CMV reshapes lymphoid immunity in aging: a single-cell atlas with predictive modeling

Abstract

Cytomegalovirus (CMV), a common herpesvirus, establishes lifelong latency and increases in prevalence with age; yet its systemic impact on the aging immune system remains incompletely understood. We profiled circulating immune cells from healthy older adults (median age: 73) who were CMV(+) or CMV(-) using single-cell RNA-sequencing and validated key findings by flow cytometry. CMV(+) individuals exhibited significant expansion of adaptive immune cells: CD4 and CD8 TEMRA T, GZMK+ CD8 T, {gamma}{delta} T, and atypical B cells. Among innate immune cells, monocytes and dendritic cells remained largely unchanged while KLRC2+ (adaptive) NK cells increased and CD56dim NK cells decreased. To facilitate CMV assessment in datasets with unknown CM serostatus, we developed CMVerify, a machine learning classifier that accurately predicts CMV serostatus from single-cell data across platforms and age groups (97% accuracy). These findings reveal extensive CMV-associated immune remodeling in older adults and underscore the importance of incorporating CMV status in studies of immune aging.