Reversal of Mitomycin C Resistance by Overexpression of Bioreductive Enzymes in Chinese Hamster Ovary Cells¹ Free

MC,³ which is produced by Streptomyces caespitosus and S. lavendulae, is a bioreductively activated chemotherapeutic agent with activity against a variety of solid tumors (1); thus, to exert its cytotoxicity, MC must be enzymatically reduced (2). A number of enzymes have been shown to be capable of activating MC in mammalian cells, including FpT (EC 1.6.2.4), FpD (EC 1.6.2.2), NAD(P)H:(quinone-acceptor) oxidoreductase (DT-diaphorase; EC 1.6.99.2), XOD (EC 1.1.3.22), and XDH [EC 1.1.1.204; see Belcourt et al. (3) for appropriate references]. FpT produces a one-electron reduction of MC to generate the semiquinone anion radical, which under aerobic conditions redox cycles in the presence of oxygen to form reactive oxygen species (i.e., superoxide, hydrogen peroxide, and/or hydroxyl radicals) and to regenerate the prodrug form of MC; under hypoxia, however, the semiquinone radical anion disproportionates to form the reactive hydroquinone intermediate, which leads to alkylation of DNA and/or other cellular nucleophiles.

In contrast to FpT, DT-diaphorase catalyzes a two-electron reduction of quinones such as MC (4). Bioactivation of MC by purified DT-diaphorase from rat liver or HT-29 human colon carcinoma cells is pH dependent and increases under acidic conditions (4, 5, 6). The pH dependence derives from the fact that, at pH 7.8, MC behaves as a suicide inhibitor of DT-diaphorase, causing irreversible inhibition of the enzyme. The metabolism of MC by purified rat liver DT-diaphorase and human lung tumor DT-diaphorase has been compared under aerobic and hypoxic conditions, with the rates of reduction of MC by both enzymes being similar in air and hypoxia (5, 6). A small difference between these enzymes exists, however, with the rate of reduction of MC by the rat enzyme being approximately five times greater than that by the human enzyme. HT-29 human colon carcinoma cells, which exhibit high levels of DT-diaphorase activity, are more sensitive to the cytotoxic action of MC than BE colon cancer cell lines, which are deficient in DT-diaphorase activity (7). However, several groups that have attempted to correlate MC cytotoxicity with DT-diaphorase activity have arrived at conflicting conclusions as to the ability of this enzyme to bioactivate MC in living cells (8, 9, 10, 11, 12).

The reduction of MC catalyzed by XOD, with NADH as the source of reducing equivalents, was essentially identical to that reported for FpT with regard to the kinetics of the reaction, the metabolites formed, and the pH dependence (13). The XDH-mediated metabolism of MC results in the generation of alkylating species under both hypoxia and aeration, with greater alkylation occurring under conditions of hypoxia (14). However, because XOD and XDH are absent in most rapidly proliferating cell lines, they do not appear to be major contributors to the activation of MC in most tumor cell lines. Nonetheless, it is conceivable that XOD and XDH may contribute to the toxic side effects of MC in normal tissues.

Although NADH is relatively ineffective in supporting the bioreductive metabolism of MC in rodent liver microsomes (15, 16), we have found that activation of this antibiotic to an electrophilic species is supported equally by NADPH and NADH in sonicates of EMT6 mammary carcinoma cells under hypoxia (17). FpD from human and rabbit erythrocytes, both of which require NADH, has also been shown to catalyze the one-electron reduction of MC (17, 18).

In an aerobic environment, molecular oxygen reacts with the one-electron reduction product of MC (i.e., the semiquinone anion radical) to regenerate the nontoxic parent compound, MC. In contrast, the two-electron reduced species (i.e., the hydroquinone species) is relatively oxygen insensitive. Thus, MC is more active against hypoxic cells where both pathways are operative. The cytotoxicity of MC correlates directly with the capacity of reduced MC to cross-link DNA, and only a few cross-links appear to be required to produce cell lethality (19, 20).

The clinical utility of MC, like that of almost all chemotherapeutic agents, is limited by the development of resistance. Various studies have suggested that MC resistance may be mediated by increased drug efflux, the repair of DNA lesions, or a deficiency in bioreductive enzymes such as DT-diaphorase (21, 22). Research in this laboratory has demonstrated that mammalian cell lines that express profound resistance to MC can be developed by incorporating a bacterial resistance gene known as mcrA (23). The mcrA gene produces a 54-kDa flavoprotein, MCRA, which functions by reoxidizing the reduced bioactive hydroquinone form of MC back to the parent molecule (23, 24, 25, 26).

In the present study we have developed cell lines highly resistant to MC (HRM cells) by continuous stepwise treatment with this agent. Similar resistance profiles for both HRM- and mcrA-transfected cell lines under oxygenated and hypoxic conditions suggested that resistance was attributable to a similar mechanism (i.e., reoxidation of the reduced hydroquinone form of MC back to the parental prodrug). This finding led us to explore whether the resistance could be altered or abrogated by overexpression of exogenously introduced cellular bioreductive enzymes known to be involved in the activation of MC.

CHO/AA8 cells were maintained under 5% CO₂ in α-MEM supplemented with 10% fetal bovine serum, penicillin (100 units/ml), and streptomycin (100 μg/ml). All tissue culture reagents were purchased from Life Technologies, Inc. (Gaithersburg, MD). The resistant phenotype of CHO/HRM cells was maintained by periodic pulsing with 100 μm MC for 1 h at 37°C.

Cells were transfected using Lipofectamine (Life Technologies) and the appropriate plasmid construct under conditions recommended by the manufacturer. Briefly, 25 μl of Lipofectamine were mixed with 5 μg of plasmid DNA in 0.6 ml of serum-free medium for 30 min at room temperature. This cocktail was added to subconfluent monolayers of CHO/AA8, CHO/HRM, or CHO/MCRA cells in 25-cm² plastic flasks with 2 ml of culture medium for 5 h at 37°C. The medium was replaced, and 48 h later cells were detached by trypsinization and seeded into 75-cm² flasks in the presence of 1 mg/ml G418 (for CHO/HRM cells) or 0.5 mg/ml Zeocin (for CHO/MCRA cells). After colony outgrowth under selection, single-cell suspensions were sorted using a flow cytometer. Overexpressing clones were identified by direct measurement of cell lysates for the appropriate enzyme activity (see “Assays of Enzyme Activity” below). Because resistance to G418 was conferred on the CHO/MCRA cell line by the introduction of the mcrA gene, the Zeocin selection system (Invitrogen) was used as a cotransfected marker to introduce the DT-diaphorase and FpT cDNAs. Plasmids containing either Zeocin or a reductive enzyme were cotransfected at a 1:5 ratio into CHO/MCRA cells.

Reductive enzyme assays were performed essentially as described previously (27, 28, 29). Briefly, exponentially growing cells were removed from flasks by trypsinization, washed in complete medium, then in PBS, and resuspended in PBS at a concentration of 10⁷ cells/ml. Cells were disrupted by sonication with three 10-s bursts, with 1 min of cooling on ice between bursts. Cell disruption was confirmed microscopically. DT-diaphorase activity was measured as the dicumarol-inhibitable reduction of dichlorophenolindophenol, measured at 600 nm with a Beckman model 6 UV/Vis spectrophotometer and calculated using an extinction coefficient of 21 mm⁻¹ cm⁻¹ at 30°C (4); the final concentration of dicumarol was 100 μm. FpT activity in cell extracts was assayed by monitoring the rate of ferricytochrome c reduction at 550 nm (extinction coefficient, 21 mm⁻¹ cm⁻¹) at 30°C. FpD activity was measured as NADH:ferricyanide reductase at 420 nm (extinction coefficient, 1.02 mm⁻¹ cm⁻¹) at 30°C.

Because both FpT and DT-diaphorase reduce dichlorophenolindophenol, in cell lines with elevated levels of FpT, the measured activity of DT-diaphorase can sometimes appear to be elevated slightly over that of control cells (27). The initial screening of CHO/MCRA cells transfected with reductive enzymes was performed using a rapid microtiter dish assay. These assays were similar to the enzyme assays described above except that overall volumes were reduced to permit the use of a 96-well plate and samples were analyzed on a Bio-Tek ELX800 universal microplate reader. XOD and XDH activities were measured by monitoring the formation of uric acid at 293 nm (extinction coefficient, 12.2 mm⁻¹ cm⁻¹) at 30°C, with oxygen and NAD⁺ as the respective electron acceptors. Protein concentrations were assayed by the BCA protein assay reagent (Pierce Chemical Co., Rockford, IL).

Clonogenic survival assays were performed as described previously (29). Cells were seeded into plastic 25-cm² Corning tissue culture flasks at 2 × 10⁵ per flask. Three days after plating, cells were pulsed for 1 h at 37°C with MC dissolved in 70% ethanol in a total volume of 2 ml of medium. After treatment, monolayers were rinsed, and cells were detached by trypsinization, suspended in cell culture medium, and counted. Sequential cell dilutions were plated in triplicate into 60-mm dishes; 10–12 days later, colonies were fixed, stained with crystal violet in 80% methanol, and counted. All analyses were corrected for plating efficiency in the presence of vehicle (70% ethanol) at concentrations equivalent to those used for the treatment with MC. Alternatively, for continuous treatment with MC, cells were maintained at 37°C in 5% CO₂ after dilution and plating in 3 μm drug until staining. Hypoxic studies were performed as described previously (29). Cells were seeded into glass milk dilution bottles at 2 × 10⁵ per flask, and hypoxia was induced by gassing each bottle with a mixture of 95% N₂-5% CO₂ at 37°C for 2 h, followed by drug treatment for 1 h with continuous gassing. Cells were then processed for clonogenic assays as described above.

Immunoblot analyses were performed using ECL (Amersham) as described previously (23). Equal numbers of cells (∼1 × 10⁵) were lysed by sonication and resuspended in protein sample buffer [3% SDS, 10% glycerol, 5% 2-mercaptoethanol, 60 mm Tris-HCl (pH 6.8)]. Equal volumes were subjected to SDS-PAGE through a 10% gel, and samples were transferred to nitrocellulose membranes. Blots were incubated with a 1:10,000 dilution of anti-MCRA antibody (25) in 5% nonfat dry milk for 2 h. After washing, the filters were incubated in a 1:1000 dilution of goat antirabbit antibody conjugated to horseradish peroxidase in 5% nonfat dry milk. Bands were visualized with enhanced chemiluminescence reagent using Kodak XAR film.

The isolation of the bacterial resistance gene mcrA and the demonstration that expression of this gene in mammalian cells conferred profound resistance to MC (23) led us to explore the possibility that a similar mechanism might be operative in mammalian cells selected for resistance to this drug. A CHO/AA8 cell line highly resistant to MC, CHO/HRM, developed by continuous aerobic culture in the presence of steadily increasing concentrations of drug up to 1 μm, followed by single-cell cloning by use of flow cytometry, resulted in clones exhibiting a pattern of sensitivity to MC under aerobic and hypoxic conditions analogous to that produced in cells transfected with mcrA (Ref. 23; Fig. 1). Thus, in selected lines of CHO/AA8 cells resistant to MC, profound resistance to the antibiotic occurred under aerobic conditions to drug concentrations as high as 400 μm, whereas under hypoxia, sensitivity to MC was much closer in magnitude to that occurring in wild-type CHO/AA8 cells. To phenotypically select against multidrug transporter resistance, we incorporated 0.01% Tween 80 into the medium during the selection process (30). In the same cell background (CHO/AA8), a cell line (CHO/MCRA) was engineered to express the bacterial MC resistance gene mcrA. Clonogenic assays were used to assess the degree of resistance to MC. As shown in Fig. 1,A and Table 1, both CHO/HRM and CHO/MCRA cells displayed a pronounced level of resistance to MC compared with parental CHO/AA8 cells. This resistance was apparent when cells were either exposed to MC in a high-pulse dose (50 μm for 1 h) or cultured continuously in the presence of 3 μm MC.

Previous studies from this laboratory have demonstrated that overexpression of DT-diaphorase and FpT, two enzymes that bioactivate MC, increases the toxicity of MC to parental CHO cells (28, 29). To gain information on the role of DT-diaphorase in the phenomenon of resistance to MC, CHO/HRM cells were generated that overexpressed DT-diaphorase, and clones were isolated by selection of single cells by flow cytometry and growth of the cloned populations in G418. Clones were initially screened and chosen by assaying for DT-diaphorase activity to detect overexpression. Resistant CHO/HRM cells were originally isolated by continuous exposure to MC up to a concentration of 1 μm. Subsequent experiments indicated that the resistant cells could readily survive for extended periods of time in 3 μm MC, a 3-fold higher concentration than the maximum concentration used for selection. Therefore, DT-diaphorase-overexpressing clones were initially tested for MC sensitivity by clonogenic assays performed using continuous treatment with 3 μm MC. As shown in Fig. 2, all of the stable CHO/HRM clones developed by transfection with and overexpression of the gene encoding DT-diaphorase were significantly more sensitive than parental CHO/HRM cells to the drug. The level of DT-diaphorase overexpression in these clones varied from 11- to 114-fold. None of the DT-diaphorase-overexpressing clones was as sensitive as the parental CHO/AA8 cells to the continuous exposure to MC, and all DT-diaphorase-overexpressing clones produced colonies of cells that survived and expanded to confluence in 3 μm MC. The growth of all of these clones under continuous treatment with 3 μm MC was unexpected. Hence, a large number of clones were analyzed to eliminate the possibility that some of the cell lines were not truly clonal. We studied this phenomenon further by allowing the CHO/HRM-DTD-114 and CHO/HRM-DTD-91 cell lines to grow under continuous treatment with 3 μm MC. These two cell lines, which were designated CHO/HRM-DTD-114X and CHO/HRM-DTD-91X, respectively, were assessed for DT-diaphorase activity and for MC sensitivity. As shown in Table 1, these cell populations regained substantial resistance to MC, approaching the level exhibited by the original CHO/HRM cell line. The levels of resistance in these cells were 875- and 275-fold greater than those of the initial CHO/HRM-DTD-114 and CHO/HRM-DTD-91 lines, respectively. Furthermore, DT-diaphorase levels decreased markedly (19- and 10-fold) in these back-selected lines, but remained slightly elevated over CHO/HRM levels. Thus, cell growth in MC correlated with a drop in the levels of DT-diaphorase. The reason for the loss of DT-diaphorase expression is unknown, but could involve epigenetic mechanisms such as DNA methylation of the transfected genes.

Most of our previous studies have used 1-h pulses of MC to assess cell sensitivity. Therefore, a representative subset of the clones was also tested using this approach, by exposing cells to a 50-μm pulse of MC for 1 h (Fig. 3). The results obtained were similar to those seen for cells continuously cultured in MC, in that all of the DT-diaphorase-overexpressing clones were more sensitive to pulsed MC than were CHO/HRM cells. At least three of the cell lines with DT-diaphorase levels elevated by 50-fold or more (i.e., CHO/HRM-DTD-57, -91, and -114) were more sensitive to the antibiotic by this assay than the parental CHO/AA8 cell line (Fig. 3 and Table 1).

One of the one-electron donating enzymes that has been implicated in the cellular bioactivation of MC is FpT. This enzyme is thought to function in MC activation mainly in environments where oxygen is limiting, such as in the hypoxic compartment of solid tumors. Previous studies from this laboratory have demonstrated that overexpression of FpT increased the sensitivity of CHO cells to MC (28). Therefore, FpT was assessed for its capacity to decrease the resistance of CHO/HRM cells to MC. Clonal cell lines were isolated after transfection with FpT and selection by flow cytometry. Individual clonal cell lines were treated with MC, and single cells were allowed to form colonies in a clonogenic assay. As shown in Figs. 4 and 5, overexpression of FpT also resulted in a decrease in resistance to MC. Loss of resistance correlated with the amount of overexpression and was demonstrated by both pulsed treatment with MC and continuous culture with the drug (CHO/HRM-FpT-3, -9, -13, and -37). Controls of clonal CHO/HRM cell lines transfected with vector alone (CHO/HRM-vector-3 and -17) displayed levels of MC resistance that were similar to the resistant CHO/HRM cell line from which they were derived (Table 1).

As noted above, a number of cellular enzymes have been implicated in the activation of MC. To verify that the expression levels of these alternate MC activators had not changed after clonal isolation of DT-diaphorase- and FpT-overexpressing cells, all cell lines were assayed for expression of DT-diaphorase and FpT (Table 2). Furthermore, a representative subgroup of the cells was assessed for FpD, XOD, and XDH activities. As shown in Table 2, in cells overexpressing DT-diaphorase activity there was no significant increase in FpT activity, and in cells overexpressing FpT activity there was no major increase in DT-diaphorase activity. As discussed in “Materials and Methods,” the method used to measure DT-diaphorase enzyme activity also detects some FpT activity, which would be expected to be most affected in cell lines with the highest levels of FpT overexpression. Nonetheless, even if the levels of DT-diaphorase varied somewhat in FpT-overexpressing clones, it was still clear that overexpression of FpT decreased the level of resistance of CHO/HRM cells. Levels of FpD reductase were similar in all cell lines, and no XOD or XDH activities were detectable in any of the cell lines tested (data not shown).

MCRA is the protein product of the bacterial gene mcrA, which confers a high level of resistance to MC when introduced into bacterial or mammalian cells. Previous studies from this laboratory have demonstrated that expression of MCRA in CHO/dhfr⁻ cells produced marked resistance under aerobic conditions, with relatively little change in sensitivity under hypoxia (23). In the present studies, MCRA was expressed in CHO/AA8 cells, the parental cells from which the resistant CHO/HRM line was derived. CHO/MCRA cells, which were found to be profoundly resistant to MC (Fig. 1 and Table 3), were also transfected with either the FpT or DT-diaphorase cDNA. Single cells were then isolated by flow cytometry, grown into clones, and analyzed for reductive enzyme expression by a rapid microtiter dish assay. Clones that overexpressed the enzyme of interest were selected for further analysis. Immunoblot analyses were conducted to demonstrate that MCRA expression had not been lost during the selection process; these measurements showed that all of the CHO/MCRA subclones continued to express MCRA abundantly (Fig. 6). The sensitivity of each clone to MC was determined by a clonogenic assay (Table 3). The sensitivity to MC varied with the levels of DT-diaphorase activity. Low-level overexpression of DT-diaphorase did not significantly sensitize CHO/MCRA cells (i.e., CHO/MCRA-DTD-6 and -13) to MC, whereas in contrast, restoration of MC sensitivity occurred with a 40-fold or greater overexpression of DT-diaphorase. Overexpression of FpT resulted in a small but consistent sensitization. Overall, the data suggested that a greater fold overexpression of DT-diaphorase was required to sensitize CHO/MCRA cells to MC than to sensitize CHO/HRM cells. The ability of DT-diaphorase to reverse high-level resistance driven by the dedicated bacterial resistance gene, mcrA, reinforced the significant role of this enzyme in modulating MC sensitivity under aerobic conditions. Finally, enzyme analyses of these cell lines demonstrated that the activity of nontransfected enzymes did not markedly change after selection of the overexpressing cells (Table 4).

A series of MC-resistant cell lines, largely human colon carcinoma and CHO cell lines, have been generated by several groups and used to evaluate the importance of some of the enzyme systems involved in the metabolic activation of MC. Resistant cell lines were produced by extensive treatment of cells with increasing concentrations of MC over a period of months; most of these cell lines probably contain an accumulation of genetic changes. The resistance of HCT116 human colon carcinoma cells to MC has been shown to be associated with an inability of the resistant cells to activate MC; resistant HCT116 cells, however, were fully sensitive to the 7-N-dimethylaminomethylene analogue of MC, BMY-25282, as were parental HCT116 tumor cells (31, 32). Pan et al. (33) also isolated a MC-resistant mutant of HCT116 cells that displayed a 50% decrease in the level of macromolecule-bound POR; the decrease was not related to the ability of the cells to accumulate POR, but rather to a defect in their ability to bioreductively activate the drug. Whereas microsomes from the MC-sensitive and -resistant cells demonstrated equal abilities to activate MC and POR, suggesting that the levels of the microsomal enzymes FpT and FpD were unchanged in the resistant cells, DT-diaphorase enzyme activity was 95% lower in resistant cells. DT-diaphorase mRNA levels in sensitive parental and resistant cells were identical, suggesting the presence of an alteration in enzyme activity. Under hypoxic conditions, the MC resistance of the mutant cell line vanished and resistant cells exhibited the same sensitivity to both MC and POR as the parental cells.

Correlating enzyme activity with mRNA levels quantified by PCR, Traver et al. (34) measured DT-diaphorase levels in 10 different human colon carcinoma cell lines, which exhibited levels of enzyme activity ranging from undetectable to 3447 nmol/min/mg of protein. The levels of DT-diaphorase activity were directly proportional to the sensitivity of the cell lines to MC. The MC-resistant cell line, BE, displayed no DT-diaphorase enzyme activity, but expressed 91% of the mRNA of the highest DT-diaphorase enzyme-expressing cell line, HT-29. Sequence analysis of the mRNA and the reverse-transcribed, PCR-amplified cDNA revealed a point mutation at position 609, which changed amino acid 187 from proline to serine, an amino acid that maps in the pyridine nucleotide binding site.

A series of CHO cell mutants with a range of sensitivities to MC suggested that several different mechanisms can lead to MC resistance (35). Whereas none of the lines showed alterations in glutathione levels or in drug transport, changes in DNA excision repair proficiency and in the capacity to bioreductively activate MC were observed. The most MC-sensitive cell line, UV-41, was UV-light sensitive and was deficient in DNA excision repair. This line was also sensitive to dicumarol, suggesting that it contained sufficient DT-diaphorase and/or FpD to activate MC. The S9-32 and S2-40 cell lines, although also being DNA repair deficient, were insensitive to dicumarol, suggesting that at best they have low levels of DT-diaphorase and/or FpD; these two cell lines were 10-fold more resistant to MC than was UV-41. Wild-type cells, designated CHO/AA8, which were 40-fold more resistant to MC than UV-41, were DNA repair proficient and were able to activate MC. The most resistant cell line, S162, which was 200-fold more resistant to MC than was UV-41, was DNA repair proficient and deficient in its ability to bioactivate MC. DT-diaphorase activity and mRNA levels in these cell lines varied as expected with their ability to bioactivate MC (36). FpT enzyme levels were roughly equivalent in all of the mutants. Of particular significance to the present report, these cell lines were all equally sensitive to MC under hypoxic conditions.

Hoban et al. (37) selected a CHO cell line that was 17-fold more resistant to MC than the parental CHO-K1 line under aerobic conditions, but was equisensitive to MC under hypoxia. The resistant cell line also displayed a 2-fold greater resistance than the parental cells to the mitomycin analogue BMY-25282 under aerobic conditions. The activity of FpT was 3–4-fold lower in resistant cells than in the wild-type parental line. Although the metabolism of MC was undetectable in air, it was detectable in hypoxia, although 2–3-fold lower in resistant cells than in the sensitive parental cells. Unlike other MC-resistant CHO mutants, this line showed no change in DT-diaphorase enzyme activity.

A human skin fibroblast cell line, 3437T, isolated from a family prone to multiple polyposis and sarcomas, demonstrated 6–7 times the MC resistance of a cell line, GM38, isolated from a non-cancer-prone family (38). Because of the insensitivity of the resistant line to dicumarol, DT-diaphorase levels were measured. 3437T cells displayed a 61-fold decrease in DT-diaphorase activity compared with GM38. Consistent with the findings in CHO cells and in human colon carcinoma cell lines, the MC sensitivities of 3437T and GM38 cells were equivalent to the parental cell line under hypoxic conditions.

Cell lines that are hypersensitive to MC typically contain mutations affecting DNA excision repair proficiency [see Buchwald and Clark (39) for appropriate references]. An exception to this finding is a clone derived from the murine lymphoblastic line L5178Y, designated L5178Y/HBM10 (40, 41). This cell line, selected for resistance to hydrolyzed benzoquinone mustard, was 4 times more sensitive than the parental cells to MC and demonstrated a 24-fold increase in DT-diaphorase enzyme levels. Dicumarol inhibited MC-induced DNA cross-link formation and cytotoxicity by 64% under aerobic conditions. The sensitivity of this line to MC under hypoxic conditions was not measured. Xu et al. (42, 43) isolated a subline of a human bladder cancer cell line by continuous exposure to MC that was 6-fold more resistant than the parental cell line to the antibiotic. Two key enzymes involved in the activation of the mitomycins, FpT and DT-diaphorase, were significantly lower in the resistant subline, suggesting that the lower aerobic sensitivity of the resistant line, denoted J82/MC, may have resulted from deficient activation of the drug.

The finding of resistance to MC and POR under aerobic conditions with no increase in resistance over parental cells under hypoxic conditions was observed in all systems in which sensitivity to the mitomycin antibiotics was measured in resistant cells under hypoxic and normoxic conditions. This finding is analogous to the results that we obtained with MC-treated CHO/dhfr⁻ (23) and CHO/AA8 cells (the present report) expressing the bacterial MC resistance protein MCRA and with CHO/AA8 cells selected for resistance to MC. Changes in drug resistance related to alterations in repair capacity do not show a similar pattern (44); instead, differences in the sensitivity of cells in air and hypoxia attributable to repair reflect the specificity of repair deficits in correcting DNA lesions and distribution of lesions (monoadducts, cross-links, and strand breaks) occurring in air and hypoxia. In an analogous manner, changes in the enzyme(s) involved in the reductive activation of the mitomycins would tend to change normoxic, but not hypoxic sensitivity only if the enzyme were operative only in air and not under hypoxia; no such enzyme has been identified. The findings in cell lines selected for resistance to MC are, however, consistent with the major mechanism of resistance being an oxygen-dependent redox cycling enzyme that functions in a manner analogous to MCRA. Such a mechanism involves the oxidation of the activated hydroquinone intermediate back to the parental prodrug by MCRA in a reaction requiring molecular oxygen.

The relatively consistent demonstration that decreases in DT-diaphorase and/or FpT activity occur in cells selected for resistance to MC, which would exacerbate an MCRA-like mechanism, prompted us to determine the effects of transfection and overexpression of DT-diaphorase or FpT cDNAs on the MC-resistant phenotype. Under aerobic conditions, both DT-diaphorase and FpT were found to partially restore the sensitivity to MC in a highly resistant selected cell line. In general, the level of enzyme overexpression corresponded to the level of MC sensitivity, with greater expression yielding greater sensitivity.

In parallel, the CHO/MCRA cell line in the CHO/AA8 background was also assessed for the ability of DT-diaphorase and FpT to reverse resistance. The mechanism of resistance in this cell line is known to be overexpression of the bacterial MC resistance protein, MCRA. The sensitivity to MC of these cells was also restored in part by overexpression of DT-diaphorase and FpT.

Taken together, results from experiments conducted on these two different cell lines, one of which has a known resistance genotype, indicate that resistance to MC can be reversed by modulation of cellular bioactivating enzymes. Even the profound resistance conferred by a bacterial protein evolved specifically to counter mitomycin toxicity in S. lavendulae, an organism that synthesizes large amounts MC, could be overcome by sufficient overexpression of DT-diaphorase. The results imply that the bioactivating enzyme DT-diaphorase can overwhelm the capacity of MCRA, and presumably a functional homologue of this enzyme existing in mammalian cells selected for resistance to MC, to redox cycle the hydroquinone two-electron reduced intermediate back to the parental prodrug (23).

We postulate that the ability of FpT to partially reverse resistance of CHO/HRM and CHO/MCRA cells is attributable to a different mechanism, the redox cycling of the semiquinone anion radical in the presence of oxygen to generate toxic oxygen radicals in addition to MC; this results in cell kill by a mechanism different from the cross-linking of DNA by MC, a phenomenon that results in reversal of the MCRA-like mechanism of resistance. Thus, these findings are in general agreement with the known and postulated roles of DT-diaphorase and FpT in the metabolism of MC when molecular oxygen is present, given that there is evidence that the MC hydroquinone is the major intermediate in the generation of DNA adducts (45). We have postulated that the semiquinone anion radical intermediate may exert its biological effects under hypoxia largely by disproportionation to the hydroquinone (46). These studies overwhelmingly support the idea that DT-diaphorase expression influences MC sensitivity and directly contradict a recently published study (12) that found that a >3000-fold overexpression of DT-diaphorase in CHO cells did not significantly increase sensitivity to low concentrations of MC.

The overall similarity in the response of CHO/HRM and CHO/MCRA cell lines to the introduction of bioactivating enzymes also supports the possibility that a mammalian homologue of MCRA contributes to the resistance of CHO/HRM cells to MC. Studies are ongoing to identify the mammalian homologue of MCRA in cells selected for resistance to MC.

We thank Philip Penketh and Rick Finch for helpful discussions.