The Role of Energy Homeostasis in Depression Pathophysiology and Its Heterogeneity

Abstract

Importance: Energy homeostatic dysregulation may constitute 1 module of the heterogeneous pathophysiology of major depressive disorder (MDD), potentially manifesting as a distinctive symptom profile.

Objective: To test whether the shared genetic liability of metabolic, interoceptive, and motivational pathways involved in energy homeostasis regulation is associated with the expression of specific MDD symptoms.

Design, setting, and participants: This study used summary-level data from large genome-wide association studies and individual-level data from 2 prospective psychiatric cohorts, the CoLaus|PsyCoLaus (population-based) and Netherlands Study of Depression and Anxiety (NESDA; clinically enriched) cohorts. Data were retrieved and analyzed from May 2023 through November 2024. A lifetime diagnosis of MDD was ascertained with semistructured diagnostic interviews. The sample comprised 1407 MDD cases and 2020 controls from CoLaus|PsyCoLaus and 1803 MDD cases and 266 controls from NESDA.

Exposures: Genomic structural equation modeling was applied to model a unique underlying factor capturing the common genetic liability shared among metabolic and interoceptive signals (body mass index, triglycerides, fasting glucose, C-reactive protein, leptin) and motivational (anhedonia) processes. From this underlying factor, a polygenic score (PGS) was derived, indexing the shared genetic liability of traits potentially involved in energy homeostasis regulation.

Main outcomes and measures: A total of 15 depressive symptoms endorsed by participants during MDD.

Results: Among 1407 MDD cases (66.2% female; median year of birth [YOB], 1956) and 2020 controls (44.3% female; median YOB, 1955) from CoLaus|PsyCoLaus and 1803 MDD cases (68.3% female; median YOB, 1962) and 266 controls (56.0% female; median YOB, 1960) from NESDA, multiple significant bidirectional mendelian randomization estimates and genetic correlations (r = 0.11-0.81) indicated a shared genetic basis between the selected traits, which was modeled as a latent homeostatic factor with genomic structural equation modeling. In cohort data, the PGS indexing the latent homeostatic factor was significantly (false discovery rate, <5%) higher in MDD cases endorsing appetite increase and hypersomnia when contrasted with both controls (appetite increase odds ratio [OR], 2.25 [95% CI, 2.00-2.53]; P = 9.03 × 10-41; hypersomnia OR, 1.22 [95% CI, 1.10-1.35]; P = 1.15 × 10-04) and other MDD cases (appetite increase OR, 1.88 [95% CI, 1.63-2.18]; P = 2.38 × 10-17; hypersomnia OR, 1.18 [95% CI, 1.05-1.33]; P = 5.80 × 10-03).

Conclusions and relevance: This study identified a module of depression pathophysiology characterized by altered energy homeostasis and associated with the expression of specific symptoms reflecting energy saving and intake responses. These findings could be used to identify patients with depression at higher metabolic risk and could pave the way for the development of targeted treatments.