Advances in Immunotherapies for Gliomas

Abstract

Purpose of Review

Immunotherapy-based treatment of glioblastoma has been challenging because of the tumor’s limited neoantigen profile and weakly immunogenic composition. This article summarizes the current clinical trials underway by evaluating the leading immunotherapy paradigms, the encountered barriers, and the future directions needed to overcome such tumor evasion.

Recent Findings

A limited number of phase III trials have been completed for checkpoint inhibitor, vaccine, as well as gene therapies, and have been unable to show improvement in survival outcomes. Nevertheless, these trials have also shown these strategies to be safe and promising with further adaptations. Further large-scale studies for chimeric antigen receptors T cell therapies and viral therapies are anticipated.

Summary

Many current trials are broadening the number of antigens targeted and modulating the microtumor environment to abrogate early mechanisms of resistance. Future GBM treatment will also likely require synergistic effects by combination regimens.

Introduction

Despite significant breakthroughs in immunotherapy for other cancers, gliomas—particularly glioblastoma (GBM)—remain stalwart models of adaptive tumor resistance that evade current methods of treatment. While advances have been made since the standard of care delineated by the original Stupp protocol involving surgical resection with chemoradiation [1], clinical progress with immunotherapy for gliomas has been difficult due to a highly immunosuppressive milieu—both systemically with global lymphodepletion and locally within the tumor microenvironment [2•, 3].

Many of the clinical advances made in immunotherapy for gliomas have largely come from adapting other successful immunotherapy treatments for tumors such as melanoma and non-small cell lung cancer [4, 5]. However, gliomas such as GBM have unique characteristics compared to other tumors: (1) a limited presence and inhibited activation of tumor infiltrative lymphocytes, (2) a predominance of myeloid-derived suppressors cells at the tumor site, (3) the blood brain barrier, (4) ineffective antigen presenting cells (APCs), (5) immunometabolic disturbances such as high glutamate levels and mitochondrial decoupling that cause further immunosuppression, as well as (6) a low diversity of neoantigens and mutational burden [3, 6]. In light of these challenges, several preclinical studies have sought to further our understanding of basic mechanisms of immune resistance endemic to GBM that in turn may reveal potential biomarkers and targets for new immunotherapies.

This review offers perspectives on current explorations of immunotherapy in clinical trials as well as promising new advances in preclinical and early clinical studies that demonstrate the latest advances in immunotherapies for gliomas. In particular, studies involving checkpoint blockade, immunomodulating vaccines, chimeric antigen receptor (CAR) T cell therapy, viral therapy, and new combinatorial immunotherapies will be highlighted (Table 1).

Table 1.

List of described clinical trials involving immune checkpoint blockade, peptide vaccinations, dendritic cell vaccinations, CAR T cell therapy, and viral immunotherapy. The table is distributed into class, phase, title of the study, and intervention. For further reference, NCT and date of last update are also included

Class	Phase	Title of study	Intervention	NCT	Last update posted
Checkpoint	III	A Study of the Effectiveness and Safety of Nivolumab Compared to Bevacizumab and of Nivolumab With or Without Ipilimumab in Glioblastoma Patients (CheckMate 143)	Biological: Nivolumab Biological: Bevacizumab Biological: Ipilimumab	NCT02017717	4/19/2021
Checkpoint	III	An Investigational Immunotherapy Study of Nivolumab Compared to Temozolomide, Each Given With Radiation Therapy, for Newly-diagnosed Patients With Glioblastoma (CheckMate 498)	Drug: Nivolumab Drug: Temozolomide Radiation: Radiotherapy	NCT02617589	2/3/2021
Checkpoint	III	An Investigational Immunotherapy Study of Temozolomide Plus Radiation Therapy With Nivolumab or Placebo, for Newly Diagnosed Patients With Glioblastoma (CheckMate548)	Drug: Nivolumab Drug: Temozolomide Radiation: Radiotherapy Other: Nivolumab Placebo	NCT02667587	9/11/2020
Checkpoint	II	Avelumab in Patients With Newly Diagnosed Glioblastoma Multiforme (SEJ)	Biological: Avelumab	NCT03047473	1/29/2021
Checkpoint	II	Phase 2 Study of Durvalumab (MEDI4736) in Patients With Glioblastoma	Drug: Durvalumab Radiation: Standard radiotherapy Biological: Bevacizumab	NCT02336165	2/24/2021
Checkpoint	II	A Study of Atezolizumab to Evaluate Safety, Tolerability and Pharmacokinetics in Participants With Locally Advanced or Metastatic Solid Tumors	Drug: Atezolizumab	NCT01375842	12/11/2018
Peptide/DC vaccine	I/II	Adjuvant Dendritic Cell-immunotherapy Plus Temozolomide in Glioblastoma Patients (ADDIT-GLIO)	Biological: Dendritic cell vaccine plus temozolomide chemotherapy	NCT02649582	1/19/2021
Peptide	III	Phase III Study of Rindopepimut/GM-CSF in Patients With Newly Diagnosed Glioblastoma (ACT IV)	Drug: Rindopepimut (CDX-110) with GM-CSF Drug: Temozolomide Drug: KLH	NCT01480479	1/16/2018
Peptide	II	ERC1671/GM-CSF/Cyclophosphamide for the Treatment of Glioblastoma Multiforme	Drug: ERC1671 Drug: GM-CSF Drug: Cyclophosphamide Drug: Oral Control (Sucrose pill) Drug: Injectable control (Sodium Chloride 0.9%) Drug: Bevacizumab/Bevacizumab Biosimilar	NCT01903330	2/23/2021
DC Vaccines	II	Study of DC Vaccination Against Glioblastoma	Procedure: Surgery Drug: Chemotherapy Radiation: Radiotherapy Biological: DC vaccination Drug: blank placebo	NCT01567202	9/9/2020
DC Vaccines	II/III	Dendritic Cell Immunotherapy Against Cancer Stem Cells in Glioblastoma Patients Receiving Standard Therapy (DEN-STEM)	Biological: Dendritic cell immunization Drug: Adjuvant temozolomide	NCT03548571	4/14/2021
DC Vaccines	II	Efficiency of Vaccination With Lysate-loaded Dendritic Cells in Patients With Newly Diagnosed Glioblastoma (GlioVax)	Biological: Autologous, tumor lysate-loaded, mature dendritic cells Drug: standard therapy	NCT03395587	5/27/2020
DC Vaccines	III	ADCTA for Adjuvant Immunotherapy in Standard Treatment of Recurrent Glioblastoma Multiforme	Biological: Autologous Dendritic Cell/Tumor Antigen, ADCTA	NCT04277221	4/17/2020
CAR T	I	IL13Ralpha2-Targeted Chimeric Antigen Receptor (CAR) T Cells With or Without Nivolumab and Ipilimumab in Treating Patients With Recurrent or Refractory Glioblastoma	Biological: IL13Ralpha2-specific Hinge-optimized 4-1BB-co-stimulatory CAR/Truncated CD19-expressing Autologous TN/MEM Cells Biological: Ipilimumab Biological: Nivolumab	NCT04003649	12/9/2020
CAR T	I	CART-EGFRvIII + Pembrolizumab in GBM	Biological: CART-EGFRvIII T cells Biological: Pembrolizumab	NCT03726515	3/3/2021
Viral Therapy	II	Immunotherapy Targeted Against Cytomegalovirus in Patients With Newly-Diagnosed WHO Grade IV Unmethylated Glioma (I-ATTAC)	Biological: Human CMV pp65-LAMP mRNA-pulsed autologous DCs containing GM CSF Drug: Temozolomide Biological: Tetanus-Diphtheria Toxoid Biological: GM-CSF Biological: 111-Indium-labeling of Cells for in vivo Trafficking Studies	NCT03927222	3/26/2021

Immune Checkpoint Blockade in the PD-1:PD-L1 Axis

The advent of therapies against the programmed cell death protein 1 (PD-1) receptor and its ligands PD-L1 (B7-H1 or CD274) and PD-L2 (B7-DC or CD273) have galvanized immunotherapy as a relatively safe and effective treatment for several cancers [7]. The PD-1:PD-L1 axis normally promotes self-tolerance by suppressing T cell inflammatory activity, though this pathway has been adopted by certain tumors to evade immune responses [8]. While IFN-γ is normally a pro-inflammatory and T cell activating cytokine, its presence actually upregulates expression of PD-L1 through the JAK1/2-STAT1/2/3-IRF1 pathway; additionally, PD-L2 expression is associated with IFN-β and IFN-γ through both IRF1 and STAT3 expression [9]. Therefore, by blocking the interaction between PD-1 and its ligands, several cancers show reinvigoration of cytotoxic T cell activity for anti-tumor response.

The success of anti-PD-1 therapy for cancers such as non-small cell lung cancer (NSCLC) and melanoma invigorated interest in exploring checkpoint blockade for GBM. While the multicenter phase III clinical trial CheckMate-143, which compared checkpoint blockade with anti-PD-1 antibody nivolumab to bevacizumab in recurrent glioblastomas, showed that PD1 therapy had more durable responses than bevacizumab in patients with first time recurrent GBM, less than 10% of patients ultimately responded to therapy, and the study failed to show improved median overall survival (OS) [10••].

Two follow-up phase III clinical trials attempted to further delineate treatment outcomes by examining (1) combination therapy with anti-PD-1 and radiation in patients with unmethylated MGMT promoter GBM (CheckMate-498) and (2) combination therapy involving anti-PD-1 with temozolomide (TMZ) and radiation in methylated MGMT promoter GBM (CheckMate-548). Both studies have failed to meet their primary endpoint for improving OS compared to current standard of care [11, 12].

Several other clinical trials have examined the efficacy of targeting the PD-1:PD-L1 axis in GBM with limited results. The single-center phase II open-label study of avelumab, a PD-L1 inhibitor, examined the effect of anti-PD-L1 therapy on patients within 3 weeks of completing combined radiation therapy and TMZ (NCT03047473). In their first interim analysis in 2019, they reported no unexpected treatment adverse events, with the most common being elevated liver or pancreatic enzymes, and median progression-free survival was 11.9 months—though still comparable with standard of care [13].

However, the open-label phase II study (NCT02336165) involving durvalumab (MEDI4736), another antibody against PD-L1, showed that in patients with unmethylated MGMT promoter GBM who received standard radiation therapy followed by durvalumab after maximum resection had a median OS of 15.1 months compared to 12.7 months in historical standard of care controls for unmethylated GBM patients [14]. This study is ongoing with tumor immunophenotype results pending. Atezolizumab, another PD-L1 inhibitor, has had limited efficacy. While well tolerated in a phase Ia PCD4989g clinical trial (NCT01375842), it demonstrated poor clinical activity in patients who had prior radiation therapy and TMZ ± bevacizumab, with median OS at 4.2 months. However, a positive association between peripheral CD4 + T cells and efficacy were observed in patients who responded to therapy [15]. A list of pertinent clinical trials related to checkpoint blockade is exhibited in Table 1.

The disappointing results of checkpoint inhibitors in GBM highlight the fact that the mechanisms of immunosuppression go beyond lymphocytes as evidenced by a lack of tumor-infiltrating lymphocytes to fulfill the desired downstream effects of checkpoint blockade. This is further upset by a disproportionate presence of immunosuppressive regulatory T cells and myeloid derivatives. Moving forward, therapeutic innovations that rebalance the T cell population and downregulate counterproductive cytokines are likely to be needed for checkpoint inhibitors to meaningfully impact clinical outcomes.

Immunomodulation with Vaccines

Therapeutic vaccines are designed with the broad goal of reprogramming the immune system to recognize and neutralize specific targets. This typically requires a (1) molecular target, usually an antigen to incite a vaccination-induced immune response and (2) immunomodulating adjuvants to promote the desired immune reactions. Overall, the selected targets and adjuvants have shown promising in early phase I/II data, but there has yet to be validating phase III data. Broadly, current vaccinations strategies include peptide-based vaccines including neoantigens, cell-based dendritic cell vaccines and CAR T cell therapies, and oncolytic viruses ± gene therapy [16].

A current phase I/II trial is investigating Wilms tumor 1 (WT1), a notable antigen with a particularly high presence in GBM (NCT02649582) [17]. WT1 expression GBM is as high as 94% and ranked first in a list of 75 cancer antigens [18, 19]. Interestingly, among current phase II trials, it is one of the few vaccine candidates with a single antigen target. While this may be concerning given that prior experience has shown that antigen escape is highly anticipated with single antigen targeting, the loss of WT1 expression leads to cessation of proliferation and/or cancer cell death, which therefore makes the risk of immune escape substantially lower [20, 21]. There are several other neoantigen targets are involved in clinical trials that are described in Table 1.

Notably, an important example of limitations using this vaccination method can be illustrated with the extensive work involving the epidermal growth factor receptor vIII (EGFRvIII) vaccine Rindopepimut (NCT01480479). Activation of wildtype (WT) EGFR of the ErbB family of tyrosine kinase receptors by its ligand epidermal growth factor (EGF) results in cell proliferation, cell migration, and apoptosis inhibition [22]. In its truncated form, EGFRvIII is constitutively activated and is implicated in uncontrolled cell growth [23]. While it is a promising therapeutic target, less than 20–30% of patients exhibit this gene mutation in EGFRvIII, making this target viable for a minority of patients. Furthermore, even when present, 82% of tumors did not express EGFRvIII at recurrence—thereby further limiting its candidacy as a sustainable target [24]. The phase III clinical trial (ACT IV) reflected this issue, with the experimental group versus control showing a median OS of 20.1 vs. 20.0 months, respectively (NCT01480479) [25].

Less selective arrangements have also gained traction as therapeutic strategies, chiefly in the form of autologous peptides derived from tumor lysates. This strategy overcomes many of the obstacles associated with pre-selected antigens. Early concerns about off-target reactions of self-non-tumor antigens have largely been addressed by years of early clinical testing. The Gliovac (ERC1671) vaccine is one such example that has been studied in a phase I/II trial and incorporates both autologous and allogenic components [26•]. In this strategy, patient-derived tumor cell lysates are irradiated and used to stimulate immune response. This is combined with three additional sets of foreign tumor cell lysates to maximize the possibility of generating an immune reaction that disrupts tolerance to a patient’s own neoantigens. Promisingly, the initial phase I trial saw a 40-week survival rate of 77% (NCT01903330) [27, 28].

Building upon the work of several studies involving melanoma, the efficacy of using multi-epitope personalized neoantigen vaccinations for GBM has been explored in a phase I/Ib clinical study [29]. In patients who did not receive dexamethasone—a potent anti-inflammatory often used to reduce vasogenic edema—the presence of these neoantigen vaccines generated polyfunctional neoantigen specific T cell activation with increased tumor infiltration. Furthermore, the phase I trial GAPVAC-101 demonstrated that integration of highly individualized vaccinations with both unmutated antigens and neoantigens in patients with HLA-A*02:01 or HLA-A*24:02 resulted in high immunogenicity with sustained production of central memory CD8 + and CD4 + Th1 cells [30]. A reference list of pertinent clinical trials associated with targeting tumor antigens is described in Table 1.

Dendritic Cell Vaccines

Approximately half of current phase II/III trials involving vaccinations encompass a cell-based strategy, primarily using a dendritic cell (DC) carrier. In these formulations, autologously derived DCs are exposed to potential antigens ex vivo to jumpstart DC immunogenicity. Early groundwork was achieved from the study involving DCVax [31]. While the associated phase III trial failed to show significant improved OS/response, antigen-loading for DC has since become a robust avenue for immunotherapy that involves exposing APCs to peptides, nuclei acids, and cell lysates [32].

Combining DC therapies with autologous glioma stem cell lysates, several phase II trials have found increased OS in select subsets of patients—especially those with IDH-WT/TERT-mutant and low B7-H4 expressing tumors (NCT01567202, NCT03548571) [33]. Another phase I/II trial explored the application of fusion cells, cells that are composed of both DC and tumor cells; the implication of this hybrid cocktail allows for the vaccine to retain both nuclei to theoretically enhance immunogenicity and produce a variety of tumor antigens [34].

Of note, vaccine formulations and protocols have been modified to optimize DC efficiency via improved patient selection. For instance, the original DC vaccine trials included recruitment of patients with large residual disease despite maximal tumor resection and were therefore thought to be exposed to an excessively immunosuppressive environment. As such, a new trial has been designed to recruit those with gross total resection or minimal residual (< 5 cc) tumor volume and utilize mature DC to maximize immune response (NCT03395587). Meanwhile, another phase III trial is evaluating if DC vaccines with tumor lysates have an immediately accessible role for patients with recurrent GBM and thus, offer it in combination with Avastin treatment (NCT04277221). A list of clinical trials involving DC vaccines can be found on Table 1.

Overall, vaccines offer a strong appeal for GBM because of their broad targeting capabilities. However, our prior experiences with large clinical trials have underscored how vaccine applications in GBM are uniquely affected by the tumor’s low number of somatic mutations, made more concerning by its ease of antigen escape [35]. Nevertheless, vaccine-based immunotherapy continues to produce robust studies. There have been an increasing number of trials combining vaccines with anti-PD-1. In one study, DNA plasmid therapy with INO-5401 (WT-1 targeting plasmids) and INO-9012 (IL-12 targeting plasmids) were used in conjunction with Cemipilmab (anti-PD-1) in hopes of boosting immune responses (NCT03491683). In another phase II study, SurvVaxM, a cancer vaccine involving Survivin, was combined with anti-PD-1 to try and improve progression-free survival in patients with recurrent GBM who either failed conventional or anti-PD-1 therapy. Current trials like these continue to assess the durability of antigen combinations and sources, and future emphasis on enhanced and rapid immunogenicity of these vaccine-based treatments will continue to be particularly important.

CAR T Cell Therapy

Due to the inactivation and relatively low prevalence of infiltrating cytotoxic T cells in GBM, strategies involving activation and recruitment of the T cell compartment have been investigated—particularly through the use of CAR T cell therapies [36]. Notably, many of the studies involving CAR T cell therapy have been inspired by the successful FDA-approved treatments in leukemia and lymphoma [37]. In brief, CAR T operates under the premise of antigen presentation for T cell activation. In normal immune responses, T cells require two signals for activation: (1) binding of the major histocompatibility complex (MHC) of APCs to the T cell receptor and (2) co-stimulatory binding of a second activating complex [38]. However, one of the adaptive tumor responses seen with GBM is loss of MHC expression, which in turn prevents activation of T cells through loss of signal 1. As such, in CAR T cells, a chimeric antigen receptor replaces the wild type T cell receptor and is able to bind to tumor antigens irrespective of MHC expression [39]. Furthermore, genes that produce co-stimulatory molecules such as CD28, 4–1BB, or Ox40 as well as additional immunostimulatory cytokines can be linked to the CAR gene, thereby boosting immune response [40]. By targeting tumor-specific antigens, activation of CAR T cells aids in T cell recruitment and anti-tumor response.

There are several clinical trials underway that involve CAR T cell targets for GBM [36]. Upregulation of interleukin-13 receptor alpha 2 (IL13Rα2) occurs in over half of GBM tumors, making it a viable target for activation of T cells. IL13Rα2 is thought to function as a decoy receptor by binding interleukin-13 (IL-13) more avidly than the immune activating IL13Rα1 receptor, which normally activates STAT6 to induce anti-tumor response [41]. Administration of CAR T cells that specifically target IL13Rα2 is hypothesized to have increased available IL-13 for IL13Rα1 activation and improved median OS in patients with GBM without severe toxicity [42••]. However, efficacy limited in recurrent presentations due to antigen loss and T cell depletion [43]. Currently, there are ongoing trials involving IL13Rα2-specific CAR T cells combined with checkpoint blockade (NCT04003649).

Another proposed application of CAR T cell therapy expands on previous work involving the EGFRvIII receptor [36]. As mentioned previously, clinical trials with EGFRvIII targeting inhibitors have shown lack of clinical efficacy [44•]. Similarly, when utilizing EGFRvIII as a specific target for CAR T cells, EGFRvIII-CAR T cells demonstrate limited anti-tumor response as well as targeted antigen downregulation. While IV delivery of EGFRvIII-CAR T cells in patients with GBM have shown detectable expansion of CAR T-EGFRvIII cells in the systemic immune milieu with trafficking of these cells to GBM, there was also increased expression of regulatory T cells [45]. While multiple phase I/II trials have demonstrated lack of clinical benefit including a phase I trial involving third-generation EGFRvIII-CAR T cells after systemic chemotherapy and interleukin-2 exposure [46], the specificity and robust immune activation associated with EGFRvIII-CAR T cells have spurred ongoing interest in this molecule as a CAR T cell target. Several trials exploring combination therapy including EGFRvIII-CAR T cell with anti-PD-1 checkpoint blockade are still underway (NCT03726515).

As with vaccines, CAR T therapies must find a way to overcome GBM heterogeneity. Thus, current efforts to engineer multiple antigen-targeted CAR T cells or identifying appropriate combination therapies are of high priority. Additionally, no phase III trial has been published to date. Given how young this therapy is relative to other strategies, the safety profile of proposed CAR T regimens in humans is also just being fully recognized.

Viral Immunotherapy

Several viral-inspired strategies can be subdivided into those involving viral antigens, oncolytic viruses, and viral vector gene therapies [47]. In regard to viral antigens, the most recent examples capitalize on the high specificity of cytomegalovirus for GBM. Cytomegalovirus antigens such as pp65 are found in over 90% of GBM, but not in normal brain [48]. In one of the most successful examples of pp65-targeted therapies, a phase I trial pulsed autologous DC cells with pp65 mRNA and, when used in combination with GM-CSF and TMZ treatment, reported a median OS of 40.1 months [48]. There is currently a phase II trial underway to further validate this strategy and its adjuvants (NCT03927222).

Oncolytic viral therapy can also be used to selectively infect tumor cells and destroy them via lytic machinery. While there are currently no active phase III trials, several candidate viruses have been studied including adenovirus, herpes simplex, reovirus, parvovirus, vaccinia, and Newcastle disease virus [47]. One of the greatest challenges involving oncolytic viruses involve achieving adequate replication for tumor cell apoptosis without triggering an excessive inflammatory response that results in their early elimination [49•]. Genetic modifications and combination regimens have been proposed to avoid premature viral clearance and sustained oncolytic response. Systemic chemotherapy, radiation, and steroids are known to temporarily suppress the innate immunity for viral infection and dissemination, which lends cautious optimism towards pursuing fine-tuned combination therapy [50, 51].

One of the latest oncolytic candidates is DNX-2401, a conditionally replicative adenovirus, designed to specifically target tumor cells based on their presence of retinoblastoma mutations. Incorporation of an E1A gene mutation in its Rb-binding domain promotes release of transcription factor E2F from the Rb-E2F complex. The displaced E2F can then drive DNA synthesis for viral replication and oncolysis [52]. Additional incorporation of αVβ3 and αVβ5 integrins allows for enhanced infection rate of the targeted cancer cell. Phase I trial results showed greater than 3-year survival for 20% of patients [53]. Other studies are ongoing for herpes simplex virus type I, which has been extensively studied for its safety and ease of genetic modifiability. Of promise is the oncolytic HSV-1 G207 immunovirotherapy for pediatric high-grade gliomas, in which a phase 1 trial examined stereotactic infusion of the HSV-1 G207 with resultant decrease in immunosuppression and increase in infiltration of T cells and median overall survival [54]. Ongoing genetic modifications involving oncolytic viral therapy are being sought to enhance its replicability [55].

Recently, there has been significant interest in virotherapy strategies with three phase III clinical trials to date, although these have still demonstrated limited clinical efficacy. Most recently, Toca 511 was evaluated in a phase III trial in which a retrovirus incorporated cytosine deaminase within infected cells for subsequent prodrug metabolism. The prodrug is converted to 5-fluorouracil locally at the tumor site, causing impaired tumor cell replication [56]. Notably, it exhibited replicative potential that spread beyond the initial injection site [57, 58]. Unfortunately, the recent phase III trial (NCT02414165) involving 271 patients with selected exposure to Toca 511 failed to demonstrate improved OS compared to standard of care (11.1 months vs. 12.2 months) [59].

Some new phase II trials are incorporating gene therapy with oncolytic viral targeting to stimulate the immune environment. One such therapy involves isolating hematopoietic stem cells for lentiviral transduction that causes tumor-infiltrating macrophages to upregulate expression of the immunogenic cytokine IFN-α2. This in turn results in an immunological cascade that inhibits pro-tumoral angiogenesis at the tumor site with minimal systemic side effects [60]. The previously described DNX-2401 vaccine has also been studied in a modified form, DNX-2440, which adds the capability of transducing the OX40L gene into infected tumor cells. This causes increased expression of the OX40 ligand, which can directly stimulate T cell activation (NCT03714334) [61]. Recently, a phase II study involving DNX-2401 in combination with anti-PD-1 showed durable ongoing responses with increased overall survival; the results of this study have spurred interest in a randomized controlled phase III trial [62•].

Importantly, over the last few years, our understanding of the GBM microenvironment has enabled viral therapy to incorporate new and relevant genetic constructs. Their direct involvement with the protein expression machinery is particularly powerful, and early trials are encouraging. Nevertheless, viral and genetic uptake efficiencies will need to be improved, and larger trials for these creative vectors will be greatly anticipated.

Combinatorial Perspectives in Immunotherapy

Combination immunotherapy has become a hallmark of checkpoint blockade, with additional checkpoint blockade involving anti-CTLA-4 with anti-PD-1 demonstrating significantly boosted response in melanoma and NSCLC. While combination therapy has increased side effect profiles including pneumonitis, colitis, vitiligo, and hypophysitis [63], its increase in efficacy by targeting multiple immunosuppressive pathways holds promise for heterogeneous tumors such as GBM. In preclinical models of GBM, there have been several studies that demonstrate synergistic response from combination checkpoint blockade involving TIM-3 [64], LAG-3 [65], and IDO-1 [66]. Of note, several early phase studies on IDO inhibitors in primary brain tumors have just completed recruitment (NCT02502708, NCT02052648, NCT02327078).

Furthermore, combination therapy with stereotactic radiation has been shown to boost immune response in checkpoint blockade, likely from release of new tumor antigens as well as increased infiltration of cytotoxic T cells and decreased immunosuppressive regulatory T cells in the tumor microenvironment [67].

It should be reiterated that much of the progress in immunotherapy for GBM has been transposed from other cancers that do not have similar barriers of immunosuppression. Recently, efforts have been made to re-establish the unique immune milieu of GBM by using single-cell sequencing data and genomic analysis. Preclinical studies have shown that in GL261 glioma-bearing mice, there is expanded heterogeneity of myeloid cells involving microglia as demonstrated by upregulation of P2ry12, Sparc, Tmem119, Gpr34, Selplg, and Cx3cr1, infiltrating monocytes/macrophages as shown through increased expression of Ly6i, Ly6c2, and Ifitm3, and CNS border-associated macrophages with Apoe, Ms4a7, and Mrc1 [68]. While immunotherapy for other cancers involving checkpoint blockade have traditionally examined targeting of T cells such as cytotoxic CD8 + T cells, ongoing research is shifting the emphasis on targeting the myeloid compartment for reversing immunosuppression [69].

In light of the predominance of myeloid cells in GBM, several studies have examined combination therapies that target the myeloid compartment. Activation of toll-like receptor (TLR) 3 by polyinosinic-polycytidylic acid poly(I:C) on both microglia and macrophages results in production of glioma specific toxic soluble factors [70]. Exposure to poly(I:C) to other APCs such as dendritic cells also results in reversal of immunosuppression with increased recruitment of T cells to the tumor site that when combined with anti-PD-1 therapy result in increased median OS in mice [71]. Similarly, targeting TLR9 on microglia and macrophages with oligodeoxynucleotides containing CpG motifs (CpG-ODN) result in decreased tumor size without high toxicity in murine models [72]. However, CpG-ODN failed to show improved progression-free survival in a phase II clinical trial (NCT00190424) [73].

Additionally, inhibitors of colony stimulating factor-1 receptor (CSF1R) have shown promise in preclinical and clinical models for targeting myeloid cells in GBM. In murine models, exposure of the myeloid compartment to anti-CSF1R showed shifts in macrophage gene signatures away from immunosuppressive phenotypes as well as consequent reduction in GBM growth [74]. However, it was found that ongoing use of CSF1R inhibitors resulted in acquired resistance to anti-CSF1R; a follow-up study focused on combining the myeloid-activating potential of anti-CSF1R with tyrosine kinase inhibitors to reduce tumor volume, with significant and lasting anti-tumor efficacy [75].

One of the barriers to human translation of GBM therapy is the discordance of tumor heterogeneity between human and murine models; in murine models, GBM is characterized by clonal expansion of the same glioma cell line, while patient tumor samples often exhibit marked heterogeneity of their tumor phenotype. As such, treatments that may work for a particular GBM phenotype may not have a durable response for another subtype. Therefore, studies have begun using single-cell sequencing to identify new biomarkers that are common to a variety of different GBM molecular signatures. One recent study examined 2343 GBM tumor and 1246 GBM peripheral cell samples that yielded 31 common genes, some of which involves tumor metastasis and progression (TMSB4X/Tβ4, which involves the immunosuppressive TGFβ cytokine), tumor activation and migration (IPCEF1), and overall poor prognosis (S100A10, a calcium-binding protein) [76]. Additional studies examining the unique gene signatures and expression patterns of GBM will need to be done for further pathways analysis in order to better understand new combinatorial targets for immunotherapy.

Conclusion

The current state of immunotherapy for GBM is one of ongoing discovery. While previous strategies that have worked for other malignancies have shown limited clinical efficacy for GBM, each study has brought us closer to understanding the immunosuppressive nature of GBM. The paradigm of GBM as a “cold” tumor, one that does not respond to traditional avenues of immunotherapy such as checkpoint blockade or vaccination efforts, indicates that its adaptive tumor response is beholden it to a different standard of immunosuppression. As such, efforts are being made to capitalize on the immune milieu specific to GBM including targeting the predominant myeloid population, finding new ways to vaccinate against a relatively low diversity of neoantigens, and fundamentally understanding immune signaling pathways through single-cell sequencing and genomic analysis. However, we remain optimistic that with each spark of knowledge we gain from ongoing clinical trials as well as new groundbreaking preclinical studies, our repertoire of knowledge regarding the cold nature of GBM will continue to grow and result in the fruition of new safe and effective treatments.