Abstract 3825: Whole genome bisulfite sequencing from plasma of patients with metastatic breast cancer identifies putative biomarkers

Abstract

Abstract Although several improvements in the management of breast cancer(BC) have been made, an estimated 90% of BC related deaths are due to the development of distant metastasis to organs such as lung, bone, liver, and brain. Numerous studies suggest that cell-free (cf)DNA methylation could serve as a useful biomarker for improving clinical management. The objective of our study was to identify novel blood-based biomarkers that can be used to predict BC patients at high risk of distant metastasis. For the first time we performed paired-end whole genome bisulfite sequencing (WGBS) using 15 ng of cf plasma DNA from three pools of samples which included metastatic BC to various organs (M, n = 40), breast cancer free survivors (BCF n = 40) and healthy individuals (H, n = 40). Sequences were aligned using the bismark tool and data analysis was conducted using the R package methylKit to identify base-pair resolution DNA methylation differences between groups. A minimum of 5 reads in each group, delta beta values ≥0.2 and p values<0.001 were considered differentially methylated loci (DML). Overall, the data revealed that H and BCF had very similar plasma methylomes with approximately 1792 DML, compared with 0.8 million and 1.1 million DML between HvsM and BFSvsM, respectively. The M pool had greater overall hypomethylation across the genome, particularly in gene bodies compared with H and BFS. However, M had greater CpG island and promoter hypermethylation compared with H and BFS. The average coverage of CpG loci across the genome was 18, 11 and 9 for M, BFS, and H, respectively indicating that plasma of M samples has greater genomic representation. Among the genes differentially methylated in plasma of M was RUNX3, which was hypomethylated overall in M compared to H and BFS except for one CpG island hotspot, where methylation and number of reads significantly exceeded that of H and BFS. Other genes with similar hotspots were also identified. We validated these data using bisulfite amplicon sequencing (Ion Torrent) in each pool and individual plasma samples. Using the HumanMethylation27K bead array, we analyzed 32 breast brain metastasis (BBM) tumor tissues compared to a series of 50 unmatched early stage BC samples and identified RUNX3 hypermethylation in the same hotspot region as WGBS. Lastly, comparing 25 early stage BC samples with known outcomes of BBM had RUNX3 hotspot hypermethylation compared with samples from patients who were BM-free for greater than 5 years. This suggested that prediction of metastasis may occur at the time of early stage of disease. Together these data identified DNA-methylation hotspots in circulation and tissue that are unique to distant MBC and which could be translated into predictive biomarkers of metastasis. Biomarkers indicating a high-risk population can shift current paradigms by allowing increased surveillance and/or prophylactic therapies aimed at preventing metastasis and improving outcomes. Citation Format: Christophe Legendre, Gerald C. Gooden, Kyle N. Johnson, Rae Anne Martinez, Bodour Salhia. Whole genome bisulfite sequencing from plasma of patients with metastatic breast cancer identifies putative biomarkers. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 3825. doi:10.1158/1538-7445.AM2015-3825