Associations between CRP-related DNA methylation, stress exposure, and depression severity in a longitudinal clinical cohort

BACKGROUND: Environmental adversity is linked to major depressive disorder (MDD), potentially via sustained low-grade inflammation. However, serum markers such as C-reactive protein (CRP) are transient and sensitive to acute states. In contrast, epigenetic signatures of inflammation may provide a more stable trace of how stress becomes biologically embedded and contributes to depression risk over time. METHODS: In a subsample of the Marburg-Münster Affective Disorders Cohort Study (MACS; N = 579; 320 healthy controls, 259 with MDD), we examined whether early life adversity (ELA; CTQ) and recent life stress (RLS; LEQ) are associated with CRP-related DNA methylation (CRPm) at baseline. We further tested whether CRPm predicts depressive symptom severity (HAMD) at baseline and at two-year follow-up (n = 407). DNA was extracted from whole blood, and CRPm scores were computed using publicly available genome-wide summary statistics. RESULTS: CRPm explained 21.3% of the variance in serum high-sensitivity CRP (hsCRP). Higher CRPm was significantly associated with both ELA (b = 0.01, SE = 0.003, p = 0.017) and RLS (b = 0.01, SE = 0.004, p = 0.032), after adjusting for age and sex. CRPm also predicted depressive symptom severity at baseline (b = 0.68, SE = 0.27, p = 0.013) and at follow-up (b = 0.79, SE = 0.25, p = 0.002). These associations remained after controlling for white blood cell-type composition but were attenuated after adjusting for BMI and smoking. In contrast, hsCRP was not associated with adversity or depressive symptoms. CONCLUSION: Our study indicates that a methylation-based index of chronic inflammation is associated with stress exposure and depressive symptoms over time, in contrast to fluctuating serum hsCRP. The findings are more consistent with an indirect pathway in which environmental adversity is linked to inflammatory biology via stress-related health behaviors, rather than with a model of direct biological embedding.