Antitumor Activity of Temozolomide Combined with Irinotecan Is Partly Independent of O⁶-Methylguanine-DNA Methyltransferase and Mismatch Repair Phenotypes in Xenograft Models¹

Temozolomide is a methylating agent that has been approved for treatment of astrocytoma and is entering various phases of clinical evaluation against other tumors. Phase II trials in Europe have also confirmed some activity against melanoma (1) and suggest activity against high-grade gliomas (2). Temozolomide is considered to exert its toxic effects primarily by generating O⁶-methylguanine in DNA (3, 4). This adduct is subject to a single-step, error-free repair reaction that simply transfers the methyl group to a cysteine residue within the repair protein MGMT,³ thus restoring the DNA to its intact state. Hence, MGMT is a major determinant of temozolomide cytotoxicity (4, 5). Furthermore, intact MMR function is critically required for the cytotoxicity of the methylating drugs. Recently, we reported the sensitivity of a series of pediatric tumor xenografts to temozolomide. Sensitivity correlated with MGMT deficiency and MMR proficiency (6).

CPT-11 is a camptothecin prodrug activated by carboxylesterases to the active topoisomerase I poison SN-38. CPT-11 has demonstrated broad activity against both murine and human tumor xenograft models (reviewed in Ref. 7) and clinically significant activity against many types of cancer (reviewed in 8, 9). Camptothecins have been reported to synergize with ionizing radiation (10, 11) and chemical agents that damage DNA, including platinum and alkylating agents (12, 13, 14). Potentially, modification to DNA can lead to recruitment of topoisomerase I, thus potentially increasing the probability of a camptothecin drug stabilizing the DNA-enzyme covalent complex (15, 16). Because topoisomerase I preferentially cuts DNA between T and G residues, we speculated that methylation of O⁶-guanine would lead to recruitment of topoisomerase I and potentially enhance the probability of inducing camptothecin-mediated damage. This formed the biochemical rationale for combining temozolomide with a camptothecin. Recently, Pourquier et al. (17) demonstrated that O⁶-alkylation of guanine induces topoisomerase-I DNA covalent complexes in N-methyl-N′-nitro-N-nitrosoguanidine-treated cells. Conceptually, this would increase the probability of a collision with the advancing replication fork and generation of a double-strand DNA break, considered the initiating event in inducing cell death(reviewed in Ref. 18).

In addition to the biochemical rationale for the interaction between temozolomide and CPT-11, in clinical trials these agents have relatively nonoverlapping toxicities. The limiting toxicity of temozolomide is noncumulative, transient myelosuppression (19), whereas when CPT-11 is given as protracted daily dosing, the limiting toxicity is primarily diarrhea (20). Here we report the significant activity of CPT-11 in combination with temozolomide. Of interest is the finding that the combination of each agent at dose levels that as monotherapy have minimal antitumor activity resulted in complete regressions of several tumors. This occurred in tumors that were MGMT proficient and MMR deficient,suggesting that the interaction between these agents may in part be independent of temozolomide-induced O⁶-methylation of guanine. In the companion paper, Patel et al. (21) have examined the sequence dependence of this combination in brain tumor xenografts.

Each of the xenografts used has been described previously (22, 23, 24). Studies used four lines of neuroblastoma, three rhabdomyosarcomas, and one glioblastoma. Tumors were grown in the s.c. space of immune-deprived, female CBA/CaJ mice, as described (25). Previously, we have reported the MGMT, MMR, and p53 phenotype of each tumor and related this to temozolomide sensitivity (6). The phenotype determined from tumor tissue for each tumor is summarized in Table 1.

For individual tumors, partial response was defined as a volume regression >50% but with measurable tumor (≥0.10 cm³) at all times. CR was defined as a disappearance of measurable tumor mass (<0.10 cm³) at some point within 12 weeks after initiation of therapy. A maintained complete response was CR without tumor regrowth within a 12-week study time frame. Methods for statistical analysis of data, and evaluation of tumor responses have been reported previously (25). Animal care was in accord with institution guidelines.

Initial studies were designed to determine the optimal schedule for administration of temozolomide and to determine any schedule-dependent antitumor activity. Tumor-bearing mice were treated with temozolomide by oral gavage for 5 days (days 1–5) or two 5-day courses (1–5 and 8–12) per 21-day cycle. Alternatively, mice received three 5-day courses per 28-day cycle. The cumulative dose in all groups was 630 mg/kg. In subsequent combination studies, temozolomide was administered daily for 5 consecutive days (days 1–5) of each cycle,because this was found optimal in initial studies. Temozolomide was administered 1 h prior to administration of CPT-11. Cycles of therapy were repeated twice at 21-day intervals. CPT-11, at doses listed for individual experiments, was administered i.v. daily for 5 consecutive days for 2 consecutive weeks, shown previously to be an optimal schedule (days 1–5 and 8–12). Temozolomide and CPT-11 were generously supplied by Schering-Plough and Upjohn-Pharmacia,respectively.

We conducted pharmacokinetic studies of the combination of temozolomide and CPT-11 to determine whether a pharmacokinetic interaction existed between the two drugs. Temozolomide was administered as a single oral dose (66 mg/kg), followed by a single i.v. dose of irinotecan (10 mg/kg). To measure temozolomide and MTIC, blood samples were collected from mice (three animals/point) at 0, 0.25, 0.5, 1, 1.5, 2, 3, and 6 h. Samples were immediately centrifuged at 5.5 × g for 2 min in a tabletop refrigerated centrifuge at 4°C. Plasma was then divided into aliquots for processing to assay either temozolomide or MTIC by isocratic high-performance liquid chromatography as described previously in detail (6). Temozolomide and MTIC were quantitated by UV detection at 325 and 318 nm, respectively. The lower levels of quantitation for temozolomide and MTIC were 0.25 and 0.5 μg/ml, respectively. All calibrators and controls were prepared in murine plasma (Hill Top Lab Animals, Inc.,Scottdale, PA).

The disposition of irinotecan and SN-38 was evaluated after administration of a single oral dose of temozolomide (66 mg/kg),followed by a single i.v. dose of irinotecan (10 mg/kg). Heparinized blood samples (∼1 ml) were collected (three animals/time point) pre,0.25, 0.5, 1, 2, 4, and 6 h after i.v. irinotecan administration. Samples were immediately centrifuged at 5.5 × g for 2 min. Plasma was separated, and proteins were precipitated by the addition of 200 μl of plasma to 800 μl of cold methanol (−30°C),followed by vigorous agitation with a vortex mixer, and centrifuged again at 5.5 × g for 2 min. The supernatant was decanted and stored at −70°C until analysis.

Irinotecan and SN-38 lactone plasma concentrations were determined by an isocratic high-performance liquid chromatography assay with fluorescence detection, as described previously in detail (24, 26). The excitation and emission wavelengths were 375 and 520 nm, respectively. The lower level of quantitation was 1 ng/ml for irinotecan and SN-38. All calibrators and controls were prepared in murine plasma (Hill Top Lab Animals, Inc.).

Temozolomide and MTIC plasma concentration-time data from oral administration were modeled using maximum likelihood estimation as implemented in ADAPT II (27). A first-order absorption,one-compartment linear model, which included first-order MTIC formation and elimination, was used to simultaneously describe temozolomide and MTIC disposition (28). AUC was calculated from the model parameters. The apparent time of maximum concentration(t_max) and maximum plasma concentration (C_max) were also noted.

The disposition of irinotecan and SN-38 was evaluated using a three-compartment model with linear distribution and elimination (29). Pharmacokinetic parameters for each set of data were initially fit by maximum likelihood estimation, as implemented in ADAPT II. Pharmacokinetic parameters calculated from these estimates included systemic clearance (CL) and AUC. The maximum observed irinotecan and SN-38 plasma concentrations(C_max) and time to maximum plasma concentration (T_max) were determined.

Previously, we and others have shown that the antitumor activity of camptothecins is highly schedule dependent. For example, the same total dose of CPT-11 given over 10 days was significantly more active than when administered over 5 days or as a single administration. To determine whether there was similar schedule dependency for temozolomide, we examined the antitumor activity of this agent given daily for 5 days, 2 × 5 days per 21-day cycle or 3 × 5 days per 28-day cycle. The total dose per cycle was constant. Results for NB-1382 xenografts are presented in Fig. 1and demonstrate the most pronounced schedule-dependent activity of temozolomide. In this experiment, mice received a cumulative dose of 210 mg/kg per cycle. Mice received three cycles of treatment (total cumulative dose was 630 mg/kg). When temozolomide was given over 5 days(42 mg/kg/dose), CRs were achieved in all mice and were maintained at week 12. Lower doses given over more protracted periods were progressively less effective. Far less schedule-dependent antitumor activity was observed in five other xenograft lines (data not shown);however, in all experiments, administration of temozolomide in the most intensive schedule (daily for 5 consecutive days/cycle) was either more active or equally active compared with the other schedules. Consequently, for combination studies we selected the daily × 5 schedule for combination with CPT-11.

Dose levels of CPT-11 and temozolomide were chosen so that neither drug alone caused CR. For CPT-11 dose, levels between 2.5 and 0.18 mg/kg daily were administered, determined by the intrinsic sensitivity of any particular xenograft line. These dose levels give SN-38 systemic exposures consistent with those achieved in patients receiving CPT-11 using the same schedule of administration (20). Temozolomide was administered at dose levels ranging from 66 to 19 mg/kg and are consistent with doses that in patients give achievable levels of parent drug and MTIC the active metabolite (6). Statistical analysis of the antitumor activity of single agents and combination treatments are summarized in Table 2. The activity of combinations was significantly better than the activity of either single agent used at the same dose level, with a few exceptions. For example, in studies where monotherapy with either CPT-11 or temozolomide at the doses used caused maintained CR, it was not possible to determine whether combinations were superior to monotherapy(e.g., Rh30 and NB-1643). However, against several tumor lines temozolomide combined with CPT-11 demonstrated significant activity against tumors at dose levels that had little activity when administered as monotherapy. For example, against NB-SD neuroblastoma xenografts, temozolomide had little activity at tolerated dose levels,and similarly tumors progressed in mice treated with CPT-11 at 0.4 mg/kg (Fig. 2). In combination these agents induced CR that was maintained at week 12. The glioblastoma line, SJ-GBM2, has a similar phenotype to NB-SD, except in this tumor MGMT is not detected (see Table 1). Although temozolomide induced some partial responses and an occasional CR during the 8 weeks of treatment at 66 or 42 mg/kg, the higher dose level was lethal during cycle 3 of treatment (Fig. 3). At 42 mg/kg, the overall effect of treatment was stasis, because at the end of treatment tumor volumes were similar to that at initiation of temozolomide. CPT-11 combined with temozolomide (66 or 42 mg/kg) resulted in CR of all tumors with no tumor regrowth during the period of observation (12 weeks). Furthermore, combination with CPT-11 decreased the toxicity of temozolomide (described below). NB-1771 tumors have detectable MGMT and appear to be MMR proficient. These tumors were relatively sensitive to temozolomide as a single agent (6); hence for this study the dose was reduced to 19 mg/kg. CPT-11 induced some CRs at 1.25 mg/kg and an occasional CR at 0.61 mg/kg; however, at week 12 all tumors had progressed. As shown in Fig. 4,temozolomide combined with CPT-11 resulted in CRs of all tumors. In groups of mice receiving the higher dose of CPT-11, there was a single tumor that regrew, whereas at the lower dose, four tumors regrew during the period of observation. Rh18 rhabdomyosarcoma expresses high MGMT levels and is deficient in MLH1 expression. Consequently, this xenograft is poorly sensitive to temozolomide as a single agent (Fig. 5 ). Ineffective doses of temozolomide combined with doses of CPT-11, which alone induced few CRs, resulted in a high frequency of CRs.

To determine whether the antitumor activity observed with the drug combination could be a result of one agent altering the systemic exposure to the other, detailed pharmacokinetic studies were performed. Temozolomide and MTIC concentrations exceeded the limit of assay sensitivity for the duration of the study. After administration, the apparent t_max was 30 min for temozolomide and 1 h for MTIC. The C_max for temozolomide and MTIC were 36 and 0.8 μg/ml, respectively. The plasma AUC_0→∞ for temozolomide and MTIC were 47 and 1.3 mg/L-hr, respectively. Irinotecan and SN-38 concentrations exceeded the limit of assay sensitivity for the duration of the study. After administration, the apparent t_max was 15 min for both irinotecan and SN-38. The C_max for irinotecan and SN-38 was 764 and 192 μg/ml, respectively. The plasma AUC_0→∞ for irinotecan and SN-38 was 993 and 463 mg/L-hr, respectively.

CPT-11 has demonstrated significant activity against various human cancers including refractory pediatric neoplasms (20). In vitro, camptothecins have demonstrated synergy with various DNA-damaging agents, including alkylating agents, cisplatin,and ionizing radiation. In vivo, 9-aminocamptothecin acts as a radiation sensitizer, and a single report indicates synergy between topotecan and temozolomide (30). Recently, we reported the activity of temozolomide in xenografts where the MGMT and MMR status had been determined (6). In the current study, we evaluated the antitumor activity of temozolomide, a DNA methylating agent, combined with CPT-11, the prodrug form of SN-38, a topoisomerase I poison. In part, the rationale for this combination was the nonoverlapping toxicities of the two agents. Dose limiting toxicity of temozolomide is cumulative myelosuppression, notably neutropenia,whereas in the schedule used here, that of CPT-11 is diarrhea with only moderate myelosuppression in children (20).

Previously, we and others have shown that the antitumor activity of several camptothecin drugs is highly schedule dependent (Ref. 22; reviewed in Ref. 7). We were interested,therefore, in initially determining whether there was a similar schedule dependency for the antitumor activity of temozolomide. Tumor-bearing mice were administered temozolomide daily for 5 days for 1, 2, or 3 consecutive weeks (i.e., days 1–5 or days 1–5 +8–12 repeated for three cycles at 21-day intervals or days 1–5 +8–12 + 15–19 repeated at 28-day intervals). In each schedule, the total cumulative dose/cycle was constant. Temozolomide demonstrated only moderate schedule-dependent antitumor activity, but in all experiments the most intensive schedule (daily for 5 days every 21 days) was either more effective or equally effective with other schedules of drug administration. Thus, for combination studies with CPT-11, temozolomide was given for 5 consecutive days at the start of each cycle of CPT-11 treatment.

Dose levels of CPT-11 were used that in mice give systemic exposures to SN-38 consistent with achievable exposures in children receiving CPT-11 using the same schedule (20). The highest dose of temozolomide (66 mg/kg) used also gives clinically relevant drug exposures (6). For most studies, the combinations had significantly greater activity than either agent administered as monotherapy at the same dose level. Exceptions were where either or both agents were effective at inducing maintained CRs. Examples were Rh30 and NB-1643, where at the dose levels administered, monotherapy resulted in CRs maintained at week 12. For NB-SD, NB-1771, SJ-GBM2, and Rh18 xenografts, combinations evaluated were significantly more effective than monotherapy. Against tumors such as Rh18 that express relatively high levels of MGMT, temozolomide induced few regressions and caused relatively little growth inhibition. However the combination resulted in more CRs than either drug given as monotherapy. The glioblastoma, SJ-GBM2, is also relatively refractory to temozolomide. Although this tumor is deficient in MGMT, it has barely detectable levels of the MMR protein MLH1. At the highest tolerated dose of temozolomide (42 mg/kg), there were tumor regressions, but at the end of treatment (week 8) tumors were similar in mass to the pretreatment values. Furthermore, there were no maintained CRs. In contrast,temozolomide combined with CPT-11 was less toxic and resulted in maintained CRs for all animals in groups receiving 66 or 42 mg/kg temozolomide combined with an ineffective dose level of CPT-11 (1.25 mg/kg). Similar results were obtained against NB-SD neuroblastoma xenografts that are resistant to temozolomide. Furthermore, all NB-1771 neuroblastoma xenografts in mice treated with the combination had CR,although some tumors had regrown by week 12. Thus, for several tumors the combination of CPT-11 with temozolomide induced superior tumor responses than either agent alone against tumors that were either MGMT proficient or MMR deficient and irrespective of wild-type or mutant p53. These results suggest that the interaction between CPT-11 and temozolomide may be, in part, independent of O⁶-methylation of guanine, the primary mechanism considered to lead to cytotoxicity of temozolomide. In addition to methylation of O⁶-guanine,temozolomide methylates N⁷-guanine or N³-adenine. These are the predominant sites of modification by temozolomide. Potentially, recruitment of topoisomerase I to DNA may be facilitated through such modifications,although this remains to be tested. Of interest also was the observation that toxicity-related death was consistently lower in groups of mice treated with temozolomide combined with higher dose levels of CPT-11.

The results of our pharmacokinetic analyses showed no difference in temozolomide or irinotecan disposition when the two agents are coadministered. We have previously studied the disposition of single-agent temozolomide and MTIC in the xenograft model after a single oral dose of 66 mg/kg (6). The temozolomide and MTIC AUC_0→∞ were 40μg/ml·hr and 1.9 μg/ml·hr, respectively, values very similar to those observed in this study when temozolomide was coadministered with irinotecan. These results are not surprising, especially because temozolomide is metabolized through nonenzymatic, pH-dependent hydrolysis (31). Likewise, the AUC_0→∞ for irinotecan and SN-38 was similar to those values reported by us previously after single-agent irinotecan (26, 32). Thus, the enhanced antitumor activity of the combination is unlikely attributable to a drug interaction between the two agents.

In summary, combination of temozolomide and CPT-11 administered on optimal schedules is effective in inducing CR in a series of xenografts derived from childhood solid malignancies. Taken together with results from pediatric brain tumor xenografts in the companion paper by Patel et al. (21), clinical evaluation of this combination may be of interest.

We thank Lorina Dudkin for technical assistance that contributed to part of this study.