Epigenetic biomarkers of ageing, physiological frailty, and mortality in people with HIV: a longitudinal cohort study

Abstract

BackgroundApproaches that provide insight into ageing mechanisms for people with HIV are needed. Epigenetic modifications such as DNA methylation have been operationalised as epigenetic clocks, providing an estimate of age acceleration in years (DNA methylation age acceleration; DNAmAA). Our objective was to determine if measures of DNAmAA predict physiological frailty, as measured by the Veterans Aging Cohort Study (VACS) Index and all-cause mortality, in people with HIV. MethodsAs part of the observational VACS, we profiled epigenome-wide DNA methylation on peripheral blood samples collected from people with HIV as part of a biomarker cohort. We generated six established measures of DNAmAA in samples from 531 participants with VACS Index data at the time of sample collection. All-cause mortality was assessed over 10 years. For each DNAmAA measure, the hazard ratio (HR) per increased unit (1 year) was determined using Cox regression. To test the association of DNAmAA measures with the VACS Index 2.0, linear regression was performed. All regression models were performed with and without adjustment for chronological age. FindingsParticipants were mostly men (98·5% [n=523]; 1·5% [n=8] women); 83·4% (n=448) were Black, 10·5% (n=56) were White, and 5·1% (n=27) were Hispanic; and their mean age was 52·4 years (SD 7·8). There were 136 deaths during a median follow-up of 8·7 years (IQR 7·7–9·4). Hannum's DNAm age acceleration (Hannum AA), extrinsic epigenetic age acceleration (EEAA), Grim age acceleration (Grim AA), and Pheno age acceleration (Pheno AA) were significantly associated with all-cause mortality, but Horvath's DNAm age acceleration (Horvath AA) and intrinsic epigenetic age acceleration (IEAA) were not in unadjusted and age-adjusted models. GrimAA was the strongest predictor (adjusted HR 1·10, 95% CI 1·06–1·14; p<0·0001). Similarly, four of the six DNAmAA measures were significantly associated with the VACS Index in unadjusted and age-adjusted models. Each year increase in GrimAA was associated with a point increase in VACS Index (adjusted β 1·16, 95% CI 0·905–1·42; p<0·0001). InterpretationAmong people with HIV, several DNAmAA measures predicted all-cause mortality and physiological frailty. Each 1-year increase in GrimAA was associated with a 10% increased risk of death. A 5-year increase in GrimAA was associated with a clinically meaningful 5·8-point increase in VACS Index. GrimAA might provide key insights into the molecular mechanisms mediating biological ageing and physiological frailty, while informing the design of novel strategies to reduce mortality and morbidity for people with HIV. FundingKKO received funding from the US Department of Veterans Affairs (grant number I01 RX002790). VCM received funding from Emory (grant number CFAR P30 AI050409). ZW received funding from NIDDK (grant number R01DK125187). KX received funding from NIDA (grant number R01DA042691). MM received funding from NIA (grant numbers P30 AG031679 and P30AI060354). ACJ received funding from NIAAA (grant numbers U24-AA020794, U01-AA020790, U24-AA022001, U10 AA013566 [completed], in kind by US Department of Veterans Affairs). YVS received funding from NIDDK (grant number R01-DK125187) and Emory (grant number CFAR P30 AI050409).