Association of Neighborhood Deprivation With Epigenetic Aging Using 4 Clock Metrics.

Importance: Neighborhood deprivation is associated with age-related disease, mortality, and reduced life expectancy. However, biological pathways underlying these associations are not well understood. Objective: To evaluate the association between neighborhood deprivation and epigenetic measures of age acceleration and genome-wide methylation. Design, Setting, and Participants: This cross-sectional study used data from the Sister Study, a prospective cohort study comprising 50 884 women living in the US and Puerto Rico aged 35 to 74 years at enrollment who had a sister with breast cancer but had not had breast cancer themselves. Cohort enrollment occurred between July 2003 and March 2009. Participants completed a computer-assisted telephone interview on demographic, socioeconomic, lifestyle, and residential factors and provided anthropometric measures and peripheral blood samples at a home examination. DNA methylation data obtained for 2630 non-Hispanic White women selected for a case-cohort study in 2014 were used in this cross-sectional analysis. DNA methylation was measured using the HumanMethylation450 BeadChips in whole blood samples collected at baseline. Data analysis for this study was performed from October 17, 2019, to August 27, 2020. Exposures: Each participants' primary address was linked to an established index of neighborhood deprivation. Main Outcomes and Measures: Epigenetic age was estimated using 4 epigenetic clocks (Horvath, Hannum, PhenoAge, and GrimAge). Age acceleration was determined using residuals from regressing chronologic age on each of the 4 epigenetic age metrics. Linear regression was used to estimate associations between neighborhood deprivation and epigenetic age acceleration as well as DNA methylation at individual cytosine-guanine sites across the genome. Results: Mean (SD) age of the 2630 participants was 56.9 (8.7) years. Those with the greatest (>75th percentile) vs least (≤25th percentile) neighborhood deprivation had higher epigenetic age acceleration estimated by Hannum (ß = 0.23; 95% CI, 0.01-0.45), PhenoAge (ß = 0.28; 95% CI, 0.06-.50), and GrimAge (ß = 0.37; 95% CI, 0.12-0.62). Increasing US quartiles of neighborhood deprivation exhibited a trend with Hannum, PhenoAge, and GrimAge. For example, GrimAge showed a significant dose-response (P test for trend <.001) as follows: level 2 vs level 1 (ß = 0.30; 95% CI, 0.17-0.42), level 3 vs level 1 (ß = 0.35; 95% CI, 0.19-0.50), and level 4 vs level 1 (ß = 0.37; 95% CI, 0.12-0.62). Neighborhood deprivation was found to be associated with 3 cytosine-phosphate-guanine sites, with 1 of these annotated to a known gene MAOB (P = 9.71 × 10-08). Conclusions and Relevance: The findings of this study suggest that residing in a neighborhood with a higher deprivation index appears to be reflected by methylation-based markers of aging.

Genetic ancestry, self-reported race and ethnicity in African Americans and European Americans in the PCaP cohort.

BACKGROUND: Family history and African-American race are important risk factors for both prostate cancer (CaP) incidence and aggressiveness. When studying complex diseases such as CaP that have a heritable component, chances of finding true disease susceptibility alleles can be increased by accounting for genetic ancestry within the population investigated. Race, ethnicity and ancestry were studied in a geographically diverse cohort of men with newly diagnosed CaP. METHODS: Individual ancestry (IA) was estimated in the population-based North Carolina and Louisiana Prostate Cancer Project (PCaP), a cohort of 2,106 incident CaP cases (2063 with complete ethnicity information) comprising roughly equal numbers of research subjects reporting as Black/African American (AA) or European American/Caucasian/Caucasian American/White (EA) from North Carolina or Louisiana. Mean genome wide individual ancestry estimates of percent African, European and Asian were obtained and tested for differences by state and ethnicity (Cajun and/or Creole and Hispanic/Latino) using multivariate analysis of variance models. Principal components (PC) were compared to assess differences in genetic composition by self-reported race and ethnicity between and within states. RESULTS: Mean individual ancestries differed by state for self-reporting AA (p = 0.03) and EA (p = 0.001). This geographic difference attenuated for AAs who answered "no" to all ethnicity membership questions (non-ethnic research subjects; p = 0.78) but not EA research subjects, p = 0.002. Mean ancestry estimates of self-identified AA Louisiana research subjects for each ethnic group; Cajun only, Creole only and both Cajun and Creole differed significantly from self-identified non-ethnic AA Louisiana research subjects. These ethnicity differences were not seen in those who self-identified as EA. CONCLUSIONS: Mean IA differed by race between states, elucidating a potential contributing factor to these differences in AA research participants: self-reported ethnicity. Accurately accounting for genetic admixture in this cohort is essential for future analyses of the genetic and environmental contributions to CaP.