Childhood maltreatment (CM) is an established major risk factor for a number of negative health outcomes later in life. While epigenetic mechanisms, such as DNA methylation (DNAm), have been proposed as a means of embedding this environmental risk factor, little is known about its timing and trajectory, especially in very young children. It is also not clear whether additional environmental adversities, often experienced by these children, converge on similar DNAm changes.

Here, we calculated a cumulative adversity score, which additionally to CM includes socioeconomic status (SES), other life events, parental psychopathology and epigenetic biomarkers of prenatal smoking and alcohol consumption. We investigated the effects of CM alone as well as the adversity score on longitudinal DNAm trajectories in the Berlin Longitudinal Child Study. This is a cohort of 173 children aged 3–5 years at baseline of whom 86 were exposed to CM. These children were followed-up for 2 years with extensive psychometric and biological assessments as well as saliva collection at 5 time points providing genome-wide DNAm levels.

Overall, only a few DNAm patterns were stable over this timeframe, but less than 10 DNAm regions showed significant changes. At baseline, neither CM nor the adversity score associated with DNAm changes. However, in 6 differentially methylated regions (DMRs), CM and the adversity score significantly moderated DNAm trajectories over time. A number of these DMRs have previously been associated with adverse prenatal exposures. In our study, children exposed to CM also presented with epigenetic signatures indicative of increased prenatal exposure to tobacco and alcohol, as compared to non-CM exposed children. These epigenetic signatures of prenatal exposure strongly correlate with DNAm regions associated with CM and the adversity score. Finally, weighted correlation network analysis revealed a module of CpGs exclusively associated with CM.

While our study identifies DNAm loci specifically associated with CM, especially within long non-coding RNAs, the majority of associations were found with the adversity score with convergent association with indicators of adverse prenatal exposures. This study highlights the importance of mapping not only of the epigenome but also the exposome and extending the observational timeframe to well before birth.

儿童期虐待 (CM) 是导致以后生活中出现许多负面健康结果的既定主要风险因素。虽然 DNA 甲基化 (DNAm) 等表观遗传机制已被提议作为嵌入这种环境风险因素的一种手段，但人们对其发生的时间和轨迹知之甚少，尤其是对于年幼的儿童。目前还不清楚这些儿童经常经历的其他环境逆境是否会导致类似的 DNAm 变化。

在这里，我们计算了累积逆境评分，除了 CM 之外，还包括社会经济地位 (SES)、其他生活事件、父母精神病理学以及产前吸烟和饮酒的表观遗传生物标志物。我们在柏林纵向儿童研究中研究了单独 CM 以及逆境评分对纵向 DNAm 轨迹的影响。这是一个由 173 名基线年龄为 3-5 岁的儿童组成的队列，其中 86 名儿童接触过 CM。这些儿童接受了为期 2 年的随访，进行了广泛的心理测量和生物学评估，并在 5 个时间点采集唾液，提供全基因组 DNAm 水平。

总体而言，只有少数 DNAm 模式在此时间范围内保持稳定，但不到 10 个 DNAm 区域显示出显着变化。在基线时，CM 和与 DNAm 相关的逆境评分均未发生变化。然而，在 6 个差异甲基化区域 (DMR) 中，随着时间的推移，CM 和逆境评分显着调节 DNAm 轨迹。其中许多 DMR 以前曾与不良产前暴露有关。在我们的研究中，与未接触过 CM 的儿童相比，接触过 CM 的儿童还表现出表观遗传特征，表明产前接触烟草和酒精的情况有所增加。这些产前暴露的表观遗传特征与 CM 和逆境评分相关的 DNAm 区域密切相关。最后，加权相关网络分析揭示了一个与 CM 专门相关的 CpG 模块。

虽然我们的研究确定了与 CM 特别相关的 DNAm 位点，特别是在长非编码 RNA 中，但发现大多数关联与逆境评分以及与不良产前暴露指标的趋同关联。这项研究强调了不仅绘制表观基因组图谱而且绘制暴露组图谱并将观察时间范围延长至出生前的重要性。