Immune microenvironment of experimental rat C6 gliomas resembles human glioblastomas

Abstract

Glioblastoma (GBM) is the most aggressive primary brain tumor, with ineffective anti-tumor responses and a poor prognosis despite aggressive treatments. GBM immune microenvironment is heterogenous and activation of specific immune populations in GBM is not fully characterized. Reliable animal models are critical for defining mechanisms of anti-tumor immunity. First we analyzed the immune subpopulations present in rat C6 gliomas. Using flow cytometry we determined kinetics of infiltration of myeloid cells and T lymphocytes into glioma-bearing brains. We found significant increases of the amoeboid, pro-tumorigenic microglia/macrophages, T helper (Th) and T regulatory (Treg) cells in tumor-bearing brains, and rare infiltrating T cytotoxic (Tc) cells. Transcriptomic analyses of glioma-bearing hemispheres revealed overexpression of invasion and immunosuppression-related genes, reflecting the immunosuppressive microenvironment. Microglia, sorted as CD11b⁺CD45^low cells from gliomas, displayed the pro-invasive and immunosuppressive type of activation. Accumulation of Th and Treg cells combined with the reduced presence of Tc lymphocytes in rat gliomas may result in the lack of effective anti-tumor responses. Transcriptional profiles of CD11b⁺ cells and composition of immune infiltrates in C6 gliomas indicate that rat C6 gliomas employ similar immune system evasion strategies as human GBMs.

Figure 1

Evaluation of the content of microglia, blood-derived macrophages and leukocytes in C6 rat gliomas. (A) Percentages of microglia, blood-derived macrophages and leukocytes in brains of sham-operated and C6 rat glioma bearing animals at day 21^st after implantation. Lower panel shows gating strategy and representative results of FACS analysis. Quadrant gates were drawn on three cell subpopulations based on differences in the surface expression of CD11b and CD45 antigens: microglia (CD11b⁺CD45^low), blood-derived macrophages (CD11b⁺CD45^high) and leukocytes (CD11b^-CD45^high). Leukocytes CD11b^-CD45^high and CD11b⁺CD45^high invading macrophages/monocytes were mostly absent in sham-operated brain samples. (B) Kinetics of accumulation of microglia, peripheral macrophages and leukocytes within the tumor. Changes in cell type content at 8^th, 14^th and 21^st day after implantation were calculated as content of each subpopulations in glioma bearing brains related to its content in sham-operated animals (N = 4–6 per time point). (C) Double staining of Iba1 + and Arginase 1 shows accumulation and the pro-tumorigenic activation of GAMs only in tumor-bearing hemispheres (N = 3–5 rats per group).

Figure 2

Global gene expression profiling of rat C6 gliomas shows overrepresentation of the immune activation signature genes. (A) Pie chart diagram shows numbers of significantly changed genes. Global gene expression in brains of naïve rats (n = 4) and glioma bearing brains (n = 3) was determined using Affymetrix microarrays hybridization. (B) For the heatmap generation, a log2 expression profile of each gene (each row) was centred on the average value in the control group, so the color indicates the log2(ratio) change in expression relative to the control and fold. (C) Gene ontology (GO) enrichment analysis was performed on all >2 fold upregulated genes that clustered into the glioma-regulated modules to identify overrepresented GO terms. GO terms were grouped according to a number of genes in given category.

Figure 3

Gene expression profiling of C6 gliomas infiltrating microglia shows markers of the pro-invasive, immunosuppressive activation. (A) Pie chart diagram shows numbers of significantly changed genes. Global gene expression in microglia (CD11b⁺CD45^low) immunosorted from brains of naïve rats (n = 4) and glioma-bearing brains (n = 4) was determined using Affimetrix microarrays hybridization. (B) For the heatmap generation, a log2 expression profile of each gene (each row) was centred on the average value in the control group, so the color indicates the log2(ratio) change in expression relative to the control. Gene ontology (GO) enrichment analysis was performed on all >2 fold upregulated genes that clustered into the glioma-regulated modules to identify overrepresented GO terms. GO terms were grouped according to a number of genes in given category.

Figure 4

Characterization of expression pattern of known M1/M2 markers in microglia infiltrating C6 gliomas shows upregulation of cell cycle, invasion, immunomodulation and immunosuppression related genes. Global gene expression in microglia (CD11b⁺CD45^low) immunosorted from brains of naïve rats (n = 4) and glioma-bearing brains (n = 4) was determined using Affymetrix microarrays hybridization. Heatmaps show a log2 expression profile of each gene (each row) which was centred on the average value in the control group, so the color indicates the log2(ratio) change in expression relative to the control; a first column indicates gene name, a second column shows fold change and a third column represents p-value. M1/M2 markers were selected based on guidelines described by Murray et al. 2015; GAMs are genes upregulated in CD11b⁺ infiltrating murine GL261 gliomas (Szulzewsky et al. 2015).

Figure 5

Accumulation of MDSCs and T cell subpopulations in rat C6 gliomas. (A) Representative graphs show the analysis of percentages of T regulatory (FOXP3⁺), T helper (CD4⁺), T cytotoxic (CD8⁺) lymphocytes and MDSCs (GR1⁺ within CD11b⁺ cell population) in tumor-bearing hemispheres at the 21st day after C6 glioma cell implantation. Histograms from different subsets isolated from gliomas (dark lines), sham groups (light lines) and IgG control (dotted lines). (B) Quantification of percentage of specific cells among all counted cells isolated from sham-operated and tumor-bearing hemispheres at the 21st day after C6 glioma cell implantation (N ≥ 6 per group). (C) Quantification of percentage of specific cells among all counted cells isolated from peripheral blood of sham-operated and tumor-bearing animals (N ≥ 6 per group).

Figure 6

Comparison of gene expression in C6 gliomas and GBM subtypes. Bars represent average correlation coefficients and whiskers show standard deviation (n = 3). The P-values were computed using two-sided Mann-Whitney test.