Plasma Proteomic Signature as a Predictor of Age Advancement in People Living With HIV

Abstract

Due to the increased burden of non-AIDS-related comorbidities in people living with HIV (PLHIV), identifying biomarkers and mechanisms underlying premature aging and the risk of developing age-related comorbidities is a priority. Evidence suggests that the plasma proteome is an accurate source for measuring biological age and predicting age-related clinical outcomes. To investigate whether PLHIV on antiretroviral therapy (ART) exhibit a premature aging phenotype, we profiled the plasma proteome of two independent cohorts of virally suppressed PLHIV (200HIV and 2000HIV) and one cohort of people without HIV (200FG) using O-link technology. Next, we built a biological age-prediction model and correlated age advancement (the deviation of the predicted age from the chronological age) with HIV-related factors, comorbidities, and cytokines secreted by immune cells. We identified a common signature of 77 proteins associated with chronological age across all cohorts, most of which were involved in inflammatory and senescence-related processes. PLHIV showed increased age advancement compared to people without HIV. In addition, age advancement in the 2000HIV cohort was positively associated with prior hepatitis C and cytomegalovirus (CMV) infections, non-AIDS-related comorbidities, ART duration, cumulative exposure to the protease inhibitor Ritonavir, as well as higher production of monocyte-derived proinflammatory cytokines and chemokines and lower secretion of T-cell derived cytokines. Our proteome-based predictive model is a promising approach for calculating the age advancement in PLHIV. This will potentially allow for further characterization of the pathophysiological mechanisms linked to accelerated aging and enable monitoring the effectiveness of novel therapies aimed at reducing age-related diseases in PLHIV.

Keywords: HIV; O‐link; non‐AIDS‐related comorbidities; premature aging; proteomics; senescence.

FIGURE 1

Proteomic age signature in people without HIV and PLHIV. (A) Volcano plot displaying the association of plasma proteins with chronological age in three independent cohorts: Left (200FG), middle: (200HIV), and right (2000HIV), respectively. (B) Venn diagram showing the comparison of the significant age‐associated proteins between the three cohorts and the signature of 77 proteins common to all cohorts. (C) Correlation of chronological age and predicted age (calculated using a lasso regression model) in the three cohorts evaluated. (D) Age advancement comparison (predicted age − chronological age) between the three cohorts. (E) Differences in age advancement across the three chronological age strata: Young (18–35 years old), middle age (36–60 years old), and older (60+). Statistical significance in (D) and (E) was estimated by Kruskal–Wallis and Wilcoxon unpaired test.

FIGURE 2

Forest plot displaying the associations between age advancement and demographics, HIV‐specific factors and comorbidities in PLHIV. For each of the associations between age advancement and clinical parameters, the standardized beta and the confidence interval derived from linear models were plotted. The significant association (FDR < 0.05) is shown with a black filled dot.

FIGURE 3

Forest plot of associations between age advancement and cumulative antiretroviral exposure in PLHIV. For each of the associations between age advancement and years of cumulative exposure to each ART medication, the standardized beta and the confidence interval derived from linear models were plotted. The model was corrected by age. The significant associations (FDR < 0.05) are shown with a black filled dot. 3TC, lamivudine; ABC, Abacavir; booster, booster drugs; COBI, Cobicistat; DTG, Dolutegravir; EFC, efavirenz; EVG, elvitegravir; FTC, Emtricitabine; INSTI, Integrase inhibitors; NNRTI, non‐nucleoside reverse transcriptase inhibitor; NVP, nevirapine; PI, protease inhibitors; RPV, rilpivirine; RTV, ritonavir; TAF, tenofovir alafenamide; TDF, tenofovir disproxil; ZDV, Zidovudine.

FIGURE 4

Heat map showing the association between (A) PBMC‐secreted SASP‐cytokines and (B) T cell cytokines with age advancement in PLHIV. (A) PBMCs from PLHIV of the 2000HIV cohort (n = 551) were stimulated with the TLR ligands Poly:IC, IMQ, LPS, the peptides HIV ENV pool and pp65 CMV, S. pneumoniae and rhIL‐1a for 24 h and cytokines (IL‐1β, IL‐6, IL‐8), and chemokines (MCP‐1, MIP‐1α) were measured by ELISA. (B) PBMCs from PLHIV of the 2000HIV cohort (n = 550) were stimulated for 7 days with C. albicans conidia, E. coli , M. tuberculosis , PHA, S. aureus , S. pneumoniae , and the cytokines (IFNγ, IL‐10, IL‐17, IL‐22, IL‐5) were measured by ELISA. Heatmaps represent the estimate from linear regression model adjusted by age and sex. Only significant associations (FDR < 0.05) are shown (*).