Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

At present, seven nucleoside or nucleotide reverse-transcriptase inhibitors (NRTIs), three nonnucleoside reverse-transcriptase inhibitors (NNRTIs), and six protease inhibitors (PIs) are available for the treatment of human immunodeficiency virus type-1 (HIV-1) infection. Highly active antiretroviral therapy (HAART) with these inhibitors has achieved high-level suppression of viral load in HIV-1-infected patients. The emergence of drug-resistant HIV-1 mutants during long-term HAART may result in the failure of therapy (Berger et al., 1998). However, some patients showed a sign of drug-resistance in the absence of drug-resistant viruses (Groschel et al., 1997). Investigations on the host cellular factors responsible for the resistance to antiviral agents revealed that the increased expression of several ATP-binding cassette (ABC) transporters might play a role in drug resistance to anti-HIV-1 agents. It was reported that overexpression of the ABC transporter P-glycoprotein (P-gp/ABCB1) was associated with the reduced antiviral activity of zidovidine (AZT) against HIV-1 replication (Antonelli et al., 1992). This transporter was also shown to interact with HIV-1 PIs and reduce their therapeutic efficacy (Washington et al., 1998). In addition, it was demonstrated that overexpression of multidrug resistance protein (MRP/ABCC) 4 severely impaired the antiviral activity of AZT and other NRTIs, including phosphonylmethoxyethyladenine (Schuetz et al., 1999).

Recently, breast cancer resistance protein (BCRP/ABCG2), a new member of the ABC transporter superfamily, was identified in the atypical multidrug-resistant human breast cancer cell line MCF-7, which was selected in the presence of doxorubicin (DOX) and verapamil (Doyle et al., 1998). BCRP is the second member of the G (white) subfamily of ABC transporters and is also known as the mitoxantrone resistance protein MXR (Miyake et al., 1999) or the placental ABC transporter ABCP (Allikmets et al., 1998). This glycosylated plasma membrane protein is a half-size transporter, which is evolutionarily distinct from other full-size ABC transporters (Rocchi et al., 2000). Cells overexpressing BCRP show resistance to mitoxantrone and, to a lesser extent, to DOX, daunorubicin, and topotecan. However, it has not been shown whether BCRP interacts with anti-HIV-1 agents and affects their antiviral activity and cytotoxicity.

In this study, we established the DOX-resistant CD4⁺ T-cell line, which expresses BCRP but not other multidrug-resistant proteins. Using this cell line, we demonstrated that a high level of BCRP expression in CD4⁺ T cells brings about reduced anti-HIV-1 activity of NRTIs, such as AZT and lamivudine (3TC).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Compounds. DOX, cisplatin, vincristine, etoposide, paclitaxel, mitoxantrone, and actinomycin D were purchased from Sigma (St. Louis, MO). 7-ethyl-10-[4-(1-piperidino)-1-piperidino]carbonyloxycamptothecin and 7-ethyl-10-hydroxycamptothecin were obtained from Daiichi Pharmaceuticals (Tokyo, Japan). AZT, stavudine, and didanosine were purchased from Sigma. 3TC and the NNRTIs emivirine (Baba et al., 1994) and nevirapine were synthesized by Mitsubishi Chemical Corporation (Yokohama, Japan). The PIs nelfinavir (NFV) and indinavir were provided by Japan Tobacco (Takatsuki, Japan) and Takeda Pharmaceutical Industries (Osaka, Japan), respectively. The BCRP-specific inhibitor fumitremorgin C (Rabindran et al., 1998, 2000) was a generous gift from Dr. Rabindran (Wyeth-Ayerst Research, Pearl Liver, NY).

Cells and Virus. The human CD4⁺ T-cell MT-4 cells (Miyoshi et al., 1982) were grown and maintained in RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum, 100 units/ml penicillin G, and 100 µg/ml streptomycin (culture medium). The DOX-resistant cell lines MT-4/DOX₁₀₀ and MT-4/DOX₅₀₀ were established by exposing MT-4 cells to increasing concentrations of the compound. MT-4/DOX₁₀₀ and MT-4/DOX₅₀₀ cells were maintained in the presence of 100 and 500 ng/ml DOX, respectively. Before cytotoxicity and antiviral assays were performed, MT-4/DOX₁₀₀ and MT-4/DOX₅₀₀ cells were cultured in the absence of DOX for at least 7 days. HIV-1_IIIB was used for the infection of MT-4 cells. The virus was propagated and titrated in MT-4 cells and stored at 80°C until use.

Cytotoxicity Assay. The cells (1 × 10⁵ cells/ml) were cultured in the presence of various concentrations of test compounds. After a 4-day incubation at 37°C, the number of viable cells was determined by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) method (Pauwels et al., 1988).

Determination of Intracellular DOX and Effect of ATP-Depletion. Intracellular accumulation and retention of DOX in MT-4, MT-4/DOX₁₀₀, and MT-4/DOX₅₀₀ cells were determined by a slight modification of the flow cytometric method described previously (Krishan et al., 1997). Briefly, the cells were exposed to 10 µg/ml of DOX for up to 180 min (accumulation phase), washed with ice-cold phosphate-buffered saline (PBS), and resuspended in warm culture medium in the absence of the compound (retention phase). At certain intervals, the cells were examined for their intracellular DOX concentrations by the use of flow cytometry (FACScan, Becton Dickinson, Franklin Lakes, NJ). The intracellular drug concentration was expressed as the mean channel fluorescence. To determine the effect of ATP depletion on the intracellular accumulation of compounds, the cells were incubated in glucose-free medium containing 50 mM 2-deoxy-D-glucose and 15 mM sodium azide for 20 min at 37°C (Doyle et al., 1998). Rhodamine 123 (100 ng/ml) (Sigma) was added and further incubated for 30 min. After washing with glucose-free medium, the cells were incubated under ATP-depleting conditions for an additional 30 min on ice, and rhodamine retention was determined by the use of flow cytometry.

Preparation of Crude Membrane Fractions, Cytosols, and Nuclear Extracts. The preparation of crude membrane fractions, cytosols, and nuclear extracts from MT-4 and the DOX-resistant MT-4 cells was described previously (Nakagawa et al., 1992; Grant et al., 1994). To prepare the crude membrane fractions, the cells were washed with 1% aprotinin-containing PBS and treated with lysis buffer [10 mM KCl, 1.5 mM MgCl₂, 10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mM p-amindinophenylmethanesulfonylfluoride, and 2 µg/ml aprotinin]. After 10 min on ice, the cells were homogenized with approximately 80 strokes with a Dounce homogenizer. The intact cells and nuclei in the homogenate were removed by centrifugation at 1,500g for 10 min at 4°C. To prepare membrane-enriched fractions, the supernatants were ultracentrifuged at 100,000g for 30 min at 4°C, and the pellets were resuspended in dilution buffer (10 mM Tris-HCl, pH 7.4, 0.25 M sucrose, and 1 mM p-amindinophenylmethanesulfonylfluoride). To prepare the cytosols and nuclear extracts, the cells were washed with ice-cold PBS, resuspended in 2 ml of buffer A [10 mM KCl, 10 mM HEPES-KOH, pH 7.8, 0.1 mM EDTA, pH 8.0, 1 mM dithiothreitol, 0.5 mM phenylmethanesulfonylfluoride, and 2 µg/ml aprotinin] and kept on ice for 10 min. The cells were resuspended with 1.2 ml of buffer A and homogenized with approximately 80 strokes of Dounce homogenizer on ice. The supernatants were harvested as the cytosol extracts after centrifugation at 500g for 5 min at 4°C. The pellets were resuspended in 500 µl of buffer B (420 mM KCl, 50 mM HEPES-KOH, pH 7.8, 0.1 mM EDTA, pH 8.0, 5 mM MgCl₂, 20% glycerol, 1 mM dithiothreitol, 0.5 mM phenylmethanesulfonylfluoride, and 2 µg/ml aprotinin). The nuclear proteins were extracted at 4°C followed by centrifugation at 24,000g for 30 min. Protein concentrations were determined by use of the methods described by Bradford (1976), and each protein was kept at 80°C until use.

Western Blot Analysis and Deglycosylation Assay. The crude membranes, cytosols, and nuclear extracts were subjected to the analyses of P-gp, MRP1, MRP2, MRP4, BCRP, lung resistance-related protein (LRP), and DNA topoisomerase II (Topo II). Membrane vesicles from KB-C2, KB/MRP, LLCPK1-cMOAT, K6, and MCF-7 AdVp 3000 cells were used as the positive controls for P-gp, MRP1, MRP2, MRP4, and BCRP (Akiyama et al., 1985; Taguchi et al., 1997; Chen et al., 1999; Lee et al., 2000). The cytosolic fraction and nuclear extract obtained from SW620 cells after a 2-week treatment with 2 mM sodium butyrate were used as the positive controls for LRP and Topo II (Kitazono et al., 1999). The anti-P-gp monoclonal antibody (mAb) C-219, the anti-MRP1 mAb MRPm6, the anti-MRP2 mAb M₂I-4, and an anti-human Topo II rabbit antibody were purchased from Zymed Laboratories (South San Francisco, CA), Kamiya Biomedical (Thousand Oaks, CA), Monosan (Uden, Netherlands), and TopoGen Inc. (Columbus, OH), respectively (Kartner et al., 1985; Flens et al., 1994). An anti-MRP4 mAb was a gift from Dr. Kruh (Fox Chase Cancer Center, Philadelphia, PA). Anti-BCRP and anti-LRP antibodies were prepared according to the procedures described previously (Kitazono et al., 1999; Kage et al., 2002). The anti-BCRP antibody was generated by immunizing rabbits with a peptide that corresponds to amino acids 340 to 359 of the human BCRP protein.

Northern Blot Analysis. Total RNA was extracted from MT-4 and MT-4/DOX₅₀₀ cells with an RNA extraction kit (RNAzol B; Tel-Test, Friendswood, TX). Total RNA was also extracted from MCF-7 AdVp 3000 cells and used as the positive control. A SacII/HincII-digested fragment of BCRP cDNA (approximately, 1200 base pairs) was used as a hybridization probe. Each fragment was labeled with a random primer labeling kit (Stratagene, La Jolla, CA) and [³²P]dCTP (ICN Biomedicals, Costa Mesa, CA). Free [³²P]dCTP was removed with MicroSpin S-300HR column (Amersham). RNA (1.5 µg) was electrophoresed on a formamide gel and transferred to a membrane (Hybond-N⁺; Amersham) in 20× sodium chloride-sodium citrate buffer overnight. After crosslinking with UV light, the membrane was prehybridized in hybridization buffer (5× sodium chloride-sodium phosphate-EDTA, 5× Denhardt's solution, 0.5% SDS, and 20 µg/ml denatured salmon sperm DNA) for 2 h at 65°C, and hybridized with each probe in hybridization buffer at 65°C overnight. After washing thoroughly, the membrane was exposed to X-ray film for 1 or 3 days.

Anti-HIV-1 Assay. The activity of the compounds against HIV-1 replication was determined from the inhibition of virus-induced cytopathicity in MT-4 and MT-4/DOX₅₀₀ cells, as described previously (Baba et al., 1991). Briefly, the cells (1 × 10⁵ cells/ml) were infected with HIV-1 at a multiplicity of infection of 0.02 and cultured in the presence of various concentrations of the test compounds. After a 4-day incubation at 37°C, the number of viable cells was determined by the MTT method.

Efflux Analysis of AZT and Its Metabolites. Intracellular AZT and its metabolites in MT-4 and MT-4/DOX₅₀₀ were analyzed by a slight modification of the high-performance liquid chromatography (HPLC), as described previously (Perno et al., 1989). Five million cells were incubated with 1 µM [methyl-³H]AZT. After a 3-h incubation, the cells were washed three times with ice-cold medium and immediately frozen in dry ice. The cells were then extracted with 60% (v/v) methanol, and the methanol extracts were further heated at 95°C for 1.5 min. The extracts were clarified by centrifugation at 12,000g for 6 min. Separation and detection of AZT and its metabolites were performed by use of a 25-cm Whatman Partisil-10 SAX column (Gilson Medical Electronics, Middleton, WI) by HPLC. After injection of the samples (25 µl), the buffer gradient was applied, starting at 0 time with 5 mM potassium phosphate and increasing linearly to 750 mM potassium phosphate over 55 min at a rate of 1 ml/min. Then, 750 mM potassium phosphate was further pumped for 10 min. The elution was fractionated at 1-min intervals (1 ml) and analyzed for radioactivity.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Cytotoxicity of Compounds in DOX-Resistant Cells. The DOX-resistant T-cell line MT-4/DOX₅₀₀ was established from MT-4 cells by exposure to increasing concentrations of DOX (up to 500 ng/ml) for 1 year. When several anticancer agents were examined for their cytotoxic effects on MT-4/DOX₅₀₀ cells and compared with those on MT-4 cells, MT-4/DOX₅₀₀ cells were found to be highly resistant to DOX and mitoxantrone, moderately resistant to 7-ethyl-10-[4-(1-piperidino)-1-piperidino]carbonyloxycamptothecin, 7-ethyl-10-hydroxycamptothecin, and etoposide, and slightly resistant to actinomycin D, whereas the cells remained sensitive to paclitaxel, vincristine and cisplatin (Table 1). When the anti-HIV-1 agents AZT and NFV were evaluated for their IC₅₀ values in MT-4/DOX₅₀₀ cells, AZT was 2.5-fold less cytotoxic to MT-4/DOX₅₀₀ cells than to MT-4 cells (Table 1). However, such reduction in cytotoxicity was not observed with NFV.

Anti-HIV-1 Activity of Compounds in DOX-Resistant Cells. Because MT-4/DOX₅₀₀ cells showed some resistance to the cytotoxicity of AZT, anti-HIV-1 assays were conducted to determine whether the activity of AZT and other anti-HIV-1 agents was also reduced. AZT proved to be 7.5-fold less inhibitory to HIV-1 replication in MT-4/DOX₅₀₀ cells than in MT-4 cells (Table 2). The 50% effective concentrations (EC₅₀) of AZT were 0.013 and 0.094 µM in MT-4 and MT-4/DOX₅₀₀ cells, respectively. Furthermore, the anti-HIV-1 activity of 3TC was severely (more than 77-fold) impaired in MT-4/DOX₅₀₀ cells. In contrast, the activity of neither NNRTIs (nevirapine and emivirine) nor PIs (indinavir and NFV) was affected in the cells (Table 2). The anti-HIV-1 activity of AZT and 3TC was also determined by the inhibition of p24 antigen production in the culture supernatants. These compounds were also less inhibitory to p24 antigen production in MT-4/DOX₅₀₀ cells than in MT-4 cells (data not shown).

Intracellular DOX Concentration in DOX-Resistant Cells. To gain insight into the mechanism of DOX resistance in MT-4/DOX₅₀₀ cells, its intracellular accumulation and retention were examined. Both MT-4/DOX₁₀₀ and MT-4/DOX₅₀₀ cells showed reduced accumulation and retention of DOX compared with MT-4 cells (Fig. 1A). The intracellular steady-state concentrations in MT-4/DOX₁₀₀ and MT-4/DOX₅₀₀ were approximately 50% and 40% of that in MT-4 cells, respectively, suggesting increased influx or decreased efflux of DOX in the resistant cells. However, when ATP was depleted, the accumulation and retention of DOX in MT-4/DOX₅₀₀ cells was found to be similar to those observed in MT-4 cells (data not shown). These results suggest that the resistance of the cells was caused by increased influx or decreased efflux of DOX and that the transport was ATP-dependent. The intracellular retention of rhodamine 123 in MT-4/DOX₅₀₀ cells was also reduced to less than 10% of that of MT-4 cells. Depletion of ATP significantly increased the retention of rhodamine 123 in MT-4/DOX₅₀₀ cells to a level comparable with that in MT-4 cells, although ATP depletion also slightly increased the retention of rhodamine 123 in MT-4 cells (Fig. 1B). Therefore, it was assumed that the efflux of DOX was markedly enhanced in the DOX-resistant MT-4 cells and that the expression of an ATP-dependent transporter might be involved in the resistance.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Kinetics of intracellular DOX and effect of ATP depletion. A, MT-4 (), MT-4/DOX100 (), and MT-4/DOX500 () cells were exposed 10 µg/ml of DOX for up to 180 min (accumulation phase), washed with ice-cold PBS, and resuspended in warm culture medium in the absence of the compound (retention phase). At certain intervals, the cells were measured for their intracellular DOX concentrations by flow cytometry. The amount of intracellular DOX was expressed as the mean channel fluorescence (MCF). All data points are the mean values of three separate experiments. B, MT-4 () and MT-4/DOX500 () cells were incubated in glucose-free medium, as described under Materials and Methods. Rhodamine 123 (100 ng/ml) was added and further incubated for 30 min. After washing with glucose-free medium, the cells were incubated under ATP-depleting conditions for an additional 30 min on ice, and rhodamine retention was determined by flow cytometry.

ABC Transporter Expression in DOX-Resistant Cells. To elucidate which ATP-dependent transporter is involved in the resistance, Western blot analyses with the anti-P-gp mAb were conducted for MT-4 and MT-4/DOX₅₀₀ cells. However, no enhancement of P-gp expression was observed in MT-4/DOX₅₀₀ cells (Fig. 2). Therefore, we further examined the expression of MRP1, MRP2, MRP4, BCRP, and LRP. As shown in Fig. 2, BCRP was highly expressed in MT-4/DOX₁₀₀ and MT-4/DOX₅₀₀ cells. Although BCRP was also expressed in MT-4 cells, the expression level was not comparable with those in MT-4/DOX₅₀₀ and MT-4/DOX₁₀₀ cells (data not shown). In contrast, no expression of MRP1, MRP2, MRP4, and LRP was detected in the DOX-resistant or in the DOX-sensitive MT-4 cells (Fig. 2). The expression of BCRP in MT-4/DOX₁₀₀ cells was lower than that in MT-4/DOX₅₀₀ cells. In accordance with the level of BCRP expression, MT-4/DOX₅₀₀ cells were more resistant to DOX than were MT-4/DOX₁₀₀ cells (Fig. 3A). A similar result was also obtained with AZT. Although the magnitude of AZT resistance in the MT-4/DOX₅₀₀ and MT-4/DOX₁₀₀ cells was not comparable with that of their DOX-resistance, MT-4/DOX₅₀₀ cells seemed to be more resistant to AZT than did MT-4/DOX₁₀₀ cells (Fig. 3B).

View larger version (53K): [in this window] [in a new window]   Fig. 2.   ABC transporter expression in DOX-resistant cells. Crude membranes (for P-gp, MRP1, MRP2, MRP4, and BCRP), cytosols (for LRP), or nuclear proteins (for Topo II) were separated by SDS-polyacrylamide gel electrophoresis and transferred to a polyvinyldenedifluoride membrane. The transferred proteins were immunoblotted with a mAb against each transporter protein. Membrane vesicles from KB-C2, KB/MRP, LLCPK1-cMOAT, K6, and MCF-7 AdVp 3000 cells were used as the positive controls for P-gp, MRP1, MRP2, MRP4, and BCRP, respectively. Cytosolic fractions and nuclear extracts from sodium butyrate-treated SW620 cells were used as the control protein for LRP and Topo II. In the analysis for P-gp, lanes 1, 2, and 3 are proteins obtained from MT-4, MT-4/DOX500, and the control cells, respectively. In the analyses for MRP1, MRP2, MRP4, and BCRP, LRP, and Topo II, lanes 1, 2, 3, and 4 indicate those obtained from MT-4, MT-4/DOX100, MT-4/DOX500, and the control cells, respectively.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Cytotoxicity of DOX and AZT in DOX-resistant cells. MT-4 (), MT-4/DOX100 () and MT-4/DOX500 () cells were cultured in the presence of various concentrations of DOX (A) or AZT (B). After a 4-day incubation, the number of viable cells was assessed by the MTT methods. Representative results for three independent experiments are shown. The viable cell numbers of MT-4, MT-4/DOX100, and MT-4/DOX500 in the presence of 200 µM AZT were 3.4 ± 1.9, 35.1 ± 20.3, and 54.1 ± 4.5% of those in the absence of the compound, respectively.

View larger version (50K): [in this window] [in a new window]   Fig. 4.   Expression of BCRP mRNA and glycosylated BCRP in DOX-resistant cells. A, total RNA was extracted from MT-4 (lane 1), MT-4/DOX500 (lane 2), and MCF-7 AdVp 3000 (lane 3) cells, electrophoresed, and transferred to a membrane. The membrane was hybridized with a [32P]dCTP-labeled SacII-HincII-digested fragment of BCRP cDNA. B, crude membranes of MT-4/DOX500 and MCF-7 AdVp 3000 cells were incubated in the absence or presence of PNGase F (500 units) for 60 min at 37°C. The samples were analyzed by immunoblotting with an anti-BCRP antibody.

Effect of the BCRP-Specific Inhibitor Fumitremorgin C. To confirm the role of BCRP in AZT resistance, the effects of fumitremorgin C on the cytotoxicity of DOX and AZT were examined in MT-4 and MT-4/DOX₅₀₀ cells. As shown in Table 3, fumitremorgin C completely abolished the resistance of MT-4/DOX₅₀₀ cells to DOX and AZT at a concentration of 5 µM. However, fumitremorgin C alone did not have any inhibitory effect on the viability and proliferation of both MT-4 and MT-4/DOX₅₀₀ cells at this concentration (data not shown). Furthermore, the inhibitor also abolished the resistance of MT-4/DOX₅₀₀ cells to mitoxantrone but not to paclitaxel (data not shown).

AZT and Its Metabolites in DOX-Resistant Cells. To determine whether the reduced anti-HIV-1 activity of AZT in the DOX-resistant cells can be attributed to the increased efflux of the compound or its metabolites from MT-4/DOX₅₀₀ by BCRP, the intracellular metabolism of AZT was investigated in MT-4 and MT-4/DOX₅₀₀ cells. An HPLC analysis showed that the AZT 5'-monophosphate (AZTMP) accumulation was significantly diminished in MT-4/DOX₅₀₀ cells (Fig. 5). Compared with AZTMP level in MT-4 cells, only 7.0% of AZTMP was retained in MT-4/DOX₅₀₀ cells. In addition, the levels of AZT, AZT 5'-diphosphate, and AZT 5'-triphosphate were also reduced to 45.7%, 30.4%, and 50.3% of those in MT-4 cells, respectively (Fig. 5). These results suggest that the impaired anti-HIV-1 activity of AZT in MT-4/DOX₅₀₀ cells was caused by the increased efflux of its metabolites, presumably AZTMP, by BCRP.

View larger version (13K): [in this window] [in a new window]   Fig. 5.   AZT and its metabolites in DOX-resistant cells. MT-4 (A) and MT-4/DOX500 (B) cells were incubated with [methyl-3H]AZT. The methanol extracts were separated by HPLC, and each fraction was analyzed for radioactivity. The intracellular concentrations of AZT and its metabolites were expressed as disintegrations per minute (DPM).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Significant advances in the treatment of HIV-1 infection have been achieved with the success of HAART, which is conducted through a combination of drugs that block different steps in the viral replication cycle, such as reverse transcription and protein processing. Unfortunately, the emergence of drug resistance frequently occurs because of rapid mutation of viral genome (Schinazi et al., 2000). In addition, it is assumed that host cellular factors are also involved in the resistance to antiretroviral drugs (Groschel et al., 1997; Swanstrom and Erona, 2000). One factor that may limit the therapeutic efficacy of PIs is the ABC transporters P-gp and MRP1. These transporters are known to interact with several PIs (Washington et al., 1998; Srinivas et al., 1998). Furthermore, if the resistance to NRTIs could be induced by host cellular factors, it would be a more serious impediment to the progress of HAART (Lavie et al., 1997). Overexpression of MRP4 has been reported to efflux some NRTIs from CD4⁺ T cells and result in their decreased anti-HIV-1 activity in vitro (Schuetz et al., 1999). The present study clearly demonstrates that the novel ABC transporter BCRP also affects the anti-HIV-1 activity of NRTIs, particularly AZT and 3TC, in cell cultures.

After prolonged treatment of MT-4 cells with DOX, the established cell line MT-4/DOX₅₀₀ displayed multidrug resistance to different classes of anticancer agents (Table 1). More interestingly, MT-4/DOX₅₀₀ cells showed some resistance to NRTIs but not to NNRTIs and PIs in terms of anti-HIV-1 activity and cytotoxicity (Tables 1 and 2). Although the anti-HIV-1 activity of 3TC was severely impaired in MT-4/DOX₅₀₀ cells, it was not possible to determine the IC₅₀ of 3TC in MT-4/DOX₅₀₀ as well as in MT-4 cells because of its low cytotoxicity (data not shown). ATP-dependent retention of rhodamine 123 in MT-4/DOX₅₀₀ cells indicated that an ABC transporter might play an important role in the resistance (Fig. 1B). It has been documented that the ABC transporter P-gp recognizes AZT as a substrate and reduces its antiviral activity against HIV-1 replication (Yusa et al., 1990; Antonelli et al., 1992). However, the P-gp-specific inhibitor verapamil could not block the efflux of DOX or rhodamine 123 in MT-4/DOX₅₀₀ cells (data not shown). Furthermore, the expression of P-gp proved to be lower than a detectable level in MT-4/DOX₅₀₀ and MT-4 cells (Fig. 2), and MT-4/DOX₅₀₀ cells did not acquire resistance to the P-gp substrates vincristine and paclitaxel (Table 1), suggesting that an ABC transporter other than P-gp is involved in the resistance to NRTIs in MT-4/DOX₅₀₀ cells. Extensive analyses revealed that, among the ABC transporters examined, BCRP was the only ABC protein that was significantly expressed in the DOX-resistant MT-4 cells (Fig. 2B). In addition, the BCRP-specific inhibitor fumitremorgin C completely abolished the resistance of MT-4/DOX₅₀₀ cells to AZT as well as to DOX (Table 3), suggesting that BCRP is the molecule responsible for the AZT resistance.

It was reported that the arginine at position 482 (R482) of BCRP was an important determinant for substrate specificity. Increased efflux of doxorubicin and rhodamine 123 was observed with the R482T or R482G mutant but not with the wild type of BCRP (Honjo et al., 2001). Our preliminary analysis for the full-length BCRP cDNA identified the R482M mutation in MT-4/DOX₅₀₀ cells (data not shown). The doxorubicin selection in the presence of the P-gp inhibitor verapamil did up-regulate the BCRP expression in human breast carcinoma cells (Doyle et al., 1998). However, we showed here that the doxorubicin selection in the absence of verapamil could also up-regulate the expression of BCRP, albeit a mutant form, in a human CD4⁺ T-cell line. Although this mutation differed from those observed previously in doxorubicin-resistant human cell lines (R482T or R482G), it could not be excluded that the NRTI resistance of MT-4/DOX₅₀₀ cells was caused by the mutation but not by the increased expression of BCRP. Further studies are in progress to determine whether the wild-type BCRP affects the anti-HIV-1 and cytotoxicity to NRTIs.

Another point that should be considered is the activity of nucleotide kinases in the DOX-resistant cells. If AZT and 3TC would be less phosphorylated in the cells because of a reduced kinase activity, the intracellular concentration of the active form (5'-triphosphate) of AZT and 3TC should be decreased. In this case, the cells might seem to be resistant to these NRTIs. However, this is unlikely, because the resistance of the cells was not induced by prolonged exposure to any nucleoside derivatives but was cultured in the presence of DOX, which is not a substrate of nucleotide kinases. Furthermore, the MT-4/DOX₅₀₀ cells showed little, if any, resistance to another pyrimidine analog stavudine (Table 2 and data not shown).

BCRP has been detected in breast, colon, and gastric cancers and in acute myeloid and lymphoblastic leukemias (Sauerbrey et al., 2002). Furthermore, BCRP is also expressed in some healthy tissues, including placenta, liver, breast, and venous and capillary endothelia (Maliepaard et al., 2001). More importantly, BCRP mRNA was detected in bone marrow and peripheral blood mononuclear cells (Sauerbrey et al., 2002). HIV-1 has been found in several tissues in vivo and can infect many different types of human cells in vitro. It is possible that the interaction of NRTIs with BCRP could reduce the intracellular drug concentrations in these tissues, resulting in insufficient suppression of HIV-1 replication.

The central nervous system disorders associated with HIV-1 infection, such as encephalopathy and dementia, occur at the late stage of the disease (Gendelman et al., 1994). Although the development of HAART has decreased the incidence rates for HIV-1-associated encephalopathy and dementia, its impact on the future incidence and course of dementia remains debatable (Geraci and Simpson, 2001; Sacktor et al., 2001). Most of the licensed anti-HIV-1 drugs have a limited capacity to enter the brain (Brouwers et al., 1997). The expression of BCRP in the capillary endothelium of the blood-brain barrier may contribute to the limited capacity of NRTIs to enter into the brain (Maliepaard et al., 2001). BCRP modulators such as fumitremorgin C may be useful to improve the entry of NRTIs into the brain and increase their concentrations in the cerebrospinal fluid (Rabindran et al., 1998, 2000).

In conclusion, we found that the novel ABC transporter BCRP is a cellular factor involved in the resistance to anti-HIV-1 NRTIs, such as AZT and 3TC. Further studies are needed to determine whether BCRP expression in the target cells is indeed related to the treatment failure and emergence of drug resistance in HIV-1-infected patients.