Chemotherapy: opportunities for optimization in high-grade glioma

Modern chemotherapy can be traced to the discovery of the antitumoral properties of nitrogen mustard, a DNA alkylating agent used in chemical warfare. The first trial of nitrogen mustard derivatives, used to treat Hodgkin lymphoma in the 1940s, followed observations of lymphosuppressive and myelosuppressive effects in soldiers exposed to mustard gas. Most historical approaches to treating cancers have incorporated agents that derive a degree of disease specificity by inducing DNA damage in rapidly dividing cells. Chemotherapeutics have traditionally been derived from broadly toxic substances that trigger cascades of programmed cell death in actively dividing tumor cells. However, the sequelae of this strategy are the many nonspecific effects in normal cells with high rates of turnover, such as those in the bone marrow, digestive tract, and hair follicles. Examples of 2 cytotoxic drugs that remain standard of care in primary and recurrent glioblastoma multiforme (GBM) are carmustine (BCNU, Gliadel) and temozolomide (TMZ; Temodar, Temodal). These 2 drugs alkylate many cellular functional groups, including sites on guanine and cytosine nucleotides, thereby triggering the DNA damage–detecting checkpoint mechanisms of mitosis that subsequently promote cellular apoptotic cascades.

DNA-damaging approaches are limited in many aggressive tumors, because of mutations resulting in prosurvival traits including defective apoptotic signaling cascades, upregulation of rates of DNA repair, and increased rates of mutation leading to drug resistance. GBM is known to be inherently resistant to nearly every standard DNA-damaging chemotherapeutic agent. The lack of significant progress with traditional cytotoxic agents has provided the impetus for a new strategy of targeting specific alterations in signaling pathways responsible for the development and maintenance of GBM. Recent genome-wide analyses of human GBM tumor samples have amassed data regarding the commonly altered, mutated, or amplified genes implicated in GBM development, many of which involve receptor tyrosine kinase (RTK) signaling pathways. Several these overactive signaling pathways, which include upstream receptors as well as downstream targets of activation, are the specific focus of many small molecule inhibitor (SMI) drugs.

What is an SMI?

A small molecule drug is a nonpolymeric organic compound, generally fewer than 800 to 1000 Da. SMIs are designed to specifically inhibit the activity of a cellular constituent for therapeutic benefit. In practice, SMIs should be soluble in aqueous solution, lipophilic enough to cross the cellular membrane, and bind specifically to a target of interest to effect some change in cellular function. A particular advantage of these compounds is their potential for high selectivity for an active region of a given target, thus minimizing potential side effects. Major additional benefits inherent to small molecule compounds include the potential for oral bioavailability and, in the specific case of brain tumors, potentially superior passage across the blood-brain barrier relative to larger compounds (such as antibodies). These properties, combined with an ability to screen both new and modified compounds in high-throughput fashion, have led to the role of SMIs as a large proportion of drugs under current clinical study for cancer.

Most SMIs find efficacy via inhibition of function; as a result, most of these drugs are targeted toward reducing flux through overactive oncogenic pathways. Many of the SMIs in current clinical use have been identified either serendipitously or through in vitro screens focusing on a desired biologic activity. Several SMIs, designed to inhibit specific kinases that are upregulated in many cancers, have recently changed standard clinical practice for several solid tumors. Examples include lapatinib (Tykerb) for metastatic breast cancer, sunitinib (Sutent) for metastatic renal cell carcinoma (RCC), and sorafenib (Nexavar), which has proved efficacious in both advanced hepatocellular carcinoma and advanced RCC. Perhaps the best-known example of a targeted agent is the tyrosine kinase inhibitor imatinib mesylate (Gleevec). This small molecule was designed to inhibit a mutated RTK fusion protein, Bcr-Abl, the constitutive activation of which has long been known to cause chronic myelogenous leukemia (CML). In what is considered to be the proof-of-principle success story for SMIs in cancer, patients with chronic-phase CML treated with imatinib generally experience dramatic remission with few side effects.

The remainder of this article focuses on specific SMIs designed to target signaling pathways previously associated with malignant glioma (reviewed in Fig. 1 ) and reviews preliminary results of clinical trials using these drugs.

Molecular targets of interest for SMIs in glioblastoma. EGFR, epidermal-derived growth factor receptor; PDGFR, platelet-derived growth factor receptor, VEGFR, vascular endothelial growth factor receptor.

What is an SMI?

A small molecule drug is a nonpolymeric organic compound, generally fewer than 800 to 1000 Da. SMIs are designed to specifically inhibit the activity of a cellular constituent for therapeutic benefit. In practice, SMIs should be soluble in aqueous solution, lipophilic enough to cross the cellular membrane, and bind specifically to a target of interest to effect some change in cellular function. A particular advantage of these compounds is their potential for high selectivity for an active region of a given target, thus minimizing potential side effects. Major additional benefits inherent to small molecule compounds include the potential for oral bioavailability and, in the specific case of brain tumors, potentially superior passage across the blood-brain barrier relative to larger compounds (such as antibodies). These properties, combined with an ability to screen both new and modified compounds in high-throughput fashion, have led to the role of SMIs as a large proportion of drugs under current clinical study for cancer.

Most SMIs find efficacy via inhibition of function; as a result, most of these drugs are targeted toward reducing flux through overactive oncogenic pathways. Many of the SMIs in current clinical use have been identified either serendipitously or through in vitro screens focusing on a desired biologic activity. Several SMIs, designed to inhibit specific kinases that are upregulated in many cancers, have recently changed standard clinical practice for several solid tumors. Examples include lapatinib (Tykerb) for metastatic breast cancer, sunitinib (Sutent) for metastatic renal cell carcinoma (RCC), and sorafenib (Nexavar), which has proved efficacious in both advanced hepatocellular carcinoma and advanced RCC. Perhaps the best-known example of a targeted agent is the tyrosine kinase inhibitor imatinib mesylate (Gleevec). This small molecule was designed to inhibit a mutated RTK fusion protein, Bcr-Abl, the constitutive activation of which has long been known to cause chronic myelogenous leukemia (CML). In what is considered to be the proof-of-principle success story for SMIs in cancer, patients with chronic-phase CML treated with imatinib generally experience dramatic remission with few side effects.

The remainder of this article focuses on specific SMIs designed to target signaling pathways previously associated with malignant glioma (reviewed in Fig. 1 ) and reviews preliminary results of clinical trials using these drugs.

Molecular targets of interest for SMIs in glioblastoma. EGFR, epidermal-derived growth factor receptor; PDGFR, platelet-derived growth factor receptor, VEGFR, vascular endothelial growth factor receptor.

Epidermal-derived growth factor receptor

Perhaps the best example of increased RTK signaling in GBM is that of the epidermal-derived growth factor receptor (EGFR), which increases activation of the downstream RAS and PI3K intracellular signaling cascades. Early studies suggested that around 40% of GBM show EGFR amplification and protein overexpression leading to increased pathway flux. Furthermore, approximately 40% of GBM with EGFR amplification also harbor activating EGFR mutations. These findings have been recently supported through integrated genome analysis from The Cancer Genome Atlas (TCGA) Research Network study, which found that 41 of 91 (45%) sequenced tumors harbored EGFR alterations.

Several SMIs designed to inhibit EGFR and its mutant variants have been or are currently under investigation in GBM, including erlotinib (Tarceva), gefitinib (Iressa), lapatinib, and AEE788, as well as a host of monoclonal antibodies outside the scope of this article. Although the drugs seem to be well tolerated, most early single-agent trials of EGFR inhibitors have failed to show significant therapeutic benefit in GBM. In one phase II study, 13% of patients remained progression free for 6 months in response to gefitinib monotherapy. However, results from trials in lung cancer, in which improved clinical and radiographic responses to gefitinib have correlated with documented mutations in the EGFR kinase regions, have not been similarly recapitulated in patients with GBM. Preliminary data from trials focusing on erlotinib were slightly more encouraging than the results from gefitinib trials, suggesting potentially greater activity of this compound against the constitutively active EGFRvIII mutant receptor frequently found in GBM, but limited overall efficacy was seen. Despite disappointing early trials comparing single-agent administration of erlotinib versus temozolomide or BCNU to treat GBMbeing, further study to understand the potential cytostatic effects of EGFR inhibitors on GBM is warranted. To this end, EGFR inhibitors have been incorporated into multidrug trials including standard therapies (radiation therapy [RT] and TMZ) to determine any synergistic effect. One recent phase I/II trial showed no benefit of adding erlotinib to standard RT/TMZ protocols, whereas another phase II trial combining erlotinib with TMZ before and after RT showed increased median survival (19.1 months) relative to historical controls (14.1 months). It is likely that the variety of alterations observed in the RTK signaling axis of GBM means that some patients will find benefit from EGFR inhibition, whereas many will not, and also that combination therapies might be designed to address the most common alterations. Further studies using erlotinib as a component of first-line therapy may continue to elucidate this concept.

Platelet-derived growth factor receptor

Platelet-derived growth factor receptor (PDGFR) is another RTK signaling molecule with documented upregulation of expression in a subset of GBM. Inhibitors of PDGFR include the drugs imatinib, dasatinib (Sprycel), tandutinib, and pazopanib (Votrient). Imatinib, an inhibitor of PDGFR as well as other selected RTKs (including KIT and ABL), is indicated for the treatment of CML and gastrointestinal stromal cell tumors. As mentioned earlier, response to imatinib in these previously untreatable tumors is often dramatic. However, imatinib has generally failed to show similar efficacy as a single-agent therapeutic drug in GBM despite studies that detected intact imatinib within the GBM tissue. A recent European Organization for Research and Treatment of Cancer (EORTC) study using imatinib monotherapy in 112 patients with recurrent gliomas showed evidence of radiological response in the form of a reduction of postcontrast T1 gadolinium enhancement but did not show a concomitant improvement in clinical outcomes. It was concluded that, in the range of 600 to 1000 mg/d, imatinib shows a good safety profile but lacks antitumor activity in most patients with recurrent glioma. Two phase II studies of recurrent GBM, using a combination of hydroxyurea (HU) plus imatinib, suggested that this combination strategy was well tolerated by patients and showed evidence of response in excess of expectations. The positive results in these phase II trials led to completion of a randomized phase III trial comparing combination imatinib plus HU therapy with HU alone in progressive patients with TMZ-resistant tumors. However, no difference in progression-free survival (PFS) was seen between the 2 arms, with median PFS in both groups being only 6 weeks. Six-month PFS (PFS-6) was also similar at 5% and 7% respectively.

Vascular endothelial growth factor receptor

Angiogenesis, a phenomenon encompassing the creation of new blood vessels from existing vasculature, is a pathologic characteristic of GBM. This process is, in part, driven by the expression of the regulatory protein vascular endothelial growth factor (VEGF) and its receptor (VEGFR). The apparent need for angiogenesis in tumors, compared with the stable vascular networks present in other tissues, has implicated VEGFR signaling as an attractive target for inhibiting tumor growth, which is particularly relevant for GBM, a tumor in which increased vascular density and VEGF levels are associated with poor prognosis. Given recent clinical success with bevacizumab (Avastin), a humanized monoclonal antibody against VEGF, there has been increasing focus on exploring SMIs targeting VEGF/VEGFR in GBM.

SMIs developed to inhibit VEGFR include the drugs vatalanib and cediranib (tentative trade name Recentin), both of which have shown promise in early clinical trials. Vatalanib inhibits both the VEGFR and platelet-derived growth factor receptor (PDGFR), and has shown moderate effect when used alone or in combination with TMZ or lomustine to treat recurrent GBM in phase I/II multicenter trials.

A recent study designed to target both EGFR and VEGFR using a combination of erlotinib and bevacizumab (a monoclonal antibody) was well tolerated in patients, but showed no benefit in increasing PFS compared with that of historical regimens containing bevacizumab.

A phase II trial of cedirinib, which inhibits many forms of VEGFR, recently showed evidence of activity with a PFS-6 of 27.6%, normalization of vasculature, and reduction of edema in patients with GBM. Because serial sampling of GBM tissue is generally not possible, this study used multiple MRI-based methods to measure functional tissue response to cedirinib over time. These methods included measurements of vessel size, permeability, gadolinium enhancement, and diffusion-weighted imaging (DWI) characteristics. The results showed rapid reduction in vessel size, blood volume, and permeability to gadolinium contrast agents, with a corresponding reduction in vasogenic edema. The noninvasive nature of MRI allowed repeated measurement and temporal characterization of the vascular changes, showing them to begin as early as 24 hours after treatment, and to begin to reverse at day 28,although effects such as the reduction in vascular permeability persisted for up to 4 months. Measurement of circulating biomarkers also provided insight into the efficacy of cedirinib in this trial: following VEGFR inhibition, the concentration of circulating VEGF ligand increased. In addition, the use of MRI and biomarker measurement provided valuable insight into the duration of effects of the small molecule inhibition of VEGFR with cedirinib, suggesting that careful timing of combinatorial therapies (including cytotoxic drugs) might be critical to their success.

Inhibition of intracellular signaling cascades

Overall, the TCGA study found that 88% of all GBM harbored 1 or more mutations increasing the activity of the RTK signaling axis and flux through downstream RAS and PI3K pathways. In addition to EGFR activity, increased signaling of the ERBB2, c-MET, and PDGFR RTKs can all result in activation of RAS and PI3K; whereas signaling through VEGF activates both pathways via PKC-β. However, increased activity of RTKs are not the only drivers of RAS and PI3K signaling. RAS and PI3K are themselves upregulated in many GBM, whereas their endogenous inhibitors NF1 and PTEN (phosphatidylinositol phosphate 3′-phosphatase) are often mutated or lost. Loss of PTEN inhibition has been shown to remove sensitivity to EGFR inhibition by erlotinib and gefitinib and is a powerful negative prognostic factor. Subsequent signaling molecules in the 2 pathways have been identified and are also deregulated and mutated, resulting in increased flux through the pathways. The characterization of these downstream alterations gives rise to approaches other than simply targeting more or different cell surface RTKs.

RAS/RAF

RAS signaling ultimately activates the transcription factor extracellular signal-regulated kinase (ERK) by way of the intermediate proteins RAF and MEK. Tipifarnib (Zarnestra) is a farnesyl transferase–inhibiting drug shown to reduce signaling through the RAS pathway. Although early phase I trials determined that tipifarnib is well tolerated by patients with GBM, early phase II trials failed to show a benefit of tipifarnib when added to TMZ and RT. The RAF protein is among those inhibited by the drug sorafenib, currently being studied in GBM. In addition to its effects on RAF, sorafenib also shows inhibitory effects on VEGFR and PDGFR. However, despite potentially complimentary antitumor effects, trials combining sorafenib with TMZ and RT have thus far failed to show benefit, although the drug combination was well tolerated in patients. Similarly, studies of recurrent GBM treated concurrently with sorafenib and TMZ have been unsuccessful in improving outcomes. There are several additional active trials of sorafenib that may provide important information about the effect of combining sorafenib with other agents including erlotinib, the RAS inhibitor tipifarnib, and the mTOR inhibitor temsirolimus (Torisel). A study proposing treatment of recurrent GBM with sorafenib plus the mTOR inhibitor evirolimus (Afinitor, Zortress) has recently been approved, but is not yet recruiting patients.