Cytomegalovirus-related immune activation and major neuropsychiatric disorders: Mendelian randomization and bulk...

Cytomegalovirus (CMV) establishes lifelong latency and induces persistent changes in peripheral immunity. Observational links to neuropsychiatric disorders are difficult to interpret because of confounding and reverse causation.

We performed two-sample, multivariable, and bidirectional Mendelian randomization (MR) on genetically predicted CMV seropositivity and antigen-specific antibody levels for Parkinson's disease (PD), bipolar disorder (BD), schizophrenia (SCZ), major depressive disorder (MDD), and Alzheimer's disease (AD). Instruments were selected at P < 5 × 10^-6 (secondary at P < 5 × 10^-7). Multivariable MR (MVMR) sensitivity analysis accounting for age-related genetic structure, pleiotropy/heterogeneity tests, and reverse MR evaluated robustness. Bulk transcriptomic validation tested whether CMV CD14+ late-response and HCMV reactivation-related host signatures were enriched in public PD and BD whole-blood datasets.

Genetically predicted anti-CMV IgG seropositivity was associated with PD (MVMR sensitivity aOR: 1.12, P = 0.020), with directional consistency in a single-instrument secondary analysis (OR: 1.25, P = 0.003). Anti-CMV pp28 antibody levels were associated with BD (aOR: 1.12, P = 0.043). The SCZ signal attenuated after adjustment, and no robust evidence was observed for MDD or AD. Both CMV-related host signatures were significantly enriched among PD-upregulated genes (CMV CD14+ NES = 1.85, FDR = 2.73 × 10^-6; HCMV reactivation NES = 2.03, FDR = 9.11 × 10^-9); BD showed no robust positive enrichment after covariate adjustment.

Integrated genetic and transcriptomic evidence prioritizes CMV-related peripheral immune activation as a plausible contributor to PD risk. The pp28-BD signal is hypothesis-generating and requires validation in datasets with direct CMV serostatus.