Association between suicidal ideation and tandem repeats in contactins

Abstract

Background: Death by suicide is one of the leading causes of death among adolescents. Genome-wide association studies (GWAS) have identified loci that associate with suicidal ideation and related behaviours. One such group of loci are the six contactin genes (CNTN1-6) that are critical to neurodevelopment through regulating neurite structure. Because single nucleotide polymorphisms (SNPs) detected by GWAS often map to non-coding intergenic regions, we investigated whether repetitive variants in CNTNs associated with suicidality in a young cohort aged 8 to 21. Understanding the genetic liability of suicidal thought and behavior in this age group will promote early intervention and treatment.

Methods: Genotypic and phenotypic data were obtained from the Philadelphia Neurodevelopment Cohort (PNC). Across six CNTNs, 232 short tandem repeats (STRs) were analyzed in up to 4,595 individuals of European ancestry who expressed current, previous, or no suicidal ideation. STRs were imputed into SNP arrays using a phased SNP-STR haplotype reference panel from the 1000 Genomes Project. We tested several additive and interactive models of locus-level burden (i.e., sum of STR alleles) with respect to suicidal ideation. Additive models included sex, birth year, developmental stage ("DevStage"), and the first 10 principal components of ancestry as covariates; interactive models assessed the effect of STR-by-DevStage considering all other covariates.

Results: CNTN1-[T]_N interacted with DevStage to increase risk for current suicidal ideation (CNTN1-[T]_N-by-DevStage; p = 0.00035). Compared to the youngest age group, the middle (OR = 1.80, p = 0.0514) and oldest (OR = 3.82, p = 0.0002) participant groups had significantly higher odds of suicidal ideation as their STR length expanded; this result was independent of polygenic scores for suicidal ideation.

Discussion: These findings highlight diversity in the genetic effects (i.e., SNP and STR) acting on suicidal thoughts and behavior and advance our understanding of suicidal ideation across childhood and adolescence.

Keywords: adolescent psychiatry; contactins; genetics; mental health; suicide; tandem repeats.

Figure 1

Spearman correlation among 12 covariates (developmental stage, biological sex, and 10 PCs of genetic ancestry) and phenotype variables. “X” indicates a correlation with non-significant p value (p > 0.05). SUI001 = “Have you ever thought a lot about death or dying?”; SUI002 = “Have you ever thought about killing yourself?”; and SUI009 = “Are you currently (within the past month) having thoughts about suicide/death/dying/killing yourself?”.

Figure 2

Manhattan plot for all genetic associations between CNTN family STRs (n = 232) and suicidal ideation phenotypes (n = 3). The gray dashed line represents a p value threshold of 0.05. The black dashed line represents a p value threshold of 0.0005 after considering a false discovery rate of 5%.

Figure 3

Manhattan plot for all genetic associations between CNTN family STRs (n = 232) and suicidal ideation phenotypes (n = 3) with respect to participant birth year. The gray dashed line represents a p value threshold of 0.05. The black dashed line represents a p value threshold of 0.0005 after considering a false discovery rate of 5%.

Figure 4

Probability of current suicidal ideation (PNC Trait ID SUI009) among participants aged 14–19 (Developmental Stage 1 OR = 1.80, SE = 1.35) and 20–21 (Developmental Stage 2 OR = 3.82, SE = 1.43) in comparison to the youngest participant group, aged 9–13 (Development Stage 0).