Camptothecin Resistance: Role of the ATP-binding Cassette (ABC), Mitoxantrone-resistance Half-Transporter (MXR), and Potential for Glucuronidation in MXR-expressing Cells

Among the most studied mechanisms of drug resistance are membrane proteins that participate in energy-dependent drug efflux and belong to the ABC² superfamily (1). The recently cloned MXR gene encodes an ABC transporter capable of conferring drug resistance (2). Unlike other ABC transporters that confer drug resistance, each of which is composed of two transmembrane regions and two ATP-binding domains, MXR encodes only one transmembrane region and one ATP-binding domain. This half-transporter structure is similar to that described previously for the human TAP and ALDP genes (1). MXR differs minimally from the recently identified half-transporter genes, ABCP1 and BCRP, cloned from placenta and drug-resistant breast cancer cells, respectively (3, 4); presumably all three clones identify the same gene. MXR/ABCP1/BCRP has the highest homology to the Drosophila white gene family, which encodes proteins involved in transmembrane transport of eye pigment precursors, and is thought to require hetero- or homodimerization for transport activity.

Camptothecin and its derivatives are topoisomerase I inhibitors. Among these, CPT-11 and topotecan, which have been approved for use as anticancer agents, are effective in lung, colon, and ovarian cancer (5). Resistance to the camptothecins has been explored in model systems, and several putative mechanisms have been identified. Alterations in topoisomerase I, through either reduced expression or mutation, were the first resistance mechanisms to be described (6). These mechanisms confer resistance to all of the camptothecins by decreasing the formation of the cleavable complexes characteristic of the interaction of camptothecins with topoisomerase I. Overexpression of Pgp results in limited resistance to topotecan but not to camptothecin (7). Two reports have linked topotecan resistance to a novel mechanism for decreased drug accumulation (8, 9). Finally, resistance to SN-38, the active metabolite of CPT-11, has been linked with glucuronidation in human lung cancer cells (10). In these cells, UGT activity and expression were increased, and an inhibitor of UGT increased sensitivity to SN-38.

The spectrum of drug resistance conferred by MXR includes mitoxantrone and the anthracyclines, with no resistance to Vinca alkaloids or taxanes. We observed that MXR-expressing cells were resistant to the camptothecin derivative topotecan and that the resistance could be modulated by the Pgp antagonist GF120918 (11). This study describes our characterization of MXR as a mechanism of resistance for the camptothecins and evaluates the potential contribution of glucuronidation to the drug-resistance phenotype.

We studied glucuronidation in the MXR-expressing cells because they are highly resistant to agents known to be glucuronidated: mitoxantrone, epirubicin, and SN-38, the active metabolite of CPT-11 (12, 13, 14). Taking normal hepatic metabolism as a model, a variety of conjugating enzymes have been shown to participate in drug metabolism and detoxification. The latter conjugate drugs to electrophilic compounds, including glutathione (glutathione-S-transferase), sulfate (sulfotransferase), and glucuronic acid (UGT1 and UGT2; Ref. 15). Conjugation with glucuronic acid has been demonstrated for both exogenous and endogenous compounds, including bilirubin, steroids, and neurotransmitters (15, 16). Bilirubin, for example, is thought to diffuse into the endoplasmic reticulum, where UGTs conjugate glucuronic acid at the inner surface of the microsomal membrane (17). The conjugate is then transported out of the endoplasmic reticulum by an unknown transport mechanism and is effluxed from the cell by the canalicular multispecific organic anion transporter (cMOAT/MRP2) (18). Whether and the extent to which these processes take place in cancer cells has not been determined.

Mitoxantrone, sulforhodamine B, d-saccharic acid-1,4-lactone (saccharolactone), β-glucuronidase (E. coli type VII A) and 4-MU were purchased from Sigma Chemical Company. 9-AC, MDO-CPT, and di-F-MDO-CPT were obtained from Dr. M. R. Wall (Research Triangle Institute, Research Triangle Park, NC). Uridine diphospho-[¹⁴C]-glucuronic acid and [2-¹⁴C]thymidine (specific activity, 287 and 50 mCi/mmol, respectively) were obtained from American Radiolabeled Chemicals Inc., St. Louis, MO. Camptothecin derivatives NSC 609959, NSC 609960, NSC 369394, and NSC 369395 were obtained from the Pharmaceutical Resources Branch, Division of Cancer Treatment, National Cancer Institute (Bethesda, MD). 3-[(3-Cholamidopropyl) dimethylammonio]-1-propanesulfonate was obtained from J.T. Baker Inc. (Phillipsburg, NJ); epirubicin was obtained from Pharmacia & Upjohn (Bridgewater, NJ).

All cells were grown as monolayer cultures at 37°C in a balanced air incubator with a humidified atmosphere of 5% CO₂ and maintained by regular passages in Improved Minimum Essential Media (MCF-7, MCF-7 AdVp3000, and MCF-7 VP-16) or in RPMI (S1, S1-M1-80, S1-B1-20) supplemented with 10% FCS, 2 mm glutamine, penicillin (100 units/ml), and streptomycin (100 μg/ml). The MCF-7 resistant sublines were derived by stepwise selection from parental MCF-7 breast carcinoma cells. The MCF-7 AdVp3000 cells were cultured in 5.5 μm Adriamycin in the presence of 5.5 μm verapamil. MCF-7/VP cells, a gift from Ken Cowan (University of Nebraska Medical Center, Omaha, NE), were maintained in 4 μm VP-16. S1-M1-80 and S1-B1-20 cells are drug-resistant sublines originally isolated from S1 human colon carcinoma cells in the laboratory of Lee Greenberger (Wyeth-Ayerst, Pearl River, NY; Refs. 19, 20). S1-M1-80 cells were obtained by exposure to increasing concentrations of mitoxantrone. The cells were continuously maintained in the selecting agents: 20 μm bisantrene for S1-B1-20 cells and 80 μm mitoxantrone for S1-M1-80 cells. The resistant sublines were placed in drug-free medium 1 week prior to each experiment.

Assays were performed in 96-well plates in the presence of increasing concentrations of the compounds. Cells seeded at a density of 1000–2000 cells/well were treated 24 h later and left in drug for an additional 96 h. After incubation, cells were fixed in 50% trichloroacetic acid, stained with 0.4% sulforhodamine B, and dissolved in 1% acetic acid as described by Skehan et al. (21). After washing, bound dye was solubilized with 10 mm unbuffered Tris base (pH 10.5). Absorbance was determined at a wavelength of 540 nm in a microplate reader (model 450; Bio-Rad, Richmond, CA) in quadruplicates. Untreated control wells were assigned a value of 100%, and the IC₅₀ was defined as the dose required to decrease the absorbance measured at 540 nm to 50% of the control value.

Cells (10⁵/ml) were grown in phenol-red-free Improved Minimum Essential Media for 48 h in 8-well chamberslides (Lab-Tek II; Nalge Nunc International, Naperville, IL) before analysis. A Zeiss LSM 410 confocal laser scanning microscope equipped with a 150 mW Omnichrome (Chino, CA) Ar-Kr laser exciting at 488 nm (all dyes except mitoxantrone) and 568 nm (mitoxantrone) was used. A drug concentration of 5 μm was used for all of the studies presented unless indicated. Emitted light passed through a 515–540 nm band-pass filter (topotecan) or a 590 nm long-pass filter (mitoxantrone, epirubicin). Images were stored on line to an optical disc drive and analyzed off line in Paint Shop Pro version 4.14 (Jasc Inc., Eden Prairie, MN). For energy depletion experiments, cells were preincubated for 20 min in PBS containing 50 mm deoxyglucose and 15 mm sodium azide at 37°C in 5% CO₂ and then incubated in topotecan or epirubicin for 30 min, continuing in ATP-depleting conditions.

RNA was extracted by the RNA STAT-60 (Tel-Test, Inc., Friendswood, TX) method according to the manufacturer’s directions. For quantitative PCR analysis, described previously by Zhan et al. (22), the following primers for UGT expression were used: PCR product = 529 bp

• 5′-end primer: 5′-TTT CCC AAG TTT GGA AAA TCT-3′

• 3′-end primer: 5′-ACT TTG CAT AAA TTA ATC AGC-3′

These primers are complimentary to sequences in exon 3, a region that is common to all UGTs (23). In this way the overall level of UGT expression was assayed, instead of limiting the assay to a single isotype. The expression of UGT was provided in units based on densitometric measurements of the ethidium-stained PCR gels and was normalized to the expression level in parental MCF-7 cells, which was arbitrarily set at 1 unit. The PCR product for parental MCF-7 cells was readily detectable after 31.25 ng of RNA was amplified for 30 cycles.

UGT activity was assayed at 0–4°C, using a previously described assay with minor modifications (24, 25). Briefly, cells were scraped and centrifuged at 750 × g for 10 min at 4°C, and the supernatants were decanted. The cell pellets were resuspended in 100 μl of ice-cold water, vortexed for 10 s, and lysed by sonication (Misonix, Farmingdale, NY). Protein was measured by the Bio-Rad assay (Bio-Rad), and 300 μg protein was added to a 100-μl reaction solution consisting of 15 μm saccharic acid lactone (an inhibitor of β-glucuronidase), 2 mm MgCl₂, 20 mm sodium phosphate (pH 7.2), 300 μm substrate, and 10 μm [¹⁴C]-UDPGA (287 mCi/mmol). In addition, 0.5 mg of 3-[(3-cholamidopropyl) dimethylammonio]-1-propanesulfonate per milligram of protein was added. Incubations were carried out at 37°C for 3 h. The reactions were terminated by adding 2 volumes of absolute ethanol. Treatment with β-glucuronidase (12.5 units) at pH 6.8 for 3 h in the absence of saccharic acid lactone was performed in some reactions. After 12 min centrifugation at 10,000 × g, the supernatants were dried and solubilized in 70% ethanol; 10-μl aliquots of the ethanolic incubation mixtures were then applied to silica gel plates (Silica gel 60 F254, 20 × 20 cm; EM Science, Gibbstown, NJ). Chromatograms were developed in n-butyl alcohol:acetone:glacial acetic acid:30% ammonia:water (70:50:18:1.5:60, v/v) as described previously (24). After air-drying, the TLC plates were processed and autoradiograms obtained after a 20-day exposure.

DNA SSBs were measured as described previously (26). Briefly, the cells were seeded at 1 × 10⁵ cells/25-cm² flask and after 24 h incubated in medium containing 0.04 μCi/ml [¹⁴C]thymidine. After 24 h, cultures were chased in unlabeled medium and incubated for at least 4 h prior to drug treatment. Camptothecin or SN-38 was added to the cultures. After treatment, flasks were placed on ice, scraped, and pipetted up and down several times to suspend the cells. An aliquot was counted, and samples containing ∼10,000 cpm were transferred to 15-ml tubes, on ice, containing HBSS plus drugs at the same concentration the cells were treated in. This step was necessary because rapid reversal of SSBs is known to occur when camptothecin is removed from the cells (26). SSBs were analyzed using DNA-denaturing (pH 12.1) alkaline elution carried out under deproteinizing conditions. Control cells were irradiated on ice with 2000 rad prior to elution. Cells were layered onto polycarbonate filters and lysed with an SDS lysis solution containing 2% SDS, 0.1 m glycine, 0.025 m disodium EDTA, and 0.5 mg/ml proteinase K (pH 10). The lysis solution was washed from the filters with 5 ml of 0.02 m EDTA (pH 10), and the DNA was eluted with tetrapropylammonium hydroxide-EDTA, 0.1% SDS (pH 12.1) at a flow rate of 0.08–0.12 ml/min. Fractions were collected at 5-min intervals for 30 min. Fractions and filters were processed, and radioactivity was determined as described previously (27, 28). The frequency of drug-induced SSBs was expressed as rad-equivalents, i.e., as the γ-radiation dose that would produce the same elution rate (rad-equivalents).

The main goals of this study were to characterize and understand the mechanisms of resistance to the camptothecins in MXR-overexpressing S1-M1-80 and MCF-7 AdVp3000 cell lines, and to investigate whether glucuronidation might contribute to drug resistance in these cells. For comparison we used parental cells and two additional drug-resistant cell lines: Pgp-expressing S1-B1-20 cells and the MRP-expressing MCF-7/VP cell line. MXR is known to confer high levels of resistance to mitoxantrone and to the anthracyclines (2, 4). As shown in Fig. 1, drug-selected S1-M1-80 and MCF-7 AdVp3000 cells are very resistant to mitoxantrone (22,000- and 5,140-fold, respectively), epirubicin (400- and 1750-fold, respectively), and SN-38 (1000- and 667-fold, respectively). The S1-B1-20 cell line, expressing moderate levels of Pgp, also showed cross-resistance to these three drugs, although the relative resistance values were markedly lower (mitoxantrone, 110-fold; epirubicin, 100-fold; and SN-38, 4-fold). The MCF-7/VP cell line, which overexpresses MRP, showed slight resistance to mitoxantrone (1.6-fold) and epirubicin (20-fold) but not to SN-38. These data underscore the similar cross-resistance profiles of S1-M1-80 and MCF-7 AdVp3000 cells, consistent with MXR overexpression as the principal mechanism of resistance.

Fig. 2 shows photomicrographs obtained by confocal microscopy of the steady-state levels and localization of mitoxantrone, epirubicin, and topotecan in the S1 and MCF-7 parental and S1-M1-80 and MCF-7 AdVp3000 resistant cell lines. After a 60-min incubation at 37°C in 5 μm drug, parental cells (S1 and MCF-7) had high intracellular drug levels in contrast to the low or absent staining in the S1-M1-80 and MCF-7 AdVp3000 sublines. In the parental cells, nuclear fluorescence was prominently observed for epirubicin, whereas mitoxantrone fluorescence was seen in nucleoli and in the cytoplasm. Topotecan fluorescence, however, appeared more cytoplasmic than nuclear. A punctuate, perinuclear, staining pattern was also observed in all cell lines, suggesting accumulation of drugs in a vesicular compartment, such as lysosomes and/or the Golgi apparatus. As described previously for mitoxantrone-selected MCF-7 MX cells and for topotecan-selected IGROV cells, topotecan accumulation is reduced relative to the level found in parental cells (8, 9).

The lower fluorescence levels present in the resistant sublines could be increased to parental cell levels by first incubating the cells in energy-depleted conditions (Fig. 3). Pretreatment with sodium azide and deoxyglucose resulted in energy depletion and increased epirubicin and topotecan accumulation in the resistant S1-M1-80 and MCF-7 AdVp3000 sublines, as evidenced by increased fluorescence. Epirubicin accumulation increased dramatically and was observed in both the cytoplasm and nucleus to a comparable extent. For topotecan, both cytoplasmic and nuclear accumulation was increased, with a marked increase in nuclear fluorescence. Enhanced accumulation in the nucleus was also observed with energy depletion of parental cells (not shown). These results imply that the reduced total accumulation observed by confocal microscopy in these cells is due to an energy-dependent transport mechanism, such as MXR. They also demonstrate an effect on nuclear accumulation of topotecan that is energy dependent but unrelated to MXR because accumulation in both parental and resistant nuclei was similarly increased by energy depletion.

To further characterize the resistance to the camptothecins, cytotoxicity assays were performed using a series of camptothecins with substitutions near or at the site of glucuronidation in SN-38 (Fig. 4). These included three compounds that have been evaluated clinically (topotecan, CPT-11, and 9-AC) plus the CPT-11 metabolite, SN-38, and six camptothecin analogues including MDO-CPT and di-F-MDO-CPT (29). Fig. 5,A displays the relative resistance values derived from cytotoxicity assays of these compounds in the two MXR-overexpressing resistant cell lines. Fig. 5 B includes these values and the relative resistance values in the Pgp- and MRP-overexpressing cell lines. To varying degrees, S1-M1-80 and MCF-7 AdVp3000 cells were cross-resistant to all of the camptothecins (14- to 1325-fold). In addition to SN-38, the two resistant cell lines demonstrated marked cross-resistance to topotecan (640- and 1325-fold) and 9-AC (1000- and 667-fold). In contrast, Pgp-overexpressing S1-B1-20 cells were markedly less resistant to all of the compounds, with resistance measured at a maximum of 3.8-fold (SN-38). The MRP-overexpressing MCF-7/VP cell line was not cross-resistant to the camptothecins. These data suggest that MXR expression in the S1-M1-80 and MCF-7 AdVp3000 cells confers significant cross-resistance to a variety of camptothecins.

The cross-resistance profiles demonstrated a large difference in sensitivity to SN-38 compared with camptothecin in the resistant cells. Differential sensitivity to various camptothecins argues against reduced expression or mutation of topoisomerase I as the etiology. However, a drug transporter may have different affinities for different compounds, resulting in diverse levels of drug resistance. Drug transport could decrease the amount of drug available for topoisomerase I inhibition; we thus determined the extent of drug-induced DNA SSBs for camptothecin and SN-38. In Fig. 6, the amount of topoisomerase-induced cleavage complexes detected as protein-linked DNA SSBs by alkaline elution are plotted as rad-equivalents (26). Consistent with the cross-resistance profile demonstrating low cross-resistance to camptothecin and high tolerance to SN-38, camptothecin produced slightly fewer SSBs in S1-M1-80 than in S1 cells, whereas SN-38 produced significantly fewer SSBs in the resistant S1-M1-80 cells. This difference was reversed by the addition of 1 μm GF120918, an agent known to reverse resistance in these cells (11). GF120918 increased the SN-38-induced SSBs in S1-M1-80 cells to a level similar to that in S1 cells (Fig. 6 C). These data are consistent with a transport mechanism of drug resistance conferred by reduced intracellular drug accumulation and argue against a topoisomerase I-mediated resistance mechanism. We also used a sensitive RNase protection assay previously shown to identify point mutations in topoisomerase I (30) but detected no mismatches in RNA from S1-M1-80 and MCF-7 AdVp3000 cells (data not shown). These results suggest that there are no acquired mutations in topoisomerase I in the resistant cells.

Previous studies have shown that SN-38, mitoxantrone, and epirubicin undergo glucuronidation during normal metabolism (12, 13, 14). This observation prompted us to evaluate whether the resistant cells possess the capability to conjugate exogenous compounds with glucuronic acid. UGT activity was evaluated by TLC, using ¹⁴C-labeled UDPGA as described previously (24, 25). Fig. 7,A shows the results of autoradiography of the TLC plates assaying UGT activity for 4-MU, a compound known to be a good substrate for glucuronidation (24). β-glucuronidase hydrolysis was used to confirm that the products observed were indeed glucuronides. Fig. 7,B shows the glucuronides for 4-MU plus SN-38, mitoxantrone, and epirubicin, which possess similar mobilities on the preparative silica gel plate chromatograms. 4-MU and epirubicin appeared to be the most effective substrates. After 2 h of β-glucuronidase hydrolysis, the radioactivity in the region corresponding to the glucuronide for each substrate was markedly diminished. We next compared glucuronidation of 4-MU, SN-38, mitoxantrone, and epirubicin in S1 parental cells and the S1-M1-80 resistant cell line. As shown in Fig. 8,A, the extent of glucuronidation was similar in both cell lines, with greater glucuronidation of 4-MU and epirubicin. The lack of up-regulation in the resistant cells suggests that glucuronidation is not rate limiting for the resistance mechanism in the S1-M1-80 cells. Because the colon cancer cells may have metabolic capabilities that cells derived from other tumor types lack, we examined the UGT activity of the parental and resistant MCF-7 cells, using 4-MU and epirubicin as substrates. Fig. 8 B shows the results of these experiments and compares UGT activity in these cells to that in the colon cancer cells. As can be seen, substantially more radioactivity was incorporated into the 4-MU [¹⁴C]glucuronide in MCF-7 AdVp3000 cells than in parental MCF-7 cells; weak glucuronide formation was observed for epirubicin in the MCF-7 AdVp3000 cells, and none was observed in the parental cells. These results suggested an increase in UGT activity in MCF-7 AdVp3000 cells relative to that in parental MCF-7 cells. However, the activity in both breast cancer cell lines is lower than that observed in the colon cancer cells.

It has been reported that cells resistant to SN-38 have increased UGT activity and also have increased expression of the UGT1A family mRNAs (10). We thus examined UGT1A family mRNA expression, using a quantitative PCR assay with primers able to amplify a 529-bp region localized to exon 5 and common to the nine or more members of the UGT1A family (23, 31). As shown in Fig. 9, MCF-7 AdVp3000 cells had 3.5-fold higher UGT1A mRNA levels than the parental MCF-7 cells, whereas UGT1A expression was not significantly different between parental S1 cells and the S1-M1-80 subline. Interestingly, the absolute difference between the two parental cell lines was not as great as would be suggested by the TLC assay.

The present study describes camptothecin resistance and the potential for drug glucuronidation in two cell lines overexpressing the ABC half-transporter, MXR. These sublines, S1-M1-80 and MCF-7 AdVp3000, show high levels of cross-resistance toward mitoxantrone, epirubicin, the CPT-11 metabolite SN-38, topotecan, and several other camptothecins (32). This cross-resistance is associated with reduced drug accumulation and increased drug efflux. Using TLC, we investigated the ability of the sensitive and resistant cell lines to glucuronidate different compounds. Lysates from both S1-M1-80 resistant cells and S1 parental cells were able to glucuronidate 4-MU and epirubicin with no evidence of increased capacity in the resistant cells. In contrast, enzymatic assays indicated an up-regulation of glucuronidating capacity in the resistant breast cancer cells, and quantitative PCR demonstrated an increase in UGT1A mRNA. These results are consistent with a role for MXR in resistance to the camptothecins as well as a potential role for glucuronidation in the resistant phenotype.

The energy-dependent transport and cross-resistance to a variety of camptothecins confirms and extends the observations in two previous reports of topotecan resistance (8, 9). MCF-MX, a subline which we now know overexpresses MXR, was shown to be >100-fold resistant to topotecan, SN-38, and 9-AC and to have no change in topoisomerase I levels (8). IGROV ovarian cancer cells, selected in topotecan, were shown to have a similar resistance profile with an energy-dependent reduction in topotecan accumulation (9). Just as in the MXR-expressing sublines reported herein, both sublines had minimal resistance to camptothecin. In view of the resistance profile, which also included mitoxantrone, it is likely that the IGROV cells express MXR.

These cell lines expand the subset of multidrug-resistant cell lines and tumors where mechanisms other than Pgp overexpression exist. The most widely studied mechanism of multidrug resistance is that mediated by Pgp, which appears to transport unmodified drugs and xenobiotics (33). However, the multidrug resistance transporter MRP is an organic anion pump able to transport conjugates of sulfates, glutathione, and glucuronic acid (34, 35). It appears to be most efficient in transporting glutathione conjugates, but it has also been suggested that cotransport of glutathione with unconjugated drug substrates occurs (36, 37). The identification of MRP as a drug conjugate transporter suggests that drug metabolism might yet play a role in drug resistance.

Although the importance of glucuronidation in hepatic drug metabolism is well known, it has received little attention as a possible mechanism of resistance in malignant cells. Glucuronides of both SN-38 and topotecan have been identified in patients treated with CPT-11 and topotecan (13, 32, 38). For SN-38, glucuronidation appears to be a clinically important mechanism of detoxification. Not only is there significant interpatient variability, but patients with Gilbert’s disease who have mutations in UGT have a marked increase in CPT-11 toxicity secondary to decreased SN-38 glucuronidation (13, 39). Isolated reports have appeared demonstrating elevated UGT activity in cancer cell lines associated with resistance to daunorubicin (40), mitoxantrone (41), SN-38 (10), and mycophenolic acid (42). In lung cancer cells resistant to SN-38, increased levels of UGT activity were also noted along with enhanced glucuronidation capacity in cell lysates (10).

Although the sensitivity of the assay is limited, the results of the TLC experiments suggest no change in UGT activity in the S1-M1-80 cells because UGT activity in parental S1 cells was as high as that in the resistant cells. However, increased UGT activity was found in MCF-7 AdVp 3000 cells. Furthermore, quantitative PCR demonstrated higher levels of UGT1A mRNA. Together, these observations suggest that resistant cells may acquire increased levels of UGT activity. The difference between S1-M1-80 and MCF-7 AdVp3000 is likely explained by the fact that the S1 cells are colon carcinoma cells and may have intrinsically higher UGT activity, as suggested in previous studies with mycophenolic acid in unselected colon cancer cells (42). The expression of UGT1A in colon carcinoma cells is supported by demonstration of UGT1A isoform expression in normal human colon cells (43). Interestingly, there was little absolute difference in UGT1A mRNA between the two parental cells. RNA studies for specific isoforms need to be performed to precisely identify the up-regulated isoform in the breast cancer cells.

A second line of evidence suggesting a role for glucuronidation in the resistance phenotype of the two cell lines can be found in the series of camptothecin analogues tested. Both MXR-expressing cell lines, but especially the S1-M1-80 cells, are preferentially resistant to camptothecins which can be glucuronidated. S1-M1-80 cells, for example, are very resistant to SN-38, topotecan, and 9-AC and have intermediate resistance to CPT-11 and NSC 369395. These compounds contain, or can be metabolized to contain, functional groups that can be glucuronidated. S1-M1-80 cells are markedly less resistant to camptothecin, MDO-CPT, di-F-MDO-CPT, NSC 609959, and NSC 369394, compounds that do not contain functional groups for glucuronidation. This distribution is not perfect, however. NSC 609960 contains a methyl group at the 10 position, and thus should not be a substrate for glucuronidation. However, S1-M1-80 cells are as resistant to NSC 609960 as to CPT-11. MCF-7 AdVp3000 cells display intermediate resistance to several of the compounds lacking sites for glucuronidation. In sum, the data implicating glucuronidation in the resistance phenotype of MXR remain suggestive but not conclusive.

Proof of the presence of glucuronidation in these cells requires demonstration of a drug-glucuronide conjugate in the medium. Such a result would support the notion that glucuronidation may be an intrinsic detoxification mechanism active in colon cancer cells, but which may be superceded by the overexpression of MXR in the setting of intense drug selection. Glucuronidation does not appear to be important in the breast cancer cells. The overriding importance of the MXR transporter in the resistance phenotype is implied by the development of gene amplification in the case of MCF-7 AdVp3000 and of gene rearrangement in S1-M1-80, conferring MXR overexpression (2, 44).

The results described herein expand our understanding of the MXR phenotype by adding a range of camptothecins to its multidrug-resistant phenotype. Except for topotecan transport by Pgp, which appears to be relatively ineffective, prior studies have not identified a drug efflux pump for the camptothecins. Further studies, including mouse knock-out studies, should help define the role of MXR in normal cells and its mechanism of action in drug resistant cells. Regardless of the role of glucuronidation, it is clear that MXR confers resistance to a variety of camptothecins, and the evidence suggests that this is mediated by enhanced drug efflux. Thus, we can add the camptothecins to the growing list of agents whose activity can be mediated by drug efflux pumps.