Understanding innate immune response in glioblastoma in search of a way forward

T cells have long been the center of focus in cancer immunology due to their role in direct killing of cancer cells. Studies that show favorable prognosis associated with higher infiltration of cytotoxic CD8 T cells¹ as well as concurrent infiltration of CD8 and CD4 T cells² have further corroborated the pivotal role T cells play in mediating tumor control or clearance. These findings, coupled with the successful use of therapeutic monoclonal antibodies directed against the programmed cell death 1 (PD-1)–programmed cell death ligand 1 (PD-L1) pathway to reinvigorate effector T cells in checkpoint blockade responsive tumors such as melanoma have shown the importance of studying T-cell activation and exhaustion to fully understand the balance between immune surveillance and tumor growth. Such efforts at using checkpoint blockade to unleash the breaks on T cells have not yielded the same level of success in gliomas, warranting the identification of key players in the tumor microenvironment that might blunt the effects of immunotherapy. In this issue, de Groot and team contribute to our understanding of these tumor-promoting myeloid cells and show us how they might mediate resistance to anti–PD-1 therapy in glioblastoma (GBM) patients.³

In a phase II “window of opportunity” trial, de Groot et al characterize the immune microenvironment of GBM specimen after pembrolizumab treatment to study anti–PD-1 therapy-induced changes in immune effector function. The study shows a paucity of T cells as well as a predominance of CD68+ immunosuppressive macrophages in the tumor that did not vary with treatment, suggesting that the presence of these immunosuppressive macrophages contribute to the blunting of anti–PD-1 efficacy. The CD68+ macrophages in the tumor express abundant immunosuppressive markers and comprise the majority of immune cells, constituting as high as 70% of the infiltrates. De Groot and colleagues show the presence of these “M2” phenotype-skewed CD68+ macrophages in tumors as well as the walls of their blood vessels, which could contribute to suppression of T-cell recruitment, effector function, or sequestration of anti–PD-1 therapy from activating exhausted T cells. Although de Groot’s team noted a modest increase in granzyme B–producing T cells in immunohistochemical analysis of pembrolizumab-treated specimen compared with untreated GBM, there was no significant increase in T-cell infiltration. However, there is variability in T-cell infiltration observed by other groups after anti–PD-1 therapy in recurrent GBM. Schalper et al⁴ and Cloughesy et al⁵ have demonstrated relative increase in T-cell infiltration and upregulation of chemokine transcripts and interferon gamma–related genes in T cells with anti–PD-1 therapy. Such variability likely highlights the heterogeneity of GBM and the need for a multipronged approach to treat it.

Since T cells cannot function entirely independently, it is imperative to understand their interaction with cells that can influence their activity, and the study by de Groot’s team points to a possible mechanism by which myeloid cells could orchestrate the failure of T cells in controlling tumor progression. Knowledge of such suppressive mechanisms is important for the development of therapies to overcome them. In the field of tumor immunology, modulating interaction between myeloid cells and T cells is being increasingly recognized as a way to influence T-cell function. Notable examples of therapies that target myeloid cells in order to affect T-cell activity include intratumoral injection of granulocyte-macrophage colony-stimulating factor to enhance the number of tumor-infiltrating dendritic cells to bolster antigen presentation, blocking activity of tumor-promoting and T cell–suppressing transforming growth factor beta cytokine as well as arginase-1 enzyme made by tumor associated macrophages in preclinical models as well as human trials.^6,7 Such efforts would be especially applicable in gliomas because they predominantly recruit macrophages and polarize them to attain tumor promoting phenotype. While studies looking at colony stimulating factor 1 receptor (CSF1-R) inhibition to target recruited macrophages have shown efficacy against murine models of glioma,⁸ clinical trials with CSF1-R inhibitors have shown limited benefits,⁹ which further necessitates the characterization of the diversity and functionality of myeloid population in glioma.

The immunophenotyping of immune cell infiltrates conducted by de Groot et al is an important step toward reducing immunosuppression in glioma, unveil new possible targets, and improve prognosis. Characterizing the immune clusters found in tumor samples identified suppressive markers on CD68+ macrophages including VISTA (V-domain immunoglobulin suppressor of T-cell activation) and B7-H3, PD-L1 as well as PD-1. Recent preclinical work by others has shown that blocking or ablating PD-1 signaling in myeloid cells can promote differentiation of effector monocytes and translate to enhanced T-cell activation and tumor clearance, while PD-1 ablation on T cells did not attain the same results.¹⁰ Although PD-1 was expressed across many macrophage clusters identified by de Groot’s team, the presence of other suppressive markers on macrophages could indicate why PD-1 blockade alone might not be enough in targeting myeloid cells in GBM. The presence of alternative checkpoints in CD68+ macrophages could compensate for PD-1–PD-L1 blockade, and the predominance of immunosuppressive myeloid cells in the tumor during the therapeutic window of anti–PD-1 administration might overwhelm or mask any therapy-mediated change in the differentiation or polarization of effector macrophage, rendering monotherapy with anti–PD-1 ineffective. Therefore, it will be beneficial to combine therapies that target inhibitory molecules like VISTA and B7-H3 on CD68+ macrophages to license T-cell activation in the context of PD-1–PD-L1 blockade.

By characterizing the changes in immune population and effector function during the therapeutic window in GBM patients treated with pembrolizumab, de Groot and team found a scarcity of T cells and the prevalence of immunosuppressive CD68+ macrophages in the tumor that did not alter with anti–PD-1 treatment. With more functional characterization of intratumoral CD68+ macrophages we might attain further insight into ways to overcome such myeloid-mediated immunosuppression. From blocking their recruitment into the tumor, to targeting alternative checkpoint markers on myeloid cells, to depleting them from the tumor, we could soon be expanding our arsenal of therapies to prolong survival of patients with gliomas.