Department of Surgery
Jeremy N Rich
Department of Medicine and Department of Neurobiology
Duke University Medical Center
Primary nervous system tumours are among the most lethal cancers, with conventional therapy offering palliation only. The recent success of small-molecule inhibitors of signal transduction pathways in other cancers has propelled rapid development of similar therapies in the treatment of patients with glioblastomas. Several molecular targets are particularly attractive in glioblastomas, including growth factor receptors and critical intracellular signalling mediators, with specific inhibitors having been developed by pharmaceutical companies in recent years. Although these agents have shown promise in preclinical and clinical trials, significant work remains to maximise the utility of these treatments to improve patient outcomes. The major challenges facing the neuro-oncology community in the use of kinase inhibitors are:
- The identification of the optimal therapeutic target(s).
- The establishment of biomarkers of tumour sensitivity or resistance.
- The optimisation of signalling inhibitor combinations with one another or with cytotoxic therapies.
The role of tyrosine kinases in gliomas
Cancer cells display a significantly greater dependency on the effects of tyrosine kinases than nontransformed cells, a characteristic termed oncogene addiction,(1) which has prompted significant efforts in cancer drug development to block particular tyrosine kinases with high selectivity and specificity. Interest has recently been raised by the remarkable success of some small-molecule tyrosine kinase inhibitors (TKIs) in selectively inhibiting tumour growth. To date, the most prominent growth factor pathways in glioma biology have been the epidermal growth factor (EGF),(2–4) the platelet-derived growth factor (PDGF)(5,6) and the vascular endothelial growth factor (VEGF)(7,8) families. Each family contains multiple ligands and receptors that variably pair to activate intracellular signalling mediators. Upon ligand binding, receptors become dimerised and transphosphorylated, which promotes the recruitment and activation of intracellular signalling mediators, including mitogen-activated protein kinase (MAPK) and Akt/protein kinase B (PKB). Growth factor pathways regulate numerous cellular behaviours – proliferation, resistance to apoptosis, motility and elaboration of angiogenic factors – that act to increase tumour malignancy. Therefore, growth factor receptors are attractive glioma targets, as they satisfy many requirements for ideal therapeutic targets – frequent expression in tumours, large differential in expression levels between tumour and normal cells, contribution to the tumour phenotype and localisation to the cell surface.
Targeting growth factor pathways
EGF receptor pathway
Several small-molecule ATP mimetics that inhibit EGF receptor (EGFR)-associated kinase activity have been identified and fall into a number of different chemical classes. Some inhibitors have been characterised as reversible, whereas others are irreversible. The receptor specificity also varies among inhibitors – some show specificity for EGFR/HER1, whereas others show broader activity. To date, the therapeutic consequences of these differences remain to be elucidated. EGFR TKIs inhibit glioma cellular proliferation, increase sensitivity to radiation and chemotherapy, decrease matrix metalloproteinase expression and invasion, and decrease VEGF.(9–11) Therapeutic trials in other solid cancers have led to the FDA approval of two anti-EGFR agents, the monoclonal antibody cetuximab in colorectal cancer and the TKI gefitinib in lung cancer. All agents remain experimental in glioma therapy.
Gefitinib (ZD1839, Iressa; AstraZeneca) is a novel, oral, low-molecular-weight ATP mimetic of the anilinoquinazoline family that reversibly inhibits the tyrosine kinase activity associated with EGFR.(12,13) We recently completed an open-label, single-centre phase II trial of gefitinib (ZD1839) in recurrent glioblastomas,(14) in which gefitinib was well tolerated with mild efficacy (see Figure 1). The six-month event-free survival rate was 13% (seven out of 53 patients). This trial included a biopsy on all patients at recurrence with characterisation of EGFR and EGFRvIII expression measurements. We found that no EGFR expression patterns correlated with either treatment response or resistance.
In a second phase I/II study of gefitinib in patients with recurrent malignant gliomas,(15) the mean time to progression for glioblastoma patients was eight weeks, and 15 weeks for anaplastic gliomas. The six-month progression-free survival was 9% for glioblastoma patients and 33% for anaplastic glioma patients. A phase I/II trial of gefitinib in combination with external beam radiation for newly diagnosed glioblastoma patients has found little toxicity, limited to infrequent transaminase elevations.(16) Expansion to the phase II component is underway for patients who are not on enzyme-inducing antiepileptic drugs (AEDs). A phase II trial for patients with newly diagnosed glioblastomas did not show increased survival compared with historical controls, and no relationship to EGFR expression.(17)
Erlotinib (OSI-774, Tarceva; OSI Pharmaceuticals) is an orally active quinazoline derivative that inhibits EGFR-specific tyrosine phosphorylation and has demonstrated antitumour efficacy similar to that of gefitinib in preclinical studies.
In addition, erlotinib has shown activity in a phase I trial either as monotherapy or in combination with temozolomide in patients with glioblastomas or anaplastic gliomas.(18) No resections were performed. In pharmacokinetic studies, concurrent P450 enzyme- inducing AED use reduced both erlotinib and active metabolite levels by 50–75%. Of 25 evaluable malignant glioma patients, six had partial responses, two had minor responses and three had stable disease.
A phase II trial of erlotinib in patients with recurrent or progressive glioblastoma has shown that three out of 10 patients achieved partial responses; one patient achieved a mixed tumour response, and one had stable disease.(19) One response was durable, with duration of over seven months.
The toxicities of gefitinib and erlotinib in brain tumour patients are the same as in other trials – prominently rash and diarrhoea – but pulmonary fibrosis has not been reported. Other EGFR TKIs are at earlier stages of development for the treatment of malignant gliomas.
PDGF receptor pathway
The role of PDGF in gliomagenesis is well recognised. The v-sis oncogene isolated from cells transformed by the simian sarcoma virus is the cellular homologue of the PDGF-B ligand. Injection of simian sarcoma virus into primates induces glioma formation.(20) Recent studies have shown that the expression of PDGF-B by retrovirus induces malignant gliomas in mice.(21,22) Gliomas of all grades express PDGF ligands and receptors, and contribute to the malignancy of gliomas.(23,24) Additional roles of the PDGF signalling axis in cancers are derived from the roles that PDGF has been shown to play in regulating the survival of pericytes associated with tumour vasculature(25) and regulation of interstitial fluid pressure.(26)
Imatinib mesylate (STI571, Glivec, Gleevec; Novartis) is an oral, small-molecule ATP mimetic that inhibits the kinase activity of several oncogenes, including BCR-ABL, c-ABL, c-KIT and PDGF receptors. Several studies have been undertaken to evaluate imatinib mesylate in patients with recurrent malignant gliomas. A phase I/II study performed in 39 patients with malignant gliomas was associated with two grade 5 toxicities (intracerebral haemorrhage and pneumocystis pneumonia) and four patients with grade 4 toxicity.(27) Fourteen out of 31 evaluable patients experienced stable disease, with four patients having more than six months of stability. Another recent study combined imatinib mesylate and the chemotherapeutic drug hydroxyurea in the treatment of patients with glioblastomas resistant to nitrosoureas and temozolomide.(28) Of 26 evaluable patients, one experienced a complete response, four had partial responses and nine had stable disease (median time 12 months). No patients experienced grade 3 or 4 toxicities.
The formation of new blood vessels is a hallmark of malignant glioma pathology. This process, called angiogenesis, occurs only in adults in rare circumstances. As tumours begin to grow they secrete angiogenic factors that promote the formation and stabilisation of the new or co-opted blood vessels essential to tumour growth. VEGF, also known as vascular permeability factor (VPF), is a major regulator of angiogenesis. The expression of both VEGF ligands and receptors is increased in malignant gliomas.(7,8) The expression of VEGF correlates with tumour grade, as glioblastomas have the highest levels of expression.(29) Preclinical evaluations of anti-VEGF therapies have shown that they inhibit malignant glioma xenograft tumour growth.(30,31) Additional interest in antiangiogenic therapies for glioma therapy has been generated by the potential to avoid delivery challenges, as endothelial cells are the primary targets. Endothelial cells also retain genomic instability, thus limiting the development of resistance.
PTK787/ZK222584 is a novel oral small-molecule ATP mimetic inhibitor of VEGF receptors that has shown antitumour activity in preclinical studies on several cancer types. A phase I multi-institutional trial of PTK787 as monotherapy found that the agent was well tolerated, with dose-limiting toxicities of deep vein thrombosis, liver enzyme elevation, insomnia, cerebral oedema, fatigue and nausea/vomiting (see Figure 2).(32) One partial response was noted in 31 evaluable patients, with 20 patients (65%) having stable disease. Another phase I trial of PTK787 is currently examining its use in combination with either temozolomide or lomustine in patients with recurrent glioblastomas,(33) with evidence of radiographic response in a subset of patients. In preliminary analysis of the PTK787 arm with concurrent temozolomide, the best response noted was three partial responses and 21 patients with stable disease out of 34 evaluable patients. Two of the patients with partial responses had been previously treated with temozolomide monotherapy without response. Of 20 evaluable patients in the lomustine and PTK787 arm, one patient achieved a partial response and 13 patients had stable disease. Dose expansion is continuing.
Conclusion and future directions
The urgent need for effective drugs with improved side-effect profiles for malignant gliomas has prompted the development of novel therapeutic modalities, aimed at the molecular targets that are frequently altered in malignant gliomas. Early results of clinical trials with TKI receptor inhibitors are promising, but indicate that most therapeutic agents have modest efficacy as monotherapies. Therapies directed towards a single target would not be sufficient to achieve control of tumour growth in a broad spectrum of patients, suggesting that TKIs may be best used in combination with other signal transduction inhibitors or cytotoxic agents. We must also develop the capacity to predict tumour sensitivity and resistance to specific agents, but clinical trials of small molecules or other targeted agents in malignant gliomas have, to date, failed to link target expression and tumour response. Success in changing the outcome of malignant glioma will probably come in halting steps that may require strong advocacy from patients and clinicians.
This work was supported by grants from Accelerate Brain Cancer Cure and the Pediatric Brain Tumour Foundation of the USA. JNR is a Damon Runyon-Lilly Clinical Investigator and a Sidney Kimmel Cancer Foundation Scholar.
- Nat Rev Cancer 2003;3;375-9.
- Nature 1985;313:144-7.
- Cancer Res 1988;48:2231-8.
- Cancer Res 1991;51:2164-72.
- Neurology 1988;38:289-93.
- Cancer Res 1992;52:3213-9.
- Nature 1992;359:845-8.
- J Clin Invest 1993;91:153-9.
- J Neurosurg 1994;81:411-9.
- Cancer Res 1996;56:3859-61.
- Neurosurgery 1997;40:141-51.
- Drugs 2002;62:2237-48.
- Clin. Cancer Res 2001;7:2958-70.
- J Clin Oncol 2004;22:133-42.
- J Clin Oncol 2004;22 Suppl: Abs 1510.
- J Clin Oncol 2004;22 Suppl: Abs 1571.
- J Clin Oncol 2004;22 Suppl: Abs 1505.
- Proc Am Soc Clin Oncol 2003;22: Abs 394.
- Neuro-oncol 2003;5:Abs TA-37.
- J Natl Cancer Inst 1971;47:1115-20.
- Cancer Res 1998;58:5275-9.
- Genes Dev 2001;15:1913-25.
- Mol Cell Biol 1993;13:7203-12.
- Cancer Res 2000;60:5143-50.
- J Clin Invest 2003;111:1287-95.
- Cancer Res 2001;61:2929-34.
- Proc Am Soc Clin Oncol 2002;21: Abs 288.
- Proc Am Soc Clin Oncol 2004;22(145) Abs 1550.
- Clin Cancer Res 2003;9:3369-75.
- Cancer Res 1999;59:99-106.
- Cancer Res 2000;60:4152-60.
- Proc Am Soc Clin Oncol 2003;22: Abs 395.
- Neuro-Oncol 2004;27(145): Abs 1513.