Transcriptomic decoding of regional cortical vulnerability to major depressive disorder

Abstract

Previous studies in small samples have identified inconsistent cortical abnormalities in major depressive disorder (MDD). Despite genetic influences on MDD and the brain, it is unclear how genetic risk for MDD is translated into spatially patterned cortical vulnerability. Here, we initially examined voxel-wise differences in cortical function and structure using the largest multi-modal MRI data from 1660 MDD patients and 1341 controls. Combined with the Allen Human Brain Atlas, we then adopted transcription-neuroimaging spatial correlation and the newly developed ensemble-based gene category enrichment analysis to identify gene categories with expression related to cortical changes in MDD. Results showed that patients had relatively circumscribed impairments in local functional properties and broadly distributed disruptions in global functional connectivity, consistently characterized by hyper-function in associative areas and hypo-function in primary regions. Moreover, the local functional alterations were correlated with genes enriched for biological functions related to MDD in general (e.g., endoplasmic reticulum stress, mitogen-activated protein kinase, histone acetylation, and DNA methylation); and the global functional connectivity changes were associated with not only MDD-general, but also brain-relevant genes (e.g., neuron, synapse, axon, glial cell, and neurotransmitters). Our findings may provide important insights into the transcriptomic signatures of regional cortical vulnerability to MDD.

Fig. 1. A schematic overview of the study design and analysis pipeline.

Abbreviations: MDD major depressive disorder, MRI magnetic resonance imaging, ALFF amplitude of low-frequency fluctuations, ReHo regional homogeneity, DC degree centrality, GMV gray matter volume, GCEA gene category enrichment analysis, AHBA Allen Human Brain Atlas. This image and every element of this image was created by J.Z.

Fig. 2. Demographic and clinical characteristics of the samples across 23 sites.

Abbreviations: MDD major depressive disorder, HAMD Hamilton rating scale for depression.

Fig. 3. Cortical functional and structural differences between MDD patients and healthy controls.

There was a mix of increased and decreased ALFF (A), ReHo (B), and DC (C) as well as reduced GMV (D) in MDD patients relative to controls (P < 0.05, cluster-level FWE corrected). Warm and cold colors respectively denote increase and decrease in imaging measurements in patients. Abbreviations: ALFF amplitude of low-frequency fluctuations, ReHo regional homogeneity, DC degree centrality, GMV gray matter volume, MDD major depressive disorder, FWE family-wise error.

Fig. 4. Gene categories associated with cortical functional changes in MDD.

A combination of transcription-neuroimaging spatial correlation and the ensemble-based GCEA revealed that ALFF (A) and ReHo (B) changes were negatively correlated with MDD-general GO categories; DC changes were negatively associated with MDD-general GO categories (C) and positively associated with brain-relevant GO categories (D). Abbreviations: ALFF amplitude of low-frequency fluctuations, ReHo regional homogeneity, DC degree centrality, GO gene ontology, BP biological process, CC cellular component, MF molecular function, ER endoplasmic reticulum, MAPK mitogen-activated protein kinase, NK natural killer, VEGF vascular endothelial growth factor, MDD major depressive disorder, GCEA gene category enrichment analysis.