Effects of genetically predicted posttraumatic stress disorder on autoimmune phenotypes

Abstract

Observational studies suggest that posttraumatic stress disorder (PTSD) increases risk for various autoimmune diseases. Insights into shared biology and causal relationships between these diseases may inform intervention approaches to PTSD and co-morbid autoimmune conditions. We investigated the shared genetic contributions and causal relationships between PTSD, 18 autoimmune diseases, and 3 immune/inflammatory biomarkers. Univariate MiXeR was used to contrast the genetic architectures of phenotypes. Genetic correlations were estimated using linkage disequilibrium score regression. Bi-directional, two-sample Mendelian randomization (MR) was performed using independent, genome-wide significant single nucleotide polymorphisms; inverse variance weighted and weighted median MR estimates were evaluated. Sensitivity analyses for uncorrelated (MR PRESSO) and correlated horizontal pleiotropy (CAUSE) were also performed. PTSD was considerably more polygenic (10,863 influential variants) than autoimmune diseases (median 255 influential variants). However, PTSD evidenced significant genetic correlation with nine autoimmune diseases and three inflammatory biomarkers. PTSD had putative causal effects on autoimmune thyroid disease (p = 0.00009) and C-reactive protein (CRP) (p = 4.3 × 10^-7). Inferences were not substantially altered by sensitivity analyses. Additionally, the PTSD-autoimmune thyroid disease association remained significant in multivariable MR analysis adjusted for genetically predicted inflammatory biomarkers as potential mechanistic pathway variables. No autoimmune disease had a significant causal effect on PTSD (all p values > 0.05). Although causal effect models were supported for associations of PTSD with CRP, shared pleiotropy was adequate to explain a putative causal effect of CRP on PTSD (p = 0.18). In summary, our results suggest a significant genetic overlap between PTSD, autoimmune diseases, and biomarkers of inflammation. PTSD has a putative causal effect on autoimmune thyroid disease, consistent with existing epidemiologic evidence. A previously reported causal effect of CRP on PTSD is potentially confounded by shared genetics. Together, results highlight the nuanced links between PTSD, autoimmune disorders, and associated inflammatory signatures, and suggest the importance of targeting related pathways to protect against disease and disability.

Fig. 1. Genetic architectures of phenotypes evaluated.

a SNP based heritability of assessed phenotypes. SNP based heritability estimates for autoimmune diseases are provided on the liability scale, assuming disease specific prevalence. Phenotypes are colored by category. Black bars ends indicate 95% confidence intervals. Confidence intervals extending beyond the plot range are indicated with arrows. b Polygenicity and discoverability components of heritability. The x-axis depicts the number of influential variants necessary to explain 90% of SNP based heritability (polygenicity ✕ constant). The y-axis depicts the discoverability of the phenotype. Circle sizes indicate relative SNP based heritability values. NMO Neuromyelitis Optica Spectrum Disorder, EGPA Eosinophilic granulomatosis with polyangiitis.

Fig. 2. Genetic correlations between PTSD and immune-related phenotypes.

Genetic correlations (r_g) are indicated by circles that are drawn along the x axis. Phenotypes are colored by domain. Hollow circles indicate SNP based heritability (h²_SNP) z-score <4 in the immune-related phenotype GWAS (r_g estimates may be unreliable). The dotted vertical bar indicates the point of zero correlation. EGPA eosinophilic granulomatosis with polyangiitis.

Fig. 3. Putative causal effects of PTSD.

Causal effects of PTSD on immune-related phenotypes, as estimated by the genetically determined PTSD instrument (gPTSD), are indicated by circles that are drawn along the x axis. Phenotypes are colored by domain. Hollow circles indicate non-significance (p > 0.0025). The dotted vertical bar indicates the point of zero effect. Confidence intervals extending beyond the plot range are indicated with arrows. EGPA eosinophilic granulomatosis with polyangiitis.