外周全血中与老化相关的DNA甲基化标记的来源

机体老化与癌症、神经退行性疾病等许多复杂疾病相关。目前,研究者已在外周全血中识别了大量的与老化相关的DNA甲基化标记,这些标记可能反映外周血白细胞在机体老化过程中发生的变化,也可能反映外周血中与年龄相关的细胞构成比例的变化。文章利用3组正常个体外周全血DNA甲基化谱,采用Spearman秩相关分析识别了与老化相关的CpG甲基化位点(age-related DNA methylation CpG sites, arCpGs)并评价了其可重复性;利用去卷积算法估计了各外周血样本中髓性和淋巴性细胞的比例并分析了其与年龄的相关性;比较了在外周全血、CD4+T细胞和CD14+单核细胞中识别的arCpGs的一致性。结果显示,在独立外周全血数据中识别的arCpGs具有显著的可重复性(超几何检验,P=1.65×10^-11)。外周血髓性和淋巴性细胞的比例分别与年龄显著正、负相关(Spearman秩相关检验,P<0.05,r≤0.22),它们间DNA甲基化水平差异较大的CpG位点倾向于在外周全血中被识别为arCpGs。在CD4+T细胞中识别的arCpGs与在外周全血中识别的arCpGs显著交叠(超几何检验,P=6.14×10^-12),且99.1%的交叠位点在CD4+T细胞及外周全血中的DNA甲基化水平与年龄的正、负相关性一致。尽管在CD14+单核细胞中识别的arCpGs与在外周全血中识别的arCpGs并不显著交叠,但是在交叠的51个arCpGs中,有90.1%的位点在CD14+单核细胞、外周全血以及CD4+T细胞中的DNA甲基化水平与年龄的正、负相关性一致,提示它们可能主要反映细胞间共同的改变。在外周全血中识别的arCpGs主要反映某些白细胞共同或特异的DNA甲基化改变,但是也有一部分反映外周血细胞比例构成的变化。

Age-related DNA methylation changes in peripheral whole blood

Aging is associated with many complex diseases such as cancer and neurodegenerative diseases. Recently, many age-related DNA methylation biomarkers in peripheral whole blood have been identified. These biomarkers may reflect DNA methylation changes derived from changes in the number of a specific leukocyte cell type during aging. To clarify the source of these age-related DNA methylation changes, we analysed DNA methylation profile of peripheral whole blood from three independent cohorts of healthy subjects and identified age-related DNA methylation CpG sites (arCpGs) using the Spearman’s rank test with high reproducibility (Hypergeometric test, P=1.65×10^-11). Using a deconvolution algorithm, we found that the proportion of myeloid lineage cells was increased while that of lymphoid lineage cells was decreased in the peripheral whole blood with age (Spearman’s rank correlation test, P<0.05, r≤0.22). The CpG sites, whose methylation levels were significantly different in myeloid cells and lymphoid cells, were preferentially recognized as arCpGs in peripheral whole blood. Moreover, the arCpGs in CD4+ T cells significantly overlapped with that in peripheral whole blood (Hypergeometric test, P=6.14×10^-12) and 99.1% of the overlapping arCpGs had consistent positive or negative correlations with age. Though the arCpGs in CD14+ monocytes did not significantly overlap with that in peripheral whole blood (Hypergeometric test, P=0.232), 90.1% of 51 overlapping arCpGs were correlated with age in CD14+ monocytes, peripheral whole blood, and CD4+ T cells consistently. In summary, most of the methylation changes in arCpGs identified in peripheral whole blood come from common or specific DNA methylation changes in leukocyte subtypes, while part of them reflect alterations in the number of specific cell types of leukocytes.