Competitive Interaction of Cyclosporins with the Vinca Alkaloid-binding Site of P-glycoprotein in Multidrug-resistant Cells*

The mechanism of reversal of resistance to Vinca alkaloids by cyclosporins is unclear. We investigated the molecular mechanism of reversal of Vinca alkaloid resistance by cyclosporin A (CsA) and its nonimmuno- suppressive analog 0-acetyl CsA 33-243) in multidrug resistant DC-3F/VCRd-5L cells. CsA at 3 PM increased sensitivity and almost totally reversed VCR resistance. analysis [‘H]VCR uptake P-glycoprotein

The mechanism of reversal of resistance to Vinca alkaloids by cyclosporins is unclear. We investigated the molecular mechanism of reversal of Vinca alkaloid resistance by cyclosporin A (CsA) and its nonimmunosuppressive analog 0-acetyl Cel CsA (   Tumor cells selected in uitro for resistance to natural product anticancer drugs such as Vinca alkaloids or anthracyclines show cross-resistance to other structurally unrelated compounds (1). This multidrug resistance (MDR)' is frequently characterized by overexpression of a 150-180 kDa transmembrane glycoprotein P-glycoprotein (P-gp) (2, 3) and concomitant decreased drug accumulation in resistant cells as compared with their drug-sensitive parental cell lines (3). P-gp is the mediator of MDR (4) and is believed to function as an energy-dependent efflux pump that transports drugs out of the cell, resulting in decreased drug accumulation within cells leading to decreased drug efficacy (3,5). Drugs involved in the MDR phenotype bind specifically to P-gp (6-10). Analysis of cDNA encoding P-gp demonstrates that this protein has two nucleotide-binding sites homologous to ATP-binding sites of bacterial transport proteins (11-13). Several reports have documented that certain lipophilic agents are able to reverse MDR in vitro (3, 13). These agents may modulate MDR by interacting with chemotherapeutic drug-binding site(s) of P-gp (8,(14)(15)(16)(17)(18)(19)(20)(21)(22). Recent work by Slater et al. (23) and Twentyman (24) showed that the immunomodulating fungal undecapeptide cyclosporin A (CsA) can reverse vincristine (VCR) and doxorubicin resistance in MDR cells. Due to the relatively low toxicity of CsA at clinically achievable concentrations (25), this agent may be useful for overcoming drug resistance in cancer patients. Since the molecular mechanisms of reversing MDR by cyclosporins are not clear and since there is increasing evidence that many lipophilic agents may interact with P-gp to reverse MDR, we examined CsA and its nonimmunosuppressive analog 0-acetyl Cgl CsA (SDZ 33-243) (24) to determine whether interaction of these agents with P-gp is the mechanism of reversing Vinca alkaloid resistance. Our data indicate that cyclosporins may reverse Vinca alkaloid resistance by competitively interacting with the Vinca alkaloid-binding site of P-gp, and demonstrate correlations between the ability of cyclosporins to inhibit binding of Vinca alkaloids to P-gp, increase cellular retention of these drugs, and reverse Vinca alkaloid resistance. in the absence or presence of 1 and 3 PM cyclosporins for 7 days. The effects of 1, 3, 5, and 10 MM cyclosporins alone on the survival of the sensitive and resistant cells were also evaluated. Colonies were fixed with 4% formaldehyde, stained with crystal violet, and colonies containing more than 50 cells were scored. Concentrations of VCR which reduced the number of colonies by 50% after 7 days treatment (I&,) in colonogenic assays were determined from triplicate experiments using 0.0001-100 PM VCR. Drug Accumulation-[3H]VCR and [3H]VBL accumulation was determined using cell monolayers grown to 70% confluence in Corning 24-well plates in growth medium containing 50 nM L3HlVBL or [3H]VCR with or without 2.5, 5, or 10 j.aM cycloiporins in a-5% CO, atmosphere. Cells in each well were incubated for 1 h at 37 "C. DN~ acc&ulation was stopped by rapidly washing the cells twice with icecold phosphate-buffered saline. The cells were then trypsinized, and the cell-associated radioactivity was counted in a Beckman LS-7500 scintillation counter.

PHI VCR Uptake by Membrane
Vesicles-Membrane vesicles were prepared by the nitrogen cavitation technique described by Horio et al. (22). The protein concentration of the membranes was determined by the method of Bradford (28) using the Bio-Rad protein assay kit. Uptake of [3H]VCR by membrane vesicles was measured by a rapid filtration method. Five ~1 of membrane vesicles containing 25 fig of protein were incubated with different concentrations of [3H]VCR (0.03-55 pM) in 10 mM Tris-HCl, pH 7.4,250 mM sucrose, 3 mM ATP or ATP-yS, and 5 mM MgC12 (final volume of 50 ~1) at room temperature. After 60 min, the membrane vesicles were collected on Millipore membrane filters (HAWP, 0.45~pm pore size; Bedford, MA) which were treated with 3% bovine serum albumin overnight. The filters were then washed two times with 3 ml of ice-cold Tris-HCl buffer containing 250 mM sucrose, 3 mM ATP, and 0.1% bovine serum albumin. The membrane filters were dissolved in scintillation fluid, and the associated radioactivity was determined by liquid scintillation counter. Specific uptake of ["H]VCR by the membrane vesicles was determined by subtracting nonspecific uptake measured in the presence of 100 or 500 $M unlabeled VCR from the total uptake. Kinetic parameters for VCR uptake were estimated from non-linear least squares analysis by using the following equation where u and s are specific uptake and VCR concentration in the reaction solution, respectively. V,, and .&,,I are maximum uptake and half-maximum uptake VCR concentrations for the high affinity component, respectively. Similarly, V,,,, and Km2 represent those values for the low affinity component, respectively. The apparent inhibitory constant (K,) of CsA was estimated from 13H]VCR uptake in the absence and presence of CsA by replacing K,, with K,,,](l + i/K), where i is the concentration of CsA in the reaction solution.

RESULTS
Cytotoricity-The cytotoxicity of VCR on DC-3F-sensitive Chinese hamster lung cells and VCR variant DC-3F/VCRd-5L cells was evaluated by colonogenic assay. The VCR concentration that inhibited cell growth by 50% (I&J was 0.026 PM for DC-3F and 62 PM for DC-3F/VCRd-5L cells, respectively (  [3H]VBL or ["HIVCR accumulation at the concentration of 50 nM was measured in culture medium maintained in 5% CO, and 37 "C for 60 min. Concentrations of CsA or SDZ 33-243 were 2.5, 5, and 10 WM. The results of DC-3F (hatched bar) and DC-3F/VCRd-5L (closed bar) cells are shown as mean f SE. of triplicate experiments.
excess VCR, specific [3H]VCR uptake was blocked, and the uptake of radioactivity to membrane vesicles increased linearly with a slope parallel to the terminal nonspecific linear portion of the biphasic curve. The specific [3H]VCR uptake was obtained by subtracting the nonspecific linear curve from the mixed biphasic profile. With this correction for nonspecific uptake, total specific uptake of The procedure for the uptake by membrane vesicles was the same as described in the legend for Fig. 2. Open or closed circles represent [3H]VCR uptake in the presence or absence of CsA (0.2 FM). The units of abscissa and ordinate are pmol/mg protein and pmol/mg protein/PM, respectively. Kinetic parameters were calculated by nonlinear least squares analysis as described in the text and used to construct the curves.  Fig. 3 also shows the effect of CsA on [ 'H]VCR uptake in membrane vesicles. CsA at 0.2 pM increased the apparent high affinity uptake K,,,, value while it did not affect the V,,,.,,,. These results demonstrate that CsA is a competitive inhibitor of VCR uptake at a high affinity site with an apparent K, of 0.126 + 0.0173 PM.
Inhibition of VBL Photoaffinity Labeling by Cyclosporins-In order to further explore the interaction of cyclosporins with Vinca alkaloid-binding site(s) of the outward drug transporter, we performed photoaffinity labeling experiments with the VBL photoactive analog ["'IINASV in the absence or presence of increasing concentrations of either unlabeled VBL or cyclosporins (Fig. 5). We have previously shown that this photoactive analog specifically binds to P-gp (6). When photoaffinity labeling was carried out with 0. 0.4 pM, respectively (Fig. 5). This is in line with our Vinca alkaloid accumulation experiments and kinetic analysis of ["H]VCR uptake in the absence or presence of cyclosporins.

DISCUSSION
In this study we examined the mechanism of reversal of VCR resistance by CsA and its nonimmunosuppressive analog, , in VCR/MDR variant DC-3F/VCRd-5L cells in uitro. As revealed by colonogenic assay, CsA was a highly effective modulator of VCR resistance and SDZ 33-243 was even more active. Furthermore, while the effects of cyclosporins are much greater in VCR-resistant cells, some degree of sensitization in the parental DC-3F cells was also evident. CsA alone partially inhibited the growth of both sensitive and resistant cells at 5 FM and SDZ 33-243 totally inhibited growth in both cell lines at 3 FM. These results are in agreement with previous studies by Saydjari et al. (29) which described growth inhibition by CsA in a hamster pancreatic carcinoma cell line, and by Twentyman (24), who demonstrated growth inhibition of a MDR subline of the human small cell lung cancer cell line NCI-H69 by O-acetyl C,' CsA. In this study, VCR resistance modulation occurred at dose levels which are not in themselves growth inhibitory. Therefore, it appears that there is no direct relationship between the growth inhibition by a given cyclosporin and its ability to reverse MDR. Moreover, our data and others (24,(30)(31)(32) demonstrate that the molecular mechanism of reversing MDR by cyclosporins is separate from the immunosuppressive properties of these agents.
Our drug uptake data demonstrate that VCR (Fig. lA), VBL (Fig. 1B)  Recently, both ATP-dependent and -independent VCR binding with apparent K,,, values of 0.24 and 9.7 pM, respectively, in membrane vesicles of the human MDR K562/ADM leukemia cells was also reported (21). The kinetic parameters obtained in the presence of ATP in the present study are close to these reported K,,, values, suggesting that the high and low affinity uptake found in this study might correspond to ATP-dependent and -independent binding to P-gp, respectively, in the process of translocating VCR across the plasma membrane. The high affinity binding site might be an activated state of P-gp. Furthermore, our kinetic analysis of ["HIVCR uptake by membrane vesicles of the resistant cells in the presence of 0.2 pM CsA revealed that this agent interacts with the drug :' I. Tamai and A. Safa, unpublished observation.
transporter at the high affinity site in the MDR cells and competitively inhibits VCR uptake with an apparent K, of 0.126 + 0.0173 PM. These results suggest that the high affinity uptake component might have an important role in reversing Vinca alkaloid resistance of MDR cells by cyclosporins.
We have previously reported that photoactive analogs of VBL bound to P-gp in MDR cells (6,9) and that many compounds which reverse MDR also bind to P-gp (8, 15-19) and inhibit binding of VBL to this protein (8, l&19). Further support for the interaction of cyclosporins with the Vincu alkaloid-binding site of P-gp was provided by photoaffinity labeling experiments with the photoaffinity analog of VBL, [12"I]NASV. When membrane vesicles from the VCR/MDR variant DC-3F/VCRd-5L cell line were photolabeled with 50 nM [1251]NASV in the presence of increasing concentrations of unlabeled VBL, CsA, or photolabeling of Pgp was inhibited with half-maximum concentrations of 0.6, 0.5, and 0.4 pM, respectively. These results suggest that cyclosporins interact directly with VBL on P-gp. Interestingly, the cyclosporins used have slightly higher potency than VBL in inhibiting [""I]NASV binding to P-gp. These results are also in agreement with the data in Fig. 4 in which 50% inhibition of VCR uptake in membrane vesicles occurred at 0.5 PM CsA. During the course of this study (32), Foxwell et al. (33) reported that a photoactive analog of CsA, [3H]cyclosporin diazirine, binds to P-gp, and its binding was competable by excess CsA and by the nonimmunosuppressive cyclosporin H. However these authors did not examine either the effect of Vinca alkaloids on CsA binding to P-gp, nor the modulating activity of cyclosporins on MDR cells. Our data provide detailed kinetic analysis of the inhibition of Vinca alkaloid interaction with P-gp by cyclosporins and combines several properties of cyclosporin-mediated reversal of P-gp-mediated MDR.
Nevertheless, the mechanisms underlying VCR sensitization by cyclosporins in sensitive cells and cellular growth inhibition by cyclosporins alone remain to be found. Since CsA and SDZ 33-243 also increased the sensitivity of the parental DC-3F cells to VCR and since these cells have a low level of P-gp (6), a proposed interaction with P-gp may indeed be their mechanism of VCR sensitization in sensitive cells. It has been proposed that cyclophillin, a specific CsA-binding protein, may mediate the immunosuppressive effect of CsA (34), and evaluation of a wide range of CsA analogs led to the conclusion that an intact Cs amino acid at position 1 of CsA structure is required for immunosuppression (35). The lack of cyclophillin binding to SDZ 33-243, which has modified Cs amino acid (36), rules out the role of cyclophillin in VCR sensitization in sensitive cells. In conclusion, our data indicate that CsA and its nonimmunosuppressive analog SDZ 33-243 interact with P-gp, competitively block binding of Vincu alkaloids to P-gp, and inhibit outward transport of these drugs. The weak correlation between the magnitude of the effect of cyclosporins on VCR accumulation and cell survival suggests that in addition to interaction with P-gp, these drugs may modulate VCR resistance by other mechanisms.
A prospective study of MDR reversal with a greater number of cyclosporins will enable definite conclusions regarding the structure-activity relationship for resistance modulation by these compounds and for the increased drug accumulation in sensitive and MDR cells by cyclosporins.
Acknowledgments-We are grateful to Dr. June L. Biedler for generously supplying the DC-3F and DC-3F/VCRd-5L cell lines. We would like to express our appreciation to Dr. Monica Morrow for helpful discussions and Drs. Richard L. Shilsky and Gini Fleming for critical review of this manuscript.
We also would like to thank Soheila