Bind Marie-Abele C, Coull Brent A, Baccarelli Andrea, Tarantini Letizia, Cantone Laura, Vokonas Pantel, Schwartz Joel
Department of Statistics, Faculty of Arts and Science, Harvard University, Cambridge, MA, USA.
Department of Biostatistics, Harvard School of Public Health, Boston, MA, USA.
Environ Res. 2016 Oct;150:38-46. doi: 10.1016/j.envres.2016.05.034. Epub 2016 May 26.
Temperature has been related to mean differences in DNA methylation. However, heterogeneity in these associations may exist across the distribution of methylation outcomes. This study examined whether the association between three-week averaged of temperature and methylation differs across quantiles of the methylation distributions in nine candidate genes. We measured gene-specific blood methylation repeatedly in 777 elderly men participating in the Normative Aging Study (1999-2010). We fit quantile regressions for longitudinal data to investigate whether the associations of temperature on methylation (expressed as %5mC) varied across the distribution of the methylation outcomes. We observed heterogeneity in the associations of temperature across percentiles of methylation in F3, TLR-2, CRAT, iNOS, and ICAM-1 genes. For instance, an increase in three-week temperature exposure was associated with a longer left-tail of the F3 methylation distribution. A 5°C increase in temperature was associated with a 0.15%5mC (95% confidence interval (CI): -0.27,-0.04) decrease on the 20th quantile of F3 methylation, but was not significantly related to the 80th quantile of this distribution (Estimate:0.06%5mC, 95%CI: -0.22, 0.35). Individuals with low values of F3, TLR-2, CRAT, and iNOS methylation, as well as a high value of ICAM-1 methylation, may be more susceptible to temperature effects on systemic inflammation.
温度与DNA甲基化的平均差异有关。然而,这些关联在甲基化结果的分布中可能存在异质性。本研究调查了九个候选基因中,温度的三周平均值与甲基化之间的关联在甲基化分布的不同分位数上是否存在差异。我们在参与规范衰老研究(1999 - 2010年)的777名老年男性中,反复测量了基因特异性的血液甲基化水平。我们对纵向数据进行分位数回归,以研究温度与甲基化(以%5mC表示)之间的关联在甲基化结果分布上是否存在变化。我们观察到在F3、TLR - 2、CRAT、iNOS和ICAM - 1基因的甲基化百分位数上,温度关联存在异质性。例如,三周温度暴露的增加与F3甲基化分布的左尾更长有关。温度升高5°C与F3甲基化第20分位数上降低0.15%5mC(95%置信区间(CI):-0.27,-0.04)相关,但与该分布的第80分位数无显著关联(估计值:0.06%5mC,95%CI:-0.22,0.35)。F3、TLR - 2、CRAT和iNOS甲基化水平较低以及ICAM - 1甲基化水平较高的个体,可能对温度对全身炎症的影响更敏感。