9-10 items), medium (11-12 items), high (13-15 items), or very high (16 or more items).
In addition to exploring the effect of demand generation activities on contraceptive use, we also examined service delivery data from NURHIsupported clinics between January 2011 through May 2013 to determine the proportion of clinicprovided family planning services attributed to outreach visits.
# Results
## Program exposure
The MLE midterm survey measured people's exposure to various NURHI messages and strategies and computed overall exposure to the NURHI program by summing up people's exposure to multiple items. Overall program exposure is presented in 4 categories:
1. No exposure 2. Low (knew of 1 or 2 NURHI activities) 3. Medium (knew 3-6 activities) 4. High (knew 7 or more activities) About 80% of women in the 4 project cities reported some exposure to the NURHI project: 24% reported low exposure, 32% medium, and 25% high, with the remaining 19% reporting no exposure.
## Myths and misconceptions
Between baseline and midterm, the percentage of women who believed in myths or had misconceptions about contraception declined. For example, the percentage of women who believed incorrectly that ''contraceptives are dangerous to your health'' dropped by 17 percentage points in Ilorin, from 37.4% to 20.4 %, and by about 15 percentage points in Ibadan, from 57.1% to 42.2%.Similarly, the percentage who believed that ''contraceptives can harm your womb'' decreased by 15.7 percentage points in Ilorin, from 33.6% to 17.9%; by 12.5 percentage points in Ibadan, from 49.8% to 37.3%; and by 9 percentage points in Abuja, from 33.4% to 24.1%.
## Intention to use contraception
In each project city, there was a significant upward trend in the percentage of women intending to use contraception. For example, in Abuja and Ibadan, the percentage of women who intended to use contraception in the next 12 months increased significantly by 10 percentage points in each city, from 13.9% to 23.5% in Abuja and from 7.5% to 17.7% in Ibadan [fig_ref] FIGURE 3: Percentage [/fig_ref]. In Ilorin and Kaduna, intention to use increased significantly by nearly 8 percentage points in each city.
## Contraceptive use at baseline and midterm
Between baseline and midterm, use of modern methods among married women increased in each city, although the change varied widely between the 4 cities [fig_ref] TABLE 1: Modern Contraceptive Prevalence Rate Among Married Women, at Baseline and Midterm, by... [/fig_ref]. For example, in Abuja, 31.9% of married women were using modern contraception at baseline; the percentage increased slightly at midterm to 34.2%, but the change was not statistically significant. On the other hand, in Kaduna, the modern CPR increased Intention to use contraception in the future increased in each project city.
Women's ideation scores were based on 32 variables across cognitive, emotional, and social interaction domains.
Using communication to lead a comprehensive family planning program www.ghspjournal.org
Global Health: Science and Practice 2014 | Volume 2 | Number 4 by 15.5 percentage points between baseline and midterm, from 19.6% to 35.1% (P , .001). One factor in those differences is the difference in the modern CPR at baseline between the cities. In Kaduna, for example, the low level of contraceptive use (19.6%) at the start of the project may have represented pent-up need for access to family planning services, resulting in the notable improvement at midterm. The cities differ demographically, politically, culturally, and religiously, and these factors may also have contributed to the different results in each city. It is interesting to note, however, that the modern CPR in the 4 cities at midterm is similar (between 34.2% and 36.9%), whereas the rates were more variable at baseline (between 19.6% and 33.3%). In addition, of note is that the modern CPR increased substantially among the poorest wealth quintiles in NURHI project cities, on average, by 8.4 percentage points. 17
## Contraceptive use by level of exposure to the nurhi program
Longitudinal data from the MLE baseline and midterm surveys show that (reported) exposure to several of the NURHI communication interventions was significantly associated with higher levels of contraceptive use. The greatest effects were associated with exposure to the locallanguage radio entertainment-education programs, social mobilization activities, and television spots. [bib_ref] Demand generation activities and modern contraceptive use in urban areas of four..., Speizer [/bib_ref] Analysis of CPR data by women's reported level of exposure to NURHI project activities shows that, among married women not using a modern method at baseline, 19.1% were using contraception at midterm among those reporting no exposure to NURHI activities compared with 32.1% among those with low exposure . Contraceptive prevalence increased positively and linearly with greater exposure (medium exposure, 34.6%; high exposure, 43.4%).
We used propensity score matching to better understand whether changes in behavior (contraceptive use) were attributable to exposure to NURHI's demand generation activities. This analysis showed that the CPR among the matched control group would have been 25.9% had the women not been exposed to the NURHI program, compared with the actual (observed) CPR of 35.8%. These data suggest that the increase in contraceptive use (ie, the treatment effect) attributed to exposure to the program was 9.9 percentage points.
## Ideational factors and contraceptive use
Analysis of longitudinal data from the baseline and midterm surveys also finds that 9 of 10 measured ideational factors increased significantly. For instance, the percentage of women who perceived there was peer support for family planning increased significantly from 22.8% to 42.4% (P , .001) between baseline and midterm [fig_ref] TABLE 2: Ten Ideation Factors at Baseline and Midterm That Predict Contraceptive Use [/fig_ref]. Similarly, the percentage of women who had positive attitudes toward family planning rose from 53.7% to 70.9% (P , .001). The data also show that level of exposure to program activities had a positive dose-response relationship with these ideational factors. For example, the percentage of women who perceived there was peer support for family planning increased by 6.2 percentage points among women reporting no program exposure, and the percen-tage increased significantly and linearly with each level of exposure: from a 17.9 percentage point increase among women with low exposure to a 26.7 percentage point increase among women with high exposure [fig_ref] FIGURE 5: Change in Perceived Peer Support for Family Planning Between Baseline and Midterm,... [/fig_ref].
Furthermore, analysis of CPR data by women's level of ideation shows that the more positive ideational factors that women had, the greater their contraceptive use. Among women not using a modern method at baseline, 15.9% of those with very low ideation at midterm were using contraception compared with 28.2% of those with medium ideation and 47.3% of those with very high ideation [fig_ref] FIGURE 6: Contraceptive Prevalence at Midterm Among Married Women Who Were Not Using a... [/fig_ref]. [fig_ref] Figure 7: shows the contribution of clinical outreach to the total number of clients... [/fig_ref]. In 2012, when NURHI started conducting outreach visits, the outreach visits contributed, on average, about 15% of these total family planning users, and the share increased to 31% in the fourth year of the project. At the end of the observation period, outreach visits were contributing nearly half of total clinical services supported by NURHI. Note that the full number of contraceptive users is better represented by data from the midterm survey since women access many sources for family planning, including private-sector providers such as pharmacists and drug shop owners.
## Contribution of clinical outreach
# Discussion
While changes to the contraceptive prevalence rate in the cities where NURHI works have been variable, sophisticated analysis of the longitudinal data indicate that NURHI's demand generation activities are indeed significantly associated with increased contraceptive use in the cities. In particular, the data support the theoretical foundation on which NURHI was based-that is, the communication theory that holds that changing ideational factors, such as knowledge, attitudes, and beliefs, increases the chances of changing people's behavior. Women's use of contraception at midterm increased linearly with increasing levels of ideation. Similar findings have been reported in Bangladesh, Burkina Faso, and the Philippines. [bib_ref] Social networks, ideation, and contraceptive behavior in Bangladesh: a longitudinal analysis, Kincaid [/bib_ref] [bib_ref] Communication, ideation, and contraceptive use in Burkina Faso: an application of the..., Babalola [/bib_ref] [bib_ref] Mass media, ideation, and behavior: a longitudinal analysis of contraceptive change in..., Kincaid [/bib_ref] Ideational factors can be thought of as positive risk factors, similar to how certain behaviors are risk factors for disease. For example, just as obesity, diet, exercise, and genetics are all risk factors for heart disease, ideational factors are ''risk factors'' for the positive behavior of family planning. And just as with risk factors for heart disease, the more factors a person has, the more likely that person is to have the outcome, in this case, contraceptive use. When designing family planning programs, this means that program planners can consider the entire scope of ideational factors that are predictive of contraceptive use and select a group of factors to target that are: (1) currently not prevalent, so there is room for growth, and (2) amenable to change. Ideal family size, for example, is an important ideational factor, but it may not always be feasible for programs to address for both practical and political reasons. The more positive ideational factors that women had, the greater their contraceptive use. reason why NURHI was designed with television, radio, social mobilization, and clinic-based communication interventions-to maximize the types of interventions people experience. Another reason to use multiple communication channels is to maximize the chance of exposure in general, as no one channel reaches everyone. In addition, different channels have different uses, for example, radio is useful for modeling change through entertainment-education while interpersonal communication helps to deepen knowledge; together, the messages communicated through multiple channels become mutually reinforcing.
Intention to use contraceptives is an important indicator for NURHI, because it gives an indication of likely future users-women who might not be ready to use contraception now due to pregnancy or other factors but who want to plan their families. In each NURHI project city, there was an upward trend in the percentage of women saying they intended to use contraception in the next 12 months. In an analysis of data from 27 Demographic and Health Surveys conducted between 1993 and 1996, for each 1% increase in intention to use contraception, there was nearly a 1% rise in contraceptive adoption. [bib_ref] Contraceptive use, intention to use and unmet need during the extended postpartum..., Ross [/bib_ref] Clinical outreach through mobile services played an important role in improving contraceptive use. Women responded enthusiastically to having family planning services brought to their own neighborhoods. By the end of the observation period, nearly one-half of the services provided by NURHI-supported publicsector clinics came from these outreach visits.
Mobile service delivery has shown great success in other projects. [bib_ref] Scaling up delivery of contraceptive implants in sub-Saharan Africa: operational experience of..., Duval [/bib_ref] In the urban slums supported by NURHI, women do, technically, have access to family planning facilities within a reasonable distance. But NURHI's mobile services put the convenience and needs of family planning users first, by having service providers travel to them, Significance of change in perceived peer support is P , .05 for zero exposure and P , .0001 for low, medium, and high levels of exposure. Significance of differences across groups: P , .001. rather than expecting women to travel to the service provider.
Many large-scale family planning programs tend to be ''service-led,'' that is, informed by a service-delivery and health systems strengthening approach. Such a program would typically be managed by a partner known for its service delivery expertise and budgeted with service delivery absorbing the majority of program funds, with service delivery needs setting the tone and pace of the project. If supported adequately by demand generation and other high-impact practices, 6 the service-led approach may be the appropriate design for a given context. However, given that demand generation and communication interventions have been shown to increase family planning use in Nigeria and elsewhere, [bib_ref] Mass media messages and reproductive behaviour in Nigeria, Bankole [/bib_ref] [bib_ref] The effects of a communication program on contraceptive ideation and use among..., Gupta [/bib_ref] it is worth exploring whether an alternative project strategy is effective.
An alternative strategy is for a ''demand-led'' program, such as that of the NURHI project, which was designed with demand generation as its driving force. What does it mean for a family planning project to be demand-led rather than service-led? It means that from the outset of design, program planners put potential and current family planning users at the forefront, along with their barriers and challenges to using family planning and their desires and hopes. With that insight as a starting point, the demand-led project would then design the appropriate systems, supplies, provider inputs, and communication interventions needed to serve the potential and current users. While using this approach, NURHI has come to see the locus of the program as the space between husband and wife, or between romantic partners, rather than at the clinic. The catalyst happens in the home; the rest of the (very substantial) work involves making sure the couple is supported and enabled, both in the community and in the clinic, to plan their family.
We cannot yet assert whether the demandled approach is effecting the CPR faster than the standard approach, but we can say that it is working. The NURHI hypothesis is that at some point, when CPR has reached a high enough level, family planning will become an ordinary part of family life, and people will feel that their community supports it to such an extent that demand for family planning will be selfmaintaining. It is at that point that demand will truly drive supply, leading to sustained demand with providers working to meet it. That does not mean that no further investment will be needed when this occurs; health systems must be By May 2013, outreach visits were contributing nearly half of total clinic family planning services supported by NURHI.
A ''demand-led'' family planning program puts potential and current family planning users at the forefront and uses demand generation as its driving force.
funded. But NURHI does believe that an investment in making family planning a social normwhereby women perceive contraceptive use is ubiquitous, approved, and supported by family, community, and influential leaders-will lead, in time, to a level of demand that will prevent the CPR from falling back down to the low levels now common in some countries such as Nigeria. We have not gotten there yet, but we are headed in the right direction.
[fig] FIGURE 1: Ideation Model of CommunicationSource: Health Communication Capacity Collaborative (2014).8 [/fig]
[fig] FIGURE 2: Nigerian Urban Reproductive Health Initiative (NURHI) Interventions Abbreviations: FP, family planning; FPPN, Family Planning Providers Network; PMVs, patent medicine vendors. [/fig]
[fig] FIGURE 3: Percentage [/fig]
[fig] Figure 7: shows the contribution of clinical outreach to the total number of clients served by the high-volume sites where NURHI has trained and supported providers. Between 2009 and 2011, the NURHI project worked with selected publicsector sites to improve their facilities and quality of services. In the third year of the project (2012), NURHI began dispatching family planning outreach staff to hard-to-reach slum areas. Between January 2011 and May 2013, the number of family planning users served by NURHIsupported facilities steadily increased, from about 1,000 users per month to about 7,000 total users in May 2013 [/fig]
[fig] FIGURE 5: Change in Perceived Peer Support for Family Planning Between Baseline and Midterm, By Level of Exposure to NURHI Activities, N54,331 [/fig]
[fig] FIGURE 6: Contraceptive Prevalence at Midterm Among Married Women Who Were Not Using a Modern Method at Baseline, by Level of Ideation at Midterm, N51,992 [/fig]
[table] TABLE 1: Modern Contraceptive Prevalence Rate Among Married Women, at Baseline and Midterm, by NURHI Project City [/table]
[table] TABLE 2: Ten Ideation Factors at Baseline and Midterm That Predict Contraceptive Use [/table]
|
Prevalence and prognostic value of the coexistence of anaemia and frailty in older patients with heart failure
Aims There have been no investigations of the prevalence and clinical implications of coexistence of anaemia and frailty in older patients hospitalized with heart failure (HF) despite their association with adverse health outcomes. The present study was performed to determine the prevalence and prognostic value of the coexistence of anaemia and frailty in hospitalized older patients with HF.Methods and resultsWe performed post hoc analysis of consecutive hospitalized HF patients ≥65 years old enrolled in the FRAGILE-HF, which was the prospective, multicentre, observational study. Anaemia was defined as haemoglobin < 13 g/dL in men and <12 g/dL in women, and frailty was evaluated according to the Fried phenotype model. The study endpoint was all-cause mortality. Of the total of 1332 patients, 1217 (median age, 81 years; 57.4% male) were included in the present study. The rates of anaemia and frailty in the study population were 65.7% and 57.0%, respectively. The patients were classified into the non-anaemia/non-frail group (16.6%), anaemia/non-frail group (26.4%), non-anaemia/frail group (17.7%), and anaemia/frail group (39.3%). A total of 144 patients died during 1 year of follow-up. In multivariate analyses, only the anaemia/frail group showed a significant association with elevated mortality rate (adjusted hazard ratio, 1.94; 95% confidence interval, 1.02-3.70; P = 0.043), compared with the non-anaemia/non-frail group after adjusting for other covariates. Conclusions Coexistence of anaemia and frailty are prevalent in hospitalized older patients with HF, and it has a negative impact on mortality.
# Introduction
Heart failure (HF) is a global public health concern estimated to affect 26 million people worldwide, and its prevalence is increasing with the aging of the population.HF adversely affects both mortality and quality of life, as well as increasing medical costs.Although drugs directly targeting aspects of the pathophysiology of HF are essential for treatment, it is also necessary to recognize and appropriately manage comorbidities to improve quality of life and prognosis.
Frailty consisting of reduced physiological reserve and vulnerability to external stressors in geriatric patients is a high-priority issue in cardiovascular medicine.Anaemia and frailty are conditions commonly encountered in hospitalized older patients with HF and are associated with poorer clinical status and higher risk of death.Anaemia and frailty have been shown to share a pathophysiology associated with chronic inflammatory processes, and coexistence of both conditions often come up on the agenda in the field of gerontology.In healthy individuals, reduced oxygen delivery is compensated by increases in both heart rate and stroke volume 9 ; however, these mechanisms already impaired in patients with HF. Consequently, anaemia reduces tissue oxygenation and impairs muscle performance in HF patients, thus increasing fatigue, cognitive decline, and weakening muscle strength.As these factors contribute to frailty syndrome, many HF patients have both anaemia and frailty, leading to a vicious cycle of physiological decline.Despite the burden imposed on healthcare systems worldwide by the coexistence of anaemia and frailty in older patients with HF, there have been no previous investigations of the prevalence rate of coexisting anaemia and frailty and the impact on mortality in this population.
The present study was therefore designed to investigate the prevalence and prognostic impact of the coexistence of anaemia and frailty in hospitalized older patients with HF based on data from the FRAGILE-HF (Prevalence and Prognostic Value of Physical and Social Frailty in Geriatric Patients Hospitalized for Heart Failure) cohort study.
# Methods
## Study design and patient population
This was a secondary analysis of the FRAGILE-HF study, a prospective, multicentre, observational study performed in 15 hospitals in Japan, as reported previously.Briefly, all consecutive patients ≥65 years old hospitalized for the first time due to decompensation of HF and capable of ambulation at discharge between September 2016 to March 2018 were evaluated for eligibility. The Framingham criteria were used for diagnosis of decompensation of HF. Patients with previous heart transplantation or left ventricular assist device implantation, those receiving either chronic peritoneal dialysis or haemodialysis, those with acute myocarditis, and those with brain natriuretic peptide (BNP) level <100 pg/mL or N-terminal-proBNP level < 300 pg/mL at admission, as well as those for whom these data were not available were excluded from the present study.
The study was performed in accordance with the tenets of the Declaration of Helsinki and Japanese Ethical Guideline for Medical and Health Research Involving Human Subjects and was approved by the Ethics Committee of each participating hospital. All participants were free to opt out of the study at any time. Study information, which included the study objectives, inclusion and exclusion criteria, and the names of participating institutes, were published in the University Hospital Information Network (UMIN-CTR, unique identifier: UMIN000023929) prior to enrolment of the first patient.
## Data collection and definitions
Baseline physical findings, blood samples, echocardiography, and medications were obtained before discharge for all patients with haemodynamic stability. A diagnosis of anaemia was made based on haemoglobin < 13 g/dL for men and <12 g/dL for women in accordance with the criteria of the World Health Organization.The formula of the Japanese Society of Nephrology was used to determine the estimated glomerular filtration rate (eGFR),and eGFR <60 mL/min/1.73 m 2 was used as a criterion for a diagnosis of renal dysfunction.
Frailty and both physical and cognitive function were evaluated by a trained personnel. Frailty status was evaluated using the Fried phenotype model,with patients considered to be frail if they had three or more of the following components: weakness (hand grip strength), slowness (gait speed), weight loss, exhaustion, and low physical activity. The questionnaires used to check and diagnose frailty were published in detail previously.The Short Physical Performance Battery (SPPB), which consists of three components (standing balance, usual gait speed, repeated chair stands), was applied according to established methods,with SPPB scores ranging from 0 to 12, that is 0-4 points for each component (0 = worst; 12 = best). The 6 min walking distance was determined according to the guidelines of the American Thoracic Society.Briefly, the patients were instructed to cover as much distance as possible within the allotted time using assistive devices if necessary. Cognitive function was evaluated using Mini-Cog, which is a composite of a three-item recall test and clock-drawing test.Instructions were provided in accordance with the Mini-Cog© website (https://mini-cog. com), and scores were given on a 5-point scale (0 = worst; 5 = best), and a score of <3 was considered abnormal.
The endpoint of this study was all-cause mortality, and the time to the endpoint was calculated as the number of days from the date of discharge to the date of the event.
# Statistical analysis
Continuous variables with a normal distribution are presented as the means ± SD, while variables with a non-normal distribution are presented as the median and interquartile range. Categorical variables are expressed as numbers and percentages. The cohort was divided into four groups: (i) non-anaemia/non-frail group; (ii) anaemia/nonfrail group; (iii) non-anaemia/frail group; and (iv) anaemia/ frail group. Differences between groups were evaluated by one-way analysis of variance or the Kruskal-Wallis test for continuous variables, and χ 2 or Fisher's exact test for dichotomous variables, as appropriate.
Survival was evaluated using the Kaplan-Meier survival method and compared using log-rank statistics. For analysis of all-cause mortality, we used the Meta-analysis Global Group in Chronic HF (MAGGIC) risk score and (log-transformed) BNP levels at discharge as adjustment variables in a multivariable prognostic model because MAGGIC score is a well validated risk score for Japanese HF patients.Multiple imputation was used to take into account the missing covariate data, excluding anaemia and frailty status, to construct multivariable Cox regression models. We created 20 datasets using a chained-equations procedure.Parameter estimates were obtained for each dataset and subsequently combined to produce an integrated result using the method described by A two-tailed P < 0.05 was taken to indicate statistical significance in all analyses. Statistical analyses were performed using R version 3.1.2 (R Foundation for Statistical Computing, Vienna, Austria; ISBN 3-900051-07-0, URL http://www. R-project.org).
# Results
A total of 1332 hospitalized HF patients ≥65 years old were registered in the FRAGILE-HF study during the study period. After excluding 115 patients due to missing data regarding anaemia or frailty status, the remaining 1217 patients were included in the analysis. The median age of the study population was 81 years, and 57.4% were male participants. The prevalence rates of anaemia and frailty were 65.7% (799/1217 patients) and 57.0% (694/1217 patients), respectively, and both increased with age (Supporting Information, . Patients with anaemia had a significantly higher rate of frail compared with those without anaemia (59.8% vs. 51.7%, respectively, P = 0.007). The study population was classified into four groups as follows:
non-anaemia/non-frail group, 202 (16.6%) patients; anaemia/non-frail group, 321 (26.4%) patients; non-anaemia/ frail group, 216 (17.7%) patients; and anaemia/frail group, 478 (39.3%) patients. More than half of the patients aged ≥85 years and with body mass index <18.5 kg/m 2 were positive for both anaemia and frailty . The anaemia/frail group was associated with older age, more severe symptoms, lower body mass index, lower diastolic blood pressure, higher left ventricular ejection fraction, lower proportion of atrial fibrillation, higher proportion of renal dysfunction, less prescription of beta blockers and mineralocorticoid receptor antagonists, and higher MAGGIC risk score than the other groups. In addition, the anaemia/frail group had lower haemoglobin, haematocrit, albumin and eGFR, and higher creatinine, blood urea nitrogen, and brain natriuretic peptide level at discharge than the other groups.
As shown in, the mean handgrip strength and gait speed for the whole study population were 20.1 ± 7.9 kg and 0.79 ± 0.30 m/s, respectively. The median SPPB score was 9 points, and lower extremity strength (chair stand test) was especially impaired, with 18% of the total cohort lacking the leg strength to stand from a seated position without using their arms even once. The mean 6MWD was 253 ± 126 m, and most patients (80.4%) had 6MWD < 400 m. Patients with anaemia had a significantly shorter 6MWD compared with those without anaemia (237 ± 121 m vs. 284 ± 129 m, respectively, P < 0.001), and patients with frailty also had a significantly shorter 6MWD compared with those without frailty (221 ± 121 m vs. 295 ± 120 m, respectively, P < 0.001). Patients in the anaemia/frail group showed greater impairment of physical function than the other groups. The median Mini-Cog score was 3 points, and 36.9% of the patients had scores <3 points. Patients with anaemia and/or frailty had a higher proportion of cognitive abnormalities compared with those without anaemia and frailty.
A total of 144 deaths occurred in the whole study population during 1 year of follow-up, and Kaplan-Meier curve analysis indicated that the incidence of all-cause mortality was higher in the anaemia/frail group than the other groups.shows the results of Cox regression analyses for all-cause death. After adjusting for other covariates, adjusted models showed that only the anaemia/frail group was associated with higher mortality (hazard ratio 1.94;95% confidence interval, 1.02-3.70; P = 0.043) compared with the non-anaemia/non-frail group.
# Discussion
This study was performed to examine the prevalence of anaemia and frailty and the coexistence of both conditions in a large cohort of hospitalized older HF patients from the FRAGILE-HF cohort study. More than 80% of the patients Anaemia and frailty in older patients with HF included in the analysis had either anaemia or frailty, and one third of the patients showed both conditions, particularly associated with more advanced age and lower body mass index. The mortality rate was higher among patients with both anaemia and frailty, even after adjusting for known risk factors. To our knowledge, this is the first report that the combination of anaemia and frailty has prognostic significance in hospitalized older HF patients and may have important implications for risk stratification and management of these two important comorbidities.
The reported prevalence rates of anaemia in HF patients vary over a wide range from 17% to 70%, which may be due to differences in the definition of anaemia and characteristics of the patient populations between studies.The prevalence rate of anaemia of 65.7% in the present study was consistent with the ADHERE and ATTEND acute HF registries, which indicated that more than half of all hospitalized HF patients are anemic.The reported prevalence of frailty in HF varies widely from 19% to 77%,likely due to differences in definitions of frailty and in characteristics of the study populations between studies. The rate of frailty was 57.0% in the present study, which was similar to the prevalence of~50% among hospitalized HF patients reported previously.The prevalence of frailty in hospitalized patients with HF is higher than in patients with stable HF, 24 because the symptoms of HF overlap with exhaustion, and hospitalization itself causes weakness and slowness. In addition, patients who are hospitalized in severe condition are more likely to have weakness and slowness, and the hospital environment itself leads to inactivity. Although fluid retention can cause weight gain in Prevalence rates of the coexisting of anaemia and frailty in the total cohort and subgroups. AF, atrial fibrillation; BMI, body mass index; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; LVEF, left ventricular ejection fraction. Values are expressed as means ± SD, n (%), or median (interquartile range).
Anaemia and frailty in older patients with HF patients with HF, the progression of HF often leads to malnutrition, resulting in weight loss.It is important to understand the condition of each component of frailty to improve the effectiveness of interventions. More than 80% of hospitalized HF patients ≥65 years old had anaemia and/or frailty in the present study, which was a markedly high proportion compared with the rate of 57.6% reported previously in patients ≥80 years old with acute coronary syndromes.Anaemia can be induced by HF via a number of related pathophysiological mechanisms, especially functional or absolute iron deficiency and erythropoietin synthesis and response.In addition, HF can result in frailty due to coordinated multisystem dysfunction as a result of its systemic nature, including systemic inflammation, high comorbidity burden, advanced age, and skeletal muscle abnormalities.The results of the present study indicated that the coexistence of both anaemia and frailty was associated with reduced functional status and elevated mortality rate, consistent with a previous study in community-dwelling older adults.Impairments in physical function were broad and severe, involving all SPPB components (balance, mobility, and strength) and 6MWD (endurance), among patients with both anaemia and frailty. Similarly to a previous study, lower extremity strength was especially impaired among each component of the SPPB,and the profoundly reduced 6MWD in the patients with both anaemia and frailty was comparable with that in patients with advanced HF awaiting left ventricular assist device implantation.In addition, handgrip strength and gait speed are not only used to assess frailty, but are also used to define sarcopenia, which is considered an important factor in the frailty cycle.International consensus on sarcopenia defines a person with sarcopenia who walks <400 m during a 6 min walk as 'sarcopenia with limited mobility,' 29 and more than 80% of the patients had 6MWD < 400 m in the present study. These conditions may be related to low physical and social activities after discharge, leading to social frailty and poor prognosis. Cognitive function was lower in patients with anaemia and/or frailty than in those without anaemia and frailty, but there was no apparent effect of the combination of anaemia and frailty on cognitive function. As the present study included patients who were able to walk at discharge, patients with relatively preserved cognitive function could have been included, and the range of the Mini-Cog score was narrow (5-point scale). These factors may have reduced the statistical power to detect such associations.
Planning of medical care and improvement of the prognosis of vulnerable populations require accurate risk stratification related to acute HF. The accurate assessment of both anaemia and frailty status in hospitalized older HF patients is important because both of these conditions are not only indicators of the severity of disease but may also be treatable and thus lead to improvement of the overall clinical outcome.Iron deficiency, which is an independent predictor of reduced exercise capacity, quality of life, and survival, is recognized as a key treatable cause of anaemia in HF patients. 30,31 FAIR-HF and CONFIRM-HF showed that intravenous iron therapy improved exercise tolerance and quality of life in patients with HF,and the European Society of Cardiology guidelines have given a class IIa recommendation for iron deficiency correction by intravenous ferric carboxymaltose.Although the role of intravenous iron therapy in acute HF is not yet clear, a new trial, AFFIRM-AHF, is currently underway with the composite of recurrent HF hospitalizations and cardiovascular mortality as the primary outcome.On the other hand, outpatient cardiac rehabilitation participation was associated with reduced risks of all-cause death and HF rehospitalization in frail patients with HF.A new trial, the REHAB-HF study, is currently underway, and this pilot study has shown that rehabilitation therapy beginning in the hospital improved 6MWD and SPPB score, with the strongest trend seen with chair stands, in hospitalized older patients with HF, and the change in SPPB score was strongly related to all-cause rehospitalization.Tissue oxygenation is decreased by anaemia and hypoxia, which may lead to functional impairment of muscle. A previous longitudinal study showed that anaemia preceded the occurrence of frailty,and a meta-analysis indicated that the odds of frailty are more than doubled by the presence of anaemia in older individuals.Although the rate of frailty was significantly higher in patients with than without anaemia in the present study (59.8% vs. 51.7%, respectively), there was no apparent difference in these rates. As HF itself can cause anaemia and frailty, HF patients may show an attenuated relation between these two conditions. On the other hand, the rate of coexisting anaemia and frailty increased markedly with age in the present study. Anaemia and frailty share pathophysiological mechanisms related to chronic inflammatory processes induced by immunosenescence-associated changes and oxidative stress.Metabolism and skeletal muscle mass are adversely affected by systemic inflammation, resulting in sarcopenia and cachexia.Cachexia is a generalized wasting process occurring at rates of 5-15% in HF patients, especially in cases of more advanced disease.Although body mass index is inversely associated with risk of HF, patients with low body mass index have poorer prognosis once HF has been established in what is referred to as the 'obesity paradox.'Patients in the present study with body mass index <18.5 kg/m 2 had a greater likelihood of having both anaemia and frailty, and therefore, careful assessment and discussion are required in the clinical management of such patients. In addition, patients with anaemia and/or frailty were more likely to have history of cancer in the present study. Cancer investigations are needed for such patients because prior cancer history was associated with high cardiac event and mortality rates in hospitalized patients with HF.Further detailed studies are required to address these issues and to guide future clinical decision making for the treatment of older HF patients.
# Study limitations
The present study had several limitations. First, iron indices were not available in this study to allow us to determine the causes of anaemia, and data were not available regarding whether patients received intravenous iron supplementation and were administered erythropoiesis-stimulating agents, folate, and vitamin B 12 . However, iron supplementation has not been confirmed to reduce all-cause mortality rates in HF populations.As our endpoint was all-cause mortality, this may have had only a limited impact on the results and conclusions of our study. Second, frailty was defined based on questionnaires, which may be susceptible to recall bias, although such questionnaires have been widely applied in population studies. In addition, we did not exclude patients with cognitive impairment from the present study, and this may have impacted the results. Third, we evaluated frailty status only once before discharge, and no information was obtained regarding changes in frailty status, which have the potential of dynamic variability after changes in medical therapy or programs of rehabilitation.Fourth, unadjusted and unmeasured factors, such as multiple comorbidities, polypharmacy, and changes in baseline variables, all of which may have an effect on mortality, leave residual bias, and the results must be replicated in future studies. Finally, this study was conducted mostly in Japanese HF patients, and further studies in other populations are required to validate the prognostic value of combined anaemia and frailty.
# Conclusions
Both anaemia and frailty have highly prevalence rates in hospitalized older patients with HF, and the coexistence of both conditions was shown to adversely affect mortality rate in this population. On the basis of their high prevalence rate and clinical impact in older HF patients, both anaemia and frailty may be important targets for therapy in these patients.
## Conflict of interest
## Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article.
## Figure s1
Prevalence of anemia and frailty in the total cohort and subgroups. |
Plant innate immunity against human bacterial pathogens
Certain human bacterial pathogens such as the enterohemorrhagic Escherichia coli and Salmonella enterica are not proven to be plant pathogens yet. Nonetheless, under certain conditions they can survive on, penetrate into, and colonize internal plant tissues causing serious food borne disease outbreaks. In this review, we highlight current understanding on the molecular mechanisms of plant responses against human bacterial pathogens and discuss salient common and contrasting themes of plant interactions with phytopathogens or human pathogens.
# Introduction
Bagged greens in the market are often labeled "pre-washed," "triple-washed," or "ready-to-eat," and look shiny and clean. But are they really "clean" of harmful microbes? We cannot be so sure. Food safety has been threatened by contamination with human pathogens including bacteria, viruses, and parasites. Between 2000 and 2008, norovirus and Salmonella spp. contributed to 58 and 11% of forborne illnesses, respectively in the United States [bib_ref] Foodborne illness acquired in the United States-major pathogens, Scallan [/bib_ref]. In those same years, non-typhoidal Salmonella alone was ranked as the topmost bacterial pathogen contributing to hospitalizations (35%) and deaths (28%) [bib_ref] Foodborne illness acquired in the United States-major pathogens, Scallan [/bib_ref]. In 2007, 235 outbreaks were associated with a single food commodity; out of which 17% was associated with poultry, 16% with beef, and 14% with leafy vegetables that also accounted for the most episodes of illnesses [bib_ref] Surveillance for foodborne disease outbreaks: United States, Cdc [/bib_ref].
Apart from the direct effects on human health, enormous economic losses are incurred by contaminated food products recalls. The 8-day recall of spinach in 2006 cost $350 million to the US economy. It should be realized that this is not the loss of one individual, but several growers, workers, and distributors. This is a common scenario for any multistate foodborne outbreak. Additionally, the skepticism of the general public toward consumption of a particular food product can lead to deficiencies of an important food source from the diet. Less demand would in turn lead to losses for the food industry. Economic analysis shows that money spent on prevention of foodborne outbreak by producers is much less than the cost incurred after the outbreak [bib_ref] Economic analysis of food safety compliance costs and foodborne illness outbreaks in..., Ribera [/bib_ref].
Contamination of plants can occur at any step of food chain while the food travels from farm to table. Both pre-harvest and post-harvest steps are prone to contamination. Contaminated irrigation water, farm workers with limited means of proper sanitation, and fecal contamination in the farm by animals can expose plants to human pathogens before harvest of the edible parts [bib_ref] The growing burden of foodborne outbreaks due to contaminated fresh produce: risks..., Lynch [/bib_ref] [bib_ref] Interrelationships of food safety and plant pathology: the life cycle of human..., Barak [/bib_ref].
After harvest, contamination can occur during unclean modes of transportation, processing, and bagging [bib_ref] The growing burden of foodborne outbreaks due to contaminated fresh produce: risks..., Lynch [/bib_ref]. Mechanical damage during transport can dramatically increase the population of human pathogens surviving on the surface of edible plants [bib_ref] Survival and dissemination of Escherichia coli O157:H7 on physically and biologically damaged..., Aruscavage [/bib_ref]. Control measures to decrease pathogen load on plant surfaces have been defined by the Food Safety Modernization Act (US Food and Drug Administration) and Hazard Analysis and Critical Control Point system (HACCP). Using chlorine for post-harvest crop handling has been approved by US Department of Agriculture (USDA) under the National Organic Program. However, some studies indicated that internalized human pathogens escape sanitization [bib_ref] Attachment of Escherichia coli O157:H7 to lettuce leaf surface and bacterial viability..., Seo [/bib_ref] [bib_ref] Surface structures involved in plant stomata and leaf colonization by shiga-toxigenic Escherichia..., Saldaña [/bib_ref]. Thus, understanding the biology of human pathogen-plant interactions is now crucial to prevent pathogen colonization of and survival in/on plants, and to incorporate additional, complementing measures to control food borne outbreaks.
We reasoned that as plants are recognized vectors for human pathogens, enhancing the plant immune system against them creates a unique opportunity to disrupt the pathogen cycle. In this cross-kingdom interaction, the physiology of both partners contribute to the outcome of the interactions (i.e., colonization of plants or not). Bacterial factors important for interaction with plants have been discussed in recent, comprehensive reviews [bib_ref] Plants as a habitat for beneficial and/or human pathogenic bacteria, Tyler [/bib_ref] [bib_ref] Human enteric pathogens in produce: un-answered ecological questions with direct implications for..., Teplitski [/bib_ref] [bib_ref] Fresh fruit and vegetable as vehicles for the transmission of human pathogens, Berger [/bib_ref] [bib_ref] Interrelationships of food safety and plant pathology: the life cycle of human..., Barak [/bib_ref] [bib_ref] Salmonella interactions with plants and their associated microbiota, Brandl [/bib_ref]. Plant factors contributing to bacterial contamination (or lack of) is much less studied and discussed. In this review, we highlight current knowledge on plants as vectors for human pathogens, the molecular mechanisms of plant responses to human bacterial pathogens, and discuss common themes of plant defenses induced by phytopathogens and human pathogens. We have focused on human bacterial pathogens that are not recognized plant pathogens such as Salmonella enterica and Escherichia coli [bib_ref] Interrelationships of food safety and plant pathology: the life cycle of human..., Barak [/bib_ref] [bib_ref] Salmonella colonization activates the plant immune system and benefits from association with..., Meng [/bib_ref] , but yet are major threats to food safety and human health.
## Plant surface: the first barrier for bacterial invaders
The leaf environment has long been considered to be a hostile environment for bacteria. The leaf surface is exposed to rapidly fluctuating temperature and relative humidity, UV radiation, fluctuating availability of moisture in the form of rain or dew, lack of nutrients, and hydrophobicity [bib_ref] Microbiology of the phyllosphere, Lindow [/bib_ref]. Such extreme fluctuations, for example within a single day, are certainly not experienced by pathogens in animal and human gut. Thus, it is tempting to speculate that animal pathogens may not even be able to survive and grow in an environment as dynamic as the leaf surface. However, the high incidence of human pathogens such as S. enterica and E. coli O157:H7 on fresh produce, sprouts, vegetables, leading to foodborne illness outbreaks indicate a certain level of human pathogen fitness in/on the leaf.
The plant surface presents a barrier to bacterial invaders by the presence of wax, cuticle, cell wall, trichomes, and stomata. All except stomata, present a passive defense system to prevent internalization of bacteria. Nonetheless, several bacteria are able to survive on and penetrate within the plant interior. The surface of just one leaf is a very large habitat for any bacteria. The architecture of the leaf by itself is not uniform and provides areas of different environmental conditions. There are bulges and troughs formed by veins, leaf hair or trichomes, stomata, and hydathodes that form microsites for bacterial survival with increased water and nutrient availability, as well as temperature and UV radiation protection [bib_ref] Appetite of an epiphyte: quantitative monitoring of bacterial sugar consumption in the..., Leveau [/bib_ref] [bib_ref] Biological sensor for sucrose availability: relative sensitivities of various reporter genes, Miller [/bib_ref] [bib_ref] Leaf age as a risk factor in contamination of lettuce with Escherichia..., Brandl [/bib_ref] [bib_ref] Internalization of Salmonella enterica in leaves is induced by light and involves..., Kroupitski [/bib_ref] [bib_ref] Colonization of tomato plants by Salmonella enterica is cultivar dependent, and type..., Barak [/bib_ref]. Indeed, distinct microcolonies or aggregates of S. enterica were found on cilantro leaf surfaces in the vein region [bib_ref] Fitness of Salmonella enterica serovar Thompson in the cilantro phyllosphere, Brandl [/bib_ref] In addition, preference to the abaxial side of lettuce leaf by S. enterica may be is an important strategy for UV avoidance . Conversion of cells to viable but non-culturable (VNBC) state in E. coli O157:H7 on lettuce leaves may also be a strategy to escape harsh environmental conditions [bib_ref] Induction of viable but nonculturable Escherichia coli O157:H7 in the phyllosphere of..., Dinu [/bib_ref]. Hence, localization to favorable microsites, avoidance of harsh environments, and survival by aggregation or conversion to non-culturable state may allow these human pathogens to survive and at times multiply to great extent on the leaf surface.
As stomata are abundant natural pores in the plant epidermis which serve as entrance points for bacteria to colonize the leaf interior (intercellular space, xylem, and phloem), several studies addressed the question as to whether human bacterial pathogens could internalize leaves through stomata. Populations of E. coli O157:H7 and S. enterica SL1344 in the Arabidopsis leaf apoplast can be as large as four logs per cm 2 of leaf after surface-inoculation under 60% relative humidity [bib_ref] Escherichia coli O157:H7 induces stronger plant immunity than Salmonella enterica Typhimurium SL1344, Roy [/bib_ref] suggesting that these bacteria can and access the apoplast of intact leaves. Several microscopy studies indicated association of pathogens on or near guard cells. For instance, S. enterica serovar Typhimurium SL1344 was shown to internalize arugula and iceberg lettuce through stomata and bacterial cells were located in the sub-stomatal space [bib_ref] Salmonella Typhimurium internalization is variable in leafy vegetables and fresh herbs, Golberg [/bib_ref]. However, no internalization of SL1344 was observed into parsley where most cells were found on the leaf surface even though stomata were partially open [bib_ref] Salmonella Typhimurium internalization is variable in leafy vegetables and fresh herbs, Golberg [/bib_ref]. Cells of S. enterica serovar Typhimurium MAE110 [bib_ref] Internal colonization of Salmonella enterica serovar Typhimurium in tomato plants, Gu [/bib_ref] , enteroaggregative E. coli [bib_ref] Interaction of enteroaggregative Escherichia coli with salad leaves, Berger [/bib_ref] , and E. coli O157:H7 [bib_ref] Surface structures involved in plant stomata and leaf colonization by shiga-toxigenic Escherichia..., Saldaña [/bib_ref] were found to be associated with stomata in tomato, arugula leaves, and baby spinach leaves, respectively. In the stem E. coli O157:H7 and Salmonella serovar Typhimurium were found to be associated with the hypocotyl and the stem tissues including epidermis, cortex, vascular bundles, and pith when seedlings were germinated from contaminated seeds .
The plant rhizosphere is also a complex habitat for microorganisms with different life styles including plant beneficial symbionts and human pathogens. Nutritionally rich root exudate has been documented to attract S. enterica to lettuce roots [bib_ref] Differential interaction of Salmonella enterica serovars with lettuce cultivars and plant-microbe factors..., Klerks [/bib_ref]. Although bacteria cannot directly penetrate through root cells, sites at the lateral root emergence and root cracks provide ports of entry for S. enterica and E. coli O157:H7 into root tissues [bib_ref] Colonization of Arabidopsis thaliana with Salmonella enterica and enterohemorrhagic Escherichia coli O157:H7..., Cooley [/bib_ref] [bib_ref] Kinetics and strain specificity of rhizosphere and endophytic colonization by enteric bacteria..., Dong [/bib_ref] [bib_ref] Physiological and molecular responses of Lactuca sativa to colonization by Salmonella enterica..., Klerks [/bib_ref] [bib_ref] Plants as a habitat for beneficial and/or human pathogenic bacteria, Tyler [/bib_ref] , and in some instances between the epidermal cells [bib_ref] Physiological and molecular responses of Lactuca sativa to colonization by Salmonella enterica..., Klerks [/bib_ref]. High colonization of S. enterica has been observed in the root-shoot transition area [bib_ref] Physiological and molecular responses of Lactuca sativa to colonization by Salmonella enterica..., Klerks [/bib_ref]. Once internalized both bacterial pathogens have been found in the intercellular space of the root outer cortex of Medicago truncatula [bib_ref] Response of Medicago truncatula Seedlings to Colonization by Salmonella enterica and Escherichia..., Jayaraman [/bib_ref]. Salmonella enterica was found in the parenchyma, endodermis, pericycle, and vascular system of lettuce roots [bib_ref] Physiological and molecular responses of Lactuca sativa to colonization by Salmonella enterica..., Klerks [/bib_ref] and in the inner root cortex of barley [bib_ref] Colonization of barley (Hordeum vulgare) with Salmonella enterica and Listeria spp, Kutter [/bib_ref]. A detailed study on the localization of E. coli O157:H7 in live root tissue demonstrated that this bacterium can colonize the plant cell wall, apoplast, and cytoplasm [bib_ref] The endophytic lifestyle of Escherichia coli O157:H7: quantification and internal localization in..., Wright [/bib_ref]. Intracellular localization of E. coli O157:H7 seems to be a rare event as most of the microscopy-based studies show bacterial cells in the intercellular space only. Bacterial translocation from roots to the phyllosphere may be by migration on the plant surface in a flagellum-dependent manner [bib_ref] Colonization of Arabidopsis thaliana with Salmonella enterica and enterohemorrhagic Escherichia coli O157:H7..., Cooley [/bib_ref] or presumably through the vasculature [bib_ref] Enterohemorrhagic Escherichia coli O157:H7 present in radish sprouts, Itoh [/bib_ref] [bib_ref] Transmission of Escherichia coli O157:H7 from contaminated manure and irrigation water to..., Solomon [/bib_ref]. The mechanism for internal movement of enteric bacterial cells from the root cortex to the root vasculature through the endodermis and casparian strips and movement from the roots to the phyllosphere through the vascular system is yet to be demonstrated.
Several outbreaks of S. enterica have also been associated with fruits, especially tomatoes. Salmonella enterica is unlikely to survive on surface of intact fruits [bib_ref] Growth and survival of Salmonella Montevideo on tomatoes and disinfection with chlorinated..., Wei [/bib_ref] raising the question: what are the routes for human pathogenic bacteria penetration into fruits? It has been suggested that S. enterica can move from inoculated leaves [bib_ref] Colonization of tomato plants by Salmonella enterica is cultivar dependent, and type..., Barak [/bib_ref] , stems, and flowers [bib_ref] Survival of Salmonellae on and in tomato plants from the time of..., Guo [/bib_ref] to tomato fruits. However, the rate of internal contamination of fruits was low (1.8%) when leaves were surface-infected with S. enterica [bib_ref] Internal colonization of Salmonella enterica serovar Typhimurium in tomato plants, Gu [/bib_ref]. The phloem has been suggested as the route of movement of bacteria to non-inoculated parts of the plant as bacterial cells were detected in this tissue by microscopy [bib_ref] Internal colonization of Salmonella enterica serovar Typhimurium in tomato plants, Gu [/bib_ref]. depicts the observed phyllosphere and rhizosphere niches colonized by bacteria in/on intact plants and probable sources of contamination.
## Figure 1 | schematic representation of human pathogen (hp) association with plants. (a)
Pathogens are introduced to soil through contaminated irrigation water, fertilizers, manure, and pesticides (1). HPs are attracted to rhizosphere (2; [bib_ref] Differential interaction of Salmonella enterica serovars with lettuce cultivars and plant-microbe factors..., Klerks [/bib_ref] and penetrate root tissues at the sites of lateral root emergence, root cracks as well as root-shoot transition area (3; [bib_ref] Colonization of Arabidopsis thaliana with Salmonella enterica and enterohemorrhagic Escherichia coli O157:H7..., Cooley [/bib_ref] [bib_ref] Kinetics and strain specificity of rhizosphere and endophytic colonization by enteric bacteria..., Dong [/bib_ref] [bib_ref] Physiological and molecular responses of Lactuca sativa to colonization by Salmonella enterica..., Klerks [/bib_ref] [bib_ref] Plants as a habitat for beneficial and/or human pathogenic bacteria, Tyler [/bib_ref]. HPs were found to live on the leaf surface near veins [bib_ref] Fitness of Salmonella enterica serovar Thompson in the cilantro phyllosphere, Brandl [/bib_ref] , in the leaf apoplast (intercellular space) [bib_ref] Fitness of Salmonella enterica serovar Thompson in the cilantro phyllosphere, Brandl [/bib_ref] [bib_ref] Transmission of Escherichia coli O157:H7 from contaminated manure and irrigation water to..., Solomon [/bib_ref] [bib_ref] Relative efficacy of sodium hypochlorite wash versus irradiation to inactivate Escherichia coli..., Niemira [/bib_ref] [bib_ref] Internalization of Salmonella enterica in leaves is induced by light and involves..., Kroupitski [/bib_ref] [bib_ref] Colonization of tomato plants by Salmonella enterica is cultivar dependent, and type..., Barak [/bib_ref] [bib_ref] Induction of viable but nonculturable Escherichia coli O157:H7 in the phyllosphere of..., Dinu [/bib_ref] [bib_ref] Internal colonization of Salmonella enterica serovar Typhimurium in tomato plants, Gu [/bib_ref] [bib_ref] Escherichia coli O157:H7 induces stronger plant immunity than Salmonella enterica Typhimurium SL1344, Roy [/bib_ref] , and sometimes with affinity for abaxial side of leaf (e.g., S. enterica; . Salmonella enterica Typhimurium can enter tomato plants via leaves and move through vascular bundles (petioles and stems) (5) into non-inoculated leaves (6) and fruits (8) [bib_ref] Internal colonization of Salmonella enterica serovar Typhimurium in tomato plants, Gu [/bib_ref]. HPs are also found to be associated with flower (7; [bib_ref] Survival of Salmonellae on and in tomato plants from the time of..., Guo [/bib_ref] [bib_ref] Colonization of Arabidopsis thaliana with Salmonella enterica and enterohemorrhagic Escherichia coli O157:H7..., Cooley [/bib_ref]. Salmonella could travel from infected leaves (4), stems (5), and flowers (7) to colonize the fruit interior (the diagram represents a cross-section of a fruit) and fruit calyx (8) [bib_ref] Survival of Salmonellae on and in tomato plants from the time of..., Guo [/bib_ref] [bib_ref] Transmission electron microscopy study of enterohemorrhagic Escherichia coli O157:H7 in apple tissue, Janes [/bib_ref] [bib_ref] Colonization of tomato plants by Salmonella enterica is cultivar dependent, and type..., Barak [/bib_ref]. Escherichia coli O157:H7 has also been observed in the internal parts of the apple and the seeds following contamination of the flower (8) [bib_ref] Attachment of Escherichia coli O157:H7 to the surfaces and internal structures of..., Burnett [/bib_ref]. Movement on the plant surface has also been observed (9; [bib_ref] Colonization of Arabidopsis thaliana with Salmonella enterica and enterohemorrhagic Escherichia coli O157:H7..., Cooley [/bib_ref]. Epiphytic Salmonella and E. coli O157:H7 can aggregate near stomata and sub-stomatal space (10; [bib_ref] Salmonella Typhimurium internalization is variable in leafy vegetables and fresh herbs, Golberg [/bib_ref] [bib_ref] Internal colonization of Salmonella enterica serovar Typhimurium in tomato plants, Gu [/bib_ref] [bib_ref] Surface structures involved in plant stomata and leaf colonization by shiga-toxigenic Escherichia..., Saldaña [/bib_ref] , reach the sub-stomatal cavity and survive/colonize in the spongy mesophyll [bib_ref] Transmission of Escherichia coli O157:H7 from contaminated manure and irrigation water to..., Solomon [/bib_ref] [bib_ref] Association of Escherichia coli O157:H7 with preharvest leaf lettuce upon exposure to..., Wachtel [/bib_ref] [bib_ref] Interaction of Escherichia coli with growing salad spinach plants, Warriner [/bib_ref] [bib_ref] Interactions of Escherichia coli O157:H7, Salmonella typhimurium and Listeria monocytogenes plants cultivated..., Jablasone [/bib_ref] [bib_ref] Quantification of contamination of lettuce by GFP-expressing Escherichia coli O157:H7 and Salmonella..., Franz [/bib_ref]. Salmonella cells were observed near trichomes (10; [bib_ref] Colonization of tomato plants by Salmonella enterica is cultivar dependent, and type..., Barak [/bib_ref] [bib_ref] Internal colonization of Salmonella enterica serovar Typhimurium in tomato plants, Gu [/bib_ref]. (B) Stem cross-section showing bacteria located in different tissues (Ep, epidermis; C, cortex; V, vascular tissue; Pi, pith) . (C) Root cross-section showing bacteria on the root surface, internalizing between the epidermal cells, and colonizing root outer and inner cortex, endodermis (En), pericycle (P) and vascular system [bib_ref] Colonization of barley (Hordeum vulgare) with Salmonella enterica and Listeria spp, Kutter [/bib_ref] [bib_ref] Response of Medicago truncatula Seedlings to Colonization by Salmonella enterica and Escherichia..., Jayaraman [/bib_ref].
## Perception of human pathogens by the plant immune system
Plants possess a complex innate immune system to ward off microbial invaders [bib_ref] The plant immune system, Jones [/bib_ref]. Plants are able to mount a generalized step-one response that is triggered by modified/degraded plant products or conserved pathogen molecules. These molecules are known as damage or pathogen associated molecular patterns (DAMP/PAMP). In many cases, conserved PAMPs are components of cell walls and surface structures such as flagellin, lipopolysaccharides, and chitin .
Examples of intracellular PAMPs exist such as the elongation factor EF-Tu. PAMPs are recognized by a diverse set of plant extracellular receptors called pattern-recognition receptors (PRRs) that pass intracellular signals launching an army of defense molecules to stop the invasion of the pathogens. This branch of the immune system known as pathogen-triggered immunity (PTI) is the first line of active defense against infection.
Human pathogen on plants (HPOP) is an emerging field that only recently has caught the attention of plant biologists and phytopathologists. A few studies have been reported in the last 5-10 years, which focused on the most well studied PAMPs, flagellin and lipopolysaccharide (LPS), in the interaction of human pathogens with plants. [fig_ref] Table 1 |: Experimental conditions used in the studies reporting plant response to pathogenic Salmonella... [/fig_ref] lists the plants, bacterial strains, and method details for such studies.
## Flagellin perception
Flagellin, the structural component of flagellum in bacteria, is involved in bacterial attachment and motility on the plant [bib_ref] Colonization of Arabidopsis thaliana with Salmonella enterica and enterohemorrhagic Escherichia coli O157:H7..., Cooley [/bib_ref] , is recognized by plant through the FLS2 receptor [bib_ref] Salmonella enterica flagellin is recognized via FLS2 and activates PAMP-triggered immunity in..., Garcia [/bib_ref] , and induces plant defenses [bib_ref] Salmonella colonization activates the plant immune system and benefits from association with..., Meng [/bib_ref] [bib_ref] Salmonella enterica flagellin is recognized via FLS2 and activates PAMP-triggered immunity in..., Garcia [/bib_ref]. Similar to the well-studied PTI elicitor flg22 [bib_ref] Plants have a sensitive perception system for the most conserved domain of..., Felix [/bib_ref] , the flg22 epitope of S. enterica serovar Typhimurium 14028 is also an effective PAMP and elicitor of downstream immune responses in Arabidopsis [bib_ref] Salmonella enterica flagellin is recognized via FLS2 and activates PAMP-triggered immunity in..., Garcia [/bib_ref] , tobacco, and tomato plants [bib_ref] Salmonella colonization activates the plant immune system and benefits from association with..., Meng [/bib_ref]. Flagellum-deficient mutants of S. enterica serovar Typhimurium 14028 are better colonizers of wheat, alfalfa, and Arabidopsis roots as compared to the wild type bacterium [bib_ref] Regulation of enteric endophytic bacterial colonization by plant defenses, Iniguez [/bib_ref] further suggesting that the Salmonella flagellum induces plant defenses that may restrict bacterial colonization of several plant organs. However, the Salmonella flg22 peptide is not the only PAMP for elicitation of plant immune response as fls2 mutant of Arabidopsis still shows a low level of PTI activation in response to this PAMP [bib_ref] Salmonella enterica flagellin is recognized via FLS2 and activates PAMP-triggered immunity in..., Garcia [/bib_ref].
Purified flagellin or derived epitopes of E. coli O157:H7 has not been used to induce plant defenses. However, flagellum-deficient mutant of this strain does not activate the SA-dependent BGL2 gene promoter as much as the wild type strain and shows larger population in Arabidopsis than the wild type strain [bib_ref] Influence of the plant defense response to Escherichia coli O157:H7 cell surface..., Seo [/bib_ref] further suggesting that surface structures in the bacterial cell are perceived by plants.
The differences in responses observed could be attributed to the presence of other microbial signatures eliciting plant defense. Variations in plant response to S. enterica flagellin could be owed to host-strain specificity as well. Although flagellin sequences from S. enterica strains and other bacteria are highly conserved, even a minor change of five amino acids in the flg22 epitope leads to reduced activation of PTI in Arabidopsis, tobacco, and tomato plants [bib_ref] Salmonella enterica flagellin is recognized via FLS2 and activates PAMP-triggered immunity in..., Garcia [/bib_ref]. Adding to the specificity, it has also been shown that Brassicaceae and Solanocecae plants recognize specific flagellin [bib_ref] Molecular identification and characterization of the tomato flagellin receptor LeFLS2, an orthologue..., Robatzek [/bib_ref] [bib_ref] Allelic variation in two distinct Pseudomonas syringae flagellin epitopes modulates the strength..., Clarke [/bib_ref]. Hence, evolving variations in flagellin sequences could be a strategy employed by the pathogens to avoid plant recognition, which in turn leads to the development of pathogen-specific immune responses in the plant.
Flagella also play an important role in bacterial behavior on the plant. Several studies have pointed out to the usefulness of flagella for attachment to leaf surfaces and movement on plant surfaces [bib_ref] Interaction of Escherichia coli O157:H7 with leafy green produce, Xicohtencatl-Cortes [/bib_ref] [bib_ref] Surface structures involved in plant stomata and leaf colonization by shiga-toxigenic Escherichia..., Saldaña [/bib_ref] [bib_ref] Flagella mediate attachment of enterotoxigenic Escherichia coli to fresh salad leaves, Shaw [/bib_ref]. [bib_ref] Plant stomata function in innate immunity against bacterial invasion, Melotto [/bib_ref]. Purified LPS from Salmonella triggers of ROS production and extracellular alkalinization in tobacco cell suspension (Shirron and Yaron, 2011) but not on tomato leaves [bib_ref] Salmonella colonization activates the plant immune system and benefits from association with..., Meng [/bib_ref] suggesting that LPS recognition may be either dependent on experimental conditions or variable among plant species. Genetic evidence suggests that the high activity of SAdependent BGL2 gene promoter in Arabidopsis is dependent on the presence of LPS in E. coli O157:H7 as higher activity of this promoter was observed in the wild type bacterial as compared to its LPS mutant [bib_ref] Influence of the plant defense response to Escherichia coli O157:H7 cell surface..., Seo [/bib_ref]. However, LPS-dependent responses seem not to be sufficient to restrict bacterial survival on plants as the population titer of E. coli O157:H7 LPS mutant or wild type in plant is essentially the same [bib_ref] Influence of the plant defense response to Escherichia coli O157:H7 cell surface..., Seo [/bib_ref]. Additionally, live S. Typhimurium cells do not induce ROS in epidermal tissue of tobacco [bib_ref] Active suppression of early immune response in tobacco by the human pathogen..., Shirron [/bib_ref] suggesting that, at least Salmonella, can suppress LPS-induced ROS and extracellular alkalinization.
## Lps perception
## Lipopolysaccharide (lps) is a component of the cell wall of
Similar to flagellin, the O-antigen moiety of LPS is not only important for plant perception of bacterial cells, but also for bacterial attachment, fitness, and survival on plants [bib_ref] The role of cellulose and O-antigen capsule in the colonization of plants..., Barak [/bib_ref] [bib_ref] Salmonella enterica strains belonging to O serogroup 1, 3, 9 induce chlorosis..., Berger [/bib_ref] [bib_ref] Differential regulation of Salmonella typhimurium genes involved in Oantigen capsule production and..., Marvasi [/bib_ref].
## Functional output of bacterium perception
One of the earliest PTI responses in plants is stomatal closure that greatly decreases the rate of pathogen entry into plant's internal tissues. This response requires molecular components of PTI including such as flagellin and LPS perception and hormone perception and signaling [bib_ref] Plant stomata function in innate immunity against bacterial invasion, Melotto [/bib_ref] [bib_ref] Role of stomata in plant innate immunity and foliar bacterial diseases, Melotto [/bib_ref] [bib_ref] Guarding the green: pathways to stomatal immunity, Sawinski [/bib_ref]. Stomatal immunity is also triggered by the presence of human pathogens S. enterica serovar Typhimurium SL1344 and E. coli O157:H7 [bib_ref] Plant stomata function in innate immunity against bacterial invasion, Melotto [/bib_ref] [bib_ref] Internalization of Salmonella enterica in leaves is induced by light and involves..., Kroupitski [/bib_ref] [bib_ref] Escherichia coli O157:H7 induces stronger plant immunity than Salmonella enterica Typhimurium SL1344, Roy [/bib_ref] , albeit at various levels. For instance, E. coli O157:H7 induces a strong stomatal immunity and Salmonella SL1344 elicits only a transient stomatal closure in both Arabidopsis [bib_ref] Plant stomata function in innate immunity against bacterial invasion, Melotto [/bib_ref] [bib_ref] Escherichia coli O157:H7 induces stronger plant immunity than Salmonella enterica Typhimurium SL1344, Roy [/bib_ref] and lettuce [bib_ref] Internalization of Salmonella enterica in leaves is induced by light and involves..., Kroupitski [/bib_ref] [bib_ref] Escherichia coli O157:H7 induces stronger plant immunity than Salmonella enterica Typhimurium SL1344, Roy [/bib_ref] suggesting that the bacterial strain SL1344 can either induce weaker or subvert stomata-based defense. Active suppression of stomatal closure by SL1344 may be unlikely because it cannot re-open dark-closed stomata [bib_ref] Escherichia coli O157:H7 induces stronger plant immunity than Salmonella enterica Typhimurium SL1344, Roy [/bib_ref]. However, it is possible that signaling pathways underlying bacterium-triggered and dark-induced stomatal closure are not entirely overlapping and SL1344 acts on immunity-specific signaling to subvert stomatal closure.
## Plant intracellular response to human pathogens
Recognition of PAMPs by PRRs leads to several hallmark cellular defense responses that are categorized based on the timing of response. [bib_ref] Pathogen-associated molecular patterntriggered immunity: veni, vidi, Zipfel [/bib_ref] have discussed that early responses occur within seconds to minutes of recognition including ion fluxes, extracellular alkalinization, and oxidative burst. Intermediate responses occur within minutes to hours including stomatal closure, ethylene production, mitogen-activated protein kinase (MAPK) signaling, and transcriptional reprogramming. Late responses occur from hours to days and involve callose deposition, salicylic acid accumulation, and defense gene transcription.
These hallmark plant cellular defenses have also been tested for both E. coli and S. enterica [fig_ref] FIGURE 2 |: Plant cellular defense responses against human pathogens [/fig_ref]. In particular, S. enterica infection results in the induction of MPK3/MPK6 kinase activity and plant defense-associated genes PDF1.2, PR1, and PR2 in Arabidopsis leaves [bib_ref] The dark side of the salad: Salmonella Typhimurium overcomes the innate immune..., Schikora [/bib_ref] as well as PR1, PR4, and PR5 in lettuce [bib_ref] Physiological and molecular responses of Lactuca sativa to colonization by Salmonella enterica..., Klerks [/bib_ref]. MPK6 activation in Arabidopsis is independent of FLS2 [bib_ref] The dark side of the salad: Salmonella Typhimurium overcomes the innate immune..., Schikora [/bib_ref] , indicating that flagellin is not the only active PAMP of Salmonella and plant response to other PAMPs may converge at MAPK signaling. Direct comparison of the PR1 gene expression in Arabidopsis indicated that both E. coli O157:H7 and Salmonella SL1344 are able to induce this defense marker gene, however at difference levels [bib_ref] Escherichia coli O157:H7 induces stronger plant immunity than Salmonella enterica Typhimurium SL1344, Roy [/bib_ref]. The PR1 gene induction is low in SL1344-infected plants indicating that immune responses are either weaker or are suppressed by Salmonella.
A few studies [fig_ref] Table 1 |: Experimental conditions used in the studies reporting plant response to pathogenic Salmonella... [/fig_ref] have addressed the role of plant hormones in response to endophytic colonization of human bacterial pathogens:
## Ethylene signaling
The ethylene-insensitive mutant of Arabidopsis, ein2, supports higher Salmonella 14028 inside whole seedlings as compared to the wild type Col-0 plants [bib_ref] The dark side of the salad: Salmonella Typhimurium overcomes the innate immune..., Schikora [/bib_ref]. Furthermore, addition of a specific inhibitor of ethylene mediated signaling, 1methylcyclopropene (1-MCP), to the growth medium resulted in increased S. enterica 14028 endophytic colonization of Medicago truncatula, but not M. sativum, roots and hypocotyls [bib_ref] Regulation of enteric endophytic bacterial colonization by plant defenses, Iniguez [/bib_ref] suggesting that the role of endogenous ethylene signaling maybe be specific to each plant-bacterium interaction. However, ethylene signaling may play a contrasting role during fruit contamination. Tomato mutants (rin and nor) with defects in ethylene synthesis, perception, and signal transduction show significantly reduced Salmonella proliferation within their fruits as compared to the wild type control [bib_ref] Ethylene signaling affects susceptibility of tomatoes to Salmonella, Marvasi [/bib_ref].
## Jasmonic acid
Similar to the ein2 mutant, the coronatine-insensitive mutant of Arabidopsis, coi1-16, also supports high Salmonella 14028 inside whole seedlings [bib_ref] The dark side of the salad: Salmonella Typhimurium overcomes the innate immune..., Schikora [/bib_ref]. Along with the induction of the jasmonate-responsive gene PDF1.2 addressed in the same study and mentioned above, it seems that jasmonate signaling is also an important component to restrict Salmonella infection in, at least, Arabidopsis. These results are surprising as coi1 mutants are well known to have increased resistant to various bacterial pathogen of plants, such as P. syringae, but not to fungal or viral pathogens [bib_ref] Arabidopsis mutants selected for resistance to the phytotoxin coronatine are male sterile,..., Feys [/bib_ref] [bib_ref] Resistance to Pseudomonas syringae conferred by an Arabidopsis thaliana coronatine-insensitive (coi1) mutation..., Kloek [/bib_ref].
## Salicylic acid
Two genetic lines of Arabidopsis has been extensively used to determine the role of salicylic acid (SA) in plant defenses against phytopathogens, the transgenic nahG plant that cannot accumulate SA [bib_ref] Characterization of tobacco plants expressing a bacterial salicylate hydroxylase gene, Friedrich [/bib_ref] and the null mutant npr1 that is disrupted in both SA-dependent and -independent defense responses [bib_ref] Characterization of Arabidopsis enhanced disease susceptibility mutants that are affected in systemically..., Ton [/bib_ref]. Both of these plant lines support higher populations of Salmonella 14028 inside their roots [bib_ref] Regulation of enteric endophytic bacterial colonization by plant defenses, Iniguez [/bib_ref] and seedlings [bib_ref] The dark side of the salad: Salmonella Typhimurium overcomes the innate immune..., Schikora [/bib_ref] as compared to the wild type plant. NPR1-dependent signaling is important reduce the population of the curli-negative strain of E. coli O157:H7 43895 but not for the curli-positive strain 86-24 in Arabidopsis leaves [bib_ref] Influence of the plant defense response to Escherichia coli O157:H7 cell surface..., Seo [/bib_ref]. Although only a few strains of Salmonella and E. coli have been used, there is an emerging patterns suggesting that SA itself and activation of SA-signaling can potentially restrict HPOP. In attempts to understand the overall cellular transcriptional response to human bacterial pathogens, global transcriptomic analyses have been used. [bib_ref] Genome-wide transcriptional analysis of the Arabidopsis thaliana interaction with the plant pathogen..., Thilmony [/bib_ref] showed that E. coli O157:H7 regulates PTI-associated genes in Arabidopsis leaves, albeit in a flagellin-independent manner. A similar transcriptomic analysis with medium-grown Arabidopsis seedlings 2h after inoculation with S. enterica serovar Typhimurium 14028, E. coli K-12, and P. syringae pv. tomato DC3000 showed a strong overlap among genes responsive to each bacterial infection suggesting a common mechanism of plant basal response toward bacteria [bib_ref] Conservation of Salmonella infection mechanisms in plants and animals, Schikora [/bib_ref]. Gene expression analysis of Medicago truncatula seedlings root-inoculated with only two bacterial cells per plant indicated that 83 gene probes (30-40% of each data set) were commonly regulated in response to S. enterica and E. coli O157:H7 [bib_ref] Response of Medicago truncatula Seedlings to Colonization by Salmonella enterica and Escherichia..., Jayaraman [/bib_ref]. All together, these studies indicate that each human pathogenic bacterium can modulate specific plant genes beyond a basal defense response; however the mechanisms for plant-bacterium specificity are largely unknown.
## Can human pathogenic bacteria induce eti in plant cells?
Successful virulent pathogens of plants are able to defeat this army plant defense by employing its own set of artillery (such as the type three secretion system effectors and phytotoxins) and cause disease in the host plant [bib_ref] Virulence strategies of plant pathogenic bacteria, Melotto [/bib_ref] [bib_ref] Pseudomonas syringae pv. tomato DC3000: a model pathogen for probing disease susceptibility..., Xin [/bib_ref]. In incompatible interactions (i.e., low bacterial colonization and no disease on leaves), the host plant already has pre-evolved molecules (R proteins) that recognize these effectors and cause a specific defense response to this pathogen. This specific response is called effector-triggered immunity (ETI). Because the type 3-secretion system (T3SS) is important for the virulence of both animal and plant pathogenic bacteria on their natural hosts as evidenced by the use of bacterial mutants, it is reasonable to expect that T3SS would be important for HPOP as well. However, animal and plant cell surfaces are structurally different; the plant cells wall seems to be impenetrable by the secretion needle of the extracellular animal pathogens (Salmonella and E. coli) as discussed by [bib_ref] Type III protein secretion mechanism in mammalian and plant pathogens, He [/bib_ref] raising the question of how these effectors can reach the plant cytoplasm and interfere with plant defenses. To date, there is no evidence for the ability of human pathogens to inject T3SS effectors inside plant cells. It is possible that the T3SS is still active on the plant cell surface and the effectors are secreted into the plant apoplast. If that is the case, however, plant membrane receptors would be necessary to recognize the effectors and trigger plant cellular responses. Nevertheless, it has been observed that the T3SS mutant of E. coli O157:H7, escN, has reduced ability to attach to and colonize baby spinach leaves similar to the fliC mutant [bib_ref] Surface structures involved in plant stomata and leaf colonization by shiga-toxigenic Escherichia..., Saldaña [/bib_ref]. Furthermore, apoplastic population of T3SS structural mutants of S. enterica serovar Typhimurium 14028 is smaller than that of the wild type bacterium in Arabidopsis leaves [bib_ref] Conservation of Salmonella infection mechanisms in plants and animals, Schikora [/bib_ref] and plant defense-associated genes are up-regulated for longer time by the prgH mutant than wild type Salmonella in Arabidopsis seedlings [bib_ref] Salmonella enterica flagellin is recognized via FLS2 and activates PAMP-triggered immunity in..., Garcia [/bib_ref]. Contrary to these findings, [bib_ref] Regulation of enteric endophytic bacterial colonization by plant defenses, Iniguez [/bib_ref] reported that two Salmonella 14028 T3SS-SPI1, the structural mutant spaS and the effector mutant sipB, hypercolonize roots and hypocotyls of M. sativum and fail to induce SA-dependent PR1 promoter in Arabidopsis leaves. More studies need to be conducted to conclude whether T3SS of Salmonella acts as "recognizable" surface structure similar to flagellum and/or as a conduit to deliver effectors in plant tissues and trigger ETI. It is worth mentioning that T3SS and effectors of the phytopathogen P. syringae pv. syringae have functions on ETI as well as bacterial fitness on plant surface [bib_ref] Type III secretion and effectors shape the survival and growth pattern on..., Lee [/bib_ref] and the filamentous T3SS protein EspA is required for E. coli O157:H7 attachment to arugula leaves.
The invA structural mutant, that is defective in all T3SS-1 system-associated phenotypes, induces high ROS and extracellular alkalinizing in tobacco BY-2 cell suspension and hypersensitive reaction (HR) in tobacco leaves as compared to the wild type strain [bib_ref] Active suppression of early immune response in tobacco by the human pathogen..., Shirron [/bib_ref] suggesting that T3SS is important for this suppression of immunity. However, [bib_ref] Active suppression of early immune response in tobacco by the human pathogen..., Shirron [/bib_ref] also reported that plant response to the regulatory mutant phoP that modulates the expression of many effector proteins and membrane components [bib_ref] Salmonellae PhoPQ regulation of the outer membrane to resist innate immunity, Dalebroux [/bib_ref] , is no different to that of the wild type bacterium. These findings raised the question whether the phenotypes observed are due to the T3SS structure itself or due to the translocated effectors. A recent report shows that transient expression of the type three effector of Salmonella 14028 SseF in tobacco plants elicits HR, and this response is dependent on the SGT1 protein . This study suggests that SseF can induce resistant-like response in plants and requires resistance (R) protein signaling components. and [bib_ref] Active suppression of early immune response in tobacco by the human pathogen..., Shirron [/bib_ref] also showed that Salmonella 14028, which is able to deliver the SseF effector, cannot induce HR or any disease-like symptoms in tobacco leaves. Thus, it remains to be determined what would be the biological relevance of ETI in the Salmonella and other human pathogenic bacteria in their interaction with plants in nature.
## Genotypic variability in plant-salmonella and plant-e. coli interactions
Although S. enterica and E. coli O157:H7 have not been traditionally known to be closely associated with plants and modulate plant's physiology, the evidence tells us otherwise. An arms-race evolution in both the human pathogen and the plant is therefore, expected. A few studies (methodology details described in [fig_ref] Table 1 |: Experimental conditions used in the studies reporting plant response to pathogenic Salmonella... [/fig_ref] have addressed whether genetic variability among plant species or within the same plant species (i.e., cultivars, varieties, and ecotypes) can be correlated with differential bacterial behavior and/or colonization of plants. [bib_ref] Colonization of tomato plants by Salmonella enterica is cultivar dependent, and type..., Barak [/bib_ref] described that different tomato cultivars can harbor different levels of S. enterica population after inoculation via water (sprinkler imitation) indicating plant factors may control the ability of bacterial to colonize the phyllosphere. However, they also found that the cultivar with the smallest S. enterica population also had the lowest number of speck lesions when infected with the tomato pathogen Pst DC3000 [bib_ref] Colonization of tomato plants by Salmonella enterica is cultivar dependent, and type..., Barak [/bib_ref] , suggesting that strong basal defense in this cultivar may account for low bacterial colonization. On a comparative study of S. enterica contamination of several crop species, [bib_ref] Differential attachment and subsequent contamination of agricultural crops by Salmonella enterica, Barak [/bib_ref] reported that seedlings from Brassicaceae family have higher contamination than carrot, tomato, and lettuce when grown on contaminated soil. Seedling contamination correlated with the Salmonella population in the phyllosphere of all crop species, except tomato. [bib_ref] Salmonella Typhimurium internalization is variable in leafy vegetables and fresh herbs, Golberg [/bib_ref] reported variations in internalization of Salmonella SL1344 in different leafy vegetables and fresh herbs using confocal microscopy. Internalization incidence (% of microscopic fields containing bacterial cells) was high in iceberg lettuce and arugula, moderate in romaine lettuce, red lettuce, basil, and low in parsley and tomato. Attraction to stomata was seen in iceberg lettuce and basil, not in arugula, parsley, and tomato. [bib_ref] Leaf age as a risk factor in contamination of lettuce with Escherichia..., Brandl [/bib_ref] reported that the age of romaine lettuce leaves is correlated with population size of E. coli O157:H7 and S. enterica Thompson on leaves. Young leaves (inner) harbor greater number of cells than middle aged leaves. These authors also observed that exudates on the surface of younger leaves have higher nitrogen content than that of older leaves, which may contribute to determining the bacterial population size on the leaf. Thus, it is tempting to speculate that the genetic variability existent among plant genotypes regarding the chemical composition of their organ exudates may be a determinant for human pathogen behavior (such as chemotaxis and tropism toward stomata and roots) and ability to colonize plants.
Finally, [bib_ref] Effect of route of introduction and host cultivar on the colonization, internalization,..., Mitra [/bib_ref] studied the effect of different methods of inoculation on internalization and survival of E. coli O157:H7 in three cultivars of spinach. Among the organs studied, the spinach phylloplane and the stem provided the most and least suitable niche for this bacterium colonization, respectively. Although the leaf surface was the best "territory" for E. coli, the leaf morphologies of each cultivar affected the ability of this bacterium to survive.
Collectively, all these studies point out that the plant genotype, age, leaf morphology, chemical composition of exudates, and the primarily infected organ affect the outcome of bacterial colonization of plants and the process may not be a generalized phenomenon, consequently shaping specific human pathogen and plant interactions.
## Concluding remarks
The fundamental understanding of plant association with human bacterial pathogens that do not cause visual or macroscopic symptom in the plant, but yet are major food contaminants, are in its infancy. Both plant and bacterial factors are critical for these cross-kingdom interactions and emerging evidence suggests an overlap between plant molecular responses to human pathogens and phytopathogens. The future challenge will be to determine how these interactions differ. As this field of research is relatively new, we see differences in conclusions from different laboratories regarding multiplication vs. decline in bacterial populations overtime and disease-like symptoms vs. HR on inoculated plants. These differences are mainly associated with differences in methods of inoculation, bacterial strains, inoculum concentration, plant age, and plant cultivation methods (e.g., growth on medium, soil, or hydroponic solutions). Standard procedures for model systems, consensus, and collaborations must be developed among food scientists, microbiologists, plant pathologists, and molecular biologists to elucidate the specificity of each plantbacterium interaction and avoid discrepancies in making general conclusions. A major point to be resolved is whether the observed plant defenses against Salmonella and its PAMPs are due to low recognition and/or active suppression. If Salmonella suppression of the plant immunity is a cause of weak defense responses, the major question becomes what is the responsible factor? This line of research might lead to a whole new paradigm that otherwise could not be revealed by only studying plant associations with its own natural pathogens.
[fig] FIGURE 2 |: Plant cellular defense responses against human pathogens. (A) Upon reception of PAMP (flagellin, LPS) through PRR (FLS2 and putatively others), Salmonella spp. trigger downstream plant defense responses which include ROS production, MPK3/6, salicylic acid (SA) signaling through NPR1, jasmonic acid (JA) and ethylene (ET) signaling, defense-associated gene induction, and extracellular alkalinization. All these cellular events ultimately lead to stomatal closure, antimicrobial activity, and plant defense. (B) Escherichia coli PAMPs (curli, LPS, flagellin, EPS) are also perceived by PRRs (FLS2 and putatively others) present on plant cell surface which triggers the induction of the SA-dependent BGL2 promoter activity and PR1 gene expression. Only components that have been directly demonstrated experimentally are included in the diagram. Plant defense responses in case of both these human pathogens are strain specific as well as plant cultivar specific. [/fig]
[table] Table 1 |: Experimental conditions used in the studies reporting plant response to pathogenic Salmonella and E. coli. [/table]
|
Subsets and Splits