M.C. Hayes, Department of Urology, Southampton General Hospital, Tremona Road, Southampton SO16 6YD, UK.
Objective To compare multidrug resistance (MDR)-1 and MDR-3 gene expression in a new urothelial cancer cell line (MGHU-1, with resistance to mitomycin C) against controls and the established (epirubicin-resistant) MDR clone, and to correlate MDR with cytotoxicity data.
Materials and methods Resistance to mitomycin C was induced by the long-term exposure of wild-type MGHU-1 cells to increasing concentrations (20–400 nmol/L) of mitomycin C. The cytotoxicity of mitomycin C or epirubicin was then compared in MGHU-1, MGHU-MMC (mitomycin C-resistant) and MGHU-1R (established MDR) cells, using the tetrazolium biomass assay. The expression of MDR-1 and -3 was investigated by the reverse transcriptase-polymerase chain reaction, using cDNA-specific primers after titration, and compared with DNA and negative controls.
Results MDR-1 and -3 were significantly and equally overexpressed in MGHU-1R, and associated with a dramatic increase in the 50% inhibitory drug concentration (P < 0.001) for mitomycin C and epirubicin against controls. In MGHU-MMC, the overexpression of MDR-1 was three times greater than that of MDR-3. The cytotoxicity profile for both agents was very similar to that of MGHU-1R. Trace amounts of MDR-1, but not MDR-3, were identified in the MGHU-1 wild-type.
Conclusions Urothelial cancer cell resistance to mitomycin C is associated with cross-resistance to epirubicin and overexpression of MDR-1, suggesting that mitomycin C falls within the MDR category. Clinical application of this methodology may allow patients to be identified who are unlikely to benefit from intravesical chemotherapy.
In recent years there has been increasing interest in rationalizing the role of intravesical cytotoxic drugs in the treatment of superficial TCC of the bladder. Although there is no substantive evidence that progression of disease in terms of grade or stage can be influenced by these agents, it is well established that the adjuvant intravesical administration of epirubicin or mitomycin C can reduce tumour recurrence rates after resection by 40–60% [1,2].
The mechanisms of action whereby these drugs exert cytotoxic effects remain poorly understood. Both agents cause DNA damage via intracellular free radical production and directly by interstrand intercalation, and are thus maximally effective during the S-phase of the cell cycle. Epirubicin is independently cytotoxic, whereas mitomycin C requires intracellular modulation via a bioreductive alkylation reaction  (catalysed by DT-diaphorase) to the active hydro/semiquinone metabolite to be cytotoxically effective. There is evidence that resistance to agents such as mitomycin C is mediated partly by failure of its activation via the DT-diaphorase and NADPH cytochrome P450 reductase pathways .
Failure of a tumour to respond to epirubicin (an anthracycline) and to some other structurally unrelated drugs (e.g. Vinca alkaloids, actinomycin D) is mediated in significant part by the phenomenon of multidrug resistance (MDR), whereby cytotoxic agents are actively excluded from compartments of the intracellular environment . Transmembrane (P-glycoprotein, P-gp , and MDR-related-protein ) and intracellular proteins (lung resistance protein)  have been implicated in this phenomenon, which is reversible in many in vitro cancer model systems . Two genes coding for P-gps have been identified in humans, i.e. MDR-1  and MDR-3 . MDR-1 codes for P-gp 1 and MDR-3 for P-gp 2, as a confusing result of a quirk of nomenclature. Only MDR-1 expression, and therefore P-gp 1 function, has been associated with resistance to lipophilic drugs in humans. The MDR locus is on chromosome 7q21.1 .
Despite its comparable clinical resistance pattern, the MDR status of mitomycin C remains equivocal. There is no published consensus on this, although MDR characteristics and the expression of membrane P-gp have been shown in, e.g. a leukaemia cell line (L1210) which had cross resistance with mitomycin C, anthracyclines and Vinca alkaloids .
We induced resistance to mitomycin C in an established TCC cell line and applied established cytotoxicity assays and PCR technology to ascertain whether this agent falls within the MDR family, and by implication examined the possibility of cross-resistance to both epirubicin and mitomycin C.
Materials and methods
The MGH-U1 urothelial cancer cell line was kindly supplied by Prof John Masters at the Institute of Urology, University College Hospital, London, in its epirubicin-sensitive (‘wild-type’ MGH-U1) and MDR (MGH-U1R) subclones (the latter previously achieved by continuous exposure to doxorubicin). Cells were grown in Dulbecco's modification of Eagles medium (DMEM) supplemented with 10% fetal calf serum, with glutamine and antibiotics, at 37 °C in humidified 5% CO2 in air. The line was subcultured using trypsin-EDTA.
To induce mitomycin C resistance, cells from the wild-type clone (MGH-U1) were cultured as above in the presence of gradually increasing concentrations of mitomycin C (20–400 nmol/L) as a maintenance concentration, for 6 months. The lines were intermittently subcultured from the outset to ensure constant clonality within the cell line (hereafter named MGHU-MMC).
Cytotoxicity was assessed using the MTT assay (thiazolyl blue; 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide). Cells from lines MGH-U1, MGH-U1R and MGHU-MMC were taken into suspension, seeded into 96-well culture-grade microtitre plates (Nunc, UK) and allowed to settle overnight (initial plating concentration 103/well) in standard conditions using 100 µL/well of supplemented DMEM. Dilutions of epirubicin and mitomycin C were made up (from stock solutions, diluted in supplemented HEPES-buffered DMEM) to obtain final assay concentrations of 0.6, 1.25, 2.5, 5.0, 10, 20, 40 and 80 µg/mL. Cells were then exposed to either agent and incubated at 37°C for 1 h. Cytotoxic drug-containing medium was then removed and the cells cultured for a further 72 h in drug-free DMEM. Subsequently the plates were incubated with MTT (0.2 mg MTT/mL HEPES-buffered DMEM, 250 µL/well) for 4 h and the resultant formazan crystals solubilized with 100% DMSO. The optical density (representing remaining viable cell biomass) was measured using a microplate reader and calculated as a percentage against control wells containing cells which had not been exposed to either cytotoxic agent.
MDR gene expression
Three gene specific primers were designed to assess the expression of MDR-1 and -3 in MGH-U1, MGH-U1R and MGHU-MMC lines. A common sequence upstream primer (MDR-U) was used in the assessment of both genes. The primer sequences for MDR-1, MDR-3 and MDR-U were TTAgACAgCCTCATATTTTg, TCAgACAACCTCAAA-TCCTC and TTCTggATggTggACAggCg, respectively. Cells from each line were trypsinized, taken into suspension, washed and pelleted in Eppendorf tubes. mRNA was extracted from each line using the RNEasy™ kit (Qiagen, West Sussex, UK). RT-PCR was used to construct cDNA therefrom and its integrity confirmed by α-actin PCR. MDR-1 and MDR-3 products were obtained by PCR with the gene-specific primers, products ligated to pGEM-T (Promega, Madison, WI) vector and OneShot™ bacteria transformed to produce MDR-1 and MDR-3-containing clones. High-purity DNA was extracted from both clone types using Qiagen™ preparations and restriction digests obtained. The resulting products were run with 3 µL bromophenol blue against 5 µL of φX Hae III DNA marker on a 8% polyacrylamide gel at 40 mA for 2 h and stained with ethidium bromide. DNA products were thereafter sequenced (ThermoSequenase™, Amersham, UK).
The expressions of MDR-1 and MDR-3 mRNA (coding for P-gp 1 and P-gp 2, respectively) by MGH-U1, MGH-U1R and MGHU-MMC are shown in Fig. 1. The expected product size for both MDR-1 and MDR-3, using the primers designed for this project, was 145 bp. Figure 1 shows that the product length falls about halfway between 118 and 194 bp, according to the standard Hae III DNA marker. MDR-1 was significantly overexpressed by both MGH-U1R and MGHU-MMC; it was also present in low concentrations in wild-type MGH-U1 but absent in the control DNA and, by definition, in the negative controls.
In contrast, MDR-3 was significantly overexpressed by MGH-U1R only; it was present, but in very low concentrations, in MGHU-MMC, and seemed to be entirely absent in wild-type MGH-U1. Again it was absent in negative controls and control DNA. A representative example of part of the MDR-1 and MDR-3 sequencing gels is shown in Fig. 2.
Cytotoxicity of epirubicin and mitomycin C
Figure 3 shows the effect of increasing doses of either epirubicin or mitomycin C on cultured cells from the cell lines MGHU-MMC, MGH-U1 and MGH-U1R. In each the abscissa is not linear, denoting drug concentrations obtained by serial dilution. The dotted lines on each plot indicate the concentration of cytotoxic drug at which half the cell population is killed (IC50). Figure 3a,b shows that although there was a comparable dose-related cytotoxic response in MGHU-MMC cells to both epirubicin and mitomycin C, it was negligible, and the IC50 was not reached for either drug at the concentrations used. This contrasts dramatically with the effects of the same concentrations of drug in MGH-U1 cells (Fig. 3c,d) where the IC50 for mitomycin C was reached at 1 µg/mL and that for epirubicin at 2 µg/mL. Interestingly, in the MGH-U1R cell line (Fig. 3e,f) there was a similar pattern to that in MGHU-MMC, whereby a dose response was still present but the IC50 for epirubicin was only approached at 80 µg/mL. However MGH-U1R cells seemed to be slightly more sensitive to mitomycin C, where the IC50 was reached at 35 µg/mL.
Classical P-gp-mediated MDR undoubtedly modulates resistance to the anthracyclines  and the failure of alkylation (drug activation) is primarily responsible for much resistance to mitomycin C . However, despite its comparable clinical resistance pattern the status of mitomycin C in the MDR family, and the extent to which MDR mechanisms modulate resistance to it, remains equivocal, and published reports are inconclusive.
In addition to the leukaemia cell line cited previously , the P-gp-expressing subline of K562, a doxorubicin-resistant erythroleukaemia clone, has cross-resistance to epirubicin and mitomycin C . Alterations in topoisomerase II expression (which facilitates DNA repair mechanisms) have also been identified in association with adriamycin-induced MDR in Chinese hamster ovary cells which were cross-resistant to topoisomerase II inhibitors . These cells were also more than 80 times as resistant to mitomycin C than the parental cell line. A resistant subline selected from rat hepatoma AH130 cells after exposure to adriamycin also showed marked cross-resistance to mitomycin C, porfiromycin and vinblastine . The use of 1,3,5-triazacycloheptanes as MDR-reversing agents was investigated for P-gp-mediated MDR in mouse P388 leukaemia cells ; these molecules potentiated the cytotoxic effect of vinblastine, adriamycin and mitomycin C in these cells.
On the other hand, the overexpression of π-type glutathione transferase, implying enhanced intracellular drug conjugation/inactivation, has been shown in a bladder cancer clone (derived from J82) which has cross-resistance between mitomycin C and cisplatin. The latter is not a drug to which resistance is mediated by classical (P-gp-associated) MDR . Ohga et al. showed that a nuclear factor (MDR-NF1) which is involved in the regulation of the MDR-1 gene is identical to the Y box-binding protein (YB-1) which binds to the control region of many genes. Transfection of a YB-1 antisense construct into human epidermoid cancer KB cells resulted in increased sensitivity to cisplatin and mitomycin C, but not to the established MDR drugs vincristine, doxorubicin or etoposide. Shibata et al. established two mitomycin C-resistant sublines of PC-9 (non-small cell lung cancer) by continuous exposure. These were not resistant to the MDR drugs vindesine, etoposide and adriamycin, amongst others, and the authors concluded that deficient drug activation (shown by the decreased cytosolic DT-diaphorase activity) was the resistance mechanism.
In the present experiments, resistance to mitomycin C was induced in an established TCC line, and established cytotoxicity and RT-PCR assays were used. By implication the possibility of cross-resistance to both epirubicin and mitomycin C was sought in a bladder cancer model. Resistance to mitomycin C was invoked by persistent low-dose exposure (400 nmol/L) of the agent to wild-type (non MDR) MGH-U1 cells in culture. This is very different to the intermittent single high-dose (1 mg/mL) exposure seen in the clinical scenario. Clearly, it is speculative whether the development of MDR in these circumstances represents the operation of ‘instructive’ or ‘selective’ mechanisms, i.e. direct gene upregulation in previously sensitive cells, or clonal expansion of a previously small population of already resistant cells. Nevertheless, from the cytotoxicity data it seems that persistent low-dose exposure not only results in a significant increase in the IC50 of MGHU-MMC for mitomycin C, but also for epirubicin. Such data obtained from one cell-line type may not be reproducible in other bladder cancer lines and must therefore be viewed in this light.
The presence of detectable levels of MDR-1 in wild-type cells is to be expected, as it can be identified by RT-PCR in up to 70% of bladder tumours and half the samples of normal urothelium . The overexpression of MDR-1 and MDR-3 in MGH-U1R, but only of MDR-1 in MGHU-MMC, may represent the interaction of different regulatory elements or binding affinities for each drug by the cell line. Alternatively, it is possible that overexpression of MDR-3 in addition to MDR-1 in MGH-U1R is a reflection of established functional MDR, in contrast to the newly induced MDR in MGHU-MMC. However, it is unclear why dual MDR-1 and MDR-3 expression is associated with a slight reduction in chemoresistance in MGHU-1R when compared to MGHU-MMC. This is probably mediated by other mechanisms, e.g. alterations in topoisomerase regulation or the process of apoptosis induction, although this remains speculative.
Intracellular epirubicin concentrations (as detected by flow cytometry) and localization patterns (determined by confocal microscopy) are very similar both in MGHU-MMC and in MGHU1-R cells (which show established P-gp-mediated epirubicin resistance). The reversibility of these patterns in the presence of verapamil (an MDR reversal agent) is similar both for MGH-U1R and MGHU-MMC (personal communication, Dr A J Cooper).
Anecdotally, it seems that some patients with recurrent tumour after exposure to one agent will benefit from substitution single-agent intravesical chemotherapy, although no clinical trials to date have sought to validate this practice. However, this suggests that an element of cross-resistance between these agents must be expected. Such suspicions are supported by the similar clinical response/recurrence rates seen in most clinical trials of both agents. However, it also shows that the phenomenon of MDR reversal, seen in the presence of verapamil and other agents which modulate P-gp-mediated drug resistance, may be applicable both to epirubicin and mitomycin C.
The complexity of drug-resistance mechanisms and the polyclonal nature of many bladder tumours makes it unlikely that this assay type alone will allow the identification of patients likely, or indeed unlikely, to respond to intravesical chemotherapy. The ability to do so would clearly be an important advance in the management of such patients and should be pursued with vigour.