Van Assche Evelien, Hohoff Christa, Zang Johannes, Knight Matthew J, Baune Bernhard T
Department of Psychiatry, University of Münster, Münster, Germany.
Discipline of Psychiatry, Adelaide Medical School, University of Adelaide, Adelaide, SA, Australia.
Front Mol Neurosci. 2023 Aug 18;16:1223216. doi: 10.3389/fnmol.2023.1223216. eCollection 2023.
The etiology of major depressive disorder (MDD) involves the interaction between genes and environment, including treatment. Early molecular signatures for treatment response and remission are relevant in a context of personalized medicine and stratification and reduce the time-to-decision. Therefore, we focused the analyses on patients that responded or remitted following a cognitive intervention of 8 weeks.
We used data from a randomized controlled trial (RCT) with MDD patients ( = 112) receiving a cognitive intervention. At baseline and 8 weeks, blood for DNA methylation (Illumina Infinium MethylationEPIC 850k BeadChip) was collected, as well as MADRS. First, responders ( = 24; MADRS-reduction of at least 50%) were compared with non-responders ( = 60). Then, we performed longitudinal within-individual analyses, for response ( = 21) and for remission ( = 18; MADRS smaller or equal to 9 and higher than 9 at baseline), respectively, as well as patients with no change in MADRS over time. At 8 weeks the sample comprised 84 individuals; 73 patients had DNA methylation for both time-points. The RnBeads package (R) was used for data cleaning, quality control, and differential DNA-methylation (limma). The within-individual paired longitudinal analysis was performed using Welch's -test. Subsequently gene-ontology (GO) pathway analyses were performed.
No CpG was genome-wide significant CpG ( < 5 × 10). The most significant CpG in the differential methylation analysis comparing response versus non-response was in the gene (cg01601845; = 1.53 × 10), linked to neurotransmission. The most significant GO-terms were linked to telomeres. The longitudinal response analysis returned 67 GO pathways with a < 0.05. Two of the three most significant pathways were linked to sodium transport. The analysis for remission returned 46 GO terms with a -value smaller than 0.05 with pathways linked to phosphatase regulation and synaptic functioning. The analysis with stable patients returned mainly GO-terms linked to basic cellular processes.
Our result suggest that DNA methylation can be suitable to capture early signs of treatment response and remission following a cognitive intervention in depression. Despite not being genome-wide significant, the CpG locations and GO-terms returned by our analysis comparing patients with and without cognitive impairment, are in line with prior knowledge on pathways and genes relevant for depression treatment and cognition. Our analysis provides new hypotheses for the understanding of how treatment for depression can act through DNA methylation and induce response and remission.
重度抑郁症（MDD）的病因涉及基因与环境之间的相互作用，包括治疗。治疗反应和缓解的早期分子特征在个性化医疗和分层的背景下具有重要意义，并能缩短决策时间。因此，我们将分析重点放在了接受8周认知干预后有反应或缓解的患者身上。
我们使用了一项针对MDD患者（n = 112）的随机对照试验（RCT）数据，这些患者接受了认知干预。在基线和8周时，采集血液用于DNA甲基化分析（Illumina Infinium MethylationEPIC 850k BeadChip）以及蒙哥马利-艾森伯格抑郁评定量表（MADRS）评估。首先，将有反应者（n = 24；MADRS降低至少50%）与无反应者（n = 60）进行比较。然后，我们分别对有反应者（n = 21）和缓解者（n = 18；基线时MADRS小于或等于9且高于9）以及MADRS随时间无变化的患者进行个体内纵向分析。在8周时，样本包括84名个体；73名患者在两个时间点都有DNA甲基化数据。使用RnBeads软件包（R语言）进行数据清理、质量控制和差异DNA甲基化分析（limma）。个体内配对纵向分析使用韦尔奇t检验。随后进行基因本体（GO）通路分析。
在全基因组范围内没有发现达到显著水平的CpG（p < 5 × 10⁻⁸）。在比较有反应者与无反应者的差异甲基化分析中，最显著的CpG位于GRIA2基因（cg01601845；p = 1.53 × 10⁻⁷），与神经传递有关。最显著的GO术语与端粒相关。纵向反应分析得出67条p < 0.05的GO通路。三个最显著的通路中有两个与钠转运有关。缓解分析得出46条p值小于0.05的GO术语，其通路与磷酸酶调节和突触功能有关。对病情稳定患者的分析主要得出与基本细胞过程相关的GO术语。
我们的结果表明，DNA甲基化可能适合捕捉抑郁症认知干预后治疗反应和缓解的早期迹象。尽管在全基因组范围内未达到显著水平，但我们比较有认知障碍和无认知障碍患者的分析中得出的CpG位置和GO术语，与先前关于抑郁症治疗和认知相关通路及基因的知识相符。我们的分析为理解抑郁症治疗如何通过DNA甲基化发挥作用并诱导反应和缓解提供了新的假设。
