Targeted Sequencing of FKBP5 in Suicide Attempters with Bipolar Disorder

Abstract

FKBP5 is a critical component of the Hypothalamic-Pituitary-Adrenal (HPA) axis, a system which regulates our response to stress. It forms part of a complex of chaperones, which inhibits binding of cortisol and glucocorticoid receptor translocation to the nucleus. Variations in both the HPA axis and FKBP5 have been associated with suicidal behavior. We developed a systematic, targeted sequencing approach to investigate coding and regulatory regions in or near FKBP5 in 476 bipolar disorder suicide attempters and 473 bipolar disorder non-attempters. Following stringent quality control checks, we performed single-variant, gene-level and haplotype tests on the resulting 481 variants. Secondary analyses investigated whether sex-specific variations in FKBP5 increased the risk of attempted suicide. One variant, rs141713011, showed an excess of minor alleles in suicide attempters that was statistically significant following correction for multiple testing (Odds Ratio = 6.65, P-value = 7.5 x 10⁻⁴, Permuted P-value = 0.038). However, this result could not be replicated in an independent cohort (Odds Ratio = 0.90, P-value = 0.78). Three female-specific and four male-specific variants of nominal significance were also identified (P-value < 0.05). The gene-level and haplotype association tests did not produce any significant results. This comprehensive study of common and rare variants in FKBP5 focused on both regulatory and coding regions in relation to attempted suicide. One rare variant remained significant following correction for multiple testing but could not be replicated. Further investigation is required in larger sample sets to fully elucidate the association of this variant with suicidal behavior.

About the Authors:

Marie E. Breen

Affiliation: Department of Psychiatry, University of Iowa Carver College of Medicine, Iowa City, Iowa, United States of America

Sophia C. Gaynor

Affiliation: Department of Psychiatry, University of Iowa Carver College of Medicine, Iowa City, Iowa, United States of America

Eric T. Monson

Affiliation: Department of Psychiatry, University of Iowa Carver College of Medicine, Iowa City, Iowa, United States of America

Kelly de Klerk

Affiliation: Department of Psychiatry, University of Iowa Carver College of Medicine, Iowa City, Iowa, United States of America

Meredith G. Parsons

Affiliation: Department of Psychiatry, University of Iowa Carver College of Medicine, Iowa City, Iowa, United States of America

Terry A. Braun

Affiliations Department of Ophthalmology and Visual Sciences, University of Iowa Carver College of Medicine, Iowa City, Iowa, United States of America, Department of Biomedical Engineering, University of Iowa College of Engineering, Iowa City, Iowa, United States of America

Adam P. DeLuca

Affiliation: Department of Ophthalmology and Visual Sciences, University of Iowa Carver College of Medicine, Iowa City, Iowa, United States of America

Peter P. Zandi

Affiliation: Department of Mental Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United States of America

James B. Potash

Affiliation: Department of Psychiatry, University of Iowa Carver College of Medicine, Iowa City, Iowa, United States of America

Virginia L. Willour

* E-mail: [email protected]

Affiliation: Department of Psychiatry, University of Iowa Carver College of Medicine, Iowa City, Iowa, United States of America

Introduction

Every hour throughout the world, approximately ninety-one people die by suicide and reports suggest suicide attempts are over twenty times higher [1]. The complexity of attempted and completed suicide, collectively termed suicidal behavior, is due to the combination of biological, behavioral and environmental factors that can increase the risk of this phenotype [2]. One of the most common risk factors is the presence of a psychiatric disorder, predominantly mood disorders such as bipolar disorder, with almost 90% of suicide attempters suffering from a preceding psychiatric condition [3, 4].

One current hypothesis in the field of suicide genetics is that suicidal behavior is associated with an altered stress response. The Hypothalamic-Pituitary-Adrenal (HPA) axis is a tightly controlled collection of interacting proteins that help regulate our response to stress. Dysregulation of the HPA axis causes an abnormal stress response and has been linked to suicidal behavior [5–7]. One critical component of the HPA axis is the co-chaperone FK506 Binding Protein 5 (FKBP5; ENSG00000096060). FKBP5 binds to the glucocorticoid receptor (GR) as part of a complex of chaperones. This complex inhibits binding of cortisol and receptor translocation to the nucleus [8]. Due to its crucial role in the HPA axis, specific focus has been placed on the FKBP5 gene and variations within it that might be associated with suicidal behavior. A family-based study found an association with FKBP5 and bipolar disorder, and a secondary analysis linked this to a history of attempted suicide [9]. European [10, 11] and Japanese [12] sample cohorts with suicidal behavior (ranging from suicidal ideation to attempt and completed suicide) have also been found to show associations with FKBP5. Adding more complexity, it has been suggested that FKBP5 variants may interact with childhood trauma to influence suicidal behavior [13, 14]. The presence of regulatory variation within FKBP5 may be functionally relevant as gene and protein expression studies have found altered FKBP5 levels in suicide completers compared to controls [15]. In contrast, genome-wide studies have not shown any association between FKBP5 and suicidal behavior [16–27].

The conflicting nature of these results suggests that a more focused interrogation of this gene including less commonly investigated regions and less common variants, may identify functionally relevant variation. Therefore, in this study, we have sequenced both coding regions and regulatory regions to identify genetic variation in FKBP5 that may increase the risk for suicidal behavior in individuals with bipolar disorder. Following stringent quality control checks, we performed single-variant, gene-level and haplotype tests on the resulting 481 variants. One rare variant remained significant following correction for multiple testing but could not be replicated. Further investigation is required in larger sample sets to fully elucidate the association of this variant with suicidal behavior.

Materials and Methods

Ethics Statement

Samples were obtained from the National Institute of Mental Health (NIMH) Genetics Initiative, (https://www.nimhgenetics.org/; S1 File) [28] and distributed through Rutgers University’s RUCDR Infinite Biologics (http://rucdr.org/). The University of Iowa Institutional Review Board considers this project to be “not human subject research” because these repository samples were sent to the University of Iowa’s laboratory in a de-identified manner.

Subjects

For this study, we utilized the same sample set as in our attempted suicide genome-wide association study (GWAS) [16]. All subjects included in the analysis were European-American and unrelated. Phenotype data was collected using the Diagnostic Interview for Genetic Studies (DIGS; versions 1.0–4.0 [29]) and diagnoses ascertained as bipolar disorder, type I (BPI) or Schizoaffective Disorder, Bipolar Type (SABP). There was no significant difference in average age between suicide attempters and non-attempters (S1 Table).

Definition of Suicide Subjects

Case subjects were defined as suicide attempters if they answered yes to the DIGS question: “Have you ever tried to kill yourself?” and had definite or serious intent to complete suicide. This resulted in a total of 476 suicide attempters and 476 non-attempters for sequencing.

Population Stratification

Principal Component Analysis (PCA) was performed using EIGENSTRAT in EIGENSOFT v3.0 [30] on our entire attempted suicide GWAS [16], and these results were included as covariates in the logistic regression analyses.

Target Determination

The UCSC Genome browser databases (UCSC Genes [31], GenCODE Genes [32], Ensembl Genes [33], RefSeq Genes [34] and ENCODE databases [35]) were used to define coding and regulatory regions in or near FKBP5 transcripts. Targets included: (i) all exons (±50 bases) from all available transcripts (generated from the previously discussed UCSC Genome browser databases), (ii) predicted DNase hypersensitivity sites with at least one overlapping transcription factor binding site within all FKBP5 transcripts (±10 kb) using ENCODE and (iii) promoter regions (determined as 2 kb upstream of all the transcript start sites). Coding regions encompass any translated exons while regulatory regions include all other loci covered.

SureSelect Technology

We used a customized SureSelect Target Enrichment System (Agilent Technologies, Santa Clara, CA, U.S.A.) to capture and sequence our chosen targets (as previously described in [36]). Briefly, 3 μg genomic DNA was sheared, prepared into a library using adaptors and hybridized to biotinylated RNA baits. Pre-designed DNA targets were collected using magnetic streptavidin beads and sequenced as 100 bp paired-end reads using the HiSeq2000 at the Iowa Institute of Human Genetics Genomics Division in pools of sixteen samples.

Sample Processing Pipeline

All samples were processed as a group. Samples were aligned to the Human Feb. 2009 Assembly (GRCh37/hg19) using the Burrows-Wheeler Aligner (BWA) tool v0.6.2 [37] and a 30 base seed length. Using SamTools v0.1.18 [38], aligned files were converted to binary (BAM) files and then sorted and indexed. Unpaired, improperly paired and unmapped reads were excluded using BamTools v2.2.3 [39] and duplicate reads removed using Picard v1.88 (http://picard.sourceforge.net). Further BamTool filtering included a mapping quality score > 20. Genome analysis toolkit (GATK) v3.1.1 [40] best practices were used to perform such tasks as realignment around insertion/deletions (indels), base score recalibration, variant calling using Haplotypecaller, and initial quality checks (such as strand bias and depth ratio). Any single nucleotide polymorphism (SNP) or indel which failed these checks was excluded. Variant calls were labeled as missing if they had a call depth < 10, or a genotyping quality Phred score < 20. Variants were excluded from the final dataset if they had missing calls in > 10% of subjects or a Hardy-Weinberg P-value < 1 x 10⁻⁶. Following a comparison to our previous GWAS genotypes, three non-attempters were excluded before analysis due to mismatches, resulting in a final total of 473 non-attempters for analysis. All variants were annotated using Ensembl v75 data-tracks via ANNOVAR (vMarch 22, 2015) [41].

Statistical Analysis

On a single-variant level, we tested 481 FKBP5 variants for an association with attempted suicide using a logistic regression test with Firth’s penalized maximum likelihood method via the logistf R package v1.21 [42].

On the gene level, we assessed the association of variation across the FKBP5 locus with suicidal behavior, with two minor allele frequency (MAF) thresholds (≤ 1% and ≤ 5%). We determined these thresholds from our own dataset as well as frequency data from the 1000 Genomes Project European samples vOct2014 [43], Exome Aggregation Consortium non-Finnish European samples (ExAC v0.3 http://exac.broadinstitute.org) and the NHLBI GO Exome Sequencing Project European samples (http://evs.gs.washington.edu/EVS/). We used two complimentary methods to test for associations with variation in the FKBP5 locus. First, we performed a SNP-set (Sequence) Kernel Association Test (SKAT) v1.0.9 [44], which employs a multivariate regression model that accommodates both risk and protective alleles in the tests of association. Second, we implemented a gene-burden test utilizing the previously described Combined Multivariate and Collapsing (CMC) method [45] to collapse all functional FKBP5 variants by subject, and Firth’s penalized logistic regression via the logistf R package v1.21 [42] to test for association. Odds ratios were calculated from the covariate-corrected logistic regression output. The gene-level analyses examined either coding or regulatory variants. Broad coding variants included any disruptive variants and non-synonymous variants predicted to be damaging by any one of six software tools (SIFT [46], Polymorphism Phenotyping v2 (Polyphen-2) HDIV, Polyphen-2 HVAR [47], LRT [48], VEST [49] or MutationTaster [50]). Disruptive coding variants included any that are considered by ANNOVAR to be frameshift, stop-gain/nonsense or essential splice site mutations. Regulatory variants in the gene-level analyses were defined by their predicted ability to alter a regulatory region as determined by RegulomeDB v1.1 [51]. Specifically, broad regulatory variants had a RegulomeDB score of 1–6 and narrow regulatory variants had a more stringent RegulomeDB score of 1 or 2.

Sex and the first three principal components from the analysis of the GWAS data were included as covariates in all primary analyses. Statistical significance was calculated using permutation testing. We performed 10,000 permutations over the suicide attempter status using custom scripts.

Genotyping

Replication of results were performed on bipolar disorder suicide attempters with definite/serious intent (N = 328) and bipolar disorder non-attempters (N = 655; S2 Table), which were obtained from the NIMH Genetics Initiative, (https://www.nimhgenetics.org/) [28] and did not overlap with the initial sample set.

Due to repetitive and homopolymer regions surrounding the top result, rs141713011, we used an RNase H-dependent (rhPCR) method (Integrated DNA Technologies, Coralville, IA, U.S.A.) to genotype this variant in the replication cohort. Briefly, we started with 20 ng genomic DNA and added SYBR GreenER qPCR SuperMix Universal (including ROX dye; Thermo Fisher Scientific, Waltham, MA, U.S.A.), specifically designed GEN2 primers for either the A allele or the C allele, and RNase H2 enzyme (Integrated DNA Technologies, Coralville, IA, U.S.A.). Real-time PCR was then performed on the ViiA7 system, and crossing threshold (Ct) values were used to identify sample genotypes.

The most commonly investigated FKBP5 variant, rs3800373, was genotyped in the replication cohort using a Taqman SNP genotyping assay (C_27489960_10; Thermo Fisher Scientific, Waltham, MA, U.S.A.).

Haplotype Analysis

To identify common haplotypes we utilized Haploview v4.2 [52]. Default values were employed except for the minimum MAF which was set at 0.05. Blocks were defined using “confidence intervals” [53].

Results

Dataset Quality

Following sequencing, 96.65% of the targeted sites in the FKBP5 loci had at least ≥ 10X coverage and 100% of targets had at least ≥ 1X coverage (S1 Table). A total of 40,922 bases were covered by the designed probes and 565 variants were detected from this. The GATK quality control filters excluded 24 variants and our stringent quality control checks excluded a further 60 variants resulting in a final high quality dataset of 481 variants. This final dataset included 13 coding variants, 217 novel variants (not found in dbSNP 142 [54] or other sequencing databases) and 35 common variants (MAF > 0.05).

Single-variant Analysis

Our primary analysis focused on 481 variants within or near FKBP5 (S3 Table), and five of them were differentially represented among suicide attempters and non-attempters at a nominally significant level (Table 1). One variant, rs141713011, showed an excess of minor alleles in suicide attempters that was statistically significant following correction for multiple testing (Odds Ratio = 6.65, P-value = 7.5 x 10⁻⁴, Permuted P-value = 0.038). This variant resides in the intron or 3’UTR of multiple transcripts (Fig 1). The minor allele was present in sixteen suicide attempters (MAF = 0.017) and two non-attempters (MAF = 0.002).

[Figure omitted. See PDF.]

Table 1. Single-variant results with a P-value < 0.05.

https://doi.org/10.1371/journal.pone.0169158.t001

[Figure omitted. See PDF.]

Fig 1. Schematic of the top variant in the FKBP5 gene alongside previous findings.

The top single variant result is displayed below the gene; variants or haplotypes with known associations to suicidal behavior are displayed above the gene. Gray variants were not covered by this study. Numbers represent references: 1. [10] 2. [14] 3. [55] 4. [12] 5. [13] 6. [11] 7. [9] 8. [56] 9. [57]. All FKBP5 transcripts have been collated to show exons (black vertical lines), introns (black horizontal lines) and untranslated regions (black boxes). A transcription factor binding site (TFBS; green box) and DNase hypersensitivity site (purple box) upstream of the top variant are labeled and are not to scale. GRE denotes glucocorticoid receptor response elements (blue boxes) and are not to scale.

https://doi.org/10.1371/journal.pone.0169158.g001

Replication of Results

To further examine the top finding, rs141713011, we attempted to replicate the significant association with suicidal behavior in an additional cohort of 328 bipolar disorder suicide attempters and 655 bipolar disorder non-attempters. In this sample we observed five suicide attempters with the minor allele (MAF = 0.015) and ten non-attempters who carried it (MAF = 0.015). Thus, we failed to replicate our initial finding (Odds Ratio = 0.90, P-value = 0.78).

Replication of Prior Findings

The two most implicated FKBP5 variants in suicide are rs1360780 and rs3800373 (Fig 1). This study did not assess rs1360780, but it was covered in our previous attempted suicide GWAS and was not significantly associated with the phenotype there [16, 58].

In comparison, rs3800373 was assessed in our initial sample, and was not found to significantly differ between suicide attempters and non-attempters (MAF in suicide attempters = 0.28, MAF in non-attempters = 0.29, Odds Ratio = 0.94, P-value = 0.52, Permuted P-value = 1.00). Similarly, it was genotyped in the additional cohort, and again the results were not significant (MAF in suicide attempters = 0.30, MAF in non-attempters = 0.28, Odds Ratio = 1.07, P-value = 0.56).

Sex-specific Analyses

To investigate if there was a sex-specific association between FKBP5 and attempted suicide, we performed female-specific and male-specific single-variant analyses (S4 and S5 Tables). Three female-specific and four male-specific variants differed between suicide attempters and non-attempters at a nominally significant level (S6 Table). The previously discussed main finding (rs141713011) was also the top female-specific finding (Odds Ratio = 6.79, P-value = 8.1 x 10⁻³, Permuted P-value = 0.28). The top male-specific variant, rs13192954, showed an increase in significance compared to the main findings (Odds Ratio = 0.52, P-value = 0.026, Permuted P-value = 0.68). No sex-specific findings were statistically significant following correction for multiple testing.

Gene-level Analysis

Due to the large amount of sequencing data produced from this study we were able to perform gene-level analyses to investigate whether FKBP5 variants collectively influence suicidal behavior. To do this, we created four groups from the 481 variants studied based on their location and their ability to be damaging: (i) Regulatory Broad, (ii) Regulatory Narrow, (iii) Coding Broad and (iv) Coding Disruptive. The inclusion of two different MAF thresholds allowed us to investigate very rare and less rare separately.

We performed two distinct gene-level tests: SKAT and gene-burden. Both tests examine all FKBP5 variants for each individual based on applied thresholds. The gene-burden test employs a logistic regression model to test the association between an indicator variable for whether individuals possess one or more rare, functional alleles of a specific category or not and suicide attempt. In comparison, SKAT employs a multivariate model that includes each variant as a covariate to look for significantly different distributions of the variants between suicide attempters and non-attempters, so it has the advantage of not being affected by the direction of effect on risk.

No gene-level tests showed evidence of association, regardless of minor allele threshold (Tables 2 and 3). The top result was from the coding broad analysis using the gene-burden approach (using either MAF threshold; P-value = 0.15). The top sex-specific P-value was identified using SKAT within the female Coding Broad subgroup (using either MAF threshold, P-value = 0.057).

[Figure omitted. See PDF.]

Table 2. Gene-level results using two minor allele thresholds.

https://doi.org/10.1371/journal.pone.0169158.t002

[Figure omitted. See PDF.]

Table 3. Sex-specific gene-level results using two minor allele thresholds.

https://doi.org/10.1371/journal.pone.0169158.t003

Haplotype Analysis

Using Haploview we identified three haplotype blocks in the FKBP5 region encompassing common variants (MAF > 0.05; Fig 2 and Table 4). Block 1 covered an 11 kb region encompassing four SNPs including the extensively studied rs3800373. Block 2 covered a 108 kb region and nine SNPs and Block 3 included thirteen SNPs over a 14 kb region. The top haplotype result was found within Block 3 (P-value = 0.097), but no significant associations between haplotypes and attempted suicide were identified.

[Figure omitted. See PDF.]

Fig 2. Haplotype block structures in the FKBP5 region.

A collated version of all FKBP5 transcripts is represented at the top of the figure. Vertical black lines represent exons, green boxes represent untranslated regions and purple boxes represent intronic regions. The numbered squares display the D’ score, unnumbered squares have a D’ score of 1.0. Three haplotype blocks were generated and are enclosed by black lines. Lines connect the variants to their approximate location within the FKBP5 locus.

https://doi.org/10.1371/journal.pone.0169158.g002

[Figure omitted. See PDF.]

Table 4. Haplotype results generated using Haploview.

https://doi.org/10.1371/journal.pone.0169158.t004

Discussion

Genetic variation can alter FKBP5 function, thereby affecting the co-chaperone complex essential to regulating translocation of the GR to the nucleus [8]. This may contribute to the over-activation of the HPA axis seen in subjects with suicidal behavior [5–7]. The goal of this study was to investigate whether variants within the FKBP5 region, separately or together, influence suicidal behavior in individuals with bipolar disorder.

We employed the novel approach of sequencing both coding and regulatory regions using state-of-the-art next-generation sequencing technology. This allowed us to capture more functional areas of FKBP5 and to focus on rare variation, thereby interrogating the gene more fully than our previous investigations could [16, 58]. Interestingly, the most significantly associated variant in this study was located near an area with experimentally demonstrated regulatory function (Fig 1) [35].

Our previous GWAS [16, 58] tested the hypothesis that common variants contributed to the risk for suicidal behavior, but no significant evidence to this effect was found in relation to FKBP5. This study, in contrast to our GWAS and most other FKBP5 genetic studies, concentrated on both common and rare variation located in coding and regulatory regions of FKBP5 and their contribution to the risk for attempted suicide. However, we were unable to find replicable associations with FKBP5 variants individually or collectively.

As secondary analyses, we considered the role of sex-specific genetic variation in attempted suicide. There are several lines of evidence that suggest that suicidal behavior differs between females and males. Specifically, nearly twice as many males complete suicide compared to females [1] while more females attempt suicide [59, 60]. Furthermore, we have previously identified a female-specific common variant associated with suicidal behavior suggesting that sex-specific variants might influence the risk of this phenotype [16]. No statistically significant sex-specific results were found in this study. The female-specific SKAT coding broad test showed the top gene burden result (P = 0.057), but larger sample sizes are required to determine whether this trend is a true association.

Several candidate variants studies have previously implicated FKBP5 variation in increasing risk for suicidal behavior. In a family study focused on bipolar disorder, four variants displayed increased significance when attempted suicide was included in the analysis as a covariate [9]. Another bipolar disorder study genotyped eight common FKBP5 variants and found that a haplotype block containing seven of the variants was associated with attempted suicide [10]. A depression study focusing on two FKBP5 variants found the TT genotype of rs1360780 and the GG genotype of rs3800373 were significantly associated with suicidal events, which included attempts, plans to attempt, and ideation, in adolescents on antidepressants; this finding remained even after controlling for treatment [11]. A similar study investigating selective serotonin reuptake inhibitor (SSRI) treatment genotyped rs1360780 and found that depressed individuals with the minor T allele were at increased risk of suicidal ideation following SSRI treatment [55]. Supporting evidence for these studies was provided by Menke et al. who investigated treatment-emergent suicidal ideation [27]. A recent study found three FKBP5 variants were associated with suicide attempt, but not with suicide completion. They further identified a haplotype block with increased risk of suicidal behavior comparing depressed suicide attempters and depressed controls. However, these findings were not corrected for multiple testing [57]. A Japanese study identified a statistically significant haplotype block with two FKBP5 variants (rs1360780 and rs3800373) in suicide completers, but did not find any single-variant associations [12]. A second, larger cohort of suicide completers was also genotyped for rs1360780 and rs3800373, and an association was found with rs3800373 [56]. It must be noted that the main limitations of all these studies was the small sample size and limited genotyping data. This was overcome in genome-wide studies which investigated suicidal behavior in larger samples and with broader coverage of the gene [16–27], but no significant associations were reported in FKBP5 across these studies.

One variant of particular interest to the field, rs3800373 [9–14, 56, 57] has repeatedly been associated with suicidal behavior, but did not show association in our study (P-value = 0.52). This may be due to our case-only approach, which differs from the trio or case-control designs used by others. We utilized this approach to ensure that our results were not influenced by an underlying psychiatric disorder, but this did increase the level of suicidal ideation in the non-attempter cohort. Ideation has been previously associated with FKBP5 variation [11, 55], and thus we may have reduced our power to detect an influence of FKBP5. Another possible explanation for the conflicting results may be that we focused on testing areas around and throughout the gene we surmised would be more likely to disrupt transcription or translation, and therefore we were unable to cover all the previously investigated variants (Fig 1). We did not cover intronic variants outside overlapping DNase hypersensitivity sites and transcription factor binding sites that have been associated with this phenotype. In particular, rs1360780, which can alter expression due to its proximity to a GR response element (GRE) [61], was not covered.

This study has other limitations. With an estimated 80% power (calculated using Quanto v1.2.4; http://biostats.usc.edu/Quanto.html), we were able to detect a single variant with an effect size as low as 2.35 assuming a study-wide significance of 1.04 x 10⁻⁴ and MAF of 0.05. For the replication analysis, we were able to detect a single variant with an effect size of 1.88 assuming a MAF of 0.05. With a MAF of 0.008 (seen in the replication results for rs141713011), an effect size of 3.70 can be detected. For the gene-level analyses, we had an estimated 80% power to detect an effect size as low as 1.95 assuming a study-wide significance of 6.3 x 10⁻³ and MAF of 0.05. In addition, it should be noted that other biological and environmental factors that impact suicidal behavior, such as epigenetics and early childhood trauma, were not included in this study, but should also be considered alongside these results. Considerable evidence has emerged for an epigenetic component to FKBP5’s role in suicidal behavior, with particular focus on demethylation within and around GREs [62]. Furthermore, interacting factors, such as high levels of early childhood trauma, have been shown to be positively associated with increased risk of suicide attempt [13, 14]. We did not have access to early childhood history for all subjects, so we could not investigate this gene-by-environment effect. An additional limitation of the replication dataset is the skewed proportion of male suicide attempters (15.85%; S2 Table) caused by the restricted number of remaining NIMH Genetics Initiative samples.

We have completed a large, targeted sequencing project focused on pursuing both coding regions and regulatory regions within FKBP5, a gene previously associated with suicidal behavior. While this study generated several nominally significant results and one statistically significant result, none of them survive correction for multiple testing and/or replication. Thus, it does not provide consistent support for a role of FKBP5 variants increasing the risk of suicidal behavior. The failure to replicate the top finding reinforces the need for larger sample sizes in genetic studies. Greater datasets will more definitively test the relationship between FKBP5 and suicidal behavior.

Supporting Information

[Figure omitted. See PDF.]

S1 Table. Demographic characteristics of sample set.

https://doi.org/10.1371/journal.pone.0169158.s001

(DOCX)

S2 Table. Demographic characteristics of the replication sample set.

https://doi.org/10.1371/journal.pone.0169158.s002

(DOCX)

S3 Table. Single-variant results of all 481 FKBP5 variants analyzed.

https://doi.org/10.1371/journal.pone.0169158.s003

(XLSX)

S4 Table. Female-specific single-variant results of all 481 FKBP5 variants analyzed.

https://doi.org/10.1371/journal.pone.0169158.s004

(XLSX)

S5 Table. Male-specific single-variant results of all 481 FKBP5 variants analyzed.

https://doi.org/10.1371/journal.pone.0169158.s005

(XLSX)

S6 Table. Single-variant sex-specific results with a P-value < 0.05.

https://doi.org/10.1371/journal.pone.0169158.s006

(DOCX)

S7 Table. Expanded dataset information for distribution and meta-analyses.

https://doi.org/10.1371/journal.pone.0169158.s007

(XLSX)

S1 File. Additional institution information for sample collection.

https://doi.org/10.1371/journal.pone.0169158.s008

(DOCX)

Author Contributions

1. Conceptualization: MEB SCG ETM KdK MGP TAB APD PPZ JBP VLW.

2. Data curation: ETM.

3. Formal analysis: MEB ETM PPZ.

4. Funding acquisition: PPZ JBP VLW.

5. Investigation: MEB SCG MGP.

6. Methodology: ETM TAB APD.

7. Project administration: VLW.

8. Supervision: VLW.

9. Validation: MEB.

10. Visualization: MEB.

11. Writing – original draft: MEB VLW.

12. Writing – review & editing: MEB SCG ETM KdK MGP TAB APD PPZ JBP VLW.

Citation: Breen ME, Gaynor SC, Monson ET, de Klerk K, Parsons MG, Braun TA, et al. (2016) Targeted Sequencing of FKBP5 in Suicide Attempters with Bipolar Disorder. PLoS ONE 11(12): e0169158. https://doi.org/10.1371/journal.pone.0169158

References

1. WHO. Preventing Suicide: A Global Imperative. 2014;http://www.who.int/mental_health/suicide-prevention/world_report_2014/en/.

2. Mann JJ, Waternaux C, Haas GL, Malone KM. Toward a clinical model of suicidal behavior in psychiatric patients. The American journal of psychiatry. 1999;156(2):181–9. Epub 1999/02/16. pmid:9989552

3. Arsenault-Lapierre G, Kim C, Turecki G. Psychiatric diagnoses in 3275 suicides: a meta-analysis. BMC psychiatry. 2004;4:37. Epub 2004/11/06. pmid:15527502

4. Beautrais AL, Joyce PR, Mulder RT, Fergusson DM, Deavoll BJ, Nightingale SK. Prevalence and comorbidity of mental disorders in persons making serious suicide attempts: a case-control study. The American journal of psychiatry. 1996;153(8):1009–14. Epub 1996/08/01. pmid:8678168

5. Coryell W, Schlesser MA. Suicide and the dexamethasone suppression test in unipolar depression. The American journal of psychiatry. 1981;138(8):1120–1. Epub 1981/08/01. pmid:7258395

6. Coryell W, Schlesser M. The dexamethasone suppression test and suicide prediction. The American journal of psychiatry. 2001;158(5):748–53. Epub 2001/05/01. pmid:11329397

7. Jokinen J, Nordstrom P. HPA axis hyperactivity and attempted suicide in young adult mood disorder inpatients. Journal of affective disorders. 2009;116(1–2):117–20. Epub 2008/12/05. pmid:19054569

8. Binder EB. The role of FKBP5, a co-chaperone of the glucocorticoid receptor in the pathogenesis and therapy of affective and anxiety disorders. Psychoneuroendocrinology. 2009;34 Suppl 1:S186–95.

9. Willour VL, Chen H, Toolan J, Belmonte P, Cutler DJ, Goes FS, et al. Family-based association of FKBP5 in bipolar disorder. Molecular psychiatry. 2009;14(3):261–8. Epub 2008/01/09. pmid:18180755

10. Leszczynska-Rodziewicz A, Szczepankiewicz A, Narozna B, Skibinska M, Pawlak J, Dmitrzak-Weglarz M, et al. Possible association between haplotypes of the FKBP5 gene and suicidal bipolar disorder, but not with melancholic depression and psychotic features, in the course of bipolar disorder. Neuropsychiatric disease and treatment. 2014;10:243–8. pmid:24523589

11. Brent D, Melhem N, Ferrell R, Emslie G, Wagner KD, Ryan N, et al. Association of FKBP5 polymorphisms with suicidal events in the Treatment of Resistant Depression in Adolescents (TORDIA) study. The American journal of psychiatry. 2010;167(2):190–7. Epub 2009/12/17. pmid:20008943

12. Supriyanto I, Sasada T, Fukutake M, Asano M, Ueno Y, Nagasaki Y, et al. Association of FKBP5 gene haplotypes with completed suicide in the Japanese population. Progress in neuro-psychopharmacology & biological psychiatry. 2011;35(1):252–6.

13. Roy A, Gorodetsky E, Yuan Q, Goldman D, Enoch MA. Interaction of FKBP5, a stress-related gene, with childhood trauma increases the risk for attempting suicide. Neuropsychopharmacology: official publication of the American College of Neuropsychopharmacology. 2010;35(8):1674–83. Epub 2010/01/22.

14. Roy A, Hodgkinson CA, Deluca V, Goldman D, Enoch MA. Two HPA axis genes, CRHBP and FKBP5, interact with childhood trauma to increase the risk for suicidal behavior. Journal of psychiatric research. 2012;46(1):72–9. Epub 2011/10/08. pmid:21978546

15. Perez-Ortiz JM, Garcia-Gutierrez MS, Navarrete F, Giner S, Manzanares J. Gene and protein alterations of FKBP5 and glucocorticoid receptor in the amygdala of suicide victims. Psychoneuroendocrinology. 2012. Epub 2012/12/12.

16. Willour VL, Seifuddin F, Mahon PB, Jancic D, Pirooznia M, Steele J, et al. A genome-wide association study of attempted suicide. Molecular psychiatry. 2012;17(4):433–44. pmid:21423239

17. Galfalvy H, Haghighi F, Hodgkinson C, Goldman D, Oquendo MA, Burke A, et al. A genome-wide association study of suicidal behavior. American journal of medical genetics Part B, Neuropsychiatric genetics: the official publication of the International Society of Psychiatric Genetics. 2015;168(7):557–63.

18. Zai CC, Goncalves VF, Tiwari AK, Gagliano SA, Hosang G, de Luca V, et al. A genome-wide association study of suicide severity scores in bipolar disorder. Journal of psychiatric research. 2015;65:23–9. pmid:25917933

19. Sokolowski M, Wasserman J, Wasserman D. Polygenic associations of neurodevelopmental genes in suicide attempt. Molecular psychiatry. 2015.

20. Bani-Fatemi A, Graff A, Zai C, Strauss J, De Luca V. GWAS analysis of suicide attempt in schizophrenia: Main genetic effect and interaction with early life trauma. Neuroscience letters. 2016;622:102–6. pmid:27109790

21. Perlis RH, Huang J, Purcell S, Fava M, Rush AJ, Sullivan PF, et al. Genome-wide association study of suicide attempts in mood disorder patients. The American journal of psychiatry. 2010;167(12):1499–507. pmid:21041247

22. Gross JA, Bureau A, Croteau J, Galfalvy H, Oquendo MA, Haghighi F, et al. A genome-wide copy number variant study of suicidal behavior. PloS one. 2015;10(5):e0128369. pmid:26010658

23. Mullins N, Perroud N, Uher R, Butler AW, Cohen-Woods S, Rivera M, et al. Genetic relationships between suicide attempts, suicidal ideation and major psychiatric disorders: a genome-wide association and polygenic scoring study. American journal of medical genetics Part B, Neuropsychiatric genetics: the official publication of the International Society of Psychiatric Genetics. 2014;165B(5):428–37.

24. Schosser A, Butler AW, Ising M, Perroud N, Uher R, Ng MY, et al. Genomewide association scan of suicidal thoughts and behaviour in major depression. PloS one. 2011;6(7):e20690. Epub 2011/07/14. pmid:21750702

25. Laje G, Allen AS, Akula N, Manji H, John Rush A, McMahon FJ. Genome-wide association study of suicidal ideation emerging during citalopram treatment of depressed outpatients. Pharmacogenet Genomics. 2009;19(9):666–74. pmid:19724244

26. Perroud N, Uher R, Ng MY, Guipponi M, Hauser J, Henigsberg N, et al. Genome-wide association study of increasing suicidal ideation during antidepressant treatment in the GENDEP project. The pharmacogenomics journal. 2012;12(1):68–77. pmid:20877300

27. Menke A, Domschke K, Czamara D, Klengel T, Hennings J, Lucae S, et al. Genome-wide association study of antidepressant treatment-emergent suicidal ideation. Neuropsychopharmacology: official publication of the American College of Neuropsychopharmacology. 2012;37(3):797–807.

28. Genomic survey of bipolar illness in the NIMH genetics initiative pedigrees: a preliminary report. American journal of medical genetics. 1997;74(3):227–37. pmid:9184304

29. Nurnberger JI Jr., Blehar MC, Kaufmann CA, York-Cooler C, Simpson SG, Harkavy-Friedman J, et al. Diagnostic interview for genetic studies. Rationale, unique features, and training. NIMH Genetics Initiative. Archives of general psychiatry. 1994;51(11):849–59; discussion 63–4. Epub 1994/11/01. pmid:7944874

30. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nature genetics. 2006;38(8):904–9. pmid:16862161

31. Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, Zahler AM, et al. The human genome browser at UCSC. Genome research. 2002;12(6):996–1006. pmid:12045153

32. Harrow J, Frankish A, Gonzalez JM, Tapanari E, Diekhans M, Kokocinski F, et al. GENCODE: the reference human genome annotation for The ENCODE Project. Genome research. 2012;22(9):1760–74. pmid:22955987

33. Cunningham F, Amode MR, Barrell D, Beal K, Billis K, Brent S, et al. Ensembl 2015. Nucleic acids research. 2015;43(Database issue):D662–9. pmid:25352552

34. Pruitt KD, Brown GR, Hiatt SM, Thibaud-Nissen F, Astashyn A, Ermolaeva O, et al. RefSeq: an update on mammalian reference sequences. Nucleic acids research. 2014;42(Database issue):D756–63. pmid:24259432

35. Encode Project Consortium. An integrated encyclopedia of DNA elements in the human genome. Nature. 2012;489(7414):57–74. pmid:22955616

36. Gaynor SC, Breen ME, Monson ET, de Klerk K, Parsons M, DeLuca AP, et al. A targeted sequencing study of glutamatergic candidate genes in suicide attempters with bipolar disorder. American journal of medical genetics Part B, Neuropsychiatric genetics: the official publication of the International Society of Psychiatric Genetics. 2016;171(8):1080–7.

37. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009;25(14):1754–60. pmid:19451168

38. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009;25(16):2078–9. pmid:19505943

39. Barnett DW, Garrison EK, Quinlan AR, Stromberg MP, Marth GT. BamTools: a C++ API and toolkit for analyzing and managing BAM files. Bioinformatics. 2011;27(12):1691–2. pmid:21493652

40. McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome research. 2010;20(9):1297–303. pmid:20644199

41. Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic acids research. 2010;38(16):e164. pmid:20601685

42. Heinze G, Puhr R. Bias-reduced and separation-proof conditional logistic regression with small or sparse data sets. Stat Med. 2010;29(7–8):770–7. pmid:20213709

43. 1000 Genomes Project Consortium, Auton A, Brooks LD, Durbin RM, Garrison EP, Kang HM, et al. A global reference for human genetic variation. Nature. 2015;526(7571):68–74. pmid:26432245

44. Wu MC, Lee S, Cai T, Li Y, Boehnke M, Lin X. Rare-variant association testing for sequencing data with the sequence kernel association test. American journal of human genetics. 2011;89(1):82–93. pmid:21737059

45. Li B, Leal SM. Methods for detecting associations with rare variants for common diseases: application to analysis of sequence data. American journal of human genetics. 2008;83(3):311–21. Epub 2008/08/12. pmid:18691683

46. Ng PC, Henikoff S. SIFT: Predicting amino acid changes that affect protein function. Nucleic acids research. 2003;31(13):3812–4. Epub 2003/06/26. pmid:12824425

47. Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, Bork P, et al. A method and server for predicting damaging missense mutations. Nature methods. 2010;7(4):248–9. Epub 2010/04/01. pmid:20354512

48. Chun S, Fay JC. Identification of deleterious mutations within three human genomes. Genome research. 2009;19(9):1553–61. pmid:19602639

49. Carter H, Douville C, Stenson PD, Cooper DN, Karchin R. Identifying Mendelian disease genes with the variant effect scoring tool. BMC genomics. 2013;14 Suppl 3:S3.

50. Schwarz JM, Rodelsperger C, Schuelke M, Seelow D. MutationTaster evaluates disease-causing potential of sequence alterations. Nature methods. 2010;7(8):575–6. pmid:20676075

51. Boyle AP, Hong EL, Hariharan M, Cheng Y, Schaub MA, Kasowski M, et al. Annotation of functional variation in personal genomes using RegulomeDB. Genome research. 2012;22(9):1790–7. pmid:22955989

52. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005;21(2):263–5. pmid:15297300

53. Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B, et al. The structure of haplotype blocks in the human genome. Science. 2002;296(5576):2225–9. pmid:12029063

54. Sherry ST, Ward MH, Kholodov M, Baker J, Phan L, Smigielski EM, et al. dbSNP: the NCBI database of genetic variation. Nucleic acids research. 2001;29(1):308–11. pmid:11125122

55. Perroud N, Bondolfi G, Uher R, Gex-Fabry M, Aubry JM, Bertschy G, et al. Clinical and genetic correlates of suicidal ideation during antidepressant treatment in a depressed outpatient sample. Pharmacogenomics. 2011;12(3):365–77. pmid:21449676

56. Fudalej S, Kopera M, Wolynczyk-Gmaj D, Fudalej M, Krajewski P, Wasilewska K, et al. Association between FKBP5 Functional Polymorphisms and Completed Suicide. Neuropsychobiology. 2015;72(2):126–31. pmid:26630184

57. Yin H, Galfalvy H, Pantazatos SP, Huang YY, Rosoklija GB, Dwork AJ, et al. Glucocorticoid Receptor-Related Genes: Genotype and Brain Gene Expression Relationships to Suicide and Major Depressive Disorder. Depression and anxiety. 2016.

58. Breen ME, Seifuddin F, Zandi PP, Potash JB, Willour VL. Investigating the role of early childhood abuse and HPA axis genes in suicide attempters with bipolar disorder. Psychiatric genetics. 2015;25(3):106–11. pmid:25714448

59. Baca-Garcia E, Perez-Rodriguez MM, Keyes KM, Oquendo MA, Hasin DS, Grant BF, et al. Suicidal ideation and suicide attempts in the United States: 1991–1992 and 2001–2002. Molecular psychiatry. 2010;15(3):250–9. pmid:18779820

60. Nock MK, Green JG, Hwang I, McLaughlin KA, Sampson NA, Zaslavsky AM, et al. Prevalence, correlates, and treatment of lifetime suicidal behavior among adolescents: results from the National Comorbidity Survey Replication Adolescent Supplement. JAMA psychiatry. 2013;70(3):300–10. pmid:23303463

61. Klengel T, Mehta D, Anacker C, Rex-Haffner M, Pruessner JC, Pariante CM, et al. Allele-specific FKBP5 DNA demethylation mediates gene-childhood trauma interactions. Nature neuroscience. 2013;16(1):33–41. pmid:23201972

62. Klengel T, Binder EB. FKBP5 allele-specific epigenetic modification in gene by environment interaction. Neuropsychopharmacology: official publication of the American College of Neuropsychopharmacology. 2015;40(1):244–6.