Abstract 3483: Single-nucleus RNA-sequencing of human bladder tumors delineates intra-tumor cellular and subtype heterogeneity

Abstract

Abstract Introduction: Single cell technologies now make it possible to study tumor ecosystems at single cell resolution and may improve our biological understanding of disease aggressiveness and tumor heterogeneity. Whereas single-cell RNA-sequencing of clinical tumor biopsies requires immediate processing after tissue acquisition, single-nucleus RNA-sequencing (snRNA-seq) allows profiling of single nuclei isolated from frozen tumor tissue from patients with long-term follow-up and known clinical outcomes. Methods: We performed snRNA-seq of frozen tumors from 48 bladder cancer (BC) patients (10 Ta, 13 T1, 25 T2-4) using an optimized DroNc-seq protocol. Nuclei were isolated from frozen biopsies using IgePal lysis buffer and droplets were created using the Dolomite Bio platform followed by library generation and sequencing on an Illumina NovaSeq 6000. All epithelial nuclei were classified according to the UROMOL classes of non-muscle-invasive BC or the consensus classes of muscle-invasive BC. Bulk total RNA-sequencing (RNA-seq) was available for 44 of the tumors for comparison. Three tumors were additionally analyzed using 10x Chromium for validation and four tumors were analyzed using 10x Visium Spatial Transcriptomics. Results: After pre-processing the raw sequencing data, we obtained data from 117,653 nuclei in total and 59,201 nuclei remained after quality control filtering (1,233 nuclei per tumor and 529 expressed genes per nuclei on average). We focused our analysis on the epithelial compartment, as it constituted the bulk of the tumors (99% of all nuclei). UMAP visualization and clustering of all tumors were mainly driven by patient origin indicating a high level of inter-tumor heterogeneity. To explore intra-tumor heterogeneity and the association to disease aggressiveness, we characterized the tumors individually using hallmark BC gene signatures. Finally, we explored the composition of transcriptomic classes for each tumor and found that 52% of tumors displayed profound intra-tumor class heterogeneity with less than 70% of all nuclei belonging to a single class. The dominating transcriptomic class of single nuclei was only consistent in 44% of the tumors when compared to the overall transcriptomic class from bulk RNA-seq. This may be explained by several levels of heterogeneity and method differences, including the technical challenges of applying a bulk classifier to single nuclei data. We are currently investigating whether specific epithelial subpopulations are associated to outcome and whether signatures derived from snRNA-seq data can be recovered in bulk RNA-seq data and used as prognostic predictors. Conclusion: Our results highlight the biological complexity of bladder tumors and underline the importance of considering the extent of intra-tumor heterogeneity in the clinical management of BC patients. Citation Format: Sia V. Lindskrog, Sofie S. Schmøkel, Iver Nordentoft, Philippe Lamy, Michael Knudsen, Jørgen B. Jensen, Lars Dyrskjøt. Single-nucleus RNA-sequencing of human bladder tumors delineates intra-tumor cellular and subtype heterogeneity [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3483.