(P 433) TUMOR HYPOXIA AND THE RISK OF METASTASIS IN HIGH-RISK LOCALIZED UPS/MFS

Objective: Hypoxia activates genetic programs that correlate with more aggressive clinical phenotypes, specifically metastasis. Undifferentiated Pleomorphic sarcomas (UPS) and Myxofibrosarcoma (MFS) are genetically complex, adult sarcomas that have high rates of distant metastasis (DM). Our group has demonstrated using FAZA/PET MRI imaging, that tumour hypoxia correlates with increased DM risk in high-risk localized UPS/MFS. This study therefore aims to delineate the mechanisms by which hypoxia contributes to metastasis in this clinical context.

Methods: Samples were collected from 20 patients with high-risk UPS/MFS enrolled onto our Phase 2 Clinical Trial (NCT03418818) at Princess Margaret Cancer Centre and Mount Sinai Hospitals, in which pimonidazole (PIMO) was administered 16-20 hours prior to surgery to allow for differentiation of hypoxic (PIMO+) and normoxic (PIMO-) cells from primary tumors. Furthermore, primary patient derived cell lines (MFS, n=4 and UPS n=3) were cultured under hypoxic (0.2% O2) and normoxic (18% O2) conditions for 72 hours, after which bulk RNA sequencing was performed to characterize hypoxic signatures. Additionally, DNA methylation was performed on hypoxic vs. non-hypoxic cells by generating single cell suspensions from primary tumor samples. These samples were freshly dissociated and sorted into PIMO+ and PIMO- cells. DNA was then extracted and analyzed using the Illumina Infinium Methylation EPIC array. Furthermore, to assess differential gene expression in patient tumors based on PIMO staining, we used spatial transcriptomic analyses of FFPE sections using the 10x Genomics Visuim platform. Analyses included the ChAMP package for differential methylation and the EdgeR package for differential gene expression from bulk RNA-seq. For the spatial transcriptomic analysis, samples were analyzed in Python using functions from the Scanpy and Squidpy packages.

Results: Analysis of UPS/MFS(n=7) cell lines exposed to hypoxic conditions identified pathways related to development, and demethylase activity DNA methylation analysis revealed differences between hypoxic and normoxic cells, with these differences primarily observed in pathways related to development. These pathways include embryonic skeletal system development, embryonic organ development and skeleton morphogenesis. Spatial transcriptomic analysis differentiated between PIMO-positive and PIMO-negative regions using a distinctive gene signature, including several genes known to be associated with hypoxia such SPP1, FCGR2B, and TREMI.

Conclusion: Our initial data suggests that gene expression signatures distinguish between hypoxic versus normoxic tumor regions from primary patient UPS or MFS. Using both in vitro hypoxia exposed UPS/MFS cell lines or freshly dissociated UPS/MFS patient tumours, we observed differential regulation of genetic pathways related to development and methylation changes. In ongoing studies, we aim to develop a hypoxia gene signature specific to UPS/MFS to facilitate patient stratification based on hypoxic tumor status and to further investigate the pathobiology of how hypoxia correlates to DM risk.