Developing ATAC-array as a novel epigenetic biomarker to guide personalized therapy in pancreatic cancer

摘要：虽然化疗仍然是全身治疗的主要手段，但大量的癌症患者， 病人没有反应。大约一半的手术切除的胰腺导管腺癌（PDAC）患者 一年内复发，尽管辅助化疗。利用全基因组ATAC-Seq技术， 分析切除的PDAC的肿瘤细胞内在染色质可及性模式，我们发现了一个标志， 1092个染色质位点显示无病生存期（DFS）&gt; 1年的患者之间的差异可及性< 1 year and patients with DFS >。基于这个特征，我们开发了“ATAC阵列”-一种新颖的，临床上 适用的微阵列平台，读取染色质可及性，而无需与Next相关的时间和成本 世代排序。使用这个新的平台，我们开发了一个基于染色质的预后评分， 从我们的训练集样本的ATAC阵列结果，这是进一步验证了一个独立的队列 PDAC患者衍生的3D类器官。通过分析过度表达的转录因子结合位点， 在DFS延长的患者中，我们发现HNF 1b是一种转录因子， 显示短DFS与长DFS患者之间核定位差异的因素。结合这两 简单的方法（即，ATAC阵列“预后评分”和免疫组织化学测定，或 HNF 1b核定位），我们开发了一种算法，在诊断时对PDAC进行分层 将患者分为DFS差异超过7倍的预后组。我们的分层算法 提供了一个简单的和临床上可实现的预测有利与不利的表观遗传状态的PDAC。 基于这项工作，我们现在假设基于ATAC阵列的染色质可及性分析将提供 一种预测局部和转移性胰腺癌预后的临床相关手段， 允许合理选择患者进行表观遗传治疗的临床试验。通过这项R21赠款，我们现在的目标是 优化ATAC阵列用于临床，并将其扩展到胰腺癌的所有阶段。我们的目标是优化 允许ATAC-阵列应用于从针穿刺活检取回的稀少肿瘤标本的方法。我们 此外，还计划在患者衍生的3D类器官中体外筛选表观遗传药物，以确定是否存在不良反应。 染色质可及性的模式可能适用于表观遗传“重编程”疗法，例如ATAC- 阵列可以提供一种治疗生物标志物，为表观遗传治疗选择患者。长期目标是 该项目是为了提高染色质可及性的全基因组评估的可行性，作为临床相关的 PDAC患者结局的指标，并促进表观遗传药物合理纳入 显示不良预后表观遗传特征的患者的治疗方案。

ABSTRACT: Although chemotherapy remains the mainstay of systemic therapy, a large number of cancer patients fail to respond to. About half of surgically resected pancreatic ductal adenocarcinoma (PDAC) patients relapse within a year despite adjuvant chemotherapy. Using genome-wide ATAC-Seq to comprehensively analyze tumor cell-intrinsic chromatin accessibility patterns of resected PDAC, we discovered a signature of 1092 chromatin loci displaying differential accessibility between patients with disease free survival (DFS) < 1 year and patients with DFS > 1 year. Based on this signature, we developed “ATAC-array” – a novel, clinically applicable microarray platform that reads chromatin accessibility without the time and cost associated with Next Generation Sequencing. Using this novel platform, we developed a chromatin-based Prognosis Score, derived from the ATAC-array results from our training set samples, which was further validated on an independent cohort of PDAC patient-derived 3D organoids. By analyzing over-represented transcription factor binding sites among chromatin loci found to be accessible in patients with prolonged DFS, we also identified HNF1b as a transcription factor displaying differential nuclear localization between patients with short vs. long DFS. Combining these two simple methodologies (i.e., the ATAC-array “Prognosis Score” and immunohistochemical determination or HNF1b nuclear localization), we have developed an algorithm that at the time of diagnosis stratifies PDAC patients into prognostic groups with more than seven-fold differences in DFS. Our stratification algorithm provides a simple and clinically achievable prediction of favorable vs unfavorable epigenetic states in PDAC. Based on this work, we now hypothesize that ATAC-array based analysis of chromatin accessibility will provide a clinically relevant means of predicting prognosis in both localized and metastatic pancreatic cancer, and further allow rational selection of patients for clinical trials of epigenetic therapy. Through this R21 grant, we now aim to optimize ATAC-array for clinical use and expand it to all stages of pancreatic cancer. Our goal is to optimize methodologies allowing ATAC-array to be applied to scant tumor specimens retrieved from needle biopsies. We additionally plan to screen epigenetic drugs in vitro in patient-derived 3D organoids to determine whether adverse patterns of chromatin accessibility might be amenable to epigenetic “reprogramming” therapy, such that ATAC- array could provide a therapeutic biomarker selecting patients for epigenetic therapy. The long-term goal of this project is to enhance the feasibility of genome-wide assessment of chromatin accessibility as a clinically relevant indicator of outcome in PDAC patients, and to facilitate the rational inclusion of epigenetic drugs into the therapeutic regimens of patients displaying poor-prognosis epigenetic signatures.