Glioblastoma Multiforme: An Overview of Emerging Therapeutic Targets

Glioblastoma multiforme (GBM) is the most common and aggressive malignant primary brain tumour in humans and has a very poor prognosis. The existing treatments have had limited success in increasing overall survival. Thus, identifying and understanding the key molecule(s) responsible for the malignant phenotype of GBM will yield new potential therapeutic targets. The treatment of brain tumours faces unique challenges, including the presence of the blood brain barrier (BBB), which limits the concentration of drugs that can reach the site of the tumour. Nevertheless, several promising treatments have been shown to cross the BBB and have shown promising pre-clinical results. This review will outline the status of several of these promising targeted therapies.


INTRODUCTION
Glioblastoma multiforme (GBM) is the most common and aggressive primary malignant brain tumour and accounts for 60% of brain tumours in adults (1). The global incidence of GBM is <10 per 100,000 persons (2) and has increased over the last decade (3). GBM patients have a poor prognosis with a 1-year survival rate of 37.2%, a 5-year survival rate of 5.1% (4) and a median survival of ∼10 months (5). GBM is divided into three subgroups based on isocitrate dehydrogenase 1 (IDH1) and IDH2 mutation status: IDH-mutant, IDH-wild-type and NOS (not otherwise specified) (6)(7)(8). However, despite this classification, the majority of GBM patients receive identical treatments, and few targeted therapies currently exist, contributing to the poor outcomes typically experienced by GBM patients. This review will outline the current treatment options for GBM and discuss some of the more recent developments in targeted therapies being investigated for the treatment of GBM.

CURRENT TREATMENT OPTIONS
The treatment of brain tumours faces unique challenges, most notably the presence of the blood brain barrier (BBB), a highly selective semipermeable barrier that separates blood from the brain. The BBB is comprised of the endothelial cells of capillaries, astrocytes surrounding the capillary, and pericytes embedded in the capillary basal lamina. Physiochemical properties including molecular weight, lipophilicity and charge affect the ability of a molecule to cross the BBB (9). The BBB prevents nearly all large molecules (>400 Da) (10) and ∼98% of small molecule drugs from entering the central nervous system (CNS) (11). The current treatment pipeline begins with surgical resection of the tumour, if applicable and safe to do so, followed by radiotherapy and concomitant chemotherapy (12,13).
The initial therapeutic approach for GBM is surgery, where maximal resection is associated with longer progression-free survival (PFS) and overall survival (OS) (14). Resection is not a curative approach; hence, patients typically undergo radiotherapy and chemotherapy as an adjunct (15). Radiotherapy, at a total dose of 60 Gy, is administered as either primary treatment or following surgery (16), both resulting in improvements to PFS and OS. Concomitant administration of temozolomide [150-200 mg/m 2 /day for 5 days each 28-days cycle (12,13)], an oral alkylating agent, significantly increases OS in patients with newly diagnosed GBM from 12.1 months with radiotherapy alone to 14.6 months with radiotherapy and temozolomide (13). However, despite this increase in survival with radiotherapy and temozolomide, tumour progression and recurrence typically occur (17,18), due to the development of resistance to temozolomide (19,20). Once GBM recurrence occurs, therapeutic options for patients are limited (21,22). Recently, tumour treating fields (TTFields; Optune), which deliver electric fields to the tumour location to disrupt cancer cell division, have emerged as an FDA-approved treatment for both recurrent and newly diagnosed GBM (23). However, the identification of new targets to facilitate the development of novel targeted therapies is warranted.

EMERGING TARGETED THERAPIES
GBM is an invasive tumour with hallmarks of neoangiogenesis and intratumour heterogeneity, contributing to the poor prognosis observed (24). A variety of genetic and epigenetic alterations have been identified in GBM that influence patient prognosis ( Table 1). Despite this heterogeneity, a large-scale analysis of genetic aberrations in GBM identified three main signalling pathways that are commonly dysregulated: activation of the receptor tyrosine kinase (RTK)/Ras/phosphoinositide 3kinase (PI3K) pathway (88%), inhibition of p53 (87%), and retinoblastoma protein (Rb) signalling pathways (78%) (34). Drugs targeting many of these commonly observed alterations have been investigated as potential targeted therapies for GBM.

EphA3 Receptor Inhibitors
The EphA3 receptor is overexpressed in 40-60% of GBM tumours and is commonly overexpressed in recurrent GBM ( Table 1) (25). EphA3 is highly expressed on the tumourinitiating cell population in glioma, maintains tumour cells in a less differentiated and stem cell-like state, and its expression mediates the tumourigenic potential in GBM cells in vitro (26), suggesting that EphA3 may be a potential target for the treatment of GBM ( Table 2).
A small molecule inhibitor of the EphA3 receptor, GLPG1790, has demonstrated superior tumour reduction in U251MG and U87MG subcutaneous xenograft models when compared to radiotherapy alone, however, GLPG1790 was not as effective as treatment with radiotherapy and concomitant temozolomide (54). Whilst GLPG1790 did not exhibit improved benefit over the current therapies, additional strategies for targeting EphA3 are being examined. An EphA3 monoclonal antibody (IIIA4) that binds the EphA3 globular ephrin-binding domain has been developed, and the humanised version (ifabotuzumab) is the subject of an investigator-sponsored Phase 0/1 clinical trial currently underway in patients with recurrent GBM (55) to identify the optimal dose for tumour penetration. The IIIA4 antibody conjugated to the cytotoxic microtubule-targeting agent maytansine (IIIA4-USAN), induced apoptosis in four primary GBM cell lines in vitro, and significantly increased survival in an orthotopic model in vivo (55). Providing further evidence for the potential suitability of monoclonal antibodies targeting EphA3, a bispecific antibody against EphA2/A3 reduced clonogenicity in vitro and decreased tumour burden in vivo (56). Taken together, these studies indicate that EphA3 receptor inhibitors may be promising treatments for EphA3 receptor-amplified GBM, including recurrent disease, however, this remains to be tested in the clinic.

EGFR Inhibitors
Epidermal Growth Factor Receptor (EGFR) amplifications and mutations are detected in 40-60% of GBM cases (28,96) (Table 1) and are generally indicative of poor prognosis (97). EGFR (also referred to as ERBB1 or HER1) is a member of the HER superfamily of RTKs, along with ERBB2, ERBB3, and ERBB4. Binding of a ligand to the ligandbinding site of these receptors induces receptor homo-or heterodimerisation, producing a conformational change that activates the intracellular tyrosine kinase domain. This results in autophosphorylation of the cytoplasmic tail and induces a variety of downstream signalling pathways. The overexpression or mutation of EGFR leads to downstream signalling that impairs apoptosis, enhances proliferation, and angiogenesis. The most common mutant form found in GBM is EGFR (EGFRvIII, or de2-7EGFR), arising through an 801 base pair in-frame deletion from the extracellular domain (98). Due to the high incidence of EGFR amplifications, a variety of EGFR inhibitors have been examined both pre-clinically and clinically ( Table 2).

Small Molecule Inhibitors
Small molecule tyrosine kinase inhibitors are the most widely studied EGFR inhibitors in GBM, and include erlotinib, gefitinib, and lapatinib. Erlotinib inhibits anchorage-independent growth of GBM cells in vitro in an EGFR expression-dependent manner and induces greater levels of apoptosis in more malignant GBM phenotypes (57). The tumour-initiating cell population, which is resistant to radiotherapy (99), is sensitive to erlotinib in a phosphatase and tensin homolog (PTEN) and Akt dependent manner (59), suggesting that erlotinib may eliminate this population in vivo. Further, treatment with erlotinib was shown to reduce tumour burden in two GBM patient-derived xenograft (PDX) models (60). However, further studies using additional GBM PDX models demonstrated that tumours overexpressing EGFR were only sensitive to erlotinib if they also expressed PTEN (61). As PTEN expression is downregulated in ∼34% of GBM patients (29), this indicates that erlotinib may not be a suitable treatment for the majority of GBM patients overexpressing EGFR. Indeed, erlotinib was not effective as a monotherapy in recurrent GBM patients and was only marginally beneficial following radiotherapy for non-progressive GBM patients (62). Despite a limited number of complete and partial responses in a Phase II study in first-relapse GBM, the 6-months PFS and median survival was similar to that previously reported for patients undergoing chemotherapy (64), suggesting that erlotinib may be useful in this setting. However, due to the non-randomised nature of this trial, these results must be interpreted cautiously. By contrast, improved survival (19.3 vs. 14.1 months) was observed when combined with temozolomide and radiotherapy (63), suggesting that erlotinib may be beneficial when combined with other treatments, rather than as a monotherapy. In contrast to erlotinib, gefitinib exhibits anti-tumour activity independent of the expression level of EGFR (100). Gefitinib inhibits GBM cell migration (65), reduces proliferation of human glioma tumour-initiating cells in vitro (59) and enhances survival in an intracranial GBM mouse xenograft model in vivo (58). Taken together, these pre-clinical studies indicate that gefitinib may be clinically beneficial. However, despite gefitinib reaching high concentrations in GBM tumour tissue (22-fold higher compared to plasma) and the significant dephosphorylation of EGFR achieved (66), limited clinical effects have been observed in Phase II trials. Several Phase I/II studies have demonstrated that whilst the addition of gefitinib to radiotherapy is well-tolerated, it has no survival benefit (67,68).

Monoclonal Antibodies
Although tumour immunotherapy has shown some success for the treatment of melanoma and haematological cancers, the applicability to GBM presents more of a challenge. Monoclonal antibodies directed against wild-type EGFR and EGFR have been developed, with the best characterised in GBM being cetuximab. Pre-clinical studies have shown that treatment with cetuximab alone and in combination with radiotherapy increases survival in vivo (69) and can also completely eliminate tumours in EGFR-amplified PDX models (70). A phase II trial examining cetuximab treatment in patients with recurrent high-grade glioma showed that cetuximab was well-tolerated, but exhibited limited activity in this patient population (71).

VEGF Inhibitors
With the largely disappointing clinical results for EGFR inhibitors, additional targets are being investigated, including vascular endothelial growth factor (VEGF), which is highly expressed in glioma cells. High VEGF expression is directly associated with the poor prognosis and malignancy of gliomas (51)(52)(53)101). VEGF is a dimeric polypeptide that binds to the VEGF receptors 1 (VEGFR-1) and VEGFR-2, and the coreceptors neuropilin 1 and 2. Following this interaction, VEGF mediates angiogenesis and cell proliferation. Under hypoxic conditions, hypoxia-inducible transcription factors translocate to the nucleus which activate VEGF leading to increased angiogenesis in an attempt to counteract hypoxia (102). GBM tumours are commonly hypoxic and have increased VEGF expression that contributes to the irregular vasculature in GBM, prompting the investigation of VEGF as a potential therapeutic target ( Table 2).

Small Molecule Inhibitors
Several VEGF inhibitors have been examined for the treatment of GBM, including the small molecule inhibitors, tivozanib, and

Metaanalysis
Four clinical trials (n = 607) No difference in overall survival, modest increase in progression-free survival when combined with chemotherapy, compared with bevacizumab or chemotherapy alone. Higher incidence of treatment-related adverse events in bevacizumab treated patients.  In vivo Mayo39 and Mayo59 subcutaneous xenografts−40 mg/kg crizotinib daily by gavage, 100 mg/kg erlotinib daily by gavage, or crizotinib + erlotinib Reduced tumour burden and vascular density (when given in combination with erlotinib) (60) pazopanib. Phase II studies of tivozanib (77) and pazopanib (78) in recurrent glioblastoma showed that these inhibitors exhibited limited anti-tumour activity and did not prolong PFS in this patient population. These trials highlight the limitations of anti-VEGF monotherapy.

Monoclonal Antibodies
Bevacizumab, a humanised monoclonal antibody against VEGF, blocks angiogenesis and thereby reduces tumour growth in a variety of GBM mouse models as a monotherapy and when combined with radiotherapy (72)(73)(74)(75). These promising preclinical studies led to the clinical investigation and subsequent approval of bevacizumab for the treatment of recurrent GBM (103). However, a meta-analysis of four clinical trials including 607 patients demonstrated that the addition of bevacizumab to standard chemo-radiotherapy in the upfront setting only improves PFS, with no improvement in OS, but with an increase in the number of treatment-related adverse events (76). Of note, a decline in neurocognitive function is more frequently observed following bevacizumab treatment, as bevacizumab impairs hippocampal synaptic plasticity and decreases dendritic spine number and length (104). The modest treatment responses combined with the increased treatment-related adverse events raises concerns about the suitability of the use of bevacizumab as a treatment for GBM.
A variety of multi-RTKs inhibitors have been examined both pre-clinically and clinically ( Table 2).
Imatinib is a small molecule that inhibits PDGFRA and PDGFRB, as well as the RTKs c-Abl and c-Kit, and is a radiation-sensitising agent for glioma cells in vitro (79) and in orthotopic GBM models in vivo (80,81). These promising preclinical studies led to the initiation of clinical trials. However, whilst imatinib was well-tolerated in recurrent GBM patients, it exhibited limited anti-tumour activity (82). Subsequent Phase III trials have examined imatinib in combination with hydroxyurea, rather than as a monotherapy (83). This trial showed that there was no PFS benefit to the addition of imatinib to hydroxyurea, or hydroxyurea alone. Further, a recent study has demonstrated that imatinib treatment can increase GBM cell migration and invasion in vitro (108), providing further potential insight as to why imatinib has failed in clinical trials for GBM.
Following the failure of imatinib in clinical trials, additional multi-RTK inhibitors have been studied. Sunitinib is an oral, small molecule that inhibits PDFR and VEGFR, and thereby reduces vascularisation and triggers apoptosis to produce tumour reduction. In pre-clinical models, sunitinib treatment induced apoptosis in vitro, and improved survival in an intracerebral GBM mouse model (84). Further, sunitinib treatment delayed tumour growth and increased survival in a PDGFF-driven mouse model, both as a monotherapy and in combination with low dose radiotherapy (85). By contrast, a Phase II trial and systematic review of the literature indicated that compared to conventional chemotherapy or bevacizumab, sunitinib has limited clinical activity in recurrent GBM as a monotherapy (86) or when combined with temozolomide or radiotherapy and temozolomide as a first-line treatment of patients with GBM (87).

PI3K Pathway Inhibitors
Genetic aberrations in GBM, including EGFR, PDGFRA, PTEN, TP53, and PIK3CA, drive the dysfunction of signalling pathways, including PI3K/Akt/mTOR, p53, and Rb1 (34). The PI3K/Akt signalling pathway is activated in most GBMs and plays a critical role in the regulation of signal transduction, and mediates a variety of cellular processes, including proliferation, survival, migration and angiogenesis in GBM. The PI3K/Akt pathway is typically initiated via the activation of RTKs or G proteincoupled receptors, were the conformational changes in the C-terminal kinase domain produced by autophosphorylation provides binding sites for the regulatory subunits of PI3K, which leads to elevated lipid kinase activity of PI3K, and activation of Akt. Despite the limited clinical efficacy of the previously described RTK inhibitors, as activation of each of these receptors leads to downstream activation of the PI3K/Akt pathway, it has therefore been suggested that PI3K pathway inhibitors may be beneficial in GBM. More than 50 PI3K inhibitors have been designed and are under investigation as treatments for a range of cancers. Several PI3K inhibitors have demonstrated pre-clinical efficacy in GBM ( Table 2) and have entered into clinical trials for GBM treatment.
Buparlisib, a pan PI3K inhibitor, reduces GBM cell growth both in vitro and in vivo (88,89). Buparlisib is the most frequently used PI3K inhibitor in clinical trials for GBM treatment, as it is well-tolerated and BBB permeable. However, single agent efficacy in Phase II trials in recurrent glioblastoma has been minimal (90). The lack of clinical efficacy was explained by incomplete blockade of the PI3K pathway in the tumour tissue. Whilst buparlisib showed minimal single-agent efficacy, the study of other PI3K inhibitors that achieve more-complete pathway inhibition may still be warranted.
Sonolisib is an irreversible wortmannin analogue that demonstrates a more persistent inhibitor effect on PI3K than wortmannin. Sonolisib inhibits invasion and angiogenesis in GBM cell lines in vitro and extends survival benefit in orthotopic xenograft models in vivo (91,92). Despite these promising preclinical results, the response rate to sonolisib in a Phase II study in patients with recurrent GBM was low, and the study did not meet its primary endpoint (109).

HGFR/MET Inhibitors
Brain tumours secrete scatter factor (SF)/hepatocyte growth factor (HGF), the activating ligand for HGFR/MET (50), which has been associated with poor prognosis for GBM patients ( Table 1) (110,111). HGF is overexpressed in 1.6-4% of GBM patients, and via activation of MET, enhances tumour growth and angiogenesis (28). The MET proto-oncogene encodes for MET, an RTK that is overexpressed in 4-6% of GBM patients. A mutated, constitutively active variant of MET, MET 7−8 , has been identified in 6% of patients with high grade gliomas, including GBM, enhancing downstream signalling to promote tumour progression and angiogenesis (46). SGX-523 is a small molecule inhibitor of HGFR/MET tyrosine kinase activity that inhibits tumour cell growth, migration and invasion in a panel of glioma cells in vitro and reduced tumour growth in a murine xenograft model of GBM using the U87MG GBM cell line ( Table 2) (93). Additionally, amuvatinib (MP470) is a small molecule inhibitor that acts on multiple tyrosine kinases, including MET, has been shown to radiosensitise GBM cell lines both in vitro and in vivo ( Table 2) (94).
Another small molecule inhibitor of MET kinase activity, crizotinib, inhibits the growth, sphere-forming capacity and expression of stem cell markers in a subcutaneous xenograft model of GBM using the U87MG cell line (95). However, in a subcutaneous xenograft model using Mayo39 and Mayo59 GBM cell lines, crizotinib was only effective at reducing tumour burden and vascular density when used in combination with the EGFR inhibitor erlotinib (60).

CONCLUDING REMARKS
GBM is an often-fatal disease and the standard treatment options available to patients are only minimally effective. There is a growing body of evidence suggesting that a personalised therapeutic approach for the stratification of GBM patients to novel treatment regimens is necessary if survival rates for GBM patients are to improve. Indeed, genetic profiling of GBM biopsies has revealed aberrant expression of several potential therapeutic targets, including a number of RTKs (EphA3, EGFR, VEGF, PDGFR, and MET), however, there has been varied and limited clinical success in the use of inhibitors of these targets as anti-cancer therapies. This highlights that a better understanding of the basic biology of GBM is required so that additional targets can be identified. Indeed, promising pre-clinical effects have been observed with EphA3 inhibitors, however, it remains to be seen whether this translates into the clinic. Whilst PI3K inhibitors have exhibited limited effects in clinical trials, they did not completely inhibit the PI3K pathway. Despite the promising outlook for personalised therapeutic approaches to treating GBM patients, the identification of therapeutics that can cross the BBB, whilst maintaining therapeutic concentrations, remains a challenge and is often not reported. Further, although targeted therapies show limited efficacy as single agents, the combination of several targeted therapies may be of benefit to GBM patients. Thus, additional research is urgently required to identify therapeutic targets in GBM and to design novel therapeutic strategies for the treatment of GBM.

AUTHOR CONTRIBUTIONS
OT and JB drafted the manuscript and reviewed the literature. OT, JB, and KS contributed to the preparation and editing of the manuscript. KS formulated the idea.