Multidrug-resistance phenotype of a subpopulation of T-lymphocytes without drug selection

Experimental

Multidrug-Resistance T-Lymphocytes ALEXANDER ALEXANDER

Cell Research

Phenotype without

18.5 (1989) 496-505

of a Subpopulation Drug Selection

of

A. NEYFAKH,“,’ ANNA S. SERPINSKAYA,* V. CHERVONSKY,? SERGEY G. APASOV,? and ALEXANDER R. KAZAROV?

*Belorersky Laboratoty of Molecular Biology and Bioorganic State University, Moscow 119899, USSR, and TAD-Union Center AMS USSR, Kashirskqve shosse 24, Moscow

Chemistry, Moscow Cuncer Research 115478, USSR

Multidrug-resistant (MDR) cells demonstrate the increased activity of the membrane transport system performing efflux of diverse lipophylic drugs and fluorescent dyes from the cells. In order to detect MDR cells we have developed a simple test consisting of three steps: staining of the cells with fluorescent dye rhodamine 123, incubation in the dye-free medium and, finally, detection by fluorescence microscopy of the cells that have lost accumulated dye. The experiments with B-lymphoma cell lines with different degrees of MDR have shown that the cell fluorescence after the poststaining incubation is indeed inversely proportional to the degree of resistance. Application of this testing procedure to normal human or mouse leukocytes revealed the presence of the cells rapidly losing the dye in these populations. Cell fractionation experiments have shown that there are Tlymphocytes (most T-killers/suppressors and a part of T-helpers) that demonstrate rapid efflux of rhodamine 123. This characteristic was detected also in T-killer clones and cell line and in some T-lymphomas. The inhibitors of the MDR transport system, reserpine and verapamil, blocked the efflux of the dye from these cells. Rhodamine-losing T-lymphoma contained large amounts of the mRNA coding P-glycoprotein, the MDR efflux pump, and demonstrated increased resistance to rhodamine 123, gramicidin D, colchicine, and vincristine, the drugs belonging to the cross-resistance group for the MDR cells. The role of the increased activity of the MDR membrane transport system in T-lymphocytes is discussed. 0 1989 Academic Press, Inc.

Cultured cells selected by cytostatic drugs often acquire resistance not only to the selecting agent but to many structurally unrelated drugs differing in the mechanisms of their action. This type of resistance was termed multidrug resistance (MDR) (for review see [l]). The basis of MDR is an active efflux of drugs from the cytoplasm of resistant cells. The transport system performing the efflux has such a broad chemical specificity that it readily excretes from cells not only drugs but also various fluorescent dyes [2]. It is believed that efflux is performed by the membrane P-glycoprotein, which is homologous to certain bacterial transport proteins [3-51, demonstrates ATPase activity [61, and binds some of the drugs 171; this binding is sensitive to inhibitors of MDR 18, 91. In MDR cells the mdr gene coding this glycoprotein is overexpressed [lo] or amplified (see [l]). As a result, these cells contain a much greater quantity of P-glycoprotein molecules than the parental nonselected cell line. ’ To whom reprint requests should be addressed. Copyright @ 1989 by Academic Press. Inc. AU rights of reproduction in any form reserved 00144827ilJ9 $03.00

496

Multidrug

resistance

in T-lymphocytes

497

Nevertheless, activity of the MDR transport system can be detected even in nonselected cells. We have found that cells of different nonselected lines slowly release the dyes, such as rhodamine 123 and phosphine 3R. More important, this release was found to be sensitive to inhibitors of MDR [ 111. Similar phenomena were observed with MDR cells, but in these cells the rate of the dye efflux is much higher [2]. The use of fluorescent dyes as probes has a serious advantage: cells rapidly releasing the dyes can be easily detected by fluorescence microscopy. Here we describe a simple test for detection of MDR cells that is based on this principle. Using this test we have found that the high activity of the MDR transport system is characteristic not only of cells selected by the drugs but also of a large fraction of normal T-lymphocytes and some nonselected T-cell lines.

MATERIALS

AND

METHODS

Chemicals and immunochemicals. Rhodamine 123, reserpine, gramicidin D, colchicine, aflinitypurified goat anti-mouse IgG, and FITC-conjugated antibodies were purchased from Sigma Chemical CO. (St. Louis, MO). Phosphine 3R and phloxin B were from Gurr, Ltd. (UK) and Chroma, Schmid & CO., respectively. Pharmaceutic preparations of verapamil (Isoptin, Yugoslavia) and vincristine (Richter, Hungary) were used. Cell culture media and sera were from Flow Laboratories (UK). AntiLy2 and L3T4 rat monoclonals were kindly provided by Dr. V. Holan (Institute of Molecular Genetics, Prague). G4 anti-Thy1.2 IgM monoclonals were obtained by the fusion of X63 Ag8.653 myeloma cells with splenocytes of AKR mice immunized with CBA splenocytes. Cells. All cell lines were cultivated in RPMI-1640 medium supplemented with 10% fetal calf serum (37”C, 5% COJ. Medium of CTLL cells and T-killer clones was supplemented with 10% of interleukin 2-enriched medium (medium conditioned by Concanavalin A-stimulated rat splenocytes). T-killer clones were obtained from the lymph nodes of BlO.AKM mice immunized with splenocytes of BlO.MBR mice. The y-irradiated BlO.MBR splenocytes were added weekly to cultures of T-killer clones for antigen stimulation. Human donor blood was obtained from the Donor Blood Service of All-Union Cancer Research Center AMS USSR. Mouse blood, lymph nodes, spleens, and thymuses were isolated from Balb/c mice by conventional procedures. The peripheral blood leukocytes were isolated by sedimentation at unit gravity of freshly collected heparinized human or mouse blood. Peripheral blood lymphocytes were purified by centrifugation of human or mouse blood samples layered on Ficoll-Hypaque isotonic solution with densities of 1.077 or 1.09, respectively. Fractionation of mouse lymph node cells to B and T subpopulations was performed by a “panning” procedure. Cell suspension in 199 medium supplemented with 5 % fetal calf serum and 20 mM Hepes (pH 7.2) was incubated for 1 h at room temperature in a Costar 90-mm plastic petri dish. The dish was preincubated overnight with affinity-purified antibody to mouse 1gG (100 ugiml in borate buffer saline, pH 8.0) at 4°C. B-cells bearing surface Ig and monocytes adhered to the dish while T-cells remained nonattached. In some experiments collected T-cells were further subdivided into LyZ-positive and L3T4-positive subpopulations. Cells were incubated for 30 min with hybridoma culture supernatants containing anti-Ly2 or L3T4 rat monoclonals, and then washed extensively by repeated centrifugations and incubated for 1 h in an anti-mouse IgG precoated dish as described above. Since anti-mouse IgG readily cross-reacted with rat monoclonals, cells that bore Ly2 or L3T4 antigens adhered to the dish. Cells remaining in the suspension were collected and studied further. In order to control the purity of cell populations, conventional immunofluorescence of living cells attached to poly-t-lysine-coated coverslips was performed. More than 400 cells were scored in every preparation to determine the percentage of Thyl.2-, Ly2-, and L3T4-positive cells. Staining of the cells with the dyes. Cells suspended in 3 ml of 199 medium supplemented with 5 % fetal calf serum and 20 mM Hepes (pH 7.2) at a density of about 10’ per milliliter were stained with rhodamine 123, phosphine 3R, or phloxin B (5 ug/ml) for 15 min at 37°C. Then cells were washed by centrifugation with 5 ml of the same dye-free medium, centrifuged again, and either examined immediately or resuspended in 5 ml of dye-free medium and further incubated for 1 h at 37°C. In some experiments dye-free medium was supplemented with reserpine (2 &ml) over verapamil (10 !&ml).

498

Neyfakh

et al.

Quantitative fluorimetric earlier [l I].

measurements of cell-associated dye were performed exactly as described

Fluorescence microscopy. Microscopy was performed on the Zeiss photomicroscope III equipped with Planapo x40 lens. In the case of cells stained by FITC-conjugated antibody, rhodamine 123, and phosphine 3R, green fluorescence under a blue excitation light was studied; in the case of phloxin B, red fluorescence under a green excitation light was studied. Quantitative measurement of cell fluorescence was performed using a LOMO photometer connected with the microscope. The fluorescence of more than 40 randomly chosen cells was measured in every preparation. In order to determine the fraction of cells containing rhodamine 123, every field of view was examined under phase-contrast and then under excitation illumination. Cells that demonstrated even hardly noticeable fluorescence were regarded as dye-containing. More than 400 cells were scored in every preparation. The values obtained by two observers differed no more than on 5 %. Determination of cell sensitivity to cytostatic drugs. Lymphoma cells were seeded to 96-well plates with a density of 5~10~ per well. Medium in the wells (200 ~1) contained 1 : 2 dilutions of drugs. In some experiments the medium was supplemented also with reserpine (1 ug/ml). Three days later cells in the wells were counted using a hemocytometer. Concentrations of drugs inhibiting cell growth on 50% (IDJo) were determined from plots of cell number versus drug concentration. Northern blot hybridization. Total RNA was isolated by the guanidine isothiocyanate method and was subsequently centrifuged through CsCl [12]. RNA electrophoresis and Northern blot with cp22 clone were done according to standard protocol with formaldehyde denaturation [12]. Clone cp22 containing cDNA of the hamster mdrl gene was the kind gift of Dr. A. Van der Bliek [13]. RNA isolated from multidrug-resistant hamster libroblasts selected by colchicine (DM5/1 line, [14]) was used as a positive control.

RESULTS Adequacy of the Testing Procedure

The initial task of this work, as has been pointed out earlier, was to develop a test for detection of MDR cells. We have chosen a simple procedure consisting of three steps: cells are stained with rhodamine 123 (5 ug/ml, 15 min), incubated in the dye-free medium at 37°C for 1 h, and then examined by fluorescence microscopy in order to find cells that had lost the accumulated dye. To demonstrate the adequacy of this procedure several cell lines of the same origin with different degrees of MDR were developed. Mouse X63 Ag8.653 Blymphoma was selected with stepwise increasing concentrations of vincristine. Cell lines growing in the presence of 2, 4, 16, and 32 ng vincristine per milliliter were obtained. Concentrations of vincristine inhibiting the growth of each line on 50% were determined and were found to be 1 rig/ml for the parental cell line and 6.5, 16,25, and 35 r&ml for the cells X63Vcr2, X63Vcr4, X63Vcrr6, and X63Vcr32, respectively. Vincristine selection usually leads to formation of cell lines with MDR phenotype [ 1J and our test was able to detect it. Cells of all five lines fluoresced brightly after staining with rhodamine. However, after a l-h incubation in dye-free medium, cells of different lines fluoresced with different intensities. An inverse correlation between cell fluorescence and degree of resistance was observed (Fig. IA). Most of the cells of X63Vcr16 and X63Vcr32 lines were practically invisible with the fluorescence optics used (Fig: 2). In contrast, stained cells incubated in a dye-free medium containing reserpine, the inhibitor of the MDR transport system E9, 151, retained the dye and no significant difference in the intensities of fluorescence of different cell lines was observed (Fig. 1B). These results show that

Multidrug

resistance

in T-lymphocytes

499

PHOTOMETER UNITS

1

:

Fig. 1. Fluorescence resistance after staining or a dye-free medium X63Vcr4, (4) X63Vcr16, poststaining incubation,,

of a series of X63 Ag8.653 cell lines with different degrees of vincristine with rhodamine 123 and 1 h poststaining incubation in a dye-free medium (A) supplemented with reserpine (1 ug/ml) (E). (1) X63Vcr”, (2) X63Vc8, (3) (5) X63Vc?. The more cells are resistant, the less dye they retain after Reserpine blocks efflux of the dye from resistant cells.

fluorescence of the cell after staining and poststaining incubation medium is a good indicator of MDR transport system activity. High Activity

in dye-free

of the MDR Transport System in Normal T-Cells

The testing proizedure described above was applied to human and mouse peripheral blood leukocytes. All leukocytes fluoresced brightly after staining with rhodamine 123. However, about 10% of the cells completely lost the dye during the l-h incubation in dye-free medium; all these cells had the appearance of lymphocytes under phase-contrast optics. We have purified lymphocytes from human and mouse blood. In both cases about 50% of lymphocytes lost the dye during poststaining incubation. All these cells could be stained again by repeated incubation with rhodamine 123, suggesting that loss of the dye was not caused by damage to mitochondria, the targets of the dye binding. More significantly, we have found that loss of the dye could be completely prevented by reset-pine and verapamil, inhibitors of the MDR transport system [8, 9, 15, 161. Mouse lymphoid organs, such as lymph nodes, spleen, and thymus, contained different quantities of rhodamine-losing cells: 35, 20, and 5%, respectively. To investigate the origin of these cells, cells of lymph nodes were divided into B and T subpopulations by the “panning” procedure. Cell suspension was incubated in a plastic dish coated with affinity-purified antibody to mouse IgG. The population adhered to the dish, which presumably included B-lymphocytes and macro-

500 Neyfakh et al.

Fig. 2. Fluorescence microscopy of the cells of X63Vcr’ (A), X63Vcr4 (B), and X63Vc? (C) cell lines stained with rhodamine 123 and postincubated in a dye-free medium for I h. In (C) six cells were present in the field. All photographic procedures were performed with similar exposition times. The brightness of cell fluorescence is inversely proportional to cell resistance.

phages, essentially did not contain dye-losing cells (not more than 5%). In contrast, 60-65% of nonadherent cells lost the dye in the reset-pine-sensitive manner. These nonadherent cells belonged to the T-subclass, since about 95 % of them bore the Thyl.2 antigen (Table 1). Mouse T-cells can be further divided by immunological means into two subclasses: Ly2-bearing cells (killers/suppressors) and L3T4-positive cells (helpers). To do so, purified lymph node T-cells were incubated with rat monoclonal antibodies to Ly2 and L3T4 and again incubated over an anti-Ig-coated plastic dish. As tested by immunofluorescence, nonadherent cells in the anti-Ly2 preincubated population contained predominantly L3T4-positive cells and vice versa (Table 1). We have found that 35-40% of L3T4-enriched cells (helpers) lost rhodamine during poststaining incubation. The proportion of such cells in the population of killers/suppressors was much higher: 90-95 % (Table 1). High Activity of the MDR Transport System in Cultured

T-Cells

We have found that killer cells retain high MDR transport system activity after transfer to cell culture. Analysis of two interleukin 2- and antigen-dependent TABLE The phenotype Subpopulation T-cells L3T4+-depleted Ly2+-depleted

1

of the cells of T-cell subpopulations” Thyl.2+ 96 97 95

L3T4+

Ly2’

Dye-losing cells

81 13 97

17 89 4

59 (65)b 90 (94) 39 (34)

a The percentage of the cells with given phenotype is indicated. b The results of the independent experiment are given in parentheses.

Multidrug DYE

resistance

in T-lymphocytes

501

RE TENTION

% 100' 75'

50

25-

1oa L

.

60 MIN 30 Fig. 3. Efflux of rhodamine 123 from three lymphomas: X63 Ag8.653 (0), EL-4 (A), and RDM-4 (0). Cells were stained with the dye, washed, and then resuspended in a dye-free medium (solid lines) or a dye-free medium containing reserpine (broken lines). Cell-associated dye was determined fluorimetrically and expressed here as percentages of the values obtained immediately after staining. RDM-4 cells release rhodamine several times faster than two other lymphomas. Reserpine inhibits efflux of the dye from all three lymphomas.

0

killer clones has shown that 95 % of these cells completely release rhodamine during the l-h poststaining incubation. This release could be blocked by reserpine. In the pseudonormal interleukin 2-dependent T-killer mouse cell line CTLL, about 20% of cells completely lost the dye and the other 80% fluoresced weakly after poststaining incubation. Among three tested mouse T-lymphomas, two lines, RDM-4 and BW-5147, demonstrated fast efflux of rhodamine 123: after poststaining incubation fluorescence could be detected only in 5 % of RDM-4 cells and in 40% of BW-5147 cells. In contrast, all cells of T-lymphoma EL-4 retained their fluorescence. These results, based on visual impression, were confirmed by quantitative measurement of cell-associated rhodamine. Figure 3 illustrates the kinetics of the dye efflux from three cell lines. RDM4 T-lymphoma cells released rhodamine several times faster than B-lymphoma X63 and T-lymphoma EL-4. The ability of the cells to release rhodamine 123 correlated with the quantity of mdr mRNA. While these transcripts could hardly be detected in X63 cells, they were abundant in dye-losing RDM4 cells (Fig. 4). To further substantiate the conclusion that the cells rapidly releasing rhodamine 123 have increased MDR transport system activity, we compared two Tlymphomas, RDM-4 and EL-4, differing in the rate of rhodamine efflux. First, we

502

Nqfakh

et al.

A

6

C

Fig. 4. Northern blot hybridization of [“PJDNA of cp22 clone (m&l cDNA) with total RNA from multidrug-resistant DMS/l cells (A, positive control, 10 ug), X63 B-lymphoma (8, 20 ug), and RDM-4 T-lymphoma (C, 20 ug). Positions of 28 and 18 S rRNA are indicated. RDM-4 cells contain a significant number of mdr transcripts in contrast to X-63 lymphoma cells.

found that not only rhodamine but also the dyes phosphine 3R and phloxin B, staining other intracellular structures, leave RDM-4 cells much faster than EL-4 cells. The efflux of these two dyes was also found to be reserpine-sensitive (not shown). These results exclude the possibility that the difference between these two lines in rhodamine retention is a result of differences in their mitochondria, the targets of rhodamine binding. Second, we compared the sensitivities of dye-losing RDM-4 cells and dyeretaining EL-4 cells to different cytostatic drugs. As shown in Table 2, RDM-4 cells are 15 times more resistant to rhodamine 123, five times more resistant to

TABLE Sensitivities of RDM4

Cell line

2

and EL4 T-lymphomas to cytostatic drugs: Influence of reserpine’

Resetpine (1 i&ml)

Rho 123 Wml)

Gram D hi4ml)

Colch @idmU

Wml)

+

25.0 0.8

25.0 4.0

20 ndb

20 nd

+

1.7 0.4

4.5 3.5

7 nd

10 nd

RDM-4 EL-4

Vcr

Nore. Rho 123, rhodamine 123; Gram D, gramicidin D; Colch, colchicine; Vcr, vincristine. ’ The concentration of the drug inhibiting growth on 50% (ID5& is indicated. b nd, Not determined.

Multidrug resistance

in T-lymphocytes 503

gramicidin D, and twice as resistant to colchicine and vincristine as EL-4 cells. Reset-pine diminished the difference between these two lines in their sensitivities to rhodamine and gramicidin D. These results show that the dye-losing T-cells are indeed multidrug-resistant. DISCUSSION In order to detect cells with high activity of the MDR transport system, we have developed a simple procedure, consisting of staining of cells with rhodamine 123, poststaining incubation, and fluorescence microscopy. The adequacy of this procedure was demonstrated in two ways. First, we studied the model series of cell lines with different degrees of resistance, and a good correlation between intensity of cell fluorescence and degree of resistance was demonstrated: the more a cell is resistant, the less dye it retains after poststaining incubation. Second, cells detected by this testing procedure as multidrug-resistant were found to be resistant indeed and to contain large amounts of mdr mRNA (see the results on T-lymphomas). It should be noted that earlier Chen and co-authors, using a similar staining protocol, have shown that transformed cells retain rhodamine 123 better than their normal counterparts [17]. The ability to release rhodamine was even used for selection of normal revertants in the population of transformed cells [ 181. Considering our results one can suppose that the difference in dye retention between normal and transformed cells is due to the difference in the activities of the MDR transport system. This interesting suggestion remains to be tested. By using the developed procedure we show here that cells of a particular differentiation type, a subpopulation of normal T-lymphocytes (practically all killers/suppressors with some helper cells) demonstrates high MDR transport system activity without prior selection. This characteristic is retained in cultivated T-cells (T-killer clones and cell line) and in some T-lymphomas. T-lymphoma EL-4, which demonstrates slow release of rhodamine 123, is possibly of helper origin or has lost its MDR phenotype during cultivation. What is the possible role of the MDR transport system in T-cell physiology? Interestingly, this system is activated only in mature T-lymphocytes: young Tcells populating the thymus release rhodamine 123 slowly. One can suggest that MDR transport system activity is necessary for some immunological functions performed by mature T-cells. In recent years it was shown that several normal tissues contain a large number of mdr transcripts [19-211. It seems likely that high MDR transport system activity is characteristic of these tissues too. Among these tissues are adrenal, colon, liver, kidney, and uterine epithelium; all these cells perform transport and secretory functions. The MDR transport system could play a significant role in these processes. A similar function may be attributed to this system in T-cells: it could play a role in the secretion of diverse immunological factors released by these cells. Another hypothesis is that the activity of this system is responsible, at least partially, for the resistance of T-killers to their own lytic agent, perform, the 33-898342

504

Neyfakh

et al.

protein forming the pores in the target cell membrane. Three observations indirectly support this hypothesis. First, Yanovich et al. have shown [22] that cells selected for MDR are resistant to the lytic action of natural killer cells that also use perforin as a weapon. Second, P-glycoprotein is highly homologous to bacterial Hly B protein [3], which plays a direct role in transmembrane transport of hemolysin, another membrane-damaging polypeptide. Third, P-glycoprotein protects cells from the membrane-damaging oligopeptides gramicidin D ([ 11, this study) and cyclosporin A [23]. Thus, it would not be surprising if P-glycoprotein could expel perforin from the membrane of T-killers. Though these considerations do not constitute proof, they indicate that this hypothesis deserves direct testing. In conclusion, our results demonstrate that the testing procedure described here, which is based on the estimation of the efflux rate of rhodamine 123, is a reliable and simple method for the detection of cells with high MDR transport system activity. This procedure can be especially useful when such cells compose the minority in the complex cell population. Its potential practical application is the detection of MDR cells in tumor cell populations after chemotherapy. This testing procedure enabled us to detect high activity of the MDR transport system in a large fraction of T-cells. Further study of this phenomenon could clarify some aspects of T-cell functioning and could help in understanding the role of the MDR transport system in normal cell physiology. We thank Professor J. M. Vasiliev and Professor I. M. Gelfand for the fruitful idea of examining normal blood for the presence of cells with MDR phenotype. We also thank Professor J. M. Vasiliev and Dr. A. V. Gudkov for helpful discussions.

REFERENCES 1. Bradley, G., Juranka, P. F., and Ling, V. (1988) Biochim. Biophys. Acta 948, 87. 2, Neyfakh, A. A. (1988) Exp. Cell Res. 174, 168. 3. Gerlach, H., Endicott, J., Juranka, P., Henderson, G., Sarangi, F., Deuchars, K., and Ling, V. (1986) Nature (London) 324, 485. 4. Gros, P., Croop, J., and Housman, D. (1986) Cell 47, 371. 5. Chen, C.-J., Chin, J. E., Ueda, K., Clark, D. P., Pastan, I., Gottesman, M. M., and Roninson, I. (1986) Cell 47, 381. 6. Hamada, H., and Tsuruo, T. (1988) J. Biol. Chem. 263, 1454. 7. Comwell, M. M., Safa, A. R., Felsted, R. L., Gottesman, M. M., and Pastan, I. (1986) Proc. Natl.

Acad.

Sci.

USA 83, 3847.

8. Comwell, M. M., Pastan, I., and Gottesman, M. M. (1987) J. Biol. Chem. 262, 2166. 9. Beck, W. T., Cirtain, M. C., Glover, C. J., Felsted, R. L., and Safa, A. R. (1988) B&hem. Biophys.

Res. Commun.

153, 959.

10. Shen, D. W., Fojo, A., Chin, J. E., Roninson, I. B., Richert, N., Pastan, I., and Gottesman, M. M. (1986) Science 232, 643. 11. Neyfakh, A. A., Dmitrevskaya, T. V., and Serpinskaya, A. S. (1988) Exp. CellRes.178,513. 12. Maniatis, T., Fritsch, E. E., Sambrook, J. (1982) in Molecular Cloning, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 13. Van der Bliek, A., Van der Velde-Koerts, T., Ling, V., and Borst, P. (1986) Mol. Cell. &ol. 6, 1671. 14. Gudkov, A. V., Chemova, 0. B., Kazarov, A. R., and Kopnin, B. P. (1987) Somatic Cell Mol. Genet.

13, 609.

15. Inaba, M., Fujikara, R., Tsukagoshi, S., and Sakurai, Y. (1981) Biochem. Pharmacol. 16. Tsuruo, T., Iida, H., Tsukagoshi, S., and Sakurai, Y. (1982) Cancer Res. 42, 4730.

30, 2191.

Multidrug

resistance

in T-lymphocytes

505

17. Chen, L. B., Summerhayes, I. C., Nadakawukaren, K. K., Lampidis, T. J., Bemee, S. D., and Shepherd, E. L. (1984) in Cancer Cells: The Transformed Phenotype (Levine, A. J., et a/., Eds.), p. 75, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 18. Zarbl, H., Latreille, J., and Jolicoeur, P. (1987) Cell 51, 357. 19. Fojo, A. T., Ueda, K., Slamon, D., Poplack, D., Gottesman, M. M., and Pastan, I. (1987) Proc. Nazi.

Acad.

Sci.

USA

84, 265.

20. Mukhopadhyay, T., Batsakis, J. G., and Kuo, M. T. (1988) .Z. Natl. Cancer Inst. US 80, 269. 21. Arceci, R. J., Croop, J. M., Horwitz, S. B., and Housman, D. (1988) Proc. Nat/. Acad. Sci. USA 85, 4350. 22. Yanovich, S., Hall, R. E., and Weinert, C. (1986) Cancer Res. 46, 4511. 23. Goldberg, H., Ling, V., Wong, P. Y., and Skorecki, K. (1988) Biochem. Biophys. Res. Commun. 152, 552.

Received February 1, 1989 Revised version received July 12, 1989

Printed

in Sweden