Abstract 6558: Defining differentiation States in well-differentiated and de-differentiated liposarcoma

Abstract

Abstract Among the many sarcoma subtypes, high-grade tumors that have failed to differentiate or undergone dedifferentiation strongly correlate with metastatic potential and poor clinical outcomes. To examine the factors that regulate cell fate, lineage plasticity, and tumor grade, we focused on two common liposarcoma variants, dedifferentiated liposarcoma (DDLS) and well-differentiated liposarcomas (WDLS), that exist at both ends of the differentiation spectrum. Since tumor grading using histology is highly subjective due to widespread tumor heterogeneity, we hypothesized that an sc/snRNA-seq-based approach would be more accurate. To determine if candidate therapies modulate differentiation, we developed a quantitative metric of differentiation using scRNA-seq. The resulting mesenchymal tissue landscape (MTL) is akin to a Waddington landscape but optimized to distinguish DDLS from WDLS. After prospectively obtaining tumors from seven chemo-naïve patients, we performed scRNA-seq on three DDLS and five WDLS specimens, totaling 59,917 cells. Classical markers were used to identify cell types. Since no pre-existing liposarcoma training sets existed, malignant cells were identified by copy number variation (CNV) inferred from the scRNA-seq data. Overexpression of MDM2, CDK4, and HMGA2, known markers of LS, were positive in clusters classified by SingleR as fibroblasts, adipocytes, and myocyte - the mesenchymal subpopulation. We also inferred CNVs and found expected amplifications in 12q13-15 in the mesenchymal subpopulation. Next, we measured cell differentiation by CytoTRACE. A subcluster of malignant fibroblast-like cells positive for MDM2 and CDK4 was the most de-differentiated. Conversely, cells labeled as adipocytes and endothelial cells were the most differentiated cells. Mapping WDLS and DDLS onto the MTL demonstrated that WDLS had more mature cells than DDLS. Surprisingly, while the pathology-assigned diagnosis falls into just two categories, our approach identified a wide range of liposarcoma cells spanning the full spectrum of possible differentiation states. Though follow-on studies are needed, our data suggest that quantitative differentiation metrics can identify malignant cells in transit from a DDLS-to-WDLS state, or possibly the reverse, whereby cells undergo dedifferentiation. Both phenomena are suspected clinically, given the frequent co-existence of WDLS and DDLS in the same tumors, and the noted emergence of DDLS from WDLS diagnosed years earlier. Summarizing our key findings, we measured the degree of differentiation in liposarcoma cells as a precursor to ongoing studies aimed at understanding the molecular mechanisms that underpin LS plasticity and differentiation. Ultimately, we aim to identify the genes and pathways linked to cell fate and discover novel biologically targeted therapies capable of pushing DDLS toward a less aggressive, slower-growing state. Citation Format: Danh Truong, Hannah C. Beird, Chia-Chin Wu, Sandhya Krishnan, Davis Ingram, Alexander Lazar, Emily Keung, Christina Roland, Wantong Yao, Robert Benjamin, Neeta Somaiah, Barry W. Feig, Joseph A. Ludwig. Defining differentiation States in well-differentiated and de-differentiated liposarcoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 6558.