Environmental Exposure and DNA Damage

癌症发病率随着年龄的增长而增加，但原因尚不清楚。 在对血液DNA的小型研究中，已经报道了与年龄相关的甲基化变化，这促使我们探索年龄和DNA甲基化之间的关系。 我们使用来自乳腺癌姐妹研究病例队列研究的数据来评估27，578个CpG的年龄相关甲基化变化。 研究中的女性年龄从35岁到76岁不等，所有女性在入组时都没有癌症。 我们发现几乎三分之一的CpG与年龄显著相关。 使用三个小型公开的血液甲基化数据集，我们构建了749个高置信度年龄相关CpG的子集，这些CpG在所有数据集中显示出一致的年龄效应。 使用来自癌症基因组图谱的新的公开数据，我们证明了我们在血液中发现的衰老位点也广泛适用于多种人体组织。 我们注意到，越来越多的甲基化的年龄相关的CpG是在一些基因的启动子区域，这些基因是细胞分化/发育程序的关键调控因子。 这些基因包括Cdx 2（滋养外胚层）、Gata 4（原始内胚层）、Sox 17（内胚层）、T（中胚层）、MyoD 1（肌肉）、Sox 1（神经外胚层）和Hox基因（A、C和D簇，胚胎体平面）。 发育调控基因对癌症的表观遗传学特别感兴趣，并且与癌症干细胞假说有关：尽管发育基因在正常细胞和癌细胞中都受到转录抑制，但它们经历了从不太稳定的基于组蛋白的基因抑制到永久DNA甲基化沉默的表观遗传转换。 这种永久沉默可能为癌细胞提供进化优势，使其增殖但不分化。 基于表观遗传转换假说，我们预测，增加甲基化的年龄相关的CpG位点将与抑制性H3 K27 me 3组蛋白标记。 为了验证这一假设，我们获得了公开可用的人类胚胎干细胞（HESC）和正常分化的淋巴母细胞系（GM 12878）中组蛋白修饰的ENCODE CHIP-seq数据。 在HESC和淋巴母细胞中，约80%的增加甲基化的年龄相关CpG具有抑制性H3 K27 me 3标记。 这显著高于（p< 10-50）阵列上其他位点的40%频率，并且提供证据表明，与许多发育基因一样，在胚胎干细胞和分化细胞中，年龄相关的甲基化位点通常与H3 K27 me 3组蛋白标记一致。 我们接下来预测，这些增加甲基化的年龄相关位点也应该是肿瘤组织中过度甲基化的靶点。 我们分析了来自癌症基因组图谱项目的正常肿瘤对的公开甲基化数据，该项目有7种癌症的数据。 我们发现，年龄相关位点的变化方向是所有7种肿瘤类型中显著甲基化变化的有力预测因素。 这些发现表明癌症干细胞假说的延伸：癌症发病率增加与年龄之间的关联可能部分归因于某些基因的中间年龄相关甲基化变化。 表观基因组范围的研究已经确定了五个基因座，其中血液DNA甲基化降低与成人吸烟相关。 我们在27 K CpG甲基化研究和450 K CpG DES研究中，检查了吸烟史与女性DNA甲基化的关系。 我们确认了先前在F2 RL 3，GPR 15，AHRR和2号染色体的基因间区域中发现的四个吸烟相关CpG的吸烟关联。 此外，我们还首次报道了粪卟啉原氧化酶（CPOX）基因中CpG之间的关联。 CPOX编码一种参与血红素生物合成的酶，该酶定位于线粒体的膜间隙，并且CPOX中的突变是遗传性粪卟啉症的原因。 为了验证该基因座，我们设计了一种焦磷酸测序测定CPOX第一内含子中的CpG海岸位点，并在来自姐妹研究的476名妇女的独立组中复制了我们的吸烟结果。 我们发现，相对于从不吸烟者的水平，前吸烟者的甲基化水平较低（p <2x 10 -6），而目前吸烟者的甲基化水平甚至更低（p <2x 10 -18）。 这些结果表明，与吸烟相关的氧化应激或血红素生物合成增加可能导致CPOX第一内含子甲基化持续降低。 此外，我们还完成了一项研究，即吸烟母亲所生婴儿血液中的表观遗传变化。 母亲在怀孕期间吸烟与婴儿发病率和死亡率显著相关，并可能影响以后的疾病风险。吸烟（和其他环境因素）可能具有长期影响的一种机制是通过表观遗传修饰，如DNA甲基化。 我们分析了从889名婴儿分娩后不久收集的血液，并使用Illumina 450 K甲基化芯片检查了DNA。 我们在吸烟者的婴儿中发现了185个具有全基因组意义的甲基化改变的CpG。这些对应于110个基因区域，其中7个先前已报道，10个是使用公开结果新确认的。其中最值得注意的是FRMD 4A、ATP 9A、GALNT 2和MEG 3，它们涉及与尼古丁依赖、戒烟以及胎盘和胚胎发育相关的过程。 在新生儿中发现的甲基化变化可能介导子宫内母亲吸烟暴露与后来的健康结果之间的关联。

Cancer incidence increases with age, but it is not clear why. Age-related changes in methylation have been reported in small studies of blood DNA, prompting us to explore associations between age and DNA methylation. We used data from our Sister Study case-cohort study of breast cancer to assess age-related methylation change at 27,578 CpGs. The women in the study ranged in age from 35 to 76 and all women were cancer-free at the time of enrollment. We found that almost a third of the CpGs were significantly associated with age. Using three small public methylation data sets on blood, we constructed a subset of 749 high confidence age-related CpGs that showed consistent age effects across all data sets. Using new publicly-available data from The Cancer Genome Atlas, we demonstrate that the aging sites we identified in blood are also broadly applicable across multiple human tissues. We noted that increasingly methylated age-related CpGs were in the promoter regions of a number of genes that are key regulators of cell differentiation/development programs. These genes included Cdx2 (Trophectoderm), Gata4 (primitive endoderm), Sox17 (endoderm), T (mesoderm), MyoD1 (muscle), Sox1 (neuroectoderm), and Hox genes (A, C, and D clusters, embryo body plan). Developmental regulatory genes are of particular interest for the epigenetics of cancer, and in relation to the cancer stem cell hypothesis: Although developmental genes are transcriptionally repressed in both normal cells and cancer cells, they undergo epigenetic switching from less stable, histone-based, gene repression to permanent DNA-methylation silencing. This permanent silencing may provide an evolutionary advantage to a cancer cell, allowing it to proliferate but not to differentiate. Based on an epigenetic switching hypothesis we predicted that increasingly-methylated age-associated CpG sites would be associated with the repressive H3K27me3 histone mark. To test this hypothesis we obtained publicly available ENCODE CHIP-seq data for histone modifications in a human embryonic stem cell (HESC), and in a normal differentiated lymphoblastoid cell line (GM12878) . In both HESC and lymphoblastic cells about 80% of increasingly-methylated age-related CpGs had the repressive H3K27me3 mark. This was significantly higher (p< 10-50) than the 40% frequency at other sites on the array, and provides evidence that, like many developmental genes, the age-related methylation sites often coincide with the H3K27me3 histone mark in both embryonic stem and differentiated cells. We next predicted that these increasingly-methylated age-related sites should also be targets for overmethylation in tumor tissues. We analyzed publicly available methylation data for normal-tumor pairs from The Cancer Genome Atlas project that had data for 7 cancers. We found that the direction of change at age-related sites was a powerful predictor of significant methylation change in all 7 tumor types. These finding suggest an extension to the cancer-stem cell hypothesis: The association between increasing cancer incidence and age may in part be attributable to intermediate age-related methylation change at certain genes. Epigenome-wide studies have identified five loci where there is decreased blood DNA methylation associated with adult smoking. We examined smoking history in relation to DNA methylation in women from our 27K CpG methylation study and in women from our 450K CpG study of DES. We confirm smoking associations for four previously identified smoking-related CpGs in F2RL3, GPR15, AHRR and an intergenic region of chromosome 2. In addition we provide the first report of an association between a CpG in the coproporphyrinogen oxidase (CPOX) gene. CPOX codes for an enzyme involved in heme biosynthesis that is localized to the intermembrane space of the mitochondria, and mutations in CPOX are a cause of hereditary coproporphyria. In order to validate this locus we designed a pyrosequencing assay for the CpG shore site in the first intron of CPOX, and replicated our smoking result in an independent set of 476 women from the Sister Study. We found that relative to the levels in never smokers, there was lower methylation in former smokers (p < 2x10-6) and even lower methylation levels among current smokers (p < 2x10-18). These results suggest that the oxidative stress or the increased heme biosynthesis associated with smoking might lead to persistent decreased methylation in the first intron of CPOX. In addition we have completed a study epigenetic changes in blood of babies born to mothers who smoke. Maternal smoking during pregnancy is associated with significant infant morbidity and mortality, and may influence later disease risk. One mechanism by which smoking (and other environmental factors) might have long-lasting effects is through epigenetic modifications such as DNA methylation. We analyzed blood collected from 889 infants shortly after delivery and examined the DNA using the Illumina 450K methylation chip. BWe identified 185 CpGs with altered methylation in infants of smokers at genome-wide significance. These correspond to 110 gene regions, of which 7 have been previously reported and 10 are newly confirmed using publically-available results. Among these 10 the most noteworthy are FRMD4A, ATP9A, GALNT2, and MEG3, implicated in processes related to nicotine dependence, smoking cessation, and placental and embryonic development. The methylation changes identified in newborns may mediate the association between in utero maternal smoking exposure and later health outcomes.