Hematopoietic expression of O⁶-methylguanine DNA methyltransferase-P140K allows intensive treatment of human glioma xenografts with combination O⁶-benzylguanine and 1,3-bis-(2-chloroethyl)-1-nitrosourea

Introduction

Chloroethylnitrosourea (CENU) agents that alkylate the O⁶-position of guanine in DNA such as 1,3-bis-(2-chloroethyl)-1-nitrosourea (BCNU or carmustine), fotemustine, lomustine, and others are used primarily to treat brain cancer, melanoma, lymphoma, and gastrointestinal malignancies (1, 2). The cytotoxic action of CENUs involves formation of an O⁶-chloroethylguanine lesion that is capable of rearranging to form a lethal interstrand cross-link (3). The effectiveness of these agents, however, is limited by tumor overexpression of the DNA repair protein O⁶-methylguanine DNA methyltransferase (MGMT), which removes cytotoxic O⁶-alkylguanine adducts (3, 4). The mechanism by which MGMT removes the adducts is a stoichiometric reaction whereby the MGMT molecule is irreversibly inactivated, and the protein is then rapidly degraded. Thus, additional repair activity requires de novo protein synthesis (5).

Several strategies have been investigated to reverse drug resistance caused by MGMT that take advantage of the ability to temporarily deplete cellular MGMT activity to very low levels. O⁶-benzylguanine (BG) is a direct substrate of MGMT that potently and rapidly depletes MGMT activity in mammalian cells, including a wide variety of human tumor cell lines. Treatment of xenograft tumors with BG has also been shown to increase sensitivity to CENUs in xenograft human tumors (6). Due to the slow kinetics involved in cross-link formation following CENU treatment (6–18 h), optimal sensitization to these agents requires sustained MGMT depletion. We have previously demonstrated that optimal sensitization to BCNU requires continuous depletion of MGMT activity for 24 h post-BCNU administration (7). Recently, we developed a treatment that used a double-bolus BG administration that depletes >95% of cellular MGMT activity for 24 h in xenograft human glioma SF767 tumors (8). However, the main limitation of clinical treatments using combination BG and BCNU therapy is enhanced myeloid toxicity that has been seen in early human safety trials (9, 10). Hence, efforts have been undertaken to protect bone marrow cells against combination BG and BCNU chemotherapy using hematopoietic gene therapy strategies.

Several human mutant MGMT proteins have been shown to confer resistance to BG while retaining the ability to remove chloroethyl adducts. The identification of these mutants has enabled the development of new strategies to increase the therapeutic index of chloroethylating agents by using BG-resistant mutant MGMT cDNAs to protect hematopoietic cells. MGMT mutants P140A and G156A expressed and purified from Escherichia coli have shown a 40- and 240-fold higher resistance to BG inactivation compared to wild-type MGMT (wt-MGMT), respectively (11). Xu-Welliver et al. (12) have generated several additional MGMT mutant proteins, including the point mutant P140K that has been shown to be greater than 1000-fold more resistant to BG. All of these mutants showed a reduced rate of repair of methylated DNA substrates in vitro compared to human wt-MGMT by factors of 2.5 (P140A), 10 (P140K), and 25 (G156A). However, each mutant protected Chinese hamster ovary cells from BCNU-induced cytotoxicity to a similar level as wt-MGMT (13). Expression of either G156A or P140A has been shown to protect hematopoietic cells from combination BG and BCNU treatment in vitro (14, 15). Mice transplanted with MGMT mutant G156A-transduced bone marrow were also protected from BG and BCNU hematopoietic toxicity (16, 17). However, the G156A mutant shows severely reduced DNA repair kinetics in addition to being relatively unstable in mammalian cells compared to wt-MGMT and P140K (14, 18). Taken together, these data indicate that the P140K mutant exhibits the most favorable biochemical characteristics for hematopoietic protection strategies. The P140K mutant was also shown to be superior to the P140A mutant and wt-MGMT in protecting bone marrow from BG/BCNU toxicity in vivo (19).

In the current study, we have employed a secondary bone marrow transplant model using in vivo-selected MGMT-P140K-transduced murine hematopoietic cells to investigate high-dose combination BG/BCNU antitumor treatment. The results demonstrate that hematopoietic expression of MGMT-P140K protects bone marrow cells against BG/BCNU treatment, enabling dose-intensified treatment of xenograft tumors.

Materials and Methods

Cell Culture

The human glioma SF767 cell line was developed by The Brain Tumor Research Center, University of California at San Francisco, and provided by Dr. Eileen Dolan at the University of Chicago. Cells were cultured in α-MEM supplemented with 10% FCS (Hyclone, Logan, UT), penicillin (100 μg/ml), and streptomycin (100 μg/ml), and maintained at 37°C in a humidified incubator with an atmosphere of 5% CO₂/95% air. Medium and antibiotics were purchased from Life Technologies, Inc. (Baltimore, MD).

Retroviral Vectors and Producer Cell Lines

The murine stem cell virus (MSCV) retrovirus vector containing the P140K cDNA was generated as previously described (19). Briefly, the full-length P140K MGMT cDNA was subcloned into the EcoRI-XhoI restriction sites upstream of the encephalomyocarditis virus internal ribosome entry site (IRES) (Clontech, Palo Alto, CA) linked to the enhanced green fluorescent protein (eGFP) sequence (Clontech) into the MSCV retroviral vector (MSCV 2.1 was kindly provided by R. Hawley, Sunnybrooks Research Institute, Toronto, Ontario, Canada). Retroviral producer cell lines were generated by transfecting plasmid DNA into GP+E86 cells (20) in the presence of 8 μg/ml polybrene (Aldrich Chemical Co., Milwaukee, WI). Producer populations were then used to harvest virus supernatant after an overnight incubation at 32°C, filtered through a 0.45-μm filter, and stored at −80°C before use (21). Virus supernatant was titered on NIH/3T3 cells as described previously. Transmission of the full-length vector genome was confirmed by Southern blot analysis of infected and selected cells.

Analysis of GFP Expression by Flow Cytometric Analysis

For analysis of GFP expression, bone marrow and peripheral blood samples were first depleted of RBC using RBC lysis buffer (Gentra Systems, Minneapolis, MN) for 30 min on ice, then blocked with 10% normal rat serum (Caltag Laboratories, Burlingame, CA) in PBS for 10 min on ice. Cells were then incubated with one or two antibodies for 30 min on ice, washed once in 0.2% BSA in PBS, then analyzed on a FACScan flow cytometer (Becton Dickinson, San Jose, CA). Antibodies used were PE-conjugated GR-1 (RB6-8C5) and PE-conjugated B220 (RA3-6B2). Both antibodies were purchased from PharMingen (San Diego, CA) and used at the concentrations recommended by the manufacturer. The percentages of GFP-expressing granulocytes (GR-1) and B cells (B220) within the bone marrow and peripheral blood were determined by dividing the number of double-positive cells by the number of lineage-positive cells (corrected for background using isotype-specific control antibody).

Bone Marrow Transduction and Primary Transplantation

Non-obese diabetic/severe combined immunodeficient (NOD/SCID) mice were maintained in microisolator cages with sterile bedding, food, and water, and all animal protocols were approved by the Animal Use Committee of Indiana University School of Medicine. Bone marrow was harvested from femurs of 8- to 10-week-old NOD/SCID mice 48 h after treatment with 5-fluorouracil (50 mg/kg; SoloPak Laboratories, Franklin Park, IL). Harvested bone marrow cells were prestimulated for 48 h at 37°C in 5% CO₂ with 100 units/ml recombinant human interleukin 6 (Pepro Tech Inc., Rock Oaks, CA) and 100 ng/ml recombinant rat stem cell factor (Pepro Tech Inc.) in α-MEM (Life Technologies, Inc.) supplemented with 20% FCS (Hyclone). Retroviral transduction was performed on fibronectin fragment CH296 (RetroNectin; Takara Shuzo, Biotechnology Group, Otsu, Japan) at a concentration of 8 μg/cm² as described previously (22, 23) with a multiplicity of infection of 0.5 in the presence of cytokines. Over 48 h, the cells were incubated with viral supernatant twice for 4 h each. After infection, hematopoietic cells were harvested using cell dissociation buffer (Life Technologies, Inc.), and 5 × 10⁶ cells were injected via tail vein into sublethally irradiated mice [300 radiation-absorbed dose (RAD); ¹³⁷Cs source, Nordion International, Kanata, Canada].

In Vivo Selection and Secondary Transplantation

Four weeks after transplantation, the primary transplant mice were randomly assigned to two treatment groups: mice were given three weekly i.p. injections of 20 mg/kg BG (Sigma Chemical Co., St. Louis, MO) followed 1 h later with either 5 or 10 mg/kg BCNU (Sigma). BG was dissolved in 40% polyethylene glycol-400 (PEG-400)/60% saline solution (v/v), and BCNU was dissolved in 10% ethanol/90% normal saline (v/v) solution. BCNU was administered within 10 min of reconstitution. Mice were euthanized 4 weeks after the final BG/BCNU treatment, and the bone marrow was harvested from the femurs. The bone marrow cells were washed in cell dissociation buffer, filtered through a 0.45-μm cell strainer, and resuspended in Iscove's modified Dulbecco's medium (IMDM; Life Technologies, Inc.). Then, 5 × 10⁶ cells were injected via tail vein into sublethally irradiated (300 RAD) 8- to 10-week-old NOD/SCID mice.

Tumor Implantation and Treatment

One week after secondary bone marrow transplantation, mice were s.c. inoculated in the flank with 10 × 10⁶ SF767 cells suspended in 0.1 ml sterile HBSS containing 1 mm HEPES. When tumors were palpable (tumor volume of 200–300 mm³), the mice were randomly assigned to groups for treatment. BG and BCNU were administered separately by i.p. bolus injections. BG was dissolved in 40% polyethylene glycol-400 (PEG-400)/60% saline solution (v/v) at a concentration of 3.75 mg/ml; BCNU was dissolved in 10% ethanol/90% normal saline (v/v) solution at a concentration of 2.5 mg/ml. BCNU was administered within 10 min of reconstitution. Mice in the control group received i.p. injections of vehicle solutions. The high-dose combination BG/BCNU treatment consisted of a double-bolus BG regimen (30 mg/kg followed 8 h later by 15 mg/kg) and a single-bolus BCNU dose (10 mg/kg) administered 1 h after the first bolus dose of BG. Tumor volume measurements were taken three times weekly by calipers and calculated according to the formula (α² × β)/2, where α is the shorter and β is the longer of two dimensions. Tumors were measured until volumes exceeded 10 times the volume of the tumor at the initiation of treatment, at which time the animals were euthanized.

MGMT Activity Assay

Measurement of cellular MGMT activity is performed by a modification of the assay described by Wu et al. (24). Cellular extracts were prepared by resuspending each sample in 1 ml cold assay buffer [50 mm Tris (pH 8.0); 1 mm DTT; 1 mm EDTA; 5% glycerol] per 10 × 10⁶ washed cells. Cells were then pulse-sonicated on ice five times for 5 s each, and centrifuged at 14,000 × g at 4°C for 30 min. Protein content was quantitated using the Bradford protein assay. To measure MGMT activity, a double-stranded 18-bp oligonucleotide containing a single O⁶-methylguanine residue nested within a PvuII restriction site (Genosys Biotechnologies, United Kingdom) and a fluorometric 5′-hexachloro-fluorescein phosphoramidite (HEX) molecule was employed, as described previously (8). MGMT-mediated repair of the O⁶-methylguanine lesion is determined by the proportion of 10-bp HEX-labeled PvuII digestion cleavage product relative to the 18-bp fragment. The HEX-labeled oligonucleotide fragments are detected and quantitated using a Hitachi FMBIO II Fluorescence Imaging System (Hitachi Genetic Systems, South San Francisco, CA). MGMT activity was measured by incubating 0.2 pmol of the HEX-labeled oligonucleotide with 1–200 μg total cellular protein at 37°C for 2 h, followed by phenol-chloroform extraction to remove cellular protein and ethanol precipitation. The purified oligonucleotide was then digested with PvuII (Roche Molecular, Indianapolis, IN) and electrophoresed on a 20% denaturing polyacrylamide gel. MGMT specific activity (femtomoles O⁶-methylguanine removed/milligram protein) was calculated according to the following equation:

Statistical Analysis

The significance of each set of values was determined by the two-tailed t test, assuming equal variance. Data are presented as the mean ± SD.

Results

Transduction and in Vivo Selection of Murine Bone Marrow Cells

Bone marrow cells were obtained from 5-fluorouracil-treated NOD/SCID mice and transduced with the MSCV-P140K-IRES-eGFP retrovirus as described above. Because the bicistronic P140K retroviral vector expresses P140K MGMT and eGFP from the same transcript, GFP expression was used to quantitate the number of transduced cells in the bone marrow and peripheral blood. Before infusion, about 6.5 ± 0.4% of the bone marrow cells expressed GFP (data not shown). Transduced cells were infused into NOD/SCID mice, and the animals were treated with combination BG/BCNU. Previous studies showed that combination BG/BCNU treatment produces greater and more sustained P140K selection than BCNU alone (19). In addition, those studies also demonstrated that selection occurred at the stem cell and progenitor level, as indicated by secondary and tertiary transplantation (19). Thus, 4 weeks after transplantation, mice received three weekly treatments of 20 mg/kg BG and either 5 or 10 mg/kg BCNU (20/5 and 20/10, respectively). The mice were euthanized 4 weeks after the third treatment, and GFP expression in the bone marrow was determined by flow cytometric analysis. The 20/5 BG and BCNU treatment resulted in 65 ± 3% GFP+ cells, while the 20/10 treatment produced 84 ± 2% GFP+ cells in the bone marrow (data not shown). In the peripheral blood, both treatments produced slight neutropenia and significant thrombocytopenia (20/5: 2.6 ± 0.6 × 10³ neutrophils/μl, 426 ± 270 × 10³ platelets/μl; 20/10: 2.7 ± 0.5 × 10³ neutrophils/μl, 258 ± 161 × 10³ platelets/μl), compared to non-transplanted, untreated mice (3.9 ± 0.5 × 10³ neutrophils/μl, 1498 ± 336 × 10³ platelets/μl). Bone marrow harvested from these primary transplant recipients treated with three weekly cycles of 20/10 BG and BCNU was subsequently used for secondary transplants and tumor studies.

Tumor Response in Secondary Bone Marrow Transplant Recipients

BG/BCNU-resistant bone marrow from primary transplant recipients (about 5 × 10⁶ cells, averaging 80–85% GFP+) was transplanted into sublethally irradiated secondary NOD/SCID mouse recipients, as described in the methods. One week later, these mice were inoculated s.c. with human glioma SF767 xenograft tumors. Non-transplanted control animals were similarly irradiated and inoculated with tumor cells. When the tumors became palpable (average volume of 200–300 mm³, about 3 weeks), high-dose combination BG/BCNU therapy was initiated. Gross animal weight and tumor volume were monitored three times weekly for 28 days posttreatment. The maximum tolerated single BG/BCNU dose in non-transplanted NOD/SCID mice is 20 mg/kg BG + 10 mg/kg BCNU.² The high-dose antitumor treatment used in these studies consisted of a double-bolus BG regimen (30 and 15 mg/kg BG administered 8 h apart) and a single-bolus BCNU treatment (10 mg/kg administered 1 h after the first BG bolus dose). This dosing regimen was uniformly lethal to non-transplanted NOD/SCID mice (see results below).

Pilot Study Using One Cycle of High-Dose BG/BCNU Treatment

In the first experiment, one cycle of the high-dose BG/BCNU treatment described above was administered on Day 0 (about 4 weeks post-bone marrow transplantation), and tumor response was monitored for 28 days posttreatment. The transplanted mice were separated into two groups: one group received high-dose BG/BCNU treatment (n = 4); the second group received only vehicle solutions (n = 3). Non-transplanted animals were separated into two similar groups: one received high-dose BG/BCNU treatment (n = 3); the other vehicle only (n = 3). Peripheral blood samples were analyzed for GFP expression 3 weeks after bone marrow transplantation (1 week before the initiation of treatment) to assess the engraftment and expansion of the transplanted bone marrow cells. Among the P140K⁺-bone marrow-transplanted mice (n = 7), four animals exhibited 10–20% GFP+ peripheral blood cells. In two of the mice, <5% of the peripheral blood cells were GFP+; and no GFP+ cells were detected in one mouse. Of the P140K⁺-bone marrow-transplanted animals that subsequently received high-dose BG/BCNU treatment (n = 4), only one animal died on the 13th day after treatment. The peripheral blood in that mouse contained only 3.3% GFP+ cells, which was the lowest level of GFP expression among the treated mice. Of the remaining three mice that survived for 28 days posttreatment, the percentages of GFP+ peripheral blood cells in these animals were 5.8%, 12.2%, and 15.2%.

Transplantation of P140K⁺-bone marrow without drug treatment had no effect on the growth rates of untreated xenograft tumors, as expected. The mean tumor doubling time in non-transplanted, non-treated animals and in transplanted, non-treated animals was 9 days (Table 1). The tumor growth curves from these two untreated groups were combined for comparison to the treated groups (Fig. 1A). Treatment groups in Fig. 1, A and B, are the same. Among the non-transplanted animals that received the high-dose BG/BCNU treatment, none survived for 28 days (Fig. 1B). Two of the mice succumbed on the 5th day, and the remaining two on the 12th day, after treatment. All of these animals exhibited severe aplasia in the spleen and bone marrow, consistent with lethal myelotoxicity. Significant regressions in tumor volume were observed in three of these animals before succumbing to treatment, averaging 18% in the first week and 23% by Day 12. Among the animals transplanted with P140K-expressing bone marrow, three of four mice survived 28 days (Fig. 1B). In the mouse that succumbed on the sixth day posttreatment, the tumor volume had regressed by 20%. In the remaining three mice, two of the tumors showed dramatic regressions in volume for the first 2 weeks posttreatment (27% and 39%), while the third tumor exhibited only a significant reduction in growth compared to untreated tumors. All of the BG/BCNU-treated tumors resumed positive growth during the third and fourth weeks posttreatment. Despite the growth of the tumor in the third and fourth weeks, the mean tumor doubling time was 32 days among the BG/BCNU-treated animals that survived for 28 days (or 3.5-fold longer than untreated tumors; P < 0.001; Table 1). The high-dose BG/BCNU treatment thus significantly delayed the tumor growth compared to untreated tumors (Fig. 1A).

Two Cycles of High-Dose BG/BCNU Treatment

Because one cycle of high-dose BG/BCNU produced significant tumor regression for the first 2 weeks after treatment but tumor growth resumed in the third and fourth weeks, a second cycle of BG/BCNU was added to the treatment regimen. This second cycle was administered 2 weeks after completion of the first cycle. Bone marrow transplantation and tumor implantation were performed as described above. Peripheral blood was collected at 3 weeks posttransplantation (before drug treatment) for GFP analysis to assess the engraftment and expansion of the transplanted bone marrow cells. In addition, neutrophil, RBC, and platelet counts were measured. The percentage of GFP expression in the peripheral blood of P140K⁺-bone marrow-transplanted mice was 17 ± 10% (n = 14). Before drug treatment, the P140K⁺-bone marrow-transplanted mice exhibited neutropenia (1.5 ± 0.6 × 10³ neutrophils/μl) compared to non-transplanted mice (4.0 ± 2.8 × 10³ neutrophils/μl; n = 12) (P < 0.01). The platelet count was also significantly lower in the transplanted mice (220 ± 139 × 10³ platelets/μl) compared to non-transplanted mice (1389 ± 156 × 10³ platelets/μl; P < 0.01). However, the transplanted mice did not exhibit any gross symptoms of toxicity, as indicated by gross weight and coat appearance.

The P140K⁺-bone marrow-transplanted mice were subsequently separated into four treatment groups: vehicle only (n = 3); BG only (n = 3); BCNU only (n = 3); and high-dose BG/BCNU (n = 5). Non-transplanted animals were separated into two treatment groups: vehicle only (n = 9) and high-dose BG/BCNU (n = 4). Two cycles of each treatment were administered: 4 weeks post-bone marrow transplantation on Day 0, and on Day 14 after the first treatment cycle. The mean tumor doubling time in the transplanted and non-transplanted, vehicle-treated animals was 9 days (Table 1). The average increase in volume of the untreated tumors at 28 days was 7.6-fold (Fig. 2A). Treatment groups in Fig. 2, A and B, are the same. Treatment with BG only (30 + 15 mg/kg) had no effect on tumor growth, and no toxic effects were observed (data not shown). Treatment with BCNU only (20 mg/kg) extended the tumor doubling time by 70% to 16 days, with an average increase in tumor volume of 5.3-fold at 28 days (P < 0.05; Fig. 2A). This BCNU treatment was severely myelotoxic to non-transplanted mice, producing about 80% lethality (data not shown). However, no toxicity was observed in the P140K⁺-bone marrow-transplanted mice from this BCNU treatment, and all of the animals survived for 28 days. The high-dose combination BG/BCNU treatment was also severely myelotoxic and highly lethal to the non-transplanted mice, with no mice surviving longer than 10 days (Fig. 2B). Among the mice transplanted with P140K-transduced bone marrow, three of five animals (60%) survived two cycles of the high-dose BG/BCNU treatment (Fig. 2B). One animal succumbed on the sixth day after the first cycle of treatment, and another succumbed 5 days after the second cycle. In the animals that survived both cycles of treatment, the tumor volume in this group on average decreased by 20% at the end of 28 days (P < 0.001; Fig. 2A); and one tumor had regressed completely, with no evidence of tumor remaining under gross pathological examination.

At the end of the experiment (about 8 weeks post-bone marrow transplantation), peripheral blood and bone marrow were harvested from the surviving mice for analysis of GFP expression and MGMT activity. All of the mice transplanted with P140K⁺-bone marrow expressed high levels of GFP in the bone marrow and in peripheral blood, regardless of drug treatment (Table 2). Hence, as expected, the GFP/drug resistance phenotype selected in the primary animals was maintained in secondary recipients.

Priming Dose and Two Cycles of High-Dose BG/BCNU Treatment

Data from the previous experiments suggest that toxicity from BG/BCNU combination therapy was excessive in spite of adaptively transferred GFP/drug resistance bone marrow cells. Because 5 mg/kg of BCNU in combination with BG was highly selective for P140K⁺-bone marrow in the primary transplanted mice, it was employed as a “priming” dose in the secondary mice before initiation of the high-dose BG/BCNU treatments. Two weeks after the priming dose (30 + 15 mg/kg BG plus 5 mg/kg BCNU; administered about 4 weeks posttransplantation), the animals received two biweekly cycles of the high-dose BG/BCNU treatment (30 + 15 mg/kg BG plus 10 mg/kg BCNU).

The mice transplanted with P140K⁺-bone marrow were separated into two treatment groups based on statistical considerations from the previous experiments: vehicle-treated (n = 3) and BG/BCNU-treated (n = 10). Non-transplanted animals were separated into the same two treatment groups: vehicle-treated (n = 9) and BG/BCNU-treated (n = 7). The low-dose priming treatment was administered about 4 weeks post-bone marrow transplantation on Day 0, and two cycles of the high-dose treatment were subsequently administered on Days 14 and 28. Gross animal weight and tumor volume were monitored three times weekly until Day 42. The mean tumor doubling time in the transplanted and non-transplanted, vehicle-treated animals was 9 days (Table 1). The average increase in volume of the untreated tumors at 42 days was 8.8-fold (Fig. 3A). None of the non-transplanted, BG/BCNU-treated mice survived beyond 20 days (Fig. 3B). Treatment groups in Fig. 3, A and B, are the same. Among the mice transplanted with P140K⁺-bone marrow, 8 of 10 animals (80%) survived all three cycles of the BG/BCNU treatment (P < 0.001; Fig. 3B). One animal succumbed on Day 28, and another on Day 42. Both of these animals exhibited severe myelotoxicity. The average increase in volume of the treated tumors of the mice transplanted with P140K⁺-bone marrow at 42 days was 2.2-fold (P < 0.01; Fig. 3A); and the mean tumor doubling time was 44 days, nearly a 5-fold increase compared to vehicle-treated tumors.

At the end of the experiment (about 10 weeks post-bone marrow transplantation), peripheral blood and bone marrow were harvested from the surviving mice for analysis of GFP expression and MGMT activity. The levels of GFP expression in the bone marrow and peripheral blood of the mice transplanted with P140K⁺-bone marrow were variable (data not shown). Half of the animals showed high levels of the GFP expression in both compartments (>60% GFP+ cells in the bone marrow; >70% GFP+ cells in the peripheral blood), while half of the animals displayed low numbers of cells in the bone marrow and peripheral blood and correspondingly low levels of GFP expression (<10% GFP+ cells in the bone marrow or peripheral blood). However, MGMT activity in the bone marrow of the P140K⁺-secondary transplanted animals was uniformly high compared to that of non-transplanted animals (Fig. 4A). The MGMT activity in the non-transplanted bone marrow was 307 ± 59 fmol O⁶-methylguanine repaired/mg protein, while that in the P140K⁺-bone marrow ranged from 2,018 to 110,400 fmol O⁶-methylguanine repaired/mg protein (average of 52,610 ± 36,760 fmol O⁶-methylguanine repaired/mg protein; n = 8; P < 0.001). Furthermore, in contrast to the bone marrow of non-transplanted animals, the P140K⁺-bone marrow was highly resistant to treatment in vitro with BG (25 μm; Fig. 4B).

Discussion

Inhibition of MGMT activity in tumor cells is a promising therapeutic strategy to sensitize MGMT-expressing tumors to agents such as BCNU that alkylate guanine at the O⁶ position. Combination BG and BCNU treatment has previously been shown to produce 10- to 15-day growth delay in tumor-bearing mice (25–27). In those studies, MGMT depletion by BG resulted in a reduction of the maximum tolerated dose (MTD) of BCNU in mice by 3-fold due to potentiation of BCNU toxicity, primarily myelotoxicity. Combination BG and BCNU treatment produced a similar effect in recent clinical studies (9, 10). Hence, combination pharmacologic plus gene therapy strategies that employ expression of BG-resistant MGMT mutants in hematopoietic cells to enhance the therapeutic index and efficacy of combination BG/BCNU antitumor treatment.

The objective of this study was to determine whether hematopoietic expression of MGMT-P140K resulted in increased tolerance to BG/BCNU treatment, thereby allowing effective antitumor treatment of a BCNU-resistant tumor. We have previously demonstrated that optimal tumor killing by BG/BCNU in MGMT+ tumor cells is accomplished by depleting MGMT activity for 24 h (7). In SF767 cells, even partial regeneration of MGMT activity within 24 h of BCNU treatment significantly attenuated the induced sensitization when compared to complete MGMT depletion for 24 h (28). We therefore developed a double-bolus BG treatment that depleted >95% of cellular MGMT activity for 24 h in xenograft SF767 tumors (8). We have employed this double-bolus BG treatment in combination with BCNU in the current study. The double-bolus BG treatment of 30 + 15 mg/kg reduced the maximum tolerated dose of BCNU by one half to 5 mg/kg compared to a single-bolus BG treatment (30 mg/kg only) in NOD/SCID mice. However, mice transplanted with MGMT-P140K+ hematopoietic cells tolerated one or two biweekly cycles of the double-bolus BG treatment in combination with 10 mg/kg BCNU. In addition, growth of the human glioma xenograft tumors was more effectively suppressed by the BG/BCNU treatment at 10 mg/kg BCNU compared to 5 mg/kg BCNU. Tumor growth was delayed by 32 days in mice that received only one cycle of the high-dose BG/BCNU treatment. Administration of a second cycle of treatment extended the growth delay to beyond 40 days; in addition, one complete tumor regression was observed by this treatment. The percentage of animals that survived for 28 days, however, was diminished to 60%. To enhance the animal survival rate while simultaneously extending the duration, a reduced dose of BG/BCNU (30 + 15 mg/kg BG and 5 mg/kg BCNU) was administered before initiation of the high-dose treatments. A tumor growth delay of 44 days was achieved, and 80% of the mice survived for 42 days. All of the mice that succumbed during the experiments exhibited severe myelotoxicity, as indicated by aplasia in the spleen and bone marrow. The surviving animals, however, exhibited no symptoms of toxicity besides slight and temporary weight loss for several days post-drug treatment. It seems likely that a period of 1 month between bone marrow transplantation and initiation of intensive combination BG/BCNU treatment is inadequate for bone marrow engraftment. Future studies will extend the time between bone marrow transplantation and drug treatment.

Suppression of the human glioma xenograft tumors for a prolonged period of time was possible only by protecting the bone marrow against BG/BCNU treatment by expression of a BG-resistant MGMT mutant protein. Several human mutant MGMT proteins have been shown to confer resistance to BG while retaining the ability to remove alkylator-induced adducts, including G156A, P140A, and P140K. Mice transplanted with G156A-transduced bone marrow were protected from BG and BCNU hematopoietic toxicity (16, 17). In addition, treatment of human colon tumor xenografts was enhanced by hematopoietic expression of G156A (17). However, the P140K mutant exhibits the most favorable biochemical characteristics for hematopoietic protection strategies. Furthermore, the P140K mutant was also shown to be superior to the P140A mutant and wt-MGMT in protecting bone marrow from BG/BCNU toxicity in vivo (14, 19). In that study, the resistance phenotype was transmitted to secondary and tertiary bone marrow transplant recipients, and provided sustained and enhanced protection. To provide a uniform and highly selected population of P140K⁺ hematopoietic donor cells, we therefore employed a secondary bone marrow transplant model in the current study. The secondary transplant model ensured that the P140K expression occurred and was maintained at the early progenitor and stem cell level.

During the high-dose BG/BCNU treatment, mice transplanted with MGMT-P140K-bone marrow cells exhibited increased survival and protection of myeloid and lymphoid lineages compared to non-transplanted mice. Bone marrow cellularity and peripheral blood counts were significantly higher in treated mice transplanted with P140K-expressing marrow compared with non-transplanted mice. However, the level of P140K expression attained in the secondary transplant recipients was not uniform. For example, half of the mice that survived the priming dose and two cycles of high-dose BG/BCNU treatment exhibited very high levels of bone marrow cellularity and peripheral blood counts, as well as GFP expression and MGMT activity in bone marrow cells. In contrast, two of the mice were depleted of cells in the bone marrow and peripheral blood. In these mice, GFP expression was undetectable and MGMT activity was dramatically lower, although still much higher than in non-transduced bone marrow cells. Thus, despite the high number of selected, P140K-expressing donor cells infused into the secondary transplant recipients, it appears that either inadequate numbers of stem cells were infused, or the level of drug resistance in these cells was below the level required to protect the marrow graft.

In conclusion, we have demonstrated that MGMT-P140K-transduced murine hematopoietic cells increase tolerance to high-dose combination BG and BCNU treatment, reducing the myelotoxicity and death associated with this therapy in xenograft tumor-bearing mice. Increasing the intensity of combination BG and BCNU treatment also increased the antitumor efficacy. In preparation for use of this strategy to treat BCNU-resistant tumors in humans, in the future we will investigate whether transduction of MGMT-P140K in human hematopoietic cells can similarly protect against intensive combination BG and BCNU antitumor therapy.