Search

Tumor cell extensions called tumor microtubes (TMs) in glioma resemble neurites during neurodevelopment and connect glioma cells to a network that has considerable relevance for tumor progression and therapy resistance. The determination of interconnectivity in individual tumors has been challenging and the impact of tumor cell connectivity on patient survival remained unresolved so far. Here, a connectivity signature from single-cell RNA-sequenced (scRNA-Seq) xenografted primary glioblastoma (GB) cells was established and clinically validated. Thirty-four of 40 connectivity genes were related to neurogenesis, neural tube development or glioma progression, including the TM-network-relevant GAP43 gene. Astrocytic-like and mesenchymal-like GB cells had the highest connectivity signature scores in scRNA-Seq data of patient-derived xenografts and patient samples. In 230 human GBs, high connectivity correlated with the mesenchymal expression subtype, TP53 wildtype, and with dismal patient survival. CHI3L1 was identified as a robust molecular marker of connectivity. Thus, the connectivity signature allows novel insights into brain tumor biology, provides a proof-of-principle that tumor cell connectivity is relevant for patients’ prognosis, and serves as a robust biomarker that can be used for future clinical trials.Statement of significanceIntegration of GB cells into functional networks drives tumor progression and resistance. Here, we established and validated a novel connectivity gene expression signature of single GB cells and whole tumors that can be easily applied to clinical and preclinical samples. It is shown that connectivity is determining prognosis combining molecular, functional and clinical insights into the disease.

Recent studies have demonstrated extensive cell-to-cell connectivity between tumor cells of gliomas with considerable relevance for tumor progression and therapy resistance. Tumor microtubes (TMs) are neurite-like tumor cell extensions that build these tumor cell networks. Measuring the extent of connectivity in individual tumors has been challenging and depended on anatomical parameters that are difficult to evaluate in patient samples. We performed bulk and single-cell (sc)RNA sequencing of connected vs. unconnected tumor cells from patient-derived xenograft tumors using a newly developed technology that exploits SR101 dye transfer within tumor cell networks. scRNA sequencing was performed with 17 human glioblastoma tumor samples. Three diffuse glioma cohorts from The Cancer Genome Atlas (n = 648), the Chinese Glioma Genome Atlas (n = 668) and the NCT Neuro Master Match (n = 38, IDH-wildtype only) were used to assess clinical properties. A connectivity signature both from bulk and scRNA sequencing data of xenografted primary glioblastoma tumor cells was established. Comparative analysis showed better performance and higher biological relevance of the single-cell derived signature that involves 71 genes. Most of the genes are related to neurogenesis and neural tube development, including several previously recognized TM-relevant genes. Highest connectivity was observed in astrocytic-like and mesenchymal-like tumor cells. Induction of connectivity in vitro was accompanied with increase of the connectivity signature. The connectivity signature was higher in astrocytic as compared to oligodendrocytic gliomas, and highest in IDH-wildtype gliomas. In accordance, connectivity correlated strongly with dismal survival in all three glioma cohorts. The connectivity signature established here is biologically plausible and associates with prognostically relevant glioma subtypes. It provides the first proof-of-principle that tumor cell connectivity is relevant for the clinical course of patients with gliomas, and at the same time serves as a robust biomarker that can be used for future studies, including prospective clinical trials.