(P 415) NEOADJUVANT RADIOTHERAPY MODULATES THE TUMOUR MICROENVIRONMENT OF UNDIFFERENTIATED PLEOMORPHIC SARCOMA

Objective: Immunotherapy resistance and tumour heterogeneity in undifferentiated pleomorphic sarcoma (UPS) are driven by an immune-cold tumour microenvironment (TME) with abundant pro-tumour macrophages and low lymphocyte infiltration. While external beam radiation (EBR) can stimulate an immune-hot phenotype in certain contexts, the effects of EBR in UPS remain poorly defined. In Alberta, two neoadjuvant EBR protocols are used depending on the treatment center: 50 Grey EBR in Edmonton, and low-dose doxorubicin with 30 Grey EBR in Calgary. Both protocols are given before surgical resection to shrink the tumour mass, however, their consequences on the TME remain unknown. Here, we aim to quantify the effects of neoadjuvant EBR on both tumour and immune cells using clinical samples, a profiling strategy that accounts for distinct histological regions within tumours, and a novel application of data analysis and integration methods.

Methods: We profiled transcriptomes from 19 UPS patients, treated with either surgery only, the Edmonton, or the Calgary neoadjuvant protocols. To improve coverage of intra-tumoral TME heterogeneity, we separately sampled multiple regions within each tumour, including the invasive margin, cellular tumour near necrosis, proliferative tumour regions, and adjacent muscle tissue. Immune and tumour cell transcriptomes were collected individually using the NanoString GeoMx digital spatial profiling (DSP) platform. To validate the results we further generated global proteomics data from the same samples, and integrated these with a large external bulk RNAseq cohort.

Results: We systematically evaluated the impact of radiotherapy on surgical UPS samples with high spatial resolution across histologically-distinct tumour regions. In Aim 1, differential expression analysis identified radiation-associated genes in both tumour and TME cells with significant changes most dramatic in the tumour margin. We identified and quantified the prevalence of TME components, showing that macrophages have significantly increased abundance in the tumour margin after radiation. These results highlight the tumour margin as a distinct ecosystem in which macrophages play a key role in the radiation response. In Aim 2, We identified gene expression programs and their usage across samples in the 3 treatment groups. Expression programs altered by radiation represented activities related to enhanced antigen presentation, upregulation of cytokine secretion, and reduced Wnt/beta-catenin signalling in the radiated samples. In Aim 3, data integration validated the existence of programs associated with an immune-hot TME, established that radiation is positively associated with this phenotype and with improved patient survival, and that the tumour margin is the site where these programs are primarily enacted. Furthermore, we confirmed a prognostic expression program related to metastatic spread and show that its usage is reduced in patients treated with radiation.

Conclusion: This study broadens our understanding of how radiation impacts UPS biology. Specifically, we identify a prognostic expression program that tumour and immune cells alter in response to high-dose radiation and show that the tumour margin is a key site where additional mechanistic insight would be of value. This work paves the way toward the identification of combination therapies that synergize with radiation to optimize the immune response in UPS.