Targeting HIF-2α in glioblastoma reshapes the immune infiltrate and enhances response to immune checkpoint blockade

Abstract

Glioblastoma (GBM) is an aggressive primary brain tumor with dismal clinical prognosis and resistance to current therapies. GBM progression is facilitated by the tumor microenvironment (TME), with an immune infiltrate dominated by tumor-associated microglia/macrophages (TAMs) and regulatory T cells (Tregs). The TME is also characterized by hypoxia and the expression of hypoxia-inducible factors (HIFs), with HIF-2α emerging as a potential regulator of tumor progression. However, its role in GBM immunosuppression remains unknown. Here, we investigate HIF-2α and the use of the HIF-2α inhibitor PT2385 to modulate the TME in the immunocompetent GL261 mouse GBM model. PT2385 administration in vivo decreased tumor volume and prolonged survival of tumor-bearing mice, without affecting GL261 viability in vitro. Notably, HIF-2α inhibition alleviated the immunosuppressive TME and synergized with immune checkpoint blockade (ICB) using αPD-1 and αTIM-3 antibodies to promote long-term survival. Comprehensive analysis of the immune infiltrate through single-cell RNA sequencing and flow cytometry revealed that combining PT2385 with ICB reduced numbers of pro-tumoral macrophages and Tregs while increasing numbers of microglia, with a corresponding transcriptional modulation towards an anti-tumoral profile of these TAMs. In vitro, deletion of HIF-2α in microglia impeded their polarization towards a pro-tumoral M2-like profile, and its inhibition impaired Treg migration. Our results show that targeting HIF-2α can switch an immunosuppressive TME towards one that favors a robust and sustained response to ICB based immunotherapy. These findings establish that clinically relevant HIF-2α inhibitors should be explored not only in malignancies with defects in the HIF-2α axis, but also in those exhibiting an immunosuppressive TME that limits immunotherapy responsiveness.

Keywords: Brain tumors; Glioma; HIF-2; Hypoxia inducible factor; Immune checkpoint inhibitors; Immunotherapy; Microglia; Treg; Tumor microenvironment; Tumor-associated macrophage.

Fig. 1

HIF-2α inhibition prolongs survival of GL261-bearing mice and decreases tumor volume. (a) Treatment schedule of mice implanted intracranially with GL261 cells. (b) Kaplan-Meier survival curve of mice administered with vehicle (n=10 females) and PT2385 (n=12 females). ms: median survival. Statistical analysis used the Log-rank Mantel-Cox test. (c–h) Proportions of lymphoid cells (CD4⁺ and CD8⁺ T cells; andCD56⁺CD3^- NK cells) and myeloid cells (dendritic cells: CD11c⁺IA-IE⁺; classical monocytes: Ly6C⁺Ly6G^- and non-classical monocytes: Ly6C⁺Ly6G^-) in the peripheral blood of vehicle (black) and PT2385 (pink)-treated mice at days 7, 14 and 21 post-implantation. Graphs show mean ± SD, n=6 per group; p-values calculated using a paired t-test. MN: monocytes. (i–l) Cytokine quantification in plasma collected from peripheral blood at day 21 post-implantation. Graphs show mean ± SD, n=12 per group; p-values calculated using an unpaired t-test. (m) Tumor volume (mm³) of vehicle (black) and PT2385-treated (pink) mice at days 7, 14 and 21 post-implantation. Tumor volume was calculated by the area of sequential coronal plane images. Graph shows mean ± SD, n=6 mice per group; p-values calculated using Tukey’s multiple comparison test. (n) Representative images of the coronal plane of T1 weighted MRI scans of vehicle and PT2385-treated mice at days 7, 14 and 21 post-implantation

Fig. 2

HIF-2α inhibition impacts infiltration and transcriptome signature of microglia and macrophages, and its knockout prevents M2 microglial polarization. (a, d) Representative flow cytometry plots of macrophages (CD49d⁺TMEM119⁻) and microglia (CD49d⁻TMEM119⁺) gated from live CD45⁺CD11b⁺CD3⁻ cells in vehicle- and PT2385-treated mice. (b, c) Quantification of macrophage and microglia populations at midterm (green, n = 4 per group). (e, f) Quantification of macrophage and microglia populations at endterm (pink, n = 6 per group). Graphs show mean ± SD. p-values were calculated using an unpaired t-test. (e–f) Vulcano plots of differentially expressed genes (DEG) in sorted microglia and macrophages from vehicle-treated mice versus PT2385-treated mice (Red: Padj < 0.05, LFC > 0.5 or LFC < -0.5, green: LFC > 0.5 or LFC < -0.5 and grey: not significant). (g) Quantification of Epas1 (blue) and Hif1a (brown) mRNA by qRT-PCR, normalized to two housekeeping genes (Gapdh and Actb). The graph shows mean ± SD, n=3; statistics used a two-way ANOVA with Tukey’s multiple comparisons test. (h) Heatmap of M1- and M2-like genes quantified by qRT-PCR. Graphs show relative expression normalized to two housekeeping genes (Gapdh and Actb), represented by the median, n=3; p-values calculated using a two-way ANOVA with Tukey’s multiple comparisons test

Fig. 3

HIF-2α inhibition decreases CD4⁺ Tregs and the CD8:Treg ratio correlates with survival. (a–d) Flow cytometry quantification of CD8⁺ and CD4⁺ T cells at midterm (green) and endterm (pink). (e–f) Treg (CD25⁺FOXP3⁺) proportions at midterm and endterm gated on CD4⁺ T cells. (g) Representative plot of Treg distribution at endterm of vehicle and PT2385-treated mice. Live cells were gated on CD45⁺CD11b^-CD3⁺. Graphs show mean ± SD, n=4 per group (midterm), n=6 per group (endterm); p-values calculated using an unpaired t-test. (h, i) CD8⁺ T cell: Treg cell ratio. Box plots line show mean, n=4 (midterm), n=6 (endterm); p-values calculated by an unpaired t-test. (j) Pearson correlation between days of survival and the CD8:Treg ratio for each mouse, the black line represents the mean of simple regression

Fig. 4

HIF-2α inhibition and ICB act synergistically to promote long-term survival of GL261-bearing mice. (a) Treatment schedule of C57BL/6 mice bearing orthotopic GL261 glioma. PT2385 (10 mg/kg) was administered from day 7 post-implantation (p.i.) for 21 days by oral gavage, every day, twice per day. αPD-1 (200 ug) or isotype control antibody or was administered intraperitoneally (i.p.) every 3 days from day 9 p.i. αTIM-3 (250 ug) or isotype control antibody or was administered i.p. every 3 days from day 14 p.i. (b) Kaplan-Meier survival curve of mice treated with vehicle (all dashed lines) or PT2385 (all solid lines) in combination with isotype control antibodies (back lines), αTIM-3 alone (yellow lines), αPD-1 alone (blue lines), or in combination with αTIM-3 and αPD-1 (pink lines) antibodies. All mice were females. Sample sizes: vehicle group (n = 11), all other groups (n = 12). ms: median survival. Comparison of survival curves performed using the Log-rank Mantel-Cox test

Fig. 5

Tumor-associated microglia and macrophages are heterogeneous, and combination therapy impacts their representation within the glioma microenvironment. (a) Uniform Manifold approximation and projection (UMAP) plot, (b) total counts, (c) proportions, and (d) dot plot of microglia clusters from GL261-bearing brains at day 21 post-implantation, following unsupervised clustering and manual annotation. (e) UMAP plot, (f) total counts, (g) proportions, and (h) dot plot representation of macrophage clusters from GL261-bearing brains at day 21 post-implantation, following unsupervised clustering and manual annotation. Dot plots show average gene expression in gradient color code and percentage of cells expressing the gene represented by circle diameter. For panels b, c, f, and g, counts and proportions of all microglia clusters in each combinatory treatment group (from left to right: control, ICB, PT2385, and PT2385+ICB)

Fig. 6

Heterogenous T cell subpopulations infiltrate the glioma microenvironment, and HIF-2α inhibition combined with ICB impacts their representation (a) UMAP projection of T cell clusters from GL261 glioma-bearing brains at day 21 post-implantation, showing unsupervised clustering and manual annotation of subpopulations. (b) Proportional representation of T cell clusters across treatment groups (vehicle, ICB, PT2385, and ICB/PT2385).(c) Dot plot illustrating average expression and percent expression of key markers across T cell subpopulations. (d) Proportion of Tregs (CD25⁺FOXP3⁺), gated on CD4⁺ T cells, in each treatment group. Data are shown as mean ± SD, with n = 5 mice per group. Statistical significance was assessed using a one-way ANOVA test. (e) Representative flow cytometry plots of Treg gating in all treatment groups. (f) Proportions of CD4⁺ T cells and (g) CD8⁺ T cells, gated on live CD45⁺CD11b^-CD3⁺ cells in each treatment group. Data are shown as mean ± SD, n = 5. (h) Ratio between CD8⁺ T cells to Tregs (CD4⁺CD25⁺FOXP3⁺) in each treatment group. Data are shown as box plots (mean ± SD, n=5); statistics by one-way ANOVA test. (i) Percentage of Treg (CD4⁺CD25⁺FOXP3⁺) migration from top chamber in a transwell assay towards the bottom chamber seeded with M2 polarized BV2 microglia pretreated with PT2385 or DMSO. PTX: Treg pretreated with pertussis toxin as negative control. Treg treated with CXCL10 and CCL22 as positive control. Data normalized by percentage of positive control are shown as mean ± SD, from 2 independent experiments with 4-5 technical replicates per group); statistics by one-way ANOVA test