Evidence suggests that alternative RNA splicing (AS) plays a critical role in tumor biology and may contribute to the generation of tumor antigens. Here, we develop a method to detect AS in short-read single-cell 5’-RNA-sequencing data, allowing us to uniquely characterize the heterogeneity and dynamic changes in AS in individual cell types within the tumor microenvironment. We identify numerous splicing events specific to either cancer cells or stromal cell types, or for triple-negative versus estrogen receptor-positive breast cancers. By correlating these splice events with expression of splicing regulators in individual cells, we also identify their potential mediators. For instance, we identify and functionally validate the Epithelial Splicing Regulatory Protein-1 (ESRP1) to drive AS in breast cancers responding to immune checkpoint blockade (ICB). Prioritization of splicing events based on their likelihood to represent tumor antigens reveals that their aggregated load also correlates with high immune activity in multiple cancers, while also predicting expansion of T-cells in breast cancers receiving ICB and prolonging long-term survival of cancer patients treated with ICB. Collectively, our method provides a framework for analyzing AS in single-cell data and defines a key role for AS in the response to ICB.

Single-Cell Mapping of Alternative Splicing Linked to Checkpoint Immunotherapy Response. Xiong et al. - 2025 - Nucleic Acids Research

Data and Code Availability

Raw sequencing reads of breast cancer 5’-scRNAseq experiments are accessible from the European Genome-phenome Archive (EGA) (study no. EGAS00001004809, data accession no. EGAD00001006608). The source code of JAseC is available on Zenodo (DOI: 10.5281/zenodo.15848594) and also on GitHub. All detected ASE structures, PSI values, and junction counts for each cell type across BC samples are deposited and freely accessible via Zenodo. The processed counts of scATACseq peaks and stLFR validated ASEs are also available on this webserver. Raw sequencing reads of bulk RNAseq for NOVA1 knockdown in BT549 cells, as well as for ESRP1 knockdown and overexpression in MCF7 cells, are deposited on Gene Expression Omnibus (ID: GSE305636).