首次發作精神病症患者之全基因體DNA甲基化異常及與治療反應有關之甲基化異常

Abstract

背景及目的 越來越多研究指出在精神病症患者（大多為思覺失調症患者）的血液中部份候選基因存在DNA甲基化異常的現象。目前僅有很少研究使用全基因體甲基化的方法針對像是首次發作精神病症患者這類對象去了解DNA甲基化在疾病早期所扮演的角色以及DNA甲基化的異常是否會影響這些患者的治療反應。因此，本研究的目的為針對首次發作精神病症患者和健康對照組的血液樣本進行全基因體DNA甲基化分析，去找出在這些患者和對照組之間甲基化程度有差異的位點，並且進一步找出在治療反應良好和治療反應較差的患者之間甲基化程度有差異的位點。 方法 本研究由兩部分組成，一部分為病例對照研究，另一部分為單純病例研究。病例對照研究的部分從台灣北部的三間醫院和兩間診所招募到49名具有基線和六個月後追蹤資料的首次發作精神病症患者，並從研究生、職員和社區中招募了43名健康對照組，並在基線時收集所有參與者的血液樣本。患者在基線和治療六個月後使用活性與負性症狀量表（Positive and Negative Syndrome Scale, PANSS）測量症狀嚴重程度，測量分數下降大於20%的患者被歸類為治療反應良好者（n=19），下降程度為20%或更低的則被歸類為治療反應較差者（n=30）。使用Illumina Infinium MethylationEPIC v1.0 BeadChip對全基因體DNA甲基化進行定量。使用線性模型並校正年齡、性別、吸菸狀況和白細胞種類比例後找出與疾病有關的差異甲基化位點；並另外再校正病人收案來源（住院/門診病人）找出與治療反應有關的差異甲基化位點。透過基因本體富集分析去了解差異甲基化位點所坐落的基因主要參與的生物路徑。分別對在病例對照研究和病例治療反應比較研究中找到的差異甲基化位點其甲基化程度進行加權，去建立兩個分類指數，並使用ROC曲線下面積評估分類指數區分不同組別的表現。 結果 通過比較首次發作精神病症患者與對照組的甲基化程度，我們共找出84個達全基因體顯著水準（p < 5.74 × 10^-6）的差異甲基化位點，以及208個達啟發性顯著水準（p < 5 × 10^-5）的差異甲基化位點。針對這些差異甲基化位點坐落的基因進行基因本體富集分析後，發現它們主要富集於紅血球相關路徑與過氧化物酶體增殖物活化受體相關路徑。比較治療反應良好和治療反應較差的患者其甲基化程度後，沒有找到差異甲基化位點達全基因體顯著水準，但有64個達啟發性顯著水準的位點，並且發現這些位點對應的基因富集於活化蛋白質激酶C的G蛋白偶合受體信號傳導路徑和肌肉細胞相關路徑。基於84個與疾病有關的差異甲基化位點的疾病指數在區分首次發作精神病症患者和健康對照組時具有好的表現，ROC曲線下面積為86.7%；基於64個與治療反應有關的差異甲基化位點所建立的反應指數其ROC曲線下面積達到96.1%，說明此指數在區分治療反應較差者和反應良好者時也有良好的表現。 結論 周邊血液DNA的甲基化異常與精神病症的致病機轉還有精神病症患者的初期治療反應有關。周邊血液的DNA甲基化可能是未來發展協助診斷與辨別精神病症患者不同治療反應的生物標記的潛在來源。

Background Accumulating studies have reported aberrant DNA methylation in certain candidate genes in peripheral blood from patients with psychosis, mostly patients with schizophrenia. There remains rarely examined the role of DNA methylation in the early phase of illness, such as first-episode psychosis (FEP), using a genome-wide methylation approach, and whether aberrant DNA methylation is also involved in these patients’ treatment responses. Hence, this study aims to conduct genome-wide DNA methylation analyses of blood samples from FEP patients and healthy controls to identify differentially methylated positions (DMPs) between these patients and controls and then to further identify DMPs between FEP patients with good treatment response and those with poor treatment response. Methods This study consisted of two parts: a case-control study and a case-only study. For the case-control part, 49 FEP patients were recruited from three hospitals and two private clinics in northern Taiwan and had both baseline and 6-month follow-up data, and 43 controls were recruited from a pool of graduate students, staff, and community volunteers. Blood samples were collected from all participants at baseline. Patients measured the symptom severity using the Positive and Negative Syndrome Scale at baseline and after six months, those with a reduction rate greater than 20% were classified as good responders (n=19), and those with 20% or even lower were classified as poor responders (n=30). The genome-wide DNA methylation patterns were quantified using an Illumina Infinium MethylationEPIC v1.0 BeadChip. Linear models with the adjustment for age, gender, smoking status, and white blood cell type proportion were used to identify disease-related DMPs, and with additional adjustment for recruitment source (inpatient/outpatient) to identify response-related DMPs. Gene ontology enrichment analyses were used to reveal the biological pathways of DMPs mapped genes. Two classification indexes were created by summing DMPs with their weighted methylation levels for the case-control and the good-vs-poor response study, respectively, and then using the area under the ROC curve (AUC) to evaluate the performance of the indexes in distinguishing different groups. Results Comparing FEP patients with controls, we identified 84 DMPs reaching genome-wide significance (p < 5.74 × 10^-6) and 208 DMPs reaching suggestive significance (p < 5 × 10^-5). When both sets of DMPs were subjected to gene ontology enrichment analysis, they were found to be mainly enriched in erythrocyte-related pathways and peroxisome proliferator activated receptor-related pathways. Comparing FEP patients with good response with those with poor response, none DMPs reached genome-wide significance and 64 DMPs reached suggestive significance level, which were found to be enriched in protein kinase C-activating G protein-coupled receptor signaling pathway and muscle cell-related pathways. The disease index based on the 84 disease-related DMPs had a high performance in distinguishing FEP patients from healthy controls, with the AUC of 86.7%. The AUC of the response index based on the 64 response-related DMPs was 96.1%, which also showed good performance in distinguishing poor responders from good responders. Conclusions Aberrant DNA methylation in peripheral blood was associated with the pathogenesis of psychosis and the initial treatment response of patients with psychosis, respectively. DNA methylation levels of blood samples might be a potential source to develop biomarkers for aiding in diagnosis or distinguishing between different treatment response groups of patients with psychosis.