Glioblastoma is a highly aggressive brain tumor characterized by diffuse growth and resistance to therapy. Angiogenesis in glioblastoma is poorly organized and therefore, tends to be associated with tumor cell necrosis, hemorrhage and thrombosis. This leads to the formation of a fibrin-rich extracellular matrix, which could provide important adhesive cues for glioblastoma growth and proliferation.
To establish that blood clotting takes place in the extracellular matrix of malignant brain tumors, we assessed fibrin formation in tumor samples from patients with astrocytoma and glioblastoma using immunohistochemistry. Compared to normal brain tissue, which is essentially fibrin-free, this analysis revealed a marked upregulation of clot formation in the interstitial spaces of patients with high-grade tumors (i.e. astrocytoma 3 and GBM). The low-grade astrocytoma 2, however, expressed 3-3.5 fold less fibrin than was found in tissues from patients with astrocytoma 3 and GBM indicating that the degree of clot formation positively correlates with tumor grade.
Paralleling these data, we found that primary GBM cells, that were freshly isolated from patients after tumor surgery, infiltrated and proliferated strongly after embedding in a three dimensional (3D) matrix of clotted plasma ex vivo. Primary tumor cells from patients with astrocytoma 2 and 3, on the other hand, infiltrated clot but were unable to proliferate in 3D. GBM proliferation in 3D depended on fibrin, which mediated upregulation of the stem cell marker nestin, whereas culturing glioblastoma cells in a 3D matrix of matrigel™ failed to promote nestin expression as well as glioblastoma proliferation. Therefore, these data suggest that the presence of clotted plasma in the tumor extracellular matrix represents a niche for glioblastoma stem cells and, as such, contributes to GBM progression.
To determine the interaction of GBM cells with fibrin on a molecular basis, we transfected GBM cells with siRNA against integrin β3, which completely abolished invadopodia formation and, at the same time, caused a sustained growth inhibition. GBM cell proliferation in 3D fibrin also depended on the formation of a fibronectin matrix as knockdown of fibronectin led to complete growth arrest. These findings appear to be clinically relevant since freshly isolated tumor cells from patients with glioblastoma colonized 3D fibrin most efficiently when they express fibronectin in combination with integrin β3. This suggests that fibrin stimulates adhesive interactions between integrin β3 and fibronectin and that these interactions in turn support glioblastoma stemness.
To assess fibrin formation in glioblastoma in vivo, mice with orthotopic U87MG xenografts were injected intravenously with the fluorescein-coupled dekapeptide CGLKIQKNEC, which is a derivative of the clot-binding peptide CLT1. Using a fluorescence endoscope in situ, we detected strong green fluorescence over the parietal lobe of the right cerebral hemisphere, where tumor growth had been established by MRI beforehand. Subsequently, we confirmed tumor binding of the peptide in isolated brain tissue by fluorescence microscopy ex vivo, which demonstrated specific green fluorescence in the tumor xenograft while adjacent normal brain tissue as well as tissues from distant organs only exhibited background fluorescence.
Together, our data demonstrate a specific upregulation of fibrin in high-grade astrocytoma, which promotes infiltration and proliferation of glioblastoma stem cells via integrin β3 and fibronectin. Moreover, we present a strategy to identify fibrin in the tumor extracellular matrix as a possible means to identify astrocytoma progression in vivo.
Disclosures
Eichler: Novo Nordisk: Membership on an entity's Board of Directors or advisory committees. Pilch:CSL Behring: Other: Grants (investigator initiated), Speakers Bureau; ASPIRE Award/Pfizer: Other: Grants (investigator initiated); Bayer: Consultancy, Speakers Bureau; Roche: Consultancy.
Glioma biopsies Classification Using Raman Spectroscopy and Machine Learning Models on Fresh Tissue Samples
