(Received for publication, May 16, 1997, and in revised form, June 17, 1997)
From the Laboratory of Cell Biology and the
¶ Laboratory of Molecular Biology, Division of Basic Sciences,
NCI, National Institutes of Health, Bethesda, Maryland 20892-4255 and the
Department of Biochemistry, Silberman Institute of
Life Sciences, Hebrew University, Jerusalem, 91904, Israel
Considerable uncertainty surrounds the
stoichiometry of coupling of ATP hydrolysis to drug pumping by
P-glycoprotein, the multidrug transporter. To estimate relative
turnovers for pumping of the drug vinblastine and ATP hydrolysis, we
began by measuring the number of P-glycoprotein molecules on the
surface of murine NIH3T3 cells expressing the human MDR1
gene. Fluorescence of cells treated with monoclonal antibody UIC2 was
determined as a function of (i) amount of antibody at a fixed number of
cells and (ii) increasing cell number at constant antibody. The two
together gives 1.95 × 106 P-glycoprotein
molecules/cell. Initial uptake rates of vinblastine ± verapamil
measure the ability of P-glycoprotein to extract vinblastine from the
plasma membrane before it enters the cell. As a function of
[vinblastine] at 37 °C, they give the maximum rate of this component of outward pumping as 2.1 × 106 molecules
s1 cell
1 or a turnover number of 1.1 s
1. Initial rates of one-way efflux as a function of
[vinblastine] at 25 °C ± glucose give the maximum rate of
this component of pumping as 0.59 × 106 molecules
s
1 cell
1. The ratio of ATPase activity of
P-glycoprotein at 37 and 25 °C is 4.6. Appropriating this ratio for
pumping, maximum one-way efflux at 37 °C is 4.6 × 0.59 = 2.7 × 106 molecules s
1
cell
1, a turnover number of 1.4 s
1. The
vinblastine-stimulated ATPase activity of P-glycoprotein has a turnover
number of 3.5 s
1 at 37 °C, giving 2.8 molecules of ATP
hydrolyzed for every vinblastine molecule transported in a particular
direction. These calculations involve several approximations, but
turnover numbers for pumping of vinblastine and for
vinblastine-stimulated ATP hydrolysis are comparable. Thus, ATP
hydrolysis is probably directly linked to drug transport by
P-glycoprotein.
Most people who die from cancer do so because their tumors have metastasized and become resistant to chemotherapy. Until we find ways to prevent cancer entirely, overcoming drug resistance is the main hope to save lives. P-glycoprotein (P-gp),1 the product of the MDR1 gene, contributes to multidrug resistance in many cell types (1, 2) and is expressed in many tumors (3-8). P-gp pumps out its drug substrate from the tumor cell, reducing the effectiveness of administered chemotherapeutic agents (9). It has ATPase (10-14) activity enhanced by numerous substrates and substrate analogs. A very wide range of substrates are pumped out of cells by P-gp (15). Major efforts have been made toward finding clinically useful reversers of P-gp that can block its action, leading to renewed accumulation of drugs within erstwhile resistant cells (16-17). Secure knowledge of the mechanism of action of P-gp is the basis for designing new and more effective reversers.
Four lines of evidence indicate that P-gp can expel its substrates directly out of the cell membrane (reviewed in Ref. 18). First, the substrates of P-gp are lipophilic and reside, most of the time, within cell membranes. Thus, it is within the membrane that the pump will find it most easy to locate its substrate. Second, kinetic analyses show that drug accumulation is reduced by the action of P-gp from the earliest times that measurements can be made, i.e. before significant amounts of the drug can enter the cell yet when it is already present within the membrane (19). Hence it seems to be pumped out from the membrane itself before it crosses it. Third, Raviv et al. (20), making use of the ability of the photosensitive probe 5-iodonaphthalene-1-azide to label membrane proteins, showed that doxorubicin was expelled from cell membranes of P-gp-containing but not drug-sensitive cells. In addition, fluorescent dyes such as Hoechst 33342 (21) and calcein-AM (22) have been used to demonstrate removal of substrate from the lipid bilayer by P-gp. These considerations favor a "vacuum cleaner" model (20) for P-gp in which this protein is an ATP-driven pump that pumps its substrates directly out of the plasma membrane. A major criticism of this model is, however, the lack of reliable information on the stoichiometry of ATP hydrolysis to drug pumping. Attempts to measure this stoichiometry in phospholipid vesicles containing pure P-gp suggested a minimum of 50 ATP molecules hydrolyzed per drug molecule transported, however, such a figure is difficult to reconcile with the metabolic potential of multidrug resistant cells (12). Recently, Eytan et al. (23) attempted to derive the stoichiometry by measuring the effect of P-gp on valinomycin-facilitated transport of 86Rb+ into proteoliposomes containing P-gp, valinomycin being a P-gp substrate. They reported 0.5-0.8 molecules of the complex valinomycin-Rb+ transported for every ATP molecule hydrolyzed. They assumed, however, that only Rb+-complexed valinomycin molecules were transported by P-gp, whereas we observed2 that valinomycin's stimulation of ATP hydrolysis by P-gp is unaffected by its charged state. Transport of charged valinomycin molecules in the experiments by Eytan et al. (23) was carried out in the presence of a substantial excess of the uncharged form, which would not contribute to the measured flux but would contribute to ATP hydrolysis. In addition, their assay was carried out in solutions that were not free of potassium, which would compete with rubidium, complicating their calculations of the stoichiometry. We deemed it important, therefore, to determine this stoichiometry using an independent approach.
Here we determine the number of P-gp molecules present on the surface of P-gp-expressing cells using an antibody (Ab) titration procedure and FACS analysis. In the same cell type, we measure the maximum capacity of P-gp to reduce the entry of its substrate vinblastine (VBL) by determining the difference between initial rates of VBL uptake in the presence and absence of the P-gp blocker verapamil. These two numbers, taken together, can determine the catalytic constant (molecules pumped per second) for the effect of P-gp on reducing VBL accumulation. Also, we measure the maximum capacity of such cells to accelerate efflux of loaded VBL in the presence of an energy source, giving another measure of the transport capacity and of the catalytic constant for the effect of P-gp on VBL transport. Comparing these catalytic constants with that obtained for the effect of VBL on accelerating ATP hydrolysis by P-gp, we compute the ratio of the maximum rate of VBL transport to the maximum rate of VBL-stimulated ATP hydrolysis. This ratio is not far from unity. Thus, only a small number of ATP molecules seem hydrolyzed for each VBL molecule transported by P-gp.
We used the following previously characterized murine NIH3T3 fibroblasts: sensitive cells (NIH3T3 cell line), wild type human P-gp-expressing MDR1-Gly-185, and mutant MDR1-Val-185 (in which glycine at position 185 is substituted by valine) as described in Stein et al. (19). Human MDR1-transfected cell lines N3-30, N3-600, and N3-2400 (resistant to vincristine at 30, 600, and 2400 ng/ml, respectively) were characterized by Germann et al. (24). Cells were grown as monolayer cultures as described earlier (19, 24).
FACS Analysis0.125 to 8 million cells (from a 10-cm dish harvested by trypsinization into PBS, washed twice with PBS containing 0.1% bovine serum albumin (PBS-BSA), and resuspended in PBS-BSA) were incubated for 30 min on ice with 0.125-6.4 µg of human P-gp-specific UIC2 monoclonal Ab (Immunotech, Westbrook, ME) (25) or with 5 µg of IgG2a isotype control Ab (PharMingen), washed twice with PBS-BSA, and reacted with fluorescein-conjugated (fluorescein isothiocyanate-labeled) goat anti-mouse IgG2a Ab (1-10 µg; Jackson ImmunoResearch Laboratories, West Grove, PA) for 30 min on ice. The cells were again washed twice in PBS-BSA and the levels of FL1 fluorescence were analyzed using a FACSort flow cytometer with Cell Quest software (Becton Dickinson FACS System, San Jose, CA).
Transport ExperimentsCells were seeded onto 5-cm dishes at 400,000 cells/2 ml of complete medium (in the absence of cytotoxic drugs) and used 2 days later at 106 cells/dish after washing in ice-cold PBS. Uptake was measured from 5 s to 120 min at 37 °C from PBS containing 5 mM glucose and 15-25 Ci/mmol 3H-VBL (Amersham) at concentrations from 1.4 nM to 80 µM and either 0 or 50 µM verapamil (to block P-gp), and stopped by washing rapidly, twice, in ice-cold PBS. Zero time uptakes were determined by adding ice-cold solutions of the 3H-VBL to the cells. Cells were removed from dishes by trypsinization, and dish contents were transferred to vials for liquid scintillation counting. Cells from dishes (PBS-washed as for the transport experiments) were trypsinized for counting in the hemocytometer. For efflux, cells were first loaded with VBL by suspending in appropriate concentrations of 3H-VBL for 40 min at 25 °C from PBS containing 10 mM sodium azide and 10 mM 2-deoxyglucose to deprive the cells of metabolic energy (and block the action of P-gp). Efflux was initiated after washing cells rapidly in PBS followed by resuspension in PBS containing either sodium azide and 2-deoxyglucose as before or 5 mM glucose. Efflux was terminated by washing the dish rapidly with ice-cold PBS, and cells were taken for scintillation counting as before. Data are reported in terms of the volume of external solution taken up by, or still remaining in, 1 million cells at the reported time (19). In all cases aliquots of the 3H-VBL solutions used were taken for scintillation counting to calculate this quantity.
ATP Hydrolysis AssaysATP hydrolysis by crude cell membranes was determined by measuring the release of inorganic phosphate from 5 mM ATP in the presence of 10 mM MgCl2, 5 mM sodium azide, 1 mM ouabain, and 0.5 mM EGTA and in the presence and absence of 0.3 mM sodium orthovanadate (26, 35).
Data AnalysisEquations 1-3 were used to obtain the best fit curves by using the program SIGMA PLOT (Jandel, Inc.), which uses the Marquard-Levenburg algorithm.
Using FACS, we first studied binding of P-gp-specific monoclonal
Ab UIC2 to cells of various cell lines engineered to express human
P-gp. Fig. 1 depicts the results of
experiments where we used control mouse IgG2a Ab (A) and
UIC2 Ab (B) binding to NIH3T3 cells transfected with the
human MDR1 gene (Gly-185 cells) at four concentrations of
Ab, 0, 0.2, 1.6, and 4.8 µg/tube containing 200 µl of medium. In
panel C, at these same concentrations with binding being to
Val-185 cells, NIH3T3 cells transfected with a mutant strain of
MDR1 where valine replaces glycine in position 185 of the
polypeptide chain of P-gp; in panel D, binding was done with
a single amount (1 µg) of antibody to five different cell strains,
using the non-transfected NIH3T3 cells, cells transfected with wild
type MDR1 gene and selected at low dose (30 ng/ml
vincristine; N3-30), with a higher dose (600 ng/ml; N3-600), and with
a still higher dose (2400 ng/ml; N3-2400), as well as the
MDR1 Gly-185 cell strain grown in the presence of 60 ng/ml
colchicine (see panel B). In each case, the data are plotted
as the number of cells on the ordinate that are labeled with the
fluorescence intensity denoted on the abscissa. Cells tested with the
control IgG Ab showed very little fluorescence with no increase as the
concentration of Ab is raised, whereas the Gly-185 and Val-185 strains,
tested with monoclonal Ab UIC2, demonstrate a clear and similar
fluorescence signal that increases in intensity as the concentration of
Ab increases. The various cell strains depicted in panel D
show a marked difference in signal intensity, consistent with
differences in the amount of P-gp present on the surface of these
different cells (24).
To determine the number of available P-gp molecules on the cell surface, we used the same monoclonal Ab UIC2 and a depletion assay. We first quantitate how the degree of reaction depends on the amount of Ab. Fig. 2 (panels A and C) depicts the median fluorescence of a sample of Gly-185 and Val-185 3T3 cells, respectively, when treated with increasing Ab concentration (0-8 µg of Ab/sample mixture of 200 µl) using a fixed number of cells, 5 × 105/reaction at 4 °C. We fitted the data by the simple binding equation.
![]() |
(Eq. 1) |
![]() |
(Eq. 2) |
For the Gly-185 cells, we found Km to be 1.87 ± 0.24 µg of Ab, while Kn was 2.97 ± 0.77 million cells/reaction sample. We calculate the number of P-gp molecules exposed to Ab as follows. In Fig. 2B, 1 µg of Ab is present per reaction sample, and Kn is 2.97. Therefore 2.97 million cells are sufficient to deplete one-half or 0.5 µg of Ab from the reaction mixture. Now 0.5 µg of Ab is 2.02 × 1012 molecules (assuming molecular mass of the Ab to be 150 kDa). Hence each cell binds 2.02/2.97 or 0.68 million molecules of Ab. But from Fig. 2A we find the Km of Ab for P-gp as 1.87 µg. Thus at the 1 µg concentration used in Fig. 2B, Ab-binding sites are only saturated to a fraction of 1/(1 + 1.87) or 0.345. Therefore, the true number of Ab-binding sites is 0.68/0.345 = 1.95 ± 0.53 million P-gp sites/Gly-185 3T3 cell on computing for the combination of errors. Performing the Ab binding procedures at 25 and 37 °C gave essentially the same values (data not shown).
For the Val-185 3T3 cells, similarly, the number of P-gp sites per cell is 2.63 ± 0.87 million, not significantly different from the number present per Gly-185 3T3 cell. We performed the same experiments also for the N3-2400 strain of human MDR1-transfected 3T3 cells (see Ref. 24 and Fig. 1D). For these cells (data not shown), the value of Km was found to be 2.9 ± 0.4 µg of Ab while Kn was 0.56 ± 0.14 × 106 cells. Calculating, as described above, gives 14.2 ± 3.8 × 106 molecules of P-gp/cell for this highly resistant strain.
In addition, with the MDR1 Gly-185-transfected 3T3 cells, we
used another human P-gp-specific monoclonal Ab with an external epitope, MRK-16 (27), and also the Fab and F(ab)2
fragments prepared from this Ab. The depletion titration gave a
Kn value of 1.90 ± 0.37 × 106 cells for the MRK-16 Ab and 2.76 ± 0.62 × 106 for the F(ab
)2 fragment, not significantly
different. Similar results were obtained with Fab (monovalent) fragment
(data not shown). Considering these data it is reasonable to assume
that 1 molecule of Ab binds to 1 P-gp molecule.
P-gp can act on its substrates in two ways; it can pump drug from the cytoplasm of the cell and also pump it out from the membrane before it reaches the cytoplasm (19). We determined the maximum rate of VBL pumping by P-gp by these two pathways. First, we studied the extraction of VBL from the membrane before it accumulated inside the cell. Fig. 3 shows the time course of uptake of VBL into Gly-185 3T3 cells at 37 °C in PBS. (In Figs. 3 and 4, the ordinate is the amount of VBL taken up and expressed as the volume of external medium cleared of VBL (in µl) per million cells at the time stated (19). Multiplying this measure by the concentration of VBL in the external medium would give the amount of VBL that enters the cell in the given time period). In Fig. 3A, the time course was measured over an extended range. The final level of uptake is dramatically increased by adding 50 µM of the reverser verapamil to the cells (filled circles). In Fig. 3B, however, uptake is measured for 12 s at 37 °C. Uptake is linear from zero time during this interval. In succeeding experiments, uptakes were performed at 10 s in the initial rate range. Fig. 4A depicts the uptake of VBL during 10 s in the presence (filled circles) and absence (open squares) of 50 µM verapamil. In the presence of verapamil there is little change in the VBL uptake as the concentration of the drug is increased. In its absence, uptake increases with the concentration of VBL. Two curves, with and without verapamil, begin to approach one another, which is consistent with increasing concentrations of VBL saturating the pumping ability of P-gp so that it cannot cope with the inflow of the drug. Thus, the difference between uptake of VBL in presence and absence of verapamil gives the component of pumping that takes place as the drug crosses the membrane but before it enters the cytoplasm (18, 19). In Fig. 4B we plot this difference, i.e. pumping of VBL by P-gp, as a function of [VBL] by combining data from different experiments. The filled circles are from Fig. 4A, open circles are from a similar experiment at other concentrations of VBL, whereas the open square is from the 10-s data points of Fig. 3B. The ordinate here, the volume of external medium cleared at the time chosen, is proportional to the velocity of uptake (v) divided by the concentration of VBL in the external medium (S). Using Michaelis-Menten kinetics to describe the pumping action of P-gp leads us to
![]() |
(Eq. 3) |
We next studied zero trans-efflux (28) from cells loaded
with VBL. In Fig. 5A, Gly-185
3T3 cells were loaded with 1.4 nM VBL at 25 °C for 40 min in the presence of 10 mM sodium azide and 10 mM 2-deoxyglucose to deplete the energy (19). (These experiments were performed at 25 rather than 37 °C since we found that during loading of the cells at 37 °C, in the absence of an energy source, they became less adherent to the dishes, preventing an
accurate determination of efflux rates.) Cells were washed rapidly with
ice-cold PBS and then exposed to a wash-out medium at 25 °C, free of
VBL but containing either PBS + 10 mM sodium azide + 10 mM 2-deoxyglucose (filled circles) or PBS + 5 mM glucose to restore energy (filled squares).
The data were fitted by an exponential efflux equation, giving values
for t1/2 of 37 and 4.4 min into
azide/2-deoxyglucose or glucose medium, respectively. Efflux into
glucose medium is thus some 8× that into azide/2-deoxyglucose. A
parallel experiment performed with [VBL] at 100 µM is
depicted in Fig. 5B. Rates into azide/2-deoxyglucose and
glucose are now similar to t1/2 values of 13 and
8.4 min, respectively. In addition, results of similar experiments with
different concentrations of VBL are depicted in Fig. 5C.
Here we plot efflux during 2 min, reported as the difference between
efflux into glucose and into azide/2-deoxyglucose and expressed as a
fraction of the zero time value, as a function of [VBL]. Data from
Fig. 5, A and B are shown as open
squares with circles being from another experiment and triangles from a third. As for Fig. 4B, we fitted
these data by Equation 3, obtaining a value for the maximum velocity of
pumping (fractional loss) as 0.179 ± 0.011/min. The volume of
external medium cleared by 106 cells at zero time of efflux
was 21.5 ± 1.5 µl (n = 5). The
Kp from Fig. 5C was 15.1 ± 3.0 µM. Thus, at a concentration of VBL in the external
medium equal to Kp, there would be 21.5 × 15.1 or 325 pmol of vinblastine present within the cell. This is pumped out
at one-half the maximum rate but represents one-half of the pumping
capacity of the cell's P-gp. Thus the maximum rate of VBL efflux is
0.179 × 325 pmol/106 cells/min or 0.59 ± 0.12 × 106 molecules/cell/s at 25 °C.
To compare these results with the turnover number for the ATPase, we
measured the ATPase activity both at 25 and 37 °C. We determined
this activity for the Val-185 3T3 cell line and the N3-2400 cell line
that is enriched for P-gp (24). We assessed the effect of the drugs
such as vinblastine, verapamil, and colchicine on stimulation of the
ATPase activity (Table I). The
VBL-stimulated ATPase activity is reduced 4.63-fold as the temperature
is decreased from 37 to 25 °C. Were this ratio to be appropriate for
the pumping function of P-gp on VBL, one would conclude that a maximum
efflux of 0.59 at 25 °C might be equivalent to a value of 2.73 ± 0.55 × 106 molecules/cell/s at 37 °C or to a
turnover number of 1.40 ± 0.43 s1, which is
indistinguishable from the value determined for the inward component of
pumping (Fig. 4B).
|
Our data provide an estimate of the number of P-gp molecules present at the surface of the cell by using an antibody depletion method. The number we find for both Gly-185 MDR1- and Val-185 MDR1-transfected cell lines appears reasonable in light of the previous determinations. Thus, Pastan et al. (29), using MRK-16 Ab and ferritin bridge labeling, determined that MDR1 virus-infected colchicine-resistant MDCK cells express about 5.5 × 106 P-gp molecules/cell. Sehested et al. (30), using freeze-fracture electron microscopy found 500 P-gp molecules/µm2 of plasma membrane, corresponding to 8 × 105 molecules/cell. Shapiro and Ling (12) estimated that multidrug-resistant Chinese hamster ovary cells contain 3 to 4 × 106 P-gp molecules/cell based on the yield of purified P-gp from these cells. Although, we compare different cell lines with different degrees of drug resistance, our estimates of 2 × 106 molecules of P-gp for the Gly-185 3T3 cells, 2.6 × 106 for the Val-185 cells, and 14 × 106 for the highly resistant N3-2400 cells are compatible with previous data. The Ab depletion method has the advantage of giving an in situ determination of the number of Ab-binding molecules per cell and requires no assumptions as to the yield of protein extracted from the plasma membrane or the yield of membranes that can be prepared from the whole cells.
Our data on the maximum velocity of drug efflux can be compared with two values in the literature. Ghauharali et al. (31) studied daunorubicin efflux from multidrug resistant human KB-8-5 epidermoid carcinoma cells using a fluorescence assay and reported a maximum efflux of 180 pmol of daunorubicin/106 cells/min at 37 °C to be compared with 58 pmol/min/106 cells at 25 °C in our study or perhaps 4.63× this value (269 pmol/106 cells/min) if we take the temperature dependence of the ATPase activity as appropriate for pumping. A far higher value of 2,000 pmol/min/106 cells was calculated by Stein (18) from data reported by Nielsen et al. (32) for daunorubicin efflux from the multidrug resistant Ehrlich ascites tumor cells. The reason for this discrepancy is not clear, but these experiments were not designed to determine the maximum value for daunorubicin pumping. The higher value for daunorubicin efflux (2 nmol/min/106 cells) gives rise to a higher drug/ATP stoichiometry in the calculations of Stein (18).
We can compare the turnover numbers that we obtain on combining our
transport data and the estimates of the number of P-gp molecules per
cell with the turnover numbers that are available for the purified
protein (12), these being in the range 1-10/s. Urbatsch et
al. (13), using purified P-gp from drug resistant Chinese hamster
ovary cells report a turnover number of 4.9 s1 for the
basal ATPase activity and 9.2 s
1 for the activity in the
presence of verapamil. The verapamil-stimulated activity would thus
have a turnover number of 4.3 s
1. Our
data2 (see also Table I)
using purified and reconstituted P-gp (isolated from KB-V1 cells) gives
turnover numbers of the basal activity of 2.9 s
1, 12.8 s
1 of the maximal verapamil-stimulated activity and 6.4 s
1 of the maximal VBL-stimulated activity. This would
give values for the verapamil component alone of 9.9 s
1
and for the VBL component of 3.5 s
1. It is difficult to
know which of these turnover numbers are appropriate to compare with
the turnover numbers for VBL pumping (1.1-1.4 s
1) that
we have determined.
In addition, our data for VBL efflux necessarily had to be determined at 25 °C whereas the values for the turnover number of the ATPase activity have been determined at 37 °C. We did, however, also measure ATPase activities at 25 °C for all the P-gp species that we studied for three different drugs that stimulate the activity of P-gp (Table I). In particular, the ATPase activity in the presence of verapamil and VBL was reduced 3.5- and 4.6-fold, respectively, as temperature was reduced from 37 to 25 °C, comparable with the reduction of 3.6-fold found in the two determinations of the VBL pumping rates in Gly-185 cells.
Our turnover numbers for VBL pumping may be underestimated. In
particular, we measured pumping rates in the living cell at prevailing
ATP concentrations. The Km for ATP for P-gp ATPase
is reported as being in the range 0.28-1.4 mM (10, 12)). With intracellular ATP concentrations being some 2 mM, P-gp
may be operating at only one-half of its maximum activity, especially in our experiments on VBL efflux, where the drug is loaded into cells
in the absence of an energy source, glucose is added at zero time, and
VBL efflux is measured from that time. Some time may be required for
ATP concentration to reach a fully saturating level so that our
turnover number of 1.4 s1 (corrected for temperature)
could well underestimate the true value.
There is no reason, however, to suggest that a large number of ATP molecules are required to expel a single molecule of VBL. The stoichiometry between VBL pumping and ATP hydrolysis seems to be of order of magnitude unity, but direct estimations of the ATP consumption by P-gp, concomitant with drug transport, will be needed to establish this. Such direct measurements have been made for a bacterial periplasmic binding protein-dependent transport system (also a member of the superfamily of ABC transporters), where a stoichiometry of almost 2 ATP molecules hydrolyzed per substrate molecule transported has been established (33).
Finally, we address the question of whether the pumping activity of P-gp, computed from the turnover number for drug pumping taken together with the number of P-gp molecules per cell, can account for the high levels of drug resistance that are found. Our data show that these numbers are perfectly compatible with the resistance parameters. We argue above that the turnover numbers that we find for drug pumping are compatible with those reported for the drug-stimulated component of the ATPase activity of P-gp. We now argue that the pump rates that we measure are compatible with the level of drug resistance. Consider first the rate of efflux pumping of VBL (Fig. 5A). The efflux rate in the presence of an active P-gp system is 8× the rate in its absence. Second, consider the numbers for the effect of P-gp on reducing the inward movement of VBL (Fig. 4A). Here P-gp reduces 3-fold the inward flow of VBL. Taken together, these two effects should bring about a reduction of 3 × 8 = 24-fold in the internal [VBL] compared with the level in the absence of an effective P-gp pumping system. The VBL resistance of Gly-185 MDR1-(wild type) transfected 3T3 cell line is 46-fold times the resistance of the recipient 3T3 drug-sensitive strain (34). Thus, the presence of P-gp acting as a drug pump with the kinetic properties revealed in this study is sufficient to account for by far the major part of the P-gp-induced drug resistance that can be measured. Interestingly, the N3-2400 cell line has a drug resistance against VBL that is 4 or 5× greater than that of the Gly-185 and Val-185 cell lines. Our analysis suggests that the number of available P-gp molecules on the surface of the N3-2400 cells is 5-7× as high as for the Gly-185 and Val-185 lines, commensurate with the increase in drug resistance.
We are grateful to Drs. Michael M. Gottesman and Ira Pastan for helpful discussions, advice, and encouragement throughout the course of this work.