2. Current Immunotherapy Options and Developments
The human immune system is a complex regulatory environment that must constantly be able to distinguish between “self” and foreign matter. The immune system can be split into “innate” and “adaptive” immunity. Innate immunity does not improve with repeated encounters and consists of phagocytic cells (neutrophils, monocytes) and pro-inflammatory cells (eosinophils, basophils, and mast cells)
[3]. Adaptive immunity learns and improves upon repeated exposure to pathogens. The main players in adaptive immunity are B and T lymphocytes, which produce antigen-specific immunoglobulins and induce foreign cell lysis
[3]. Following activation, part of the immune system′s natural response is to return the hyperactive immune response to basal levels. Cells such as regulatory T-cells (Tregs) release anti-inflammatory cytokines leading to a diminished immune response
[4]. Similarly, cell–cell signaling via inhibitory immunoreceptors such as PD-1, CTLA-4, LAG3, TIM3, TIGIT, and BTLA can attenuate an upregulated immune response
[5]. The most promising immunotherapy approaches to treating glioblastoma are immune checkpoint inhibition
[6][7], T-cell transfer therapy
[8], vaccination
[9], and oncolytic virus therapy (OVT). These methods harness the immune system to recognize and focally target tumor cells.
2.1. Immune Checkpoint Inhibitors
Immune checkpoint inhibitors (ICIs) avert the inactivation of CD8+ T-cells by preventing checkpoint receptors from binding with their ligands (
Figure 1A). The critical immune checkpoint targets for ICIs in glioma include programmed cell death protein-1 (PD-1), programmed cell death ligand-1 (PD-L1), and cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4). In the canonical pathway, when the PD-L1 ligand on the target cells interacts with the PD-1 receptor on the T-cells, intracellular tyrosine residues on the PD-1 cytoplasmic region lead to the recruitment of Src homology 2 domain-containing protein tyrosine phosphatase-2 (SHP-2)
[10]. This causes spleen tyrosine kinase (Syk) and phospholipid inositol-3-kinase (PI3K) to be phosphorylated, resulting in T-cell exhaustion and a suppressed immune response
[10][11]. Glioblastoma cells can co-opt this machinery by overexpressing PD-L1, thereby evading the immune response
[12]. Through a similar mechanism, glioblastoma cells also can upregulate CTLA-4, which promotes T-cell anergy through blockade of the B7/CD28 co-stimulatory signal
[13].
Figure 1. (
A) Immune checkpoint inhibitors bind to and inhibit immunosuppressive molecules on either T-cells or tumor cells. This dampens tumor cells′ ability to evade the immune system. (
B) (1) In tumor-infiltrating lymphocyte (TIL) therapy, T-cells from the tumor microenvironment are isolated following surgical resection. (2) Isolated T-cells are clonally expanded by using IL-2 stimulation. (3) Expanded T-cells are reintroduced to the patient. (
C) (1) In the vaccine approach, a resected tumor biopsy is taken from the patient and sequenced to identify neoantigens. (2) Neoantigens are then delivered via a vaccine. (3) At the site of injection, neoantigens stimulate antigen-presenting cells (APCs). (4) In the lymph node, APCs present T-cells with neoantigens. (5) Activated T-cells attack cancer cells. Created with
BioRender.com.
2.2. T-Cell Transfer Therapies
T-cell transfer therapy or adoptive T-cell therapy is a type of immunotherapy that encompasses two main approaches: tumor-infiltrating lymphocyte (TIL) therapy and chimeric antigen receptor (CAR) T-cell therapy (Figure 1B).
In TIL therapy, T-lymphocytes invading the TME are collected via routine biopsy or surgery, isolated using fluorescence-activated cell sorting (FACS), and then selectively expanded using IL-2 stimulation
[14][15][16]. The logic behind this approach is that T-cells found in or near the tumor already have a “proven track record” for identifying cancerous cells, but there are too few of them to overcome immunosuppression. Moreover, TIL therapy significantly reduces off-target effects due to their inherent specificity to the tumor
[16]. Mathewson et al. performed single-cell transcriptome sequencing in a group of patients with isocitrate dehydrogenase (IDH)-mutant and IDH-wildtype glioblastoma
[17]. They described the potential effectors of anti-tumor immunity in a population of cytotoxic TILs expressing several natural killer (NK) cell genes, including the CD161-encoding gene KLRB1. The inactivation of KLRB1 or antibody-mediated CD161 blockade resulted in increased T-cell cytotoxicity against tumor cells in vitro and an enhanced response in vivo.
CAR T-cell therapy introduces synthetic T-cell receptors into T-cells, which confer the ability to recognize tumor-specific surface antigens and initiate an MHC-independent immune response
[18][19][20]. CAR T-cell therapy has had great efficacy in hematogenous malignancies but has been difficult to implement in solid tumors due to the immunosuppressive environment of the TME
[2][19]. Moreover, solid tumors lack highly specific surface antigens, which can lead to numerous off-target effects when using CAR T-cell therapy. Two small Phase I trials tested CAR T-cell therapy in EGFRvIII-positive recurrent glioblastoma. Although EGFRvIII-targeted CAR T-cells found their way from peripheral blood to the tumor, no meaningful response was detected
[21][22]. This lack of response to anti-EGFRvIII CAR T-cells may be attributable to the significant intra- and inter-tumoral heterogeneity of EGFRvIII expression in glioblastoma as well as to adaptive changes in the local TME, which include changes in antigen expression over time. For instance, following treatment, EGFRvIII was lost in a group of patients.
2.3. Vaccination
Tumor vaccines elicit an immune response against one or several tumor antigens (Figure 1C). Vaccines usually consist of peptides or proteins, but may also constitute antigen-laden dendritic cells. Immunostimulants such as poly ICLC are often co-administered with tumor vaccines to enhance adaptive immunity.
In a single-arm, multicenter, open-label Phase I trial performed in patients with newly diagnosed Grade 3 and 4 IDH1-mutant astrocytoma, an IDH1-specific peptide vaccine induced an immune response in 30 out of 32 (93.3%) patients
[23]. The 3-year progression-free and overall survival rates were 63% and 84%, respectively. The 2-year progression-free rate among patients with an immune response was 82%, while the two patients without an immune response had tumor progression within 2 years of diagnosis.
Another vaccine approach involves a vaccination against survivin, an antiapoptotic protein expressed by many tumor types
[24][25][26]. Survivin expression in GBM has been associated with increased recurrence, chemotherapy resistance, and poor overall prognosis
[24][25][26][27][28]. The SurVaxM vaccine contains a synthetic long peptide mimic that spans the human survivin protein sequence; it expresses MHC Class I epitopes and stimulates the MHC Class II-restricted T-cell responses required for cytotoxic CD8+ T-cell activity against tumors
[24]. A Phase I trial of SurVaxM against recurrent GBM demonstrated no serious adverse events and prolonged overall survival following vaccination (86.6 weeks) compared with historical overall survival (30 weeks)
[24]. A subsequent study identified that glioma patients routinely expressed elevated serum levels of CD9+/GFAP+/SVN+ exosomes, associated with tumor progression, compared with healthy controls
[29]. Patients treated with antisurvivin therapy showed decreased levels of these exosomes. Monitoring of CD9+/GFAP+/SVN+ exosomes may be a promising adjunct to the use of MRI in disease surveillance. Current trials are underway to evaluate SurVaxM’s efficacy in newly diagnosed GBM
[30]. However, identifying plausible new vaccine targets for GBM remains difficult due to the heterogeneity of GBM tumors.
2.4. Oncolytic Virus Therapy
In recent years, the use of OVT has shown promise in the treatment of GBMs. OVT utilizes intratumoral delivery of viral vectors to either deliver oncolytic gene therapy into the TME or to cause direct cytotoxicity through viral infection and replication
[31][32]. OVT also has pro-immunogenic effects due to the induction of immunogenic cell death (ICD) in infected tumor cells. In ICD, the destruction of tumor cells by OVT leads to the release of antigenic molecules into the TME which both recruits and activates local dendritic cells, with the subsequent stimulation of specific T-cells
[32].
The earliest trials of oncolytic therapy in GBM used murine fibroblasts to deliver the replication-defective herpes simplex virus 1 (HSV1) thymidine kinase (tk) gene to GBMs, which conferred increased chemosensitivity to antiviral agents such as acyclovir, ganciclovir, and valganciclovir
[31][33]. However, this trial failed to show prolonged survival in the OVT group, which was hypothesized to be the result of low gene transduction rates due to the nonmigratory nature of murine fibroblasts
[33]. More recently, a genetically engineered replication selective HSV1 virus, G207, has shown safety and efficacy in clinical trials. G207 contains a deletion of the diploid γ134.5 neurovirulence gene and has viral ribonucleotide reductase (UL 39) disabled by the insertion of Escherichia coli lacZ. This allows for conditional replication in tumor cells while preventing the infection of normal cells
[34]. A Phase I trial showed a median survival of 15.9 months in 13 GBM patients treated with intratumoral G207, with no evidence of HSV encephalitis
[35][36]. A separate Phase I trial demonstrated the safety of G207 administration in conjunction with radiotherapy
[35], while a more recent trial showed its safety in the treatment of pediatric high-grade gliomas
[37]. HSV-vector mediated delivery of gene therapy offers significant promise in the treatment of GBM, and a current Phase I trial is investigating the use of a new drug, rQnestin34.5v.2, after a preclinical study suggested its low toxicity to humans
[38][39].
Another development in OVT was the use of intratumoral injection of aglatimagene besadenovec (GliatakTM), a replication-defective adenovirus vector-mediated delivery of HSV1-tk (AdV-tk), in conjunction with subsequent valaciclovir therapy. Phase I trials of Gliatak conducted by Chiocca and colleagues demonstrated the safety of the therapy and an impressive radiographic response
[40], while the Phase II trial showed a statistically significant improvement in the MOS of GBM patients treated with Gliatak after gross total resection (GTR) compared with patients treated with the standard of care after gross total resection (25.1 months vs. 16.3 months, respectively)
[41]. Importantly, the survival benefit was even further improved at 2 and 3 years compared with the standard of care treatment, but no difference was noted if the resection was subtotal
[41]. However, another Phase III clinical trial named the Aspect trial, which utilized AdV-tk, showed no significant improvement in overall survival when patients were treated with intratumoral injections of AdV-tk compared with the standard of care treatment group
[42]. It should be noted that the ASPECT trial had uneven use of temozolomide, and radiotherapy was not administered concomitantly with the gene therapy
[42]. Yet another Phase I trial evaluated the use of a human interferon-β-expressing adenovirus vector (Ad.hIFN-β). Intratumoral injection of Ad.hIFN-β was associated with a dose-related induction of apoptosis within tumors, but several patients experienced adverse effects and one patient experienced two serious dose-related adverse effects
[43]. Ultimately, further investigation into adenovirus vectors is required.
The use of a live attenuated form of poliovirus has recently been studied as well. A Phase II clinical trial demonstrated that PVSRIPO, a live attenuated poliovirus Type 1 vaccine with its cognate internal ribosome entry site replaced by that of human rhinovirus Type 2 conferred an overall survival benefit
[44]. Specifically, this randomized controlled trial (RCT) showed that the group treated with PVSRIPO had an overall survival rate of 21% at both 24 and 36 months, compared with 14% and 4% in the control group, respectively
[44]. The foreign ribosomal entry site on PVSRIPO causes neuronal incompetence and ablates neurovirulence
[45]. The effects of PVSRIPO are mediated by CD155, a Type 1 transmembrane glycoprotein receptor that is more commonly known as the poliovirus receptor
[44][46][47][48]. CD155 is almost ubiquitously upregulated in solid tumors, including GBM, and it regulates natural killer (NK) cells and is part of the Ig-superfamily adhesion family response for cell motility and invasiveness
[46][48][49]. When the PV capsid binds to CD155, the capsid protein is extruded and ultimately initiates the transfer of the viral RNA genome to the cytoplasm, then subsequently allows for the translation of the RNA and mediates the viral oncolytic effects
[50]. Additional Phase II studies for PVSRIPO in conjunction with additional drugs are underway, with Phase III studies likely to commence in the foreseeable future.
Translating the success of early Phase I and Phase II trials to widespread clinical use has been challenging. Phase I and Phase II trials of the drug Toca 511 (Vocimagene amiretrorepvec), a γ retroviral replicating vector encoding a transgene for an optimized yeast cytosine deaminase, demonstrated both early safety and efficacy, with prolonged overall survival and complete responses in recurrent high-grade glioma and GBM compared with accepted survival rates in the literature
[51]. However, in the Phase III arm of the clinical trial, the overall survival for patients treated with Toca 511 was 11 months compared with 12 months in the patient group receiving the standard-of-care treatment, with no significant difference between the two groups
[52]. Toca 511′s Phase III failure underscores how challenging the introduction of new GBM therapies into the market has been. Several obstacles underlie these challenges in translating OVT into widespread clinical use. Pre-existing antibodies and the circulating complement in the peripheral vasculature may neutralize OVT particles before they are successfully delivered into the TME
[53]. Moreover, uptake into nontarget organs (e.g., the liver) is a common barrier to efficient delivery
[53]. As with the other therapeutic modalities, the BBB is a major obstacle to the effective delivery of any exogenous therapeutics.
3. Strategies to Enhance Immunotherapy’s Effectiveness
Physical modalities, such as noninvasive microbubble-enhanced focused ultrasound (MB-FUS) (
Figure 2), can safely and transiently alter the permeability of the BBB/BTB without directly causing changes in the tumor cells. This technology has been demonstrated preclinically in numerous species, including nonhuman primates
[54][55][56][57][58][59][60][61][62][63][64], and in multiple successful Phase I and IIa clinical trials executed by several different groups
[65][66][67][68][69][70][71][72][73][74]. Ultrasound-mediated BBB disruption has been observed in normal brains
[54][75], brains affected by neurodegenerative diseases (e.g., Parkinson′s disease and Alzheimer′s disease)
[74], and brains with tumors
[65][66][76]. Together, these studies have demonstrated the robustness of the technique. The temporary increase in permeability lasts between a few hours and several days, and depends on the type and dose of the microbubbles used and the ultrasound parameters
[77][78][79][80][81][82][83]. The increased permeability occurs both through the opening of the tight junctions of the endothelium and through increased transcytosis
[84]. These effects are nucleated by the gentle volumetric oscillation of the microbubbles when they are exposed to low-amplitude ultrasound, with the ultrasound amplitude being within the range used for diagnostic ultrasound imaging. Care must be taken to identify the appropriate ultrasound amplitude. If the amplitudes are too low, the barrier will not be disrupted, and for amplitudes that are too high, petechial hemorrhage may occur
[85]. The emissions from the oscillating microbubbles can be used to identify the appropriate amplitudes in real time, providing patient- and treatment-specific guidance and control
[86][87][88][89][90][91]. Phase I clinical trials have demonstrated the safety of this technology. Oscillation of the microbubbles not only increases the permeability of the blood–brain/tumor barrier but can also establish a convective flow that enhances the delivery of chemotherapeutics
[92][93][94][95].
Figure 2. Cartoon illustrating how microbubbles can induce a focal disruption or opening of the blood–brain barrier (BBB), thus enabling the delivery of a biologic such as a monoclonal antibody. Microbubbles flow through the normal vasculature or vasculature supplying the glioblastoma tumor microenvironment (TME). Only the microbubbles in the vasculature exposed to ultrasound insonation enable BBB/BTB disruption following ultrasound insonation. Created with
BioRender.com.
Delivery of a wide range of potential therapeutics has been demonstrated in preclinical models, including chemotherapeutics
[72][96][97], adenoviruses
[98][99], antibodies
[100][101], nanoparticles (NPs)
[94][102][103][104], and whole cells
[105][106]. Guo et al. demonstrated that NPs as large as 50 nm can achieve significant extravasation into the TME with the application of focused ultrasound
[94]. NPs have a wide variety of formulations. Guo et al. used them as a lipid-based encapsulation method to protect therapeutic payloads from degradation as they traversed the vasculature to the TME. Their study also demonstrated that focused ultrasound delivery of RNA-loaded NPs significantly downregulated the expression of an oncogenic mRNA
[104][107]. NPs have a use in immunotherapy, as they can be combined with anti-PD-L1 antibodies to focally target drug delivery to the TME
[108][109][110]. Similarly, groups have used NPs to deliver CAR-T-cells in a mouse model of glioma
[109][111]. NPs could also have a use in the delivery of vaccines or OVT, given the previously discussed barriers to the effective delivery of these therapies. Ultrasound-mediated delivery to specifically induce immune modulation and therapy has been previously described
[112]. Approaches include the passage of IL-12
[113], immune checkpoint inhibitors
[55][67][114][115][116][117], and natural killer cells
[118]. In addition to transient disruption facilitating the diffusion of therapeutics into the brain, disruption of the BBB can also enable the release of tumor biomarkers, which can assist in assessing the treatment response
[119].
While there has been a significant and deserved emphasis on focused ultrasound to transiently permeabilize the blood–brain barrier, ablative ultrasound therapies can also enhance immune checkpoint inhibition
[120]. Thermal ablative ultrasound therapy uses high-intensity focused ultrasound to increase the local temperature to 60 °C or higher to induce coagulative necrosis. It has been safely used in the brain to ablate neuronal tracks underlying the pathogenesis of essential tremor
[121][122][123] and also to treat chronic neuropathic pain
[124]. Preclinical evidence has indicated that ultrasound thermal ablation may work adjunctively with immune checkpoint inhibitors
[125][126]. Furthermore, mechanically ablative ultrasound therapy (histotripsy) can also potentially enhance immune checkpoint inhibitors by stimulating nonimmunogenic “cold” tumors into becoming “hot” immunogenic tumors
[127][128]. Common obstacles to this treatment approach are the interference of uniform ultrasound wave propagation through bone and gas, and organ movement during treatment, leading to collateral tissue damage
[129].