Current Status and Challenges of Vaccination Therapy for Glioblastoma

Glioblastoma (GBM) is the leading cause of death from primary brain tumors worldwide. It is characterized by the loss of cell-cycle control in glial cells and the uncontrolled growth of astrocytes in the brain. Despite recent advances in GBM treatment, such as chemotherapy, surgery, and radiation, GBM is a lethal brain tumor that can lead to death within 14 to 17 months or less (1, 2). Tumor heterogeneity, tumor “stemness,” complex metabolic reprogramming, and highly invasive behavior all contribute to the initiation, progression, and evolution of GBM (3). There is an unmet need to develop innovative strategies to treat this deadly cancer and improve overall survival (OS).

Increasing evidence positions immunotherapy as an important strategy for this and other malignancies (4). Cancer vaccination is considered a viable therapeutic method for immunotherapy against solid tumors. In-depth mechanistic studies and advances in vaccination platforms have led to improved cancer vaccines with immediate translational relevance. Cancer vaccines boost antitumor immunity by delivering tumor antigens through whole cells, peptides, nucleic acids, and other molecules. One central goal of cancer vaccines is to overcome tumor immunosuppression and boost both humoral and cellular immunity. In theory, vaccination provides an effective means to bolster several arms of the immune system against cancer. Vaccines are developed based on the exact molecular characteristics of the cancer in question, conferring a level of specificity not observed in many modalities. The preclinical success of cancer vaccination, in general, has inspired the development of numerous GBM vaccine platforms.

Despite positive results from the phase I/II clinical trials, no successful phase III clinical trials for vaccination against glioblastoma have been conducted (5). This article summarizes recent findings and ongoing clinical research on GBM immunotherapy by focusing on peptide vaccines, genetic vaccines (DNA, RNA), dendritic, and tumor cell vaccines (Fig. 1). Several barriers to effective cancer vaccine therapy are reviewed, and future perspectives on cancer vaccine development and optimization are discussed.

Peptide vaccines are 8 to 25 amino acids that mimic antigenic epitopes found on tumor cells and often elicit durable antitumor T-cell responses. These short peptides are frequently coupled to large carrier proteins to increase stability, immunogenicity, and efficacy. In general, unconjugated peptide antigens are ineffective and susceptible to enzymatic degradation. Diphtheria toxin, tetanus toxoid, bovine serum albumin or ovalbumin (OVA), and H. influenza protein D are the most common carrier proteins used (6). Peptide vaccines are attractive due to their ease of synthesis, low cost, improved stability, and relative safety in a variety of preclinical studies. However, compared with other forms of vaccination, peptide-based vaccines have not yet been clinically approved for cancer therapy (7).

Peptide design is an important factor influencing efficacy. During a native immune response, activated B cells differentiate into antibody-secreting plasma cells. Commitment to differentiation occurs when variable domains of the B cell receptor (BCR) identify and respond to unique antigenic targets (8). If the peptide is too short, it may not adequately resemble the shape or conformation of the target epitope. It may associate with the MHC of nonprofessional APCs, resulting in poor T cell–mediated immunity (9, 10). In addition, short peptides, unless modified, are more prone to enzymatic degradation and rapid elimination from the body.

Conversely, longer peptides encompass more epitopes and provide better coverage of human leukocyte antigen (HLA) haplotypes. Longer peptides are superior in activating CD8⁺ and CD4⁺ T cells (11). Cancer-related challenges such as tumor heterogeneity and tumor antigen escape can also be addressed with multiepitopes vaccines (7).

GBM expresses several tumor-specific antigens (TSA) and tumor-associated antigens (TAA) that are relevant for vaccine development (12). TSAs are restricted to tumor cells and are often patient-specific. They arise from a variety of genetic alterations that can lead to the formation of novel peptide fragments (i.e., neoantigens), improper posttranslational modifications, or oncolytic viral infections (oncoviruses). TAAs are more abundant than TSAs in tumor cells but can also be expressed in normal tissues (13). There are concerns about redirecting immune cells to TAAs and TSAs in case reactivity is prompted against “self" antigens and precipitates autoimmune events. Interestingly, the overall response to TAAs in clinical trials has been more encouraging than that to TSAs (14).

AXL, Gas6, EGFR, IGFBP2, RAS, B-RAF, PDGFR, NKX2-2, and OLIG2 are frequently mutated or overexpressed in GBM (12, 15). A limited number of mutated gene products (WT1, Survivin, IDH1, and EGFRIII) have been investigated as potential peptide vaccines for GBM patients due to a lack of conservation among GBM tumors (16). We have summarized the ongoing clinical trials in which these peptides are being used as vaccines in GBM patients in Table 1.

A rare form of pediatric kidney cancer is associated with mutations in the Wilms tumor 1 gene (WT1) located on chromosome 11p13. This gene encodes the WT1 protein, a zinc finger transcription factor that promotes cell proliferation and differentiation (17). Overexpression of WT1 promotes carcinogenesis and has been associated with a number of leukemias and solid tumors, including gastric cancer and cancer of the central nervous system (brain and spinal cord). Most GBM specimens overexpress WT1, and immunotherapy against WT1 has been shown to be effective in treating recurrent GBM (17). The National Cancer Institute (NCI) has identified the WT1 protein as a primary validated target for cancer vaccine development. Several clinical trials using peptide vaccines against WT1 have observed therapeutic benefits and reduced detectable WT1 transcript levels (18). IZUMOTO and colleagues recently completed a phase II clinical trial of WT1 vaccine treatment in patients with recurrent GBM who did not respond to conventional therapy (19, 20). Although the response rate of the trial (9.5%) was not particularly high compared with the results of chemotherapy trials, the disease control rate (57.1%) was acceptable and well tolerated in patients with recurrent GBM. The median progression-free survival (PFS), 6-month PFS rate (20.0 weeks), and disease control rate were similar to chemotherapeutic agents, with less of adverse side effects (19, 20). Oji and colleagues found that WT1 vaccination stimulates T-lymphocyte-dependent cellular immune responses and induces antibody-dependent humoral immunity (21). They detected WT1-235 IgG antibodies (unmeasurable before vaccination but abundant in 50.8% of patients) 3 months after WT1 peptide vaccination. Serum WT1-235 IgG levels significantly correlated with longer PFS and OS duration (21, 22). Other clinical trials, such as NCT01291420 and NCT02649582, have also been conducted to evaluate the efficacy and safety of WT1 vaccination in glioblastoma patients (23).

WT1 expression is elevated in a number of malignancies, including GBM, but normal cells can also express low levels (21). This raises the possibility that anti-WT1 T-cell responses, particularly those mediated by high-avidity T cells, may become tolerogenic (24). Recently, Molldrem and colleagues suggested that the generation of low-avidity CTLs may be part of the solution to circumvent immune tolerance in leukemia patients (24, 25). However, low-avidity TCRs for WT1-MHC are less effective than high-avidity T cells in killing CML.

Apoptosis is a form of regulated cell death that occurs under normal and pathologic circumstances, characterized by distinct morphologic changes, as well as biochemical changes, including proapoptotic caspase activation. Caspases are known mediators in the transmission of cell death signals. IAPs (inhibitors of apoptosis) are proteins containing one or more IAP repeats (BIR domains) that interrupt this critical process (26). These proteins perform a variety of biological functions, including caspase degradation through ubiquitination, control of cell-cycle arrest, regulation of receptor-mediated signal transduction, and cell division (27). Survivin belongs to the family of IAP proteins. Survivin expression is regulated by Wnt and EGFR signaling (cross-talk is not required for dysregulated expression). It has a single IAP repeat at the N-terminus and a coiled-coil domain (CC) at the C-terminus. Survivin is overexpressed during embryogenesis but downregulated in most normal, differentiated adult cells (28). Survivin expression is associated with poor prognosis and low OS in CNS malignancies such as gliomas and other cancers. In cancer cells, survivin can disrupt apoptosis, contribute to chemotherapy resistance, support cancer stem cell renewal, and enhance tumor cell invasion (29). Inhibition of survivin has little effect on the regular physiologic activity of healthy cells, implying that survivin-based therapies have a low toxicity profile (29).

Survivin is a potent tumor antigen that binds to MHC class I molecules on the surface of cancer cells, providing a stimulatory ligand for T cells (30, 31). The vaccine could potentiate preexisting immunity as wild-type survivin has low immunogenicity (32). SurVaxM [SVN53-67/M57-keyhole limpet hemocyanin (KLH)], a newly produced peptide vaccine derived from the human survivin protein sequence (including amino acids 53–67), promoted more robust antitumor immune response against tumor cells than the corresponding wild-type survivin peptide (32). SurVaxM treatment with Montanide ISA 51 (as a vaccine adjuvant with immunostimulatory activity) in combination with GM-CSF (sargramostim) was evaluated in phase I clinical trial (NCT01250470) in patients with malignant glioma (32). The adjuvant, sargramostim, is a recombinant human granulocyte-macrophage colony-stimulating factor (GM-CSF) used to improve cancer treatment outcomes and minimize vaccine-associated toxicity (33). Sargramostim is also used to increase the number of WBCs and platelets in the blood (34). The results revealed that treatment was both safe and well tolerated for all nine patients who participated in the clinical trial. The patients had a higher OS rate (12–20 months or more) than the control group (7 months), and most responded positively to vaccination. The vaccine elicited antibody and T-cell responses specific to survivin-expressing cancer cells (32, 35). Following these promising results, a phase II-in-human trial of the SurVaxM vaccine in combination with conventional treatment (radiotherapy, surgery, and temozolomide) was initiated under NCT02455557 (36). Six-month progression-free survival (PFS-6) and 12-month OS (OS-12) rates of 96.3 and 90.9% were observed, respectively (32, 37). Several survivin-specific vaccines, including EMD640744 (38), DPX-Survivac (39), and Sur1M2 (40), have been developed in addition to the SurVaxM vaccine (LMLGEFLKL). However, there are limited data on their safety and efficacy in GBM patients, and further studies are needed.

Isocitrate dehydrogenase1/2 (IDH1/2) are metabolic enzymes encoded by the IDH1 and IDH2 genes on chromosomes 2 and 15, respectively. IDH1/2 catalyzes the reversible conversion of isocitrate to alpha-ketoglutarate (α- KG; ref. 41). Mutations in the genes encoding these enzymes have been discovered in a number of human cancers, including gliomas (42). These mutations occur uniformly in the catalytically active regions of these enzymes and lead to oncometabolite 2-hydroxyglutarate (2HG) production, genome instability, and neoplastic transformation. The D enantiomer of 2HG competitively inhibits α-KG–dependent dioxygenases such as prolyl hydroxylase and histone lysine demethylase. Consequently, this can impair cellular metabolism and restrict cell differentiation through epigenetic mechanisms (43). In one form of this mutation (mismatch; heterozygous), arginine (R) replaces histidine (H) in the catalytic region of IDH1 at codon 132. IDH1-R132H is the most common type of mutation and is present in 70% of grade II–III and secondary gliomas (44). This mutation produces a common clonal neoepitope that is recognized by MHC class II and stimulates CD4⁺ T helper-1 (TH1) responses and the formation of specific antibodies that are naturally detectable in patients with IDH1(R132H)-mutated gliomas (45). Because IDH1-R132H is present only in tumor cells and not in normal cells, it is a true tumor-specific antigen suitable for mutation-specific vaccination and mutation-specific T-cell responses (45). Preclinical studies with the IDH1 vaccine in mice resulted in an efficient MHC class II–restricted mutation-specific antitumor immune response and reduced the development of gliomas expressing IDH1 (R132H; refs. 46, 47). Vaccinated mice had higher levels of peripheral IDH1(R132H)-specific T cells, IFNy, and IDH1(R132H)-specific antibodies, resulting in reduced growth and, in some cases, remission of intracranial glioma (46–48). Currently, three different clinical trials are evaluating the safety and efficacy of IDH1 peptide vaccination alone (NCT02454634 and NCT02771301; ref. 49) or in combination with temozolomide (NCT02193347). In NCT02454634, Platten and colleagues developed a vaccine directed against the IDH1 R132H neoepitope peptide and elicited an immunologic response in more than 90% of vaccinated patients. No serious adverse events were reported (50). In the phase I trial, NCT02193347, radiotherapy, surgery, and temozolomide were combined with PEPIDH1M (a peptide vaccine covering the altered IDH1R132H region). Despite a higher rate of serious adverse events with the combined therapy (16.67%), patients exhibited higher immunogenicity compared with the IDH1 vaccine alone, and vaccination increased the number of splenic white blood cells specific for the mutant peptide (44, 51). Interestingly, patients with low-grade and anaplastic gliomas with high penetrance of the IDH1 (R132H) mutation may benefit most from IDH1 vaccination (44). Research on IDH vaccines is in its nascent stages, and further studies are needed to evaluate their safety and immunogenicity.

The epidermal growth factor receptor III variant (EGFRvIII) is a constitutively active wild-type tyrosine kinase mutant uniquely found on a subset of tumor cells (approximately 33%). This degree of specificity makes it a promising target for the development of targeted immunotherapies (52). The EGFRvIII variant is caused by an in-frame deletion of exons 2–7 of the EGFR gene. The deleted region codes for 267 amino acids that form the extracellular domains I and II. This culminates in a restructured, aberrant form of the EGFR gene with significant pathologic consequences (53).

Mechanistically, EGFRvIII enhances glioma infiltration through both overexpression of the antiapoptotic protein Bcl-xL and STAT3-dependent activation of HIF1a (54). EGFRvIII is constitutively active. Despite the absence of a functional ectodomain that would otherwise confer ligand specificity, it displays continuous ligand-independent activity. Dysregulated EGFR signaling is associated not only with enhanced tumorigenesis but also with metastasis and resistance to radiotherapy as well as chemotherapy (55). Because this variant (EGFRvIII) supports proliferation, this oncogenic protein is a very attractive tumor-specific target for GBM vaccine development (56). The truncated extracellular domain generates a novel tumor neoantigen, resulting in a GBM cell-specific antigen with proven immunogenicity and efficacy in both mice and humans (57). This motivated the development of rindopepimut (CDX-110), a peptide-based vaccine. Rindopepimut is a peptide vaccine consisting of 14 amino acids of the EGFRvIII fusion site that leads to the eradication of EGFRvIII-expressing GBM cells (57). This vaccine is often combined with KLH because of its (i) high immunogenicity, (ii) low toxicity, and (iii) widespread availability. KLH activates T- and B-cell–dependent immunologic responses in vivo, promoting antigenic immune responses (58).

Longer-term follow-up of a phase II trial (ReACT trial; NCT01498328) of rindopepimut in patients with recurrent EGFRvIII-positive GBM showed that the addition of rindopepimut to standard therapy with bevacizumab (a VEGF inhibitor) improved 6-month PFS (28%) compared with the control group (16%; bevacizumab plus KLH adjuvant; ref. 59). Twenty percent of patients showed PFS at 24 months compared with bevacizumab alone (3%), and more than half of patients treated with rindopepimut had higher anti-EGFRvIII antibodies, which was associated with improvement in OS (59, 60). In addition, the combination group (rindopepimut plus bevacizumab) benefited from secondary outcome criteria such as OS, duration of response, and corticosteroid requirement. Another clinical trial (ACT-IV study; NCT01480479) evaluated rindopepimut in combination with radiation and routine adjuvant treatment (temozolomide [TMZ]) in patients with newly diagnosed GBM who expressed EGFRvIII and whose tumor was surgically removed (60). There was no significant difference in OS for patients with minimal residual disease at the nadir of the study: the median OS in the rindopepimut group (rindopepimut with TMZ) was 21 months versus 20 months in the control group (TMZ plus KLH). Furthermore, when comparing the rindopepimut group with the control group, the rindopepimut group had a higher grade of 3–4 adverse events in all treated participants (61). Of note, antibody levels were higher in the ReACT group than in the ACT-IV group, possibly due to TMZ in ACT-IV or BEV increasing T-cell priming. BEV has the potential to shift the local milieu to a more immunologically favorable or even stimulating environment (62).

Overall, these results support the principle that EGFRvIII can be a potential therapeutic target for vaccination (56). Zebertavage and colleagues recently demonstrated that a Listeria monocytogenes–based vaccine vector induced superior EGFRvIII-specific CD8⁺ T cells and antitumor immunity (56). To increase the number of potential MHC class I binding epitopes, a larger 21 amino acid peptide (compared with the existing 14 amino acid vaccine) was selected and inserted into the genome of a live-attenuated vaccine vector (63). The inflammation induced by microbe-based cancer vaccines promotes APC maturation, antigen processing, and T-cell proliferation, resulting in strong antigen-specific CD8⁺ T-cell responses (64). The microbe-based vaccine produced 3–5 times more EGFRvIII-specific CD8⁺ T cells in mice than the previously described rindopepimut-KLH vaccination and effectively reduced the growth of EGFRvIII-expressing squamous cell carcinomas.

Genetic vaccines (also gene-based vaccines) are the fastest growing area in vaccine technology, in which cells take up nucleic acids such as DNA (as plasmids) or RNA (as mRNA) and translate them into proteins in accordance with the nucleic acid template. Adhering to this blueprint can elicit a tumor-specific immune response. Interestingly, genetic vaccines, such as live viruses or attenuated viruses, activate MHC class I and II pathways so that both CD8⁺ and CD4⁺ T cells can be activated without the inherent risk associated with live vaccines (65, 66). In addition, genetic vaccines can circumvent many of the issues associated with recombinant protein-based vaccines, such as high manufacturing costs, purification problems, improper antigen folding, and insufficient activation of CD8⁺ T cells (66).

Several groups actively develop personalized cancer treatment approaches using mRNA-based vaccines. The effectiveness of mRNA vaccines against gliomas remains unknown due to tumor heterogeneity and the immunosuppressive milieu (67). Unlike live-attenuated or inactivated viral vaccines, there is minimal risk of infection or insertional mutation with mRNA vaccines because they cannot integrate into chromosomal DNA (68). mRNA vaccines are rapidly degraded in the body, reducing their toxicity risk. However, their vivo half-life can sometimes be prolonged by various modifications and delivery methods, and this is an active area of investigation (68, 69). Due to the high-efficiency in vitro transcription, mRNA vaccines for any pathologic antigen can be developed and scaled up using standardized techniques (69). This approach assumes that a genetically modified mRNA can be taken up by patient cells, considered genetic material, and translated into the appropriate antigen protein by the patient cell machinery (69, 70).

Zhong and colleagues recently identified four glioma antigens, including ANXA5, FKBP10, MSN, and PYGL, as promising targets for mRNA vaccine development. Increased expression of these genes has been associated with glioma development and progression, as well as a decrease in OS and disease-free survival (DFS; ref. 71). Importantly, their abundance correlates with increased recruitment of B cells, macrophages, and DCs, implying that APCs can process these antigens and present them on the cell surface in association with II MHC class molecules, where T-cell receptors and B cells specifically recognize them to activate a tumor response (71, 72). Other studies have identified FCGBP, FLNC, ARPC1B, TLR7, CSF2RA, and HK3 as potential TAAs for developing mRNA vaccines against GBM (73). Most of these overexpressed proteins are involved in cell adhesion pathways and are essential for tumor vascularization and proliferation, and also inhibit T-cell responses (74). NCT04573140, NCT02709616, NCT00846456, NCT00890032, NCT00626483, NCT00639639, NCT03548571, NCT02366728, NCT02649582, and NCT02465268 are all clinical trials using mRNA as a potential vaccine for the treatment of glioblastoma. In Table 2, we have summarized these clinical trials and the interventions used to improve antitumor efficacy.

NCT02465268 (ATTAC II) is a phase II clinical trial currently under way to evaluate the efficacy of this vaccine in patients with newly diagnosed GBM (75). This clinical trial showed that vaccination of newly diagnosed GBM patients with CMV pp65 RNA-loaded DCs in combination with GM-CSF as an adjuvant, followed by higher-dose routine chemotherapy (TMZ), resulted in improved PFS (31.4 months) and OS (34 months) compared with a control group that received a high dose of TMZ cycles alone after 6 weeks of conventional chemoradiation (76). Patients in NCT00846456 received DC-based vaccination targeting cancer stem cells (CSC; ref. 77). This clinical trial extracted mRNA from autologous glioblastoma stem cells and transfected them into autologous immature DCs to elicit an immunologic response against patient antigens (77). The modified mature DCs were administered subcutaneously at specific intervals after the completion of 6 weeks of postoperative radiochemotherapy (77, 78). They detected a GSC-specific immune response in all vaccinated patients, and PFS was significantly longer than in the control group (median 694 days vs. 236 days). No serious adverse events occurred (similar to standard therapy), and vaccination was well tolerated and safe.

DNA vaccine development is a new method being investigated in GBM patients (Table 3). Previous studies emphasized viral sources as a biological vector to deliver predefined antigens into host cells. More recently, the focus has shifted to synthetic DNA plasmids (4). DNA vaccines have a number of properties that make them ideal for cancer vaccine development. Concerns of replication competency and immunogenicity are less in plasmid vaccines, making them safer than other platforms (recombinant proteins and viral vectors). Other advantages include stability, ease of large-scale production, absence of infectious agents, and superior compatibility with humans (79). Plasmid-based DNA vaccines encoding TAAs and immunostimulatory molecules (IL12, TNF, NO, IL2, GM-CSF, and others) can induce robust CD4 and CD8 T-cell responses by presenting antigens on both MHC class I and II, as well as a humoral response, both of which correlate with protection against the specific cancer antigen in host cells. DNA vaccines expressing transgenic antigens can be produced rapidly and inexpensively. Moreover, immune cells do not initiate an immune response against the DNA backbone, allowing for highly targeted host immunity directed against the transgene rather than the DNA structure itself (80).

Recently, Lopes and colleagues developed pTOP (plasmid to deliver T-cell epitopes), a novel DNA vaccine formulation that delivers the coding sequence of the vesicular stomatitis virus glycoprotein (VSV-G), where antigen epitopes can be embedded in permissive sites without affecting protein functions. VSV-G acts as an adjuvant to DNA vaccine and has been shown to induce both cell-mediated and humoral immunity (81). Due to the presence of VSV-G, this modified DNA vaccine can elicit antiepitope T-cell responses while retaining some innate viral immunogenicity (82). This strategy can potentially improve outcomes in many tumor models, especially when combined with a sensible treatment plan. Lopes and colleagues demonstrated that pTOP vaccination resulted in an acceptable cure rate (78%) and increased life expectancy in melanoma mice models (30 days vs. 19 days in untreated mice).

Moreover, this strategy led to a decrease in immunosuppressive cells and an increase in immunologically active CTLs in the brain (81). It has been proposed that tumor removal facilitates vaccine-triggered immune response (in the brain) by creating a local inflammatory milieu and further disrupting the blood–brain barrier (83, 84). When tumor resection and pTOP vaccination were combined, 78% of mice with melanoma survived at least 250 days, compared with 40 days in control mice (resection or pTOP7), suggesting that tumor resection may increase the effectiveness of pTOP vaccines.

A phase I clinical trial (NCT02718443) evaluated the efficacy of the DNA vaccine VXM01 in patients with recurrent glioblastoma who had failed to respond to standard therapy (radiochemotherapy with TMZ; ref. 85). VXM01 is a plasmid vaccine derived from an attenuated strain of the gram-negative bacteria Salmonella typhi that encodes VEGFR-2 (vascular endothelial growth factor receptor 2; ref. 86). This study aimed to evaluate the safety and tolerability of VXM01 at escalating doses in GBM patients, as well as the immunomodulatory and antiangiogenic effects of VXM01 (87). The results showed that VXM01 was well tolerated at all doses, and most of the vaccinated patients exhibited a VEGFR-2–specific T-cell immune response and higher rates of tumor-infiltrating T cells in tumor tissue after vaccination (86, 87).

A similar phase I trial of a DNA vaccine (NCT04015700) in patients with newly diagnosed IDH1 or IDH2 mutants in glioblastoma is under way. This trial uses INO-9012 (synthetic DNA plasmid encoding the proinflammatory cytokine interleukin-12) and a DNA plasmid encoding the tumor-specific antigen GNOS-PV01, amenable for treating patients with this mutation. IL12 acts as a molecular adjuvant to activate the immune system by promoting the generation of specialized T cells against patient-specific antigens (88).

Dendritic cells (DC) are a specialized form of APCs that control both immune tolerance and immunity. They exist in most tissues as immature (resting) cells, positioning them as a promising target to generate immune responses against cancer. After antigen capture and processing, they present mature peptides on their HLA class I and II receptors to generate MHC–peptide complexes. Activated DCs (mature) relocate from peripheral tissues to lymph nodes and secondary lymphoid organs to physically interact with and stimulate T-cell responses (89).

Dendritic cell vaccination (DCV) as a therapeutic adjuvant in GBM is intensely debated. Numerous clinical trials have used DCV as active immunotherapy and vaccinated hundreds of GBM patients to elicit an antitumor immune response. Overall, the efficacy of DCV in GBM is variable, ranging from no clinical response to significant responses. DCVs have reduced tumor progression, prolonged life expectancy, stimulated tumor-specific IFNγ, and activated cytotoxic T lymphocyte (CTL). Vaccination appeared safe and well tolerated with no serious adverse effects (≥grade 3; refs. 90, 91). Inogés and colleagues conducted a phase II clinical trial in which they administered autologous DCs vaccination to patients with newly diagnosed GBM following resection and combined it with adjuvant radiochemotherapy to improve patients’ survival (92). Autologous DCs were produced from blood monocytes and treated with autologous whole tumor lysate. They found an increase in the proliferation whole tumor cell vaccines and number of IFN-producing cells after antigen stimulation of monocytes, before and after vaccination (92). In phase I clinical trial in patients with newly diagnosed and recurrent GBM, Hu and colleagues used an autologous DC vaccine pulsed with a lysate derived from a GBM stem-like cell line (93). Patients who had just been diagnosed with GBM also received conventional radiation and TMZ. The primary goal of this clinical research was to evaluate the autologous DC vaccine's safety and tolerability. This study found the vaccine safe and well tolerated, with a median PFS of 8.75 months for patients with newly diagnosed GBM and 3.23 months for patients with recurrent disease. Patients with newly diagnosed GBM had a median OS time of 20.36 months, whereas those with recurrent disease had a median OS of 11.97 months. A fraction of patients showed a cytotoxic T-cell response, indicating that vaccination might help with GBM therapy (93). Li and colleagues conducted a phase I study in patients with surgically accessible recurrent GBM. They evaluated the efficacy and safety of pembrolizumab/ MK-3475 (as PD-1 antibody) in combination with autologous tumor lysate-pulsed DC vaccination (ATL-DC). Administration of pembrolizumab and the ATL-DC vaccine may be more effective than ATL-DC alone in patients with recurrent GBM (94). Table 4 summarizes clinical trials using DC vaccination in GBM patients.

As described above, the immune system recognizes tumors using two types of antigens: TSAs and TAAs. TAAs are classified into five types: (i) mutant antigens, which are expressed exclusively by tumors; (ii) highly expressed antigens, which are normal proteins whose expression is increased in tumors; (iii) oncofetal antigens, which are typically present only during fetal development; (iv) differentiation antigens, which usually are found only at certain stages of differentiation of a cell type; and (v) cancer-testis antigens, which are typically expressed in human testes and placenta (95). Characterizing the immunogenicity of TAAs in humans, identifying the most immunogenic epitopes, and evaluating their role as tumor defense antigens that can trigger tumor regression is a highly active area of research. A promising alternative to single TAAs is developing and implementing vaccines from whole tumor cells without the need to define antigens. Tumor cells possess a variety of TAAs that contain epitopes recognized by both CD8⁺ CTLs and CD4⁺ T helper cells, activating both the innate and adaptive immune systems (96). One potential benefit of using tumor cell vaccines that simultaneously target multiple tumor antigens and a strong polyclonal T-cell response is the induction of a robust antitumor response and long-term memory.

In contrast to defined tumor-derived peptides and proteins, whole tumor vaccine therapy applies to all patients regardless of HLA type, facilitates an effective targeting of the majority of tumor cells, and eliminates the need for definition, testing, and selection of immunodominant epitopes (97). Simultaneously targeting multiple tumor antigens (compared with single-epitope vaccines) can reduce the likelihood of tumor escape. Patients can be vaccinated with either autologous or allogeneic tumor cells. Autologous whole-cell cancer vaccines use the patient's tumor cells, which contain unique mutations that encode neotumor antigens from the patient's tumor, making them attractive for developing personalized cancer vaccines. Personalized vaccination without HLA restriction is possible with autologous whole-cell cancer vaccines (97, 98). The main disadvantage of these vaccines is that they are suitable only for personalized antitumor immunotherapies in a single patient and may cause problems in collection, processing, reproducibility, and interpatient variability (97). In contrast, using allogeneic tumor cell lines from different tumors in the vaccine that share one or more of the TAAs produced by the patient's tumor provides an easier way to introduce antigens in the vaccination process. Cell factories can produce allogeneic cell lines in enormous quantities, and their quality can be quickly evaluated and controlled in good manufacturing practice facilities (99). However, there is limited evidence in animal studies that allogeneic vaccines cause tumor rejection. In addition, T lymphocytes require antigen presentation by cells with a matching HLA profile. Allogeneic tumor cells, by definition, have an HLA profile that is foreign to T lymphocytes (100).

Living tumor cells are poorly immunogenic as they release soluble factors that can suppress DC development and function and trigger lymphocyte apoptosis (101, 102). The immunosuppressive factors released in the local tumor microenvironment include prostaglandin E2 (PGE2), transforming growth factor beta (TGF-b), IL10, IL27, NO (nitric oxide), and IDO (indoleamine 2,3-dioxygenase; refs. 103, 104). Combining whole-cell tumor vaccines with other immunotherapeutic techniques is a potential method to improve their immunogenicity. Integrating tumor cells with potent adjuvant haptens is an efficient strategy to elicit a robust inflammatory response. Berd and colleagues have developed an innovative active immunotherapy technique based on modifying autologous cancer cells with hapten dinitrophenyl (DNP; ref. 100). They have recently obtained encouraging results in ovarian cancer and other cancers, including melanoma. Administration of modified autologous cancer cells with the DNP to patients with metastatic melanoma promoted inflammation in the metastatic masses. There was an infiltration of T lymphocytes, most of which were CD8⁺. In most cases, these T cells produced interferon-γ in situ. Although no signs of acute toxicity were observed, no clinically significant reactions were seen in most patients (100).

As monotherapy, vaccination against tumors has not shown a survival benefit in multiple cancers, including GBM. As a monotherapy, the effectiveness of immune-checkpoint blockade is limited as GBM has an immunosuppressive microenvironment, vast inter- and intratumoral heterogeneity and a low mutational burden at diagnosis compared with other immunogenic tumors (103, 105). Combination immunotherapy has emerged as an alternative technique to enhance the antitumor immune response in GBM and circumvent the complicated immunosuppression in the tumor microenvironment. The combination of cancer vaccines, which activate a specific immunologic response, and immune-checkpoint inhibitors, which recover T-cell function in hostile tumor environments, has attracted interest. GBM is characterized by inflammation in and around the tumor that enhances the expression of immune-checkpoint ligands. Thus, the combination of vaccines and immune-checkpoint ligand/receptor-targeted therapy may synergize in reversing GBM immunosuppression (106). An ongoing phase I clinical trial (NCT02287428) combines a personalized neoantigen vaccine/NeoVax with pembrolizumab/MK-3475 (anti–PD-1 antibody) and radiotherapy in newly diagnosed patients with MGMT-unmethylated GBM. The objective of the clinical trial is to determine whether it is possible to improve the efficacy of the vaccine by combining it with pembrolizumab (107).

There is growing evidence that combining cancer vaccines with standard treatments that regulate the immune response may provide the greatest clinical benefit. It is possible to augment the antitumor activity elicited by cancer vaccines through various immunomodulatory effects of conventional therapies (108). Radiotherapy increases tumor antigens, costimulatory molecules, cytokines, and chemokines and downregulates regulatory T cells (Treg). Chemotherapy induces immunogenic death of tumor cells and leukopenia, favors tumor-specific T cells, and upregulates tumor antigens. Interest in combining radiotherapy and chemotherapy with GBM vaccines is growing in proportion to our understanding of the immunomodulatory properties of radiotherapy and chemotherapy, which can be used to enhance vaccine-mediated antitumor effects (108). However, a recent comparison between chemoradiotherapy in combination with the GBM vaccine and chemoradiotherapy alone for the treatment of newly diagnosed GBM showed that this combination did not significantly improve patient survival.

Furthermore, despite the apparent trend, the combination was not significantly associated with an increased incidence of grade 3 or higher serious adverse events. These results suggest that the inclusion of the GBM vaccine in standard therapy for GBM is relatively safe. The investigators believe that standardization of clinical trials of the GBM vaccine with chemotherapy and radiotherapy for GBM treatments is needed to allow more accurate comparison and analysis of these combination treatments.

GBM is classified into two different immune subtypes based on the characteristics of the tumor microenvironment, immunologic subtype 1 (IS1) and immunologic subtype 2 (IS2; ref. 109). There is a significant difference between the subtypes: (i) patients with GBM of IS1 have lower OS than those of IS2, (ii) IS2 has higher immune infiltration, and better outcomes than IS1, and (iii) IS1 has an immunosuppressive feature, whereas IS2 has a proinflammatory type (109, 110). IS2 has a significantly increased abundance of monocytes and macrophages than IS1, as well as CD8⁺ T-cell anticancer immune responses (111). These findings present a new challenge in defining the immune subtypes of GBM patients and identifying potential tumor antigens that may be favorable for the immunotherapy of each immune subtype (112).

Numerous clinical trials for GBM have been conducted to investigate the efficacy of GBM vaccination. Although tumor vaccines have resulted in significant tumor rejection and long-term tumor immunity in animals, antitumor efficacy in human studies has been disappointing for all cancers, including GBM. Several factors may limit the efficacy of current GBM vaccines. These include the continued tumor dedifferentiation by glioblastoma stem cells (GSC), limited access to the central nervous system and tumor due to limitations in drug choice and route of vaccination, profound tumor-associated mechanisms of immunosuppression and -evasion, tumor heterogeneity, and low mutational burden (104). Immune evasion originates from both tumor and tumor microenvironmental modifications. Tumor evasion is believed to be caused by tumor cells' low production of immunogenic target antigens, making them "poorly accessible" to T cells. In contrast, immune suppression is defined by multiple mechanisms that limit the adaptive antitumor immune response (92). Another factor contributing to vaccine resistance in GBM is the high level of immunosuppressive factors released in the local tumor microenvironment by tumor cells and tumor-associated cells.

In addition to the factors mentioned above, GBM has evolved to resist and escape CTL-mediated lysis. CTL can only execute their tumor lysing effector functions by recognizing the HLA class I attached to tumor-specific antigenic peptides. Mutations in the peptide processing machinery, complete loss of genomic DNA encoding specific HLA alleles, and alterations in the chromatin structure of HLA gene promoters are common in GBM. Taken together, these mechanisms impede T-cell activation and reduce the efficiency of vaccine-induced CTL-mediated lysis (4, 92, 104).

Another limitation in vaccination therapy against cancer, particularly GBM, is vaccine-associated toxicity (113). In general, cancer vaccines are often well tolerated and associated with minimal toxicity. The spectrum of antigens, the variety of preparations, the adjuvants used, and their combination with immunomodulators present a significant challenge in evaluating the toxicity of cancer vaccines. Common toxicity induced by cancer vaccines is a mild autoimmune reaction to normal tissues (113, 114). The autoimmune response to specific self-antigens may be tissue-specific or systemic if a variety of expressed autoantigens are targeted (115).

GBM is an aggressive form of brain tumor with few therapeutic options. Despite technological advances in standard treatment, the average patient survival is 15 months. The most commonly used vaccination strategies for treating GBM are peptide vaccines, genetic vaccines (RNA and DNA), and cell-based vaccines (DC and whole tumor cell vaccines). Dendritic and whole tumor cell vaccines are the primary cell-based vaccines whose efficacy in GBM has been studied. In general, cell-based vaccines are safe, but objective clinical responses are limited, and in most patients, prolonged survival has not been routinely demonstrated. To date, developing vaccines for GBM has been confined to a limited set of patients and early-phase clinical trials. Given the innovative advances in immunotherapy, including the use of immune-checkpoint blockade, the benefit of vaccines against cancer will likely be combinatorial approaches.

R.S. O'Connor reports grants from the NIH during the conduct of the study. No disclosures were reported by the other authors.

The authors declare that no funds, grants, or other supports were received during the preparation of this manuscript.