5-hydroxymethylcytosine as a liquid biopsy biomarker in mCRPC.

Martin Sjöström,Shuang Zhao,Eric Jay Small,Yuhong Ning,Corinne Maurice-Dror,Adam Foye,Jun Jie T. Hua,Haolong Li,Tomasz M. Beer,Christopher P. Evans,Matthew Rettig,Kim N. Chi,Joshi J. Alumkal,Rahul Raj Aggarwal,Alan Ashworth,Samuel Levy,Housheng H. He,Alexander W. Wyatt,David A. Quigley,Felix Y Feng
DOI: https://doi.org/10.1200/jco.2021.39.6_suppl.148
IF: 45.3
2021-02-20
Journal of Clinical Oncology
Abstract:148 Background: 5-hydroxymethylcytosine (5hmC) is an epigenetic modification which regulates gene expression and is associated with active transcription. The optimization of 5hmC sequencing in cell-free DNA (cfDNA) could therefore enable assessment of gene activity through a liquid biopsy. We aimed to investigate the 5hmC landscape of metastatic castration-resistant prostate cancer (mCRPC) and to evaluate the potential of 5hmC modifications in cfDNA as biomarkers of outcome in mCRPC patients. Methods: Genome-wide 5hmC modifications were analyzed with a low-input whole-genome 5hmC sequencing method based on selective chemical labeling in DNA from 93 mCRPC tissue biopsies previously profiled with whole-genome sequencing (WGS), RNA-sequencing and whole-genome bisulfite sequencing (WGBS). In addition, we analyzed 64 cell-free DNA (cfDNA) samples, from men with mCRPC before first-line abiraterone or enzalutamide, with both 5hmC sequencing and a conventional targeted ctDNA panel assessing common genomic alterations. Results: In mCRPC tissue samples, 5hmC enrichment was more strongly associated with gene expression than promoter methylation or copy number. Among cancer hallmark pathways, the androgen response genes had the strongest association between 5hmC and gene expression, suggesting a disease specific marking of gene activation. 5hmC patterns in cfDNA could be used to estimate the circulating tumor DNA fraction (ct-fraction), which was prognostic for overall survival (tertiles of ct-fraction, HR = 1.6 95%CI 1.1-2.3, p = 0.007). Further, 5hmC levels were indicative of gain of oncogene activity (such as AR, MYC, and PIK3CA) and loss of tumor suppressor gene activity (such as RB1, TP53 and BRCA2). The number of alterations, by 5hmC levels, of common drivers of mCRPC was prognostic for overall survival, also after adjusting for ct-fraction (adjusted p = 0.00001), and the prognostic value of common alterations detected by 5hmC sequencing versus conventional targeted ctDNA sequencing was similar. Finally, 5hmC levels in cfDNA of genes not significantly altered by copy number gain or loss (and thus not routinely included in targeted ctDNA sequencing assays), such as TOP2A and EZH2, identified a high-risk subgroup of mCRPC, which was highly prognostic for overall survival independent of ct-fraction (adjusted HR = 1.8 95%CI 1.2-2.8, p = 0.007). Conclusions: 5hmC in mCRPC tissue demonstrated an association with gene expression that was highest for prostate cancer driver genes, highlighting the ability to track disease-specific biology. 5hmC in cfDNA from men with mCRPC can be used to estimate the ct-fraction of the sample, infer activity gain and loss of common drivers of mCRPC, and identify high-risk groups of mCRPC based on alterations not commonly detected with conventional ctDNA sequencing, showing its potential as a liquid biomarker. Further studies are aimed at optimizing and validating 5hmC-based biomarkers in larger cohorts.
oncology
What problem does this paper attempt to address?