- © 2005 by American Society of Clinical Oncology
Phase II Study of Imatinib Mesylate Plus Hydroxyurea in Adults With Recurrent Glioblastoma Multiforme
- David A. Reardon,
- Merrill J. Egorin,
- Jennifer A. Quinn,
- Jeremy N. Rich Sr,
- Idharan Gururangan,
- James J. Vredenburgh,
- Annick Desjardins,
- Sith Sathornsumetee,
- James M. Provenzale,
- James E. Herndon II,
- Jeannette M. Dowell,
- Michael A. Badruddoja,
- Roger E. McLendon,
- Theodore F. Lagattuta,
- Kimberly P. Kicielinski,
- Gregor Dresemann,
- John H. Sampson,
- Allan H. Friedman,
- August J. Salvado and
- Henry S. Friedman
- From the Departments of Medicine, Pharmacology and Cancer Institute, University of Pittsburgh, Pittsburgh, PA; Franz-Hospital Dülmen, Dülmen, Germany; Novartis Pharmaceuticals, Florham Park, NJ; and Departments of Surgery, Neurology, Pediatrics, Radiology, Pathology, and Cancer Center Biostatistics, Duke University Medical Center, Durham, NC
- Address reprint requests to David A. Reardon, MD, the Brain Tumor Center at Duke, Duke University Medical Center, Box 3624, Durham, NC 27710; e-mail: reard003{at}mc.duke.edu
Abstract
Purpose We performed a phase II study to evaluate the combination of imatinib mesylate, an adenosine triphosphate mimetic, tyrosine kinase inhibitor, plus hydroxyurea, a ribonucleotide reductase inhibitor, in patients with recurrent glioblastoma multiforme (GBM).
Patients and Methods Patients with GBM at any recurrence received imatinib mesylate plus hydroxyurea (500 mg twice a day) orally on a continuous, daily schedule. The imatinib mesylate dose was 500 mg twice a day for patients on enzyme-inducing antiepileptic drugs (EIAEDs) and 400 mg once a day for those not on EIAEDs. Assessments were performed every 28 days. The primary end point was 6-month progression-free survival (PFS).
Results Thirty-three patients enrolled with progressive disease after prior radiotherapy and at least temozolomide-based chemotherapy. With a median follow-up of 58 weeks, 27% of patients were progression-free at 6 months, and the median PFS was 14.4 weeks. Three patients (9%) achieved radiographic response, and 14 (42%) achieved stable disease. Cox regression analysis identified concurrent EIAED use and no more than one prior progression as independent positive prognostic factors of PFS. The most common toxicities included grade 3 neutropenia (16%), thrombocytopenia (6%), and edema (6%). There were no grade 4 or 5 events. Concurrent EIAED use lowered imatinib mesylate exposure. Imatinib mesylate clearance was decreased at day 28 compared with day 1 in all patients, suggesting an effect of hydroxyurea.
Conclusion Imatinib mesylate plus hydroxyurea is well tolerated and associated with durable antitumor activity in some patients with recurrent GBM.
INTRODUCTION
Although a modest survival benefit was recently demonstrated for newly diagnosed glioblastoma multiforme (GBM) patients treated with temozolomide plus external beam radiotherapy (XRT), recurrence remains nearly universal.1 Because of ineffective salvage therapies,2 most GBM patients die within 1 to 2 years of diagnosis. Therefore, innovative, more effective treatments are desperately needed for this patient population.
Novel therapeutics targeting activated signal transduction pathways are currently being widely evaluated in oncology. Imatinib mesylate (Gleevec, formerly STI-571; Novartis Pharmaceuticals, East Hanover, NJ), a selective receptor tyrosine kinase inhibitor of Bcr-Abl, c-KIT, c-fms, and platelet-derived growth factor (PDGF) receptors, has established activity against both hematologic and solid organ cancers.3-12 Imatinib has been evaluated for malignant glioma (MG) patients based on the frequent, increased expression of both PDGF and PDGFR.13,14 Although preclinical studies confirm substantial antiglioma activity,15 therapeutic benefit of imatinib in MG trials has been modest.16,17 In contrast, a marked benefit was reported in a small series of recurrent GBM patients treated with imatinib plus hydroxyurea, a ribonucleoside diphosphate reductase inhibitor.18 Although the mechanism of imatinib plus hydroxyurea is not known, preclinical studies demonstrate that signal transduction inhibitors can enhance the activity of cytotoxic agents.19 Our phase II study, conducted to further define the clinical activity of imatinib plus hydroxyurea among recurrent GBM patients, provides the first report of a regimen combining a signal transduction inhibitor with a cytotoxic agent for this population. Furthermore, we demonstrate that this regimen is active and well tolerated among GBM patients.
PATIENTS AND METHODS
Patient Eligibility
Patients were required to have recurrent GBM; be ≥ 18 years old; have measurable, contrast-enhancing tumor on magnetic resonance imaging (MRI); have a Karnofsky performance status ≥ 60%; be on a stable corticosteroid dose for ≥ 1 week; have satisfactory hematologic (hematocrit > 29%; ANC > 1,000 cells/μL; platelet count > 100,000 cells/μL) and biochemical results (serum creatinine, blood urea nitrogen [BUN], AST, and bilirubin < 2.0 times the upper limit of normal); have recovered from all expected toxicity related to previous therapy; and provide written informed consent. No restriction was placed on either the number of prior recurrences or treatments.
Patients were excluded for pregnancy or nursing; lack of effective, appropriate contraception; prior imatinib or hydroxyurea treatment; intratumoral hemorrhage (except postoperative grade 1); significant concurrent medical illness or prior malignancy; concurrent warfarin use; ≥ grade 2 peripheral edema, pulmonary or pericardial effusions, or ascites; and prior stereotactic radiosurgery or radioimmunotherapy, unless there was obvious radiographic disease progression or biopsy-proven recurrent tumor.
Treatment Design
Imatinib and hydroxyurea were given orally on a continuous dosing schedule for 28-day cycles until either unacceptable toxicity, tumor progression or consent withdrawal occurred. Imatinib was administered at 400 mg/d to stratum A patients (no enzyme-inducing antiepileptic drugs [EIAEDs]) and at 500 mg twice a day to stratum B patients (receiving EIAED). All patients received hydroxyurea at 500 mg twice a day. Patients undergoing pharmacokinetic studies began hydroxyurea on day 2.
Dose Modification and Re-Treatment
Physical examinations and MRI scans were performed pretreatment and before every cycle. A weekly CBC and monthly biochemistry profile were also obtained, and patients recorded their weight weekly. Toxicity was graded using National Cancer Institute Common Toxicity Criteria version 3. Imatinib was reduced (by 100 mg/d for stratum A patients and by 200 mg/d for stratum B patients) for related grade ≥ 3 nonhematologic or grade 4 hematologic toxicities. Hydroxyurea was reduced by 25% for grade 4 hematologic toxicity that developed after imatinib dose modification. Re-treatment required adequate hematologic and biochemical parameters (defined in eligibility criteria) and resolution of any related grade ≥ 3 toxicity to grade ≤ 1.
Response Evaluation
Study investigators determined response by neurologic examination and contrast-enhanced MRI. A complete response (CR) was defined as disappearance of all enhancing tumor on consecutive MRIs at least 6 weeks apart, with corticosteroid discontinuation and neurologic stability or improvement. A partial response (PR) was defined as ≥ 50% reduction in size (product of largest perpendicular diameters) of enhancing tumor with stability or improvement of neurologic status and corticosteroid requirement. Progressive disease was defined as ≥ 25% increase enhancing tumor or new lesion. Stable disease was defined as any assessment not meeting CR, PR, or progressive disease (PD) criteria.
Pharmacokinetic Analysis
Blood samples were collected from patients on days 1 and 28 of cycle 1 before treatment and 0.5, 1, 1.5, 2, 4, 6, 8, and 24 hours after their morning dose. Plasma supernatants were separated by centrifugation and immediately frozen (−20°C). Plasma concentrations of imatinib and its metabolite, CGP74588, were determined by high-pressure liquid chromatography/mass spectrometry.20 Day 1 data were used to calculate the maximum plasma imatinib concentration (Cmax) and time achieved (Tmax). Noncompartmental analysis21,22 was used to calculate the area under the concentration versus time curve from time zero to the last sampling point before the next dose of imatinib (area under the plasma concentration time curve [AUC]; AUC0-tend), AUC0-inf, and terminal plasma half-life (t1/2). Clearance (Clapp) was calculated as dose/AUC0-inf. For day 28, AUC0-24 hours and AUC0-12 hours were calculated for stratum A and B patients, respectively, and used in place of AUC0-inf to calculate Clapp.
Imatinib plasma protein binding determined by equilibrium dialysis as previously described.23 Initial studies used control human plasma (Central Blood Bank, Pittsburgh, PA) to which was added 5,000 ng/mL of imatinib and 0, 25, 50, or 100 μmol/L of hydroxyurea. Subsequent studies evaluated imatinib protein binding in day 1 and 28 plasma samples from each patient.
Plasma hydroxyurea concentrations were quantitated by high-pressure liquid chromatography as described with modification.24 BUN acid reagent (No. 535-3) and BUN color reagent (No. 535-5) were prepared as originally described.25 Separation was performed on a Waters μBondpak C18 column (10-μm, 3.9 × 300 mm; Waters Corp, Milford, MA) perfused at 1.7 mL/min with an isocratic mobile phase of acetonitrile:distilled water (13:87, v/v). Column eluate was monitored at 449 nm with a Waters 2487 dual absorbance detector and detector output was collected with Chromperfect Software (Justice Innovations, Mountain View, CA) to integrate the area under the hydroxyurea and internal standard peaks. Patient hydroxyurea concentrations were determined by calculating the ratio of area under the hydroxyurea peak to area of internal standard peak and comparing that ratio to a duplicate standard curve containing 0, 1, 3, 5, 10, 15, and 30 μg/mL hydroxyurea concentrations. Each run included duplicate quality control samples containing 3, 9, and 20 μg/mL. Under these conditions, hydroxyurea and methylurea internal standard eluted at approximately 6.3 and 11.8 minutes, respectively, and the lower limit of quantitation was 3 μg/mL.26 Plasma hydroxyurea Cmax and Tmax were determined from the concentration versus time curves of the day 28 hydroxyurea dose. Noncompartmental analysis21,22 was used to calculate the area under the concentration-time curve from time zero to the last sampling point before the next dose of hydroxyurea (AUC0-tend), AUC0-inf, Clapp, and t1/2.
Statistical Considerations
The primary objective was to evaluate the 6-month progression-free survival (PFS) rate. Because Yung et al27 reported a 6-month PFS of 21% (95% CI, 13% to 29%) among patients with GBM treated at first relapse with temozolomide, we used a single-stage, phase II design to differentiate between a 5% and 20% 6-month PFS rate with type I and II error rates of 0.074 and 0.093, respectively.
Time to progression and overall survival were measured from treatment initiation and analyzed by the Kaplan-Meier method including 95% CIs.28,29 The exact Wilcoxon rank sum test examined the relationship between pharmacokinetic parameters to EIAED use and measurement day.
Univariate logistic regression established whether each covariate (age, Karnofsky performance status, EIAED use, number of prior progressions, number of prior chemotherapy agents, time from original diagnosis, and day 28 imatinib AUC [hours 0 to 24 for stratum A patients; 2 × hours 0 to 12 for stratum B patients]) predicted PFS. If the P value was ≤ .25, then the characteristic was considered a possible multivariate Cox regression model covariate.30
RESULTS
Patient Characteristics
Thirty-three patients with recurrent GBM were enrolled, including 18 patients (55%) on stratum A and 15 patients (45%) on stratum B (Table 1). All patients received prior XRT and chemotherapy that included at least temozolomide.
Outcome
With a median follow-up of 58.1 weeks, the 6-month PFS rate for all patients was 27% (95% CI, 16% to 48%; Table 2; Fig 1A). Patients on stratum B had a significantly better PFS rate than stratum A patients (P = .0217; Table 2; Fig 1B). PFS was also better among patients with only one prior episode of progressive disease (P = .0043; Fig 1C). The median PFS rates for all patients and for those on strata A and B were 14.4 weeks (95% CI, 8.3 to 16.6 weeks), 8.5 weeks (95% CI, 7.9 to 15.9 weeks), and 16.6 weeks (95% CI, 14.4 to nonestimable [NE] weeks), respectively. Five patients (16%) continue to receive treatment on study in cycles 11 (n = 2), 12 (n = 1), 13 (n = 1), and 14 (n = 1). The median overall survival rates for all patients and for those in strata A and B were 48.9 weeks (95% CI, 25.7 to 71.1 weeks), 32.7 weeks (95% CI, 21.1 to 71.1 weeks), and 56.6 weeks (95% CI, 33.7 to NE), respectively.
One patient achieved a complete radiographic response, two patients achieved a partial response (9% overall response rate), and 14 patients achieved stable disease (Table 2; Fig 2).
Prognostic Implications
A multivariate Cox regression model for PFS identified EIAED use (P = .0387), one prior episode of progressive disease (P = .0131), and time from original diagnosis to study enrollment (P = .0126) to be jointly predictive of PFS (Table 3). Specifically, each independent variable retained statistical significance after adjustment for the other two variables as follows: patients receiving EIAED had a 59% decreased risk of progression than patients not treated with EIAEDs; patients with more than one recurrence had a 2.9-fold higher risk of progression; and each week from original diagnosis to study enrollment conferred a 1.01-fold greater risk of progression.
Toxicity
The most frequent grade 3 toxicities (Table 4) were hematologic and included neutropenia (n = 5, 15%) and thrombocytopenia (n = 2, 6%). Nonhematologic grade 3 toxicities included edema (n = 2; 6%) and reversible transaminase elevation (n = 2; 6%). Individual patients experienced grade 3 lower-extremity deep venous thrombosis, fatigue, and pulmonary edema. No grade 4 or 5 events occurred.
Pharmacokinetic Analyses
Eight patients from stratum A and nine patients from stratum B had plasma imatinib and CGP74588 concentration versus time profiles (Table 5). Day 1 imatinib Cmax, Tmax, Clapp, and t1/2 values for stratum A patients agree with previous reports.31,32 In contrast, patients on EIAED had a significantly shorter t1/2 and lower AUC and Clapp than stratum A patients. The impact of EIAED on CGP74588 pharmacokinetics was less striking. In addition, comparison of day 1 and 28 imatinib pharmacokinetic parameters revealed decreased imatinib Clapp on day 28, regardless of strata. Hydroxyurea pharmacokinetics were comparable between the strata (Table 5) and were compatible with previous reports.33-35
Plasma hydroxyurea concentrations representative of those produced by the 500-mg twice daily dosing used in our study had no effect on plasma protein binding of a 5,000-ng/mL imatinib solution in control human plasma (data not shown). Furthermore, hydroxyurea had no consistent effect on plasma protein binding of imatinib (Table 5). Moreover, there was no difference in percentage of unbound imatinib on days 1 and 28 between patients receiving and not receiving EIAEDs (Table 5).
Dose Escalation for Patients Receiving EIAEDs
After completion of accrual, we amended our phase II study to enroll six additional recurrent GBM patients treated with EIAEDs to evaluate the safety of continuous treatment with 600 mg of imatinib twice a day with 500 mg of hydroxyurea twice a day. Three patients experienced significant toxicity. Individual patients developed abdominal pain (grade 3) and neutropenia (grade 4), whereas a third patient developed transaminitis (grade 4) and pericardial (grade 4) and pleural (grade 3) effusions. Of note, each of these patients recovered fully after treatment interruption.
DISCUSSION
Our study demonstrates that imatinib plus hydroxyurea is modestly active and well tolerated for patients with GBM who have experienced disease progression after standard of care therapy with XRT and temozolomide. Potential limitations of our study include the following: a nonblinded, nonrandomized, single-institutional design; modest study size; and enrollment of relatively young patients with good performance status. Although our study was not powered to detect a statistically different outcome compared with published reports, our results are nonetheless encouraging relative to two benchmark studies for patients with recurrent GBM.2,27 Specifically, temozolomide administered at first relapse achieved a median PFS of 12.4 weeks, a 6-month PFS rate of 21% (95% CI, 13% to 29%), and a 5% radiographic response rate,27 whereas Wong et al2 reported a median PFS of 9 weeks and a 6-month PFS rate of 15% (95% CI, 10% to 19%) with eight consecutive salvage regimens. In contrast, treatment with imatinib plus hydroxyurea on our study achieved a median PFS of 14.4 weeks, a 6-month PFS rate of 27% (95% CI, 16% to 48%), and a radiographic response rate of 9%.
Dresemann18,36 first described durable antitumor activity and minimal toxicity in patients with recurrent GBM treated with imatinib and hydroxyurea. Although we corroborate his findings, this regimen’s therapeutic benefit is unexpected for two reasons. First, separate administration of either imatinib or hydroxyurea is minimally active among malignant glioma patients.16,17,37-41 Second, the imatinib and hydroxyurea dose levels were selected empirically. To validate Dresemann’s results, we used the same dosing schedule but increased the imatinib dose to 1,000 mg/d for patients receiving EIAEDs, based on the known effect of EIAEDs on imatinib metabolism and a report that a 1,000-mg/d dose was well tolerated.16 We increased the dose of imatinib among patients concurrently taking EIAEDs to 1,200 mg a day, but observed significant toxicity, thereby confirming that the maximum-tolerated dose of imatinib, when combined with hydroxyurea among patients receiving EIAEDs, is 500 mg twice a day.
There are several possible mechanisms of action for imatinib plus hydroxyurea. First, given the frequency of increased PDGF/PDGFR expression and activity in MG,13-15,42,43 correlation of response to either tumor PDGF/PDGFR expression or PDGFR mutations should be evaluated. Although epidermal growth factor receptor (EGFR) expression does not predict response to EGFR inhibitors in patients with recurrent GBM,44 target expression and inhibition should be evaluated for GBM patients undergoing PDGFR-directed therapies. Furthermore, examination for PDGFR mutations should also be considered, based on the association of somatic activating EGFR mutations with responsiveness to EGFR inhibitors in lung cancer patients.45,46 Unfortunately, tissue samples were not available for our study. Nonetheless, PDGF/PDGFR expression and PDGFR mutations can be evaluated in preclinical models and in patient tumor samples in future studies. However, if response to imatinib regimens is critically linked with either PDGF/PDGFR expression or PDGFR mutations in GBM patients, one would expect greater antitumor activity with imatinib monotherapy than previously reported.16,17
Imatinib can also diminish tumor interstitial pressure, leading to increased capillary-to-interstitium transport in vivo and enhanced chemotherapy delivery.47-49 Data from preclinical studies support this potential mechanism.50-52 Thus imatinib may enhance hydroxyurea cytotoxicity by improving its delivery into the tumor microenvironment. As a logical next step to further investigate this mechanistic interaction, we are currently conducting a clinical trial of imatinib with temozolomide, a cytotoxic agent with more established single-agent activity against malignant glioma than hydroxyurea.
PDGFR inhibitors also exhibit significant antiangiogenic activity in preclinical models,53,54 primarily by targeting perivascular cells.55 In addition, protracted dosing of several chemotherapeutics (metronomic chemotherapy) suppresses tumor angiogenesis, enhances the antitumor activity of vascular endothelial growth factor inhibitors, and limits tumor growth, including GBM.56-59 Therefore, PDGFR inhibition combined with metronomic chemotherapy may provide complementary antiangiogenic activity.
Imatinib also diminishes tumor cell DNA repair after chemotherapy or XRT by reducing Rad51 expression, a critical component of the DNA double-strand break pathway.60,61 Thus imatinib-related decreased DNA repair may potentate the cytotoxicity of hydroxyurea.
A final potential mechanism of this regimen may be enhanced drug delivery via modulation of adenosine triphosphate–dependent transporter proteins responsible for regulating uptake and efflux of agents at both the blood-brain barrier and tumor cell membrane.62-70 Thus imatinib plus hydroxyurea may interact to enhance delivery into the CNS and possibly tumor cells by interacting as either competitive substrates or direct inhibitors of critical regulatory transporter proteins.
Multivariate analysis demonstrated that concurrent EIAED use, no more than one prior episode of PD, and time from original diagnosis to study enrollment were independent prognostic predictors of PFS. Although the latter two factors are not surprising among patients with recurrent MG, the association of EIAED use with improved PFS was unexpected, because EIAED use significantly lowered imatinib exposure. The explanation for this paradoxical association is not clear and warrants further investigation. Our protein-binding studies indicate that EIAEDs do not increase the percentage of unbound imatinib, which argues against the better response to treatment seen in that cohort being due to an increased imatinib free fraction.
Although the increased Clapp and shorter t1/2 of imatinib observed in patients taking EIAED was expected, we were surprised that, despite a 2.5-fold higher daily imatinib dose, imatinib exposures remained significantly lower in such patients. This finding further attests to the powerful effect of EIAEDs on hepatic drug metabolism. Another unexpected finding was the consistent decrease in day 28 imatinib Clapp observed among all patients.71 This decrease was not due to increased imatinib protein binding. Furthermore, the 98% bioavailability of imatinib implies that decreased day 28 imatinib Clapp reflects inhibited CYP3A metabolism. Although hydroxyurea is not a known cytochrome p450 (CYP) substrate, recent studies of the 5-lipoxygenase inhibitor, zileuton, the structure of which includes a hydroxyurea moiety, indicate that it inhibits CYPs, including CYP3A.72 We are actively performing in vitro studies to test the effect of hydroxyurea on imatinib metabolism. Furthermore, in that hydroxyurea inhibits ribonucleotide reductase via interaction with a catalytic iron, inhibition by hydroxyurea of another iron-dependent enzyme, such as a CYP isoform, is feasible. If correct, demonstration of CYP inhibition by hydroxyurea would have clinical implications broader than those related to the current study.
Our study is the first to report that a signal transduction inhibitor combined with a chemotherapeutic has modest activity in the treatment of patients with GBM. Given the dire need for effective therapies for this patient population, our study results provide impetus for two important efforts. First, validation of our results in a multicenter trial for recurrent GBM patients that ideally incorporates evaluation of potential clinical prognostic factors and correlative tumor biomarkers is warranted. Second, a better understanding of the mechanism underlying the activity of this regimen is imperative. Once better understood, additional strategies to exploit this mechanism further should be investigated.
Authors’ Disclosures of Potential Conflicts of Interest
Although all authors completed the disclosure declaration, the following author or immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed discription of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
| Authors | Employment | Leadership | Consultant | Stock | Honoraria | Research Funds | Testimony | Other |
|---|---|---|---|---|---|---|---|---|
| David A. Reardon | Novartis Pharmaceuticals (A) | Novartis Pharmaceuticals (A) | ||||||
| Merrill J. Egorin | Novartis (A); Bristal-Myers Squibb (A) | Novartis (A); Bristal-Myers Squibb (A); Novartis (A) | ||||||
| August J. Salvado | Novartis Pharmaceuticals, (N/R) | Novartis Pharmaceuticals (B) |
Dollar Amount Codes (A) < $10,000 (B) $10,000–99,000 (C) ≥ $100,000 (N/R) Not Required
Kaplan-Meier estimate of progression-free survival for (A) all patients, (B) patients stratified by concurrent use of enzyme-inducing antiepileptic drugs, and (C) patients with one versus more than one prior episode of progressive disease (PD).
Contrast-enhanced magnetic resonance images of representative responses to imatinib mesylate plus hydroxyurea, including (A) a patient who achieved a complete response and (B) a patient who achieved a partial response.
Patient Characteristics
Parameters of Efficacy
Summary of Univariate and Multivariate Cox Regression Analysis Results: Progression-Free Survival
Number of Patients Experiencing Grade 3 Toxicity
Comparison of Pharmacokinetic Measures of Imatinib, CGP74588, and Hydroxyurea by Stratum
Footnotes
-
Supported by National Institutes of Health Grants No. NS20023, CA11898, MO1 RR 30, General Clinical Research Center Program, National Center for Research Resources; NCI Specialized Program of Research Excellence P30CA47904 and a grant from Novartis Pharmaceuticals.
Authors’ disclosures of potential conflicts of interest are found at the end of this article.
- Received June 22, 2005.
- Accepted September 30, 2005.
