Introduction

Current treatment of high-grade glioma (HGG) consists of a multi-modality approach including surgery, radiotherapy, and chemotherapy [1]. Despite optimal therapy, the prognosis remains poor with a median overall survival (OS) for newly diagnosed glioblastoma (GBM) of 14.6 months and for anaplastic glioma (AG) of 2–5 years [2–5]. Survival outcomes for recurrent HGG remain grim with 6 month progression-free survival (PFS6) of 9–16% for GBM and 28–31% for AG [6–8]. These dismal outcomes have prompted the search for more effective forms of therapy.

Advances in molecular biology have uncovered clues to glioma pathogenesis. Common aberrations in cellular signaling involved in growth, survival, angiogenesis, and invasion have been identified, and new treatments are aimed at targeting specific molecules in these pathways. Much interest has surrounded the use of anti-angiogenic agents as a novel anticancer therapy. The Food and Drug Administration (FDA) recently granted accelerated approval for bevacizumab, a humanized monoclonal antibody to vascular endothelial growth factor (VEGF), for the treatment of recurrent GBM [9–11]. Unfortunately, the effect on survival is modest with eventual progression of tumors.

As recently reviewed, mechanisms of resistance to anti-angiogenic therapies include upregulation of alternative pro-angiogenic signaling pathways; recruitment of pro-angiogenic bone marrow derived endothelial progenitor cells and monocytes; increase of protective pericyte coverage along tumor vasculature; and activation of invasion [12, 13]. There is increasing evidence that endothelial cells can induce pericyte recruitment for protection when survival signals, normally conveyed by VEGF, are inhibited [12, 14–18]. Platelet derived growth factor-B (PDGF-B) and platelet derived growth factor receptor beta (PDGF-Rβ) have been implicated in pericyte recruitment during developmental vasculogenesis, and PDGF-B expression has been linked to pericyte recruitment in gliomas and other solid tumors [19–22].

Another proposed mechanism of resistance of tumors to anti-angiogenic agents is activation of a more invasive phenotype that is mediated in part by the SRC pathway [12, 23–26]. Dysregulation of the SRC pathway is implicated in tumorigenesis, with Src proteins functioning as a central regulator between extracellular signaling and intracellular activation [27]. Recent work has shown that the Src family kinases, including FYN and SRC, are effectors of oncogenic EGFR signaling, promoting invasion and tumor cell survival [27, 28]. Pharmacologic inhibition of FYN and SRC result in decreased tumor invasion, tumor regression, and tumor cell apoptosis in cell culture [28].

Dasatinib is an oral small molecule inhibitor of SRC, BCR-ABL, c-KIT, EPHA2, and PDGFRβ [29]. It has demonstrated efficacy in malignancies with mutant BCR-ABL, KIT, and PDGFR [30, 31]. Dasatinib is FDA-approved for treatment of chronic myelogenous leukemia (CML) that is resistant or intolerant to other therapies and Philadelphia chromosome positive acute lymphoblastic leukemia (ALL). The role of this drug in solid tumors has yet to be identified, though data from phase I and II studies in advanced melanoma, gastrointestinal stromal tumors (GIST), and prostate cancer have been encouraging [32, 33]. In this report, we review our experience with dasatinib in conjunction with bevacizumab in a heavily pretreated group of patients with recurrent GBM who have failed prior therapy with bevacizumab.

Methods

A retrospective review of 14 adult patients with histologically confirmed GBM and radiographically documented recurrent disease was conducted. All patients had prior treatment with radiotherapy, age 18 or older, Karnofsky Performance Status (KPS) of 50 or greater, absolute neutrophil count (ANC) greater than or equal to 1500/Ul, normal renal function, and prior failure of bevacizumab therapy. There was no limit on the number of previous therapies. Endpoints included response rate, progression free survival (PFS), six-month progression free survival (PFS6), overall survival (OS), and toxicities.

Between February 2007 and August 2009, patients were treated with dasatinib 70–100 mg twice daily, in combination with bevacizumab until tumor progression or unacceptable toxicity. In general, patients with a history of myelosuppression received dasatinib 70 mg twice daily, while patients without such a history received 100 mg twice daily. Brain MRI scans were obtained every 8 weeks, or more often if clinically indicated. Response was determined using Macdonald criteria [34]. Toxicities were graded according to the National Cancer Institute Common Toxicity Criteria (CTCAE 3.0). The retrospective analysis was approved by the Institutional Review Board (IRB) of the Dana Farber Cancer Institute. The Kaplan–Meier method was used to estimate survival function.

Results

Patient characteristics are presented in Table 1. Fourteen patients were treated (13 males and 1 female). The median age was 55 years (range 32–66) and median KPS was 80 (range 50–90). All patients (100%) had GBM. The median number of prior regimens was 4 (range from 2 to 6) and included prior treatment with bevacizumab, either alone or in combination with another agent. These fourteen patients received dasatinib in combination with bevacizumab 10 mg/kg every 2 weeks.

Table 1 Patient characteristics
Full size table

Thirteen patients were evaluable. One patient was censored for enrollment in an impending vaccine trial. Of the thirteen evaluable patients, there were no radiographic responses. One patient (7%) had stable disease after an 8 week interval. Median PFS was 28 days (95% CI 26–35 days), and median OS was 78 days (95% CI 41–137 days).

The treatment was fairly well tolerated (Table 2). The most common toxicities included fatigue, diarrhea, and thrombocytopenia. Two patients (14%) stopped dasatinib therapy because of toxicity. These toxicities included grade 1 fever and rigor (1) and grade 4 hemorrhage (1). Overall, most toxicities were graded 1. The patient who sustained a grade 4 cerebral hemorrhage was not on anticoagulation at the time hemorrhage.

Table 2 Toxicities
Full size table

Discussion

This retrospective study showed that dasatinib, a SRC, BCR-ABL, c-KIT, EPHA2, and PDGFRβ inhibitor, in conjunction with bevacizumab, does not have significant activity in patients with recurrent, heavily pretreated GBM following failure of bevacizumab therapy. Median PFS was only 28 days, and PFS6 was 0%. Median OS was 78 days. The treatment was moderately well tolerated with only two patients experiencing adverse effects requiring discontinuation of the drug. The main toxicities included fatigue, diarrhea, rash, fever, thrombocytopenia, and hemorrhage. The observation of one grade 4 cerebral hemorrhage in this small cohort is concerning, and further evaluation of dasatinib in this population warrants caution.

Several reasons could account for these dismal outcomes. The cohort consisted of heavily pre-treated patients (median of 4 prior therapies) who had failed bevacizumab. Clinical deterioration has been reported after disease progression on bevacizumab, secondary to aggressive tumor growth and rebound edema [35]. Attempts to add a chemotherapeutic agent to bevacizumab in these patients have been generally ineffective [36–38]. One study reported median PFS of 2 months (95% CI 1.2–3.3 months), PFS6 of 0%, and median OS of 5.2 months (95% CI 3.3–8.4 months) with salvage chemotherapy after bevacizumab failure [38]. Another recent study reported median PFS of 37.5 days (95% CI 34–42 days) and PFS6 of 2% with a second bevacizumab regime [37]. Our results were slightly worse with a median PFS of 1 month, PFS6 of 0%, and median OS slightly greater than 2 months.

There is increasing animal and human data suggesting that treatment with bevacizumab induces a particularly invasive tumor phenotype in a subset of patients [36, 39]. Several reviews on glioma invasion and migration have recently been published implicating SRC [27, 40, 41]. Although dasatinib inhibits SRC, it is possible that other signaling pathways play a greater role in tumor invasion.

Another potential reason for the lack of response to dasatinib is the inability of the drug to penetrate the blood–brain barrier and achieve adequate concentrations in the tumor. Some murine and human studies suggest that dasatinib’s efficacy against intracranial leukemia is secondary to improved central nervous system (CNS) penetration over imatinib [42]. Unfortunately, pharmacokinetic analysis shows that its CNS penetration, which ranges from 5 to 28% of plasma levels, is still relatively low when compared to other agents known to have good penetration. CNS penetration of this drug remains a real challenge, potentially preventing its effective use in gliomas. Multidrug efflux transporters such as P-glycoprotein (P-gp; ABCB1), breast cancer resistance protein (ABCG2), and multidrug resistance protein 2 (ABCC2) mediate drug resistance to a variety of drugs [43, 44]. Recent animal studies have found that brain accumulation of dasatinib is restricted by P-gp and ABCG2 at the blood brain barrier; inhibition of these transporters increased the brain uptake of this drug [44]. It is possible that the lack of response to dasatinib is partly a consequence of ineffective drug delivery to its molecular target. In addition, the concomitant administration of bevacizumab may reduce disruption of the blood–brain barrier and further decrease CNS penetration of dasatinib. Co-administration of the drug with an inhibitor of P-gp and ABCG2 could theoretically overcome this issue. Preclinical studies demonstrate that P-gp inhibitors, such as elacridar, can increase the concentrations of drugs in brain tissue, although its efficacy in humans is unknown [45, 46]. Despite these arguments, it is well known that the blood–brain barrier is disrupted in the center of high-grade gliomas, which may indicate that tumor penetration is not the major obstacle to dasatinib’s efficacy in this setting.

The activity of single agent dasatinib in recurrent GBM is currently being evaluated by the Radiation Therapy Oncology Group (RTOG 0627). It remains possible that dasatinib could be effective in a particular GBM subgroup defined by molecular genetics. For instance, there is report of increased sensitivity to dasatinib in glioma cells with functional PTEN [47]. However, it must be noted that the applicability of these in vitro data to glioma patients is uncertain. The same study also found significant augmentation of autophagy when dasatinib was combined with temozolomide [47]. Thus, if confirmed in other preclinical models, combination therapy with dasatinib may also be worth exploring. Because of the small number of patients evaluated in this retrospective study and the lack of a uniform treatment approach, subgroup analysis was not pursued in the current dataset.

In summary, this retrospective analysis did not find any significant activity of dasatinib, in combination with bevacizumab, in our cohort of heavily pretreated patients who had failed prior bevacizumab therapy. Whether there may be a role for this drug in the therapy of a selected group of less heavily pretreated patients with recurrent GBM is unclear and will require larger prospective studies. Any future clinical studies regarding this agent should include correlative biomarkers to explore the biological basis of success or failure.