PERIVASCULAR FIBROBLASTS ARE ASSOCIATED WITH INCREASED TUMOR STIFFNESS AND POOR IMMUNOTHERAPY RESPONSE IN GLIOBLASTOMA PATIENTS

Introduction: Glioblastoma (GBM) is a highly aggressive type of glioma accounting for more than 60% of all brain tumors in adults. Modern therapies have allowed for a modest increase in average survival time; however, GBM remains a terminal diagnosis with poor outcomes due to resistance to approved treatment modalities. Poor therapeutic response is mediated in part by the tumor microenvironment (TME) that harbors aberrant vasculature that promotes hypoxia and immunosuppression, and limits the entry of drugs and effector immune cells. Although much is known about GBM vasculature, the role of rare mural cell types is poorly understood.

Recent studies have uncovered the presence of fibroblasts in the perivascular spaces of the brain. In some extracranial cancers, local tissue-resident fibroblasts are reprogrammed by the tumor into cancer-associated fibroblasts that can remodel extracellular matrix (ECM), thereby accelerating tumor growth and impairing response to immunotherapies. Understanding whether this mechanism is also conserved within the central nervous system (CNS) will reveal novel insights into therapeutic resistance and guide the design of new strategies to improve outcomes in brain tumor patients.

To this end, we used publicly available single cell and bulk RNA-seq datasets to identify CNS fibroblasts in GBM and characterize their gene expression profiles, as well as their clinical correlates. In particular, we sought to answer whether CNS fibroblasts may be responsible for ECM remodeling and tumor stiffening, and whether CNS fibroblasts may hinder response to immune checkpoint blockade (ICB) immunotherapy in GBM patients.

Materials and

Methods: The scRNA-seq data (GSE182109, GSE163577) were obtained from Gene Expression Omnibus. Individual samples were log-normalized and integrated using Seurat’s (v4.3.0.1) reciprocal PCA/ Doublets were identified and removed using Scrublet (v0.2.3) algorithm. Next, cells were clustered and annotated based on previously reported marker genes as well as copy number alterations (CNAs) inferred using CopyKat (v1.1.0) algorithm. To exclude poor quality cells within each cell type, we applied median absolute deviation (MAD)-based outlier detection approach, as described previously. Clusters characterized by low UMI counts, high fraction of mitochondrial reads, and uninformative marker genes were removed. scRNA-seq datasets were integrated using anchor-based Seurat (v3) integration.
Publicly available RNA-seq data were downloaded from TCGA and CGGA databases. Magnetic resonance elastography (MRE) data as well as biopsy-matched RNA-seq data were obtained from GitHub repository (SkabbiVML/stiffR). TCGA and CGGA datasets were were TMM-normalized (edgeR v3.40.2), and logCPM values were used for all downstream analyses. Signature scoring was performed using ssGSEA (GSVA v1.50.0). Cell type deconvolution was performed using MuSiC (c.1.0.0).

Results, Conclusions, and Discussions: We employed a marker-based approach to identify distinct clusters of perivascular fibroblasts (P-FB), meningeal fibroblasts (M-FB), pericytes (PC), and smooth muscle cells (SMC) in GBM (Fig. 1a), which aligned with analogous populations from a human brain vascular atlas (Fig. 1b). Notably, perivascular fibroblasts in GBM exhibited a phenotype similar to cancer-associated fibroblasts (CAFs), expressing markers suggestive of a pro-tumorigenic role.

Analysis of magnetic resonance elastography (MRE) and biopsy-matched RNA-seq data from GBM patients revealed an enrichment of perivascular fibroblast and pericyte gene signatures in stiff tumor biopsies (Fig. 2a-b). Furthermore, retrospective analysis of anti-PD1 immunotherapy trials indicated a significant reduction in overall survival for patients with tumors enriched in perivascular fibroblasts (Fig. 3a-b), indicating a potential role of perivascular fibroblasts in immunotherapeutic resistance.

This study highlights the role of perivascular fibroblasts in ECM remodeling, which is associated with increased tumor stiffness and resistance to immunotherapy. While previous studies identified pericytes as the source of ECM in GBM, our comprehensive gene expression profiling suggests a distinct role for perivascular fibroblasts. This study underscores therapeutic implications of reducing ECM deposition and addressing immunosuppression, with perivascular fibroblasts emerging as potential targets for intervention. However, further research is needed to elucidate the precise mechanisms underlying the resistance to immunotherapy mediated by perivascular fibroblasts.
In conclusion, our findings indicate that perivascular fibroblasts play a significant role in promoting tumor stiffness and immunotherapeutic resistance in GBM, resembling the behavior of CAFs in other cancer types. This study sheds light on the complex interplay between the tumor microenvironment, ECM remodeling, and therapeutic response in GBM, offering potential avenues for novel therapeutic strategies targeting GBM vasculature.

Acknowledgements (Optional): This work was supported in part by NIH K22CA258410 (to M.D.), and BIPH Graduate Fellowship (to M.Z.); X.L. is supported by NIH grant R01CA248033 and DoD grants W81XWH2010312 and W81XWH2010332; Part of the analysis was performed on the high performance computing clusters provided by the Notre Dame Center for Research Computing. We thank K. E. Emblem, R. M. Prins, and R. Rabadan for kindly providing their data.