Introduction:

Bladder cancers exhibit a wide spectrum of morphologic heterogeneity, with one-third of tumors containing regions of urothelial carcinoma adjacent to regions of divergent differentiation or histologic subtypes. Select subtypes are associated with increased risk of metastatic progression and are overrepresented in patients with lethal metastatic disease. The approval of combination enfortumab vedotin (targeting Nectin-4) and pembrolizumab has altered the treatment paradigm in this space, but not all patients will respond to this combination therapy. Moreover, tumors with predominant variant histologic subtypes have historically been excluded from therapeutic clinical trials and large molecular studies of bladder cancer, limiting our understanding of the molecular biology of tumors with divergent differentiation. We therefore performed an analysis of transcriptomic data from a large cohort of bladder tumors displaying phenotypic plasticity to define the molecular features associated with morphologic heterogeneity and characterize the expression of standard-of-care and investigational antibody drug conjugate targets in these tumors.

Methods:

Bladder tumor samples collected at radical cystectomy or transurethral resection were centrally reviewed by a board-certified genitourinary pathologist (H.A.) and histomorphologically classified according to the WHO classification of tumors.  The specimens were microdissected to isolate areas of uniform histomorphology and subsequently underwent bulk whole transcriptome RNA sequencing. We identified 341 specimens from 254 patients; histologic subtypes represented included urothelial carcinoma, not otherwise specified (UC-NOS, N = 130), neuroendocrine (N = 50), squamous (N = 43), glandular (N = 36), sarcomatoid (n = 22), nested (N = 17), poorly differentiated (N = 16), plasmacytoid (N = 14), and micropapillary (N = 13). RNA expression for selected genes of interest was compared between areas of different histomorphology.

Results:

Application of the consensus molecular classification system revealed variation in the distribution of molecular classes between histologic subtypes (Figure 1). Some were composed of a single molecular class in >80% of cases—including sarcomatoid (81% basal/squamous), squamous (89% basal/squamous), neuroendocrine (84% neuroendocrine-like), and adenocarcinoma (100% stroma-rich)—while the remaining subtypes were heterogeneous. Evaluation of transcriptional patterns of markers of divergent differentiation also revealed significant heterogeneity of luminal (e.g. GATA3, PPARG, FOXA1), basal (e.g. KRT5) neuroendocrine (e.g. CHGB, SYP, TUBB2B, NCAM1) and epithelial to mesenchymal transition (e.g. VIM, FOXC2) markers (Figure 2). There was heterogeneity in expression of antibody-drug conjugate targets between different histologic subtypes. NECTIN-4 and TACSTD2 (TROP2) expression was downregulated among sarcomatoid, neuroendocrine, and adenocarcinoma tumors whereas other subtypes (including NOS) showed high expression. ERBB2 expression was low in sarcomatoid, neuroendocrine, adenocarcinoma, squamous, and nested tumors, heterogeneous among NOS, glandular, and plasmacytoid tumors, and high expression among micropapillary tumors.

Conclusion:

In a transcriptional analysis of a large cohort of bladder cancer specimens, we identified significant molecular heterogeneity among different variant histologic subtypes including expression of novel predictive biomarkers. These data highlight key challenges in bladder cancer precision oncology and could be used as a guide for potential rational clinical trial design for patients with variant histology bladder cancer.

Funding: This work was supported by the Mark Foundation, Sidney Kimmel Center for Prostate and Urologic Cancers, the Michael and Zena Wiener for Therapeutics Program in Bladder Cancer, for Survival, the Marie-Josee and Henry R. Kravis Center for Molecular Oncology, P01-CA221757 and SPORE in Bladder Cancer P50-CA221745, and P30-CA008748. It was also supported in part by an Astra-Zeneca ASCO Conquer Cancer Young Investigator Award