Structural landscape of DNA and precision medicine, firm handshake

From the Ha lab, Public Health Sciences and Human Biology Divisions

Most cases of localized or pre-metastatic prostate cancer can be cured by surgically removing the cancerous tissue or treating with radiotherapy. However, this is not true for aggressive metastatic castration-resistant prostate cancer (mCRPC). To understand what drives these aggressive forms of prostate cancer, Dr. Gavin Ha’s lab at the Fred Hutchinson Cancer Center teamed up with Dr. Srinivas Viswanathan and colleagues at the Dana Farber Cancer Institute to take a deeper dive into these tumors to analyze their DNA. This deep dive was spurred by the observation that structural changes to the genomic DNA occur proximal to oncogenes in localized, non-metastatic prostate cancer cells, but these studies did not investigate metastatic prostate cancer. To determine if structural variations in cancer cell DNA, particularly changes in non-coding regions, could explain the differences between localized and metastatic prostate cancer types, Dr. Ha and his lab investigated the presence of alterations in the genomic landscape of both. The research team observed distinct, recurrent breakpoints in the DNA from metastatic cancer cells that occurred close to known oncogenes: androgen receptor (AR), Myc, FOXA1 and LSAMP. These findings prompt a need to further dissect the differences in the genomic architecture of prostate cancer subtypes, as this type of cancer profiling may have translational potential with precision medicine approaches to treat these and other aggressive cancers. This work was recently published in the JCI Insight journal.

The DNA or genetic code within cells includes regions that code for proteins, while other regions are termed “non-coding” and function as structural features of the DNA that regulate gene expression. In prostate cancer, there are many known cancer driver genes that have mutations within the protein coding regions. One prominent example is AR. Despite this pivotal finding of AR and others as drivers of prostate cancer, mutations within these protein coding regions of the genome alone cannot distinguish curative, localized from non-curative, metastatic prostate cancer subtypes. Additionally, the protein coding regions of the genome only make up 1-2% of the genome, begging the question, can alterations to the DNA outside the protein coding region differentiate subcategories of cancers and prioritize alternative targeted therapies for these more aggressive types? “We wanted to better understand the chaotic genetic alterations that drive this deadly disease,” said Dr. Ha. This idea that non-protein coding DNA regions may contribute to prostate cancer outcomes is supported by studies in which changes to the genomic structure have been observed in non-metastatic prostate cancer as well as others, and these alterations occur near to genes with known cancer-causing functions. The work completed by Dr. Ha’s lab extended this investigation to metastatic prostate cancer. Dr. Ha explains their recent work, “This study provides a comprehensive view of the genomic architecture and patterns of DNA breaks that distinguish molecular subclasses of [the lethal] mCRPC [from the localized cancer].” 

To characterize the structural features of the non-protein coding regions in genomes from CRPC cells, the Ha lab performed linked-read whole genome sequencing on 36 metastatic CRPC tumors with normal tissue paired samples. They combined these data with previously published genomic and transcriptomic data for an integrative analysis of 531 localized and 143 metastatic CRPC samples. With these datasets, the authors identified genomic “hotspots” in which DNA structural changes occurred frequently. The researchers focused on hotspots that contained one of the 159 genes known to cause prostate cancer and included the following alterations: nucleotide deletion, tandem duplication, inversion, interchromosomal translocation, intrachromosomal translocation. Each of these features were characterized at DNA sites straddling the coding and non-coding regions (gene transecting events) while only tandem duplication was documented in non-coding regions surrounding a coding region (gene flanking events). From these genomic analyses, “we identified distinct hotspots of genomic rearrangements involving underappreciated oncogenes in mCRPC that differed from localized [non-metastatic] genomes,” said Dr. Ha. Specifically, higher rates of DNA structural variations transecting or flanking AR, FOXA1, MYC and LSAMP genes were observed for the metastatic prostate cancer samples, as compared to localized, non-metastatic prostate cancer samples. These findings suggest that these genes are of potential interest for determining metastatic outcome of prostate cancer and should be investigated mechanistically. Confirming the connection between these changes in the genomic architecture and disease outcomes is needed and would provide support for the use of genomic architectural profiling of tumors to enhance precision medicine treatment strategies.

A) DNA breakpoint hotspots were mapped across all chromosomes for the CRPC samples, and the colors indicate proximity of these breakpoints to known gene drivers of prostate cancer. B) These hotspots of structural variation were compared between primary (non-metastatic) and mCRPC (metastatic) tumor samples and revealed differences in the frequency of breakpoints near known gene drivers of prostate cancer.
A) DNA breakpoint hotspots were mapped across all chromosomes for the CRPC samples, and the colors indicate proximity of these breakpoints to known gene drivers of prostate cancer. B) These hotspots of structural variation were compared between primary (non-metastatic) and mCRPC (metastatic) tumor samples and revealed differences in the frequency of breakpoints near known gene drivers of prostate cancer. Figure taken from primary publication

The deep dive into the genomic structure of metastatic prostate cancer cells revealed many newly prioritized genes of interest for cancer pathology. To confirm that these structural alterations result in changes to the disease outcome of these metastatic prostate cancers, “we have started to reconstruct the genomic architecture of some hotspots using long-read sequencing and optical mapping technologies,” stated Dr. Ha. “We hope that these studies will provide insights into the dysregulated mechanisms that lead to these abnormal genomes.” In the future, Dr. Ha plans to conduct “studies using larger cohorts to validate these new genome rearrangement signatures, particularly with clinical outcomes.”

The spotlighted research was funded by the National Institutes of Health, the Department of Defense Prostate Cancer Research Program, PCF Young Investigator awards, PCF-Movember Challenge Award, Brotman Baty Institute for Precision Medicine, the Fund for Innovation in Cancer Informatics Major Grant, the B Foundation Scholar Grant, the Wong Family Award in Translational Oncology and Dana-Farber Cancer Institute Medical Oncology grant, the Pan-Mass Challenge team IMAGEINE and American Cancer Society Research Professor grant.

Fred Hutch/University of Washington/Seattle Children's Cancer Consortium members Gavin Ha and Peter Nelson contributed to this work.

Zhou M, Ko M, Hoge AC, Luu K, Liu Y, Russell ML, Hannon WW, Zhang Z, Carrot-Zhang J, Beroukhim R, Van Allen EM, Choudhury AD, Nelson PS, Freedman ML, Taplin ME, Meyerson M, Viswanathan SR, Ha G. 2022. Patterns of structural variation define prostate cancer across disease states. JCI Insight. 7(17):e161370.