O⁶-Benzylguanine-mediated Enhancement of Chemotherapy 1 This work was supported by NIH Grants NS30245 (to H. S. F.), NS20023 (to H. S. F), CA57725 (to A. E. P., H. S. F., and M. E. D.), and CA81485 (to M. E. D.). Drs. Pegg, Moschel, and Dolan have a financial relationship with Access Oncology, the company that is presently licensing O⁶-benzylguanine. Dr. Friedman is a paid consultant for Access Oncology. ¹

Introduction

Several preclinical and clinical studies have demonstrated that the DNA repair protein AGT³ is responsible for resistance to chlorethylation or methylation damage at the O⁶ position of guanine in DNA (1–15), and several studies have shown that this enhances BCNU, temozolomide, and cyclophosphamide resistance. O⁶-BG inactivates AGT activity both in vitro and in vivo (16–27). More recently, we have demonstrated that the combination of temozolomide plus CPT-11 displays a schedule-dependent enhancement of antitumor activity secondary to the trapping of topoisomerase I by O⁶-methylguanine residues in DNA (28, 29). These studies suggested that there might be favorable therapeutic interactions between O⁶-BG, and combinations of BCNU plus cyclophosphamide or temozolomide plus CPT-11, respectively. We now report the marked increase in antitumor activity produced by these combinations after O⁶-BG- mediated AGT inactivation.

Materials and Methods

Animals.

Male and female athymic BALBc mice (nu nu genotype, 6 weeks of age or older) were used for all of the studies and were maintained as described previously (30). Athymic mice were housed in an isolation facility with high-efficiency particulate air-filtration. All of the mice were maintained under a controlled 12-h light/12-h dark cycle and were provided food and water ad libitum.

Xenografts.

Human central nervous system tumor- derived xenografts were used for in vivo studies. The xenografts were maintained as described previously (31). The derivation and AGT status of these xenografts is detailed in Table 1 from a previous publication (25).

Drugs.

Temozolomide was provided by Schering-Plough Research Institute (Kenilworth, NJ). CPT-11 was provided by Pharmacia and Upjohn Global Distribution Center (Kalamazoo, MI). Cyclophosphamide and BCNU were purchased from Sigma (St. Louis, MO). O⁶-BG was synthesized as described previously (32).

s.c. Xenograft Transplantation.

s.c. tumor transplantation was performed into the right flank of the animals with an inoculation volume of 50 μl using a brei prepared from xenografts (33).

Tumor Measurements.

Tumors were measured twice weekly with hand-held vernier calipers (Scientific Products, McGraw, IL). Tumor volume was calculated according to the following formula: V = [(width)² × (length)]/2.

Drug Toxicity.

Mice were weighed twice weekly to assess weight loss and were checked daily for survival.

Xenograft Therapy.

Groups of 10 randomly selected mice were treated when the median tumor volume was in the range of 100–500 mm³ and were compared with control animals receiving drug vehicle. Cyclophosphamide was administered via i.p. injection at a single dose of 300 mg/m² (100 mg/kg) in 0.9% saline, which represents 22% of the LD₁₀. BCNU was administered via i.p. injection at a single dose of 24.0 mg/m² (8 mg/kg) in 5% ethanol, which represents 23.8% of the LD₁₀. Temozolomide was administered via i.p. injection at a single dose of 105 mg/m² (35 mg/kg) in 0.9% saline, which represents 10% of the LD₁₀. CPT-11 was administered via i.p. injection at a dose of 12 mg/m² (4 mg/kg) in 0.9% saline on days 1–5 and 8–12, which represents 10% of the LD₁₀. O⁶-BG was administered at a dose of 240 mg/m² (80 mg/kg) in polyethylene glycol-400/0.9% saline (4:6) 1 h before any of the other agents. At this dose, O⁶-BG depletes tumor AGT levels by >90% within 1 h.

In additional studies, groups of five non-tumor-bearing mice were treated with the combination of temozolomide at 25% of the LD₁₀ plus CPT-11 at 25% of the LD₁₀ given alone or with O⁶-BG. Similar studies with five non-tumor-bearing mice/group using temozolomide at 50% of the LD₁₀ plus CPT-11 at 50% of the LD₁₀ given alone or with O⁶-BG were also conducted.

Assessment.

The response of the s.c. xenografts was assessed by delay in tumor growth and by tumor regression. Growth delay, expressed as T − C, is defined as the difference in days between the median time required for tumors in treated (T) and control (C) animals to reach a volume five times greater than that measured at the start of treatment. Tumor regression is defined as a decrease in tumor volume over two successive measurements. Statistical analyses were performed using a personalized SAS statistical analysis program, the Wilcoxon rank order test for growth delay, and Fisher’s exact test for tumor regression as described previously (34).

Results

Toxicity

The toxic deaths and mean nadir weight loss produced by the individual agents and their combinations in tumor-bearing mice are indicated in Tables 2 and 3. No treatment arm displayed more than one toxic death in 10 treated mice. The combination of temozolomide (0.25 LD₁₀) plus CPT-11 (0.25 LD₁₀), with or without O⁶-BG, produced no toxic deaths. The combination of temozolomide (0.50 LD₁₀) plus CPT-11 (0.5 LD₁₀) produced two of five deaths with O⁶-BG and two of five deaths without O⁶-BG.

Activity

Cyclophosphamide/BCNU/O⁶-BG.

The combination of cyclophosphamide plus O⁶-BG produced increased antitumor activity in five AGT-positive tumors, compared with cyclophosphamide alone (Table 2). Cyclophosphamide produced growth delays of 8.0, 12.0, 4.6, 15.3, and 16.3 days, whereas cyclophosphamide plus O⁶-BG produced growth delays of 9.5, 21.9, 13.1, 18, and 17.7 days against d-245 MG, d-341 MED, d-456 MG, d-528 EP, and d-612 EP, respectively. The addition of O⁶-BG to BCNU increased growth delays of 11.7, 1.0, 2.0, 6.5, 9.2, and 2.3 days with BCNU alone to 24.7, 20.2, 14.4, 42.5, and 45.6 days with BCNU plus O⁶-BG. The combination of BCNU plus cyclophosphamide produced growth delays of 23.9, 13.3, 12.0, 10.8, and 32.7 days. The use of BCNU, cyclophosphamide, and O⁶-BG together produced the longest growth delays of 46.6, 60.1, 70.4, 49.0, and 55.4 days.

Temozolomide/CPT-11/O⁶-BG.

The administration of temozolomide plus O⁶-BG produced an increased antitumor activity in the AGT-positive xenograft D-456 MG compared with temozolomide alone (Table 3). Temozolomide (0.1 LD₁₀) produced growth delays of 7.9 and 2.9 days, whereas temozolomide (0.1 LD₁₀) plus O⁶-BG produced growth delays of 11.3 and 7.3 days against d-456 MG. CPT-11 (0.1 LD₁₀) produced growth delays of 23.7 and 31.1 days, with the addition of only O⁶-BG changing the growth delays to 26.1 and 20.4 days. The combination of temozolomide (0.1 LD₁₀) plus CPT-11 (0.1 LD₁₀) produced the expected marked increase in antitumor activity, with growth delays of 43.2 and 37.0 days. However, the combination of temozolomide (0.1 LD₁₀), CPT-11 (0.1 LD₁₀), and O⁶-BG produced apparent cures with growth delays of >150 and >150 days. The use of CPT-11 (1.0 LD₁₀) or temozolomide (1.0 LD₁₀) also produced growth delays of >150 days and >150 days.

Discussion

The use of O⁶-BG-mediated AGT depletion to augment the antineoplastic activity of chemotherapeutic agents remains an exciting possibility. The combination of BCNU plus O⁶-BG, a very effective strategy in vitro and in rodents (16–24, 26, 27) has not yet produced similar results in humans, presumably because of the marked reduction of BCNU dose required to deliver this regimen safely. The dose of BCNU when given with O⁶-BG must be reduced by 80%, from 200 to 40 mg/m², because of dose-limiting myelosuppression (35). The use of O⁶-BG plus this reduced dose of BCNU was not effective against BCNU-resistant glioblastoma multiforme (36). Nevertheless, AGT depletion is a very enticing strategy if rational and safe regimens can be designed.

The current laboratory studies focused on two different strategies to exploit O⁶-BG-mediated AGT depletion. BCNU and cyclophosphamide are frequently administered in combination, particularly in high-dose chemotherapy regimens using stem cell support. Our results confirm previous work in our laboratory that demonstrated modestly enhanced activity of cyclophosphamide and substantially enhanced activity of BCNU produced by O⁶-BG (21, 37). The concomitant use of three agents produced markedly enhanced antitumor activity, even compared with the combination of cyclophosphamide and BCNU. A precise molecular explanation for this interaction remains somewhat speculative. The interaction between BCNU and O⁶-BG is clearly a result of reduced removal of BCNU-induced mono-adducts with higher formation of the lethal DNA interstrand cross-link (11). The interaction between cyclophosphamide and O⁶-BG is more complex.

Evidence has emerged recently that suggests a role for AGT in the repair of certain cyclophosphamide-induced lesions. Cyclophosphamide is metabolized to acrolein and phosphoramide mustard. The antitumor effect is thought to be mainly mediated by interstrand cross-links formed from the reaction of phosphoramide mustard and DNA (38). We demonstrated that AGT-expressing CHO cells were significantly less sensitive to the toxic and mutagenic effects of both 4-HC (activated form of cyclophosphamide) and 4- hydroperoxydidechlorocyclophosphamide (4-HDC; a generator of acrolein and a nonalkylating form of phosphoramide mustard) than CHO cells without detectable AGT (37, 39). However, neither the toxic nor the mutagenic effects of phosphoramide mustard were altered in the presence or absence of AGT. These results suggest that inactivation of AGT by O⁶-BG in AGT-expressing tumors is likely to result in an increase in acrolein-induced lesions in DNA and unlikely to impact on antitumor activity produced by phosphoramide mustard. This is consistent with results shown in Table 2, in which there may be a slight advantage, albeit not dramatic, of AGT depletion (O⁶-BG) on tumor growth of animals treated with cyclophosphamide.

Interestingly, we observe a more dramatic tumor growth inhibition when O⁶-BG is combined with BCNU and cyclophosphamide. In D-245 MG, D-341 MED, and D-456 MG, the growth delay with the three-drug regimen was more than the additive effect of O⁶-BG with cyclophosphamide or O⁶-BG with BCNU. The critical toxic lesion formed in DNA by BCNU is the 1-(3-cytosinyl)-2-(1-guanyl)ethane interstrand cross-link after chloroethylation at the O⁶ position of guanine (40). Cyclophosphamide produces interstrand N7-N7 cross-links involving the two guanines in GNC•GNC (5′→3′/5′→3′) sequences (41). We have observed that, in some cell lines deficient of AGT, O⁶-BG enhances the toxicity of 4-HC (42). The mechanism of enhancement is unrelated to AGT inactivation because the known toxic DNA lesions associated with nitrogen mustards occur at nucleophilic nitrogens of guanine, not the O⁶ position of guanine, and there is no evidence to suggest that AGT can repair these lesions. In addition, enhancement is observed in cell lines such as CHO and SQ20b (squamous cell carcinoma) that are apparently devoid of the AGT protein. ⁴ Preliminary results from our laboratory indicate that the formation or processing of intrastrand and/or interstrand DNA cross-links is critical for the interaction of O⁶-BG with cyclophosphamide (43). The dramatic growth delay when O⁶-BG is combined with cyclophosphamide and BCNU could be attributable to the abundance of structurally different cross-links on DNA. O⁶-BG might act by two mechanisms: (a) inactivation of AGT, thus producing more BCNU cross-links and (b) a more elusive mechanism involving less repair of cyclophosphamide cross-links. Both of the interstrand cross-links formed on DNA may act in concert, resulting in greater tumor growth inhibition. Nevertheless, the concomitant use of the three agents produced marked increases in antitumor activity, which suggests that translation to the clinic, presumably in a program using stem cell support to offset dose-limiting myelosuppression, may be warranted.

The striking increases in antitumor activity seen with the combination of temozolomide, CPT-11, and O⁶-BG are better understood and may provide a rational combination for the clinic. Temozolomide resistance is produced, at least in part, by tumor AGT (44). However, it is also possible that altered drug pharmacokinetics may be contributing to the increase in antitumor activity. A current Phase I trial of temozolomide plus O⁶-BG will ultimately culminate in a Phase II trial of this combination in patients with temozolomide-resistant malignant glioma. Nevertheless, it is possible that dose-limiting toxicity will, similar to the BCNU plus O⁶-BG trial (36), minimize the dose and potentially limit the benefit of temozolomide when given in combination with O⁶-BG. Accordingly, we are also conducting a Phase I trial of CPT-11 plus temozolomide designed to use methylation of the O⁶ position of guanine produced by temozolomide to trap topoisomerase I and enhance CPT-11 activity. Unfortunately, AGT removal of the methyl adduct may minimize any therapeutic gain if it occurs before CPT-11-induced cytotoxicity.

Therefore, although a Phase II trial of CPT-11 plus temozolomide will be warranted, a design of strategies to reduce the impact of AGT-mediated methyl adducts before CPT-11 cytotoxicity may be necessary. Our current results suggest that the use of temozolomide (to methylate tumor cell DNA), O⁶-BG (to prevent AGT-mediated adduct removal), and CPT-11 (to target trapped topoisomerase I) may be a highly effective clinical intervention. A Phase I trial of this three-drug regimen is currently being designed that will reflect the schedule-dependent interaction between CPT-11 and temozolomide (28) and the need to deplete AGT before temozolomide-induced methylation.