Transcriptional dissection of symptomatic profiles across the brain of men and women with depression

Major Depressive Disorder (MDD) is a significant global health concern, affecting over 280 million people annually, with women experiencing higher prevalence and severity. Current treatments are often inefficient, highlighting the need to understand MDD's clinical heterogeneity. MDD is characterized by core symptoms like depressed mood and anhedonia, along with varying cognitive impairments, anxiety, sleep disturbances, and suicidal ideations. The current diagnostic approach, based on subjective behavioral measures, limits our ability to develop effective treatments. Functional imaging studies have linked specific MDD symptoms to distinct brain regions and circuits, but the underlying molecular mechanisms remain unclear. Previous research has shown transcriptional changes in brain regions of MDD patients and animal models, but the association between these changes and symptom expression hasn't been fully explored. This study employed data-driven, system-based network analyses to investigate the potential association between transcriptional signatures across brain regions and the expression of distinct symptom domains in males and females with MDD. The aim was to identify whether transcriptional signatures across different brain regions are associated with specific symptomatic profiles in males and females.

Functional neuroimaging studies have identified associations between specific MDD symptoms and the activity of distinct brain regions and circuits. For example, hyperconnectivity within the default mode network has been linked to rumination, sadness, and hopelessness, while hypoconnectivity in other networks is associated with anxiety, anhedonia and cognitive deficits. Previous transcriptional studies have revealed gene expression changes in various brain regions in MDD patients and animal models. These studies have highlighted the role of gene networks in mediating stress susceptibility in a sex-specific manner and revealed transcriptional signatures associated with trait versus state depression in males. However, none had directly linked these signatures to specific MDD symptom domains in both sexes. Network-based analyses have proven useful in characterizing the molecular architecture of complex neuropsychiatric disorders, including MDD.

This study analyzed RNA sequencing data from post-mortem brain tissues of 89 individuals (25 males and 25 females with MDD, and 17 males and 22 females as healthy controls) across six brain regions: anterior insula (aINS), orbitofrontal cortex (OFC), ventromedial prefrontal cortex (vmPFC), dorsolateral prefrontal cortex (dIPFC), nucleus accumbens (NAc), and ventral subiculum (vSub). The data included in this study comes from two cohorts - one previously published and another newly sequenced. Differential expression analysis (using limma) identified genes significantly up- or downregulated in each brain region for males and females with MDD compared to controls. Rank-rank hypergeometric overlap (RRHO) analysis assessed transcriptional overlap between sexes. Gene ontology analysis identified enriched functional pathways. Weighted gene co-expression network analysis (WGCNA) constructed gene co-expression networks in each brain region, separately for males and females. Module differential connectivity (MDC) measured changes in intramodular connectivity between MDD and control groups. Module preservation assessed the similarity of network organization between sexes. Finally, associations between gene modules and MDD symptom domains (insomnia/hypersomnia, change in appetite/weight, psychomotor agitation/retardation, low self-esteem, and difficulty in concentration/indecision) were determined using point-biserial correlations. Data preprocessing included normalization using voom, batch effect correction using RUVseq, and adjustment for covariates identified through principal component analysis (PCA), including postmortem interval (PMI), pH, age, drug abuse, and RNA integrity number (RIN).

Differential expression analysis revealed substantial sex differences in gene expression profiles across all six brain regions, with limited overlap between males and females, except in cortical regions. RRHO analysis confirmed significant overlap in cortical regions but not in limbic structures. WGCNA identified numerous gene modules, with a significant proportion of modules preserved across sexes, particularly in the nucleus accumbens, anterior insula, and ventromedial prefrontal cortex. Many unique modules in each sex showed significant module differential connectivity (MDC) compared to controls, reflecting changes in intramodular connectivity. Analysis of associations between modules and specific symptoms revealed a complex pattern of sex-specific relationships. For example, in males, the OFC's 'Ivory' module strongly associated with changes in appetite/weight, insomnia/hypersomnia, and difficulty in concentration/indecision. This module was enriched for genes involved in GABAergic neurotransmission. In females, the anterior insula's 'Darkorange' module associated with changes in appetite/weight, and the dorsolateral prefrontal cortex's 'Saddlebrown' module associated with psychomotor agitation/retardation. A small number of modules showed similar associations in both sexes but often with opposite directions, suggesting that while common functional pathways may be involved, the specific genes and their effects differ between the sexes. The study also identified several modules which exhibited opposite correlations between gene networks and symptoms in males versus females.

This study provides a comprehensive analysis of sex-specific transcriptional signatures in MDD, linking specific gene networks in distinct brain regions to individual symptom domains. The findings demonstrate that although some network organization is preserved across sexes, the association of these networks with symptoms differs significantly between males and females. This suggests that common functional pathways may be disrupted in both sexes, but through the action of different genes. The identified gene networks may influence symptom expression by altering brain region activity and functional connectivity within larger circuits. For example, the GABAergic system alterations identified in males are consistent with the established role of this system in cognitive and emotional processing. The study highlights the importance of considering sex-specific molecular mechanisms in understanding and treating MDD.

This study reveals a complex interplay between sex, brain region, gene networks, and symptom expression in MDD. Future research should focus on replicating these findings in larger cohorts, investigating the functional roles of the identified gene networks and their effects on brain circuit activity, and exploring the implications for sex-specific therapeutic interventions. The integration of systems biology approaches, like the one used here, with large-scale genetic studies, could greatly improve the precision of MDD diagnosis and treatment.

This study used post-mortem brain tissue, limiting the ability to assess symptom intensity and recurrence. The reliance on psychological autopsies for clinical information might introduce some subjectivity. The relatively small sample size, while larger than many previous studies, may limit the power to detect subtle effects or complex interactions. The study focused on a limited set of symptom domains, and future research should consider a broader range of symptoms and their interactions. Finally, the study did not empirically confirm the functional roles of predicted gene networks.