Abstract

INTRODUCTION

Despite recent advances in therapeutic regimens, the diagnosis of malignant glioma still carries a poor prognosis. Although alkylators such as temozolomide and polifeprosan 20 with carmustine implant (Gliadel, Guilford Pharmaceuticals-MGI Pharma, Bloomington, MN) are US Food and Drug Administration–approved for treatment of malignant glioma, their ability to prolong survival is short-lived. Thus, innovative therapeutic agents and strategies are imperative.

One novel approach to fighting this deadly disease is to target the mechanisms of resistance to chemotherapy. Because O⁶-alkylguanine-DNA alkyltransferase (AGT) has been shown to be a major factor in the resistance of tumor cells to alkylating agents such as carmustine and temozolomide,^1–3 perhaps targeting this DNA repair protein can help restore chemotherapy activity.

One such agent that can irreversibly inactivate AGT is a low molecular weight substrate, O⁶-benzylguanine (O⁶-BG). In vitro^1,4,5 and in vivo^6,7 studies using tumor cell lines and subcutaneous brain tumor xenografts demonstrate that O⁶-BG increases the therapeutic effectiveness of temozolomide.

These studies suggest that combining O⁶-BG with temozolomide may circumvent resistance and restore chemotherapy sensitivity. However, before one can combine these two agents, one must answer the following two important questions: what dosing schedule of O⁶-BG is necessary to completely suppress AGT activity in brain tumors for at least 48 hours, and what is the maximum-tolerated dose (MTD) of temozolomide in a 1-day dosing regimen that can be combined with O⁶-BG?

Prior research endeavors have answered both of these questions. Weingart et al⁸ answered this first question in a phase I trial of polifeprosan 20 with carmustine implant plus continuous infusion of intravenous O⁶-BG in adults with recurrent malignant glioma by establishing the dose of O⁶-BG to be 120 mg/m² for 1 hour followed by a continuous infusion of 30 mg/m²/d for 2 days. Quinn et al⁹ answered the second question in a phase I trial of temozolomide plus O⁶-BG in adults with recurrent malignant glioma by establishing the MTD of temozolomide to be 472 mg/m² when administered as a single dose in conjunction with the above dosing schedule of O⁶-BG.

The objectives of this phase II trial were twofold: to define the role of O⁶-BG in restoring temozolomide sensitivity in patients with recurrent or progressive, temozolomide-resistant malignant glioma and to evaluate the safety of administering O⁶-BG in combination with temozolomide.

PATIENTS AND METHODS

RESULTS

Patient Characteristics

Sixty-seven patients were enrolled. Sixty-six patients received at least one dose of temozolomide and O⁶-BG. One patient with AA did not receive any protocol treatment because of withdrawal of consent and is excluded from all summaries. Patient demographics and baseline disease characteristics of the 66 patients who received protocol treatment are listed in Table 1. Central pathologic review confirmed that 34 patients had GBM, 28 patients had AA, and four patients had AO. The median age was 51 years, and the majority of patients were male. Fifty-five percent of patients had a Karnofsky performance score of ≤ 80. All patients had progressive or recurrent disease after prior therapy with temozolomide, and the majority of patients (59%) had at least two prior progressions. The median time from diagnosis to the start of therapy with temozolomide plus O⁶-BG was 60 weeks (range, 12 to 72 weeks).

Table 1.

Patient Demographics and Clinical Characteristics

Characteristic	No. of Patients (N = 66)	%
Age, years
Median	51
Range	21-69
Sex
Male	44	67
Female	22	33
Karnofsky performance score, %
< 70	3	4
70-80	33	51
90-100	30	45
No. of progressions
Median	2
Range	1-5
Time from diagnosis, weeks
Median	60
Range	12-702
Histological diagnosis
AA	28	42
AO	4	6
GBM	34	52

Open in a separate window

Abbreviations: AA, anaplastic astrocytoma; AO, anaplastic oligodendroglioma; GBM, glioblastoma multiforme.

Response

The overall ORR for all histologies was 9% (95% CI, 3% to 19%). One (3%) of 34 patients with GBM (95% CI, 0.1% to 15%) and five (16%) of 32 patients with anaplastic glioma (95% CI, 5% to 33%) were responders to the combination of temozolomide and O⁶-BG. Table 2 details the response in each histology subgroup. One patient died as a result of a perforated cecum 2 days after treatment with this drug combination. Five patients with anaplastic glioma who had a response completed at least a full year of therapy. Of these five patients, disease progression did not occur in three patients for 131, 93, and 89 weeks, respectively, and the other two patients remain disease-free 185 and 164 weeks from enrollment, respectively. Of interest, one patient with an anaplastic glioma who completed a year of therapy remained progression free for 143 weeks but never realized a response by the modified Macdonald criteria. However, this patient's positron emission tomography scan was hypometabolic on completion of a full year of therapy.

Table 2.

Overall Survival, Progression-Free Survival, and Objective Response Rate According to Histology

Variable	Overall		GBM		AA/AO
	No. of Patients	95% CI	No. of Patients	95% CI	No. of Patients	95% CI
Overall survival
Median, weeks*	24.7	20.9 to 31.0	19.4	13.7 to 24.3	33	28 to 49.6
At 6 months, %	47	35 to 58	26	13 to 42	69	50 to 82
Progression-free survival
Median, weeks*	7.9	7.6 to 9.6	7.5	4.2 to 7.9	9.6	7.9 to 16.1
At 6 months, %	17	9 to 27	9	2 to 21	25	12 to 41
Objective response rate,* %	9	3 to 19	3	0.1 to 15	16	5 to 33

Open in a separate window

Abbreviations: AA, anaplastic astrocytoma; AO, anaplastic oligodendroglioma; GBM, glioblastoma multiforme.

^*There was a statistically significant difference between the two histologies.

PFS

The 6-month PFS was 17% overall (95% CI, 9% to 27%), 9% for GBM (95% CI, 2% to 21%), and 25% for anaplastic glioma (95% CI, 12% to 41%). The median PFS was 7.9 weeks overall (95% CI, 7.6 to 9.6 weeks), 7.5 weeks for GBM (95% CI, 4.2 to 7.9 weeks), and 9.6 weeks for anaplastic glioma (95% CI, 7.9 to 16.1 weeks). There was a significant statistical difference between the two histologies (P = .0049). The PFS data are illustrated by the Kaplan-Meier curve in Figure 1.

Open in a separate window

Fig 1.

Kaplan-Meier estimates of progression-free survival according to histology. AA, anaplastic astrocytoma; AO, anaplastic oligodendroglioma; GBM, glioblastoma multiforme.

OS

The 6-month OS was 47% (95% CI, 35% to 58%), 26% for GBM (95% CI, 13% to 42%), and 69% for anaplastic glioma (95% CI, 50% to 82%). The median OS was 24.7 weeks (95% CI, 20.9 to 31.0 weeks), 19.4 weeks for GBM (95% CI, 13.7 to 24.3 weeks), and 33 weeks for anaplastic glioma (95% CI, 28.0 to 49.6 weeks). There was a significant statistical difference between the two histologies (P = .0004). The OS data are presented as a Kaplan-Meier curve in Figure 2.

Open in a separate window

Fig 2.

Kaplan-Meier estimates of overall survival according to histology. AA, anaplastic astrocytoma; AO, anaplastic oligodendroglioma; GBM, glioblastoma multiforme.

Toxicity

Table 3 summarizes the adverse events (AEs). The most commonly reported AEs were grade 4 hematologic events experienced in 48% of the patients, for whom a 25% temozolomide dose reduction was required for subsequent dose administrations. The most frequent hematologic AEs were grade 4 neutropenia (47%), followed by grade 4 thrombocytopenia (12%). One patient died as a result of a perforated cecum 2 days after initiation of treatment with this drug combination. This event was unlikely to be drug related, given that before enrollment she complained of abdominal pain. One patient was not assessable for toxicity because the patient never received this drug combination.

Table 3.

Adverse Events

Adverse Event	Grade 3		Grade 4		Grade 5
	No. of Patients	%	No. of Patients	%	No. of Patients	%
Hematologic
Anemia			1	1
Lymphopenia			1	1
Thrombocytopenia			8	12
Neutropenia			31	47
Neutropenia, febrile	1	1
Nonhematologic
Infection without neutropenia	2	3
Seizure	4	6
Thrombosis/embolism			2	3
GI perforation					1	1

Open in a separate window

DISCUSSION

Temozolomide and polifeprosan 20 with carmustine implant are both US Food and Drug Administration–approved drugs because of their ability to prolong survival in patients diagnosed with malignant glioma. Unfortunately, all patients with this disease will eventually experience treatment failure with these agents.

Resistance to both temozolomide and carmustine is mediated in part through the DNA repair protein AGT. Three clinical trials^13–15 reported an inverse relationship between AGT levels and survival for patients with malignant glioma treated with carmustine. Esteller et al¹⁶ also confirmed this relationship, albeit using methylation of the AGT promoter gene in lieu of quantization of AGT levels. Likewise, this same inverse relationship between AGT levels and response was seen in patients with malignant glioma treated with preradiation temozolomide.¹⁷ Furthermore, Hegi et al¹⁸ have shown a relationship between inactivation of the AGT gene by promoter methylation and survival in patients with newly diagnosed GBM treated with surgery, radiotherapy, and temozolomide. These studies suggest that the efficacy of temozolomide and carmustine could be enhanced by depletion of tumor AGT.

Depletion of tumor AGT activity by a selective inhibitor, O⁶-BG, enhances the cytotoxicity of chloroethylators and methylators including temozolomide and carmustine.^{1,4,7,17,19–21} However, the main limitation in systemic administration of alkylating agents in combination with O⁶-BG is their potential for dose-related toxicity to the hematopoietic system. This was seen in animal studies^22–24 and confirmed in a phase I clinical trial,²⁵ where systemic administration of O⁶-BG and carmustine markedly enhanced myelosuppression and reduced the carmustine MTD from 200 to 40 mg/m², representing an 80% dose reduction. This profound reduction in the carmustine dose was most likely the underlying factor in the failure of this drug combination to cause frank tumor regression in the phase II trial.²⁶

Given that temozolomide is inherently less toxic than carmustine, particularly to hematopoietic cells, it was hypothesized that perhaps temozolomide in combination with O⁶-BG would produce less myelosuppression than carmustine in combination with O⁶-BG. Thus, a phase I clinical trial was performed in patients with malignant glioma to determine the MTD of single-dose temozolomide in combination with O⁶-BG.⁹ When combined with O⁶-BG, the MTD of temozolomide was found to be 472 mg/m². The requirement for an 80% dose reduction seen with carmustine when combined with O⁶-BG was seen to a lesser extent (50%) with temozolomide when combined with O⁶-BG.

Thus, one of the aims of this study was to evaluate the safety of administering O⁶-BG in combination with temozolomide, especially to the hematopoietic system. As expected, myelosuppression was the most commonly reported AE. Grade 4 hematologic events were experienced in 48% of the patients for whom a 25% temozolomide dose reduction was required for subsequent dose administrations. The most frequent hematologic AEs were grade 4 neutropenia (47%) followed by grade 4 thrombocytopenia (12%).

Another aim of this study was to identify the role of O⁶-BG in restoring temozolomide sensitivity in patients with recurrent or progressive temozolomide-resistant malignant glioma. Unfortunately, the lack of a clinical rationale to justify a biopsy at recurrence precluded quantitation of tumor AGT levels, which could have provided insights if other non-AGT mechanisms of resistance were operational. We determined that O⁶-BG when added to a 1-day dosing regimen of temozolomide was able to restore temozolomide sensitivity in patients with temozolomide-resistant anaplastic glioma, but there seemed to be no significant restoration of temozolomide sensitivity in patients with temozolomide-resistant GBM.

Despite the encouraging response to temozolomide and O⁶-BG in patients with anaplastic glioma, it is somewhat disappointing that a greater response and improvement in PFS and OS were not realized in GBM, especially if AGT was indeed the primary mechanism responsible for tumor resistance to temozolomide. Although O⁶-BG in combination with temozolomide was able to reverse the resistance in a minority of patients with recurrent or progressive, temozolomide-resistant malignant glioma, it was unable to reverse this resistance in the majority of these patients. The explanation for the inability of this combination to reverse resistance may lie in the two following reasons.

First, the dose reduction in temozolomide necessary to avoid hematologic toxicity when combined with O⁶-BG may be an underlying factor in the failure of this drug combination to restore sensitivity. As a 1-day dosing regimen, temozolomide required a 50% dose reduction when delivered with O⁶-BG. Alternative routes of administration of either O⁶-BG or alkylators have been investigated in hopes of reducing systemic toxicity while maximizing efficacy. Administration of O⁶-BG locally by intracerebral infusion through an Ommaya reservoir with concomitant oral administration of temozolomide was attempted in one patient.²⁷ This therapy was well tolerated and caused tumor stabilization for four months, but its success at depleting AGT and thus prolonging survival above and beyond the effect of temozolomide alone is unknown at this time. The administration of carmustine locally in the form of polifeprosan 20 with carmustine implant in combination with systemic O⁶-BG was performed in a phase I⁸ and a phase II clinical trial.²⁸ Both of these trials support the idea that local administration of carmustine in the form of polifeprosan 20 with carmustine implant in combination with systemic O⁶-BG may mitigate the systemic toxicities of alkylators with improved efficacy.²⁸

Second, combining O⁶-BG with a 5-day dosing regimen of temozolomide may be more effective than a 1-day dosing regimen, especially if the total dose of temozolomide can be increased. To optimize a 5-day dosing regimen of temozolomide in combination with O⁶-BG, one could hypothesize that suppressing AGT beyond 48 hours for at least 5 days while administering temozolomide may improve efficacy. Because we know that the above O⁶-BG regimen completely suppresses O⁶-BG for at least 48 hours, perhaps repeating this exact regimen every 48 hours for a period of 5 to 7 days may optimize AGT suppression while administering a 5-day dosing regimen of temozolomide. Although other O⁶-BG regimens have been used in combination with temozolomide²⁹ and polifeprosan 20 with carmustine implant,⁸ there is no evidence to suggest that these dosing regimens are adequate to suppress AGT levels in brain tumors.

Although the current trial of O⁶-BG when added to a 1-day dosing regimen of temozolomide was able to restore temozolomide sensitivity in patients with temozolomide-resistant anaplastic glioma, no significant restoration of temozolomide sensitivity was seen in patients with temozolomide-resistant GBM. The limited response seen in GBM propels the research in a direction that explores alternative dosing regimens of temozolomide and O⁶-BG and alternative administration routes for chemotherapy. Thus, additional investigation in combining temozolomide and O⁶-BG in a 5-day dosing schedule needs to be undertaken in a phase I trial. Also, additional exploration of the safety and efficacy of combining polifeprosan 20 with carmustine implant with O⁶-BG should be pursued. In addition, the role of drug delivery needs to be addressed, given that it is possible that potentially higher interstitial pressure seen in GBMs compared with anaplastic gliomas may be limiting drug delivery. This question might be addressed in future studies by incorporation of dynamic contrast-enhanced MRIs to evaluate tumor blood flow and potentially drug delivery.

Footnotes

REFERENCES