Distinct Neurodevelopmental Signatures of Sex-Specific Transcriptome-Based Polygenic Risk Scores for Depression

Abstract

Major depressive disorder (MDD) impacts females and males differently and post-mortem gene expression profiling reveals distinct transcriptomic signatures of the disorder in each sex. Using genes that are transcriptionally altered in MDD in both sexes, we recently developed a novel transcriptome-based polygenic risk score (tPRS), which had sex-specific associations with brain structure and depressive symptoms in both adolescents and adults. Identifying the neurodevelopmental signatures of genetically-induced shifts toward a depression-like brain transcriptome in each sex during a crucial stage, when sex differences in MDD vulnerability initially manifest, could provide useful information about the developmental pathways of early MDD risk. Leveraging sex-specific MDD gene expression data, we sought to develop female- and male-specific tPRSs (tPRS-F and tPRS-M, respectively) and evaluate their impact on regional cortical thickness, cortical surface area, subcortical volume, and depressive symptoms at baseline and 2-year follow-up in a developmental sample of 5002 adolescents (46.6% female, aged 8.9-11.0). In males, tPRS-M was associated with higher depressive symptoms at both timepoints and thicker left posterior cingulate at follow-up. In females, tPRS-F was associated with lower volumes of the right accumbens area, right caudate, and bilateral hippocampi at follow-up. Subcortical volumes of the right caudate and right hippocampus further mediated an indirect effect of tPRS-F on depressive symptoms. For each sex-specific PRS, no effects emerged in the opposite sex. Our findings suggest that sex-specific depression-like shifts in gene expression may contribute to unique vulnerability phenotypes for future MDD risk via distinct mechanisms in each sex.

Figure 1.. tPRS-M and depressed symptoms in males at T1 and T2.

A significant association occurred at T1 between tPRS-M and higher (a) withdrawn/depressed symptoms, while similar associations with anxious/depressed symptoms occurred at both T1 (b) and at T2 (c). Age, 10 genetic principle components, and study site were regressed out of the CBCL scale measures.

Figure 2.. tPRS-F and regional subcortical volume in females at T1 and T2, and associated regional mediations.

Nominally and pFDR significant lower volume in subcortical structures are depicted at T1 (a) and T2 (b). Indirect effects of tPRS-F on withdrawn/depressed symptoms were significantly mediated by the right accumbens (c) and right hippocampus (d). 10 genetic principle components, eTIV, and study site were regressed out of the subcortical volume measure in each plot. p < 0.1 (‘); p < 0.05 (*); p < 0.01 (**).

Figure 3.. tPRS-M cortical thickness effects in males at T1 and T2.

Associations between tPRS-M and cortical thickness in males are depicted at T1 (a) and T2 (b). At T2, the right posterior cingulate demonstrated pFDR significant increased cortical thickness. 10 genetic principle components, and study site were regressed out.