Vaccination in the immunotherapy of glioblastoma

ABSTRACT

Glioblastoma remains one of the most common central nervous system tumors with an extremely poor prognosis. Recently, rapid progress in immunotherapy has provided new options for the treatment of glioblastoma. Vaccination, the primary method of immunotherapy, stimulates the body's tumor-specific immune response by the injection of foreign antigens. Peptide vaccines involve the injection of tumor-specific antigens, such as EGFRvIII or heat-shock proteins. Cell-based vaccines, which primarily include dendritic cell vaccines and tumor cell vaccines, involve injections of ex vivo-modified cells. Despite the encouraging results of phase I/II clinical trials, no successful phase III clinical trials involving glioblastoma immunotherapy, including glioblastoma vaccinations, have been reported to date. In this review, the authors summarize the published outcomes of glioblastoma vaccine therapy, explore its future prospects based on ongoing clinical trials, and discuss combined therapy as a future direction for glioblastoma treatment.

KEYWORDS:

• Glioblastoma
• immunotherapy
• vaccination
• peptide vaccine
• cell-based vaccine

Glioblastoma multiforme (GBM) is an extremely malignant grade IV central nervous system (CNS) tumor,¹ accounting for 30% of all gliomas with an incidence of approximately 3.19 patients per 100,000 people annually.^2,3

Stupp et al. proposed the following standard treatment for GBM: patients underwent tumor resection, followed by radiochemotherapy (during the radiotherapy, the tumors were irradiated at a dose of 2 Gy/day, 5 times per week for 6 weeks, and the total dose was 60 Gy; during the chemotherapy, temozolomide (TMZ) was administered at 75 mg per square meter of body surface area daily, 7 days per week from the first to the final day of radiotherapy) and six cycles of adjuvant TMZ (150 to 200 mg per square meter for 5 days during each 28-day cycle).⁴ The prognosis of GBM patients remains poor: The median overall survival (OS) is 14.6∼16.7 months with the current standard-of-care therapy.^5–8 Treatment- and prognosis-related molecular markers, such as the methylation of O6-methylguanine-DNA methyltransferase (MGMT), 1p/19q co-deletion, mutation of IDH, mutation of the TERT promoter, overexpression of epidermal growth factor receptor (EGFR), and new therapies such as immunotherapy, are continuously emerging.^5,9

Immunotherapy is defined by the National Cancer Institute (NCI) as ‘a type of biological therapy that uses substances to stimulate or suppress the immune system to help the body fight cancer or other diseases’.¹⁰ Immunotherapy is mainly divided into active immunotherapy and passive immunotherapy. Active immunotherapy elicits a tumor-specific immune response through an injection of foreign antigens, primarily via a vaccine injection (including peptide and cell-based vaccines), whereas passive immunotherapy, including antibody therapy and adoptive immunotherapy, achieves anti-tumor responses through an injection of novel immunemodulating biologics (e.g., antibodies) instead of directly activating the body's immune system.¹¹ Recently, immunotherapy has become a new antineoplastic regimen in addition to surgery, radiotherapy, chemotherapy and targeted therapy in cancers such as melanoma and non-small cell lung cancer, yet none immunotherapy approaches have been approved for clinical use to treat GBM. This review aims to summarize the current clinical and experimental outcomes of vaccination-based active immunotherapy for GBM and, thus, provide a reference for the clinical management of GBM.

Immunology of the central nervous system (CNS)

Previously immune-privileged site

The CNS was traditionally considered an immune-privileged site due to its lack of a lymphatic system and its separation from the blood circulatory system by the blood-brain barrier.¹² However, several studies have revealed the mechanism underlying the circulation of immune cells and systems in the CNS. Microglia constitute one type of CNS immune cell, accounting for 5% to 20% of all glial cells and playing important roles in innate and acquired immunity (e.g., antigen presentation).¹³ Interestingly, Fonseca et al. investigated the aggregation of abundant microglia and macrophages in GBM and hypothesized that microglia promoted tumor progression by interacting with the tumor.¹⁴ Activated T-lymphocytes, which are located in cervical lymph nodes, can penetrate the blood-brain barrier, enter the CNS and contribute to the interchange between the CNS and lymphatic system.¹⁵ The immune system can also communicate with the brain parenchyma via cerebrospinal fluid (CSF) and interstitial fluid (ISF) and can transport antigens and immune cells.¹⁶

Immunosuppression in GBM

Although the CNS has a unique immune mechanism and sequestration, immunotherapy for GBM continues to face immunosuppression. GBM cells create immunosuppressive microenvironments mainly by secreting certain cytokines and expressing membrane proteins.¹⁷

GBM can inhibit immune system attacks by secreting cytokines. The tumor cell secretes chemokines (e.g., CCL2), which recruit regulatory T cells (Treg) and inhibit the proliferation and activity of effector T cells and antigen-presenting cells (APCs) through anti-inflammatory cytokines, e.g., interleukin 10 (IL-10) and transforming growth factor beta (TGF-β).^18,19 In addition, tumor-associated macrophages (TAMs) and myeloid-derived suppressor cells (MDSCs) can inhibit the immune response in the GBM microenvironment.¹⁹

GBM can directly or indirectly attenuate the function of immune cells by expressing membrane proteins. By expressing the FASL, B7 and PDL1 receptors, GBM cells bind to negative regulatory factors of immune functions (e.g., FAS, CTLA-4 and PD-1), thereby inhibiting the activity of T cells or NK cells and promoting the immune escape of GBM.^20–23 The downregulation of classical MHC class I molecules and the overexpression of certain non-classical MHC class I molecules (e.g., HLA-G and HLA-E) can inhibit the cytotoxicity of NK cells and cytotoxic T cells, which contributes to tumor immune escape.^24,25

Additionally, the irregular vascular structure of GBM and the low perfusion of the tumor could result in a reduction in immune and treatment responses.²⁶

Origins of vaccine immunotherapy and non-specific vaccines

The earliest report of a cancer treatment using a vaccine was reported in 1891 by William B. Coley, who injected streptococcal organisms into a patient with inoperable cancer and observed the disappearance of the tumor.²⁷ Bacillus Calmette-Guerin, which was developed and licensed for the treatment of superficial urothelial carcinoma of the bladder,²⁸ was applied to the treatment of GBM in 1980,²⁹ and soon after, another clinical study reported that this therapy could prolong patient survival.³⁰ Subsequently, physicians treated glioma patients concurrently with ImuVert (which is derived from the bacterium Serratia marcescens) and radiotherapy, and those patients tolerated this therapeutic regimen well with a certain degree of efficacy.³¹ At the turn of the century, infecting human tumor cells with Newcastle Disease Virus (NDV) (act as an oncolytic virus in addition to immunomodulation) resulted in the upregulation of HLA, which induced the production of interferon and ultimately resulted in apoptosis; NDV was then applied for treating patients with recurrent GBM.^32,33 These non-specific vaccines achieved an anti-tumor response by stimulating the immune system, which was similar to a postoperative infection as previously reported.³⁴ The existing vaccine studies exploring GBM are listed in Table 1.

Table 1. Published results of the vaccination in glioblastoma.

Year	Pts No.	Disease	WHO grade	Vaccine	Result	Toxicity
1993^31	15	nAA, nGBM	Grade III, IV	ImuVert	PFS = 7.8 m	57% grade 3 or 4 flu-like symptoms, 15% grade 3 hypotension
					OS = 18.2 m
2006^33	14	rGBM	Grade IV	NDV	1 patient with CR	Grade 1/2 constitutional fever
2009^48	12	nGBM	Grade IV	PEP-3-KLH	PFS = 6.8 m	Limited to grade 2 toxicities; no allergic reactions or serious adverse events
		(EGFRvIII)			OS = 18.7 m
2010^49	18	nGBM	Grade IV	PEP-3-KLH	PFS = 14.2 m	Limited to grade 2 toxicities, mostly related to injection-site reactions; no autoimmune reactions
		(EGFRvIII)			OS = 26 m
2011^50	22	nGBM	Grade IV	PEP-3-KLH	PFS = 15.2 m	Allergic reactions, with grade 3 skin (14%) and gastrointestinal allergies (5%)
		(EGFRvIII)			OS = 23.6 m
2015^51	65	nGBM	Grade IV	PEP-3-KLH	PFS = 12.3 m	Mild-to-moderate injection-site reactions occurred in nearly all patients (≥86%). Fatigue (26%), rash (17%), nausea (12%), pruritus (62%), and headache (8%) of any grade also occurred. Grade 3 or 4 events were relatively rare and limited to single patients.
		(EGFRvIII)			OS = 24.6 m
2015^53(retracted)	36	rGBM	Grade IV	PEP-3-KLH	OS = 11.3 m	Frequent grade 1–2 injection-site reactions. 25% ≥grade 3 adverse events, including back pain (6%), convulsion (11%), fall (3%), headache (3%), and hypertension (3%).
		(EGFRvIII)
2016^56	371	nGBM	Grade IV	PEP-3-KLH	OS = 17.4 m	Mostly limited to grade 2 toxicities. 80% of patients showed injection-site reactions (primarily erythema, pruritus, and rash). Fatigue, nausea, headache, and constipation also occurred in ≥20% of patients. 5% of patients showed brain edema.
		(EGFRvIII)
2013^72	12	rHGG	Grade III, IV	HSPPC-96	OS = 11.0 m	Mild injection-site erythema and/or induration (50%) related to vaccine
2014^73	41	rGBM	Grade IV	HSPPC-96	OS = 10.0 m	The toxicity associated with the vaccine was minimal and mostly related to injection-site reactions (41%). 2% grade 3 constitutional events (fatigue) related to vaccine
2016^84	9	rGBM, recurrent anaplastic glioma	Grade III, IV	SurVaxM	PFS = 4.1 m	67% experienced injection-site reaction (all grade 1). Myalgia (22%), lymphopenia (33%), and leukopenia (33%) of grade 1 or 2 were also observed.
		(survivin-positive)			OS = 20.2 m
2001^92	7	nGBM	Grade IV	DCs vaccine	OS = 15.2 m	No autoimmune reactions. Fever (14%), nausea (14%) and vomiting (14%), and enlarged lymph nodes (14%) were observed.
2004^93	8	rGBM	Grade IV	DCs vaccine	OS = 31.1 m	No autoimmune reactions. Headache (38%), fatigue (25%), erythema (12%), and seizures (25%) were observed and limited to grade 2 toxicities.
2007^94	7	rAA, rGBM		DCs vaccine	–	Skin reactions (28%) and headache (14%) were observed and limited to grade 2 toxicities.
2005^95	12	nGBM, rGBM	Grade IV	DCs vaccine	PFS = 19.9 m	Fever and/or flu-like symptoms (33%), nausea and vomiting (25%), and enlarged lymph nodes (17%) were observed and limited to grade 2 toxicities.
					OS = 35.8 m
2008^96	13	nAA, rAA, nGBM, rGBM	Grade III, IV	DCs vaccine	OS = 11.6 m	No adverse events related to the vaccination process
2008^97	34	nGBM, rGBM	Grade IV	DCs vaccine	Time to death = 21.4 m in responders	Limited to grade 2 toxicities. 3% had enlarged lymph nodes.
2010^98	8	nGBM	Grade IV	DCs vaccine	PFS = 18 m	12% grade 4 ischemic stroke. Other adverse events were limited to grade 2.
					OS = 24 m
2011^99	16	nGBM, rGBM	Grade IV	DCs vaccine	OS = 12.7 m (for nGBM)	No reports of severe constitutional symptoms or injection-site reactions. 52.9% of patients showed lymphopenia, and 47.1% of patients had elevated levels of serum AST/ALT.
					OS = 32.2 m (for rGBM)
2011^100	23	nGBM, rGBM	Grade IV	DCs vaccine	PFS = 15.9 m	No autoimmune reactions. Nausea/vomiting (14%), headache (5%), fatigue (14%), diarrhea (9%), fever (5%), and injection-site reaction (18%) were limited to grade 2 toxicities.
					OS = 31.4 m
2011^101	10	nGBM	Grade IV	DCs vaccine	PFS = 9.5 m	10% of patients had grade 2 unilateral neck pain.
					OS = 28 m
2012^102	13	nGBM	Grade IV	DCs vaccine	PFS = 11.92 m	15% of patients developed mild fever, and 8% of patients had red papules. All disappeared without special treatment.
					OS = 17 m
2012^103	18	nGBM	Grade IV	DCs vaccine	PFS = 8.5 m	Transient abnormal liver function (6%), lymphopenia (6%), pancytopenia (6%), and increased intracranial pressure (6%).
					OS = 31.9 m
2012^104	77	nGBM	Grade IV	DCs vaccine	PFS = 10.4 m	39% of patients developed a series of adverse events (grade 3/4/5), including 23.3% of patients with hematological adverse events.
					OS = 18.3 m
2012^105	9	rAA, rAO, rGBM	Grade III, IV	DCs vaccine	–	11% of patients had mild hepatic dysfunction, which was limited to grade 2 toxicities.
2013^106	21	nGBM, rGBM	Grade IV	DCs vaccine	PFS = 16.9 m	Diarrhea (9%), fatigue (27%), flushing (9%), pruritus (18%), rash (18%), redness on skin (9%), and vomiting (9%) were observed and limited to grade 2 toxicities.
					OS = 38.4 m
2013^107	7	nGBM	Grade IV	DCs vaccine	PFS = 23.1 m	Grade 1 or 2 fatigue (86%), anorexia (71%), and headache (57%) were most often reported, and 14% of patients had grade 3 fatigue.
					OS = 25.3 m
2013^110	34	nGBM, rGBM, newly diagnosed or recurrent anaplastic glioma	Grade III, IV	DCs vaccine	—	Flu-like symptoms, injection-site reactions, and lymphadenopathy were common and limited to grade 2 toxicities. Grade 3 adverse events (i.e., seizure) were rare and likely related to tumor progression.
2014^108	13	rGBM	Grade IV	DCs vaccine	22% PR	—
					56% SD
2015^109	10	rAA, rAO, rGBM	Grade III, IV	DCs vaccine	OS = 26 m	Injection-site reactions (100%), fever (60%), and fatigue (30%) were observed and were all limited to grade 2 toxicities.
2017^111	31	nGBM	Grade IV	DCs vaccine	PFS = 12.7 m	No adverse events or toxicities attributable to the DC vaccine were documented.
					OS = 23.4 m
2007^113	12	nGBM, rGBM	Grade IV	Formalin -fixed tumor vaccine	OS = 10.7 m	No allergic dermatitis or anaphylaxis was observed. Skin irritation (including erythema, induration, and swelling of the inoculation site) and low-grade fever with mild fatigue (50%) were observed and limited to grade 2 toxicities.
2014^114	24	nGBM	Grade IV	Formalin -fixed tumor vaccine	PFS = 8.2 m	Skin irritation (68%, including local erythema, induration, and swelling at the injection site), fever (17%), appetite loss (4%), seizure (8%), and headache (8%) were observed and were all limited to grade 2 toxicities.
					OS = 22.2 m
2016^125	45	nGBM	Grade IV	IMA950	OS = 15.3 m	Injection-site reactions were the most common and limited to grade 2 toxicities. Grade 4 neutrophil count decrease (4%), grade 3 fatigue and anaphylaxis (4%) were observed.
2010^122	97	nGBM	Grade IV	Poly-ICLC	OS = 17.2 m	The most frequent grade 3–4 toxicities were neutropenia (20.6%), leukopenia (16.5%), and thrombocytopenia (9%).

Notes: Pt. No.: patient number; nGBM: newly diagnosed glioblastoma multiforme; rGBM: recurrent glioblastoma multiforme; nAA: newly diagnosed anaplastic astrocytoma; rAA: recurrent anaplastic astrocytoma; rAO: recurrent anaplastic oligodendroglioma; BCG-CW: Bacillus Calmette-Guerin cell wall; HGG: high-grade glioma; NDV: Newcastle Disease Virus; CR: complete response; AST: aspartate aminotransferase; ALT: alanine aminotransferase.

Peptide vaccines

Tumor-specific antigens are antigens that are present on tumor cells and generally absent from normal tissues. These antigens are often proteins that are encoded by mutant genes in the tumor, which are relatively conserved among different types of cancers and patients and can serve as targets for immunotherapy.³⁵ Peptide vaccines induce immune responses through an injection of tumor-specific antigens, which leads to the destruction of tumor cells.³⁶

EGFRvIII-targeted peptide vaccines

EGFR expression plays a key role in tumorigenesis. EGFR, along with ErbB-2 (HER2), ErbB-3 (HER3) and ErbB-4 (HER4) constitute the ErbB family of receptor tyrosine kinases³⁷ and form dimers when activated.³⁸ The ErbB family is closely related to many downstream pathways (e.g., the PI3K/AKT/mTOR and RAS/RAF/MEK pathways), and the mutation or continuous activation of these pathways results in the development of GBM.³⁹ EGFR is overexpressed in 50% to 60% of patients with GBM,⁴⁰ but EGFR-targeted drugs, such as nimotuzumab, do not prolong the progression-free survival (PFS) or OS of patients.⁴¹

EGFRvIII (type III epidermal growth factor receptor mutation), which is present in 20% to 30% of patients with GBM, is formed due to the deletion of exons 2∼7 of EGFR, which leads to an extracellular truncation of EGFR and allows EGFR to be continuously activated in the absence of ligand.⁴⁰

The earliest study investigating an EGFRvIII-targeted peptide vaccine was reported in 1997 by Moscatello et al., who obtained the peptide vaccine PEP-3-KLH (generic name: Rindopepimut, also known as CDX-110) by conjugating PEP-3, which is a 14-amino-acid peptide that comprises an EGFRvIII-specific mutant part (amino acid sequence: LEEKKGNYVVTDHC), to keyhole limpet hemocyanin (KLH).⁴² They induced the corresponding antibody and observed an immune response in cytotoxic T cells in a mouse model.⁴² Heimberger et al. observed that the PEP-3-KLH vaccine significantly prolonged the median survival time in an EGFRvIII-expressing murine brain tumor model.^43,44 In 2008, Schmittling et al. first reported that the PEP-3-KLH vaccine could induce newly diagnosed patients with GBM to produce an EGFRvIII-specific antibody.⁴⁵ Sampson et al. found a significant inverse correlation between the frequency of Treg and PEP-3-KLH-stimulated humoral immunity during combined treatment with the anti-IL-2Rα MAb daclizumab, TMZ and PEP-3-KLH, which explains the role of the PEP-3-KLH vaccine.⁴⁶ Saraswathula et al. observed that the administration of the vaccine plus TMZ did not stimulate regulatory B cells and further inhibited the secretion of IL-10 and TGF-β to a certain extent, which disinhibited the activation of effector T cells and APCs.⁴⁷

The first clinical trial investigating the PEP-3-KLH vaccine was reported in 2009. Twelve patients with newly diagnosed GBM received three consecutive intradermal vaccinations with PEP-3-KLH and dendritic cell injections. The toxicity was minimal and limited to grade 2 toxicities. Although some patients showed a small increase in the erythrocyte sedimentation rate (33% of patients) and rheumatoid factor level (10% of patients) after vaccination, no associated clinical symptoms were observed.⁴⁸ In a subsequent phase II clinical trial (ACTIVATE), 18 patients with newly diagnosed EGFRvIII-expressing GBM received a PEP-3-KLH vaccination, and the median PFS and OS in the treatment group were 14.2 and 26 months, respectively, which were significantly longer than those in the control groups who were on TMZ (PFS: 6.3 months; OS: 15.0 months).⁴⁹ In a phase II clinical trial (ACT II) investigating PEP-3-KLH vaccination in 22 patients with newly diagnosed EGFRvIII-expressing GBM, the median PFS and OS were 15.2 and 23.6 months, respectively, which was also significantly longer than historical data (PFS: 6.3 months; OS: 15.0 months).⁵⁰ Furthermore, the authors observed that when patients were simultaneously treated with TMZ and vaccines, the dose intensity of TMZ caused lymphopenia, while the vaccine resulted in increased tumor-specific lymphocytosis.⁵⁰ No autoimmune reactions were documented in ACTIVATE or ACT II, and adverse events in the two trials were mostly limited to grade 2, most of which were related to injection-site reactions (e.g., erythema and pruritus). Some patients in ACT II had allergic reactions, including grade III skin and gastrointestinal allergic reactions.^49,50

In 2015, a multicenter, single-arm phase II clinical trial (ACT III) investigated the efficacy of PEP-3-KLH plus standard adjuvant TMZ chemotherapy in 65 patients with newly diagnosed EGFRvIII-expressing GBM.⁵¹ As the treatment progressed, the expression of the serum anti-EGFRvIII antibody increased significantly, but no T cell immune response was observed due to the cytotoxic effect of TMZ on immune cells.⁵¹ The median PFS and OS were 12.3 and 24.6 months, respectively, which was similar to the results observed in ACTIVATE and ACT II.⁵¹ The methylation status of the MGMT promoter, which was a predominant predictor of GBM prognosis and the response to TMZ,⁵² also serves as a prognoses factor in ACT III.⁵¹ A phase II clinical trial (ReACT) investigating PEP-3-KLH plus bevacizumab in patients with recurrent GBM reported an improved prognoses at the Society of Neuro-oncology Annual Meeting in 2015⁵³. However, the article was withdrawn due to subsequent data analysis.^54,55 Adverse events in ACT III and ReACT were similar to those in ACTIVATE and ACT II, with mild-to-moderate injection-site reactions occurring in nearly all patients. Grade 3 or 4 events, including back pain, convulsion, headache, and hypertension, were relatively rare and limited to single patients.^51,53

A multicenter, double-arm phase III clinical trial (ACT IV) investigating PEP-3-KLH was recently completed. This clinical trial enrolled 700 patients with newly diagnosed GBM. Patients in the treatment arm were treated with PEP-3-KLH plus TMZ, and individuals in the control arm were treated with KLH plus TMZ, with the patients' OS as the primary efficacy endpoint. However, at the Society of Neuro-oncology Annual Meeting in November 2016, Weller et al. reported a result of experimental failure. PEP-3-KLH was sufficiently safe in ACT IV, with 80% of patients showing grade 1 or 2 injection-site reactions (primarily erythema, pruritus, or rash). Fatigue, nausea, headache, and constipation also occurred in ≥20% of patients but were mostly limited to grade II toxicities. The patients showed a good humoral immune response, but compared to the control group, patients with minimum residual disease (MRD) and all intention-to-treat (ITT) patients exhibited only a small difference in the median OS (PEP-3-KLH vs. control: MRD: 20.1 months vs. 20 months; ITT: 17.4 months vs. 17.4 months). Only patients with bulky disease exhibited a significant difference in the 2-year OS between individuals who received a therapeutic regimen of PEP-3-KLH plus TMZ and the controls (30% vs. 19%, P = 0.029).⁵⁶ Weller et al. also found that the prognosis of the control group was significantly beyond expectations, which suggested there was an existing problem with the use of historical data as a control in the previous single-arm phase II clinical trials.⁵⁶ Although PEP-3-KLH exhibited sufficient safety in the clinical trial, its efficacy in the phase III clinical trial was unsatisfactory. The extent of tumor resection and the threshold for ‘EGFRvIII-positive’ criteria may be critical factors that impeded the effectiveness of the PEP-3-KLH vaccine.⁵⁷ Nevertheless, the current results of ACT IV failed to demonstrate the efficacy of the PEP-3-KLH vaccine, and subgroup analyses of ACT IV are necessary to determine the clinical role of the peptide vaccine.

The failure of ACT IV terminated the development of EGFRvIII-targeted peptide vaccine. Despite, EGFRvIII still serves as a tumor-specific antigen for GBM in many other immunotherapy studies, such as the antibody-drug conjugates (ADCs) AMG 595 and ABT-414, although does not constitute vaccination therapy. In addition, a phase I clinical trial of ADU-623, a live-attenuated Listeria Monocytogenes Strain expressing the EGFRvIII-NY-ESO-1 vaccine, is underway (NCT01967758).⁵⁸ The ongoing clinical trials investigating therapies with GBM vaccines are listed in Table 2.⁵⁸

Table 2. Ongoing clinical trials of the vaccination in glioblastoma.⁵⁸

Trail code	Disease	WHO grade	Vaccine	Phase	Start time	Status
NCT01967758	rAA, rGBM	Grade III, IV	ADU-623	Phase I	2013	Recruiting participants
NCT00905060	nGBM	Grade IV	HSPPC-96	Phase II	2009	Completed
NCT02122822	nGBM	Grade IV	HSPPC-96	Phase I	2013	Ongoing, but not recruiting participants
NCT02722512	Newly diagnosed or recurrent pediatric HGG, ependymoma	—	HSPPC-96	Phase I	2016	Recruiting participants
NCT01814813	rGBM	Grade IV	HSPPC-96	Phase II	2013	Recruiting participants
NCT03018288	nGBM	Grade IV	HSPPC-96	Phase II	2017	Recruiting participants
NCT02454634	nAA, nAO, nGBM	Grade III, IV	Derived from IDH1 R132H	Phase I	2015	Recruiting participants
NCT02960230	Newly diagnosed DIPG, newly diagnosed glioma	—	K27M peptide	Phase I	2016	Recruiting participants
NCT02455557	nGBM	Grade IV	SurVaxM	Phase II	2015	Recruiting participants
NCT00045968	nGBM	Grade IV	DCs vaccine (DCVax)	Phase III	2006	Ongoing, but not recruiting participants
NCT02146066	nGBM, rGBM	Grade IV	DCs vaccine (DCVax)	Expanded access	2014	Currently available
NCT00890032	rGBM	Grade IV	DCs vaccine (brain tumor stem cells mRNA loaded)	Phase I	2009	Completed
NCT01171469	rAA, rGBM, recurrent medulloblastoma, recurrent ependymoma	—	DCs vaccine (brain tumor stem cells as antigen)	Phase I	2010	Completed
NCT02820584	rGBM	Grade IV	DCs vaccine (tumor stem cell-loaded)	Phase I	2016	Not yet open
NCT01522820	rGBM	Grade IV	DCs vaccine (fusion peptide loaded)	Phase I	2012	Completed
NCT02709616	nGBM	Grade IV	DCs vaccine (tumor mRNA loaded)	Phase I/II	2016	Recruiting participants
NCT00639639	nGBM	Grade IV	DCs vaccine (pp65 RNA loaded)	Phase I	2006	Ongoing, but not recruiting participants
NCT02465268	nGBM	Grade IV	DCs vaccine (pp65 RNA loaded)	Phase II	2016	Recruiting participants
NCT02366728	nGBM	Grade IV	DCs vaccine (pp65 RNA loaded)	Phase II	2015	Recruiting participants
NCT01567202	nGBM, rGBM	Grade IV	DCs vaccine (autogeneic glioma stem-like cells (A2B5+) loaded)	Phase II	2012	Recruiting participants
NCT01204684	nAA, rAA, nAO, rAO, nGBM, rGBM	Grade III, IV	DCs vaccine (tumor lysate loaded)	Phase II	2010	Ongoing, but not recruiting participants
NCT00626483	nGBM	Grade IV	DCs vaccine (RNA loaded)	Phase I	2007	Ongoing, but not recruiting participants
NCT01957956	nGBM	Grade IV	DCs vaccine (tumor lysate loaded)	Phase I	2013	Recruiting participants
NCT02049489	rGBM	Grade IV	DCs vaccine (peptide loaded)	Phase I	2016	Ongoing, but not recruiting participants
NCT02649582	nGBM	Grade IV	DCs vaccine (Wilms' tumor 1 mRNA loaded)	Phase I/II	2015	Recruiting participants
NCT01808820	rAA, rGBM	Grade III, IV	DCs vaccine + tumor lysate boost	Phase I	2013	Recruiting participants
NCT02010606	nGBM, rGBM	Grade IV	DCs vaccine (allogenic GBM stem-like cell lysate loaded)	Phase I	2013	Recruiting participants
NCT03014804	rGBM	Grade IV	DCs vaccine (tumor lysate loaded)+ nivolumab	Phase II	2017	Not yet open
NCT01400672	Recurrent brain stem GBM	Grade IV	Tumor lysate vaccine	Phase I	2012	Completed
NCT01759810	rGBM	Grade IV	HSCs, DCs vaccine, CTLs	Phase II/III	2012	Enrolling by invitation
NCT02754362	rGBM	Grade IV	TAA, Poly-ICLC, KLH, bevacizumab	Phase II	2016	Not yet open
NCT02078648	rGBM	Grade IV	SL-701, poly-ICLC, bevacizumab	Phase I/II	2014	Ongoing, but not recruiting participants
NCT02546102	nGBM	Grade IV	ICT-107	Phase III	2015	Recruiting participants
NCT01920191	nGBM	Grade IV	IMA950, Poly ICLC	Phase I/II	2013	Completed
NCT01222221	nGBM	Grade IV	IMA950, GM-CSF	Phase I	2010	Completed
NCT02510950	nGBM	Grade IV	Personalized peptide vaccine, Poly ICLC	Phase 0	2015	Recruiting participants

Notes: nGBM: newly diagnosed glioblastoma multiforme; rGBM: recurrent glioblastoma multiforme; nAA: newly diagnosed anaplastic astrocytoma; rAA: recurrent anaplastic astrocytoma; nAO: newly diagnosed anaplastic oligodendroglioma; rAO: recurrent anaplastic oligodendroglioma; HGG: high-grade glioma; TAA: tumor-associated antigen.

Heat shock protein vaccines

Heat shock proteins (HSPs) are a family of proteins that respond to temperature changes and are widely distributed in bacteria, plants and animals. Due to their molecular chaperone activity, HSPs can inhibit the biomacromolecular denaturation induced by temperature, oxygen, and ions.⁵⁹ The molecular weights of HSPs range from 10 to over 100 kDa and are closely related to the locations of HSPs in the cell.⁶⁰ The expression of HSPs is regulated by many pathways,^61,62 and HSPs are overexpressed in many cancers and are associated with cancer cell proliferation, differentiation, invasion and metastasis.^62,63 HSPs can serve as therapeutic targets and prognostic indicators for several tumors.^63–65 HSP27, HSP72, HSP73/HSP70, and HSP90 have been confirmed to be overexpressed in glioblastoma cell lines and mouse models.^66,67

HSP vaccines are composed of HSPs and autologous tumor-antigenic peptides.⁶⁸ Neither individual HSPs nor autologous tumor-antigenic peptides can elicit an immune response, but HSP-peptide complexes (HSPPCs) can mediate endocytosis by binding to APC membrane receptors and activating CD4⁺ and CD8⁺ T cells and certain APC signaling pathways to trigger immune responses to tumor-antigenic peptides by antigen presentation.^68,69 Compared to antigen-specific peptide vaccines (e.g., EGFRvIII-targeted vaccines), HSP vaccines can harbor multiple tumor antigens and respond to the poor outcomes induced by tumor heterogeneity and immunoediting, but their specificity for tumor antigens differs from that of antigen-specific peptide vaccines.⁶⁸

HSPPC vaccines mainly use HSPPC-96 (generic name: Vitespen), which is a composite of a tumor antigenic peptide and HSP glycoprotein-96, and clinical trials investigating its efficacy in renal cell carcinoma and melanoma have been performed.^70,71 The first clinical trial investigating HSPPC-96 in autologous tumor-derived peptides was reported in 2013 by Crane et al., who conducted a phase I clinical trial involving 12 patients with high-grade recurrent glioma.⁷² The authors obtained non-necrotic tumor tissues from a sample of surgically resected tumors from each patient, and the HSPPC-96 protein was mixed with the tumor antigens, purified and developed into a vaccine.⁷² No distinct toxicity was found in the patients after the vaccination, and a significant peripheral immune response specific to the tumor antigens was observed in the peripheral blood and tumor in 11 of 12 patients after the vaccination, which suggests that the HSPPC-96 vaccine was effective.⁷²

In 2014, the first phase II clinical trial investigating the HSPPC-96 vaccine in GBM was completed. Investigators performed tumor resections in patients with recurrent GBM and produced the HSPPC-96 vaccine. The vaccine was administered every week for 4 weeks, followed by every 2 weeks thereafter until tumor recurrence; the median PFS and OS reached 19.1 and 42.6 weeks, respectively.⁷³ Compared to the efficacy of the bevacizumab/CCNU regimen for recurrent GBM (the median PFS and OS were 7.2–24.4 and 24.8–33.1 weeks, respectively),^74–76 HSPPC-96 displayed a certain degree of efficacy.⁷³ Investigators also studied the relationship between patients' absolute lymphocyte count and their prognosis and found that patients with an above-average lymphocyte count had a significantly better prognosis.⁷³ The patient's enrollment status (operation tolerance and no relapse during the first follow-up period) may also have affected the improvement in the prognosis.⁷³ The toxicity associated with the vaccine was minimal, mostly grade 1 or 2 injection-site reactions (erythema or induration). A single-arm, phase II clinical trial investigating HSPPC-96 (NCT00905060) was completed, but the results have not been published.⁵⁸ In this trial, 46 patients with newly diagnosed GBM were treated with HSPPC-96 plus TMZ, and the safety profile of this therapy and the patients' OS were considered the primary outcome measures. The results of this trial will provide an important reference for the application of HSPPC-96 and a phase III clinical trial.

To date, several ongoing phase I and II clinical trials have investigated HSPPC-96 (see Table 2).⁵⁸ The purpose of the phase I clinical trial (NCT02722512) was to determine the safety profile of the HSPPC-96 vaccine in 20 patients with newly diagnosed high-grade gliomas or recurrent resectable high-grade gliomas and ependymomas. The clinical trial NCT02122822, which was conducted in Beijing Tiantan Hospital in 2013, investigated the safety and effectiveness of an autologous gp96 treatment in 20 patients with GBM. Furthermore, a phase II clinical trial (NCT01814813) investigated the effectiveness of HSPPC-96 when combined with bevacizumab to treat patients with recurrent GBM. In this trial, 165 patients were assigned to the following three groups: HSPPC-96 + concomitant bevacizumab; HSPPC-96 with bevacizumab at progression; and bevacizumab alone, with the patients' OS serving as the primary outcome. The conclusion of this trial may provide insight into resolving the difficulty of treating recurrent GBM. Clinical trial NCT03018288 investigated whether HSPPC-96 improves the efficacy of radiotherapy, TMZ and pembrolizumab during combination therapy. In addition to HSPPC-96, HSP47, which is a glioma-associated antigen, could be a potential target for vaccine therapy.⁷⁷ HSP70, however, induced an immune response in a murine model.⁷⁸

Peptide vaccines specific to other targets

IDH mutation, the most important molecular marker in gliomas, has important effects on multiple pathways, metabolisms and prognoses of gliomas.⁷⁹ IDH mutations have been observed in approximately 80% of low-grade gliomas, of which the most common was the R132H mutation in IDH1 (accounting for 70% of all IDH mutations).⁷⁹ In GBM, mutations in IDH1 tend to indicate that the tumor has evolved from a low-grade glioma, and a small fraction of primary GBM cases exhibit mutations in IDH1.⁸⁰ In 2014, Schumacher et al. designed a 15-amino-acid polypeptide targeting R132H mutations in IDH1, found that it was present on MHC II and induced mutation-specific CD4⁺ responses, and detected its antibody in a mouse model.⁸¹ Two phase I clinical trials investigating the IDH1 peptide vaccine (NCT02193347 and NCT02454634) are ongoing, and their aims are to verify the safety of the vaccine and the immune response in patients with gliomas.⁵⁸

Due to the specific location and limited treatment, patients with diffuse intrinsic pontine gliomas (DIPG) exhibit a median OS⁸² of less than one year. Since the H3.3 K27M mutation acts as a molecular marker of DIPG, and H3.3-derived peptides that encompass this mutation can induce immune responses in HLA-A2⁺ mice,^82,83 the phase I clinical trial NCT02960230 investigating H3.3 K27M is currently being conducted.⁵⁸ Another phase I clinical trial investigating a long peptide vaccine (SurVaxM) targeting Survivin (BIRC5) was recently completed,⁸⁴ and the phase II clinical trial NCT02455557 is underway.⁵⁸

Cell-based vaccines

Cell-based vaccines have long attracted researchers' attention. Autologous or allogeneic cells are re-injected into patients' bodies after in vitro modifications and trigger specific immune responses for tumor killing. This method mainly comprises dendritic cell (DC) vaccines, tumor cell vaccines and other cell-based vaccines.

DC vaccines

DCs are cells with the strongest antigen presentation in the human body, which stimulate the conversion from innate immunity to acquired immunity. DCs are generated from hematopoietic stem cells, mature during migration to the lymph nodes, and play important roles in the immune response, differentiation, and exogenous/endogenous antigen presentation of lymphocytes.⁸⁵

The production of DC vaccines includes the isolation of DCs from the patient's body, the loading of the tumor antigen, the maturation of DCs via cytokines, and the re-injection of DCs into the patient's body.⁸⁶ Sipuleucel-T is the only approved APC vaccine for the treatment of tumors and has been shown to prolong the median OS in patients with prostate cancer by 4 months.⁸⁷ The culture of DCs for a DC vaccine for the treatment of GBM began with CD-14-positive monocytes isolated from the peripheral blood. The monocytes differentiated into immature DCs in the presence of granulocyte macrophage colony-stimulating factor (GM-CSF) and IL-4.⁸⁸ Tumor antigens, including polypeptides, RNAs, DNAs and tumor lysates, can be loaded onto immature DCs and presented on MHCs,⁸⁹ and they mature in the presence of many cytokines (e.g., GM-CSF, IL-4, TNF-α and IL-6).⁸⁸ In addition to tumor antigens, DCs loaded with Cytomegalovirus phosphoprotein 65 (pp65) RNA show a significant improvement in lymph node homing after the vaccine site is pre-conditioned with tetanus/diphtheria, which is a potent recall antigen.⁹⁰ The migration process and the effects of pp65 DCs are currently being investigated in phase I and II clinical trials (NCT00639639, NCT02465268, and NCT02366728).⁵⁸

The use of DC vaccines in GBM was first reported in 2000 by Liau et al., who enabled a patient with recurrent brainstem glioblastoma to survive for 21 months.⁹¹ In 2001 and 2004, Yu et al. loaded DCs with peptides eluted from the surface of autologous glioma cells and tumor lysates and separately investigated their safety profiles and immune responses,^92,93 and the median OS for the 8 recurrent GBM patients in the 2004 study reached 133 weeks.⁹³ In 2007, Okada et al. loaded tumor lysates into DCs and created a vaccine along with TFG-IL4-Neo-TK-transfected fibroblasts, which produced a better safety profile and a certain immune response.⁹⁴ Currently, the results of more than 10 phase I/II trials (see Table 1) have indicated the feasibility of DC vaccines, most of which were tumor lysate-pulsed DCs as well as DCs pulsed with polypeptides and nucleic acids.^95–111 Grade 1 or 2 toxicities, including fever (and/or flu-like symptoms), nausea and vomiting, and enlarged lymph nodes, were the most frequently observed adverse events in these studies. Lymphopenia and hematological adverse events were also reported in some studies and were most likely the result of maintenance TMZ therapy.^95–111 The clinical trial involving the largest number of patients was the HGG-2006 trial, which was published in 2012. Ardon et al. prepared DC vaccines with tumor lysates and treated 77 patients with newly diagnosed GBM.¹⁰⁴ The authors integrated the vaccination into the Stupp regimen as follows: the patients first received 60 Gy of radiotherapy for 6 weeks and daily administration of TMZ, followed by vaccination once per week for 4 weeks; the patients received a 5-day administration of TMZ and were vaccinated once during the subsequent six cycles of a 28-day course.¹⁰⁴ The median PFS and OS were 10.4 and 18.3 months in all patients, respectively, and patients with a lower EORTC RPA classification had significantly better outcomes than patients with a higher EORTC RPA classification; furthermore, patients with a methylated MGMT promoter had better outcomes.¹⁰⁴ The adverse events of HGG-2006 were more severe than in other DC vaccines studies, with 38 series adverse events (grade 3/4/5) found in 30 patients (39%), including 19 hematological adverse events in 18 patients.¹⁰⁴ In 2013, by comparing the research results by Prins et al. with the therapeutic effects of autologous tumor lysate (ATL)-pulsed DC vaccination and glioma-associated antigen (GAA) peptide-pulsed DC vaccination, the authors found that both vaccines caused similar adverse events and activated similar numbers of lymphocytes, but significantly higher frequencies of NK cells were activated by the GAA vaccine than those activated by the ATL vaccine.¹¹⁰ However, patients treated with the ATL-pulsed DC vaccine had significantly better outcomes than those treated with the GAA peptide-pulsed DC vaccine (PFS = 18.1 vs. 9.6 months; OS = 34.4 vs. 14.5 months).¹¹⁰

Dozens of clinical trials investigating DC vaccines are ongoing (see Table 2), most of which are phase I or II trials investigating various DC loads.⁵⁸ The most noteworthy study is the DCVax clinical trial (a DC vaccine project by Northwest Biotherapeutics), which began in 2006¹¹² and has entered a phase III clinical trial NCT00045968.⁵⁸ In this trial, 348 patients with newly diagnosed GBM underwent surgical resection with concurrent radiotherapy and chemotherapy, and tumor lysate proteins were used to prepare DCVax(R)-L. The treatment cohort was vaccinated on days 0, 10, and 20 and on weeks 8, 16, 32, 48, 72, and 120, and PFS was measured as the primary outcome. Expanded access to DCVax is underway (NCT02146066), which will play a critical role in the application of the DC vaccine for the treatment of GBM.⁵⁸

Tumor cell vaccines

In addition to the re-injection of tumor antigen-loaded immune cells into patients' bodies, investigators have directly injected tumor lysates or fixed tumor cells (as antigens) into patients' bodies.^113,114 Ishikawa et al. reported re-injecting formalin-fixed GBM cells into patients' bodies to treat GBM after the standard Stupp regimen.^113,114 In one study, 24 patients were treated with nGBM and exhibited a median OS of 22.2 months.¹¹⁴ The authors concluded that the patients with unchanged p53, higher MHC-I expression or a delayed-type hypersensitivity (DTH) response had better outcomes.^113,114 Clinical trial NCT01400672 is currently investigating the treatment effect of imiquimod/tumor lysate vaccine in brain stem glioma.⁵⁸

Other cell-based vaccines

Other immune cells or stem cells are also promising for the treatment of GBM. Bryukhovetskiy et al. reported that cancer stem cells (CSCs) were very difficult to destroy and could recruit other types of stem cells, and they proteomically identified an interaction between CSCs and hematopoietic stem cells (HSCs).^115,116 An animal model has confirmed this hypothesis,¹¹⁷ and a clinical trial (NCT01759810) investigating concurrent HSCs, DC vaccines and cytotoxic lymphocytes in the treatment of GBM has been conducted.⁵⁸ In addition to cell-based vaccines, other cell-based immunotherapies, such as adoptive cellular therapy (ACT), have shown potential feasibility for clinical application.¹¹⁸

Others

In addition to peptide and cell-based vaccines, immune adjuvants are sometimes used in immunotherapy to enhance efficacy. Problems such as immunoediting must be overcome to allow immunotherapy to have a continuous effect.

Immune adjuvants

An immune adjuvant is a substance that does not initially trigger an immune response but enhances the acquired immune response or changes the type of immune response. If injected with a vaccine, the body's immune response can be enhanced, and the seroconversion rate increases in populations with reduced responsiveness, which simultaneously allows for the use of smaller antigen doses and fewer immunizations.¹¹⁹ Immune adjuvants also alter the type of immune response and increase the generation of memory lymphocytes, the speed of the initial response, and the specificity of the response.¹¹⁹ Common immune adjuvants include the double-stranded RNA derivative Poly-ICLC, oligonucleotide chain CpG, and imiquimod for Toll-like receptors (TLRs) 3, 9 and 7/8.¹²⁰ In the treatment of GBM, oligonucleotide CpG has been confirmed to enhance the immune response in an animal model.¹²¹ Moreover, poly-ICLC plus radiochemotherapy increased the median OS in patients with newly diagnosed GBM to 17.2 months,¹²² and this treatment has been used in several clinical trials (NCT02754362, NCT02078648, NCT01920191, and NCT02510950). Imiquimod has been used in several clinical trials investigating DC and tumor cell vaccines (NCT01171469, NCT01204684, NCT01808820, and NCT01400672).⁵⁸

Immunoediting

Immunoediting refers to the remodeling of tumor cells to promote their motions, invasions and metastases while eliminating by the immune system.^123,124 The relationship between tumors and the immune system is believed to consist of the following three phases: during the ‘elimination phase’, the tumor cells are killed by the immune system; during the ‘equilibrium phase’, there is a state of equilibrium between the tumor cells and the immune system; and during the ‘escape phase’, the tumor cells evade the surveillance of the immune system and eventually form tumors.^123,124 Immunoediting is one of the greatest obstacle in tumor immunotherapy. In phase II clinical trials investigating the EGFRvIII-specific peptide vaccine PEP-3-KLH (ACTIVATE, ACT II and ACT III), 82%, 91.6% and 66.7% of patients lost the EGFRvIII mutation at recurrence, respectively, which suggests that these patients were no longer sensitive to the PEP-3-KLH vaccine.^49–51 Multitarget immunotherapy or multidrug therapy may be a solution to the immunoediting problem: clinical trial NCT02078648 investigated the feasibility of SL-701, poly-ICLC plus bevacizumab for the treatment of GBM; DCs loaded with six tumor-associated antigens (TAAs), including HER2, TRP-2, gp100, MAGE-1, IL13Rα2 and AIM-2, in the ICT-107 vaccine, which produced good phase I results (median PFS in newly diagnosed GBM was 16.9 months, and median OS in newly diagnosed GBM was 38.4 months) and is currently being investigated in a phase III clinical trial (NCT02546102).^58,106 Although clinical trials investigating the IMA950 multi-peptide vaccine plus poly-ICLC or GM-CSF in GBM are ongoing (NCT01920191 and NCT01222221),⁵⁸ a clinical trial in the UK investigating IMA950 did not find a difference in prognosis between multi-tumor-associated peptide responders and single-tumor-associated peptide responders.¹²⁵ The clinical trial NCT02510950 involved designing polypeptide vaccines for all individual patients to increase the intensity and specificity of the immune response, but this trial is still in the stage of dose exploration and safety experiments.⁵⁸ In addition to mixed polypeptides, researchers have proposed that tumor patients could be treated with vaccine therapy in combination with vaccine adjuvants and immune checkpoint-specific antibodies (e.g. ipilimumab for anti-CTLA-4 antibodies, pembrolizumab and nivolumab for anti-PD-1 antibodies, atezolizumab and avelumab for anti-PD-L1 antibodies),^126,127 and several clinical trials investigating combined therapies are currently ongoing (NCT01176474, NCT02515227, and NCT02897765).⁵⁸ However, no such clinical trials for the treatment of GBM have been conducted to date.

Summary and prospects

GBM, which was once considered an immune-privileged site and harbored an immunosuppressive microenvironment, remains a cancer with an extremely poor prognosis and quality of life in patients. Immunotherapy is increasing in applicability as a therapeutic regimen for the treatment of GBM in addition to surgery, radiotherapy, chemotherapy and targeted therapy. The failure of the phase III clinical trial investigating EGFRvIII (ACT IV) set back the prospects for immunotherapy, although several phase I/II clinical trials investigating active immunotherapy (dominated by vaccinations) have shown some promising results. The study and popularization of immunotherapy, especially cell-based vaccines, requires robust infrastructure, including hospitals with sufficient experience in the management of glioma and research centers with adequate investigations in immunotherapy. In addition, expenses may be a barrier to the study and adoption of immunotherapy (especially cell-based vaccines) due to the vaccines being produced individually.

In addition to the resistance of vaccine therapy in GBM that is caused by the tumor heterogeneity and immunoediting, vaccine therapy continues to face many challenges. Patients require the surgical removal or biopsy of the lesion to identify their molecular pathology, prepare the vaccine, and increase the tumor response to the vaccination. Therefore, it is difficult to administer therapeutic vaccines to patients without indication based on surgery or biopsy. Molecular imaging can help identify molecular markers of the tumor to a certain extent (e.g., PET imaging of D-2HG levels to determine whether IDH mutations are present in the patient) and may help patients with surgical limitations in terms of vaccine therapy.

Chemotherapy was once thought to inhibit immune system function. However, researchers studying other tumors recently found that although TMZ chemotherapy significantly reduced the number of T lymphocytes, the ratio of CD8⁺ T cells to total T cells increased.¹²⁸ TMZ chemotherapy also generated chemokines (e.g., CXCL9 and CXCL10), which resulted in significant T-cell accumulation in metastatic melanoma, but not in other cancers, such as skin tumors.¹²⁹ The complex relationship between chemotherapy and the immune system may also affect the efficacy of immunotherapy.

Despite the great success achieved in other cancers, immunotherapy for the treatment of glioblastoma continues to face difficulties. The phase III clinical trial CheckMate 143 (NCT02017717) failed to demonstrate improved OS with nivolumab (anti-PD-1 monoclonal antibody) compared to bevacizumab in patients with recurrent GBM, highlighting the obstacle of passive immunotherapy in glioblastoma.¹³⁰ Although non-phase III clinical trials investigating immunotherapy for GBM have succeeded, studies investigating immunotherapy will continue to be conducted (ongoing clinical trials investigating vaccination in GBM are summarized in Table 2). Two ongoing phase III clinical trials investigating GBM vaccines, i.e., the DCVax (NCT00045968 and NCT02146066) and ICT-107 vaccines (NCT02546102), as well as the ADC drug ABT-414 (NCT02343406 and NCT02573324) and the phase III clinical trial CheckMate 498 (NCT02617589, investigating nivolumab in combination with radiotherapy ± TMZ in patients with newly diagnosed GBM),⁵⁸ represent a silver lining for GBM immunotherapy. Combinations of different therapies, including combinations of different vaccination strategies, combinations of vaccinations with immune checkpoint blockade, and combinations of surgical resection, chemotherapy, radiotherapy, targeted therapy, and immunotherapy, are possible future directions for GBM treatment.

Disclosure of potential conflicts of interest

The authors report no conflict of interest.

Funding

This work was supported by the National Health and Family Planning Commission of the People's Republic of China under Grant number 201402008; and Chinese Academy of Medical Sciences under Grant number 2016-I2M-2-001.