Single-Cell Transcriptomics Resolves Cellular Heterogeneity And Oncogenic Networks During Malignant Transformation Of Peripheral Nerve Sheath Tumors

Abstract Malignant peripheral nerve sheath tumors (MPNSTs) are highly aggressive Schwann cell (SC)-derived sarcomas with strong metastatic proclivity and resistance to radiation and chemotherapy. Molecular events driving SC-to-MPNST transformation are incompletely understood. MPNST are heterogeneous tumors with a distinct microenvironment that contributes to the establishment of a tumor niche. Cancer behaviors including initiation, progression, metastasis and relapse cannot be fully defined by genetic mutations alone, but are critically dependent on intercellular communication between tumor cells and the microenvironment. However, currently, a comprehensive understanding of tumor cell lineage evolution, heterogeneity and cell-cell communication network in MPNST is lacking. A better understanding of cellular heterogeneity and tumor-initiating cells during tumorigenesis at a single-cell level will facilitate identifying molecular and cellular determinants underlying malignant transformation in MPNST. We have established a murine model of aggressive peripheral nerve malignancy with HIPPO-TAZ/YAP activation (Lats1/2-def mice), which provides a unique opportunity for dissecting mechanisms of SC malignant transformation. To identify potential PNST-driven progenitor populations, we compare single-cell transcriptome profiles of Lats1/2-def sciatic nerve tumors during pre-neoplastic and advanced stages. Our data reveal that early sciatic nerve tumors contain primarily mature SCs, whereas full-blown tumors harbor neoplastic SCs composed of different subsets: fast-proliferating, mesenchymal-like and neural crest stem cell-like cells, indicative of a developmental trajectory from differentiated SCs to tumor-like SCs with distinct molecular and transcriptional signatures. Interestingly, transformed SCs exhibit inflammation gene signature, which may suggest cell-autonomous inflammation-driven reprogramming that drives tumorigenesis. To characterize cellular heterogeneity in MPNST-like tumors, we further identify cell-cell communication networks between SC-derived tumor cells and cancer-associated stromal cells within the tumor microenvironment. In summary, our studies using single-cell transcriptomics in clinically relevant mouse models of PNST reveal tumor cellular heterogeneity and the microenvironment niche for driving SC tumorigenesis. It will provide principles for identifying potential targets or treatment strategies to cure MPNST.