Project 2: Primary melanoma DNA Methylation profiling for evaluating subtypes and survival

项目总结/摘要 虽然已经确定黑色素瘤经常具有异常的DNA甲基化模式，但这是不可能的。 不清楚原发性黑色素瘤的DNA甲基化是否定义了生物学相关的亚类或预测了疾病- 具体的生存。该项目的目标是识别和表征DNA甲基化为基础的主要 黑色素瘤亚类，并训练，测试和验证CpG签名预后的疾病特异性生存， 原发性黑素瘤我们的假设是，原发性黑色素瘤的DNA甲基化将定义亚组 （包括预后不良的CIMP-"CpG岛甲基化表型"亚型），并且CpG标签将 是黑色素瘤生存的预后指标。将进行Illumina Infinium甲基化EPIC（850 K）阵列分析 同样的1,000例原发性黑色素瘤，将用于整个项目。这些黑色素瘤 来自AJCC IIA/IIB/IIC/IIIA/IIIB期患者，包括500例死于黑色素瘤的患者 5年和500从那些谁住了至少5年匹配的阶段。在目标1中，使用1000个 对于原发性黑色素瘤，我们将识别基于甲基化的黑色素瘤亚类，其特征是5年生存率 和其他结局、人口统计学因素、集中审查的组织病理学特征、CD3/CD8/FOXP3 免疫特征、PD1和PDL1蛋白表达、体细胞突变、拷贝数改变和mRNA 和miRNA表达数据。进一步表征将确定子类是否与 生物相关的DNA甲基化特征，我们的小组和其他人以前已经从 独立的数据集。在目标2中，使用统计建模，我们将训练原发性黑色素瘤CpG签名， 使用660例原发性黑色素瘤预测5年内黑色素瘤死亡（主要结局） 用340个原发性黑色素瘤进行测试。我们将确定这个CpG存活信号是否增加了 通知AJCC staging。作为次要结局，我们还将制定5年内复发的签名 年、前哨淋巴结（SLN）阳性和SLN活检阴性后远处复发。在目标3中，我们 使用定量PCR技术验证存活特征中CpG/基因的DNA甲基化水平， 甲基化特异性聚合酶链反应（PCR）。对于选择的标记物，我们将定量测量其 蛋白质表达差异使用多重免疫组化，比较原发性黑色素瘤， 存活的患者和诊断后5年内死亡的患者，并评估最具预后的功能 项目3的目标。我们的目标是识别和描述基于DNA甲基化的原发性黑色素瘤 亚类，并发现和验证一个CpG签名预后的生存从黑色素瘤，增加 通知AJCC staging。我们将开始将该特征应用于临床可行的检测预后 为了生存该项目应允许识别预期生存率较差的黑色素瘤患者， 他们可以从更密切的随访和辅助治疗中受益。我们的结果还应该指出哪些目标 在预后不良的黑色素瘤中存在治疗反应的预测因子。

Project Summary/Abstract Although it has been established that melanomas frequently have aberrant DNA methylation patterns, it is unknown if DNA methylation of primary melanoma defines biologically relevant subclasses or predicts disease- specific survival. This project's goals are to identify and characterize DNA methylation-based primary melanoma subclasses, and train, test, and validate a CpG signature prognostic for disease-specific survival in primary melanoma. Our hypothesis is that DNA methylation in primary melanoma will define subgroups (including a poor-prognostic CIMP - `CpG island methylator phenotype' subtype) and that a CpG signature will be prognostic for melanoma survival. Illumina Infinium MethylationEPIC (850K) array profiling will be performed on the same 1,000 primary melanomas to be utilized throughout this program project. These melanomas will be from AJCC stage IIA/IIB/IIC/IIIA/IIIB patients, including 500 from patients who died due to melanoma within 5 years and 500 from those who lived at least 5 years matched on stage. In Aim 1, using data from the 1,000 primary melanomas, we will identify methylation-based melanoma subclasses, characterized by 5-year survival and other outcomes, demographic factors, centrally reviewed histopathological features, a CD3/CD8/FOXP3 immune profile, PD1 and PDL1 protein expression, somatic mutations, copy number alterations, and mRNA and miRNA expression data. Further characterization will determine if the subclasses are associated with biologically relevant DNA methylation signatures that our group and others have previously produced from independent datasets. In Aim 2 using statistical modeling, we will train a primary melanoma CpG signature prognostic for death from melanoma within 5 years (the primary outcome) using 660 of the primary melanomas and test it using 340 of the primary melanomas. We will determine whether this CpG survival signature adds information to AJCC staging. As secondary outcomes, we will also develop signatures for recurrence within 5 years, sentinel lymph node (SLN) positivity, and distant recurrence after negative SLN biopsy. In Aim 3, we will technically validate DNA methylation levels of CpGs/genes in the survival signature using quantitative methylation-specific polymerase chain reaction (PCR). For select markers, we will quantitatively measure their protein expression differences using multiplexed immunohistochemistry, comparing primary melanomas from patients alive and those who died within 5 years of diagnosis and assess the function of the most prognostic targets in Project 3. Our goal is to identify and characterize DNA methylation-based primary melanoma subclasses and discover and validate a CpG signature prognostic for survival from melanoma that adds information to AJCC staging. We will initiate application of the signature to a clinically viable assay prognostic for survival. This project should allow identification of melanoma patients expected to have worse survival and who could benefit from closer follow-up and adjuvant treatments. Our results should also indicate which targets predictive of treatment responses are present in poor prognostic melanomas.