As glioma cells infiltrate the brain they associate with various microanatomic structures such as blood vessels and myelinated white matter tracts. How distinct invasion patterns coordinate tumor growth and influence clinical outcomes remains poorly understood. We have investigated how perivascular invasion affects glioma growth patterning and response to anti-angiogenic therapy within the highly vascularized brain. Orthotopically implanted rodent and human glioma cells are shown to commonly invade and proliferate within brain perivascular space. This form of brain tumor growth and invasion also characterizes de-novo endogenous mouse brain tumors, biopsies of primary human glioblastoma and peripheral cancer metastasis to the human brain. Perivascularly invading brain tumors are vascularized by preexisting normal brain microvessels as individual gliomas cells use perivascular space as a conduit for tumor invasion. Agent-based computational modeling recapitulated biological perivascular glioma growth without the need for neoangiogenesis. We tested the requirement for neoangiogenesis in perivascular glioma empirically by treating animals with the angiogenesis inhibitors bevacizumab and DC101. These inhibitors induced the expected vessel-normalization, yet failed to reduce tumor growth or improve survival of mice bearing orthotopic or endogenous gliomas while exacerbating brain tumor invasion. The results of these studies provide compelling experimental evidence in support of the recently described failure of clinically used anti-angiogenics to extend the overall survival of human patients with newly diagnosed glioblastoma.
To identify molecular factors driving perivascular invasion, we examined genome-wide microarray data from mouse GL26 gliomas harvested from the brains of syngeneic C57BL/6 mice. This analysis revealed that galectin-1 (mLgals1) was the most abundantly expressed mRNA transcript in mouse GL26 glioma. We tested the role of this gene in perivascular invasion based on its known role in mediating cell-extracellular matrix interactions and its correlation with increasing malignancy and poor patient prognosis in clinical glioma. Unexpectedly, shRNA-mediated galectin-1 knockdown did not inhibit perivascular glioma invasion, but led to the complete eradication of intracranial tumor implants in RAG1-/- mice, which grew normally in more severely immunocompromised NOD-scid IL2Rgnull mice. Natural killer cell depletion in both RAG1-/- and C57BL/6J mice permitted the growth of galectin-1-deficient glioma. Orthotopic galectin-1-deficient glioma implants in RAG1-/- mice contained more granzyme B+ cells. Interferon gamma ELISpot assays were used to assess the level of tumor-specific adaptive immunity raised against galectin-1-deficient glioma. These experiments showed that galectin-1-deficient glioma was eradicated prior to, and independent of, adaptive anti-tumor immunity. Further in vitro experiments demonstrated that galectin-1-deficient GL26-Cit glioma cells were approximately 3-fold more sensitive to natural killer cell-mediated tumor lysis compared to their galectin-1 expressing counterparts. These findings suggest that galectin-1 suppression in human glioma cells could improve patient survival by restoring natural killer cell immune surveillance, which we have demonstrated to be capable of eradicating these tumors.
Further experiments were performed to elucidate the mechanism by which galectin-1-deficient glioma is rejected by natural killer cells at the cellular level. Our results indicate that a population of Gr-1+/CD11b+ myeloid cells from blood are crucial for stimulating natural killer cell-mediated eradication of galectin-1-deficient glioma. We show that immunodepletion of Gr-1+ cells in RAG1-/- mice permits galectin-1-deficient glioma growth and significantly reduces the amount of tumor-infiltrating granzyme B+ cells. In vitro experiments demonstrate that monocytic Gr-1+/CD11b+ myeloid cells isolated from peripheral blood upregulate NK1.1 and granyzme B expression in response to the presence of galectin-1-deficient glioma cells, and that these cells enhance natural killer cell-mediated tumor lysis. Since we demonstrate that monocytic Gr-1+/CD11b+ myeloid cells lack tumor lytic potential themselves, we postulate that these cells act by stimulating anti-tumor cytotoxic potential in natural killer cells.
In vivo experiments reveal that 7-fold more Gr-1+/CD11b+ myeloid cells infiltrate galectin-1-deficient gliomas 48 hrs post-tumor implantation compared to gliomas expressing normal levels of galectin-1. On the contrary, the number of tumor-infiltrating natural killer cells is found to be the same between the two tumor types, suggesting that galectin-1 specifically inhibits the entry of Gr-1+/CD11b+ myeloid cells.
Our most recent experiments indicate that glioma-derived galectin-1 acts as a molecular switch for monocytic Gr-1+/CD11b+ myeloid cells to become anti-inflammatory. We demonstrate that the application of recombinant mouse galectin-1 protein onto monocytic Gr-1+/CD11b+ myeloid cells co-cultured with glioma cells and natural killer cells inhibits the enhanced tumor lysis attributed to the addition of the monocytic myeloid cells. This view of galectin-1 as a protein with anti-inflammatory effects on monocytic myeloid cells helps explain why the few Gr-1+/CD11b+ myeloid cells that do infiltrate galectin-1-replete gliomas fail to active natural killer cell-mediated anti-tumor cytotoxicity.
Together, our data suggest that glioma-derived galectin-1 acts to inhibit anti-tumor natural killer cell immune surveillance through a tripartite mechanism: (1.) by blocking the infiltration of Gr-1+/CD11b+ myeloid cells into the tumor microenvironment which stimulate cytotoxic activity in natural killer cells, (2.) by provoking immunosuppressive function in those Gr-1+/CD11b+ myeloid cells that do infiltrate the gal-1-replete tumor microenvironment, thus inhibiting the stimulation of natural killer cells by Gr-1+/CD11b+ myeloid cells, and (3.) by imparting an intrinsic resistance to natural killer cell-mediated tumor cell lysis as demonstrated by us previously. We now plan to perform enzyme-linked immunosorbent assays, quantitative reverse transcription polymerase chain reaction, and genome-wide gene expression profiling on monocytic Gr-1+/CD11b+ myeloid cells to identify genes/proteins responsible for the enhancement of natural killer cell-mediated lysis of galectin-1-deficient glioma. Ultimately, through a deeper understanding of the mechanisms of natural killer cell-mediated anti-tumor immune surveillance, we hope to design novel innate immunotherapeutic strategies to be tested in patients suffering from glioblastoma.