Large-scale proteomic analyses before depression diagnosis reveal novel pathophysiological insights

The early pathophysiology of depression is poorly understood. We elucidated the decadal temporal evolution of plasma proteomic alterations before depression diagnosis and evaluated their associations with comorbid conditions and neuroanatomical changes.

This study analyzed 31,114 depression-free participants and identified 1555 incident depression cases after a median follow-up of 7.8 years. Cox regression was used to identify depression-associated proteins, adjusting for sociodemographic, lifestyle, and genetic factors. Subsequent analyses of depression-related proteins included exome-wide association analysis (EWAS), temporal change modeling of pre-diagnostic protein dynamics via LOESS regression, association analyses with eight comorbid conditions and 58 regional brain volumes, and LightGBM-based predictive modeling.

We found 64 depression-related proteins, with PIGR, HAVCR2, and IL4R validated in EWAS. For example, PIGR exhibited risk effects for depression (HR = 1.26, 95%CI: 1.13-1.40) and comorbid conditions, particularly diabetes (HR = 2.34, 95%CI: 2.12-2.58). Temporal profiling identified three protein clusters: one (cell-matrix adhesion) characterized by initial stability and subsequent decline, another (including PIGR and HAVCR2) characterized by MAPK cascade activation, and the third characterized by increased apoptosis and immune response. Neuroimaging correlations confirmed that elevated PIGR levels were associated with reduced volume in the bilateral ventral diencephalon (β = -0.062--0.061). A predictive model combined proteins and clinical features, achieving superior accuracy in depression prediction after 15 years (area under the curve = 0.74). Our findings reveal early peripheral pathophysiological changes in depression, suggesting a progression that may involve early apoptotic processes, an intermediate inflammatory phase, and later proteolytic dysregulation.

These insights hold significant potential for developing early biomarkers and precision therapies.