Dysfunctional adaptive immune response in adolescents and young adults with suicide behavior

Journal Pre-proof Dysfunctional Adaptive Immune Response in Adolescents and Young Adults with Suicide Behavior Manish K. Jha, Ling Cai, Abu Minhajuddin, Cherise Chin Fatt, Jennifer L. Furman, Bharathi S. Gadad, Brittany L. Mason, Tracy L. Greer, Jennifer L. Hughes, Guanghua Xiao, Graham Emslie, Betsy Kennard, Taryn Mayes, Madhukar H. Trivedi

PII:

S0306-4530(19)30613-4

DOI:

https://doi.org/10.1016/j.psyneuen.2019.104487

Reference:

PNEC 104487

To appear in:

Psychoneuroendocrinology

Received Date:

24 June 2019

Revised Date:

17 October 2019

Accepted Date:

18 October 2019

Please cite this article as: Jha MK, Cai L, Minhajuddin A, Fatt CC, Furman JL, Gadad BS, Mason BL, Greer TL, Hughes JL, Xiao G, Emslie G, Kennard B, Mayes T, Trivedi MH, Dysfunctional Adaptive Immune Response in Adolescents and Young Adults with Suicide Behavior, Psychoneuroendocrinology (2019), doi: https://doi.org/10.1016/j.psyneuen.2019.104487

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier.

Dysfunctional Adaptive Immune Response in Adolescents and Young Adults with Suicide Behavior

Manish K. Jha, MD,1,2 Ling Cai, PhD,3 Abu Minhajuddin, PhD,1 Cherise Chin Fatt, PhD,1 Jennifer L. Furman, PhD,1 Bharathi S. Gadad, PhD,1,5 Brittany L. Mason, PhD,1 Tracy L. Greer, PhD,1 Jennifer L. Hughes, PhD, MPH1 Guanghua Xiao, PhD,3 Graham Emslie, MD,4 Betsy Kennard, PsyD,4 Taryn Mayes, MA,1 Madhukar H.

ro

of

Trivedi, MD1*

1. Center for Depression Research and Clinical Care, UT Southwestern Medical Center, Dallas, TX

-p

2. Department of Psychiatry, Icahn School of Medicine at Mount Sinai, New York, NY

re

3. Quantitative Biomedical Research Center, UT Southwestern Medical Center, Dallas, TX 4. Department of Psychiatry, UT Southwestern Medical Center, Dallas, TX

6.

ur na

Address correspondence to:

lP

5. Department of Psychiatry, Texas Tech University Health Science Center, El Paso, Texas

Madhukar H. Trivedi, M.D. Professor of Psychiatry

Julie K. Hersh Chair for Depression Research and Clinical Care

Jo

Betty Jo Hay Distinguished Chair in Mental Health Director, Center for Depression Research and Clinical Care Peter O’Donnell Jr. Brain Institute University of Texas Southwestern Medical Center Email: [email protected]

1

Highlights  

ro

of

 

Interleukin 4 (IL-4) levels are lower in adolescents and young adults with depression and suicidality. Other inflammation-related markers didn’t differ in those with recent suicidality vs. healthy controls and those at-risk. IL-4 levels were negatively correlated with self-reported depression severity. Lower IL-4 levels were associated with higher autoantibody levels in an unrelated sample of depressed outpatients.

-p

Abstract

re

Background: Immune system dysfunction has been implicated in the pathophysiology of suicide behavior. Here, we conducted an exploratory analysis of immune profile differences of three groups of adolescents

lP

and young adults (ages 10-25 years): healthy controls (n=39), at risk of major depressive disorder (MDD; at-risk, n=33), and MDD with recent suicide behavior/ ideation (suicide behavior, n=37).

ur na

Methods: Plasma samples were assayed for chemokines and cytokines using Bio-Plex Pro Human Chemokine 40-plex assay. Log-transformed cytokine and chemokine levels were compared after controlling for age, gender, body mass index, race, ethnicity, and C-reactive protein (CRP) levels. In posthoc analyses to understand the effect of dysregulated immune markers identified in this exploratory

Jo

analysis, their association with autoantibodies was tested in an unrelated sample (n=166). Results: Only levels of interleukin 4 (IL-4) differed significantly among the three groups [false discovery rate (FDR) adjusted p=0.0007]. Participants with suicide behavior had lower IL-4 [median=16.8 pg/ml, interquartile range (IQR)=7.9] levels than healthy controls (median=29.1 pg/ml, IQR=16.1, effect size [ES]=1.30) and those at-risk (median=24.4 pg/ml, IQR=16.3, ES=1.03). IL-4 levels were negatively 2

correlated with depression severity (r= -0.38, p=0.024). In an unrelated sample of outpatients with MDD, levels of IL-4 were negatively correlated (all FDR p<0.05) with several autoantibodies [54/117 in total and 12/18 against innate immune markers]. Conclusions: Adolescent and young adult patients with recent suicide behavior exhibit lower IL-4 levels. One biological consequence of reduced IL-4 levels may be increased risk of autoimmunity.

of

Key words: Depression; Suicidality; Suicide behavior; Adaptive Immunity; Autoimmunity; IL-4; Type 2

Jo

ur na

lP

re

-p

ro

immune response

3

1

Introduction

Deaths due to suicide have been increasing in the United States over the last two decades (NIMH, 2019). Notably, between the ages of 10-34 years, suicide is the second leading cause of mortality (Heron, 2018; NIMH, 2019). Inflammation has gained recent attention for its role in etiology of suicide (Brundin et al., 2017). This is consistent with reports of elevated markers of inflammation such as interleukin 6 (IL-6) and

of

C-reactive protein (CRP) in patients with major depressive disorder (MDD) who report high suicidal ideation than those with lower suicidal ideation and healthy controls (O'Donovan et al., 2013). Similarly,

ro

individuals who attempt suicide also exhibit increased markers of inflammation, both in peripheral circulation (Janelidze et al., 2011; Nassberger and Traskman-Bendz, 1993) and in cerebrospinal fluid

-p

(Lindqvist et al., 2011; Lindqvist et al., 2009). Increased gene and protein expression levels of

re

proinflammatory cytokines have been reported in postmortem brain tissue from teenage suicide victims as compared to controls (Pandey et al., 2012). Similarly, altered levels of cytokines and markers of innate

lP

immune response have also been reported in postmortem brain tissue from adults who died by suicide compared to controls (Pandey et al., 2019; Pandey et al., 2014; Pandey et al., 2018; Tonelli et al., 2008).

ur na

Contrary to these reports, other studies have reported decreased levels of pro-inflammatory cytokines in actively suicidal adolescent patients with MDD versus non-suicidal adolescents with MDD (Gabbay et al., 2009).

These somewhat inconsistent findings may be related to the complexity of immune system which often

Jo

involves reciprocal interactions between different arms to maintain homeostasis. Within the context of T-cell mediated immune response, the differentiation of naïve T-cell after T-cell receptor activation depends on the milieu of cytokines with the presence of IL-4 resulting differentiation to T-helper 2 (Th2) cells which in turn may suppress the activity of T-helper 1 (Th1) cell-mediated and vice versa (Mosmann and Coffman, 1989). Thus, there is a need to systematically study a broad set of immune markers

4

(Pariante, 2019). Additionally, confounding effects of obesity should be considered, as recent findings suggest that obesity partly explains dysfunctional immune response in depression (Shelton et al., 2015). Thus, in this report, we conducted an exploratory analysis of immune dysfunction associated with recent suicide behavior. We focused on adolescents due to the mortality burden of suicide in this age group. We compared the levels of immune markers in peripheral circulation of three groups of adolescents and

of

young adults: 1) those who were admitted to an intensive outpatient program (IOP) for suicidality (hereafter referred to as suicide behavior), 2) healthy controls, and 3) those not currently depressed but

ro

considered to be at risk for depression (hereafter referred to as at-risk). We operationalized the at-risk group as individuals who were currently physically and mentally healthy but either were diagnosed with

-p

or treated for any anxiety, conduct, or alcohol/substance use disorder during their lifetime or had a family

re

history (mother or father) of mood disorders, alcohol/substance use disorders, or suicide behaviors (attempts, deaths). Healthy controls did not have any of the aforementioned at-risk characteristics. In

lP

addition to controlling for key variables that may affect immune function [age, sex, race and body mass index (BMI)], we also controlled for CRP to account for any acute inflammatory changes.

ur na

We used a sample of convenience (n=166) to further understand the biological effects of immune markers that were identified as altered in our exploratory analysis. Plasma specimens from this sample of outpatients with major depressive disorder (MDD) were assayed for levels of immune markers and autoantibodies [immunoglobulin (Ig) G against self-antigens] for unrelated projects (Gadad et al., 2017;

Jo

Jha et al., 2017b). We focused on autoimmunity as previous reports have suggested that suicidal depressed patients have activation of Th1 cells (Mendlovic et al., 1999). Additionally, previous reports have found not only an increased cross-sectional association between autoimmune disorders and depression (Siegmann et al., 2018) but that history of depressive disorder earlier in life is associated with greater prevalence of autoimmune disorders later in life (Andersson et al., 2015). Suppression of Th2 cell mediated immune response may be a potential mechanism linking depression and risk of autoimmunity. 5

For example, a recent report found that treatment naïve outpatients with MDD had markedly reduced IL4 as compared to healthy controls (Syed et al., 2018). Hence, we conducted post-hoc analyses testing for association between altered immune markers (identified in above-described exploratory analysis) and levels of autoantibodies in the unrelated sample of convenience after controlling for age, sex, race, BMI, and CRP.

of

2 Methods All participants were recruited and all study related procedures were completed only after the

ro

documentation of informed consent. In case of participants of age less than 18 years, assent was obtained

-p

from the participants in addition to the informed consent from their parents/legal guardians. The studies involving adolescents and young adults (suicide behavior, at-risk, and healthy controls) were approved by

re

the Institutional Review Board (IRB) at the UT Southwestern Medical Center at Dallas. The study involving outpatients with MDD was conducted as a multicenter clinical trial [Combining Medications to Enhance

lP

Depression Outcomes (CO-MED)] which was reviewed and approved by the Institutional Review Boards at UT Southwestern Medical Center at Dallas, the University of Pittsburgh Data Coordinating Center, each

ur na

participating regional center, and all relevant clinics. 2.1 Participants

2.1.1 Participants with suicide behavior

Jo

Participants included adolescents (ages 12-18 years) who were admitted to the suicide prevention intensive outpatient program (IOP) of a large urban hospital who provided an informed consent for a biomarker study. The IOP has been described previously by Kennard et al. (Kennard et al., 2019). Briefly, admission was based on the clinical judgment of licensed psychologists, therapists or post-doctoral fellows after completing evaluations within a few days after the suicidal event (suicide attempt or worsening of ideation that warranted emergency services; together referred to as suicide behavior). The length of 6

treatment, comprised of 3 hours of group therapy twice weekly, was individualized based on clinical need but generally varied from 4-8 weeks in duration. Use of psychotropic medications was per the clinical necessity and guided by the clinicians working with individual patient. Participants for this study (n=37) were recruited from the IOP from July 2016 to January 2018. Participants completed the adolescent selfreport version of the Quick Inventory of Depressive Symptom Severity (QIDS-A) as a measure of

of

depression severity (Bernstein et al., 2010). 2.1.2 Healthy controls and at-risk participants

ro

Healthy controls (n=39) and at-risk (n=33) participants were recruited from an ongoing study, titled the Resilience in Adolescent Development (RAD, NCT03458936). The RAD study enrolls children, adolescents

-p

and young adults (ages 10-24 years) who may be at-risk for mood disorders in a 10-year longitudinal

re

prospective study designed to uncover socio-demographic, lifestyle, clinical, psychological and neurobiological factors that contribute to resilience against MDD. For this report, participants were

lP

recruited during Dec 2016 to Nov 2017. This report includes only the initial baseline visit, at which time participants were not diagnosed with depressive disorder or suicidality. All participants in RAD study were

ur na

medically healthy and did not have any ongoing psychiatric illness at the time of study participation. 2.1.3 Combining Medications to Enhance Depression Outcomes trial Participants of CO-MED trial (NCT00590863) who provided plasma samples at baseline as part of an optional sub-study (n=166) were included in this report. Described in detail by Rush et al. (Rush et al.,

Jo

2011), the CO-MED trial recruited treatment-seeking adult outpatients (ages 18-75 years) with chronic or recurrent MDD who were not currently taking psychotropic medications for a 3-arm single-blind randomized controlled trial of escitalopram monotherapy versus combinations of bupropion plus escitalopram and venlafaxine plus mirtazapine. Inclusion was restricted to 18-75-year-old treatmentseeking MDD outpatients with nonpsychotic chronic (current episode ≥2 years) or recurrent depression with current episode ≥2 months and a baseline 17-item Hamilton Rating Scale (HRSD17) ≥16 (21). 7

2.2 Biomarkers 2.2.1 Biospecimen collection and biomarker measurements for suicide behavior, at-risk, and healthy control groups Peripheral venous samples were collected in EDTA tubes at a single center using the same collection and processing protocol. The timing of blood draw depended on the availability of participants. For most of

of

the participants (72.7% of healthy control, 89.3% of at-risk, and 81.6% of participants with suicide behavior), blood draw was done between 10 AM and 2 PM. Within two hours of blood collection, samples

ro

were centrifuged at 1200 rpm at room temperature for 15 minutes to isolate plasma. Extracted plasma

-p

was aliquoted and stored at -80°C until further assay. Levels of 40 cytokines and chemokines were measured in de-identified samples by the Genomics and Microarray Core at the UT Southwestern Medical

re

Center (UTSW Core) using a luminescence multiplex bead-based sandwich immunoassay kit (Bio-Plex ProTM 40-plex Human Chemokine Panel, Bio-Rad), according to the manufacturer’s protocol. Cytokines

lP

and chemokines, detailed in Table 2 along with assay range, were measured in pg/mL using Bio-Plex Manager software version 5.0 (Bio-Rad) and interpreted only if the intra- and inter-assays coefficients of

ur na

variation were less than 10% of detection limits (or precision range) specified by manufacturer. 2.2.2 Biospecimen collection and biomarker measurements in CO-MED trial Peripheral venous samples from CO-MED trial participants (n=166) were collected in EDTA tubes and

Jo

transported overnight to the Biologic Core of the National Institute of Mental Health Repository and Genomics Resource (NIMH RGR) where plasma was extracted by centrifuging blood samples at 2500 rpm for 10 minutes at room temperature, aliquoted, and stored at -80°C. All samples from CO-MED trial for this report were obtained from the NIMH RGR core and transported to UT Southwestern on dry ice for storage at -80°C until immediately prior to assays without any freeze/thaw cycles. Cytokine levels were measured in de-identified samples by the UTSW Core using a Bioplex Pro™ human cytokine standard 278

plex kit (Bio-Rad Laboratories, Hercules, CA, USA) as reported previously (Jha et al., 2017c). Cytokine levels were measured in pg/mL and interpreted only if the intra- and inter-assays coefficients of variation were less than 10% of detection limits (or precision range) specified by manufacturer. Levels of autoantibodies were measured with an autoantigen microarray (Autoantigen Microarray Super Panel) available through the UTSW Core. The microarray chip used for this panel contains 117 autoantigens and various internal controls and can assay 15 samples/chip. Levels of IgG autoantibodies were measured at laser wavelengths

of

532nm using the Genepix Pro 6.0 software. Net fluorescence intensity was defined as the spot minus

ro

background fluorescence intensity. Detection of autoantibodies using similar microarray methods have been reported previously in patients with lupus (Li et al., 2007; Quan et al., 2014).

-p

2.3 Statistical analyses

re

Descriptive statistics were used to report the clinical and sociodemographic features as well as to present the levels of each biomarker. Biomarkers that were not normally distributed were log2-transformed. In a

lP

post hoc power analysis, the study had a 79% power to detect two-fold difference in immune markers with two-tailed alpha of 0.005 and a sample size of 33 in each group, using the concentration of immune

ur na

markers (IL-6 as the reference) reported recently (Syed et al., 2018). For each of the 40 cytokines and chemokines, separate linear regression analyses tested for differences in levels among the three groups (suicide behavior, at-risk, and healthy control groups) with age, gender, BMI, CRP, race, and ethnicity as covariates. Post-hoc analyses tested for between group differences. Association with depression severity

Jo

was tested with Pearson’s correlation only in the suicide behavior group. Clinical significance of differences in biomarkers was estimated by computing Cohen’s d effect size. In the sample of convenience from CO-MED trial, the effect of interleukin 4 on levels of IgG autoantibodies (listed in supplementary table 1) were tested with separate linear regression analyses after controlling for age, gender, BMI, CRP, race, and ethnicity. Threshold for statistical significance was set at p <0.05 after false discovery rate (FDR) adjustment. 9

3 Results Participants with recent suicide behavior (n=37) were more likely to be female, white, and younger in age as compared to at-risk (n=33) and healthy control (n=39) groups, see Table 1 for details. Notably, the three groups did not differ significantly on BMI or levels of CRP. As described earlier (Jha et al., 2017a; Jha et al., 2017c), participants who did not provide plasma in CO-MED trial (n=499) differed from those included in

of

this report in age (mean age included=44.51 years vs. excluded=42.11, p=0.03) and use of statin medication (included=20.5% vs excluded=13.6%, p=0.03). Summary statistics (median and interquartile

ro

range as distribution was skewed) for each biomarker are presented in Table 2.

-p

3.1 Differences in immune markers among suicide behavior, at-risk and healthy control groups After FDR adjustment and controlling for age, gender, BMI, race, ethnicity, and CRP, only IL-4 (p=0.0007)

re

was significantly different among the three groups of participants (Table 3). Participants with recent

lP

suicide behavior had significantly lower IL-4 levels as compared to those at-risk and healthy controls (Figure 1). The effect size (95% confidence interval) of differences in IL-4 between suicide behavior and at-risk group was 1.03 (0.53, 1.54) and between suicide behavior and healthy control group was 1.30 (0.79,

ur na

1.80). In post-hoc analysis, levels of IL-4 were moderately and negatively correlated with self-reported depression severity in participants with recent suicide behavior (Pearson’s r=-0.38, p=0.024). Due to study criteria, range of depression severity in healthy controls and those at-risk were significantly limited.

Jo

Hence, association of IL-4 with depression severity were not checked in these groups. 3.2 Association of IL-4 with markers of autoimmunity in an unrelated sample of convenience In CO-MED trial participants (n=166), 54 out of 117 autoantibodies were significantly and negatively associated with IL-4 levels even after controlling for age, gender, race, ethnicity, BMI, and CRP, see Figure 2 and Supplementary Table 1 for details. The autoantibodies (location of self-antigen detailed in Supplementary Table 1) that were significantly associated with IL-4 were grouped as those against 10

extracellular self-antigens related to innate immunity (Figure 2, Panel 1), other extracellular self-antigens (Figure 2, Panel 2), intracellular nuclear self-antigens (Figure 2, Panel 3), and intracellular cytoplasmic selfantigens (Figure 2, Panel 4). The strongest association was noted for Beta 2 microglobulin. Also, among the 18 self-antigens included in the microarray that were related to innate immunity, autoantibodies against 12 (66.67%) were significantly and negatively associated with IL-4 levels. Thus, lower levels of IL4 were associated with higher levels of autoantibodies against Complement C3b, C4, C5, and C8. Among

of

antibodies against intracellular self-antigens, 21/48 (43.8%) against nuclear self-antigens and 7/14 (50.0%)

ro

against cytoplasmic/mitochondrial self-antigens were significantly associated with IL-4.

-p

4 Discussion

In this study of adolescents and young adults, those with suicide behavior had significantly lower IL-4

re

levels as compared to those at-risk and healthy controls. Using Cohen’s convention for effect sizes (Cohen, 2013), the magnitude of these differences was large. Levels of pro-inflammatory cytokines and CRP (a

lP

non-specific marker of acute infection) were not significantly different among the three groups. Lower levels of IL-4 were associated with greater depressive symptom severity.

ur na

Several findings of this report are consistent with those reported earlier. Lack of association of proinflammatory cytokines such as IL-1β and TNF-α with suicide behavior are consistent with the findings of Lindqvist et al. who did not find any difference in these cytokines in CSF of patients with recent suicide

Jo

attempt vs. healthy controls (Lindqvist et al., 2009). In the study by Lindqvist et al., IL-6 levels were elevated in those with recent suicide attempt (Lindqvist et al., 2009). This may be related to the use of violent means for suicide attempt in their study, especially as aggressive behaviors have been associated with elevated levels of IL-6 in a previous report (Coccaro et al., 2014). Negative correlation of IL-4 with measures of depressive symptom severity is also consistent with a previous study of patients with MDD (Kim et al., 2008). Reduced levels of IL-4 in this study differs from a previous report which found no

11

significant difference in IL-4 levels between adolescent with active suicidal ideation (n=12) and healthy controls (n=15) (Gabbay et al., 2009). Further, a previous report had found elevated IL-4 mRNA levels in postmortem brain tissue from orbitofrontal cortex of women victims of suicide as compared to normal controls (Tonelli et al., 2008). There was no significant elevation of IL-1β or IL-6 in this report contrary to the finding of a recent meta-analysis, where these immune markers were elevated in both blood and postmortem samples of patients with suicidality as compared to those with no suicidality (Black and Miller,

of

2015). Additionally, unlike this previous meta-analysis, there was no significant elevation of CRP or IL-10

ro

in this report (Black and Miller, 2015). Notably, the meta-analysis found reduced levels of IL-4 production in vitro in suicidal patients as compared to healthy controls (Black and Miller, 2015).

-p

There may be therapeutic implications for the association of suicide behavior and reduced IL-4 levels.

re

These include additional caution with the use of dupilumab, a monoclonal antibody against IL-4 that is approved by the FDA for treatment of atopic dermatitis (Shirley, 2017). Adults and adolescents with atopic

lP

dermatitis have higher severity of depressive symptoms in general and suicidality in particular as compared to controls (Slattery et al., 2011; Yu and Silverberg, 2015). While use of dupilumab was

ur na

associated with overall improvement in depressive symptoms among patients with atopic dermatitis, one patient receiving this medication in the phase 3 study completed suicide (Simpson et al., 2016). Biological mechanisms underlying the association of reduced IL-4 with suicide behavior remain unclear. Role of IL-4 in brain function has gained recent attention (Gadani et al., 2012). Reduced levels of IL-4 and

Jo

reduced responsiveness of microglial cells to IL-4 have both been associated with depression-like behaviors in animal models (Wachholz et al., 2016; Wachholz et al., 2017). Thus, future studies using peripheral markers and positron emission tomography (PET) ligands are needed to evaluate if reduced IL4 levels are associated with microglial activation. Additionally, reduced levels of IL-4 have been associated with decreased production of brain derived neurotrophic factor by astrocytes which in turn has been linked to impaired performance on learning tasks (Derecki et al., 2010). Additionally, IL-4 in meningeal 12

space also promotes alternate activation of meningeal myeloid cells which are anti-inflammatory and restrict the effect of pro-inflammatory cytokines (Derecki et al., 2011). An additional mechanism linking reduced IL-4 with suicide behavior is the increased susceptibility to an autoimmune process (Van Dyken and Locksley, 2013). This may occur as IL-4 serves to limit the activation of Th1 cells in systemic circulation (Maggi et al., 1992). Consistent with this notion, in an unrelated cohort of outpatients with MDD, we found that lower levels of IL-4 were associated with higher levels of a wide range of autoantibodies,

of

especially against immune markers involved in innate immune response. Due to the cross-sectional nature

ro

of our data and lack of information regarding autoantibody levels in groups of adolescents and young adults included in this report, the temporal relationship of suicidal behavior, IL-4, and autoimmune

-p

process is unclear. It is noteworthy that previous reports have identified higher rates of subsequent depression in large-scale epidemiological studies among patients with autoimmune disorder (Benros et

re

al., 2013; Euesden et al., 2017). Thus, it remains unknown whether suicidality precedes the dysfunctional

lP

immune response or vice-versa. However, converging evidence from clinical and pre-clinical studies have implicated immune dysfunction as one of the pathophysiological mechanisms underlying depression (Felger et al., 2018; Hodes et al., 2015; Jha et al., 2018). Our study proposes Th2 suppression as a novel

ur na

mechanistic hypothesis that links inflammation and depression in patients with suicide behavior. There are several limitations of this report including lack of validation of multiplex assays in separate singleplex or enzyme-linked immunosorbent assays. Due to the exploratory nature of this report, these

Jo

findings should be considered hypothesis-generating and necessitate future confirmatory studies. An ongoing large observation study, DFW5000, offers the opportunity to validate and extend these findings in a separate large sample of youths and young adults with additional assays such as levels of cortisol and kynurenine metabolites and characterization of stressful life events (Trivedi et al., 2019). These future analyses should also incorporate the role of hypothalamic pituitary axis and whether genetic or epigenetic factors interact with stress in modulating the immune changes associated with suicidality (Trivedi et al., 13

2019). While including a broad panel of immune markers, this study did not include some key cytokines driving T-cell differentiation (such as IL-12) which restrict the interpretation of the biological mechanisms resulting in reduced IL-4 (such as shift towards Th1 instead of Th2 from naïve T cells). Future studies should characterize T cells in greater detail to extend these findings beyond levels of circulating immune markers (Trivedi et al., 2019). Non-inclusion of adolescents and young adults with MDD but no previous history of suicide behavior precludes drawing inference about whether reduced IL-4 is associated specifically to

of

suicide behavior versus associated more broadly with MDD. As the study recruited volunteers,

ro

generalizability to general population may be limited. dLongitudinal studies are needed to evaluate whether improvement in suicidality and depression is associated with normalization of IL-4 levels.

-p

Additionally, the analyses of relationship between IL-4 and markers of autoimmunity are limited by the

re

lack of healthy control group. 5 Conclusion

lP

We have identified suppression of Th2-mediated immune response as a potential biological mechanism associated with suicide behavior in MDD. Th2 suppression in turn may be associated with increased

Jo

ur na

markers of autoimmunity, especially those involved in innate immunity.

14

Funding The studies of adolescents and young adults included in this report were funded in part by the W.W. Caruth Jr. Foundation, the Elizabeth Jordan Harris Foundation, REDCap (UL1 TR001105), Center for Depression Research and Clinical Care (Principal Investigator: Madhukar Trivedi) and the Hersh Foundation. The CO-MED trial was funded by the National Institute of Mental Health (NIMH, N01 MH-

of

90003). The content is solely the responsibility of the authors and does not necessarily represent the

ro

official views of the various funding organizations.

-p

Declarations of Interest

re

Dr. Jha has received contract research grant from Acadia Pharmaceutical and Janssen Research. Dr. Trivedi has received research support from NIMH, NIDA, J&J, Janssen Research and Development LLC;

lP

has served as a consultant for Alkermes Inc., Allergan, Arcadia Pharmaceuticals Inc., AstraZeneca, Lundbeck, Medscape, MSI Methylation Sciences Inc., Merck, Otsuka America Pharmaceuticals Inc., and

ur na

Takeda Pharmaceuticals Inc. has received an honorarium from the University of Texas San Antonio. Dr. Greer has received research funding from NARSAD and contracted research support from Janssen Research & Development, LLC. She has received honoraria and/or consultant fees from H. Lundbeck A/S and Takeda Pharmaceuticals International, Inc. Dr. Hughes has received funding as a Youth Aware of

Jo

Mental Health (YAM) trainer, consulting for Mental Health in Mind International. Dr. Hughes has also received royalties from Guilford Press and serves as a board member for the American Psychological Association (APA) Division 53, Society for Clinical Child and Adolescent Psychology (SCCAP). She also served as the Chair and is the current Past-Chair of the Association for Behavioral and Cognitive Therapies (ABCT) Child and Adolescent Depression Special Interest Group. Dr. Emslie has received research funding from Duke University, Forest Research Institute, Inc., and Janssen Pharmaceuticals. He 15

has been a consultant for Assurex Health, Inc., INC Research, Inc., Lundbeck, Neuronetics, Inc., Otsuka, Pfizer, Inc., and the Texas Department of State Health Services. Dr. Kennard receives royalties from Guilford Press and currently serves on the Board of Trustees for the Jerry M. Lewis, M.D. Mental Health Research Foundation. Drs. Cai, Chin Fatt, Furman, Gadad, Mason, Minhajuddin, Xiao, and Mrs. Mayes

of

have no conflicts to disclose.

ro

Acknowledgements

-p

We wish to thank the study participants, families, staff, and colleagues who made this project possible. We also acknowledge the editorial assistance of Georganna Carlock, B.A., of the Center for Depression

Jo

ur na

lP

re

Research and Clinical Care.

16

References Andersson, N.W., Gustafsson, L.N., Okkels, N., Taha, F., Cole, S.W., Munk-Jørgensen, P., Goodwin, R.D., 2015. Depression and the risk of autoimmune disease: a nationally representative, prospective longitudinal study. Psychological medicine 45, 3559-3569. Benros, M.E., Waltoft, B.L., Nordentoft, M., Østergaard, S.D., Eaton, W.W., Krogh, J., Mortensen, P.B.,

of

2013. Autoimmune Diseases and Severe Infections as Risk Factors for Mood Disorders: A Nationwide StudyAutoimmune Diseases and Mood Disorder RiskAutoimmune Diseases and Mood Disorder Risk.

ro

JAMA psychiatry 70, 812-820.

Bernstein, I.H., Rush, A.J., Trivedi, M.H., Hughes, C.W., Macleod, L., Witte, B.P., Jain, S., Mayes, T.L., Emslie,

-p

G.J., 2010. Psychometric properties of the Quick Inventory of Depressive Symptomatology in adolescents.

re

International journal of methods in psychiatric research 19, 185-194.

Black, C., Miller, B.J., 2015. Meta-Analysis of Cytokines and Chemokines in Suicidality: Distinguishing

lP

Suicidal Versus Nonsuicidal Patients. Biol Psychiatry 78, 28-37.

Brundin, L., Bryleva, E.Y., Thirtamara Rajamani, K., 2017. Role of Inflammation in Suicide: From

ur na

Mechanisms to Treatment. Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology 42, 271-283.

Coccaro, E.F., Lee, R., Coussons-Read, M., 2014. Elevated plasma inflammatory markers in individuals with intermittent explosive disorder and correlation with aggression in humans. JAMA psychiatry 71, 158-165.

Jo

Cohen, J., 2013. Statistical power analysis for the behavioral sciences. Routledge. Derecki, N.C., Cardani, A.N., Yang, C.H., Quinnies, K.M., Crihfield, A., Lynch, K.R., Kipnis, J., 2010. Regulation of learning and memory by meningeal immunity: a key role for IL-4. J Exp Med 207, 1067-1080. Derecki, N.C., Quinnies, K.M., Kipnis, J., 2011. Alternatively activated myeloid (M2) cells enhance cognitive function in immune compromised mice. Brain, behavior, and immunity 25, 379-385.

17

Euesden, J., Danese, A., Lewis, C.M., Maughan, B., 2017. A bidirectional relationship between depression and the autoimmune disorders - New perspectives from the National Child Development Study. PLoS One 12, e0173015. Felger, J.C., Haroon, E., Patel, T.A., Goldsmith, D.R., Wommack, E.C., Woolwine, B.J., Le, N.A., Feinberg, R., Tansey, M.G., Miller, A.H., 2018. What does plasma CRP tell us about peripheral and central inflammation in depression? Molecular psychiatry.

of

Gabbay, V., Klein, R.G., Guttman, L.E., Babb, J.S., Alonso, C.M., Nishawala, M., Katz, Y., Gaite, M.R.,

ro

Gonzalez, C.J., 2009. A preliminary study of cytokines in suicidal and nonsuicidal adolescents with major depression. Journal of child and adolescent psychopharmacology 19, 423-430.

-p

Gadad, B., Jha, M., Yun, Y., Kim, M., Xiao, G., Emslie, M., Carmody, T., Trivedi, M., 2017. 317. Screening for Auto-Antibodies with Antidepressant Medications Using Glomerular Proteomic Microarray Arrays:

re

Findings from CoMed Clinical Trial. Biological Psychiatry 81, S130-S131.

lP

Gadani, S.P., Cronk, J.C., Norris, G.T., Kipnis, J., 2012. IL-4 in the brain: a cytokine to remember. Journal of immunology (Baltimore, Md. : 1950) 189, 4213-4219. Heron,

M.,

2018.

Deaths:

Leading

Causes

for

2016.

Accessed

on

03MAY2019

at

ur na

https://www.cdc.gov/nchs/data/nvsr/nvsr67/nvsr67_06.pdf. National Vital Statistics Reports 67. Hodes, G.E., Kana, V., Menard, C., Merad, M., Russo, S.J., 2015. Neuroimmune mechanisms of depression. Nature neuroscience 18, 1386-1393.

Jo

Janelidze, S., Mattei, D., Westrin, A., Traskman-Bendz, L., Brundin, L., 2011. Cytokine levels in the blood may distinguish suicide attempters from depressed patients. Brain, behavior, and immunity 25, 335-339. Jha, M.K., Miller, A.H., Minhajuddin, A., Trivedi, M.H., 2018. Association of T and non-T cell cytokines with anhedonia: Role of gender differences. Psychoneuroendocrinology 95, 1-7.

18

Jha, M.K., Minhajuddin, A., Gadad, B.S., Greer, T., Grannemann, B., Soyombo, A., Mayes, T.L., Rush, A.J., Trivedi, M.H., 2017a. Can C-reactive protein inform antidepressant medication selection in depressed outpatients? Findings from the CO-MED trial. Psychoneuroendocrinology 78, 105-113. Jha, M.K., Minhajuddin, A., Gadad, B.S., Greer, T.L., Mayes, T.L., Trivedi, M.H., 2017b. Interleukin 17 selectively predicts better outcomes with bupropion-SSRI combination: Novel T cell biomarker for antidepressant medication selection. Brain, Behavior, and Immunity 66, 103-110.

of

Jha, M.K., Minhajuddin, A., Gadad, B.S., Greer, T.L., Mayes, T.L., Trivedi, M.H., 2017c. Interleukin 17

ro

selectively predicts better outcomes with bupropion-SSRI combination: Novel T cell biomarker for antidepressant medication selection. Brain, behavior, and immunity 66, 103-110.

-p

Kennard, B., Mayes, T., King, J., Moorehead, A., Wolfe, K., Hughes, J., Castillo, B., Smith, M., Matney, J., Oscarson, B., Stewart, S., Nakonezny, P., Foxwell, A., Emslie, G., 2019. The Development and Feasibility

re

Outcomes of a Youth Suicide Prevention Intensive Outpatient Program. The Journal of adolescent health

lP

: official publication of the Society for Adolescent Medicine 64, 362-369. Kim, Y.K., Lee, S.W., Kim, S.H., Shim, S.H., Han, S.W., Choi, S.H., Lee, B.H., 2008. Differences in cytokines between non-suicidal patients and suicidal patients in major depression. Progress in neuro-

ur na

psychopharmacology & biological psychiatry 32, 356-361. Li, Q.Z., Zhou, J., Wandstrat, A.E., Carr-Johnson, F., Branch, V., Karp, D.R., Mohan, C., Wakeland, E.K., Olsen, N.J., 2007. Protein array autoantibody profiles for insights into systemic lupus erythematosus and

Jo

incomplete lupus syndromes. Clinical and experimental immunology 147, 60-70. Lindqvist, D., Janelidze, S., Erhardt, S., Traskman-Bendz, L., Engstrom, G., Brundin, L., 2011. CSF biomarkers in suicide attempters--a principal component analysis. Acta psychiatrica Scandinavica 124, 52-61. Lindqvist, D., Janelidze, S., Hagell, P., Erhardt, S., Samuelsson, M., Minthon, L., Hansson, O., Bjorkqvist, M., Traskman-Bendz, L., Brundin, L., 2009. Interleukin-6 is elevated in the cerebrospinal fluid of suicide attempters and related to symptom severity. Biol Psychiatry 66, 287-292.

19

Maggi, E., Parronchi, P., Manetti, R., Simonelli, C., Piccinni, M.P., Rugiu, F.S., De Carli, M., Ricci, M., Romagnani, S., 1992. Reciprocal regulatory effects of IFN-gamma and IL-4 on the in vitro development of human Th1 and Th2 clones. The Journal of Immunology 148, 2142-2147. Mendlovic, S., Mozes, E., Eilat, E., Doron, A., Lereya, J., Zakuth, V., Spirer, Z., 1999. Immune activation in non-treated suicidal major depression. Immunology letters 67, 105-108.

to different functional properties. Annual review of immunology 7, 145-173.

of

Mosmann, T.R., Coffman, R.L., 1989. TH1 and TH2 cells: different patterns of lymphokine secretion lead

suicide attempters. Acta psychiatrica Scandinavica 88, 48-52.

ro

Nassberger, L., Traskman-Bendz, L., 1993. Increased soluble interleukin-2 receptor concentrations in

-p

NIMH, 2019. National Institute of Mental Health. Suicide. Accessed on 03MAY2019 at https://www.nimh.nih.gov/health/statistics/suicide.shtml.

re

O'Donovan, A., Rush, G., Hoatam, G., Hughes, B.M., McCrohan, A., Kelleher, C., O'Farrelly, C., Malone,

lP

K.M., 2013. Suicidal ideation is associated with elevated inflammation in patients with major depressive disorder. Depression and anxiety 30, 307-314.

Pandey, G.N., Rizavi, H.S., Bhaumik, R., Ren, X., 2019. Innate immunity in the postmortem brain of

ur na

depressed and suicide subjects: Role of Toll-like receptors. Brain, behavior, and immunity 75, 101-111. Pandey, G.N., Rizavi, H.S., Ren, X., Bhaumik, R., Dwivedi, Y., 2014. Toll-like receptors in the depressed and suicide brain. Journal of psychiatric research 53, 62-68.

Jo

Pandey, G.N., Rizavi, H.S., Ren, X., Fareed, J., Hoppensteadt, D.A., Roberts, R.C., Conley, R.R., Dwivedi, Y., 2012. Proinflammatory cytokines in the prefrontal cortex of teenage suicide victims. Journal of psychiatric research 46, 57-63.

Pandey, G.N., Rizavi, H.S., Zhang, H., Bhaumik, R., Ren, X., 2018. Abnormal protein and mRNA expression of inflammatory cytokines in the prefrontal cortex of depressed individuals who died by suicide. Journal of psychiatry & neuroscience : JPN 43, 376-385.

20

Pariante, C.M., 2019. Did Spider-Man Work in the NESDA Cohort? In Immunopsychiatry, With Great Power Comes Great Responsibility. Biological Psychiatry 85, 787-788. Quan, J., Lakhanpal, A., Reddy, M.M., Zaman, S., Li, Q.Z., German, D.C., Olsen, N.J., Kodadek, T., Karp, D.R., 2014. Discovery of biomarkers for systemic lupus erythematosus using a library of synthetic autoantigen surrogates. Journal of immunological methods 402, 23-34. Rush, A.J., Trivedi, M.H., Stewart, J.W., Nierenberg, A.A., Fava, M., Kurian, B.T., Warden, D., Morris, D.W.,

of

Luther, J.F., Husain, M.M., Cook, I.A., Shelton, R.C., Lesser, I.M., Kornstein, S.G., Wisniewski, S.R., 2011.

ro

Combining medications to enhance depression outcomes (CO-MED): acute and long-term outcomes of a single-blind randomized study. The American journal of psychiatry 168, 689-701.

-p

Shelton, R.C., Falola, M., Li, L., Zajecka, J., Fava, M., Papakostas, G.I., 2015. The pro-inflammatory profile of depressed patients is (partly) related to obesity. Journal of psychiatric research 70, 91-97.

re

Shirley, M., 2017. Dupilumab: first global approval. Drugs 77, 1115-1121.

lP

Siegmann, E.M., Muller, H.H.O., Luecke, C., Philipsen, A., Kornhuber, J., Gromer, T.W., 2018. Association of Depression and Anxiety Disorders With Autoimmune Thyroiditis: A Systematic Review and Metaanalysis. JAMA psychiatry 75, 577-584.

ur na

Simpson, E.L., Bieber, T., Guttman-Yassky, E., Beck, L.A., Blauvelt, A., Cork, M.J., Silverberg, J.I., Deleuran, M., Kataoka, Y., Lacour, J.-P., Kingo, K., Worm, M., Poulin, Y., Wollenberg, A., Soo, Y., Graham, N.M.H., Pirozzi, G., Akinlade, B., Staudinger, H., Mastey, V., Eckert, L., Gadkari, A., Stahl, N., Yancopoulos, G.D.,

Jo

Ardeleanu, M., 2016. Two Phase 3 Trials of Dupilumab versus Placebo in Atopic Dermatitis. New England Journal of Medicine 375, 2335-2348. Slattery, M.J., Essex, M.J., Paletz, E.M., Vanness, E.R., Infante, M., Rogers, G.M., Gern, J.E., 2011. Depression, anxiety, and dermatologic quality of life in adolescents with atopic dermatitis. Journal of Allergy and Clinical Immunology 128, 668-671.e663.

21

Syed, S.A., Beurel, E., Loewenstein, D.A., Lowell, J.A., Craighead, W.E., Dunlop, B.W., Mayberg, H.S., Dhabhar, F., Dietrich, W.D., Keane, R.W., de Rivero Vaccari, J.P., Nemeroff, C.B., 2018. Defective Inflammatory Pathways in Never-Treated Depressed Patients Are Associated with Poor Treatment Response. Neuron 99, 914-924.e913. Tonelli, L.H., Stiller, J., Rujescu, D., Giegling, I., Schneider, B., Maurer, K., Schnabel, A., Moller, H.J., Chen, H.H., Postolache, T.T., 2008. Elevated cytokine expression in the orbitofrontal cortex of victims of suicide.

of

Acta psychiatrica Scandinavica 117, 198-206.

ro

Trivedi, M.H., Chin Fatt, C., Jha, M.K., Cooper, C.M., Trombello, J.M., Mason, B.L., Hughes, J., Gadad, B., Czysz, A.H., Toll, R.T., Fuller, A.K., Sethuram, S., Mayes, T.L., Minhajuddin, A., Carmody, T., Greer, T.L.,

-p

2019. Comprehensive phenotyping of depression disease trajectory and risk: Rationale and design of DFW 5000. Journal of Psychiatric Research In press.

re

Van Dyken, S.J., Locksley, R.M., 2013. Interleukin-4- and interleukin-13-mediated alternatively activated

lP

macrophages: roles in homeostasis and disease. Annu Rev Immunol 31, 317-343. Wachholz, S., Esslinger, M., Plumper, J., Manitz, M.P., Juckel, G., Friebe, A., 2016. Microglia activation is associated with IFN-alpha induced depressive-like behavior. Brain, behavior, and immunity 55, 105-113.

ur na

Wachholz, S., Knorr, A., Mengert, L., Plumper, J., Sommer, R., Juckel, G., Friebe, A., 2017. Interleukin-4 is a participant in the regulation of depressive-like behavior. Behavioural brain research 326, 165-172. Yu, S.H., Silverberg, J.I., 2015. Association between Atopic Dermatitis and Depression in US Adults. Journal

Jo

of Investigative Dermatology 135, 3183-3186.

22

Figure 1. Difference in levels of interleukin 4 (IL-4) in the following groups of adolescents and young adults:

Jo

ur na

lP

re

-p

ro

of

healthy control, at-risk for depression, and with recent suicide behavior.

23

Figure 2. Association between levels of autoantibodies and levels of Interleukin 4 (IL-4) in CO-MED trial Legends: This figure presents the autoantibodies (Immunoglobulin G) that were significantly associated with IL-4 after controlling for age, body mass index, sex, race, ethnicity, and c-reactive protein (CRP) in the

Jo

ur na

lP

re

-p

ro

of

Combining Medications to Enhance Depression Outcomes (CO-MED) trial.

24

Table 1: Baseline sociodemographic and clinical factors of adolescents and young adults with recent suicide behavior/severe ideation and those at-risk of depression and healthy controls.

Categorical variables

Suicide behavior

At-risk

N

N

%

37

Healthy controls %

33

N

%

8.15 (2)

0.017

of

8

21.6

15

45.5

20

51.3

Female

29

78.4

18

54.5

18

46.2

Race 94.6

25

75.8

Black

2

5.4

5

15.2

Other

0

0.0

3

17

lP

9.1

25.48(4)

<0.001

6.36 (2)

0.04

43.6

-p

35

11

re

White

ro

Male

11

28.2 28.2

Hispanic

5

13.5

4

12.1

13

33.3%

Non-Hispanic

31

83.8

29

87.9

26

66.7%

ur na

p value

39

Gender*

Ethnicity

χ2 (df)

Continuous variables

Mean

SD

Mean

SD

Mean

SD

F statistics

p value

Age

14.8

1.3

18.5

4.1

18.3

4.2

12.70

<0.001

24.6

7.2

24.6

7

24.5

5.1

0.00

>0.99

-0.8

0.8

-1.2

0.8

-0.9

0.7

1.81

0.17

Body mass index

Jo

Log CRP

* missing n=1 for at-risk group, CRP is C-reactive protein

25

Table 2. Comparison of cytokines and chemokines among three groups of adolescents and young adults. Assay

limits Suicide

(pg/mL) LD Chemokine (C-C) ligand 21 21.9

behavior At-risk (n=33)

(n=37)

Healthy controls (n=39)

UD

Median

IQR

Median

IQR

Median

IQR

3,923

3383.2

549.3

3532.5

527.7

3532.5

712

1,200

17.5

13.8

21.8

5,000

540.4

255.5

120,000

611.3

Chemokine (C-X-C) ligand 13 0.7

Chemokine (C-C) ligand 27 1.2

(CXCL5) 1.5

3,859

Eotaxin-2 (CCL24)

6.2

Eotaxin-3 (CCL26)

0.9

183.7

491.9

231.8

681.6

360.2

695.1

441

520.6

9.4

49.2

24.7

50.8

17

4,073

216.8

161.2

203.4

189.7

243.1

191.6

12,109

6.2

2.3

4.9

4.1

5.9

2.8

ur na

Fractalkine (CX3CL1)

8

48.2

lP

Eotaxin-1 (CCL11)

206.3

re

Chemokine (C-X-C) ligand 5 7.3

19.2

-p

(CCL27)

14.9

ro

(CXCL13)

of

(CCL21)

4.0

11,463

150.5

43.3

153

51.8

177.6

102.8

chemotactic 0.8

11,135

34.1

23.2

41.8

36.2

42.2

34.2

macrophage 5.3

35,000

43.2

37.2

50.8

51

68.8

61.7

7,024

167.2

55.8

179.3

76.7

201

90.8

Granulocyte

protein 2 (CXCL6)

Jo

Granulocyte colony

stimulating

factor

(GM-CSF)

Chemokine (C-X-C) ligand 1 3.1 (CXCL1)

26

Chemokine (C-X-C) ligand 2 4.6

13,257

292.3

232.5

321.3

381.7

313.3

575.3

1,015

59.4

13.2

60.4

18.2

61.2

14.6

(CXCL2) Chemokine (C-C) ligand 1 1.8 (CCL1) 2.3

20,236

48.7

22.4

63.2

48.3

67.8

37.9

Interleukin 1 β (IL-1β)

0.4

7,000

3.3

2.1

3.2

2.3

3.7

2.5

Interleukin 2 (IL-2)

0.8

13,000

13.4

5.5

14.8

10.4

16.7

8.8

Interleukin 4 (IL-4)

1.2

4,804

16.8

7.9

24.4

16.3

29.1

16.1

Interleukin 6 (IL-6)

0.7

12,000

7.6

3.8

7,640

5.8

3.4

(IL-8,

also 0.5

re

CXCL8)

ro

8

9.6

8.7

11.4

7.5

7.6

5.1

8

5.3

-p

Interleukin

of

Interferon γ (IFN-γ)

1.3

18,708

18.1

8.8

23.2

18.2

23.5

17.9

Interleukin 16 (IL-16)

2.1

34,000

812.9

278.2

980.5

799.6

1005.3

589.3

80.7

36

99.8

69.8

93.7

42.1

2,298

12.9

7.2

12.8

11.3

13.9

13.1

4,812

26.1

7.2

25.3

8.8

26

9.4

lP

Interleukin 10 (IL-10)

IFN-γ induced protein 10 (IP- 1.6

ur na

10)

Chemokine C-X-C) ligand 11 0.1 (CXCL11) Monocyte

7,714

chemoattractant 0.3

Jo

protein (MCP) 1 MCP-2

0.3

4,056

21.4

9.2

21.2

9.6

24.3

23

MCP-3

1.9

20,133

68.9

29.4

81.4

58.6

90.1

48.3

MCP-4

0.2

3,368

26.4

18.2

31.2

26.2

39.6

31.8

27

Chemokine (C-C motif) ligand 0.9

14,649

800.3

325.7

853.7

448.2

825.5

262.8

377,724

1148.8

672.8

1023.9

1185.2

1258.7

2133.6

19,600

173.3

85.2

166.5

143.4

228.9

133.5

1,543

5.9

2.1

6.8

3.3

6.7

2.4

9,100

4885.4

2158.1

4455

4,675

11.8

6.2

11.3

48,494

125.3

63.7

235.8

22 (CCL22) Macrophage inhibitory factor Chemokine

(C-X-C

23.1

motif) 1.8

ligand 9 (CXCL9) Chemokine (C-C motif) ligand 0.4

5183.8

3102.6

ro

Chemokine (C-C motif) ligand 1.7

5.1

12.1

6.8

125.6

144.7

134.3

106.3

122.7

207.3

209.4

171.5

160.9

Chemokine (C-C motif) ligand 3.0

re

20 (CCL20)

lP

19 (CCL19) Chemokine (C-C motif) ligand 1.0

14,450

ur na

23 (CCL23) Chemokine

-p

15 (CCL15) Chemokine (C-C motif) ligand 0.3

(C-X-C

of

3 (CCL3)

977.7

motif) 0.5

2,867

464.8

130.3

424.2

123.4

410.1

149.1

motif) 8.3

115,730

2365.1

579.5

2348.4

1187.8

2354.7

841.4

430

47.4

21

54

56.5

64.3

49.7

114,493

278

86.7

205.8

143.2

253.7

134.1

ligand 16 (CXCL16) Chemokine

(C-X-C

Jo

ligand 12 (CXCL12)

Chemokine (C-C motif) ligand 1.7 17 (CCL17)

Chemokine (C-C motif) ligand 20.6 25 (CCL25)

28

Tumor necrosis factor (TNF) α 0.9

13,879

24.2

9.6

28.9

15.9

28.2

14.5

pg is pictogram, mL is milliliter, LD is lower limit of detection, UD is upper limit of detection, IQR is

Jo

ur na

lP

re

-p

ro

of

interquartile range.

29

Table 3. Results of regression analyses comparing levels of cytokines and chemokines after controlling for C-reactive protein, age, gender, body mass index, race, and ethnicity Overall ANCOVA

Healthy control vs. at- Healthy control vs. risk groups

p*

t stats

Chemokine (C-C) ligand 21 (CCL 21)

2.1

0.86

0.97

Chemokine (C-X-C) ligand 13 (CXCL 13)

0.6

0.86

0.17

Chemokine (C-C) ligand 27 (CCL 27)

0.15

0.95

Chemokine (C-X-C) ligand 5 (CXCL5)

0.65

0.86

Eotaxin-1

0.55

0.86

Eotaxin-2

0.3

Eotaxin-3

p*

t stats

of

F value

suicide behavior p*

-1.28

0.95

0.96

-0.92

0.95

0.93

0.68

0.95

-0.7

0.93

-0.18

0.95

-0.51

0.93

-0.33

0.95

0.95

-0.38

0.93

-0.12

0.95

4.51

0.27

-1.26

0.93

1.77

0.95

0.86

0.86

-0.6

0.93

-0.73

0.95

Granulocyte chemotactic protein 2 (CXCL6) 0.95

0.86

-0.5

0.93

-0.88

0.95

Granulocyte

colony 0.18

0.95

-0.45

0.93

-0.23

0.95

0.81

0.86

-0.8

0.93

0.17

0.95

Chemokine (C-X-C) ligand 2 (CXCL2)

0.13

0.95

0.38

0.93

0.19

0.95

Chemokine (C-C) ligand 1 (CCL1)

0.57

0.86

-0.78

0.93

0.37

0.95

Interferon γ (IFN-γ)

1.25

0.86

-0.91

0.93

-1.06

0.95

Interleukin 1 β (IL-1β)

0.88

0.86

-1.14

0.93

-0.14

0.95

macrophage

0.37

-p

re

lP

ur na

Fractalkine (CX3CL1)

ro

0.93

stimulating factor (GM-CSF)

Jo

Chemokine (C-X-C) ligand 1 (CXCL1)

30

Overall ANCOVA

Healthy control vs. at- Healthy control vs. risk groups

suicide behavior

0.51

0.86

-0.65

0.93

-0.5

0.95

Interleukin 4 (IL-4)

12.49

0.0007

-1.19

0.93

-3.45

0.035

Interleukin 6 (IL-6)

0.84

0.86

-0.85

0.93

-0.79

0.95

Interleukin 8 (IL-8, also CXCL8)

0.59

0.86

0.11

0.96

-0.39

0.95

Interleukin 10 (IL-10)

0.62

0.86

-0.25

0.93

-0.88

0.95

Interleukin 6 (IL-16)

1.46

0.86

0.24

0.93

-0.99

0.95

IFN-γ induced protein 10 (IP-10)

0.06

0.97

Chemokine C-X-C) ligand 11 (CXCL11)

1.04

0.86

Monocyte chemoattractant protein (MCP) 0.02

0.98

ro

of

Interleukin 2 (IL-2)

0.93

0.29

0.95

-0.33

0.93

-0.04

0.96

1.11

0.93

1.07

0.95

0.86

-0.25

0.93

-0.27

0.95

1.15

0.86

-0.39

0.93

-0.91

0.95

1.72

0.86

-0.07

0.96

-0.47

0.95

Chemokine (C-C motif) ligand 22 (CCL22)

0.16

0.95

0.8

0.93

-0.61

0.95

Macrophage inhibitory factor

0.39

0.91

-0.05

0.96

-0.37

0.95

Chemokine (C-X-C motif) ligand 9 (CXCL9)

0.6

0.86

-0.61

0.93

-0.8

0.95

Chemokine (C-C motif) ligand 3 (CCL3)

0.42

0.91

-0.45

0.93

-0.25

0.95

Chemokine (C-C motif) ligand 15 (CCL15)

1.45

0.86

-1.25

0.93

0.06

0.96

Chemokine (C-C motif) ligand 20 (CCL20)

2.38

0.86

-2

0.93

-1.32

0.95

Chemokine (C-C motif) ligand 19 (CCL19)

0.25

0.95

-0.24

0.93

-0.79

0.95

Chemokine (C-C motif) ligand 23 (CCL23)

2.02

0.86

1.43

0.93

-0.38

0.95

re

-p

1.19

1

0.99

lP

MCP-2 MCP-3

Jo

ur na

MCP-4)

31

Overall ANCOVA

Healthy control vs. at- Healthy control vs. risk groups

Chemokine

suicide behavior

(C-X-C

motif)

ligand

16 0.05

0.97

0.49

0.93

0.29

0.95

(C-X-C

motif)

ligand

22 0.14

0.95

0.74

0.93

1.11

0.95

0.93

-1.04

0.95

0.93

1.45

0.95

-0.4

0.95

(CXCL16) Chemokine

0.64

0.86

-0.7

Chemokine (C-C motif) ligand 25 (CCL25)

4.09

0.27

-1.06

Tumor necrosis factor (TNF) α

0.52

0.86

ro

Chemokine (C-C motif) ligand 17 (CCL17)

of

(CXCL22)

-0.06

0.96

-p

Bolded are unadjusted p <0.05 for overall test in Analysis of Covariance (ANCOVA). Abbreviations, p* is

Jo

ur na

lP

re

FDR adjusted p value

32

33

of

ro

-p

re

lP

ur na

Jo