High-throughput global transcriptional profiling to identify the STAT3 signaling pathway as a potential biomarker for immune checkpoint inhibitor resistance in metastatic/advanced urothelial carcinoma.

Abstract

474 Background: Advanced/metastatic urothelial carcinoma (UC) is a significant public health burden with median overall survival of 15 months. Although, immune checkpoint inhibitors (ICI) have provided an additional second line treatment option, only 15-40% of patients will respond. There has been much effort in determining the mechanisms of immunotherapy resistance and predictive biomarkers to further improve these treatments. Methods: Pre-treatment genomic sequencing data derived from FFPE samples from the IMVIGOR210 clinical trial (n=298) was accessed for analysis. Briefly it was a single arm phase II clinical trial where advanced/metastatic UC patients refractory to platinum chemotherapy treatment received the ICI atezolizumab. This study has been published with detailed methods (PMID: 28950298). The raw sequencing data was pre-processed using standard QC measures and aligned to the human reference genome (hg38). The resulting outputs were then normalized and processed to generate the gene level counts for differential gene expression (DGE). We did DGE analysis comparing patients who had clinical benefit (CR, PR, SD) vs non-clinical benefit (ie. PD) to atezolizumab. The list of differentially expressed genes were then analyzed using various gene ontology, pathway and systems biology tools (IPA, Enrichr, and X2Kweb). Further subset analysis was done using gene-gene correlations (ie. PD-L1 and STAT3) and clinicopathologic features (eg. gender, race, smoking history). Results: Among the 298 patients in this study, there were 25 with CR, 43 with PR, 63 with SD, and 167 with PD based on clinical response to atezolizumab. Subgroup analysis for CR vs PD patients found that approximately 847 genes were differentially expressed with statistical significance (p ≤ 0.05). IPA analysis for this list of differentially expressed genes found among the top signaling pathways were “primary immunodeficiency” and “sirtuin signaling”. Further subset analysis of 39 genes (p ≤ 0.01) enriched in PD patients using Enrichr and X2kweb found that there was an overrepresentation of STAT3 signaling genes (hypergeometric p-val 6.32x104). Conclusions: Our results found that when the transcriptional profiles of CR vs PD there was differential gene expression in STAT3, primary immunodeficiency, and sirtuin signaling pathways. Of note it has been reported that STAT3 signaling can modulate immune activity and its expression is correlated with poor prognosis in urothelial carcinoma patients. These results warrant a larger study to see if STAT3 signaling is a potential biomarker for ICI resistance. If validated this may indicate that the STAT3 pathway is a potential therapeutic target to overcome ICI resistance and improve the efficacy of these agents.