Characterization of Castration-Resistant Prostate Cancer Subtypes Through Their Genomes and Epigenomes
Access to this document is restricted. Some items have been embargoed at the request of the author, but will be made publicly available after the "No Access Until" date.
During the embargo period, you may request access to the item by clicking the link to the restricted file(s) and completing the request form. If we have contact information for a Cornell author, we will contact the author and request permission to provide access. If we do not have contact information for a Cornell author, or the author denies or does not respond to our inquiry, we will not be able to provide access. For more information, review our policies for restricted content.
Castration-resistant prostate cancer (CRPC) is a highly plastic disease exhibiting two histological classes and four epigenomic subtypes. By analyzing tissue RNA-seq and cell- free DNA (cfDNA) whole-genome sequencing (WGS) data from 486 patients, we find significant heterogeneity of molecular subtypes in both histological classes, with CRPC- SCL (stem cell-like) generally co-existing with other subtypes and enriched in bone metastases. CRPC-SCL tumors show upregulation of YAP/TAZ target genes, and WGS of tissue and cfDNA show enrichment of genomic alterations in the upstream genes, revealing novel associations of genomic events with lineage plasticity. Specifically, we find structural variants on chr4, supported by matched Hi-C data from patient samples, that decrease the access of MOB1B promoter, a YAP/TAZ pathway inhibitor, to its enhancers. Integration of genomic events in the pathway using a classifier allows prediction of CRPC-SCL tumors with 79% accuracy. We find that lineage plasticity in CRPC tends to be associated with high genomic heterogeneity. Our study shows that molecular subtype annotations can aid therapeutic decisions, and joint inference of epigenomic state and genomic variants can reveal novel biology. Moreover, leveraging the rich information of chromatin accessibility inherent in cfDNA also allowed us to build a computational model to predict gene expression from coverage signals and fragmentation patterns. This approach, coined cfOncoPath, provides a method to predict oncogenic pathway activity and outperforms other existing methods of gene expression prediction. Since cfOncoPath relies on cfDNA alone, which can be collected in a minimally invasive fashion, it can be easily translated for multiple clinical uses, including monitoring disease progression, especially when collection of tissue biopsies is infeasible.