Hypoxia potentiates glioma-mediated immunosuppression

Abstract

Glioblastoma multiforme (GBM) is a lethal cancer that exerts potent immune suppression. Hypoxia is a predominant feature of GBM, but it is unclear to the degree in which tumor hypoxia contributes to this tumor-mediated immunosuppression. Utilizing GBM associated cancer stem cells (gCSCs) as a treatment resistant population that has been shown to inhibit both innate and adaptive immune responses, we compared immunosuppressive properties under both normoxic and hypoxic conditions. Functional immunosuppression was characterized based on production of immunosuppressive cytokines and chemokines, the inhibition of T cell proliferation and effector responses, induction of FoxP3+ regulatory T cells, effect on macrophage phagocytosis, and skewing to the immunosuppressive M2 phenotype. We found that hypoxia potentiated the gCSC-mediated inhibition of T cell proliferation and activation and especially the induction of FoxP3+T cells, and further inhibited macrophage phagocytosis compared to normoxia condition. These immunosuppressive hypoxic effects were mediated by signal transducer and activator of transcription 3 (STAT3) and its transcriptionally regulated products such as hypoxia inducible factor (HIF)-1α and vascular endothelial growth factor (VEGF). Inhibitors of STAT3 and HIF-1α down modulated the gCSCs' hypoxia-induced immunosuppressive effects. Thus, hypoxia further enhances GBM-mediated immunosuppression, which can be reversed with therapeutic inhibition of STAT3 and HIF-1α and also helps to reconcile the disparate findings that immune therapeutic approaches can be used successfully in model systems but have yet to achieve generalized successful responses in the vast majority of GBM patients by demonstrating the importance of the tumor hypoxic environment.

Figure 1. Hypoxia increased immunosuppressive cytokine production in gCSCs.

Under hypoxic conditions, there was an increase in production of: VEGF in 3 of 4 gCSCs, galectin-3 in 3 out of 4 gCSCs, sCSF-1 in 3 of 4 gCSCs, and CCL-2 in 3 of 4 gCSCs, as assessed by ELISA. Error bars are derived from three separate experiments representing the deviation of the associated means between experiments throughout the manuscript. * P<0.05.

Figure 2. Hypoxia enhances the gCSC-mediated immunosuppression on human T cells.

A. When healthy donor peripheral blood mononuclear cells (PBMCs) were cultured in the presence of supernatant medium from cultures of each of the gCSCs, T cell proliferation was inhibited, as demonstrated by fluorescence-activated cell sorting (FACS) analysis of CD3+ T cell carboxyfluorescein diacetate succinimidyl ester (CFSE) labeling that measures T cell proliferation. For each of the matched gCSCs under hypoxic conditions, T cell proliferation was further inhibited compared to normoxia. Representative FACS histograms or dot plots are shown on the left, and the summary bar graphs of percentage changes are shown on the right. T cells cultured with neurosphere cell medium served as a control. B. The supernatants from the gCSCs induced an increase in the number of FoxP3⁺ regulatory T cells in the gated CD4⁺ T cells, which is further enhanced under conditions of hypoxia. C. The inhibition of IFNγ production in CD3+ T cells was enhanced under hypoxic condition compared to normoxia. * P<0.05.

Figure 3. The ability of gCSCs to inhibit phagocytosis in monocytes is enhanced by hypoxic conditions.

A. Supernatants from gCSCs cultured in normoxic and hypoxic conditions inhibited phagocytosis of fluorescent microbeads (red) in monocytes exposed to the supernatants for 72 h, with hypoxic supernatants inhibiting phagocytosis to a greater extent than normoxic supernatants for all 4 gCSCs tested. Representative high-power fluorescent microscope images (40X) of monocytes exposed to (i) neurosphere medium alone as a positive control, (ii) normoxic gCSC supernatant, and (iii) hypoxic gCSC supernatants. Cell nuclei were stained with DAPI (blue). B. *P<0.05 indicates a significant difference between phagocytosis inhibition by hypoxic versus normoxic gCSC supernatants.

Figure 4. Hypoxia increased the intracellular pSTAT3 and HIF-1α expression in gCSCs.

A and B. The pSTAT3 expression in the gCSCs was measured by intracellular pSTAT3 (pY705) staining via flow cytometry. The mean fluorescent intensity of positive staining was determined by flow cytometry analysis under both normoxic and hypoxic conditions for each gCSC (representative histogram was shown as A), and then the percentage change was calculated compared with the expression of pSTAT3 in each gCSC under normoxic condition (shown as B). C and D. The HIF-1α expression in the gCSCs was measured by intracellular HIF-1α staining via flow cytometry after culture in either hypoxic or normoxic culture conditions: representative histogram was shown as B, and percentage changes of all 4 gCSCs under hypoxia compared to normoxic condition was plotted as bar graphs in D. E. Confirmatory representative western blot demonstrating that pSTAT3 and HIF-1α expression is increased within gCSCs after hypoxia exposure. There was no change for total STAT3 expression.

Figure 5. WP1066 inhibits pSTAT3 and HIF-1α expression, whereas WP744 only inhibits HIF-1α expression in gCSCs.

A and B. WP744 or WP1066-treated, hypoxia and normoxia exposed gCSCs express lower levels of HIF-1α than gCSCs exposed to hypoxia and normoxia alone. The HIF-1α expression in the gCSCs cultured under normoxia and hypoxic conditions was measured by intracellular staining via flow cytometry in the presence or absence of WP744 or WP1066. Representative histograms were shown as A. The percentage changes of all 4 gCSCs under hypoxia and normoxia were compared to untreated gCSCs and then the corresponding percentage change of expression was calculated and plotted as bar graphs in B. C and D. WP1066 inhibited pSTAT3 expression in gCSCs under both normoxia and hypoxia, whereas WP744 enhances pSTAT3 expression. The pSTAT3 expression in the gCSCs under normoxia and hypoxic conditions was measured by intracellular pSTAT3 staining via flow cytometry in the presence or absence of WP744 or WP1066. Representative histograms were shown as C, and percentage changes of all 4 gCSCs under hypoxia and normoxia were plotted as bar graphs in D. In histograms, dashed line represented isotype staining and the black line staining for either HIF-1α or pSTAT3. The number in histograms represented positive percentage for pSTAT3 or HIF-1α relative to the isotype. *P<0.05.

Figure 6. Inhibition of pSTAT3 and HIF-1α reverses hypoxia-gCSCs induced immunosuppression A.

WP1066 reverses gCSC-mediated immunosuppression in human T cells to a greater degree than WP744. When healthy donor PBMCs were cultured in the presence of each of the supernatants conditioned by the gCSCs, T cell proliferation was inhibited, as demonstrated by FACS analysis on CFSE labeled T cell division. This inhibitions was further potentiated by hypoxia as previous shown but could be reversed with WP1066 to the greater degree than WP744. B. The supernatant culture mediums from the gCSCs induces an increase in the number of FoxP3⁺ regulatory T cells among the gated CD4⁺ T cells that is further enhanced under conditions of hypoxia as previously shown. This was reversed with WP744 to a greater degree than with WP1066. The controls were T cells cultured in respective neurosphere medium of normoxia and hypoxic conditions. * P<0.05 versus untreated gCSCs exposed to normoxia or hypoxia C. Both WP1066 and WP744 reversed the inhibition of phagocytosis in monocytes by hypoxia exposed gCSCs in all 4 gCSCs. Shown are representative high-power fluorescent microscope images (40X) of monocytes cultured with (i) hypoxia-exposed gCSC supernatant, (ii) WP1066-treated, hypoxia-exposed gCSC supernatant, and (iii) WP744-treated, hypoxia-exposed gCSC supernatant. * P<0.05 versus untreated hypoxia-exposed gCSCs.