Breast Cancer Resistance Protein Is Localized at the Plasma Membrane in Mitoxantrone- and Topotecan-resistant Cell Lines¹

Tumor cells can be intrinsically resistant to drugs or they can acquire resistance to structurally and functionally unrelated drugs on drug exposure. This phenomenon is known as MDR³(reviewed in Ref. 1). In human tumor cells, several transporter proteins can be involved in MDR. These proteins—MDR1 P-gp(ABCB1; reviewed in Ref. 2), MRP1 (ABCC1; reviewed in Ref.3), MRP2 (ABCC2; Ref. 4), MRP3 (ABCC3; Ref.5), and BCRP (ABCG2; Ref. 6)—all belong to the ABC transporter family (7). They act as efflux pumps,which result in decreased intracellular concentrations of cytotoxic drugs.

BCRP is a recently discovered half-transporter that probably acts as a homo- or heterodimer in transporting cytotoxic agents (6). The transporter molecule is capable of transporting several anticancer drugs but has thus far been found mainly in MX-resistant cell lines (8, 9).

To date, all studies on BCRP expression have reported on BCRP mRNA levels. Because no polyclonal or monoclonal antisera that would detect BCRP are yet available, studies at the protein level have not yet been described. Therefore,information regarding the presence and localization of BCRP in (tumor)cells is still lacking.

The MCF-7 MR breast cancer cell line is one of several MX-resistant cell lines described with a non-P-gp, non-MRP1 phenotype and elevated levels of BCRP mRNA (8). To characterize the resistance mechanism in cells with these characteristics, we set out to produce Mabs reactive to proteins elevated in this cell line, as compared with sensitive cells. Mice were immunized with MCF-7 MR cells and, using a cytospin-based screening method with MX-resistant and-sensitive cell lines, we isolated a Mab named BXP-34 that specifically reacts with the BCRP protein. To study the presence and subcellular localization of BCRP in human tumor cell lines and tumor samples,panels of parental and MX-, TPT-, and multidrug-resistant cell lines as well as primary and chemotherapy-treated breast cancer and AML samples were tested for BCRP using the BXP-34 Mab.

All of the cell lines that we used have been described previously:(a) the drug-sensitive breast cancer cell line MCF-7, the Dox-selected subline MCF-7 Dox40, the MX-selected subline MCF-7 MR, and the mock- and BCRP (clone 8)-transfected sublines (6, 10); (b) the myeloma cell line 8226, the Dox-selected subline 8226 Dox40, and the MX-selected subline 8226 MR20(11); (c) the ovarian carcinoma cell line Igrov1, the TPT-selected subline T8, the partial revertant of this cell line, the T8rev, and the MX-selected subline MX3 (9);(d) the ovarian carcinoma cell line A2780 and the Dox-selected subline 2780AD (12); (e) the non-small cell lung cancer cell line SW1573 and the Dox-selected sublines SW1573/2R120 and SW1573/2R160 (13);(f) the small cell lung cancer cell line GLC4, the Dox-selected GLC4/ADR subline, and the MX-selected subline GLC4 MIT(14); and (g) the leukemia cell line HL60 and the Dox-selected HL60/ADR subline (15). The ovarian carcinoma cell line 2008 and the MRP1-,MRP2-, and MRP3-transfected sublines were described in Ref. 5 and by Scheffer et al.⁴All of these cell lines are human cell lines. The mouse fibroblast cell line MEF3.8, the MX-selected subline M32, and the TPT-selected subline T6400 were described in Ref. 16. Monkey kidney CV-1 cells,transformed by an origin-defective mutant of SV40 that codes for wild-type T antigen, COS7 cells, were described in Ref.17.

All of the cell lines were grown in Dulbecco’s modified essential medium or RPMI (Life Technologies, Inc. Europe, Paisley, Scotland),supplemented with 10% heat-inactivated FCS, 2 mml-glutamine, penicillin, and streptomycin. Resistant cell lines were cultured in the presence of drugs until 3–10 days before the experiments. All of the cells were negative for Mycoplasma as tested by the Gene-Probe rapid Mycoplasma detection system (Gene-Probe, San Diego, CA).

Three 8–12-week-old female BALB/c mice (Harlan, Zeist, the Netherlands) received footpad injections of sonicated MCF-7 MR cells emulsified in Freund’s complete adjuvant (Difco, Detroit, MI). Approximately 2 × 10⁶ cells were used per injection. After 10 days, the animals received a first booster injection with sonicated MCF-7 MR cells in PBS. Similar booster injections were given at day 20 and 30. Four days after the last booster, 3 days before fusion, a final booster injection was given. The mice were housed and treated in accordance with current regulations and standards of the Institutional Animal Ethics Committee.

The mice were killed and draining popliteal lymph nodes were removed and used for fusion with mouse myeloma Sp2/0 cells as described previously (5). Hybridoma supernatants containing monoclonal antibodies were screened on octo-spins containing eight cytospins of a mixture of MCF-7 MR and MCF-7 parental cells per slide. Antibody binding was detected as described in the“Immunohistochemistry” section. Hybrid cells that secreted antibodies of interest were selected and subcloned three times by limiting dilution. The isotype of the selected Mabs was determined using IsoStrips (Boehringer Mannheim).

Cytospin preparations and cryosections (4 μm) were air-dried overnight and fixed for 7 min in acetone at room temperature. The slides were incubated with undiluted hybridoma supernatant for 1 h at room temperature. Biotinylated rabbit-antimouse serum (1:150, Zymed,San Francisco, CA) and HRP-labeled streptavidin (1:500, Zymed), diluted in PBS/1% BSA, were used as secondary reagents. Color development was with 0.4 mg/ml AEC and 0.02%H₂O₂ as a chromogen.

Monkey kidney COS7 cells were transfected with pcDNA3-BCRP(6) or (control) pCDM8-LRP plasmids by the DEAE dextran (Promega Corporation, Leiden, the Netherlands) method as described by Aruffa and Seed (18). Three days after transfection, the cells were harvested, cytospin preparations were made, and transient gene expression was examined using BXP-34 and(control) LRP-56 Mabs.

A panel of human tumor samples comprising tumors of different origin and relative drug sensitivity was selected from our frozen tissue bank(see Table 1). For most tumor types, two patients were selected. The samples were mainly primary untreated adenocarcinoma samples. In the breast cancer group, more patients were included. Besides 10 patients with untreated adenocarcinoma, 7 patients with locally advanced breast cancer were entered. These latter patients had received standard combined Dox/CyPhos chemotherapy (six cycles with dosages starting at 90 mg/m² Dox and 1000 mg/m²CyPhos, decreasing to 75 mg/m² Dox and 750 mg/m² CyPhos in the last period).

Cytospins from paired samples from 8 patients with AML de novo and at relapse were made. Patients were treated with standard induction chemotherapy treatment that consisted of a first cycle of daunorubicin (45 mg/m²) or idarubicin (12 mg/m²) plus 1-β-d-arabinofuranosylcytosine (200 mg/m²) for 7 days, followed by a second cycle of amsacrine (120 mg/m²) plus 1-β-d-arabinofuranosylcytosine (2000 mg/m²) for 6 days and a third cycle of MX (10 mg/m²) plus etoposide (100 mg/m²) for 5 days (6 patients) or an autograft or HLA-matched allograft (2 patients).

To study the transporter(s) involved in MX-selected, non-P-gp non-MRP1 MDR cell lines, we set out to produce Mabs detecting proteins up-regulated in MX-resistant cell lines. Because BCRP mRNA levels were reported to be up-regulated in these cell lines, particular attention was given to this transporter. Hybridoma supernatants obtained from mice that had been immunized with BCRPmRNA-positive breast cancer MCF-7 MR cells were screened on cytospins for reactivity with MCF-7 MR cells and absence of reactivity with parental MCF-7 cells and P-gp-positive, BCRP mRNA-negative MCF-7 Dox40 cells. Several clones were selected and were further examined for reactivity with a similar panel of 8226 myeloma cell lines. This approach caused us to discard most of the clones because they failed to react with MX-resistant, BCRP mRNA-positive myeloma 8226 MR20 cells but resulted in the isolation of one Mab,called BXP-34, most likely to detect the BCRP protein (Fig. 1 A).

Specificity of reactivity of the BXP-34 Mab to BCRP was confirmed by reactivity with monkey kidney COS7 cells transiently transfected with BCRP cDNA (Fig. 1,B). The Mab does not cross-react with known MDR transporters like P-gp, MRP1, MRP2, or MRP3, as shown by the absence of staining with cells overexpressing these transporter molecules (Table 1). The BXP-34 Mab was unreactive in Western blots with protein extracts of BCRP-positive cells (not shown), which indicates that the Mab reacts with a nonlinear epitope of the protein. The Mab was unable to stain viable unfixed BCRP-positive cells, which showed that the Mab detects an internal epitope of the BCRP protein. The Mab was also unreactive on cytospins of Bcrp mRNA-positive mouse M32 and T6400 cells,which showed that the BXP-34 epitope is not preserved in mouse Bcrp. Isotyping of the Mab showed that BXP-34 is of IgG1 subclass.

Using the BXP-34 Mab on cytospins and frozen sections of a panel of human tumor cell lines and tumor samples, we examined the presence of BCRP (Figs. 1,2 and Tables 1,2). First, cytospins of a panel of parental cell lines, cell lines selected for resistance to TPT, MX, or Dox, and cell lines transfected with MDR transporters were examined. As shown in Table 1, the BXP-34 Mab detected no, or very low, amounts of BCRP in the parental cell lines or in resistant cell lines with high levels of P-gp or MRP1. In contrast, in the MX-selected MCF-7 MR, the 8226 MR20, and the MX3 cell lines as well as in the TPT-selected T8 cell line, elevated levels of BCRP were detected. Staining of the cell membranes of these cells is most pronounced, with some additional cytoplasmic staining. BCRP levels in the T8rev, a partial revertant of the T8 cell line, were markedly decreased, in accordance with decreased BCRP mRNA levels and reduced resistance to TPT and MX (9). No staining was observed in the MX-selected GLC4 MIT cells. Also, no BXP-34 staining was observed in the 2008 sublines transfected with MRP1, MRP2, or MRP3. In the BCRP-transfected MCF-7/BCRP (clone 8) cells, as expected, clear membranous staining of BCRP was observed.

Then, BCRP levels were examined in cryosections of a panel of primary and chemotherapy-treated human tumor samples. As shown in Table 2, BCRP proved to be undetectable in all of the samples examined, except for a case of small-intestinal adenocarcinoma, in which weak staining of the apical membrane of the epithelial cells of the neoplastic crypts was observed (Fig. 2,A). In some tumor samples, occasional staining of endothelial cells and desmoplastic stroma and inflammatory cells was observed (Fig. 2, B and C). Clear positive staining of tumor cells was observed in control cryosections from normal human kidney, preinjected with a cell suspension of MCF-7 MR and parental cells (Fig. 2 D). In breast tumor samples of patients treated with combined Dox/CyPhos chemotherapy and in cytospins of de novo and relapsed AML patients, no detectable levels of BCRP were observed.

To date, of at least two ABC proteins, MDR1 P-gp and MRP1, the contribution to MDR in human tumor cells is well established. Some family members, including the recently cloned BCRP protein, are also able to transport anticancer drugs (5, 6, 19, 20). Whether they also play a role in drug resistance in patients remains to be established. We have recently described a panel of transporter-specific Mabs, facilitating such investigations.⁴ Extending our studies we set out to produce Mabs detecting the putative MX transporter, the BCRP protein. Using MX-resistant, BCRPmRNA-positive, MCF-7 MR cells for immunization and a cytospin-based screening system with MX-resistant and -sensitive cells, we selected one BCRP-specific Mab, BXP-34. Several other Mabs that were initially selected for reactivity to MCF-7 MR cells were not reactive to other MX-resistant cell lines. These results indicate that (a)human BCRP is only weakly immunogenic in mice; and (b)proteins other than BCRP are selectively up-regulated in MCF-7 MR cells. Notwithstanding, using the BXP-34 Mab, all of the MX- and TPT-selected human cell lines tested were clearly found to be BCRP-positive, except for the GLC4 MIT cell line. In these cells, other mechanisms of MX resistance should be operative. The absence of BCRP mRNA as well as changes other than BCRP overexpression were described in some MX-resistant cell lines (8, 11). As anticipated from drug efflux studies (6, 9, 11), we show here that BCRP is localized most prominently at the plasma membrane rather than at internal vesicular membranes. The BCRP transporter is,therefore, more likely to be involved in active transport from the cell than in transport into internal vesicles.

Except for a case of small-intestine cancer that was weakly positive,cryosections of a panel of human primary tumors were BCRP-negative. Even in tumors, such as renal adenocarcinoma, that have a relatively high level of intrinsic resistance, BCRP could not be detected. Moreover, also in samples from patients with drug-treated breast cancer or AML, no BCRP was detected. Although these data may point to a limited clinical relevance of BCRP in drug resistance, it should be taken into account that the breast cancer patients had shown a pronounced decrease in tumor mass after chemotherapy. Furthermore, in the AML patients, recurrence was independent of the drug types used,and, therefore, the results in the tumor cells tested thus far may not reflect true drug resistance. Because no BCRP mRNA data in human tumors have been published until now, we cannot compare the present results to data reported earlier. BCRP mRNA data should, however, be considered with caution in light of the presently observed BCRP staining in endothelial cells and desmoplastic stroma cells in the tumor samples.

In conclusion, our results indicate that BCRP is highly expressed in MX- and TPT-selected cell lines but not, or only at very low levels, in human tumors. Nevertheless, to reveal the potential contribution of BCRP to MDR, more extended studies are required. The BXP-34 Mab may be a valuable tool in these studies. Nevertheless, other BCRP-specific Mabs will be required that will allow BCRP detection with techniques suited for large-scale screening of tumor samples. We are currently making efforts to obtain such Mabs. Furthermore, studies are ongoing to examine the detailed normal tissue distribution of the BCRP protein. These studies may give clues regarding tumors that arise from BCRP-positive tissues and that are perhaps more likely to use the BCRP protein as a drug resistance mechanism.