Immunotherapy for glioblastoma: the promise of combination strategies

Glioblastoma (GBM) is the most common and aggressive malignant tumor of the central nervous system (CNS) [1, 2]. GBM is a grade IV diffuse astrocytoma that is thought to arise from neural stem cells or progenitor cells, such as oligodendrocyte precursor cells [2‐4]. Approximately 90% of GBM cases are considered primary GBM, with fast and de novo expansion and without any sign of less malignant precursor tumors. Primary GBM often develops in elderly patients and shows a much poorer prognosis than secondary GBM, which originates from grade II and III astrocytomas, oligodendrogliomas or oligoastrocytomas and most likely manifesting in younger patients [3, 5].

Standard of care (SOC) therapy aims at increasing patient life expectancy and focuses on maximal and safe surgical resection combined with radiotherapy (RT) and adjuvant chemotherapy in the form of oral delivery of temozolomide (TMZ) [6]. Despite this treatment, the median survival of patients diagnosed with GBM is approximately 15 months, with a 2-year life expectancy of less than 30% [7]. For patients with unresectable GBM (up to 35–40% of patients), the prognosis is even poorer [8‐10]. Indeed, microscopic infiltration of GBM cells and the tumor location render total resection difficult or even impossible and produce inevitable recurrences [9]. Finding new therapies for GBM is therefore an urgent unmet need, although it is very challenging because of unique characteristics of GBM and its tumor microenvironment (TME). GBM are characterized by intratumoral and intertumoral heterogeneity, highly invasive and infiltrative cell properties and an immunosuppressive TME promoting GBM growth via complex interactions [11].

Immunotherapies, by re-educating and harnessing the patient’s immune response against tumors, hold great promise for cancer treatment. These methods have become increasingly used in the treatment of different kinds of cancers, including brain cancers [12, 13]. Current immunotherapy strategies used to treat cancers are mainly based on immune checkpoint blockade (ICB) agents [14, 15], but also therapeutic vaccines [16, 17], adoptive cell therapy [18, 19], monoclonal antibodies (mAbs) [20] and oncolytic viruses [21].

Treatment with ICBs has shown remarkable success in a population of patients with melanomas and other tumors [22‐25]. Therapeutic vaccines have also emerged as promising cancer treatments, with currently 3 therapeutic cancer vaccines approved by the Food and Drug Administration (FDA) [26]. However, even though immunotherapies have shown survival benefit for some proportions of patients with solid tumors, most patients still do not respond to immunotherapy. Less than 15% of cancer patients currently respond to ICBs [27]. Furthermore, these strategies are not as effective as would be desired for GBM treatment. To date, phase III clinical studies with ICB and vaccine therapies have shown no major benefit of immune modulation for GBM treatment or patient survival [13, 28].

A recent clinical study, however, demonstrated that administration of anti-programmed cell death protein 1 (PD-1) mAbs prior to tumor resection increased local and systemic antitumor immune responses [29]. Additionally, interim results of a phase II clinical study evaluating the combination of an allogeneic/autologous therapeutic GBM vaccine in combination with granulocyte-macrophage colony-stimulating factor (GM-CSF), cyclophosphamide and bevacizumab demonstrated a significant survival benefit [30]. These results raise hope for research on GBM immunotherapy treatments. Immune modulation in combination with other treatments has shown encouraging preclinical results. This review summarizes some of these promising combination strategies for the treatment of GBM. We will particularly focus on combinations including ICBs, as they are the most studied combination strategies including immunotherapy for GBM.

To ensure proper neuronal function, the brain has to be maintained in a homeostatic state. This implies regulation of the influx/efflux of cells, molecules and ions [31]. Two major barriers contribute to separating the CNS from the variable environment of blood: (i) the blood−cerebrospinal fluid barrier, formed by the choroid plexus epithelium and separating the cerebrospinal fluid from the blood; and (ii) the blood−brain barrier (BBB), formed by endothelial cells of the capillary of brain parenchyma and separating the blood from the brain interstitial fluid [32, 33].

Due to its anatomical structure and vascular organization, the BBB is the most selective barrier [34]. For this reason, the BBB is also an important obstacle for the development of successful GBM treatments. While it has been shown to be disrupted in GBM, an intact BBB is still found peritumorally [35, 36]. This heterogeneous disruption leads to protection of most infiltrative components of GBM and limits the delivery of the majority of therapeutics to the tumor [35‐37].

Additionally, due to the presence of the BBB and its tight junctions, the brain has long been considered an immune-privileged site. The identification of an absence of classic lymphatics and tolerance of foreign tissue transplants in the brain also suggested that the brain was immunologically unique [38‐40]. However, it is now known that the brain is not isolated immunologically [41]. It has been proven that functional lymphatic vessels are present in the CNS, that activated T cells can traffic to the CNS and that CNS antigens can reach the peripheral lymph nodes [38]. This emerging evidence suggests that immunotherapy can be applied to GBM and other brain cancers. However, several characteristics of GBM, e.g., its heterogeneity, BBB, low tumor mutation burden, low infiltration of T cells and microenvironment (which, for example, features a high infiltration of immunosuppressive cells) induce very complex immunosuppression, which is one major hindrance in finding new treatments and translating immunotherapies for GBM (Fig. 1).

GBM tumors develop in an immunosuppressive microenvironment that stimulates tumor cell growth and aggressiveness. The GBM TME is composed of various cell types: infiltrating tumor cells and cancer stem cells as well as noncancerous cells, such as myeloid cells (including resident microglia and bone marrow-derived macrophages), tissue-resident cells (such as neurons and astrocytes), and lymphocytes, and all these cells can interact together [42, 43].

GBM cells are known to secrete chemokines, growth factors and cytokines into the TME. Liberation of these molecules will attract and stimulate immunosuppressive cells [43‐45]. In turn, immunosuppressive cells interact with GBM cells through different immunosuppressive receptors, resulting in tumor growth promotion and tumor cell resistance and evasion of immune surveillance [43, 44, 46, 47]. One of the most well-known and described immunosuppressive receptors involved in GBM immune escape is PD-1 [48‐50]. The PD-1 receptor is mostly expressed on activated T cells, and binding of PD-1 with its ligand PD-L1 leads to inactivation of those T cells [51]. In GBM, PD-L1 has been shown to be overexpressed by GBM and myeloid cells, leading to effective binding of PD-L1 to PD-1 and therefore inhibition of the immune response [48, 52, 53].

The predominant non-neoplastic cells are tumor-associated microglia and macrophages (TAMs), which constitute approximately 30% of GBM TME [54, 55]. Microglia and macrophages are the main innate immune cells in the healthy CNS, where they play a major role in maintaining homeostasis and immune surveillance [56, 57]. Microglia are the resident macrophages of the CNS localized in the parenchyma. They arise from yolk-sac macrophage precursors at early embryonic stage [58‐61]. Three non-parenchymal macrophage populations, or border-associated macrophages, are also found in the CNS under normal conditions: perivascular, meningeal and choroid macrophages [56, 57, 62]. In GBM, the partial disruption of the BBB leads to peripheral bone marrow-derived macrophages infiltration. This macrophage population account for approximately 85% of GBM TAMs [63].

Two TAM phenotypes are commonly described: (i) the classical inflammatory and anti-tumoral M1 phenotype and (ii) the anti-inflammatory and pro-tumoral M2 phenotype [64]. Recent studies also indicate that heterogeneous populations of TAMs expressing both M1- and M2-associated markers are found in human and murine GBM [65, 66]. A high infiltration of M2-like TAMs has been associated with poor prognosis in GBM [55, 67]. Moreover, a higher proportion of M2-like TAMs are found in high grade gliomas, as compared to low grade gliomas [68]. Indeed, GBM cells recruit microglia and macrophages in the TME by means of chemoattractants (e.g., C-C motif chemokine ligand 2 (CCL2), CSF-1) [69, 70] and induce a switch to pro-tumoral subtype [65, 71, 72]. In return, TAMs promote tumor cell proliferation and angiogenesis as well as inhibition of effector T cells proliferation and attraction of T regulatory cells (Tregs) and MDSCs through cytokines secretion [65, 72, 73].

MDSCs constitute a heterogeneous population of cells that play an important role in maintaining an immunosuppressive environment in GBM. They inhibit the immune response by interacting with different cells in the TME, promoting Treg function, limiting antigen presentation and inhibiting effector T cell activity, among other things [74‐77]. In addition, as already mentioned, myeloid cells (including TAMs and MDSCs) have been shown to overexpress the negative checkpoint molecule PD-L1 in GBM patients, therefore promoting negative regulation of the immune response by inducing T cell dysfunction [78].

Compared to other tumors, the number of tumor-infiltrating lymphocytes (TILs) is relatively low in GBM [79]. Additionally, those that are present express high levels of exhaustion markers such as the inhibitory coreceptors T cell immunoglobulin and mucin domain containing-3 (TIM-3), lymphocyte activation gene 3 protein (LAG-3), and PD-1 [80‐83]. Moreover, the fraction of Tregs among TILs is increased in GBM patients [84]. Tregs contribute to the TME immunosuppression by inhibiting effector T cells and antigen-presenting cells (APCs) [83]. Beside a low density of TILs, a low number of infiltrating Natural killer (NK) cells and B cells are found in GBM [85]. Besides, GBM cells have developed mechanisms to escape from NK immune surveillance through inhibitory bindings [86].

In addition to the complex interplay between all the cells present in the TME of GBM, resulting in multifactorial and complex immunosuppression, low tumor mutation burden is an additional likely reason why applying immunotherapy in GBM is difficult. Indeed, tumor-associated antigen production is quite low in GBM, which may lower the efficacy of immunotherapy [87].

Moreover, unique characteristics of GBM also impede the development of new treatments. Indeed, GBM is a very infiltrative tumor that exhibit inter- and intratumoral heterogeneity. Intratumoral heterogeneity refers to molecular diversity within a same tumor, leading to differences in growth rate, cellular morphology, histopathology and differentiation markers expression, among other things [88]. Therefore, different subpopulations of GBM cells coexist and cooperate to promote tumor growth and progression, contributing to divergence in response and resistance to SOC [89, 90]. Intertumoral heterogeneity refers to cellular and genetic differences between GBM tumors from different patients and leads therefore to different molecular subtypes, and heterogeneity in patients’ response to therapy [91].

Immune checkpoints are regulators of the immune system that control immune effector function by maintaining an equilibrium between inhibitory and costimulatory signals. Their role is to protect tissues from damage due to excessive immune response but also to prevent autoimmunity [14, 92]. However, cancer cells have been shown to take advantage of this system to escape immune surveillance by upregulating inhibitory immune checkpoint expression and activating these negative regulators on tumor-specific immune cells [14, 93, 94]. The main example of this mechanism in GBM is the appropriation of the PD-1/PD-L1 pathway [48‐50].

The main immune checkpoints that have been successfully targeted in cancer treatments are PD-1, its ligand PD-L1, and cytotoxic T-lymphocyte antigen 4 (CTLA-4). Inhibition of these T cell negative regulators with mAbs increases the immune response in many cancers (e.g., melanoma and non-small-cell lung cancer) [25, 95‐97]. In 2011, the FDA for the first time approved the use of the checkpoint inhibitor ipilimumab – a monoclonal antibody inhibiting CTLA-4 – as a frontline cancer treatment in advanced melanoma [22, 98]. Since 2014, the first mAb directed toward PD-1, nivolumab, has been FDA-approved for the treatment of metastatic melanoma [99]. Nivolumab was further approved for other cancers, such as non-small-lung cancer and squamous cancer of the head and neck [23, 24]. The combination of ipilimumab and nivolumab was also approved for advanced renal cell carcinoma, metastatic melanoma and metastatic colorectal cancer [100‐102].

By generating and amplifying specific T cell responses against tumors, therapeutic cancer vaccines hold a major place among strategies to fight cancer. In the past decade, research on therapeutic cancer vaccine has extended, thanks to recent progress in delivery technologies and target selection methods as well as continuous advances in understanding tumor immune response mechanisms [103]. Today, there are 3 FDA-approved therapeutic cancer vaccines. The first immunotherapy ever approved by FDA was intravesical BCG (bacillus Calmette-Guérin) in 1990 [26]. This live attenuated vaccine is now part of the standard treatment of early-stage bladder cancer and reduces risks of cancer progression [104, 105]. In 2010, the FDA approved Sipuleucel-T (Provenge®) – an autologous dendritic cell (DC) vaccine that reduces the risk of death in prostate cancer patients [106]. Intralesional vaccination with an oncolytic herpes virus encoding GM-CSF, talimogene laherparepvec (T-VEC; Imlygic®), improved mOS in patients with advanced melanoma [107]. T-VEC was approved by the FDA in 2015 [108].

However, both ICBs and vaccines as monotherapies are still ineffective in multiple cancers, such as GBM. To date, there is no FDA-approved immunotherapy for GBM, even though some are being tested in clinical trials [109]. Moreover, some phase III clinical trials have failed, including trials of immunotherapy treatments such as ICBs or therapeutic vaccines (Table 1).

ICBs, and anti-PD-1 therapy more particularly, have been extensively studied in GBM treatment given their promising results in other solid tumors. Anti-PD-1 agents were the first type of ICB tested in a clinical trial for GBM treatment (CheckMate 143, NCT02017717) [110, 113]. Phase III of this study was conducted on 369 patients with recurrent GBM randomized to receive either nivolumab (anti-PD-1 antibody) or bevacizumab (an antibody targeting and inhibiting vascular endothelial growth factor (VEGF); an antiangiogenic treatment). The results of this study showed no improvement in the median overall survival (mOS) for patients treated with anti-PD-1 antibody compared to those in the other treatment arms [110]. Similarly, the phase III CheckMate 498 trial (NCT02617589) failed to meet its primary endpoint of improving mOS. The study was conducted on 560 newly diagnosed patients and compared nivolumab and TMZ, each of which was given with RT. More recently, Bristol-Myers Squibb (BMS) announced that the phase III trial Checkmate 548 (NCT02667587) did not meet its primary endpoint. The study evaluated the addition of nivolumab to the SOC and was conducted on 693 patients with newly diagnosed GBM.

Likewise, two phase III studies with therapeutic vaccines added to standard treatments failed to produce convincing results [111‐113]. In the ACT IV phase III study (NCT01480479), the addition of rindopepimut – a vaccine targeting EGFRvIII – to standard chemotherapy did not improve survival of EGFRvIII-positive glioblastoma patients [111]. The NCT00045968 phase III trial evaluating the addition of an autologous tumor lysate-pulsed DC vaccine (DCVax®-L) to SOC for newly diagnosed GBM patients was put on hold by the FDA for unidentified reasons [112, 113].

The main commonality between the agents used in these disappointing phase III trials is that they stimulate only one arm of antitumor immunity: they reduce the immunosuppression exerted on T cells (ICBs), stimulate the immune response against a specific antigen or activate a specific dendritic cell (DC) response against the tumor. Even when combined with SOC, the immune response is not enhanced, mainly because of the immunosuppressive characteristics of systemic chemotherapy and RT (which induce lymphopenia and hypoxia, respectively), as well as corticotherapy [114, 115]. Indeed, the blood of patients treated with dexamethasone has a reduced number of immune cells [116].

Despite these disappointing clinical trial results, the randomized open-label pilot study from Cloughesy and colleagues raises hope for the use of immunotherapy in GBM [29]. In this study, an anti-PD-1 mAb (pembrolizumab) was given as a neoadjuvant drug to patients with recurrent GBM. Administration of pembrolizumab before resection significantly improved overall and progression-free survival with induction of TIL functional activation and production of an interferon (IFN)-γ response within the TME [29]. Besides, a phase II study on the therapeutic GBM vaccine ERC1671 (Gliovac™) showed promising preliminary results. ERC1671 is a vaccine composed of tumor cell lysate from allogeneic and autologous GBM patients mixed with primary irradiated/inactivated whole tumor cells [30]. Combination of ERC1671 with GM-CSF and cyclophosphamide plus bevacizumab resulted in a significant survival benefit with a mOS of 12 months, while placebo plus bevacizumab mOS was 7.5 months. Interim results suggested that the benefit was correlated with the CD4 helper T lymphocytes counts in the peripheral blood [30].

The use of immunotherapy for GBM is therefore not a dead end. According to us, combination strategies targeting different arms of the cancer immunity cycle have great potential to overcome GBM multifactorial immunosuppression and increase antitumor immune response. These strategies are being tested to a large extent in preclinical and clinical studies, giving encouraging results.

The immune cycle in GBM is a multistep process that can be targeted by treatments at different levels (Fig. 2). Antigens released from dying GBM cells are captured by APCs, processed and displayed on major histocompatibility complex (MHC)-I and -II molecules for presentation to T cells. Effector T cells are primed and activated in response to tumor antigen presentation. Activated T cells kill GBM cells after binding to GBM tumor antigen on MHC-I through the T cell receptor (TCR). However, the immunosuppressive microenvironment can hinder immune control in GBM.

ICBs act by restoring the activity of effector cells so that they can recognize and attack cancer cells again. However, if there is not enough T cell infiltration at the tumor site and/or too many immunosuppressive cells in the TME (e.g., Tregs and TAMs), ICBs may not be enough to promote a strong antitumor response inducing the death of cancer cells. Targeting different arms of the cancer immunity cycle is of great interest, particularly in the context of GBM, in which it could offer better chances to overcome multifactorial immunosuppression [117, 118]. Indeed, it is possible to stimulate the immune system at different times and locations during immune response generation. One possibility to enhance the immune response against GBM (or other cancers) is by stimulating antigen release from dying tumor cells and their presentation to APCs. Immunogenic cell death (ICD) inducers, such as RT and some chemotherapeutics, induce dying cancer cells to release danger signals that stimulate the recruitment of APCs and antigen presentation [117]. Antigen presentation may also be stimulated by using vaccines (e.g., whole-cell tumor vaccines or peptide vaccines) that enhance the recognition of antigens and increase the production of antigen-specific T cells [118]. Another way to improve the immune response against GBM is to decrease the immunosuppression occurring at different stages of the generation of effector T cells. The use of ICBs is helpful to block the inhibition of effector T cells and therefore restore their activity (either early in the course of their activation or in the TME) [119]. Costimulatory agonist mAbs might also be of great interest to potentiate effector T cell function [120] (Fig. 2).

In this chapter, we describe some of the combination strategies that are used in preclinical studies. Most of the described results rely on GL261(−Luc) models, as they are the most common models used nowadays for studying the immunotherapy of GBM. We focus first on combination of chemotherapy with immunotherapy and then on combination of different immunotherapies. ICBs are the most studied immunotherapeutic approach; for this reason, the majority of the combinations presented hereafter includes inhibitory checkpoint molecules. However, we also describe other strategies, including vaccines and agonist costimulatory checkpoints.

Many immunotherapy strategies that have shown successful results in preclinical studies have failed to produce convincing results in clinical trials, revealing limitations and inadequacies of current GBM preclinical models [110, 111, 172]. Translational impact may be improved by developing new relevant preclinical models. In the literature, murine GL261 GBM cell lines are the most commonly used. However, they are highly immunogenic unlike human GBM [172, 173]. Additionally, luciferase-expressing GL261 cells display even more immunogenic features than GL261, with a prolonged median survival time and an elevated inflammatory cytokines production [174]. Transitioning to other models that are less immunogenic and closer to human GBM is highly encouraged [175]. To that purpose, the recently developed SB28 and 005 GSC models are interesting alternatives and are among the best syngeneic models to represent human GBM TME [147, 171, 176‐178]. Indeed, they are poorly immunogenic, with absence of MHC molecules expression as well as low immune cell infiltration and activation [177‐179]. These models are moreover resistant to ICBs [147, 178].

An alternative to syngeneic immunocompetent mouse models is the use of genetically engineered mouse models (GEMs). GEMs of GBM reproduce more closely the histology and biology as well as the development of human GBM [172, 180]. Another advantage of these models is that they usually do not require intracranial injections as they are generated through genetic modifications [180]. However, GEMs still do not reflect GBM heterogeneity. Moreover, setting up GEMs requires a lot of expertise and can be expensive [172].

Using patient-derived xenograft models is the best approach to gather human GBM histology and heterogeneity as well as intratumoral heterogeneity [172]. However, these kinds of models generally require immunodeficient mice, which are not suitable for immunotherapy studies [181]. Yet, an immunocompetent mouse model of human GBM was recently developed. By transiently blocking T cell costimulation, the researchers managed to grow human GBM xenograft while keeping intact the mouse immune system [182]. This could be a step forward in the preclinical research for immunotherapy against GBM.

However, the perfect model imitating exactly all characteristics of human GBM (e.g., intratumoral heterogeneity, invasive properties, low immunogenicity, resistance to radio- and chemotherapy) doesn’t exist [172, 175]. It is therefore crucial to design the experiment and select the models according to their unique characteristics. In addition, studies on different models will help to develop a more general therapy against GBM by (i) covering multiple aspects of GBM immunology and (ii) addressing GBM heterogeneity (intratumoral and interpatient) [175].

Nevertheless, to improve immunotherapy for GBM, novel strategies need to be studied. Looking at the development of novel therapeutic targets (e.g., STING, Toll-like receptor (TLR) agonists or old combination therapies applied to new targets) is important to assess the utility of targeting underexplored pathways useful for improving combination immunotherapy [183‐187]. In pursuit of this goal, targeting immunosuppressive factors and/or cells within GBM TME is highly valuable [39]. To do so, targeting angiogenesis signaling factors (e.g., VEGF, Ang-2) [146] and adenosinergic pathways components (e.g., adenosine receptors, CD73, CD39) [160, 165, 173, 188] has shown great potential in preclinical studies. Moreover, targeting TAMs are particularly interesting since these cells are the major non-neoplastic cellular components of GBM TME. To this end, many strategies are tested in preclinical studies such as (i) depleting TAMs using CSF-1R inhibitor [189], (ii) activating TAMs by using agonist anti-CD40 mAbs [148, 190], (iii) reprogramming TAMs to induce pro-inflammatory and antitumor immune response by using TLR agonists [186], (iv) promoting TAMs phagocytic activity by using anti-CD47 mAbs [130, 191] and (v) inhibiting TAMs recruitment by targeting the C-C motif chemokine ligand 2/C-C receptor 2 (CCL2/CCR2) axis [192]. CCL2 chemokine production in GBM TME has also been shown to be essential for the recruitment of MDSCs and Tregs [193]. Blockade of CCL2 is therefore a promising approach to overcome GBM TME immunosuppression [194]. Therapies targeting myeloid cells are even more interesting to use in combination with ICBs knowing that they express high levels of PD-L1. This imply furthermore the great potential of targeting both innate and adaptive immunity for GBM treatment.

In addition, development of technologies that improve immunotherapeutic combination against different targets is necessary. As an example, bispecific antibodies targeting two different antigens has shown to be a valuable approach for combination strategies [195]. Antibodies hold a fundamental place in cancer immunotherapy strategies. However, the BBB limits the passage of most of the large molecules such as mAbs [196]. Many physicochemical properties determine the ability of a compound to pass the BBB (e.g., size, lipophilicity, molecular weight, degree of hydrogen bonding) and the delivery of many drugs across the BBB is consequently insufficient [198]. Therefore, improving delivery systems needs to be considered in order to increase delivery at the tumor site. In this regard, novel cyclic peptides modulating the BBB enhanced the brain delivery of mAbs [196]. Similarly, focused ultrasound-mediated BBB disruption improved anti-CD47 mAbs delivery to GBM tumors [199].

To improve local administration, the use of drug delivery systems enhancing brain penetration for intratumoral administration is highly valuable to improve drug distribution and sustained-release. It was demonstrated that densely PEGylated PLGA-based nanoparticles enhanced the penetration of paclitaxel in the brain tissues and therefore improved the treatment efficacy [200, 201]. Similarly, development of novel implantable biomaterials to improve administration in GBM resection cavity and prevent tumor recurrences are encouraged [202, 203]. Among other technologies, a thermoreversible biodegradable chitosan-based hydrogel containing therapeutic T cells showed encouraging results in GBM, offering an interesting platform for local immunotherapy [204]. Furthermore, nanocarriers have to be developed to enhance the local immune response, turning GBM “cold” tumors into “hot” tumors and therefore promoting the infiltration of immune cells. To do so, inducing ICD of GBM cell by using immunostimulant nanocarriers can facilitate the antitumor immune response and improve therapeutic effects [205, 206]. In addition, new administration routes need to be explored. In this regard, intranasal administration is providing promising results [207, 208]. Finally, scalability and clinical translation must be considered when developing such systems.

It was long thought that immunotherapy could not be applied to GBM (e.g., because of the BBB, multifactorial immunosuppressive TME, tumor heterogeneity, etc.). The use of ICBs as immunotherapy during the past decade has revolutionized cancer treatments given their ability to improve patient outcomes. However, ICBs for GBM are still ineffective, as proven by the recent phase III clinical trials that did not reach their primary endpoints when administered in monotherapy. Likewise, vaccines added to the SOC did not improve survival.

In contrast, preclinical studies on combination immunotherapy showed encouraging results. The most successful strategies in terms of survival and immune benefits are the ones that impact the cancer-immunity cycle at different time and/or locations, inducing both stimulation of the immune response and inhibition of immunosuppressive components. Combination of ICBs with chemotherapy, radiotherapy inducing ICD or vaccines have been extensively studied and demonstrated significant improvement over monotherapies. Besides, while immunotherapy for cancer is mainly focusing on factors regulating T cell activation, concomitant targeting of both innate and adaptive immunity holds great promise for GBM treatment. The simultaneous use of immunotherapeutic agents targeting different arms of the immune system is starting to be largely tested in GBM clinical trials, given the success of combination immunotherapy in other cancers and in GBM preclinical studies. There are currently nine ongoing clinical trials (phase I or II) studying the combination of multiple ICBs for GBM treatment [209]. The combination of ICBs with radiotherapy or vaccination is also being tested, with twelve and seven ongoing clinical trials, respectively [209].

We believe that currently ongoing clinical trials and preclinical research on combination strategies will provide key information and better survival without major side effects and that immunotherapies will be added to the SOC for patients facing GBM in the near future.