Data_Sheet_1_Longitudinal multi-omics alterations response to 8-week risperidone monotherapy: Evidence linking cortical thickness, transcriptomics and epigenetics.docx

BackgroundAntipsychotic treatment-related alterations of cortical thickness (CT) and clinical symptoms have been previously corroborated, but less is known about whether the changes are driven by gene expression and epigenetic modifications.
MethodsUtilizing a prospective design, we recruited 42 treatment-naive first-episode schizophrenia patients (FESP) and 38 healthy controls. Patients were scanned by TI weighted imaging before and after 8-week risperidone monotherapy. CT estimation was automatically performed with the FreeSurfer software package. Participants' peripheral blood genomic DNA methylation (DNAm) status, quantified by using Infinium® Human Methylation 450K BeadChip, was examined in parallel with T1 scanning. In total, CT measures from 118 subjects and genomic DNAm status from 114 subjects were finally collected. Partial least squares (PLS) regression was used to detect the spatial associations between longitudinal CT variations after treatment and cortical transcriptomic data acquired from the Allen Human Brain Atlas. Canonical correlation analysis (CCA) was then performed to identify multivariate associations between DNAm of PLS1 genes and patients' clinical improvement.
ResultsWe detected the significant PLS1 component (2,098 genes) related to longitudinal alterations of CT, and the PLS1 genes were significantly enriched in neurobiological processes, and dopaminergic- and cancer-related pathways. Combining Laplacian score and CCA analysis, we further linked DNAm of 33 representative genes from the 2,098 PLS1 genes with patients' reduction rate of clinical symptoms.
ConclusionsThis study firstly revealed that changes of CT and clinical behaviors after treatment may be transcriptionally and epigenetically underlied. We define a “three-step” roadmap which represents a vital step toward the exploration of treatment- and treatment response-related biomarkers on the basis of multiple omics rather than a single omics type as a strategy for advancing precise care.

研究背景：既往已有研究证实抗精神病药物治疗相关的皮层厚度（cortical thickness, CT）改变与临床症状存在关联，但目前尚不清楚此类改变是否由基因表达及表观遗传修饰所驱动。
研究方法：本研究采用前瞻性设计，招募了42名初治首发精神分裂症患者（first-episode schizophrenia patients, FESP）及38名健康对照。在接受8周利培酮单药治疗前后，分别对受试者进行T1加权成像扫描。采用FreeSurfer软件包自动完成皮层厚度估算。同时在T1扫描时采集受试者外周血基因组DNA甲基化（DNA methylation, DNAm）数据，使用Infinium® Human Methylation 450K BeadChip芯片进行定量检测。最终共收集到118名受试者的皮层厚度数据，以及114名受试者的基因组DNA甲基化数据。采用偏最小二乘（partial least squares, PLS）回归分析检测治疗后纵向皮层厚度变化与从艾伦人脑图谱（Allen Human Brain Atlas）获取的皮层转录组数据之间的空间关联。随后进行典型相关分析（canonical correlation analysis, CCA），以鉴定PLS1相关基因的DNA甲基化水平与患者临床症状改善程度之间的多变量关联。
研究结果：本研究检测到与皮层厚度纵向改变相关的显著PLS1成分（包含2098个基因），该PLS1基因集显著富集于神经生物学过程以及多巴胺能通路与癌症相关通路。结合拉普拉斯得分（Laplacian score）与CCA分析，我们进一步将2098个PLS1基因中的33个代表性基因的DNA甲基化水平与患者临床症状减分率相关联。
研究结论：本研究首次揭示，治疗后皮层厚度与临床行为的改变可能存在转录与表观遗传层面的调控基础。我们提出了一套“三步式”研究框架，该框架为基于多组学（而非单一组学类型）探索治疗及治疗反应相关生物标志物提供了关键路径，有望助力精准医疗的发展。