Multidrug transporters as drug targets.

Transport molecules can significantly affect the pharmacodynamics and pharmacokinetics of drugs. An important transport molecule, the 170 kDa P-glycoprotein (Pgp), is constitutively expressed at several organ sites in the human body. Pgp is expressed at the blood-brain barrier, in the kidneys, liver, intestines and in certain T cells. Other transporters such as the multidrug resistance protein 1 (MRP1) and MRP2 also contribute to drug distribution in the human body, although to a lesser extent than Pgp. These three transporters, and especially Pgp, are often targets of drugs. Pgp can be an intentional or unintentional target. It is directly targeted when one wants to block its function by a modifier drug so that another drug, also a substrate of Pgp, can penetrate the cell membrane, which would otherwise be impermeable. Unintentional targeting occurs when several drugs are administered to a patient and as a consequence, the physiological function of Pgp is blocked at different organ sites. Like Pgp, MRP1 also has the capacity to mediate transport of many drugs and other compounds. MRP1 has a protective role in preventing accumulation of toxic compounds and drugs in epithelial tissue covering the choroid plexus/cerebrospinal fluid compartment, oral epithelium, sertoli cells, intesticular tubules and urinary collecting duct cells. MRP2 primarily transports weakly basic drugs and bilirubin from the liver to bile. Most compounds that efficiently block Pgp have only low affinity for MRP1 and MRP2. There are only a few effective and specific MRP inhibitors available. Drug targeting of these transporters may play a role in cancer chemotherapy and in the pharmacokinetics of substrate drugs.

Plasma membrane biophysical properties linked to the antiproliferative effect of interferon-alpha.

The relationship of plasma membrane biophysical properties to the anti-proliferative effect of interferon-alpha (IFN-alpha) was investigated in Daudi lymphoblasts cell lines with sensitivity to growth inhibition, parallel clonal variants selected for resistance, and one revertant subclone. Lateral mobility of surface differentiation antigens (I2, CD19, CD20, and sIgM-kappa) were measured by fluorescence recovery after photobleaching (FRAP). The mean diffusion coefficients, D, values for two clones of IFN-alpha resistant Daudi cells were significantly higher (D = 8.1-11 x 10(-10) cm2/sec) than for parental sensitive cells (D = 4.9-7.4 x 10(-10) cm2/sec). Microviscosity of the plasma membranes were probed by electron spin resonance (ESR) spectrometry. These results also indicate a greater degree of molecular motional freedom in resistant cells. Treatment of sensitive lymphoblasts with IFN-alpha (100-400 U/10(6) cells) for 5-30 min consistently increased mean values of D and the degree of spin-probe motional freedom, whereas no significant differences were detected in resistant cells. The effect of IFN-alpha on the membrane potential (Em) of Daudi cells was quantitated by flow cytometry using a voltage-sensitive oxonol dye. Membrane potential of all clones was similar (-50 to -56 mV). Treatment with IFN-alpha for 8-10 min caused hyperpolarization in the sensitive cells (deltaEm up to 45 mV), but only minimal hyperpolarization in the resistant ones (deltaEm up to 7 mV). We concluded that sensitivity to IFN-alpha and treatment with IFN-alpha are related to the biophysical status of plasma membranes.

Influence of beta-adrenergic antagonists, H1-receptor blockers, analgesics, diuretics, and quinolone antibiotics on the cellular accumulation of the anticancer drug, daunorubicin: P-glycoprotein modulation.

BACKGROUND: Treatment of patients with several drugs simultaneously may result in modulation of the naturally expressed P-glycoprotein (Pgp) at different tissues. With this possibility in mind, we have assessed the ability of different classes of drugs to modulate Pgp function in vitro. Modulation of the Pgp function was studied at in vitro drug concentrations comparable to therapeutic blood levels of the drugs. MATERIALS AND METHODS: Human blood brain barrier endothelial cells and human colon adenocarcinoma cells were transduced or transfected with the multidrug resistance gene (MDR1) to express Pgp. The uptake of fluorescent substrates of Pgp, Rhodamine 123 and daunorubicin, into these cells and NIH3T3/MDR1 and MDCK/MDR1 cells was measured by flow cytometry and in monolayers in the presence and absence of the different drugs. RESULTS: From the tested six H1-receptor blockers, seven beta-adrenergic antagonists, four analgesics, ten diuretics and five quinolons, five drugs inhibited Pgp at therapeutic blood levels and two at somewhat higher concentrations. Significant synergism for blocking Pgp could be demonstrated for several drugs. CONCLUSION: We conclude that administration of several drugs which modulate the function of Pgp to patients may adversely affect the natural function of this efflux pump and may cause drug-drug interactions induced side effects.

Induction of apoptosis in MDR1 expressing cells by daunorubicin with combinations of suboptimal concentrations of P-glycoprotein modulators.

The application of most agents with the capacity to reverse multidrug resistance (MDR) via modulation of the multidrug transporter P-glycoprotein (Pgp) was shown to be associated with toxic side-effects. For this reason, we have investigated the effect of combinations of suboptimal concentrations of Pgp blockers on the induction of apoptosis and growth arrest in daunorubicin (D) treated, MDR1 gene transfected cells. We used verapamil, PSC833 and Cremophor EL as Pgp modulators, which affect the function of Pgp by different mechanisms. Treatment of NIH3T3/MDR1 cells with combinations of suboptimal concentrations of Pgp modulators in the presence of D caused apoptosis and G(2) arrest to the same extent as optimal concentrations of singly used blockers. We conclude that combinations of suboptimal concentrations of Pgp modulators may cause effective sensitization of resistant tumor cells, and at the same time, may avoid the frequently observed toxic effects experienced in clinical trials with a single modifier applied at the optimal dose.

Influence of antipsychotic, antiemetic, and Ca(2+) channel blocker drugs on the cellular accumulation of the anticancer drug daunorubicin: P-glycoprotein modulation.

We investigated the effect of antiemetic, antipsychotic, and Ca(2+) blocker drugs on the function of P-glycoprotein (Pgp) in vitro and compared inhibitory concentrations with therapeutic blood levels. Human colon adenocarcinoma (Caco-2) and human blood-brain barrier endothelial cells were transfected or transduced to express Pgp, and the uptake of rhodamine123, calcein AM, or daunorubicin was measured by flow cytometry in the presence of the drugs. NIH3T3/MDR1 cells were used for reference testing. Results of the flow cytometric studies were supported by cell proliferation and monolayer permeability studies. Thirty-five drugs are included in this study, of which 13 modulate the function of Pgp at the therapeutic blood concentration and 8 at a concentration 2 to 4 times higher. Two drugs, which block the function of Pgp only partially at therapeutic blood concentrations, blocked the function of Pgp completely if used concomitantly. Based on these in vitro experiments, we conclude that administration of several drugs that modulate the function of Pgp simultaneously may adversely affect the natural function of this efflux pump and may cause drug-induced side effects in patients.

Anti-psychotic drugs reverse multidrug resistance of tumor cell lines and human AML cells ex-vivo.

Anti-psychotic drugs are used in cancer patients undergoing chemotherapy frequently and the concomitantly used drugs may alter the pharmacokinetics of each other. One reason for the alteration of pharmacokinetics may be the modulation of the function of P-glycoprotein, whose efflux pump occurs in resistant cancer cells, in human intestine and in the blood-brain barrier. For this reason we tested the effect of several anti-psychotic drugs on the multidrug-resistant pump, P-glycoprotein. We found that in the MDR gene transfected L121C MDR, L5178 MDR and in the KB-V-1 cells selected for resistance some antipsychotic drugs block the function of P-glycoprotein. Blood cells of two treatment-resistant leukemic patients also showed increased uptake of daunorubicin if treated ex vivo with the anti-psychotic drugs. Our results suggest that pharmacokinetic studies should be performed prior to concomitant clinical use of such drugs which block P-glycoprotein function.

Combinations of P-glycoprotein blockers, verapamil, PSC833, and cremophor act differently on the multidrug resistance associated protein (MRP) and on P-glycoprotein (Pgp).

Clinical studies are currently in progress to evaluate functional modifiers of P-glycoprotein (Pgp), an efflux pump associated with resistance to cancer chemotherapy. However, the effects of these modifiers on a more recently discovered efflux pump, the multidrug resistance associated protein (MRP), have not yet been fully characterized. MRP is expressed in most human tissues and is overexpressed in several tumor types. For these reasons, we have investigated the effects of three prototype Pgp modifiers, which act by different modes on the function of Pgp, on the function of MRP in two MRP-overexpressing cell lines: UMCC/VP lung and MCF-7/VP breast cancer cells. Clinically optimal plasma levels of verapamil, cremophor, and PSC833 have been shown to completely block the function of Pgp in Pgp-over expressing cells. However, in the two MRP-over expressing cell lines, these modifiers only partially blocked the function of MRP and combinations of these optimal concentrations acted antagonistically. Similar antagonistic effects were seen with combinations of suboptimal concentration levels of these blockers, while these combinations resulted in synergistic effects in Pgp overexpressing cells. For two biophysical parameters measured at the plasma membrane, membrane fluidity and membrane potential, the effects of these modifiers were essentially similar in Pgp and MRP expressing cells. We suggest that the 170 kD Pgp and the 190 kD MRP glycoproteins, imbedded in the plasma membranes, respond differently to simultaneous effects of the investigated prototype resistance modifiers. These results also suggest that the identification of the specific mechanism of drug resistance is important for the selection of chemotherapeutic strategies to block the efflux pump on the cancer cell.

Biochemical and clinical aspects of efflux pump related resistance to anti-cancer drugs.

This review paper will focus on the molecular biological, biochemical and biophysical aspects of the following three efflux pumps: P-glycoprotein (Pgp), Multidrug Resistance Protein (MRP) and Lung cancer Resistance-related Protein. Since suppression of the function of these pumps has clinical implications, novel approaches developed in our laboratory for blocking the function of Pgp and MRP are also discussed. The second part of this review will summarize the clinical significance and the results of clinical trials designed to suppress the effects of these efflux pumps.

Inhibition of HIV infection of H9 cells by chlorpromazine derivatives.

The binding between the HIV surface protein, gp120, and the CD4 coreceptor is known to be initiated by electrostatic interactions. Because of the ability of chlorpromazine to interact with proteins by charge transfer, we tested several derivatives for their ability to block binding of HIV to CD4+ cells. We have shown that 7,8-dioxo-chlorpromazine blocks binding of fluorescein isothiocyanate-labeled anti-Leu3a and rgp120 to peripheral human blood T4 cells and blocks syncytia formation between gp120- and CD4-expressing cells. We also found that 7,8-dioxo-chlorpromazine blocks HIV infectivity of H9 cells and acts synergistically with zidovudine.

Inhibition of the transport function of membrane proteins by some substituted phenothiazines in E. coli and multidrug resistant tumor cells.

Efflux-pumps mediated by P-glycoprotein increase the level of resistance to antibiotics in bacteria and to cytostatics in tumor cells due to decreased drug accumulation, and are also involved in the operation of blood brain barrier. Different compounds are able to enhance drug retention in the cells by inhibiting the efflux-pump mechanism of multidrug resistant (mdr) cancer cells and bacteria. The effects of substituted chlorpromazines were studied on a hemolysin producing and antibiotic resistant plasmid carrying E coli, and rhodamine uptake of multidrug resistant (mdr 1 gene expressing) mouse lymphoma cells. Hemolysin transporter protein encoding plasmids were eliminated from E. coli by a representative phenothiazine namely promethazine. Minimal inhibitory concentrations of tetracyclin and promethazine were lower for plasmidless bacteria as compared to the parent, plasmid carrying strains. The antibiotic resistance plasmid was cured of the R-plasmid of E. coli JE 2571, however, the ring substituted derivatives were less effective then parent compounds. The effect of some substituted phenothiazines on P-glycoprotein efflux-pump of mouse lymphoma cells were studied. The majority of ring substituted derivatives reversed the mdr of tumor cells. The 3,7,8-trihydroxy- and 7,8-dihydroxy derivatives of chlorpromazine were effective as P-glycoprotein blockers, however, 7,8-diacetoxy-, 7,8dimetoxy-, 7-semicarbazone-, and 5-oxo-chlorpromazine derivatives had only moderate effect. A tomato lectin, specific for blood brain capillary endothelium was able to modify the activity of P-glycoprotein in tumor cells. Phenothiazine and tomato lectin had some antagonism in tumor cells. Our results suggest that the inhibition of P-glycoprotein function in murine tumor cells and inhibition of transporter protein in E. coli bacteria may depend on pi-electron superdelocalizibility and electrophile binding of the compounds to the transporter proteins. The intracellular accumulation of antibiotics or chemotherapeutics increased as a consequence of decreased drug efflux in both bacterial and tumor cell systems. The inhibition of the drug effux-pump is the same for all individual cells of the population. These results can be realized by combination chemotherapy, however, antiplasmid effect itself cannot be exploited in this respect because the resistance was reversed in a part of the population only. The similarity with mdr P-glycoprotein in tumor cells and brain capillary endothels provides a good model for molecules opening the blood brain barrier.

Intracellular pH does not affect drug extrusion by P-glycoprotein.

The intracellular pH (pH(i)) of cells exhibiting multidrug resistance (MDR) related to the expression of the P-glycoprotein (Pgp) is often more alkaline than that of the parental cells, as also observed for the KB-V1/KB-3-1 system in this paper. The possible role of an elevated pH(i) in Pgp-related MDR has been investigated by shifting back the pH(i) of the MDR+ cells to a more acidic value using the mobile proton ionophore carbonylcyanide m-chlorophenylhydrazone (CCCP). The influence of CCCP-evoked delta pH(i) on relative daunorubicin (DNR) accumulation was similar in the case of several Pgp positive and negative cell lines, in view of flow cytometric and radioactive drug accumulation studies and measuring DNR levels in the medium in a flow-through system. Our data argue against a significant effect of pH(i) on Pgp pumping efficiency. However, an indirect connection between pH(i) regulation and the MDR phenotype is suggested by the fact that acidification of the external medium in the presence of verpamil could be observed exclusively in MDR+ cells.

MDR1/P-glycoprotein function. I. Effect of hypotonicity and inhibitors on rhodamine 123 exclusion.

The MDR1 protein (P-glycoprotein) is a membrane ATPase whose expression results in resistance to several anti-tumor drugs. It has been proposed that the MDR1 protein, in addition to its pumplike properties, can function as (Gill et al. Cell 71: 23-32, 1992; Altenberg et al. Cancer Res. 54:618-622, 1994) or mediate the activity of (Hardy et al. EMBO J. 14: 68-75, 1995) a hypotonic stress-induced Cl- current. In addition, one study found that drug transport and Cl- channel-associated functions of MRD1 were separable and mutually exclusive and that, when cells were swelled, the MDR1 protein could not transport substrate. This hypothesis was tested in four pairs of isogenic cell lines with MDR1 transfectants expression 8,000-55,000 MDR1 antibody binding sites per cell. Cytoplasmic exclusion of rhodamine 123 was used as an indicator of MDR1 function to measure the effect of hypotonic stress, MDR1 inhibitors, and Cl- channel blockers on MRD1 transport function. It was found that MDR1 activity and its inhibition by cyclosporine A or flufenamic acid were unaffected by hypotonicity alone or in combination with Cl- channel blockers.

MDR1/P-glycoprotein function. II. Effect of hypotonicity and inhibitors on Cl- efflux and volume regulation.

Resistance to anti-tumor drugs can be mediated by overexpression of the multidrug resistance 1 (MDR1) protein (P-glycoprotein). In three MDR1-transfected cell lines (Gill et al. Cell 71: 23-32, 1992; Altenberg et al. Cancer Res. 54: 618-622, 1994), a hypotonic stress-induced Cl- current has been demonstrated that can be inhibited by MDR1 substrates and Cl- channel blockers. We tested the hypothesis that MDR1 expression confers additional Cl- conductance by measuring regulatory volume decrease (RVD) in four pairs of isogenic cell lines and 36Cl efflux in two cell lines with and without hypotonic stress. The kinetics of RVD and response to Cl- channel blockers were indistinguishable in MDR and parental cells. Additionally, no significant difference was seen between 36Cl efflux rate constants under hypotonic conditions between NIH/3T3 and L1210 parental and MDR cells. We conclude that, in intact cells, the expression of MDR1 does not alter the rate of volume regulation or the rate 36Cl efflux under hypotonic conditions between parental and MDR cells.

Effect of combination of suboptimal concentrations of P-glycoprotein blockers on the proliferation of MDR1 gene expressing cells.

Pharmacologically active in vivo doses of P-glycoprotein (Pgp) blockers, specifically verapamil, Cremophor EL and PSC833 cause toxicity in addition to that from the concomitantly used cancer chemotherapeutic drugs. It was shown before that these blockers cause different types of toxicities in vivo. We found that these 3 chemically distinct Pgp blockers exert different biophysical effects on the membranes of L1210 MDR cells. They also affect the general metabolism of these cells differently, but all block affinity labeling of Pgp. We could also show that the combination of suboptimal doses of these blockers can restore the uptake of the Pgp substrate rhodamine 123 into L1210MDR, 3T3MDR and KB-VI cells and can reduce the survival rate of these cells when treated in combination with daunorubicin. Our results suggest that the combination of suboptimal doses of these Pgp blockers may be advantageous in clinical practice.

Behavior of N-acylated daunorubicins in MDR1 gene transfected and parental cells.

The substrate specificity of the P-glycoprotein (P-170), a multidrug transporter, was studied using N-acylated daunorubicin derivatives and four MDR1 cDNA transfected cell lines. Results showed that N-acetyl-daunorubicin is a substrate, but the longer fatty acid derivatives, N-octanoyl and N-dodecanoyl daunorubicins, are not. This conclusion was reached by flow cytometric drug uptake assay, cell proliferation assays, and confocal microscopy. It was concluded that the longer fatty acid derivatives interact with plasma membranes in a way that affected P-glycoprotein function.

The effect of ion channel blockers, immunosuppressive agents, and other drugs on the activity of the multi-drug transporter.

The MDRI protein is an energy-dependent transport protein responsible for the multi-drug resistance seen in many tumors. A variety of drugs have been shown to inhibit the function of this pump, including compounds known to block various ion channels. The mouse lymphoma cell line L5178Y has been transduced with the human mdrI gene. Using this cell line, we have tested a number of compounds to determine whether there is a correlation between the ability to block a specific type of ion channel, or shift membrane potential, and the ability to act as an MDR-reversing agent using the fluorescent substrates Rhodamine 123 and daunorubicin as test compounds. Our results show no apparent correlation between the ability to block a specific ion channel and reversal of MDR transport ability. We have found active MDR inhibitors in compounds that affect K+, Na+, Ca++, H+, but not Cl- channels. Our data suggest that Cl- channel activity may be distinct from MDR activity. Several immunosuppressive compounds and analogs were also tested and found to be active reversing agents. Measurements suggest a significant difference in resting membrane potential between the L5178YvMDR line and the L5178Y parental cell line used in these experiments. No correlation was found between the ability of drugs to alter membrane potential and to inhibit MDR transport activity. Our results suggest that MDR transport function may be independent of the physiological movement of ions and show that a wide variety of compounds can inhibit MDR transport.

Modulation of the anti-proliferative signal of interferon-alpha by tamoxifen in U937 cells.

Several clinical protocols are attempting to utilize the combined anti-proliferative signal of interferon-alpha (INF-alpha) and tamoxifen on cancer cells. We demonstrated here that the effect of these two agents on the growth of the premacrophage U937 cells is antagonistic. This antagonistic effect is paralleled by the ability of tamoxifen to modulate the INF-alpha-induced hyperpolarization in these cells. INF-alpha-induced hyperpolarization was shown before to be an integral part of the anti-proliferative signal of this agent. Tamoxifen liberates Ca2+ from intracellular stores of U937 cells but this effect of the drug is not the cause of its antagonistic effect with the anti-proliferative signal of IFN-alpha. We suggest therefore, that the combined effect of these two anti-cancer drugs could also be advantageous for macrophage proliferation.

Cytoskeletal modulation of plasma membrane events induced by interferon-alpha.

Cytochalasin B, a drug that alters microfilament structure, was found to modulate interferon-alpha (IFN-alpha)-induced changes in ion fluxes, in motional freedom of spin probes, and lateral diffusion of surface antigens. These changes occur in Daudi cells inherently sensitive to the antiproliferative signal of IFN-alpha, but not in insensitive cells, and were associated with the antiproliferative signal previously. The biophysical effects of cytochalasin B were detected by flow cytometric quantitation of membrane potential using an oxonol dye, by electron spin resonance (ESR) spectrometry, and by measurements of fluorescence recovery after photobleaching (FRAP) of surface antigens using a laser-interactive cell imaging system. Cytochalasin B treatment increased an IFN-alpha-induced membrane potential shift by -5 mV. The motional freedom of 5-doxyl-stearic acid changed from 0.67 to 0.63, as expressed by the order parameter, S, with IFN-alpha treatment and was prevented by cytochalasin B. Changes in the lateral diffusion of surface antigens induced by IFN-alpha treatment, D = 5.3 x 10(-10) without treatment and D = 7.8 x 10(-10) cm2/s with treatment, was blocked by cytochalasin B. In contrast, the microtubule stabilizers taxol and D2O and the microtubule depolymerizing colcemid were ineffective at dose levels sufficient to cause the characteristic cell physiological alterations of these agents. These results implicate microfilaments but not the microtubule system in transduction of the antiproliferative signal by IFN-alpha in Daudi cells.

Epitope mapping by photobleaching fluorescence resonance energy transfer measurements using a laser scanning microscope system.

The donor photobleaching method (T. M. Jovin and D. J. Arndt-Jovin. 1989. Annu. Rev. Biophys. Biophys. Chem. 18:271-308.) has been adapted to an ACAS 570 (laser scanning microscope) system to measure fluorescence resonance energy transfer (FRET) on individual human peripheral blood T cells. Photobleaching was completed in approximately 100 ms in our case and it followed double-exponential kinetics. The energy transfer efficiency (E) was approximately 20% between the CD4 epitopes OKT4-FITC and Leu-3a-PE as well as between OKT4E-FITC and OKT4-PE. E was approximately 8% between OKT4-FITC and Leu-4-PE (alpha CD3) and barely detectable (approximately 4%) from OKT4-FITC to Leu-5b-PE (alpha CD2). The E values obtained by the photobleaching method were highly reproducible both in repeated measurement of identical samples and in experiments with different batches of cells and were in agreement with the flow cytometric donor quenching measurements. As expected, E measured between primary and secondary layers of antibodies increased (from approximately 14% to approximately 28%) when F(ab')2 fragments were substituted for whole antibody molecules as the donor. On a T cell line we mapped the distance between the idiotypic determinant of the T cell receptor (TcR) and the Leu-4 epitope of CD3 as proximal as E = 28%, as compared to E = 4% between a framework TcR epitope and Leu-4. In the latter case, however, approximately 40% less Leu-4 was bound suggesting that the antigen binding site of TcR is in close proximity with one of the two CD3 epsilon chains, which hence are not equivalent.