Dissection of tumor-intrinsic and tumor-extrinsic features of MiT/TFE translocation renal cell carcinoma via single-cell RNA sequencing

Abstract Background Translocation renal cell carcinoma (tRCC) is a rare and aggressive subtype of renal cell carcinoma driven by oncogenic gene fusions involving MiT/TFE family transcription factors, most commonly TFE3. Currently, there are no molecularly tailored treatments for tRCC, and standard-of-care therapies utilized for other RCC subtypes are typically less effective in tRCC. Emerging data suggest that tRCC is molecularly distinct from more common RCC subtypes. However, an incomplete understanding of both tumor-intrinsic drivers and the tumor microenvironment (TME) features of tRCC presents barriers to developing effective therapeutics for this cancer. Methods We used single-nucleus RNA sequencing (snRNA) to profile ten samples of tRCC tumors and one adjacent normal tissue. Five of the samples also underwent multiome profiling with snATAC-seq while nine of the tumor samples underwent whole genome sequencing and RNA-sequencing. We preprocessed and obtained the raw fasta files using cellranger and extracted RNA counts and chromatin accessibility peaks using the Seurat/Signac algorithm. Scrublet was used to exclude data from droplets containing more than one cell by performing doublet detection and removal on gene-barcode matrices. Cells with fewer than 200 genes detected or more than 5% of counts attributed to mitochondrially-encoded transcripts were removed before across-sample integration. We also excluded genes detected in fewer than three cells across all samples. The combined cohort was analyzed using the MergeData function in Seurat, and cancer cells were selected based on known tRCC gene markers and inferred transcriptional copy number variations estimated via the inferCNV package and verified from WGS data. Non-tumor cell types were determined through manual annotation via known marker genes. We analyzed patterns of gene expression at the single-cell level using the Seurat V4 package and module score functions. Differential expression analysis comparing cells from different clusters or treatment exposure groups was performed using a two-sided Wilcoxon rank-sum test with Bonferroni FDR correction. Results Following quality control and integration, 71,124 single-cells were used for analysis (67,762 from tumor samples) – representing 43,214 tumor cells, 8,352 monocytes, 4,692 T cells, 4,614 endothelial/stromal cells, and normal kidney cells. The normal-adjacent sample represented multiple cell types from the normal kidney, including cells from distal connecting tubule, connecting duct, proximal tubule, thick ascending limb, podocytes, and endothelial cells. Analysis of chromatin accessibility profiles from snATAC-seq and snRNA-seq data of tumor cells was employed to identify unique cell states and a putative cell of origin for tRCCs. Tumor cell analysis also demonstrated intra- and inter-tumoral transcriptional variability in genetically similar tumor subclusters, each with distinct regulon activity. Immune subpopulation analyses revealed higher proportions of resting T cells in treatment-naive tRCC samples and higher proportions of progenitor-exhausted and terminal-exhausted populations in an immune checkpoint inhibitor (ICI) treated sample, indicating immune reprogramming in response to ICI in this mutationally quiet tumor. To further understand the role of, and factors permissive to, anti-tumor T cell responses in tRCC (which is characterized by a low tumor mutational burden) we explored the hypothesis that MiT/TFE fusions may constitute tumor neoantigens. We computationally identified fusion-associated neoantigens corresponding to these tRCC samples, as well as across a larger aggregate dataset of tRCC samples profiled by bulk RNA-Seq. We predicted binding affinities between fusion-associated neoantigens and HLA types and identified multiple peptide candidates with high HLA-binding affinity, which will be validated for their immunogenicity in vitro. Conclusions Overall, our results highlight tumor-intrinsic and tumor-extrinsic determinants of immunogenicity in tRCC and may guide the development of immunotherapeutic strategies with strong mechanistic rationale in tRCC. CDMRP DOD Funding: yes