Development of a Reactive Stroma Gene Signature (RSGS) Score to Characterize the Longitudinal Cellular and Molecular Alterations in the Tissue Microenvironment of Locally Advanced Cervical Cancer in Response to Radiotherapy at Single-Cell Resolution

Abstract

<h3>Purpose/Objective(s)</h3> Understanding how the tumor-associated tissue microenvironment impacts and reflects the response and effectiveness of chemoradiotherapy in cervical and vaginal cancer will open avenues for future clinical interventions and appropriate selection of radiation dose delivered via brachytherapy. <h3>Materials/Methods</h3> Biopsies of tumor and adjacent normal tissue from 7 cervical and vaginal cancer patients were collected longitudinally at baseline and after 45 Gy of ionizing radiation (IR) and analyzed by single-cell RNA sequencing to better understand the tumor, immune, and stromal determinants of response and resistance. Five of the 7 patients had biopsies available at 3 months and were included in this analysis. The single-cell transcriptomes of 4 cervical cancer patients (adenocarcinomas N=3, squamous cell carcinoma N=1) and one patient diagnosed with vaginal cancer (squamous cell carcinoma) were characterized. <h3>Results</h3> Of 5 patients evaluated at 3 months following high-dose-rate brachytherapy, 3 had a complete response, one had persistent disease and one had distant relapse in the peritoneum. Collectively, over 70,000 cells were analyzed in this study. We compared the single-cell transcriptomes of 2 poor responders versus 3 patients with a complete clinical response. The tumor tissue of poor responders had a molecular and cellular signature indicative of a reactive stroma (VCAM1+ fibroblasts, CXCR4+ endothelial cells, and elevated expression of select collagens), with infiltration of immunosuppressive lymphoid and myeloid cells. We developed a reactive stroma gene signature (RSGS) score, which was enriched in cells from irradiated tumor biopsies collected from subjects with poor outcomes. In contrast, the RSGS score was consistently low in normal biopsies at both baseline and post-IR across all subjects, with one exception. The subject with distant relapse showed RSGS score enrichment in the normal post-IR biopsy. When evaluating the cell proportions of immune cell populations, we observed that the abundance of mast cells was consistently lower (< 0.05%) in normal biopsies at both baseline and post-IR. Likewise, the proportion of mast cells in post-IR biopsies was also low in subjects with complete responses. However, in subjects with poor outcomes, the mast cell abundance increased at least three-fold compared to matched baseline tumor biopsies. <h3>Conclusion</h3> Importantly, the findings stemming from this study suggest that treatment-refractory disease upregulated tumor-promoting stromal and immune responses after initiating radiotherapy that was not observed in the tumor microenvironments of complete clinical responders; thus nominating molecular factors that can potentially be therapeutically exploited to improve the efficacy of radiation-based therapies in patients diagnosed with cervical and vaginal cancer. Future studies with larger cohorts enabled by more scalable in situ approaches will be necessary to assess the generalizability of these findings.