Abstract Human cancer tissues are known to release their own DNA into body fluids such as plasma. These DNA molecules derived from tumors, collectively known as circulating tumor DNA (ctDNA), contain genetic information about disease progression at the time of sampling. Importantly, ctDNA has recently been utilized for capturing the whole-body disease profile and when sampled at various time points over the treatment, can inform the dynamic of clonal heterogeneity and its contribution to therapy failures. Metastatic castration-resistant prostate cancer (mCRPC) is a class of cancerous diseases with a high degree of clonal heterogeneity and high mortality which is largely due to the frequent emergence of therapy resistance. Therefore, reconstructing the clonal history for mCRPC diseases using ctDNA has great potential in identification of novel and targetable molecular mechanisms that drive mCRPC progression. To assess this possibility, we hypothesize that tumor clones that are responsible for therapy resistance carry distinct genomic profiles, which are captured in plasma ctDNA and could be computationally resolved through genomic sequencing and clonal reconstruction. To address this question, we obtained plasma cell-free DNA (cfDNA), a mixture of ctDNA and other tissue-derived DNA, from 38 individuals treated with combination PD-L1 and PARP inhibition involved in a recent clinical trial (NCT02484404). Whole-genome sequencing was performed using cfDNA and fragmented buffy coat DNA as germline control. cfDNA samples with a tumor fraction of lower than 10% were excluded. Various computational strategies were used to model the clonal structures of diseases and estimate the temporal order in which small and large structural variants emerged. Through this approach, we observed a negative association between cfDNA tumor fraction and therapy response. Multiple clonal evolutionary patterns leading to therapy failures were observed. For example, both gain and loss of activating androgen receptor mutations were detected in different cases following the treatment, which represents two distinct mechanisms of evading immunotherapy targeting. As a result, changes in the relevant transcriptional activities were also detected through ctDNA fragment-based computational inference. Additionally, clonal persistence was detected in multiple cases that exhibited stable diseases upon the treatment, which involves loss-of-function mutations in p53, ATM, and CDK12. To estimate the potential contribution to plasma ctDNA by metastatic diseases, multi-region whole genome and exome sequencing data of the tumors that are available for some cases prior to the treatment were also analyzed. Clones resolved from analyzing ctDNA were computationally mapped to the primary disease or metastatic tissue with a shared set of mutations in a subset of cases and may explain the patterns of treatment response. Our novel findings from this study will fill in a critical gap of knowledge about the complex genetic mechanisms driving mCRPC immunotherapy resistance. Citation Format: Chennan Li, Anna Baj, Clara C. Y. Seo, Nicholas T. Terrigino, John R. Bright, S. Thomas Hennigan, Isaiah M. King, Scott Wilkinson, Shana Y. Trostel, William D. Figg, William L. Dahut, Jung Min Lee, David Y. Takeda, Fatima Karzai, Adam G. Sowalsky. Tracing the clonal dynamic of metastatic castration-resistant prostate cancer over immunotherapy using circulating tumor DNA [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Translating Cancer Evolution and Data Science: The Next Frontier; 2023 Dec 3-6; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(3 Suppl_2):Abstract nr PR005.
Read full abstract