Isolation of antigenic mimics of MDR1-P-glycoprotein by phage-displayed peptide libraries

To identify an MC57 epitope which is more efficiently expressed on inactivated forms of P-glycoprotein we utilized peptide libraries displayed on filamentous phage. Using this technology, we selected specific phage clones blocking the binding of the murine monoclonal (MAb) MC57 with live human multi-drug-resistant (MDR) cells, and sequenced and analyzed their DNA. The results we obtained indicate that MAb MC57 epitope could be formed by 2 regions localized on the predicted fourth and sixth extracellular loops of the current 12-transmembrane-domain model predicted for MDR1-P-glycoprotein. Surprisingly, a third region, defined by residues 800-807 of the MDR1-P-glycoprotein sequence and postulated to be intracellular, was also identified as a putative part of the MC57 epitope. This finding adds weight to the interesting hypothesis that a P-glycoprotein structure different from the current model may exist.

Identification of a sister gene to P-glycoprotein

The P-glycoproteins (Pgps) are a small family of transport proteins associated with the multidrug resistance phenotype of cell lines selected for growth in cytotoxic drugs. Utilizing low stringency screening, we have identified a novel gene closely related to the Pgps expressed in the pig and other mammalian liver which we have called Sister of P-glycoprotein (spgp). Sequence of this gene shows it to be a member of the ATP-binding cassette family of transporters and the gene most closely related to Pgp identified to date. The function of spgp is not known, but it can be recognized by at least one Pgp mAb, C219. This cross-reactivity has implications for expression studies in tissues and tumors utilizing this and other Pgp antibodies.

Inhibition of cell-mediated cytolysis and P-glycoprotein function in natural killer cells by verapamil isomers and cyclosporine A analogs

We have previously shown that among normal leukocytes, CD56+ and CD8+ cells express relatively high levels of P-glycoprotein (P-gp), a transmembrane efflux pump. While the physiologic significance of P-gp expression in leukocytes is unknown, the relatively high levels of P-gp in CD56+ and CD8+ cells suggest that P-gp may function in cell-mediated cytolysis. To explore this possibility we examined the effect of four inhibitors of P-gp efflux [(R)-verapamil (R-ver), (S)-verapamil (S-ver), cyclosporine A (CsA), and PSC833 (PSC)] on both the inhibition of natural killer cell (NK) function and on P-gp efflux. NK function was assayed by measuring the lysis of 51Cr-labeled K562 target cells in the presence and absence of inhibitors. All four P-gp efflux inhibitors inhibited NK-mediated cytolysis in a dose-dependent manner. The stereoisomers of verapamil were more potent inhibitors of cell-mediated cytolysis than the cyclosporines CsA and PSC. In contrast, CsA and PSC were more potent as inhibitors of P-gp-mediated rhodamine 123 dye efflux than the verapamil isomers. Both CsA and PSC maximally inhibited P-gp efflux at 3 microM, but only minimally inhibited cell-mediated cytolysis. The verapamil compounds demonstrated closer correlation between efflux inhibition of NK-mediated cytolysis. The data support a role for P-gp in NK-mediated cytolysis; however, these studies also suggest that the NK cytolytic process is multifaceted and that inhibition of the P-gp-mediated efflux mechanism only partially abrogates this process.

Immunohistochemical detection of P-glycoprotein in teleost tissues using mammalian polyclonal and monoclonal antibodies

Mammalian P-glycoprotein is a highly conserved 170-kD integral plasma membrane protein functioning as an energy-dependent efflux pump of exogenous and endogenous lipophilic aromatic compounds entering the cell by diffusion. In this study, the tissue specificities of one polyclonal (pAb) and three monoclonal (mAbs) antibodies to mammalian P-glycoprotein were identified in paraffin-embedded, parasagittal whole-body sections of the guppy Poecilia reticulata. Polyclonal antibody mdr(Ab-1) and mAbs C219, C494, and JSB-1 demonstrated differential staining patterns in the following tissues: bile canaliculi in the liver, exocrine pancreas, lumenal surface of the intestinal epithelium, renal tubules, interrenal tissue, branchial blood vessels, gas gland, pseudobranch, and the gill transverse septa. Positive P-glycoprotein expression in P. reticulata correlates well with published results for homologous mammalian tissues of secretory and excretory function. These data indicate that one or more highly conserved members of the P-glycoprotein transporter family exist in a teleost species and can be detected using commercially available mammalian antibodies.

Analysis of a novel cDNA encoding a C219-reactive peptide isolated from methotrexate-selected multidrug-resistant human leukemic cells

We have isolated and sequenced a novel cDNA clone (D320), from methotrexate-selected multidrug-resistant (MDR) human leukemic CCRF-CEM cells, which encodes a peptide reactive with monoclonal antibody (mAb) C219. Despite having been cloned using this ostensibly P-glycoprotein-specific mAb, clone D320 has no significant homology with either the MDR-encoding gene or, at the amino acid (aa) level, with its product, P-glycoprotein. A putative C219-binding site has been identified in D320, which differs at two positions from the 6-aa C219 epitope previously described in P-glycoprotein.

Multidrug resistance

Since we found verapamil as a multidrug resistance (MDR) reversing agent in 1981, many MDR reversing compounds have been reported. This type of drug must have strong effects with little side effects. We recently found MS-209 and PSC-833 as reversing agents. These two compounds interacted directly with P-glycoprotein, and showed a good MDR reversing effect in vitro and in vivo. MRK16, an antibody against P-glycoprotein, also showed a good therapeutic effect against drug resistant human tumors. MS-209, PSC-833 and the antibody against P-glycoprotein are interesting candidates for clinical use in the future.

P-glycoprotein epitope mapping. II. The murine monoclonal antibody MM6.15 to human multidrug-resistant cells binds with three distinct loops in the MDR1-P-glycoprotein extracellular domain

A new murine monoclonal antibody (MAb), MM6.15, to human MDR1 P-glycoprotein was found to be reactive in ELISA with synthetic peptides selected from the predicted sequences of the first, fourth and sixth extracellular loop of MDR1-P-glycoprotein. In order to precisely define the MM6.15-binding site, a peptide library of overlapping 5- to 9-mer residues covering the entire sixth extracellular loop of both human and rodent class-1 P-glycoproteins was synthesized on polyethylene pins and tested for MAb binding. The results of this ELISA demonstrated that the MAb MM6.15 reacts only with human synthetic peptides and that the critical component of the MAb recognition is made up of the amino-acid sequence LVAHKL (residues 963-968 of the MDR1-P-glycoprotein) with histidine (H), lysine (K) and possibly leucine (L), key residues of this immunogenic domain.

P-glycoprotein expression in acute myeloblastic leukemia analyzed by immunocytochemistry and flow cytometry

Using the APAAP technique, we assessed the reactivity of five monoclonal antibodies, JSB1, C219, Mab 57, 2F8 and MRK16, to gp 170 in 60 cases of de novo acute myeloid leukemia (AML) and 13 relapses. Reactivity, varied between the five antibodies, and positivity was obtained with 2F8 > JSB1 > MRK16 > Mab57 > C219. Sixteen of the 60 cases were also studied by flow cytometry. In 10 cases, the results with the two techniques corresponded, in the other 6 cases, flow cytometry proved more sensitive than APAAP in detecting small amounts of gp170. In the flow cytometry analysis, the cells fixed in methanol and paraformaldehyde were more fluorescent than unfixed samples or those fixed in paraformaldehyde or methanol alone. Our results thus reveal that positivity for gp 170 depends on various factors, including the specificity of the monoclonal antibodies, the techniques used and the preservation of the samples. This suggests the need for a clear standardization of the methods to detect gp170.

Isoform I (mdr3) is the major form of P-glycoprotein expressed in mouse brain capillaries. Evidence for cross-reactivity of antibody C219 with an unrelated protein

P-glycoprotein (P-gp) is expressed in various non-cancerous tissues such as the endothelial cells of the blood-brain barrier. We used several monoclonal antibodies (mAbs) and isoform-specific polyclonal antibodies to establish which P-gp isoforms are expressed in isolated mouse brain capillaries. P-gp class I isoform was detected in capillaries with a Western immunoblotting procedure using a specific antiserum. No immunoreactivity was observed with either class II- or class III-specific antisera. Immunoreactivity was observed with mAb C219. However, this antibody detected two distinct immunoreactive proteins (155 and 190 kDa) in the isolated brain capillaries. These two proteins comigrated as a broad band when the samples were submitted to heat prior to gel electrophoresis. The glycoprotein nature of these two antigens was evaluated by their sensitivity to N-glycanase treatment. Following this treatment, the size of the proteins was reduced from 190 and 155 kDa to 180 and 120 kDa, respectively. Triton X-114 phase-partitioning studies showed that the 190 kDa immunoreactive protein was poorly solubilized by Triton X-114, while the 155 kDa protein was partitioned in the detergent-rich phase. In labelling experiments, only the 155 kDa protein was photolabelled with [125I]iodoarylazidoprazosin. These results show that a 190 kDa protein detected by antibody C219 is an antigen unrelated to the three P-gp isoforms presently known. Cross-reactivity of C219 with an unrelated protein emphasizes the fact that more than one antibody should be used in the assessment of P-gp expression in cell lines and tissues.

An M(r) 7-kDa membrane protein overexpressed in human multidrug-resistant ovarian cancer cells

We have developed a monoclonal antibody (designated 1D7) which recognizes an M(r) 7-kDa plasma membrane protein overexpressed in ovarian MDR cancer cells. The expression of the M(r) 7-kDa protein in various human multidrug-resistant and drug-sensitive cell lines was analysed by Western blot and flow cytometry methods. The small molecular weight protein was overexpressed in the human ovarian carcinoma cell line, SKVLB which was selected for vinblastine resistance from SKOV3 cells and in OVCAR 4/ADR100 and OVCAR 4/VBL200 which were generated from NIH:OVCAR4 by stepwise selection against adriamycin and vinblastine, respectively. Only a minor amount of the M(r) 7-kDa protein was found in the parent cell line, SKOV3. It was not found in other drug-resistant human cell lines such as the vinblastine-resistant CEM cells (CEM/VLB300), the intrinsic MDR colon cell line HCT15 and the human MDR breast cancer cell line, MCF7/AdrVp. 1D7 specifically inhibited the proliferation of the resistant cells. Our results suggest that the expression of the M(r) 7-kDa protein on the plasma membrane of ovarian MDR cancer cells may be involved in a mechanism related to the proliferation of the drug resistant cancer cells.

P-glycoprotein of blood brain barrier: cross-reactivity of Mab C219 with a 190 kDa protein in bovine and rat isolated brain capillaries

P-glycoprotein (P-gp), an active efflux pump of antitumor drugs, is strongly expressed in endothelial cells of the blood brain barrier (BBB). Two proteins (155 and 190 kDa) were detected by Western blot analysis of beef and rat capillaries with the monoclonal antibody (MAb) C219. In order to characterize the nature of these proteins, their profile of solubilization by different detergents was established and compared with that of P-gp from the CHRC5 tumoral cell line. The 155 kDa protein (p155) of capillaries and the P-gp of CHRC5 cells were well solubilized by deoxycholate and Elugent, whereas the 190 kDa kDa protein (p190) was only solubilized by sodium dodecylsulfate (SDS). Both proteins have different patterns of extraction by Triton X-114, p155 partitioning as a membrane protein, while p190 was insoluble. Deglycosylation of capillary proteins resulted in a 27-28 kDa decrease in the apparent molecular weight of p155, similar to that observed for the P-gp of CHRC5 cells, but a decrease of only 7-8 for p190. Only p155 was immunoprecipitated by MAb C219. These results suggest that only p155 is the P-gp in BBB and that MAb C219 cross-reacts with a 190 kDa MDR-unrelated glycosylated protein. Consequently, the use of this antibody, which is frequently used to detect P-gp in tumors, could be a pitfall of immunohistochemistry screening for cancer tissues and lead to false positive in the diagnosis of MDR.

Enhancement by recombinant human interferon alfa of the reversal of multidrug resistance by MRK-16 monoclonal antibody

BACKGROUND

The anti-P-glycoprotein monoclonal antibody MRK-16 mediates the reversal of multidrug resistance. Recombinant human interferon alfa (rHuIFN alpha) enhances the cytotoxic activity of diverse chemotherapeutics and may modulate multidrug resistance.

PURPOSE

Our purpose was to determine the outcome of combination treatment with MRK-16, rHuIFN alpha-2a, and cytotoxic agents on tumor cells that express P-glycoprotein (Pgp).

METHODS

Three Pgp-expressing, multidrug-resistant human tumor cell lines were used: the MDR1 retrovirus-infected HT-29 colon adenocarcinoma (HT-29mdr1), the doxorubicin (Adriamycin)-resistant MCF-7 (AdrR MCF-7) breast carcinoma, and the de novo Pgp-acquired, HCT-15 colon carcinoma. The parental cell lines HT-29par and MCF-7 were used as controls. The in vitro effects of MRK-16 and rHuIFN alpha-2a were studied on: (a) chemosensitivity of parental and multidrug-resistant cell lines to vincristine, doxorubicin, or paclitaxel (Taxol); (b) intracellular drug concentrations; and (c) Pgp expression. The efficacy of vincristine alone or in combination with MRK-16 and/or rHuIFN alpha-2a was assessed against HT-29mdr1 cells in female, athymic NCr-nu/nu mice.

RESULTS

For vincristine, the IC50 (i.e., the concentration that causes 50% inhibition of cell growth) was 7.0 ng/mL in HT-29mdr1 cells. Pretreatment of HT-29mdr1 cells with MRK-16 partially restored vincristine sensitivity (IC50 = 4.8 ng/mL), which was enhanced by noncytotoxic concentrations of rHuIFN alpha-2a (IC50 = 2.9 ng/mL) via a mechanism independent of Pgp modulation or [3H]vincristine efflux. rHuIFN alpha-2a potentiated MRK-16 reversal of multidrug resistance with both doxorubicin and paclitaxel on HT-29mdr1 cells and with vincristine on AdrR MCF-7 and HCT-15 tumor cells. Treatment of mice with 1 mg/kg vincristine weekly for 3 weeks, beginning 10 days after tumor injection, significantly increased the median survival times of the HT-29par tumor-bearing mice (60 days versus 35 days; P < .0001) but was only marginally therapeutic for HT-29mdr1 tumor-bearing mice (52 days versus 46 days). Pretreatment with MRK-16 (500 micrograms) and rHuIFN alpha-2a (5 x 10(4) U), alone or in combination, 24 hours before vincristine therapy did not affect the survival of HT-29par tumor-bearing mice. In contrast, the survival of mice bearing HT-29mdr1 tumors was significantly increased following treatment with MRK-16 before vincristine (80 days; P < .0001). Administration of a nontherapeutic dose of rHuIFN alpha-2a (5 x 10(4) U) with MRK-16 before vincristine treatment further increased the median survival times of HT-29mdr1 tumor-bearing mice (116 days; P < .0001).

CONCLUSIONS

MRK-16 used in combination with rHuIFN alpha-2a was significantly more effective than MRK-16 in overcoming multidrug resistance.

Characterization of the M(r) 190,000 multidrug resistance protein (MRP) in drug-selected and transfected human tumor cell

Overexpression of multidrug resistance-associated protein (MRP) has been detected in resistant cell lines derived from a variety of tumor types. The deduced amino acid sequence of MRP suggests that it is a member of the ATP-binding cassette transmembrane transporter superfamily that may be glycosylated and/or phosphorylated [S. P. C. Cole et al., Science Washington, DC), 258: 1650-1654, 1992]. Recently, transfection of HeLa cells with MRP expression vectors has demonstrated that the protein is capable of increasing resistance to natural product drugs such as anthracyclines, Vinca alkaloids, and epipodophyllotoxins (C. E. Grant et al., Cancer Res., 54: 357-361, 1994). Although the resistance phenotype of the transfectants is similar to that of the human small cell lung cancer cell line, H69AR, from which MRP was originally cloned, the transfectants differ in their drug accumulation characteristics, relative resistance to certain drugs, and MRP mRNA:protein ratio. Such differences have also been observed among drug-selected cell lines that overexpress MRP, and the underlying causes of these variable phenotypes are presently not known. We have utilized polyclonal anti-MRP-peptide antibodies to compare MRP post-translational modification, stability, processing, and subcellular distribution in the HeLa transfectants and in the drug-selected H69AR cells. These studies establish that MRP in both the transfected and selected cells is an ATP-binding, integral membrane glycophosphoprotein with an apparent molecular weight of 190,000. No obvious differences were detected in the extent or type of glycosylation or the kinetics of processing and turnover of the protein that might contribute to the different characteristics of the transfected and drug-selected cells. Analyses of the subcellular distribution of MRP by isopyknic density gradient centrifugation revealed that approximately 80% of MRP in the HeLa transfectants was associated with a low density plasma membrane fraction while the comparable fraction in the drug-selected H69AR cells contained only approximately 50% of the protein. The remaining MRP and plasma membrane markers were codistributed in higher density fractions consistent with the presence of MRP in endocytotic vesicles. The relatively high proportion of MRP associated with these fractions in H69AR cells may contribute to the lack of an observable accumulation defect in these cells when compared with the transfectants.

Liposomal vaccines

Liposomes have been used therapeutically to deliver drugs to certain anatomical sites. The use of liposomes to deliver antigens, although not a new concept, has received less attention. At least two vaccines of nearly identical liposome base composition to our vaccines have been tested in humans. A malaria vaccine study showed that the liposomal preparation is quite safe: reaction profiles of volunteers receiving the vaccine demonstrated little reactivity and virtually no pyrogenicity (14). The concentration of MPLA in the vaccine was substantially higher (nearly 50,000 times) than the pyrogenic dose of free lipid A. The same vaccine, but different antigen (gp120, an HIV protein), was tested in volunteers and had the same lack of toxicity (27). In both studies, antibodies and cytotoxic cells specific for the respective antigens were produced. We have several subunit vaccines under development for infectious diseases (gram negative sepsis, fungal infections, protozoan infections), metabolic disorders (hypercholesterolemia, diabetic retinopathy, macular degeneration), and neoplastic diseases (multi-drug resistant cancer, primary and metastatic tumors, and angiogenic hyperproliferative disorders). In each case, one or more antigens were identified that might be useful in immunologic control of biologic proliferation (i.e., pathogen or tumor growth, rise in serum cholesterol, growth of blood vessels). We anticipate that at least one of these vaccines will be ready for testing in humans in the next calendar year.

ATP-activated chloride channel inhibited by an antibody to P glycoprotein

In this report, we present the characteristics of a Cl- channel found in lens fiber cells. The single channel has a conductance of 17 pS, a linear current-voltage curve, is activated by ATP or strong depolarization and is blocked by verapamil, quinidine, 4,4'-diisothiocyanostilbene-2,2'-disulfonic acid, 5-nitro-2-(3- phenylpropylamino)benzoate, dideoxyforskolin, and tamoxifen. These properties are similar to those reported for a volume-activated Cl- channel associated with the multidrug resistance (MDR) gene product, P glycoprotein (24). Confirming this connection, we demonstrate that our lens Cl- channel is inhibited by an antibody to P glycoprotein. The data we present here may, therefore, be the first characterization of the single channel activity of the Cl- channel associated with P glycoprotein.

Monoclonal antibodies against MDR1 P-glycoprotein inhibit chloride conductance and label a 65-kDa protein in pancreatic zymogen granule membranes

The regulation of Cl- and cation conductances by the nonhydrolyzable ATP analog adenosine 5'-(beta,gamma-methylene)triphosphate (AMP-PCP) was characterized in isolated zymogen granules (ZG) from pancreatic acinar cells. ZG were purified from rat pancreas homogenate by Percoll gradient centrifugation. Cl- conductance was assayed by suspending ZG in isotonic KCl buffer and measuring osmotic lysis induced by maximal permeabilization of ZG membranes (ZGM) for K+ with the K+ ionophore valinomycin (Val). This resulted in influx of K+ through the artificial pathway and of Cl- through endogenous channels. To measure cation conductances ZG (pHi approximately 6) were suspended in pH 7 buffered isotonic monovalent cation acetate salts. The pH gradient was converted into an outside-directed H+ diffusion potential by maximally increasing H+ conductance of ZGM with the protonophore carbonyl cyanide p-chlorophenylhydrazone. Osmotic lysis of ZG was induced by H+ diffusion potential driven influx of monovalent cations through endogenous channels and non-ionic diffusion of the counterion acetate. In the absence of Val, ZG were stable in KCl buffer up to 2 h. AMP-PCP enhanced osmotic lysis approximately 4-fold compared to control, due to activation of Cl- conductance by AMP-PCP and K+ influx through an AMP-PCP-insensitive nonselective cation pathway, which could be blocked by 0.1 mM Ba2+, 0.5 mM quinine, or 0.2 mM flufenamate. In addition, a K+ and Rb+ selective cation conductance was found which was completely blocked by 0.5 mM AMP-PCP or 0.5 mM quinine. AMP-PCP induced Cl- conductance was strongly inhibited by two monoclonal antibodies against MDR1 P-glycoprotein (JSB-1 and C219; 5-10 micrograms/ml), but not by a monoclonal antibody against the cystic fibrosis transmembrane conductance regulator (M3A7; 5 micrograms/ml) or by mouse IgG. The AMP-PCP insensitive nonselective cation conductance was not blocked by monoclonal antibodies against MDR1 P-glycoprotein (MDR1). Immunoblot studies of ZG membranes revealed the presence of a major immunoreactive protein band of approximately 65 kDa with both monoclonal antibodies against MDR1, but no protein of the approximate size of MDR1 (approximately 170 kDa) was detected. We propose that the Cl- channel or a regulator of the channel, that is activated by the non-hydrolyzable ATP analog AMP-PCP in ZG membranes, is a member of the ATP binding cassette superfamily of transporters and may have homology to MDR1 P-glycoprotein.