The future looks bright for immunotherapies and novel targeted approaches for patients with gliomas, many of which have already been developed for other types of cancer.
Andrew S. Chi, MD, PhD
The heterogenic nature of brain tumors has historically made the malignancy a difficult one to treat, with very few therapies panning out in large clinical trials. However, despite the letdown of prior research, the future looks bright for immunotherapies and novel targeted approaches for patients with gliomas, many of which have already been developed for other types of cancer.
“We have so much knowledge about the biology of these tumors and a proliferation of different drugs and therapies that have been developed over the past few years are working in other cancers,” Andrew S. Chi, MD, PhD, head of Neuro-Oncology at NYU Langone’s Laura and Isaac Perlmutter Cancer Center, told OncLive. “We have to figure out how to get these drugs to work in brain cancers.”
Therapies targeted against IDH1/2 represent a potential treatment paradigm for patients with glioma. Additionally, novel approaches, such as those using alternating electric fields have demonstrated efficacy. In October 2015, Optune, a device using tumor treating fields, became the first FDA-approved therapy in more than a decade to demonstrate a statistically significant extension in overall survival (OS) for patients with newly diagnosed glioblastoma multiforme (GBM).
Immunotherapy has also generated a great deal of excitement. Recently, the PD-1 inhibitor nivolumab (Opdivo) demonstrated early signs of activity as a monotherapy for patients with recurrent GBM. Furthermore, other immune-based therapies targeted against EGFRvIII have shown exceptional signs of activity and tolerability.
Adding to the excitement, preclinical research conducted on tumor samples and mouse models showed that the protein FGL2, which is secreted by GBM, causes the upregulation of immune suppression mechanisms. In early studies, an anti— FGL2 antibody demonstrated therapeutic potential, suggesting the possible role for a novel immunotherapy.
Checkpoint Inhibitors Show Early Promise
“FGL2 is not just working on immune checkpoints. It’s also working on other cells that are also contributing to the tumormediated immune suppression,” Amy Heimberger, MD, professor in the Department of Neurosurgery, University of Texas MD Anderson Cancer Center, told Targeted Oncology. “We think this is going to impact more patients because it hits more than one mechanism and we think it will have a more profound impact.”The CheckMate-143 study is comparing nivolumab alone (3 mg/kg) versus bevacizumab alone (10 mg/kg) in patients with recurrent GBM. Preliminary results from an earlier phase of the study showed that nivolumab monotherapy had a median OS of 10.5 months, with 40% of patients alive after one year of follow-up.1 Traditionally, bevacizumab has shown a median OS between 7 and 10 months and a 1-year OS rate of around 25%.
The study is now proceeding to a phase III expansion, which is already under way. The estimated completion date for collecting data on the primary outcome measures of safety and OS is June 2017. An estimated 440 patients are expected to enroll (NCT02017717).
“Immunotherapies are really riding a big wave in research and therapeutic development and they are generating a lot of excitement in a lot of cancer types, as well as brain cancer,” said Chi. “When these immunotherapies work in other cancers, they have worked phenomenally well.”
In addition to nivolumab, the PD-1 inhibitor pembrolizumab (Keytruda) and the PD-L1 inhibitor durvalumab are also being explored in phase II studies for patients with glioma. These trials are looking at the agents as monotherapy and in combination with bevacizumab or radiotherapy. The pembrolizumab trial plans to enroll 80 patients (NCT02337491) and the durvalumab trial is looking to enroll 100 patients (NCT02336165).
Chimeric Antigen Receptors in Glioma
“The excitement for brain cancers has been high because, when you look at the biology of these tumors, there is significant rationale that these immunotherapies should work,” said Chi. “Whether it be vaccines or immune checkpoint inhibitors, there has been a tremendous amount of research into the immunobiology, gliomas in particular, that suggests that some of these immune therapies should work.”Chimeric antigen receptor (CAR) T-cell therapies targeted against CD19 have shown exciting promise for patients with hematologic malignancies, where they have demonstrated impressive efficacy in both chronic lymphocytic leukemia and acute lymphoblastic leukemia. Now, following this proof of principle, these agents are being explored across a variety of solid tumors.
“CAR involves designer T-cells that hone in one specific mutation or abnormal proteins on the surface of glioma cells, and ones that are only on glioma cells and not normal brain cells,” said Chi. “This would be a very specific attack and one without much toxicity.”
A pilot study has assessed CAR T-cell therapies for those with EGFRvIII-positive GBM. In the ongoing pilot study, which had the primary endpoints of safety and feasibility, 8 patients with multifocal disease were treated with CAR T-cells. Overall, 5 of the 8 patients received surgery and were refractory to prior therapies.2
Overall, a peak in T-cell engraftment was observed by day 7 following treatment. In those treated with surgery, 80% experienced a significant reduction in EGFRvIII levels following treatment with the CAR T-cell therapy. In 3 of the 5 patients, there was a near complete elimination of EGFRvIII observed.
Edema and seizure occurred in two separate patients. The seizure was grade 3 and the edema was grade 4 but transient. Both events were quickly reversed with steroids and a single dose of IL-6 blockade.
“Some of the problems we need to overcome in brain cancer therapy are if the T-cells are actually getting into the brain and actually getting into the tumor,” said Chi. “Another problem we deal with is that maybe the T-cells that are designed to hit an abnormal molecule aren’t because those specifically abnormalities might not be on every single glioma cell.”
In the pilot study, the CAR T-cell therapy did not disrupt other mutations, such as PI3K or wild-type EGFR. Moreover, there were indications that some of the activated EGFR alleles, which were coexpressed in these tumors before CAR-T cells, disappeared in addition to EGFRvIII.
EGFRVIII Inhibition Hits Snag
“Some parts of the tumor are very different than other parts on a molecular and biological level,” said Chi. “If you design a T-cell to attack one specific protein, then you may not get the effectiveness of essentially killing the entire tumor.”In the phase II ReACT trial, the combination of the EGFRvIII-targeted immunotherapy rindopepimut (Rintega) and bevacizumab demonstrated promising signs of activity for patients with relapsed GBM.3 In updated findings for OS, 25% of patients treated with rindopepimut remained alive at 2-years compared with none in the control arm. The median OS with rindopepimut was 11.3 versus 9.3 months in the control arm (HR, 0.53; CI, 0.32-0.88; P = .0137).
“Most of us were taught, in medical school, that the brain is immuno-privileged and, therefore, immunotherapy may not have as much of an impact for patients with brain cancer,” said David Reardon, MD, clinical director at the Center for Neuro-Oncology, Dana-Farber Cancer Institute. “From research, we now know that that is not the case; the relative privilege of the central nervous system is not definitive by any means. There is a dynamic interaction between the systemic immune system and what is reactive in the brain.”
Unfortunately, results from the ReACT trial could not be duplicated in the phase III ACT IV study. In early March, the developer of the immunotherapy, Celldex Therapeutics, announced that treatment with rindopepimut plus temozolomide failed to improve OS compared with temozolomide and a control for patients with newly diagnosed EGFRvIII-positive GBM.
In the topline findings from the study, the median OS with rindopepimut was 20.4 months compared with 21.1 months in the control arm (HR, 0.99). Based on this assessment, which was conducted by an independent data safety and monitoring board, the study was discontinued, according to Celldex; however, the immunotherapy will continue to be offered in the phase III trial and compassionate use programs.
Rindopepimut consists of an EGFRvIII peptide conjugated to keyhole limpet hemocyanin. The vaccine works by generating a specific immune response against tumors that express EGFRvIII, which is a tumor-specific oncogene expressed in approximately 30% of GBMs.
“We know that tumors have a number of immunosuppressive factors associated with them,” said Reardon. “Therefore, combining a treatment that has a very potent immunogenic effect, such as a tumor-specific vaccine like rindopepimut, with treatments that can decrease some of the immunosuppressive factors in the tumor microenvironment, like the immune checkpoint inhibitors, is a logical and exciting potential to build on this.”Early phase findings have described the ability to overcome immunosuppression, specifically the biological significance of FGL2 expression. In a study using The Cancer Genome Atlas (TCGA) glioma database and tumor lysates analysis found that FGL2 promotes the development and growth of GBM by inducing multiple immune suppression mechanisms, such as PD-1 and CD39.
“Right now there is a great deal of enthusiasm on things like immune checkpoints, but we think there are only going to be a subset of patients who respond to that,” said Heimberger. “Therefore, we are going after key targets and key hubs of tumor-mediated immune suppression. These are particular pathways or mechanisms that deal with more than one mechanism of immune suppression.”
Findings from the TCGA were further explored in immunecompetent mice.4 The study found that 72.5% of low-grade gliomas maintained two copies of the FGL2 gene whereas in GBM 83.8% had FGL2 gene amplification or copy gain. Overall, those with high levels of FGL2 mRNA in glioma tissue had a lower OS (P = .009). Levels of FGL2 in lysates were high in GBM versus low-grade glioma.
Mice treated with an anti-FGL2 antibody had a median OS of 27 days compared with only 17 days for mice treated with a control antibody (P = .01). The investigational antibody reduced CD39+ Tregs, M2 macrophages, PD-1, and myeloidderived suppressors cells.
“In the case of FGL2, this is a secreted protein that glioblastoma makes. What it does is it exploits multiple mechanisms of immune suppression to stay behind the detection of the immune system,” said Heimberger. “It specifically increases T regulatory cells, which are known to inhibit immune responses.
It also induces M2, or the tumor-supported macrophage, as well as recruits a myeloid-derived suppressor cell.” Given the heterogeneity in most high-grade gliomas, a multipronged mechanism of action could prove useful. Additionally, an anti—FGL2 antibody could have utility in combination strategies, specifically with agents that hit other targets. In most cases, when one immune target is hit another becomes upregulated.
“There needs to be an appreciation of the fact that immune suppression is complex and simply hitting a single mechanism is not going to be a home run,” said Heimberger. “We have to look at the complexity of the biology that we’re dealing with and realize that it’s not just about immune checkpoints and it’s not just about T regulatory cells. It’s about the whole picture of how the tumor exploits the immune system to stay behind detection.”Mutations in the IDH1 and IDH2 genes have been described in studies for the past 5 years, especially for their potential role in personalizing therapy. Following the characterization of the genes, clinical trials are now assessing peptide vaccines and targeted therapies against IDH1/2.
“One of the subtypes of glioma that has emerged over the past 5 years are ones defined by the mutation of either IDH1 or IDH2,” said Chi. “These are genes and proteins that are a core part of a cell’s metabolism and the mutations of IDH were discovered in late 2008. It took a while for us to understand the biological and clinical impact of that mutation in gliomas and other cancers.”
In a preclinical study,5 Chi and colleagues found that IDH1- mutant cancers were extremely vulnerable to coenzyme NAD+ depletion. The study assessed metabolic therapeutic targets through a systematic profiling of metabolites in endogenous IDH1-mutant cancer cells following IDH1 inhibition.
In the study, they found that NAD+ levels were decreased in IDH1-mutant tumors through the inhibition of an alternate salvage pathway, which created sensitivity to NAD+ depletion. When NAD+ was blocked in the trial, IDH1-mutant cells underwent autophagy.
“Ultimately, what we tried to do with NAD was manipulate NAD and see what the effect was on the tumor,” said Chi. “We took drugs that are known to deplete NAD from the tumor and it turned out that, in the dish, and we found that they potently killed IDH1-mutant glioma cells. These drugs did not have any effect on tumors that were not IDH1-mutant.”
A number of clinical trials are currently assessing IDH1- and IDH2-targeted therapies and peptide vaccines across a variety of cancer types. These agents have also demonstrated efficacy in patients with acute myeloid leukemia, representing a broad application.