Abstract 6835: Profiling of new prostate cancer vulnerabilities through single-nucleus RNAseq implicates the CAF-specific TGF-β pathway in tumor progression

Abstract Background: Transcriptomic profiles of primary high grade prostate cancer (hgPCa) have been extensively studied and prognostic gene expression signatures to classify patients for risk have been identified. However, such studies have employed bulk genomic techniques that blur the contributions of the different cell types. Indeed, tumor microenvironment (TME) cells can significantly contribute to these transcriptomic signatures thereby confounding the lineage of potential therapeutic targets. These challenges underscore the potential of single-nucleus (sn) RNAseq to elucidate new PCa vulnerabilities. We set out to assess whether the TME, particularly cancer-associated fibroblasts (CAFs) drive poor outcome in PCa. Methods: We obtained 15 treatment-naïve frozen primary high risk PCa tissue samples, mostly Gleason =8. A PCa nuclei dissociation protocol was adapted for single nucleus snRNAseq (10X Genomics). Sample profiling included all cell populations in the tissue including tumor and TME. Output was analyzed using Seurat (v. 4. 1. 1), corrected for ambient RNA expression with CellBender (v. 0. 3. 1) and detected and removed doublets with Scds (v. 3. 9). Tumor cells were distinguished from normal epithelium cells by a combination of inferCNV (v. 1. 12. 0) and supervised learning with random forests classification. Results: Through snRNAseq, we show that CAFs, and not tumor cells, entirely drive the expression of previously validated poor prognostic signatures (Mortensen, 2015). Our analysis revealed that the TGFβ pathway expression, TGF-β1 in particular, is upregulated in CAFs rather than PCa cells. This pathway, along with poor prognosis signatures are selectively elevated in CAFs from patients with Gleason high compared to low, implicating the TGFb axis as a PCa vulnerability. We also derived a snRNAseq CAF signature that differentiates non-responders to immune checkpoint therapy in mCRPC (Nivolumab+Ipilimumab). Furthermore, cluster profiling of tumor nuclei revealed that hgPCa harbors high heterogeneity with distinct niches of hgPC, mCRPC or neuroendocrine phenotypes for each patient. Conclusions: Through snRNAseq we show that CAFs are the main contributors to prior bulk RNAseq classifiers of poor response, resulting in a refined CAF signature that allows for direct investigation of the contribution of CAFs in other patient segments (mCRPC). Our findings point to a hgPCa patient segment with high CAF representation that may benefit from the addition of a CAF targeting agent (ie. TGFbi) to androgen deprivation therapy. This data has the potential to help refine our prognostic classifiers for patient selection and clinical applicability. Further study of the reported tumor heterogeneity will inform rational combination therapies to tackle resistance mechanisms. Citation Format: Alexandre P. Alloy, Sharmila Chamling Rai, Brendan O'Brien, Anna Lyubetskaya, Arnaud Amzallag, Benjamin Chen, Ana Lako. Profiling of new prostate cancer vulnerabilities through single-nucleus RNAseq implicates the CAF-specific TGF-β pathway in tumor progression abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts) ; 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84 (6Suppl): Abstract nr 6835.