Clinical implication of cellular vaccine in glioma: current advances and future prospects

High-grade glioma, especially recurrent GBM, is almost inevitable, and the prognosis remains poor, with a median survival of 12–15 months [22]. Immunotherapies using autologous vaccines are now considered promising approaches for treating patients with malignant gliomas. The microenvironment of brain tumors contains unique tissue-resident cell types, in addition to cancer cells, including microglia, astrocytes, neurons and immune cells. The normal brain is considered an “immune privileged” organ that is sheltered from immune cell entry through the integrity of the BBB, which physically protects the brain from inflammatory factors that can be cytotoxic and cause neurodegeneration [23, 24]. On the other hand, peripheral immune cells infiltrating from the circulation can cross the BBB and enter the CNS after injury and, in disease, then join resident immune cells to modulate neuroprotective lymphocyte responses and brain function via their interactions with glia [25]. Indeed, current data indicate that therapeutic vaccines based on T cells, DC cells, tumor cells and NK cells are feasible and generally well tolerated. Furthermore, their clinical efficacy has been demonstrated in several randomized clinical trials.

Adoptive T cell transfer can overcome the in vivo progression of gliomas by using autologous cells with engineered receptor specificities for TAAs [26]. Antigen-specific T cells can be expanded ex vivo for subsequent administration, which produces the cytokines that are essential for T cell expansion and sustained anti-tumor activity [27]. CAR-engineered T cell (CAR T cell) therapy is a promising therapeutic approach genetically generated with modified T cells to express recombinant protein CARs that may be effectively and safely applied to GBMs to reduce recurrence rates [28, 29]. Several cell surface proteins, such as interleukin 13 receptor α2 (IL13Rα2), epidermal growth factor receptor variant III (EGFRvIII), ephrin type-A receptor 2 (EphA2), and human epidermal growth factor receptor 2 (HER2), have been found to actively target CAR T cell therapy in preclinical models [30‐33], but only a few of these cell-surface receptors have been validated in clinical trials. Accordingly, a phase I/II clinical study of adoptive immunotherapy suggests that anti-EGFRvIII CAR-engineered T cells effectively produced the effector cytokines and interferon-γ, contributing to lyse the antigen-expressing glioma cells [34]. Meanwhile, another completed phase I clinical trial program (NCT01109095) reveals that anti-HER2 CAR CMV-specifc T cells seem to be able to inhibit HER2 + glioma growth [35]. Here, to improve anti-glioma responses, we discuss the use of TAA-engineered T cells through their clinical strategies and outcomes under investigation.

IL13Rα2, a cell-surface receptor positively expressed in 82% of GBM samples and > 70% of glioma stem-like cancer initiating cells [36, 37], was previously thought to be directly associated with increased mesenchymal signature gene expression and poor patient survival [38]. For the treatment of recurrent GBM, Christine et al. showed the first-in-human clinical experience for CAR-engineered IL13Rα2-specific CD8⁺ CTL and observed significant tumor regression. Briefly, for autologous IL13-zetakine⁺ CD8⁺ CTL manufacturing, the peripheral blood mononuclear cells (PBMCs) were stimulated with anti-CD3 antibody, followed by DNA electroporation, drug selection and ex vivo expansion using OKT3 and irradiated feeders. In three patients with recurrent GBM, the feasibility of repetitive intracranial administration of first-generation IL13Rα2-specific CD8⁺ CAR T cells was demonstrated and transient anti-tumor activity for some patients was reported in the absence of serious adverse events, such as occlusion, malfunction, or infection [30]. Building on these results, the modified IL13Rα2-targeted CAR T cells were further reported to improve anti-tumor potency and T cell persistence by 4-1BB co-stimulation and IgG4-Fc linker mutation [39]. A patient with recurrent multifocal GBM who received treatment with modified IL13Rα2-targeted CAR T cells had regression of all intracranial and spinal tumors, along with significant increases in the levels of cytokines C-X-C motif chemokine ligand 9 (CXCL9) and CXCL10, as well as immune cells in the cerebrospinal fluid [28]. Comparing the ability to abrogate tumor growth at local and distant sites, Christine et al. suggested intraventricular administration of CAR T cells is better than intracavitary therapy for the treatment of malignant brain tumors. Nevertheless, the above evidence of the safety and anti-tumor activity of IL13Rα2-targeted CAR T cell immunotherapy still needs to be evaluated in a larger cohort of patients.

Negative prognostic indicator EGFRvIII is expressed in about 25–33% of all patients with GBMs [40] and is the most commonly mutated gene among the EGFR family in glioma [41]. In EGFRvIII-expressing newly diagnosed GBM, a peptide vaccine targeting EGFRvIII (rindopepimut) was previously evaluated and found to be well tolerated, providing immune responses with prolonged progression-free survival [42, 43]. More recently, O’Rourke et al. conducted a phase I safety study of autologous CAR T cells targeted to EGFRvIII (CART-EGFRvIII) in 10 patients with recurrent GBMs. Intravenous infusion of a single dose of CART-EGFRvIII cells was found to be feasible and safe, without off-tumor toxicity or cytokine release syndrome [44]. For vaccine delivery, CART-EGFRvIII cells were detected transient expansion in peripheral blood. Trafficking of CART-EGFRvIII cells were also found in regions of active GBM in 7 patients with surgical intervention. Compared to pre-CART-infusion, tumors had markedly induced expression of immunosuppressive molecules (IDO1 and FoxP3) post-infusion. However, marked tumor regression was not observed by MRI over 18 months of follow-up after CART infusion. It is possible that this invalid clinical benefit of CART-EGFRvIII, which can be definitively determined from a large sample size, may improve the responsiveness of tumors through intraventricular therapy.

Elevated HER2 expression has been reported in 41% of primary GBM samples and in 81.4% of GBM primary cell lines and were correlated with impaired survival [45, 46]. In preclinical models of GBM, bispecific CAR molecules that incorporated 2 antigen recognition domains for HER2 and IL13Rα2 showed the functional superiority of T cell expressing antigens ex vivo and in an orthotopic GBM xenograft model [47]. However, the safety concerns of autologous CART-HER2 cells were raised by the report of a serious adverse event following the administration 1 × 10¹⁰ T cells of vaccine based on trastuzumab [48]. While administration of up to 1 × 10⁶/m² CART-HER2 cells showed no evident toxicities, unfortunately, the expansion and persistence of CART-HER2 cells was limited [49]. To treat HER2-positive GBM, Nabil Ahmed et al. developed HER2-specific CAR-modified virus-specific T cells (CAR VS Ts-HER2) with an FRP5-based exodomain and a CD28.ζ endodomain [50]. Up to 1 × 10⁸/m² CAR VS Ts-HER2 were infused intravenously without dose-limiting toxic effects in 17 patients with progressive GBM. After the infusion, CAR VS Ts-HER2 was detectable in the peripheral blood for up to 12 months, with no observed expansion in peripheral blood (but expansion at GBM sites). Of 16 evaluable patients, 50% of patients had clinical benefit, as defined by a partial response (N = 1, over 9 months) and stable disease (N = 7, 8 weeks–29 months). Despite the feasibility and safety of CAR VS Ts-HER2, a clinical strategy using it alone or in combination with other immunomodulatory approaches is warranted to improve the anti-GBM activity of CAR VSTs-HER2 by augmenting their expansion and persistence.

In cancer cells, epigenetic regulation by DNA methylation facilitates cancer/testis (CT) antigen expression and constitutes an attractive immunological target for cancer immunotherapy [51‐53]. Recently, a novel approach was reported based on the adoptive transfer of cytotoxic lymphocytes with specificity for antigens induced by a DNA demethylating agent [54]. In this report, Kirkin et al.. found that after treatment with 5-aza-2’-deoxycytidine, a DNA-demethylating agent, activated CD4⁺ T helper cells from the patients with recurrent GBM could express a broad repertoire of endogenous CT antigens and serve as antigen-presenting cells to generate autologous CTL and NK cell responses. This group also revealed that injections of tumor-reactive lymphocytes generated by this DNA demethylation-mediated procedure into peripheral vein of GBM patients generated persistent anti-tumor immune response without treatment-related adverse effects. Treatment of cancer cells with 5-aza-2’-deoxycytidine also leads to the increased levels of MHC and costimulatory molecules (CD80, CD86, and CD40) necessary for the antigen-presenting function of DC cells [55, 56]. In addition, as an important determinant of anti-tumor immune response, DNA methylation aberration implicates epigenetic modulation as a combination regimen for potential precision immunotherapy with checkpoint blockade [57].

Overall, TAA-engineered T cells, along with DNA-demethylated T helper cells against recurrent GBMs, initially confirm the safety and anti-tumor activity of autologous T cell immunotherapy, providing a minimally invasive strategy for glioma treatment and prevention.

DC therapy is a safe and well tolerated immunotherapeutic method, and its clinical effectiveness has been validated in melanoma, prostate cancer, malignant glioma, and renal cell carcinoma [58]. Previous studies have demonstrated that microglia are the predominant antigen-presenting cells (APCs) in the brain [59], while DC vaccines are likewise gaining significant clinical attention as a complementary strategy to stimulate T cell responses [60]. For instance, DC derived from blood monocytes could further enhance tumor-specific CD8⁺ T cell polyfunctionality in vivo when administered as a vaccine [61]. Recently, the presence of human cytomegalovirus (HCMV) antigens has been specifically found in gliomas, but not surrounding normal brain tissue. These unique and immunogenic HCMV antigens provide an attractive opportunity to leverage HCMV antigens as tumor-specific immunotherapy targets whilst minimizing toxicity [62]. To prove this conjecture, a randomized pilot trial in patients with newly diagnosed GBM was conducted to demonstrate that HCMV pp65 RNA-loaded DC vaccination experienced enhanced tumor-specific CD8⁺ T cell polyfunctionality and significantly increased overall survival [63]. The initial results obtained from clinical trials of autologous antigen or peptide-pulsed DC appear to be encouraging for a variety of tumors [64, 65]. The clinical trials testing DC vaccines after modified or adjuvant treatment with novel chemotherapy for gliomas defined as malignant or special subtype are underway at the NIH Clinical Center (Table 1). Here, we mainly review the feasibility and safety of autologous DC vaccines in phase I-III clinical trials and discuss recent progress in immunotherapy using autologous DC-based vaccination (Fig. 2).

DC can sample tumor antigens through capturing the ‘eat me’ or ‘do not eat me’ signal of tumor cells [66]. Molecules such as milk fat globule-EGF factor 8 (MFG-E8) bridge the phosphatidylserine of dying cells with integrin αvβ3 of DC, then link to downstream signals through integrin receptors on the phagocyte [67]. Tumor lysates include poorly identified high-grade glioma-specific tumor antigens, which have practicality in terms of personalized medicine [68]. Previous studies have indicated that vaccination with glioma cell lysate-pulsed DC elicited a stronger specific CTL response, thereby preventing glioma formation in C57BL/6 mice model [69]. More importantly, combinatorial treatment of tumor lysate-pulsed DC vaccines and other therapeutic strategies, such as checkpoint inhibitors [70], angiogenesis inhibitors [71] and cytokine gene therapy [72], conferred a greater survival advantage and significantly increased the therapeutic anti-glioma efficacy. In addition, counteracting the immunosuppressive environment before vaccination is requisite to facilitate the long-term anti-glioma immune responses [9]. Nowadays, using DC from the peripheral blood of patients pulsed with an autologous tumor lysate is currently being evaluated in phase I/II clinical trials for gliomas [73], but the objective response rate still has been relatively low in some research.

Vaccination with tumor lysate-pulsed DC elicits antigen-specific and CTL responses in patients with malignant glioma. The percentage of NK cells, such as CD16⁺ and CD56⁺ cells, in peripheral blood lymphocytes were safely increased after immunization [73, 74]. Accordingly, a significant positive correlation was observed between activated NK cell populations and overall survival in patients administrated with autologous tumor lysate-pulsed DC vaccination [75]. For newly diagnosed GBM, autologous vaccination with tumor lysate-pulsed DC enhanced the precursor frequency of CD4⁺ T and CD4⁺ interferon (IFN) γ-producing cells, suggesting an induced tumor-specific CTL response [76]. A recent phase I study was conducted to evaluate the safety and bioactivity of vaccination with tumor lysate-pulsed DCs in patients with GBM and anaplastic astrocytoma. In response to tumor lysate after vaccination, most of patients displayed robust systemic cytotoxicity as indicated by peripheral IFN γ accumulation and intratumoral CD8⁺ T cell infiltrate. Furthermore, no clinical evidence of autoimmune diseases was detected, suggesting that tumor lysate-pulsed DC vaccination was safe [77, 78]. By comparing specific immune responses before or after vaccination, up-regulation and/or cytoplasmic accumulation of chemoresistance-associated peptides, including Wilms tumor protein (WT1), glycoprotein 100 (gp100), and MAGE family member A3 (MAGEA3), were noted in tumors being treated with immunotherapy [79]. Yu et al. noted significant expansion in CD8⁺ antigen-specific T cell clones against TAAs, including melanoma-associated antigen (MAGE) 1, gp100, and HER2, and the CD8⁺ T cell intratumoral infiltrate was increased in 50% of patients [77]. Moreover, tumor lysate and IL-18 loaded DC vaccines can elicit a specific CD8⁺ cytotoxic T lymphocyte response in GBM patients. The cytotoxic responses were augmented by transfecting DC with the gene for IL-18, but significantly inhibited by anti-human leukocyte antigen (HLA) class I antibody [80].

In addition, several small size studies showed biased results with or without a pro-survival effect from tumor lysate-pulsed DC vaccination in glioma patients. Ryuya et al. showed the maturation of DC with OK-432, granulocyte-macrophage colony-stimulating factor (GM-CSF) and IL-4 that were pulsed with autologous tumor lysate showed better survival in 24 patients with recurrent GBMs. Also, patients injected with both intratumoral and intradermal administrations had longer survival times as compared with intradermal administration only [81]. A phase II trial demonstrated 53% of GBM patients (N = 32) exhibited over a 1.5-fold increase in vaccine-enhanced cytokine (IFN γ) and had significantly longer times to tumor progression and survival [82]. Furthermore, intradermal administration of fusion cells and subcutaneous injection of recombinant human interleukin 12 (rhIL-12) at the same site showed a greater than 50% reduction in tumor size in some patients without adverse effects and safely induced clinical anti-tumor effects with malignant gliomas [83]. Accounting for the narrow therapeutic index of rhIL-12 [84], the alteration of chemoresistance-associated peptides may achieve susceptibility of TAA-expressing glioma cells to the specific immune response in DC-based immunotherapy. However, vaccination with Audencel, a tumor lysate-charged autologous DC vaccine, failed to improve progression-free survival and median overall survival, although no severe toxicity was observed in those newly diagnosed with GBM (N = 76) [85]. The factors dictating the efficacy of DC vaccines may represent a viable strategy to improve anti-tumor immunotherapy. The enzyme-linked immune absorbent spot (ELISPOT) is one of the most commonly used to understand the frequency of cytokine-secreting cells at the single-cell level [86]. Increased ELISPOT and delayed-type hypersensitivity responses after vaccination could provide good laboratory markers to predict the clinical outcome of patients receiving DC vaccination [73, 81]. The content of tumor-infiltrating lymphocytes (TILs) and T cell receptor (TCR) repertoire in brain tumors and peripheral blood have been proved to be correlated with improved clinical outcome in GBM patients [87]. TCR repertoires are widely shared sequences that have been suggested to be over-represented due to their potential immune functionality or their ease of generation by V(D)J recombination [88]. After treatment, tracking TCR repertoire shifts in tumors and peripheral blood can be used to monitor the treatment-associated immune response without the need to know the specificity of receptors [89]. Hsu and colleague revealed a statistically significant correlation between higher degrees of TCR repertoires and prolonged overall survival in autologous tumor lysate-pulsed DC vaccines-treated GBM patients [87]. In a prospective case control study that enrolled 47 GBM patients with DC vaccine adjuvant therapy, better outcomes were predicted with younger age and a lower programmed cell death protein 1 (PD-1)⁺/CD8⁺ ratio in TILs and PBMCs [90]. Further analysis of the immune system factors demonstrated that the patients with an immune system equipped with favorable pre-existing or post-vaccination anti-tumor capabilities, such as IFN γ secretion and CD8⁺ cells, are more likely to live longer [91]. Also, another clinical trial protocol of DC vaccination, including inclusion/exclusion criteria, was proposed and may increase the immune response and safety in pediatric and adult subjects [92, 93]. Thus, even DC immunotherapy against GBM has some exciting outcomes, and the bias caused by sample size, mode of administration and non-specificity of vaccine-target interactions indicate that investigating combination therapies or developing meaningful biomarkers should be studied in further phase II/III clinical trials.

Standard therapy for GBMs post-surgery includes radiotherapy and chemotherapy with temozolomide. For DC vaccination as adjuvant therapy, up to now, most of the clinical trial showed no serious vaccine-related adverse events, and it may extend survival [76]. Early insight into tumor lysate-pulsed DC vaccines was provided in a phase III randomized, double-blinded, placebo-controlled clinical trial (N = 331) and suggested the addition of the vaccination to standard therapy is feasible and safe [94]. Only 2.1% of patients experienced any grade 3 or 4 adverse events that were at least possibly related to treatment with tumor lysate-pulsed DC vaccines. Autologous DC vaccines benefit patients with malignant glioma but may cause transient and reversible elevation of serum AST/ALT levels [95].

DC are the most potent professional APCs capable of engulfing foreign antigens from invading pathogens. Through activation of pathogen recognition receptors (PRRs), mature and activated DC via complex downstream signaling pathways (such as cell surface co-stimulatory molecules) assemble the antigen peptide-major histocompatibility complex (MHC) and promote antigen-specific T cell expansion [96]. In early phase I trials, activated DC vaccines showed good safety, but had a weak anti-tumor response, which was not impressive compared with chemotherapeutic regimens. The α-type-1 polarized DC activated by maturation reagents (tumor necrosis factor α, IL 1β, IFN α, IFN γ and polyinosinic acid-polycytidylic acid) and pulsed with cocktail of 5 synthetic peptides (WT-1, HER2, MAGE-A3, and MAGE-A1 or gp100) were well tolerated, except for transient liver dysfunction. Long-term recurrence-free and positive immunological responses were only observed in 1.3% of HLA-A24⁺ glioma patients with stable disease (SD) [97]. The valuable target antigens of immunogenic synthetic peptides that are in existence in tumor lysate still seem to have an advantage to improve the vaccine-induced benefits and relapse-free period and optimal combinations need to be developed.

Vaccination using autologous tumor cells benefit existing humoral and cell-mediated immunity to antigenic epitopes, as well stimulates polyclonal immune attack against multiple, even undetected, TAAs. In separate reports, preclinical evidence shows that tumor cell immunotherapy can enhance anti-glioma immunity and can be effective in intracranial glioma models [98, 99] (Fig. 3). Autologous tumor-derived peptides from GBM can be used to safely immunize patients with recurrent GBM. After vaccination, brain biopsies revealed significant infiltration of focal CD4⁺, CD8⁺, CD56⁺ and IFN γ producing T cells against autologous tumor-derived peptides bound to HSP-96 [100]. Additionally, the NIH Clinical Center provided the promising attempt of administration of tumor lysate vaccine with a potent immune response modifier, such as imiquimod or poly-polyinosinic-polycytidilic acid (ICLC) adjuvant, to stimulate cell-mediated immune responses (Table 2). However, even though the feasibility and safety of vaccination were noted, the relatively weak anti-tumor activity of vaccination with irradiated autologous glioma has been demonstrated in recent clinical trials.

Following the observance of the effective anti-tumor ability of GM-CSF in vitro and in vivo, the US Food and Drug Administration has approved GM-CSF for use with dose-intensive chemotherapy. A strategy referred to as “GVAX”, composed of GM-CSF-engineered irradiated autologous tumor cells, has demonstrated an immunostimulatory effect related to GM-CSF in extensive preclinical results [101]. Using autologous tumor vaccine plus GM-CSF as an adjuvant, 89% of the vaccinated patients developed an autologous tumor-associated delayed type hypersensitivity (DTH) response; 42% of patients showed radiological evidence of a response, while only 26% patients showed clinical improvement after vaccination [102]. In a phase I study of recurrent malignant glioma (N = 10), irradiated autologous glioma cells were mixed with irradiated K562 cells, which over-expressed GM-CSF, to strengthen its DTH responses and humoral immunity. Vaccination revealed activation of T-lymphocytes, with increasing cytotoxic T-lymphocyte-associated protein 4 (CTLA4), PD-1, 4-1BB, and OX40 expressed CD4⁺ T cells, as well as PD-1 and 4-1BB expressed CD8⁺ T cells. No dose-limiting toxicities were seen when vaccinating patients with subcutaneous and intradermal injections of irradiated autologous glioma cells [103]. Transforming growth factor-beta 2 (TGF-β2) is a secreted protein known as an immunosuppressive molecule, whose dysregulated signaling contributes to the initiation and progression of many cancers, including glioma [104]. Preclinical study demonstrated that inhibition of TGF-β2 expression significantly enhances tumor-cell immunogenicity in eliminating previously implanted tumors [105]. Administration of TGF-β2 modified tumor cells showed recoverable and low-grade treatment-related toxicities and may be safe for glioma patients. In a phase I clinical trial, injection of TGF-β2 antisense-modified autologous tumor vaccine increased survival in some patients and induced generation of cellular and humoral anti-tumor immune responses in stage IV astrocytoma [106].

Vaccination-generated Abs may contribute to the in vivo immunogenicity of selected tumor antigens and may contribute to the therapeutic efficacy of cellular vaccines mainly designed to induce/up-regulate a tumor-specific CTL response. Recently, therapeutic vaccines based on genetically modified allogeneic tumor cells have been evaluated in clinical trial, regarding its advantage of acceptable construction costs, and the host of special antigen presenting. Clinical support for an allogeneic cellular vaccine approach has been validated in melanoma and has shown the extent of immunization against selected tumor antigens [107], but such is not the case in glioma. These findings support the clinical application of autologous tumor vaccines comprised of systemic immunosuppression antigens in glioma. Further randomized trials with allogeneic tumor cell immunotherapy may take a chance to establish efficacy as an alternative approach.

The risk of local recurrence occurs highly at and around the site of injury after surgical removal of the tumor [108]. Previous studies demonstrated fibroblasts may be applicable to cancer-targeting gene therapy for local control of the tumor around the injured tissue [109]. Using autologous glioma cells and IL-4 gene transfected fibroblasts, the feasibility and safety of adjuvant vaccinations were evaluated in patients with newly diagnosed and recurrent GBM, and it demonstrated encouraging immunological and clinical responses without allergic encephalitis [110]. In recurrent GBM, IL-4 dose-dependent infiltration of CD4⁺ and CD8⁺ T cells was observed at local vaccine sites. Vaccinations in HLA-A2⁺ patients demonstrated systemic T cell responses against an HLA-A2-restricted GAA epitope, EphA2883-89. In newly diagnosed GBM, monocyte-derived DC produce high levels of IL-12 p70 followed by two intradermal vaccinations with transfected fibroblasts admixed with DC loaded with autologous tumor lysate. However, detectable IFN γ post-vaccine responses or prolonged progression-free survival was not observed in these participants. To improve the intensity of the vaccine regimen, future studies will need to identify the potential vaccine intensification approach to enhance therapeutic efficacy without detrimental counterbalancing of autoimmunity.

NK cells are highly efficient in the cellular immune response against diseases, including malignancies. The past several years have seen tremendous advances in the selection and expansion of NK cells, and they have been used in clinical trials as adaptive immunotherapy for cancer [111]. Current NK cell-based cancer immunotherapy was designed to improve NK cell paralysis using several approaches, which are more practical for quality control and large-scale production by using stable autologous NK cell lines, for adoptive cellular immunotherapy [112]. In a phase I clinical trial, autologous NK cell therapy was demonstrated to be safe and partially effective in patients with recurrent GBM. The NK cell-rich effecter cells were manufactured from PBMCs through co-culturing with irradiated human feeder cell line (HFWT) and IFN β [113]. To a certain degree, autologous NK cell therapy could not yet exhibit their full cytotoxic capacity in vivo due to MHC class I expression in cancer patients that suppresses autologous NK cells [114]. A previous study demonstrated NK cell cytotoxicity is up-regulated by killer cell immunoglobulin-like receptors (KIRs), which interact with HLA class I ligands [115], while this finding was not tested in phase I clinical trial in human patients with glioma.