Transcriptomic dysfunction disparities: greater burden in female brains highlights immune and synaptic pathways for psychiatric disorders

Many psychiatric disorders exhibit strong sex differences, yet the underlying mechanisms of such sex differences remains poorly understood. Inspired by the classic genetic liability model, which hypothesized that the genetic liability threshold differs between sexes, we derived a novel hypothesis: burden of transcriptome dysfunction of psychiatric disorders differs between sexes. The burden, defined as the amount, degree, and organization of changes between cases and controls, of transcriptome dysfunction may reveal sex-related contributions to the disorders through both genetic and environmental factors. We analyzed transcriptome data of 2160 postmortem brain samples from the PsychENCODE project in a sex-stratified design in schizophrenia (SCZ), bipolar disorder (BD), and autism spectrum disorder (ASD). We found that the females have a higher burden of transcriptome dysfunction than males in SCZ, ASD, and BD at three strategies: 1) The overall expression changes in female patients were greater than in male patients, and more than 37% of genes show the opposite direction of changes comparing males to females in SCZ, BD, and ASD. The 768 differentially expressed genes shared by female and male SCZ have a significantly larger effect size (1.37 times) in females than in males (p < 2.2e-16). 2) The changes in the connectivity of co-expressed modules were also greater in females than in males, which indicates the organization disruption of co-expression in the patients was more severe in females than in males. Integrating the differential expression and co-expression analysis, we found the up-regulated genes enriched for immune-related pathways in females with SCZ and ASD, while these genes’ connectivity was significantly reduced in female patients but not in male patients. 3) We explored sex-specific genetic regulation of multi-omics data that included chromatin accessibility, gene expression, splicing, translation, and protein. More than 15% of the quantitative trait loci (QTL) detected in females are female-specific. About 35% of the QTLs are male-specific. More SNP-based heritability in SCZ can be explained using sex-specific QTLs in females than in males for all the five-omics measures. The average enrichment of heritability for SCZ is bigger in female than male (1.4 vs 1.2). Such a difference cannot be explained by the difference of sample size. Lastly, the sex differences in transcriptome predicted distinct drugs for SCZ in males and females, suggesting the treatment should be optimized differentially for males and females. This is the first study of sex-biased burden of risk of psychiatric disorders at brain transcriptome. We explored the sex difference in the burden of transcriptome dysfunction in brain tissue and different cell types (pending) focusing on differential expressed genes, co-expression network, and genetic regulation. We discovered disparities between men and women by the number of genes or networks involved and by the effect size of specific genes and organization of their regulatory networks. It provided the most comprehensive dissection for the sex difference in brain transcriptome of psychiatric disorders.