Chromatin modifications on plasma ctDNA as a tool to phenotype castration-resistant prostate cancer.

Asli Munzur,Joonatan Sipola,Clara Seo,Edmond Michael Kwan,Karan Parekh,Cecily Q. Bernales,Gráinne Donnellan,Ingrid Bloise,Gillian Vandekerkhove,Matti Annala,Corinne Maurice Dror,Kim N. Chi,Cameron Herberts,David Yoshio Takeda,Alexander William Wyatt
DOI: https://doi.org/10.1200/jco.2024.42.4_suppl.214
IF: 45.3
2024-02-01
Journal of Clinical Oncology
Abstract:214 Background: Existing clinical ctDNA assays focus on mutations and copy number changes but may incompletely inform tumor biology. Epigenomic modifications influence mCRPC cell phenotype, and are implicated in treatment resistance and emergence of neuroendocrine prostate cancer (NEPC). Intriguingly, plasma cell-free DNA (cfDNA) can be bound to histones retaining cell-of-origin posttranslational modifications—and can be characterized via chromatin immunoprecipitation followed by sequencing (cfChIP-seq)—potentially simplifying routine tumor epigenomic profiling. Here, we tested the potential for cfChIP-seq to clinically stratify mCRPC. Methods: We analyzed 63 cfDNA samples from 33 pre-treated mCRPC patients (pts) and 19 controls. Pts were selected to represent varied clinical phenotypes, including location and burden of metastases (defined via assessment of bone scintigraphy and computed tomography imaging) and evidence of biopsy-confirmed NEPC. To enhance cfChIP-seq tumor specificity, samples were also pre-selected to include those with high ctDNA fraction (ctDNA%). cfDNA was subjected to H3K4me2 (canonically marking active gene promoters and enhancers) cfChIP-seq plus simultaneous deep targeted sequencing (including matched white blood cells) to inform on driver genotypes. Results: Median age at first cfDNA collection was 69.5 (IQR: 65-72). Prior to first cfDNA collection, 47% of pts had received ≥1 line of AR-targeted therapy and 17% had received ≥1 line of taxane chemotherapy. 51% of pts had >10 bone lesions, 20% had liver, 15% had lung, and 58% had bone metastases without visceral involvement. Frequent TP53 (61% of pts), RB1 (18%), and PTEN (19%; homozygous deletion only) disruption mirrored clinically aggressive disease. Promoter H3K4me2 counts in established (PC) genes (e.g. KLK3, HOXB13) were markedly higher in ctDNA-positive samples than in ctDNA-negative and healthy controls. Conversely, ctDNA-negative samples had comparatively elevated promoter H3K4me2 counts in neutrophil- and leukocyte-related gene sets reflecting the hematopoietic origin of most non-tumor cfDNA. Leveraging a public pan-cancer ATAC-seq atlas, H3K4me2 density was highest in PC-specific open chromatin regions relative to other cancers. H3K4me2 was enriched at AR transcription-factor binding sites except in pts with NEPC. High burden of liver metastases correlated with elevated promoter H3K4me2 counts in liver-associated genes. Finally, KLK3 (encodes prostate specific antigen [PSA]) promoter H3K4me2 count was strongly correlated with time-matched serum PSA (p<0.01) independent of ctDNA%. Conclusions: cfChIP-seq captures mCRPC-specific epigenomic features, indicating a new opportunity for minimally invasive disease phenotyping. cfChIP-seq may augment conventional ctDNA profiling for discovery of predictive and prognostic biomarkers and detection of treatment resistance.
oncology
What problem does this paper attempt to address?