SHORT COMMUNICATION Selective growth-inhibition of multidrug-resistant CHO-cells by the

The development of drug resistance is a major problem in effective chemotherapy of cancer. During the past few years, the phenomenon of multidrug resistance (MDR) has been described (Ling et al., 1983; Volm et al., 1987). Frequently, MDR-cells overexpress a membrane protein (P-glycoprotein, P-170) which is thought to function as an efflux pump for different cytostatics and thereby cause the development of drug resistance (Juliano & Ling, 1976; Chen et al., 1986). P-glycoprotein could serve as a target for the selective killing of MDR-cells. In an effort to devise an effective treatment for drug-resistant tumours we have evaluated the therapeutic potential of the monoclonal antibody (MAb) 265/F4 against P-glycoprotein (Lathan et al., 1985) with regard to its ability to inhibit growth of MDR-cells. Furthermore, we have constructed an immunotoxin by coupling ricin-alpha chain (RAC) to MAb 265/F4. Our results indicate that MAb 265/ F4 and 265/F4-RAC conjugate may be important weapons for the selective killing of MDR-cells. -For this investigation sensitive and colchicine-resistant Chinese hamster ovary (CHO) cells were used (obtained from Dr V. Ling, Ontario Cancer Institute, Toronto, Canada). The CHO-cells were cultured in a-MEM medium with ribonucleosides and deoxyribonucleosides (Biochrom, Berlin, Germany) supplemented with 10% foetal calf serum. The colchicineresistant subline (CHO-C5R) was maintained continuously in the presence of 5 fig ml-' colchicine. The MDR-phenotype of these cells has been reported earlier (Bech-Hansen et al., 1976). The detection of P-glycoprotein by Western blotting

The specificity of MAb 264/F4 for P-glycoprotein was demonstrated by Western-blotting. Membrane-fractions of The binding capability of 265/F4-RAC conjugate and the presence of ricin in the conjugate was tested by radioimmuno-assay (Table I). 265/F4-RAC immunotoxin bound strongly to resistant cells, whereas the binding to sensitive cells was low. Thus, 265/F4-RAC is able to detect P-glycoprotein in resistant CHO-cells. An anti-RAC antibody was used to detect ricin in the conjugate. The c.p.m.-values detecting RAC and MAb 265/F4 in the conjugate are comparable indicating that one ricin molecule is bound to one antibody molecule. Since the coupling of RAC to antibodies may reduce the binding ability to the target protein, we compared the binding of 265/F4-RAC and uncoupled 265/F4 to resistant CHO cells. As can be seen in Table I, there are no differences in the reactivity between the immunotoxin and the native antibody. In control measurements, the anti-RAC antibody failed to detect uncoupled 265/F4 indicating the specificity of the radioimmuno-reaction.
In order to examine whether MAb 265/F4 possesses an inhibitory effect on cell growth, we exposed sensitive and resistant CHO-cells to different concentrations of this antibody in drug-free medium. As can be seen in Figure   MAb 265/F4 inhibited the growth of multidrug-resistant CHO-cells in a dose-dependent manner. Maximal effect (75% inhibition of growth) was observed at a concentration of 1OOLgml-l MAb 265/F4 after a 7-day exposure of the antibody. In contrast to resistant cells, MAb 265/F4 had only little effect on the growth of sensitive CHO-cells. We measured a growth inhibition up to 20% as compared to untreated control cells.
The antitumour activity of the 265/F4-RAC conjugate on sensitive and multidrug-resistant CHO-cells is shown in Figure 2b. Again, the growth of resistant cells could be inhibited selectivity dependent on the concentration of 265/ F4-RAC conjugate. At a concentration of 3 gig ml ' cell growth was reduced by 90%. In contrast, the same concentration of immunotoxin caused only a growth inhibition of 10% in sensitive cells. In control experiments an inhibitory effect of uncoupled ricin alpha chain was not observed (data not shown).
Furthermore, we analysed the effect of MAb MRK 16 which detects P-glycoprotein in human cells, but not in CHO cells on cell growth of sensitive and resistant CHO-cells. In this experiment we proved whether a specific targetting of Doses (,ug ml -') P-glycoprotein is required to inhibit cell growth or whether increased unspecific membrane turnover and protein internalisation in resistant cells result growth inhibition. As can be seen in Figure 3, MAb MRK 16 had only marginal effect on cell growth of both sensitive and resistant CHO cells. a Thus far, several monoclonal antibodies (265/F4 (Lathan et al., 1985); C219, C494 (Kartner et al., 1985), 32G7 (Danks et al., 1985), MRK 16, MRK 17 (Hamada & Tsuruo, 1986), JSB-1 (Scheper et al., 1988), HYB-612 (Meyers et al., 1989), and polyclonal antibodies P7 (Richert et al., 1988), anti-P, anti-C (Yoshimura et al., 1989), anti-PO, anti-P4, 4007, 4077 (Bruggemann et al., 1989)) against P-glycoprotein has been prepared. However, only few of these antibodies (265/F4, MRK 16, MRK17, and HYB-612) were developed against intact tumour cells and recognised external epitopes of Pglycoprotein. The external targetting of surface proteins by monoclonal antibodies can be used for immunotherapeutical approaches. Therefore, antibodies recognising external epitopes of P-glycoprotein might be suitable for an efficaceous treatment of MDR-cells. Indeed, our results and the experiments of Hamada and Tsuruo (1986) indicate that the antibodies 265/F4, MRK 16 and MRK 17 possess biological activity against MDR-cells in vitro. Both MRK 16 and MRK 17 have been used for the selective growth inhibition of human MDR-tumours in nude mice (Tsuruo et al., 1989). In contrast to the growth inhibitory feature of 265/F4, application of this antibody to drug-containing medium showed that 265/F4 is unable to modulate drug accumulation (unpublished results). The biochemical mechanism of the selective growth-inhibitory effect of MAb 265/F4 on multidrug-resistant CHO-cells is until now unclear. We found a marginal inhibition effect of MAb 265/F4 on sensitive CHO-cells. MAb MRK 16 which does not cross-react with P-glycoprotein of CHO-cells also inhibited growth of sensitive and resistant cells to a small extent. We suggest that these marginal effects were not due to binding of P-glycoprotein. This might reflect unspecific pinocytosis of MAb 265/F4 and MAb MRK 16, respectively, together with culture medium.
Furthermore, we showed that a MAb 264/F4-ricin alpha chain conjugate is able selectively to inhibit the growth of multidrug-resistant CHO-cells. These data are in accordance with Fitzgerald et al. (1987), who used MAb MRK 16 coupled to Pseudomonas exotoxin as immunotoxin to kill multidrug-resistant KB-cells in vitro. Effective doses of MRK 16-Pseudomonas exotoxin are lower (1-10ngml-', Fitzgerald et al., 1987) than those of 265/F4-RAC (0.3-3ljg ml-'). Similarly, biological activity of uncoupled MRK 16 was also observed at lower concentrations (1 lgml-', Hamada & Tsuruo, 1986) as compared to uncoupled 265/F4 (10-I00 ig ml-'). This might reflect differences in the binding affinities of the two antibodies to P-glycoprotein. The selective growth inhibition by MAb 265/F4 and 265/F4-RAC immunotoxin in vitro raises the possibility. of using this antibody for the treatment of human tumours which are unresponsive to conventional therapy by doxorubicin or vincristine. However, before such immunotoxicological approaches for the treatment of MDR-cells can be established, several problems remain to be solved. One substantial obstacle is the expression of P-glycoprotein in certain normal tissues, e.g. kidney, liver, colon and brain capillaries (Sugawara et al., 1988). Nevertheless, the selective killing of MDR-cells by 265/F4 should encourage further investigation which may provide guidelines for the improvement of conventional chemotherapy.
We thank Dr T. Tsuruo for the generous provision of MRK 16. We are indebted to Dr J.M. Walsh for critically reading the manuscript.