Given the current lack of positive phase 3 immunotherapy trials for glioblastoma, a misconception held by oncologists and neuro-oncologists is that immunotherapy has no place in this malignancy.
Immunotherapies approved by the FDA in the past decade have become standard of care for many malignancies. Despite the ability of these agents to slow or eradicate numerous tumors, even those metastasized to the brain,1-3 none thus far have extended survival in glioblastoma (GBM). Given the current lack of positive phase 3 immunotherapy trials for GBM, a misconception held by oncologists and neuro-oncologists is that immunotherapy has no place in this malignancy. In reality, tumor immunology is far too complex to determine infeasibility based on these initial completed trials. In fact, these negative trials provide valuable lessons on designs of future trials based on the hurdles limiting anti-GBM immune responses (Figure).
Clinicians treating patients with GBM are crucial to the future of immunotherapy development. Patients must continue to be enrolled into we newer generations of trials that build on past lessons by testing novel therapeutics, combinations, and longitudinal biomarker analyses.
Antagonism of T-cell immune checkpoint molecules PD-1, PD-L1, and CTLA4 unleashes immune responses against many tumors. The remarkable efficacy of antibodies blocking these checkpoints has brought immunotherapy to thousands of patients.4,5 Three recently completed randomized phase 3 trials in GBM—CheckMate 143 (NCT02017717),6 CheckMate 498 (NCT02617589),7 and CheckMate 548 (NCT02667587)8—tested PD-1 blockade, but each demonstrated an objective response rate of less than 10%.
CheckMate 143 evaluated the anti–PD-1 monoclonal antibody nivolumab (Opdivo) in patients with a first recurrence of GBM.6 CheckMate 498 tested radiation and nivolumab vs radiation with standard-of-care temozolomide (Temodar) in patients with newly diagnosed GBM with an unmethylated MGMT promoter.7 CheckMate 548 tested the addition of nivolumab to radiation and temozolomide in newly diagnosed patients with a methylated MGMT promoter.8 These well-designed trials demonstrated that anti–PD-1 therapy lacks efficacy as a single agent or in combination with standard-of-care therapy for patients with recurrent or newly diagnosed GBM. Importantly, tumors responsive to anti–PD-1 in CheckMate 548 expressed more than 5% PD-L1.8 Therefore, inhibition of local glioblastoma-reactive T cells via PD-L1 is a rare mechanism of immune suppression (Figure). Low PD-L1 expression characterizes immunologically “cold” tumors such as GBM, which is further reflected by a low mutational burden, relatively fewer neoepitopes for T cells, heterogeneity in antigen expression, and a paucity of T cells infiltrating the tumor bed (Figure).
Trials moving forward in GBM are attempting to address the lessons learned from the negative phase 3 trials with new checkpoint combinations, timings, and agents targeting multiple antigens. Dozens of other T-cell checkpoints have been identified, many of which are being tested in novel combinations. Examples of these trials include urelumab (antiLAG3) with nivolumab (anti–PD-1; NCT02658981); sabatolimab (anti–TIM-3) with spartalizumab (anti–PD-1) and radiosurgery (NCT03961971); and domvanalimab (anti-TIGIT) with zimberelimab (anti–PD-1; NCT04656535).
These agents potentially modulate T cells in the periphery or in the tumor, with the potential for synergy with radiation or surgery to release tumor antigens and increase T-cell proliferation. Some neoadjuvant or window-of-opportunity trials include surgical resection of recurrent tumor among patients who do or do not receive a preoperative dose of experimental immunotherapy; therefore, pharmacodynamic measures of intratumoral or systemic immune activation can be compared between responders and nonresponders (NCT04656535 and NCT04606316). Clinical trials are increasingly available using novel combinations of immune checkpoint blocking antibodies separately or in combination with standard therapies for patients with recurrent GBM.
Vaccination strategies have been attempted in GBM for many years. Prior strategies failed to encompass the heterogeneity in antigen expression9 or expand enough T cells into the tumor.10 Three ongoing trials employ novel strategies to vaccinate against more potent GBM antigens in combination with immune checkpoint blockade to assist with initial expansion11 in addition to cytolytic activity of T cells.12
The multicenter, multinational ROSALIE study (NCT04116658) combines pools of peptides derived from GBM that resemble peptides from the gut microbiome, with the rationale that T cells reactive to these peptides are already relatively prevalent because of prior expansion and activation physiologically in the gut, and thus they could be capable of enhanced tumor killing. This multipeptide vaccine is combined with the anti–PD-1 nivolumab for patients with recurrent glioblastoma with or without the anti–VEGF-A agent bevacizumab (Avastin). Another trial (NCT03491683) incorporates a novel DNA plasmid approach that vaccinates against 3 GBM-associated antigens and adds IL-12 in combination with the anti–PD-1, cemiplimab-rwlc (Libtayo). This regimen is added to chemoradiation with temozolomide for patients with newly diagnosed GBM. Another example includes a personalized neoantigen peptide vaccination strategy in combination with the anti–PD-1 antibody pembrolizumab (Keytruda) plus standard-of-care therapy for patients with newly diagnosed GBM (NCT02287428). Of note, this trial is further evaluating the timing of PD-1 blockade relative to vaccine priming based on emerging data implicating PD-1 signaling on T-cell priming and memory responses.13,14
Clinical trials using other approaches to address the hurdles of intratumoral heterogeneity and other immune suppressive cells as well as factors in the GBM microenvironment (Figure) are also in development. For example, chimeric antigen receptor (CAR)-positive T cells, which show evidence of efficacy in select patients,15 are being engineered to target more than 1 antigen to account for heterogeneity in GBM.16 Along these lines, City of Hope in Duarte, California, is preparing to launch a trial in which CAR T cells engineered against 2 distinct tumor antigens are coadministered. Such approaches may help overcome therapeutic limitations associated with intratumoral heterogeneity and antigen escape by providing supraphysiologic expansion of distinct tumor-targeting T cells ex vivo that are infused into the tumor or ventricular space.
In addition, oncolytic viral therapy remains an attractive approach for reversing the poorly understood complexities of the tumor microenvironment, which likely arise in great part from myeloid and other prevalent cell populations in GBM.17 Oncolytic viruses induce proinflammatory changes in the microenvironment by introducing of viral proteins and nucleic acids and lysis-mediated release of powerful danger signals, in addition to releasing GBM antigens in the microenvironment. Combinations of oncolytic virus and immune checkpoint blockade now seek to potentiate T-cell responses to glioblastoma antigens,18 including some trials incorporating serial, longitudinal oncolytic virus injection19 (NCT03636477; NCT02798406; NCT03152318). Although specific mechanisms of immune suppression in the GBM microenvironment are being elucidated, oncolytic viral therapy represents a promising approach to potentially reverse this immensely complex problem.
Single agent blockade of PD-1 has failed to extend survival in patients with GBM as it has in other cancers. However, advances in our understanding of the key hurdles to immune-mediated tumor killing are informing the current generation of clinical trials using rational combinations. Patients in clinical trials contribute to the understanding of immune suppression in this disease through immune monitoring. Given the complexity of the immune system and an equally complex set of mechanisms for therapeutic targeting, the current landscape of trials is only the beginning of the story of immunotherapy for GBM.
David A. Reardon, MD, is clinical director of the Center for Neuro-Oncology and institute physician at Dana-Farber Cancer Institute, and a professor of medicine at Harvard Medical School, both in Boston, Massachusetts.
Brian M. Andersen, MD, PhD, is an instructor in neurology in the Department of Neurology at Brigham and Women’s Hospital, Harvard Medical School, in Boston, Massachusetts, and in the Department of Neurology at Veterans Affairs Medical Center in Jamaica Plain, Massachusetts.