Abstract:Abstract Background: Prostate cancer (PCa) is the most commonly diagnosed cancer for men in the United States. Most of the PSA screening-detected PCa are slow-growing and not life-threatening. However, the majority of PSA-detected PCa receive aggressive treatment, which is often associated with significant morbidity. Clinical variables, such as Gleason Score (GS), PSA level, and tumor stage, are not sufficient to distinguish aggressive from indolent diseases at diagnosis. Biomarkers are urgently needed for more accurate prediction of individual tumor behavior and patient prognosis. DNA methylation has shown profound impact on cancer development and progression. In this study, we performed a genome-wide DNA methylation profiling of peripheral blood leukocytes to identify predictors of aggressive PCa. Methods: We performed leukocyte DNA methylation profiling in 140 GS=6 PCa patients and 147 GS ≥ 8 patients using the Illumina's Methylation 450K BeadChip. The differential methylated CpG position (DMP) was identified using the ChAMP analysis package implemented in R. Top differentially methylated CpG sites between less aggressive (GS=6) and more aggressive (GS≥ 8) patients were used to develop a signature to best cluster these two groups using semi-supervised recursively partitioned mixture modeling (SS-RPMM) algorithm. The white blood cell subtype composition was analyzed with RefbaseEWAS method. Ingenuity pathway analysis was used to find the enriched pathways that may determine aggressive prostate cancer. Results: Analyzing all the assayed CpG sites together, the mean methylation level appeared to be lower in GS ≥ 8 patients than in GS=6 patients. This global DNA hypomethylation was only significant at CpG sites located within gene body, shore, shelf, opensea and 3' UTR regions, but not in promoter and regions close to transcription start sites. There were many individual hypermethylated CpG sites in GS ≥ 8 patients that are mostly located in CpG islands within promoter and in the vicinity of transcription start sites. We then split the patients into a training (70 GS=6 and 73 GS ≥ 8 patients) and testing set (70 GS=6 and 74 GS ≥ 8 patients) to identify and validate a CpG methylation signature that best differentiates GS=6 and GS≥ 8 patients. The signature can identify a subgroup of aggressive PCa patients with higher risks of biochemical recurrence (HR=2.57, 95% Cl, 0.99-6.68). We also analyzed the white blood cell subtypes in patients and found that natural killer (NK) cell proportion is significantly decreased in patients who experienced biochemical recurrence (P=0.045), indicating weakened immune response. Conclusion: These data suggest that DNA methylation in peripheral blood leukocytes may serve as valuable biomarkers to supplement clinical variables to predict aggressive PCa at diagnosis and worse prognosis of PCa patients. Citation Format: Yuyan Han, Jia Li, Mutian Zhang, Jeri Kim, Xifeng Wu, Deqiang Sun, Jian Gu. Genome-wide DNA methylation profiling of peripheral blood leukocytes identifies predictors of aggressive prostate cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 3332.

Abstract 3332: Genome-wide DNA Methylation Profiling of Peripheral Blood Leukocytes Identifies Predictors of Aggressive Prostate Cancer