对不同族裔前列腺组织进行DNA甲基化分析以鉴定预测前列腺癌进展的新型生物标志物

Prostate cancer (PCa) is the most common non-dermatological malignancy and the second leading cause of cancer-related deaths among men in the United States. Disparate clinical outcomes are especially observed between patients of African American (AA) and European American (EA) ancestry. This difference may reflect the long-term biological consequences of systemic racism, including differential environmental exposures and unequal access to healthcare. Additionally, these factors, along with genetic variation across ancestries, can contribute to epigenetic changes in PCa, including DNA methylation. To date, whole-DNA-methylome data has overrepresented EA individuals, and was used to develop DNA methylation-based biomarkers for PCa. In this study, we comprehensively characterize the genome-wide DNA methylome of PCa in a cohort of ethnically diverse individuals to identify new PCa-specific epigenetic features and a novel set of DNA methylation biomarkers. We obtained formalin-fixed paraffin-embedded (FFPE) prostate tissues from AA and EA PCa patients and generated DNA methylation profiles across normal controls and tumors spanning multiple Gleason grades, molecular subtypes, and pathological features. FFPE tissue regions were macrodissected based on pathologic annotations, allowing for recovery of Gleason grade-specific DNA. DNA methylation assays included both the Illumina Infinium Methylation EPIC array and Whole-Genome Bisulfite Sequencing (WGBS). We identified thousands of hypomethylated and hypermethylated CpG sites between normal and tumor samples across prostate tissue samples first profiled by both the EPIC array that were also covered by WGBS. Based on the most variable differentially methylated regions (DMRs), we identified clusters of tumor samples exhibiting distinct DNA methylation patterns. When we developed a model to identify DMRs that are predictive of tumor grade and other clinicopathological features; interestingly, we found that some DMRs were located within intergenic regions, including enhancers of genes that are typically upregulated in PCa. Additionally, we identified PCa-specific partially methylated domains, which are progressively hypomethylated throughout tumor progression. The average methylation levels at these regions predicted tumor grade. Overall, this study identifies PCa-specific epigenetic features, revealing potential new biomarkers that predict PCa severity and progression, with the inclusion of an ethnically diverse cohort.