Background

Glioblastoma (GBM), WHO grade IV glioma, is the most devastating brain tumor in adults [55]. O’Rourke et al. [41] reported the first study in humans of intravenous delivery of a single dose of EGFRvIII-CAR T cells in 10 recurrent GBM patients which showed safety and limited anti-tumor response as well as targeted antigen downregulation. Goff et al. [56] reported a pilot phase I trial with third generation EGFRvIII-CAR T cells administered after lymphodepleting chemotherapy and intravenous interleukin-2 injections. This study did not show significant toxicity but failed to show clinical efficacy. Another design of EGFRvIII-CAR T cell using humanized scFv again proved safe but also failed to show clinical benefit [41, 57]. Currently, there are six EGFRvIII-CAR T cell clinical trials ongoing with two in combination with chemotherapy or ICI (Table 1).

HER2

HER2 is a member of the EGFR family, also named as ERBB2. Increased levels of HER2 protein in GBM patients is linked to poor survival [14, 58, 59]. Unlike EGFRvIII, which is unique to GBM, HER2 is overexpressed in many cancer types including breast, ovarian and GBM; it also is expressed in some normal tissues, leading to safety concerns. An early case report with HER2-CAR T cell therapy has documented a metastatic colon cancer patient who experienced respiratory distress within 15 min of HER2-targeting CAR T cell administration and died 5 days after treatment. Serum samples after cell infusion showed marked increases in interferon γ (IFNγ), granulocyte macrophage-colony stimulating factor (GM-CSF), tumor necrosis factor-alpha (TNF-alpha), interleukin-6 and -10 (IL-6 and IL-10), consistent with a cytokine storm [60]. Since then, efforts have been made to overcome these safety concerns [61]. Liu et al. [62] showed that lowering the binding affinity between scFv and HER2 may increase the differential binding of CAR T cells to tumor versus normal tissue in preclinical cancer models, providing a good strategy for targeting the antigens that are not specific to solid tumors. A more recent clinical trial (NCT00902044) reported no significant toxicities of a second generation HER2-specific CAR T-cell therapy for sarcoma [63]. Compared with the original 4D5 HER2-CAR which binds to the juxtamembrane region of HER2 protein and caused severe toxicity, this trial used a FRP5-HER2 CAR which recognizes a discontinuous epitope between residues 11–169 of HER2 that is more distant from the cell surface. Thus, the binding site of the epitope also determines the activity of the CARs. As for GBM, investigators at Baylor College of Medicine have conducted a clinical trial (NCT01109095) to evaluate the safety and efficacy of HER2-specific CARs using virus-specific T cells (CAR-VSTs) [59]; previous studies have shown that virus engineered cytotoxic T cells demonstrate better persistence and T cell expansion through appropriate co-stimulatory signaling activation [64]. In this trial, overexpression of FRP5 HER2-CARs in virus-specific T cells not only improved the specific targeting of HER2+ tumor cells, but also prolonged the HER2-CAR VST cell persistence and antitumor activity. While this trial has demonstrated treatment safety, the clinical benefit was limited.

EphA2

EphA2 protein is overexpressed in gliomas and is associated with malignancy, thus it became a good molecular target in GBM [65, 66]. Preclinical studies demonstrated that a second generation EphA2-CAR T cell induced glioma xenograft tumor regression in vivo [46, 67]. So far one clinical trial was initiated to evaluate the safety and effectiveness of CAR T cell immunotherapy in treating patients with EphA2+ malignant glioma but has been withdrawn recently for unknown reason (NCT02575261).

GD2

GD2 is a glycosphingolipid expressed at low levels on the surface of healthy cells, but highly expressed on several tumor types including gliomas and is associated with increased tumor proliferation and invasion [68]. Pre-clinical studies have demonstrated a robust antigen-dependent cytokine production and killing of GD2+-positive glioma cells in vitro and in vivo using patient-derived orthotopic xenograft models overexpressing GD2 [16, 17]. Currently, two phase I clinical trials (NCT04099797 and NCT03423992) are evaluating the safety and efficacy of GD2-specific CAR T cells in high grade glioma and diffuse intrinsic pontine glioma (DIPG).

B7-H3

B7-H3 (CD276) is an immune checkpoint molecule and a member of B7 protein superfamily. B7-H3 binds to the majority of neuroepithelial tumors but not to normal glia or tissues, making it a promising target for therapeutics [69]. Clinical trials for two mAbs targeting B7-H3 (8H9 and MGA271) have been shown to be safe and promising for metastatic CNS neuroblastoma and DIPG in children.[70, 71]. Using a B7-H3 (MGA271) 41BBζ CAR, Majzner et al. [18] have reported a robust anti-tumor activity in multiple solid, liquid, and CNS tumor types. In addition, Nguyen et al. [72] demonstrated that B7-H3-specific (MGA271) CD28ζ CAR-T cells have potent anti-tumor activity in U373 glioma model. To date, two GBM-related clinical trials focused on B7-H3 CAR T-cell therapy (NCT04385173, NCT04077866) are in the recruitment phase.

Chlorotoxin

Chlorotoxin (CLTX) is a 36-amino acid peptide first isolated from scorpion venom that specifically binds to GBM but not to normal tissue. Pharmacologically, CLTX binds to and blocks small-conductance chloride channels [73]. Unlike the common CARs that are designed to recognize surface TAAs to kill tumor cells, the CLTX-CAR was designed as a peptide-based CAR to recapitulate the GBM-binding potential of CLTX. With this approach, Wang et al. [19] showed that CLTX-CAR T cells mediate potent anti-GBM activity and efficiently targeted tumors lacking expression of other GBM-associated antigens, resulting in tumor regression in orthotopic xenograft GBM tumor models. Importantly, the CLTX-CAR-T cells exhibited minimal off-target effects, without showing toxicity following systemic or regional delivery into mice. Given the finding that effective targeting by CLTX-CAR T cells requires cell surface expression of matrix metalloproteinase-2 (MMP2), a clinical trial has been initiated for treating MMP2+ recurrent GBM (NCT04214392). More importantly, this study also opened a new avenue to repurpose the use of a natural toxin for CAR T cell engineering.

Other potential CAR T cell targets for GBM

While the current CAR T cell targets continue to be developed and improved, progress also is being made in exploring new targets specific for GBM. Below we review other promising brain TAAs that have been tested pre-clinically.

CD70

CD70 is a type II transmembrane protein and a member of the tumor necrosis factor family. CD70 was not detected in normal peripheral and brain tissues but was constitutively overexpressed in isocitrate dehydrogenase (IDH) wild-type primary low-grade gliomas, and GBMs in the mesenchymal subgroup and recurrent tumors. CD70 is associated with poor survival in GBM patients making it a good candidate for CAR T-cell therapy [74, 75]. Several pre-clinical studies with differently designed CD70-specific CAR-T cells have shown robust anti-tumor response against CD70+ mouse models [74, 76]. While there is no current clinical trial for testing CD70-specific CAR-T cells in GBM, phase I/II clinical trials are underway to study the safety and efficacy of CD70-specific CAR-T cell therapy in B cell cancers (NCT03125577, NCT04429438) as well as pancreatic, renal, breast and ovarian cancers (NCT02830724).

CD133

CD133 is a marker for self-renewing cancer stem cells (CSCs) in solid tumors. CD133+ tumor cells, including GBM, are known to be highly resistant to chemo- and radiotherapy. Recently, CD133 has been identified as a potential CAR T cell target for treatment of GBM. While CD133 also is expressed on hematopoietic stem and progenitor cells (HSPC), Vora et al. [77] demonstrated that intra-tumoral injections of CD133-specific CAR T cells are effective at eliminating GBM and that this treatment does not cause systemic toxicity in humanized mouse models. In addition, in a phase 1 clinical trial (NTC 02541370) CD133-specific CAR T cells showed feasibility and efficacy against advanced metastatic malignancies with no serious adverse effects, further confirming the therapeutic potential of targeting CD133 [78].

MET

MET is the receptor of hepatocyte growth factor (HGF), and a well-known RTK being developed as an anti-cancer target. HGF/MET overexpression frequently occur in GBM patients and is associated with poor prognosis [54, 79, 80]. Using a transgenic mouse model, Qin et al. [81] showed that overexpression of HGF and MET may transform neuro stem cells into glioma stem cells (GSCs), leading to GBM initiation. GBM harboring MET amplification or HGF autocrine activation are sensitive to MET inhibitors in preclinical models [82, 83], recent clinical trials further showed that a combination of MET and VEGF inhibitors (ontarzucimab plus bevacizumab vs. placebo plus bevacizumab) significantly improved progression free survival (PFS) and overall survival (OS) in the mesenchymal subtype of recurrent GBM patients with high tumor HGF expression [84]. These results suggest that MET maybe a good target for CAR T-cell therapy in GBM.

Summary for CAR T cell targets in GBM

Although a large number of CAR T cell targets under development for treating GBM are showing promising preclinical results, limited antitumor response has been observed in clinical trials, largely due to limited T cell persistence and antigen-negative relapses. While technical innovations to improve the CAR T cell expansion, survival, efficacy and safety remain the key next steps to pursue, combination with other therapies also have been comprehensively considered to improve the CAR T cell therapy [85]. Below, we introduce the combination with ICIs as a promising approach to overcome T cell exhaustion which in turn may improve CAR T cell efficacy in clinical trials.

CAR T cell therapy in combination with immune checkpoint inhibitors

The tumor microenvironment is orchestrated by immunosuppressive cytokines, regulatory modulators and co-inhibitory receptors that regulate the responsiveness of immune cells [86]. While active CAR T cells elicit specific recognition and killing activities against tumor cells, the chronic exposure to tumor cells results in T cell exhaustion mediated by immune checkpoint pathway activation that is initiated by tumor cells, leading to a reduced ability to proliferate, produce cytokines and attack the tumor (Fig. 1c). Exhausted T cells upregulate immune inhibitory receptors, such as programmed death ligand-1 (PD-L1) and cytotoxic T lymphocyte-associated antigen 4 (CTLA-4). Blocking this immune suppressive signaling has led to the development of ICIs such as nivolumab and durvalumab (PD-1/PD-L1 inhibitors) and ipilimumab (CTLA-4 inhibitors). These ICIs are FDA-approved for treating several types of cancer including melanoma, hepatocellular carcinoma and lymphoma and have shown significant clinical results [87]. Since ICIs eliminate T cell exhaustion, combining their use may further enhance CAR T-cell therapy efficacy through extended T cell proliferation and sustained tumor-killing activity. A recent case report has shown that administration of the PD-1/PD-L1 inhibitor pembrolizumab after CD19-specific CAR T-cell therapy in refractory diffuse large B-cell lymphoma induced a clinically significant antitumor response, suggesting that the PD-1 pathway may be critical in determining the response to CAR-modified T-cell immunotherapy [88]. However, despite the promising clinical results in other types of cancer, the use of ICIs in GBM clinical trials remain controversial. Nivolumab, a PD-1/PD-L1 inhibitor, failed to prolong overall survival of patients with recurrent GBM, leading to a discussion of whether ICIs may benefit GBM patients after all [89,90,91]. Nevertheless, a recent trial with 35 patients with recurrent, surgically-removable GBM showed that patients who receive neoadjuvant PD-1 blockade, with continued adjuvant therapy following surgery had significantly better overall survival compared to those who received adjuvant or post-surgical PD-1 blockade alone [92]. Another study also suggests that combination of local chemotherapy with PD-1 blockade enhanced antigen-specific T effector cell expansion and improves survival in GBM models [93]. These studies suggest that anti-PD-1 blockade alone may not directly benefit GBM patients but may improve the efficacy of other therapeutics when used in combination. Currently, clinical trials are on-going to evaluate the therapeutic efficacy of EGFRvIII-CAR (NCT03726515) and IL13Rα2-CAR (NCT04003649) T-cell therapy in combination with ICIs such as pembrolizumab or nivolumab in recurrent or refractory GBM patients (Table 1). We anticipate that anti-PD1 blockage may improve the CAR T-cell efficacy for treating GBM patients.

Future perspectives

Over the past decades, we have learned key principles to manage CAR T-cell efficacy: the specificity of the targeted antigen, the sufficient TCR activation by the specific antigen binding domain, the level of T cell activation and longevity, and the hostile micro-environment for T cells to penetrate, which also serves a source of immunosuppressive factors [94]. While many types of CAR T cells are now in clinical trials (Table 1), further optimization with advanced approaches are expected to improve their overall clinical efficacy.

Since genetic modification of CAR vectors is showing promising results in improving CAR T-cell efficacy, gene editing of the T cell is now a new strategy to further improve CAR T cells. Recently, the CRISPR/Cas9 mediated gene editing system has been applied to delete or mutate checkpoint-associated genes in CAR T cells, and has shown improvement in T cell persistence and survival [95,96,97]. Using CRISPR/Cas 9-mediated deletion, Ren et al. [95] generated CD19-CAR T cells simultaneously deficient in TCR, HLA class I molecule and PD1, which showed more potent antitumor activities than non-edited CAR T cells when tested in xenograft mouse models. Notably, the TCR and HLA class I double-deficient T cells further reduced MHC restriction and eliminated alloreactivity and potential graft-versus-host disease (GVHD), thus making the strategy particularly useful for treating different patients. A different study using CRISPR/Cas9-mediated deletion of TCR α subunit constant (TRAC) region, beta-2 microglobulin (B2M) and PD-1 also showed profound anti-tumor activity in vivo [97]. Recently, Li et al. [98] investigated the approaches to improve CAR T transduction efficacy and CRISPR/Cas9 mediated PD-1 deletion using a 2-in-1 lentivirus vector and found that blocking anti-viral signaling in human primary T cells enhanced lentiviral-mediated transduction efficacy of a combinatory CAR/CRISPR vector for targeting HER2 and inhibiting PD-1 at the same time. Most importantly, the first phase 1 clinical trial using CRISPR/Cas9 for multiplex gene editing on T cells (NCT03399448) has demonstrated safety and feasibility in three patients with refractory cancer [99]. Given that the CRISPR/Cas9 approach has been widely applied to functional gene screening and validation, it also opens a new avenue to discover and target novel modulators of immune suppression. Through CRISPR/Cas9-mutagenesis screening, Wei et al. [100] have identified Regnase-1 as a major negative regulator of anti-tumor responses in CD8+ T cells, and that CD19-CAR T cells with deleted Regnase-1 showed greater longevity and a more vigorous anti-tumor response in B16 ovarian and B6 melanoma mouse models. We anticipate that future GBM-specific studies also will explore these novel ideas and approaches, which hopefully will further enhance CAR T-cell proliferation, longevity and long-term anti-tumor activity.

Conclusion

Brain tumors are considered one of the “most-difficult-to- treat” solid tumors. With the most recent advances in immunotherapy, the CAR T-cell therapy has become a revolutionary approach for treating hematological malignancies and it has great potential for brain tumors. This review intends to introduce interpretations of the most relevant papers addressing the use of CAR T cells for the treatment of GBM. We discussed the CAR designs and optimization, the major CAR T cell targets in clinical trials, as well as the strategies developed to improve CAR-T cell efficacy in the context of GBM. We anticipate that future clinical trial designs will not only focus on efficacy and safety, but also on the mechanisms involved in the immune response and resistance, and that the next generation of CAR T cell therapy will become an effective and safe therapeutics for treating malignant GBM.