Abstract
Objective(s):
To quantify the rate of change in epigenetic age compared to chronological age over time in youth with perinatally-acquired HIV (YPHIV) and youth who are perinatally HIV-exposed uninfected (YPHEU).
Design:
Longitudinal study of 32 YPHIV and 8 YPHEU with blood samples collected at two time points ≥3 years apart.
Methods:
DNA methylation was measured using the Illumina MethylationEPIC array and epigenetic age was calculated using the Horvath method. Linear mixed effects models were fit to estimate the average change in epigenetic age for a one year change in chronological age separately for YPHIV and YPHEU.
Results:
Median age was 10.9 and 16.8 years at time 1 and 2, respectively. Groups were balanced by sex (51% male) and race (67% Black). Epigenetic age increased by 1.23 years (95%CI: 1.03,1.43) for YPHIV and 0.95 years (95%CI: 0.74,1.17) for YPHEU per year increase in chronological age. Among YPHIV, in a model with chronological age, a higher area under the curve (AUC) VL was associated with an increase in epigenetic age over time [2.19 years per log10 copies/mL, (95%CI: 0.65,3.74)], whereas a higher time-averaged AUC CD4+ T-cell count was associated with a decrease in epigenetic age over time [−0.34 years per 100 cells/mm3, (95%CI: −0.63,−0.06)] in YPHIV.
Conclusions:
We observed an increase in the rate of epigenetic aging over time in YPHIV, but not in YPHEU. In YPHIV, higher VL and lower CD4+ T-cell count were associated with accelerated epigenetic aging, emphasizing the importance of early and sustained suppressive treatment for YPHIV, who will receive lifelong ART.
Keywords: epigenetics, epigenetic aging, biological aging, perinatal HIV, HIV exposure
Introduction
Due to the efficacy of combination antiretroviral therapy (ART), HIV infection is now a chronic condition. Yet, health risks remain for people living with HIV, including many age-related comorbidities, such as cardiovascular disease, renal disease, osteoporosis, and dementia.1 The onset of these conditions occurs earlier in the life course than typically observed in uninfected individuals, suggesting accelerated aging is occurring in people living with HIV.2,3
Recently, research has begun to characterize molecular profiles that reflect this accelerated aging, which can be inferred when the estimated biological age exceeds chronological age.4 In particular, patterns of DNA methylation associated with age across a variety of tissues5-8 are proposed as a marker of biological age in an “epigenetic clock”.5 Accelerated epigenetic age has been linked to aging-related outcomes, including all-cause mortality,9-11 frailty,12 cognitive decline and physical fitness in the elderly.13
Cross-sectional studies have reported epigenetic age acceleration in adults living with HIV on ART, finding a higher difference in the gap between chronological age and epigenetic age among those with HIV compared to adults without HIV at a single timepoint.14-16 Two recent studies have examined this question amongst youth with perinatally-acquired HIV (YPHIV) in South Africa, finding disparate results. In a cross-sectional study comparing YPHIV to HIV-unexposed uninfected age-matched controls 9-12 years of age, Horvath et al. found significant epigenetic age acceleration in YPHIV compared to typically developing controls.17 In a second cross sectional study of YPHIV, youth perinatally HIV-exposed uninfected (YPHEU), and HIV-unexposed uninfected children aged 4-9, Shiau et al. did not detect differences in epigenetic age acceleration between the three groups calculated using the Horvath method.18
Discrepancies in the findings of these studies may be due to the method(s) used to calculate epigenetic age, distinctions in the comparison groups chosen, different age ranges of the participants, and different covariates considered. In addition, both of these studies focus on cross-sectional, single time point measures of discrepancies in the gap between chronological age and epigenetic age, and in general, longitudinal evidence on acceleration is limited. Here, we quantified the rate of change in epigenetic age compared to chronological age over time in YPHIV and YPHEU, and among YPHIV, examined associations with time-averaged area under the curve (AUC) HIV RNA viral load (VL) and CD4+ T-cell count.
Methods
Study Population
Participants for this study were selected from the Adolescent Master Protocol (AMP) of the Pediatric HIV/AIDS Cohort Study (PHACS) network, which enrolled YPHIV and YPHEU from March 2007 through November 2009 at clinical sites in the United States and Puerto Rico, with active follow up.19 Inclusion criteria included youth with peripheral blood mononuclear cell (PBMC) samples collected at two time points ≥3 years apart, and all YPHIV who had HIV RNA VL measured within 1 year of birth and within 1 year of each of the sample time points. Participants were selected to have an equal distribution of male and female participants in each group and participants with the largest time difference between the first and last sample were prioritized for selection. Among a total population of 451 YPHIV and 227 YPHEU enrolled in AMP, 32 YPHIV and 8 YPHEU that met inclusion criteria were selected. One YPHEU was excluded from the analyses due to a mislabeling of one of the specimens.
Participating sites and the Harvard T.H. Chan School of Public Health obtained Institutional Review Board approvals. Written informed consent was obtained from the parent or legal guardian and assent was obtained from participants according to local Institutional Review Board (IRB) guidelines.
Demographic, Clinical, and Laboratory Measurements
Demographic, clinical, and laboratory data were collected through self-report and medical chart abstraction, including information on history of ART, HIV RNA VL, and CD4+ T-cell lymphocyte measurements.19 Time-averaged area under the curve (AUC) CD4 T-cell count and HIV RNA VL were calculated using the trapezoidal rule from the first VL measurement to each of the sample time points.20,21
DNA Methylation Assessment and Data Pre-Processing
PBMC samples stored at −80°C were provided to the Integrated Genomics Core Resource at Emory University, where they underwent DNA isolation using the Omega M6399 extraction kit (Omega BIO-TEK, Norcross, GA) on a KingFisher Flex instrument (Thermo Fisher Scientific, Waltham, MA). The DNA was quantified fluorescently using the Quant-iT PicoGreen system (Thermo Fisher Scientific) on a Tecan Infinite M200 Pro plate reader (Tecan Systems Inc., San Jose, CA). DNA quality was assessed by agarose gel. The resulting DNA was subsequently bisulfite modified using the Zymo EZ-96 DNA Methylation kit (Zymo Inc, Irvine, CA). Modified DNA was then profiled for genome-wide DNA methylation using the Illumina MethylationEPIC (850K) BeadArray following the manufacturer’s protocol (Illumina Inc., San Diego, CA).
DNA methylation data were processed via a standardized pipeline implemented by the minfi Bioconductor package using R statistical software, version 3.4.4 (R Foundation for Statistical Computing, Vienna, Austria).22,23 Specifically, we performed quality control and filtering procedures on the data. We excluded 3070 CpG probes that had poor detection p-values (p-value > 0.01) in any sample. No samples had poor detection p-values in more than 5% of CpG probes. The final analytic sample included 39 samples and 482,694 probes. None of the excluded probes were part of the CpG probes needed for the calculation of epigenetic age.
The data were normalized and background correction of raw signals was implemented.22,23 DNA methylation data were exported from minfi as β-values (proportion of methylated alleles at each CpG site) and standardized across probe types with beta mixture quantile (BMIQ) normalization.24 Normalized β-values were adjusted for potential batch effects using the combat function in the R package, sva.25,26 Batch corrected β-values were logit transformed (M-values) to better approximate a normal distribution.27 DNA methylation-estimated cell type percentages (B-cells, CD4 T-cells, CD8 T-cells, natural killer cells, granulocytes, monocytes) were calculated using the Houseman method.28
Calculation of Epigenetic Age
Epigenetic age was calculated using the Horvath method (353 CpGs) via the online calculator.29 As recommended in the calculator instructions, raw CpG data were normalized with the noob function in the minfi Bioconductor package and filtered to only the CpGs required by the calculator. The normalized and filtered CpGs were uploaded along with a sample manifest containing sample IDs, chronological age at collection of each sample, sex, and tissue.
Statistical Analysis
Statistical analyses were all performed using R statistical software, version 3.4.4 (R Foundation for Statistical Computing, Vienna, Austria) and SAS, version 9.4 (SAS Institute, Cary, North Carolina, USA). First, descriptive statistics were used to summarize demographic and clinical characteristics of the subset of YPHIV and YPHEU selected for this analysis at the two time points. Given the modest sample size with repeated measures, we evaluated the 32 YPHIV and 7 YPHEU separately or combined, rather than attempting to statistically compare the rate of change in epigenetic age between these two groups.
At each of the two timepoints, an epigenetic age acceleration residual was calculated as the difference between epigenetic age and chronological age. Box and whisker plots were used to graphically depict the epigenetic age acceleration residual at timepoint 1 and timepoint 2 separately for YPHIV and YPHEU.
Next, scatter plots were used to graphically depict epigenetic age by chronological age for YPHIV and YPHEU. To examine longitudinal changes in epigenetic age compared to chronological age, linear mixed effects models were fit to estimate the average change in epigenetic age for a one year change in chronological age separately for YPHIV and YPHEU. The linear mixed effects model included a random intercept to account for participant level variation. Given the small sample size, a limited number of additional covariates (HIV status, sex, geographic region, race, and estimated cell type percentages) were investigated as additional predictors of epigenetic age in separate linear mixed effects models including chronological age.
Finally, among YPHIV only, ART regimen, time-averaged AUC CD4 T-cell count and HIV RNA VL were evaluated as additional predictors of epigenetic age in separate linear mixed effects models and together in a single model including chronological age. Statistical significance was defined using two-sided P-value less than 0.05.
Results
Characteristics of the 32 YPHIV and 7 YPHEU included in the analysis are displayed in Table 1. Median age was 10.9 (range 7.0 – 16.9) and 16.8 (range 14.9 – 20.8) years at timepoints 1 and 2, respectively. At timepoint 1, median age was higher in the YPHIV group than the YPHEU group (11.4 vs. 8.1 years). Median duration of time between timepoint 1 and timepoint 2 was 6.0 years (interquartile range [IQR]: 4.0-7.7). Groups were balanced by sex (51% male) and race (67% Black). Approximately half (46%) of participants were from the Midwest, 23% from the Northeast, 23% from the West, and 8% from the South. A third of participants (32%) had an annual household income below $20,000 per year, although the YPHEU were more often in the low income group than the YPHIV (67% as compared to 25% with income < $20,000). YPHIV started combination ART at a median of 0.3 years of age (IQR: 0.2 – 1.2 years). Median time-averaged AUC CD4 T-cell count was 1,228 and 1,140 cells/mm3 and time-averaged AUC HIV RNA VL was 2.5 and 2.3 log10 copies/ml at timepoints 1 and 2, respectively. At timepoint 1 the majority of YPHIV (78%) were on a protease inhibitor (PI)- containing ART regimen and 2 children (6.3%) were not on ART. At timepoint 2, 25% were on an integrase strand transfer inhibitor (INSTI)-containing regimen, 22% on a PI-containing regimen, and 31% on a non-nucleoside reverse transcriptase inhibitor (NNRTI)- containing regimen, and 4 (12.5%) were not on ART.
Table 1:
Characteristics of 32 youth with perinatally-acquired HIV (YPHIV) and 7 youth perinatally HIV-exposed uninfected (YPHEU) by group
| Characteristic | YPHIV (N=32) | YPHEU (N=7) | Total (N=39) | |
|---|---|---|---|---|
| Chronological age at Timepoint 1 (years) | Min, Max | 7.0, 16.9 | 7.2, 9.5 | 7.0, 16.9 |
| Median (Q1, Q3) | 11.4 (9.9, 12.8) | 8.1 (7.4, 9.0) | 10.9 (8.7, 12.4) | |
| Chronological age at Timepoint 2 (years) | Min, Max | 14.9, 20.8 | 15.0, 17.6 | 14.9, 20.8 |
| Median (Q1, Q3) | 17.0 (16.2, 17.6) | 16.0 (15.1, 17.0) | 16.8 (16.0, 17.5) | |
| Time between Timepoint 1 and 2 (years) | Min, Max | 3.2, 8.6 | 6.9, 8.1 | 3.2, 8.6 |
| Median (Q1, Q3) | 5.2 (3.9, 7.1) | 8.0 (7.6, 8.1) | 6.0 (4.0, 7.7) | |
| Year of Birth | Min, Max | 1992, 2001 | 1999, 2002 | 1992, 2002 |
| Median (Q1, Q3) | 1991 (1997, 2000) | 2001 (2000, 2002) | 1999 (1997, 2000) | |
| Sex | Male | 16 (50.0%) | 4 (57.1%) | 20 (51.3%) |
| Female | 16 (50.0%) | 3 (42.9%) | 19 (48.7%) | |
| Race | White | 10 (31.3%) | 2 (28.6%) | 12 (30.8%) |
| Black | 21 (65.6%) | 5 (71.4%) | 26 (66.7%) | |
| Other | 1 (3.1%) | 0 (0.0%) | 1 (2.6%) | |
| Site Region | Northeast | 9 (28.1%) | 0 (0.0%) | 9 (23.1%) |
| South | 3 (9.4%) | 0 (0.0%) | 3 (7.7%) | |
| West | 6 (18.8%) | 3 (42.9%) | 9 (23.1%) | |
| Midwest | 14 (43.8%) | 4 (57.1%) | 18 (46.2%) | |
| Annual Household Income | ≤$20,000 | 8 (25.0%) | 4 (66.7%) | 12 (31.6%) |
| $21,000 – $50,000 | 16 (50.0%) | 0 (0.0%) | 16 (42.1%) | |
| >$51,000 | 8 (25.0%) | 2 (33.3%) | 10 (26.3%) | |
| Unknown | 0 | 1 | 1 | |
| Age at First ARV (years) | Min, Max | 0.2, 2.0 | ||
| Median (Q1, Q3) | 0.2 (0.2, 0.5) | |||
| Age at First HAART (years) | Min, Max | 0.2, 9.5 | ||
| Median (Q1, Q3) | 0.3 (0.2, 1.2) | |||
| Time-averaged AUC CD4 T-cell count at Timepoint 1 (cells/mm3) | Min, Max | 596, 2,544 | ||
| Median (Q1, Q3) | 1,228 (1,000, 1,530) | |||
| Time-averaged AUC CD4 T-cell count at Timepoint 2 (cells/mm3) | Min, Max | 480, 2,077 | ||
| Median (Q1, Q3) | 1,140 (879, 1,359) | |||
| Time-averaged AUC HIV RNA viral load at Timepoint 1 (log10 copies/ml) | Min, Max | 1.5, 4.2 | ||
| Median (Q1, Q3) | 2.5 (2.1, 3.2) | |||
| Time-averaged AUC HIV RNA viral load at Timepoint 2 (log10 copies/ml) | Min, Max | 1.5, 4.3 | ||
| Median (Q1, Q3) | 2.3 (2.0, 2.9) | |||
| ART Regimen at Timepoint 1 | INSTI-based ART | 1 (3.1%) | ||
| PI-based ART | 25 (78.1%) | |||
| NNRTI-based ART | 3 (9.4%) | |||
| Other ARV | 1 (3.1%) | |||
| No ARV | 2 (6.3%) | |||
| ART Regimen at Timepoint 2 | INSTI-based ART | 8 (25.0%) | ||
| PI-based ART | 7 (21.9%) | |||
| NNRTI-based ART | 10 (31.3%) | |||
| Other ARV | 3 (9.4%) | |||
| No ARV | 4 (12.5%) |
Figure 1a and Figure 1b shows the distribution of the epigenetic age acceleration residual (difference between epigenetic age and chronological age) at timepoints 1 and 2 by group. The median epigenetic age acceleration residual was −0.1 year (IQR: −1.8, 1.9) at timepoint 1 and 0.1 year (IQR: −3.7, 4.1) at timepoint 2 for YPHIV. The median epigenetic age acceleration residual was −0.6 year (IQR: −3.7, 0.7) at timepoint 1 and −0.9 year (IQR: −7.5, −0.7) at timepoint 2 for YPHEU.
Figure 1:
Violin plots of epigenetic age acceleration residual at timepoint 1 (a) and timepoint 2 (b) for youth with perinatally-acquired HIV (YPHIV) and HIV-exposed uninfected youth (YPHEU). Scatterplot of epigenetic age by chronological age at two timepoints for YPHIV and YPHEU (c)
As shown in Figure 1c, epigenetic age increased for all YPHIV and YPHEU from time point 1 to time point 2.
In linear mixed effects models, epigenetic age increased by 1.23 years (95% confidence interval [CI]: 1.03, 1.43) for YPHIV and 0.95 years (95% CI: 0.74, 1.17) for YPHEU per year increase in chronological age (Table 2). Except geographical region, HIV status, sex, and race were not associated with an increase or decrease in epigenetic age in separate models including chronological age (Table 3 models 1-4).
Table 2:
Change in epigenetic age for a year increase in age among 32 youth with perinatally-acquired HIV (YPHIV) and 7 youth perinatally HIV-exposed uninfected (YPHEU) in linear mixed effects models
| Population | Covariate | Estimate | (95% CI) | P-value |
|---|---|---|---|---|
| YPHIV | Chronological age (years) | 1.23 | (1.03, 1.43) | 0.027 |
| YPHEU | Chronological age (years) | 0.95 | (0.74, 1.17) | 0.606 |
Table 3:
Predictors of epigenetic age in 32 youth with perinatally-acquired HIV (YPHIV) and 7 youth perinatally HIV-exposed uninfected (YPHEU) in linear mixed effects models including chronological age
| Covariate | Level | Estimate | (95% CI) | P-value | |
|---|---|---|---|---|---|
| Model 1 | Chronological age (years) | 1.15 | (0.99, 1.31) | 0.067 | |
| HIV status | YPHIV | 2.47 | (−0.60, 5.55) | 0.11 | |
| YPHEU | Ref. | ||||
| Model 2 | Chronological age (years) | - | 1.17 | (1.01, 1.33) | 0.038 |
| Sex | Female | −1.54 | (−3.92, 0.84) | 0.20 | |
| Male | Ref. | ||||
| Model 3 | Chronological age (years) | - | 1.15 | (0.99, 1.31) | 0.068 |
| Site Region | Midwest | 2.38 | (−0.60, 5.36) | 0.11 | |
| Northeast | 3.76 | (0.32, 7.20) | 0.033 | ||
| South | 2.98 | (−1.87, 7.83) | 0.22 | ||
| West | Ref. | ||||
| Model 4 | Chronological age (years) | - | 1.16 | (1.00, 1.32) | 0.051 |
| Race/Ethnicity | Black | −0.16 | (−2.82, 2.49) | 0.90 | |
| Other | −3.42 | (−11.35, 4.51) | 0.39 | ||
| White | Ref. | ||||
| Model 5 | Chronological age (years) | - | 0.91 | (0.71, 1.10) | 0.32 |
| Percentage of B Cells1 | - | −0.15 | (−0.32, 0.02) | 0.084 | |
| Percentage of CD4 T cells1 | - | −0.13 | (−0.24, −0.03) | 0.014 | |
| Percentage of CD8 T cells1 | - | −0.0002 | (−0.0003, −0.00004) | 0.013 | |
| Percentage of Natural Killer cells1 | - | −0.09 | (−0.24, 0.06) | 0.23 | |
| Percentage of Granulocytes1 | - | −0.005 | (−0.04, 0.03) | 0.76 | |
| Percentage of Monocytes1 | - | −0.03 | (−0.17, 0.12) | 0.72 |
DNA methylation estimated cell type percentages. Estimates are given as a 1% increase in the given cell type.
In a model with chronological age and DNA methylation-estimated cell type percentages (B-cells, CD4 T-cells, CD8 T-cells, natural killer cells, granulocytes, monocytes), a 1% increase in DNA methylation-estimated percentages of CD4 T-cells and CD8 T-cells were associated with a lower predicted epigenetic age over time: −0.13 years, (95% CI: −0.24, −0.03) and −0.0002 years (95% CI: −0.0003, −0.00004), respectively (Table 3 model 5). When limited to YPHIV only, DNA methylation-estimated percentages of CD4 T-cells were associated with lower predicted epigenetic age over time [−0.16 years, (95% CI: −0.29, −0.04)].
Among YPHIV, the chronological age-adjusted difference in epigenetic age comparing youth not on ART to those on PI-containing ART was 3.32 years (95% CI: −0.11, 6.75) (Table 4 model 1). We explored whether percentage of follow-up time not on ART from birth was associated with a change in epigenetic age, and observed a 0.11 year increase in epigenetic age (95% CI: −0.03, 0.26) for each 1% of time since birth that an individual was not on any ART.
Table 4:
Predictors of epigenetic age in 32 youth with perinatally-acquired HIV (YPHIV) in linear mixed effects models including chronological age
| Model | Covariate | Level | Estimate | (95% CI) | P-value |
|---|---|---|---|---|---|
| Model 1 | Chronological age (years) | - | 1.15 | (0.85, 1.44) | 0.31 |
| ART Regimen | INSTI-based ART | 0.18 | (−2.69, 3.05) | 0.90 | |
| PI-based ART | Ref. | ||||
| NNRTI-based ART | 0.66 | (−1.96, 3.28) | 0.61 | ||
| Other ARV | 1.29 | (−2.38, 4.95) | 0.48 | ||
| No ARV | 3.32 | (−0.11, 6.75) | 0.058 | ||
| Model 2 | Chronological age (years) | - | 1.27 | (1.08, 1.45) | 0.007 |
| Time-averaged AUC HIV RNA viral load (log 10 copies/ml) | - | 2.72 | (1.09, 4.34) | 0.002 | |
| Model 3 | Chronological age (years) | - | 1.09 | (0.87, 1.31) | 0.39 |
| Time-averaged AUC CD4 T-cell count (100 cells/mm3) | - | −0.44 | (−0.73, −0.14) | 0.005 | |
| Model 4 | Chronological age (years) | - | 1.15 | (0.94, 1.36) | 0.17 |
| Time-averaged AUC CD4 T-cell count (100 cells/mm3) | - | −0.34 | (−0.63, −0.06) | 0.021 | |
| Time-averaged AUC HIV RNA viral load (log 10 copies/ml) | - | 2.19 | (0.65, 3.74) | 0.007 |
In a model with chronological age, higher time-averaged AUC HIV-RNA VL was associated with an increase in epigenetic age over time [2.72 years per log10 copies/mL, (95% CI: 1.09, 4.34) model 2], while a higher time-averaged AUC CD4 T-cell count was associated with a decrease in epigenetic age over time [−0.44 years per 100 cells/mm3, (95% CI: −0.73, −0.14), model 3]. In a multivariable model with chronological age, AUC HIV-RNA VL, and AUC CD4, a higher AUC HIV-RNA VL was associated with an increase in epigenetic age over time [2.19 years per log10 copies/mL, (95% CI: 0.65, 3.74)], whereas a higher time-averaged AUC CD4 T-cell count was associated with a decrease in epigenetic age over time [−0.34 years per 100 cells/mm3, (95% CI: −0.63, −0.06)] in YPHIV (Table 4 model 4).
Discussion
To our knowledge, this is the first study to examine longitudinal changes in epigenetic age compared to chronological age over time in YPHIV, making use of repeated samples over time. We observed accelerated epigenetic aging over time in ART-treated YPHIV, but not in YPHEU. However, the sample size was small for YPHEU (N=7) and findings should be interpreted with caution. In addition, few YPHIV were not receiving ART, which limited comparisons with other regimens. This longitudinal observation is consistent with findings from cross-sectional studies of individuals with HIV, which have reported a higher increase in the gap between epigenetic age and chronological age among ART-treated individuals with HIV and HIV-uninfected controls at cross-sectional time points,14-17 as well as a study of children with perinatally-acquired HIV ages 9-12 in South Africa.17
Our findings add to the existing research suggesting accelerated biological aging in individuals living with HIV.14 Older epigenetic aging relative to chronological age has been found to be associated with a wide spectrum of aging outcomes, most consistently mortality.30 Our finding of an increased rate of epigenetic age over time in YPHIV provides preliminary evidence of accelerated aging at a young age and may have future consequences. Age-related comorbidities such as cardiovascular disease, renal disease, osteoporosis, and dementia contribute to a large burden of disease in people living with HIV. Some studies have reported findings that older epigenetic age relative to chronological age may be associated with poorer measures of cognitive functioning in YPHIV and HIV-associated neurocognitive disorders (HAND) in adults with HIV.17,31 Our small sample size limited our ability to explore whether epigenetic age acceleration in YPHIV may be associated with early markers of other age-related comorbidities, such as dyslipidemia, impaired renal function, or low bone mineral density, but future larger longitudinal studies to explore these associations, including studies utilizing banked specimens from the PHACS, are warranted.
In the context of HIV, a high CD4 count and a low/undetectable HIV RNA VL are desirable. An undetectable VL is indicative of successful ART, meaning that the individual is adherent to their HIV medications. We found that a lower cumulative CD4 T-cell count and a higher cumulative HIV RNA VL were associated with accelerated epigenetic aging. We also observed that those not on ART, albeit a small number (N=2 at timepoint 1 and N=4 at timepoint 2; 5 unique individuals), had a higher epigenetic age compared to those on PI-based regimens. Consistently, those with a higher percentage of time not on ART tended to have higher epigenetic ages. Taken together, this evidence suggests that those less adherent to ART over time and with subsequently higher HIV RNA VL may have accelerated epigenetic aging. A positive correlation between HIV RNA VL and higher epigenetic age relative to chronological age has been reported in some but not all studies of adults with HIV.16,31 In the South African study of YPHIV ages 9-12 years, no significant association was observed between HIV RNA VL and higher epigenetic age relative to chronological age, but those on a second or third-line regimen were more likely to have higher epigenetic age relative to chronological age.17 In that same study, lower CD4 T-cell count was associated with increased epigenetic age relative to chronological age, after adjustment for HIV RNA VL and other factors.17 In addition, evidence of alterations to immune aging-related markers (e.g. circulating hematopoietic progenitors and lymphocyte population phenotypes) has been reported in young adults ages 18-25 living with HIV since childhood who presented with high viral loads.32,33
Since our methylation data were generated from PBMCs, it is possible that our findings may be related to changes in cell composition. Although we only had unfractionated blood and could not study cell-type specific methylation, we did use the method by Houseman et al. to evaluate cell type percentages, finding that percentage of CD4 T-cells in particular had a large influence on the longitudinal estimates of epigenetic age in YPHIV.28 Other characteristics (e.g. sex, race) had minimal impact on the longitudinal epigenetic aging estimates, but future studies are needed with a sample sufficiently large to stratify on sex to evaluate whether the relationship is different for males vs. females. Other characteristics and early life factors, such as measures of adversity or stressful life events have been linked to accelerated epigenetic aging.34-36 Incorporating these measures in future larger studies would be useful to determine potential opportunities for intervention.
In conclusion, we observed descriptive differences in the rate of epigenetic aging in YPHIV and YPHEU. A major strength of this study was its longitudinal design with repeated measures of epigenetic age. These analyses will directly inform effect sizes and potential confounders to consider in a larger study. Further, these findings emphasize the importance of early and sustained suppressive treatment for YPHIV, who will age on lifelong ART. In addition, future work should examine the dynamic nature of epigenetic age, through examinations of differences in viral load over time, or how interventions leading to improved adherence impact epigenetic age.
Acknowledgements:
We thank the participants and families for their participation in PHACS, and the individuals and institutions involved in the conduct of PHACS. The study was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development with co-funding from the National Institute on Drug Abuse, the National Institute of Allergy and Infectious Diseases, the National Institute of Mental Health, the National Institute of Neurological Disorders and Stroke, the National Institute on Deafness and Other Communication Disorders, the National Institute of Dental and Craniofacial Research, the National Cancer Institute, the National Institute on Alcohol Abuse and Alcoholism, the Office of AIDS Research, and the National Heart, Lung, and Blood Institute through cooperative agreements with the Harvard T.H. Chan School of Public Health (HD052102) (Principal Investigator: George R Seage III; Program Director: Liz Salomon) and the Tulane University School of Medicine (HD052104) (Principal Investigator: Russell Van Dyke; Co-Principal Investigator: Ellen Chadwick; Project Director: Patrick Davis). Data management services were provided by Frontier Science and Technology Research Foundation (PI: Suzanne Siminski), and regulatory services and logistical support were provided by Westat, Inc (PI: Julie Davidson).
Appendix
The following institutions, clinical site investigators and staff participated in conducting PHACS AMP and AMP Up in 2018, in alphabetical order: Ann & Robert H. Lurie Children’s Hospital of Chicago: Ellen Chadwick, Margaret Ann Sanders, Kathleen Malee, Yoonsun Pyun; Baylor College of Medicine: William Shearer, Mary Paul, Chivon McMullen-Jackson, Mandi Speer, Lynnette Harris; Bronx Lebanon Hospital Center: Murli Purswani, Mahboobullah Mirza Baig, Alma Villegas; Children's Diagnostic & Treatment Center: Lisa Gaye-Robinson, Sandra Navarro, Patricia Garvie; Boston Children’s Hospital: Sandra K. Burchett, Michelle E. Anderson, Adam R. Cassidy; Jacobi Medical Center: Andrew Wiznia, Marlene Burey, Ray Shaw, Raphaelle Auguste; Rutgers - New Jersey Medical School: Arry Dieudonne, Linda Bettica, Juliette Johnson, Karen Surowiec; St. Christopher’s Hospital for Children: Janet S. Chen, Maria Garcia Bulkley, Taesha White, Mitzie Grant; St. Jude Children's Research Hospital: Katherine Knapp, Kim Allison, Megan Wilkins, Jamie Russell-Bell; San Juan Hospital/Department of Pediatrics: Midnela Acevedo-Flores, Heida Rios, Vivian Olivera; Tulane University School of Medicine: Margarita Silio, Medea Gabriel, Patricia Sirois; University of California, San Diego: Stephen A. Spector, Megan Loughran, Veronica Figueroa, Sharon Nichols; University of Colorado Denver Health Sciences Center: Elizabeth McFarland, Carrie Chambers, Emily Barr, Mary Glidden; University of Miami: Gwendolyn Scott, Grace Alvarez, Juan Caffroni, Anai Cuadra
Note: The conclusions and opinions expressed in this article are those of the authors and do not necessarily reflect those of the National Institutes of Health or U.S. Department of Health and Human Services.
Footnotes
Conflicts of Interest: None
References
- 1.Shah ASV, Stelzle D, Lee KK, Beck EJ, Alam S, Clifford S, Longenecker CT, Strachan F, Bagchi S, Whiteley W, Rajagopalan S, Kottilil S, Nair H, Newby DE, McAllister DA, Mills NL. Global Burden of Atherosclerotic Cardiovascular Disease in People Living With HIV. Circulation. 2018. September 11;138(11):1100–1112. PMCID: PMC6221183 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.High KP, Brennan-Ing M, Clifford DB, Cohen MH, Currier J, Deeks SG, Deren S, Effros RB, Gebo K, Goronzy JJ, Justice AC, Landay A, Levin J, Miotti PG, Munk RJ, Nass H, Rinaldo CR, Shlipak MG, Tracy R, Valcour V, Vance DE, Walston JD, Volberding P, OAR Working Group on HIV and Aging. HIV and aging: state of knowledge and areas of critical need for research. A report to the NIH Office of AIDS Research by the HIV and Aging Working Group. J Acquir Immune Defic Syndr. 2012. July 1;60 Suppl 1:S1–18. PMCID: PMC3413877 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Sundermann EE, Erlandson KM, Pope CN, Rubtsova A, Montoya J, Moore AA, Marzolini C, O’Brien KK, Pahwa S, Payne BAI, Rubin LH, Walmsley S, Haughey NJ, Montano M, Karris MY, Margolick JB, Moore DJ. Current Challenges and Solutions in Research and Clinical Care of Older Persons Living with HIV: Findings Presented at the 9th International Workshop on HIV and Aging. AIDS Res Hum Retroviruses. 2019. September 10; PMCID: PMC6862962 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hastings WJ, Shalev I, Belsky DW. Comparability of biological aging measures in the National Health and Nutrition Examination Study, 1999-2002. Psychoneuroendocrinology. 2019;106:171–178. PMCID: PMC6599717 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Horvath S DNA methylation age of human tissues and cell types. Genome Biol. 2013;14(10):R115. PMCID: PMC4015143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Christensen BC, Houseman EA, Marsit CJ, Zheng S, Wrensch MR, Wiemels JL, Nelson HH, Karagas MR, Padbury JF, Bueno R, Sugarbaker DJ, Yeh R-F, Wiencke JK, Kelsey KT. Aging and environmental exposures alter tissue-specific DNA methylation dependent upon CpG island context. PLoS Genet. 2009. August;5(8):e1000602. PMCID: PMC2718614 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bollati V, Schwartz J, Wright R, Litonjua A, Tarantini L, Suh H, Sparrow D, Vokonas P, Baccarelli A. Decline in genomic DNA methylation through aging in a cohort of elderly subjects. Mech Ageing Dev. 2009. April;130(4):234–239. PMCID: PMC2956267 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Rakyan VK, Down TA, Maslau S, Andrew T, Yang T-P, Beyan H, Whittaker P, McCann OT, Finer S, Valdes AM, Leslie RD, Deloukas P, Spector TD. Human aging-associated DNA hypermethylation occurs preferentially at bivalent chromatin domains. Genome Res. 2010. April;20(4):434–439. PMCID: PMC2847746 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Christiansen L, Lenart A, Tan Q, Vaupel JW, Aviv A, McGue M, Christensen K. DNA methylation age is associated with mortality in a longitudinal Danish twin study. Aging Cell. 2016. February;15(1):149–154. PMCID: PMC4717264 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Marioni RE, Shah S, McRae AF, Chen BH, Colicino E, Harris SE, Gibson J, Henders AK, Redmond P, Cox SR, Pattie A, Corley J, Murphy L, Martin NG, Montgomery GW, Feinberg AP, Fallin MD, Multhaup ML, Jaffe AE, Joehanes R, Schwartz J, Just AC, Lunetta KL, Murabito JM, Starr JM, Horvath S, Baccarelli AA, Levy D, Visscher PM, Wray NR, Deary IJ. DNA methylation age of blood predicts all-cause mortality in later life. Genome Biol. 2015. January 30;16:25. PMCID: PMC4350614 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Perna L, Zhang Y, Mons U, Holleczek B, Saum K-U, Brenner H. Epigenetic age acceleration predicts cancer, cardiovascular, and all-cause mortality in a German case cohort. Clin Epigenetics. 2016;8:64. PMCID: PMC4891876 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Breitling LP, Saum K-U, Perna L, Schöttker B, Holleczek B, Brenner H. Frailty is associated with the epigenetic clock but not with telomere length in a German cohort. Clin Epigenetics. 2016;8:21. PMCID: PMC4768341 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Marioni RE, Shah S, McRae AF, Ritchie SJ, Muniz-Terrera G, Harris SE, Gibson J, Redmond P, Cox SR, Pattie A, Corley J, Taylor A, Murphy L, Starr JM, Horvath S, Visscher PM, Wray NR, Deary IJ. The epigenetic clock is correlated with physical and cognitive fitness in the Lothian Birth Cohort 1936. Int J Epidemiol. 2015. August;44(4):1388–1396. PMCID: PMC4588858 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Gross AM, Jaeger PA, Kreisberg JF, Licon K, Jepsen KL, Khosroheidari M, Morsey BM, Swindells S, Shen H, Ng CT, Flagg K, Chen D, Zhang K, Fox HS, Ideker T. Methylome-wide Analysis of Chronic HIV Infection Reveals Five-Year Increase in Biological Age and Epigenetic Targeting of HLA. Mol Cell. 2016. 21;62(2):157–168. PMCID: PMC4995115 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Nelson KN, Hui Q, Rimland D, Xu K, Freiberg MS, Justice AC, Marconi VC, Sun YV. Identification of HIV infection-related DNA methylation sites and advanced epigenetic aging in HIV-positive, treatment-naive U.S. veterans. AIDS. 2017. 20;31(4):571–575. PMCID: PMC5263111 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Horvath S, Levine AJ. HIV-1 Infection Accelerates Age According to the Epigenetic Clock. J Infect Dis. 2015. November 15;212(10):1563–1573. PMCID: PMC4621253 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Horvath S, Stein D, Phillips N, Heany S, Kobor M, Lin D, Myer L, Zar H, Levine A, Hoare J. Perinatally acquired HIV infection accelerates epigenetic aging in South African adolescents. Aids. 2018. July 1;32(11):1465–1474. PMID: 29746298 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Shiau S, Strehlau R, Shen J, Violari A, Patel F, Liberty A, Foca M, Wang S, Terry MB, Yin MT, Coovadia A, Abrams EJ, Arpadi SM, Kuhn L. Biomarkers of Aging in HIV-Infected Children on Suppressive Antiretroviral Therapy. J Acquir Immune Defic Syndr. 2018. August 15;78(5):549–556. PMCID: PMC6037570 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Van Dyke RB, Patel K, Siberry GK, Burchett SK, Spector SA, Chernoff MC, Read JS, Mofenson LM, Seage GR, Pediatric HIV/AIDS Cohort Study. Antiretroviral treatment of US children with perinatally acquired HIV infection: temporal changes in therapy between 1991 and 2009 and predictors of immunologic and virologic outcomes. J Acquir Immune Defic Syndr. 2011. June 1;57(2):165–173. PMCID: PMC3125481 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Lesosky M, Glass T, Rambau B, Hsiao N-Y, Abrams EJ, Myer L. Bias in the estimation of cumulative viremia in cohort studies of HIV-infected individuals. Ann Epidemiol. 2019;38:22–27. PMCID: PMC6812622 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Cole SR, Napravnik S, Mugavero MJ, Lau B, Eron JJ, Saag MS. Copy-years viremia as a measure of cumulative human immunodeficiency virus viral burden. Am J Epidemiol. 2010. January 15;171(2):198–205. PMCID: PMC2878100 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, Irizarry RA. Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics. 2014. May 15;30(10):1363–1369. PMCID: PMC4016708 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Fortin J-P, Triche TJ, Hansen KD. Preprocessing, normalization and integration of the Illumina HumanMethylationEPIC array with minfi. Bioinformatics. 2017. 15;33(4):558–560. PMCID: PMC5408810 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Teschendorff AE, Marabita F, Lechner M, Bartlett T, Tegner J, Gomez-Cabrero D, Beck S. A beta-mixture quantile normalization method for correcting probe design bias in Illumina Infinium 450 k DNA methylation data. Bioinformatics. 2013. January 15;29(2):189–196. PMCID: PMC3546795 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Johnson WE, Li C, Rabinovic A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics. 2007. January;8(1):118–127. PMID: 16632515 [DOI] [PubMed] [Google Scholar]
- 26.Leek JT, Johnson WE, Parker HS, Jaffe AE, Storey JD. The sva package for removing batch effects and other unwanted variation in high-throughput experiments. Bioinformatics. 2012. March 15;28(6):882–883. PMCID: PMC3307112 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Du P, Zhang X, Huang C-C, Jafari N, Kibbe WA, Hou L, Lin SM. Comparison of Beta-value and M-value methods for quantifying methylation levels by microarray analysis. BMC Bioinformatics. 2010. November 30;11:587. PMCID: PMC3012676 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Houseman EA, Accomando WP, Koestler DC, Christensen BC, Marsit CJ, Nelson HH, Wiencke JK, Kelsey KT. DNA methylation arrays as surrogate measures of cell mixture distribution. BMC Bioinformatics. 2012. May 8;13:86. PMCID: PMC3532182 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Horvath S DNA methylation age of human tissues and cell types. Genome Biol. 2013;14(10):R115. PMCID: PMC4015143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Jylhävä J, Pedersen NL, Hägg S. Biological Age Predictors. EBioMedicine. 2017. July;21:29–36. PMCID: PMC5514388 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Levine AJ, Quach A, Moore DJ, Achim CL, Soontornniyomkij V, Masliah E, Singer EJ, Gelman B, Nemanim N, Horvath S. Accelerated epigenetic aging in brain is associated with pre-mortem HIV-associated neurocognitive disorders. J Neurovirol. 2016;22(3):366–375. PMCID: PMC4900944 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Fastenackels S, Sauce D, Vigouroux C, Avettand-Fènoël V, Bastard J-P, Fellahi S, Nailler L, Arezes E, Rouzioux C, Warszawski J, Viard JP, Appay V, ANRS Co19 COVERTE Study Group. HIV-mediated immune aging in young adults infected perinatally or during childhood. AIDS. 2019. 01;33(11):1705–1710. PMID: 31149945 [DOI] [PubMed] [Google Scholar]
- 33.Williams DW, Elahi S. Profound immune consequences for young adults infected with HIV perinatally or during childhood: a cautionary tale regarding adherence to antiretroviral therapy. AIDS. 2019. 15;33(14):2251–2252. PMCID: PMC6839777 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Carter SE, Ong ML, Simons RL, Gibbons FX, Lei MK, Beach SRH. The effect of early discrimination on accelerated aging among African Americans. Health Psychol. 2019. November;38(11):1010–1013. PMCID: PMC6800763 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Palma-Gudiel H, Fañanás L, Horvath S, Zannas AS. Psychosocial stress and epigenetic aging. Int Rev Neurobiol. 2020;150:107–128. PMID: 32204828 [DOI] [PubMed] [Google Scholar]
- 36.Marini S, Davis KA, Soare TW, Zhu Y, Suderman MJ, Simpkin AJ, Smith ADAC, Wolf EJ, Relton CL, Dunn EC. Adversity exposure during sensitive periods predicts accelerated epigenetic aging in children. Psychoneuroendocrinology. 2020. March;113:104484. PMID: 31918390 [DOI] [PMC free article] [PubMed] [Google Scholar]