Single cell RNA sequencing of upper tract urothelial carcinoma to reveal significant heterogeneity of the tumor and immune microenvironment.

484 Background: Upper tract urothelial carcinoma (UTUC) comprises 5-10% of urothelial malignancies but demonstrates unique clinical and molecular characteristics compared to urothelial carcinoma of the bladder. Prior investigations have used bulk profiling of tumor tissue to identify molecular subtypes, classifying the majority of UTUC as luminal and T-cell depleted. However, bulk sequencing does not allow for analysis of the significant heterogeneity known to be present in urothelial tumors. Single-cell RNA sequencing (scRNA-seq) allows examination of intra-tumoral heterogeneity, clonality, and the complex interactions of the immune tumor microenvironment (TME). We sought to apply this technology to better characterize UTUC and the TME. Methods: Single cell RNA sequencing (scRNA-seq) was performed on nine UTUC tissue specimens from six different patients collected fresh via ureteroscopic biopsy using an established institutional process and the 10X Genomics platform. Sequencing reads were normalized and analyzed using R/Seurat package. We assessed the composition of each tumor specimen with known marker genes for molecular subtypes (luminal, basal, squamous, EMT, and claudin-low). We then assessed the composition of immune cells in each specimen using known marker genes. We compared high- and low-grade specimens by subtype composition and immune cell infiltrates. Results: Lineage density analyses demonstrate the intra- and inter-tumoral heterogeneity of the nine endoscopic samples analyzed by molecular subtype composition. There is higher expression of luminal and claudin-low subtypes across all samples. The high-grade samples have higher expression of squamous markers. There is significant heterogeneity of immune cell infiltrates in seven specimens (two specimens were excluded due to low CD45+ cell counts). There is higher macrophage infiltration in high-grade samples, which was the only significant difference (Wilcoxon two-sided p-value = 0.05). Conclusions: This is the first known study using scRNA-seq expression analysis to characterize the notable heterogeneity of high and low-grade UTUC and the associated TME. Lineage density analysis demonstrates high luminal gene expression across samples, which has been demonstrated on prior bulk sequencing studies. The immune TME is also heterogeneous, with notable increased infiltration of macrophages in high-grade disease. There are unique limitations to performing and analyzing scRNA-seq of fresh UTUC tissue specimens, thus data should be interpreted cautiously. However, this study demonstrates the marked heterogeneity of UTUC tumors and frames our current approaches to bulk molecular subtyping of urothelial cancers and immune deconvolution. Further high-resolution studies are needed to characterize UTUC and inform bulk-sequencing efforts.