Resistance to multiple chemotherapeutic agents in human cancer cells

Resistance to multiple chemotherapeutic agents in human cancer cells

TIPS - February 1988 iiiol. 91 54 importance of very large trials in detect reliably moderate benefits or ills from a particular In the past many tr...

661KB Sizes 1 Downloads 60 Views

TIPS - February 1988 iiiol. 91

54 importance

of very large trials in detect reliably moderate benefits or ills from a particular In the past many treatmenP3. trials have been hopelessly optimistic. If we are to evaluate the many possible treatments for heart attack properly we must recognize that trials involving about 20000 patients at a time are necessary. We therefore need to make them simple and cheap. We can reduce much of the work involved by telephone randomization and data entry, with a one page form at discharge to record death and/or complications in hospital, together with drugs given both in hospitai and on discharge home. The use of a hard endpoint (death) avoids bias. Subsequent follow-up (dead/ alive) can be done relatively through Government cheaply records without extra work for the local physicians. The trials can be made more efficient by the use of order


a factorial design which tests two or more interventions in the same patient population, as in ISIS-2 which tests aspirin as well as streptokinase2. q



Oral P-adrenoceptor blockade is widely used in coronary care units but often in inadequate dose, so that the achieved blockade is only partial and,. importantly, it is also late. Both MIAMI and ISIS-l show that the greatest benefit is seen in the first day or so. It therefore seems logical that this inadequate oral blockade should be replaced by the rapidly effective intravenous route.

References 1 GISSI Collaborative Group (1986) Lncet i, 397402

Resistance to muitiple chemotherapeutic agents in human cancer cells Michael M. Gottesman and Ira Pastan Human cancer cells selected for resistance to multiple hydrophobic cytotoxic drugs express the cell surface P-glycoprotein, which is the product of the MDRl gene. In this review Michael Gottesman and Ira Pastan discuss the genetic and biochemical data indicating that P-glycoprotein is a major component of an energy-dependent efflux system antagonized by specific agents that prevent binding of cytotoxic drugs to this protein. Expression of Pglycoprotein occurs in normal liver, kidney, colon, jejunum and adrenal gland, suggesting that this protein functions as a transport system for cytotoxic compounds. Many multidrug-resistant tumors arising from these tissues express elevated MDRl levels, as do some tumors which become drug-resistant during chemotherapy. The selection of cultured cells resistant to a variety of cytotoxic natural products such as colchitine, the Vinca alkaloids (c,g. vinblastine and vincristine), epipodophylotoxins, actinomycin D, Mickncl Cottesnm

is Clrirf of tlrc Mo/~TII/~~

Lnborntoij of Molcrrttnr Biology, kniiorrnl Carrcer frrsfitrrtr, Nntiotlnl hstitrrres of Henttk Brthsdn, Mnr!/ln,ld 20892, USA.

taxol and anthracyclines (e.g. doxorubicin) (some examples are shown in Fig. 1) allows the isolation of mutants which are crossresistant to all of these agents and to others. This tissue culture phenomenon has been called multidrug resistance (see Refs 1 and 2). Recently, it has been appreciated that the resistance of human tumors to chemotherapy follows a similar pattern in two

2 ISIS Steering Committee (1987) Lancet i, 502 Rossi, I’. R. F., Yusuf, S., Ramsdale, D., Furse, L. Y. and Sleight, P. (1983) Br. Med. 1.286,506-510 Yusuf, S., Ramsdale, D., Pete, R., Furse, L., Bennett, D., Bray, C. and Sleight, P. (1980) Lancet ii, 273-276 Ramsdaia, D. R., Faragher, E. B., Bennett, D. H., Bray, C. L., Ward, C., Cruickshank, J. M., Yusuf, S. and Sleight, P. (19&2) Am. Heart 1. 103,459467 Yusuf, S., Pete, R., Lewis, J., Collins, R. and Sleight, P. (1985) Prog. Cardiovasc. Dis. 27,335371 Sleight, P., Yusuf, S., Ramsdale, D., Rossi, P., Pete, R., Bennett, D., Bray, C. and Furse, L. (1982) Br. 1. CIEn. Pharmacol. 14,37s-41§ 8 Waagstein, F. and Hjalmarson, A. C. (1975) Acta. Med. Scund. 587,201-211 9 ISIS-1 (1st International Study of Infarct Survival) Collaborative Group (19%) Lancet ii, 57-66 10 The MIAMI Trial Research Group (1985) Eur. Heart 1. 6. 199-226 11 Sleight, P.- (1986) Annu. Rev. Med. 37, 41-25 12 Kostis, J. B.. and Rosen, R. C. (1987) Circuktion 75,204-212 13 Yusuf, S., Collins, R. and Pete, R. (1984) Stat. Med. 3, 409-420

situations. For many tumors, including adenocarcinomas of the colon and kidney, malignant cells are primarily unresponsive to multiple chemotherapeutic agents (intrinsic drug resistance). In other cases* such as childhood leukemias and neuroblastomas, there is a good response to initial chemotherapy, but tumors may become refractory to therapy because they become resistant to the drugs used for treatment as well as to many others (acquired drug resistance). The drugs of proven chemotherapeutic value to which multidrug-resistant mutar& are resistant include vinblastine, vincristine, etoposide (VP-16), teniposide (VM-26), doxorubicin (adriamycin), daunorubicin, plicamycin (mithramycin) and actinomycin D. These drugs do not share common cytotoxic targets within cells, but they are al! hydrophobic drugs derived from natural products. A model system to study multidrug resistance in human cancer cells To determine the molecular basis of drug resistance in human cancer, it was necessary to develop a model system in tissue

culture using a human carcinoma cell line. Previous studies by Beck

UPS - Februa y 1988 [Vol. 91 et al. had established such a system for human leukemic cellsj. KB carcinoma cells (actually a subclone of HeLa) were chosen because they were drug-sensitive, grew rapidly, cloned with high efficiency, and could be selected for the multidrug resistance phenotype with relative ease. KB cells were selected in multiple steps with drugs known to give rise to multidrug-resistant mutants: colchicine, doxorubicin and vinblastine4a5. For each drug, and at each selection step, crossresistance to the kinds of drug listed above was found. A genetic analysis of these human KB multidrug-resistant cells provided evidence that mutations affecting a single step, a small cluster of linked genes or a coordinately regulated complex of genes are responsible for multidrug resistance. For example, in the absence of selective media the cells revert and simultaneously lose resistance to .a11 the agents ([email protected] The complete multidrug resistance phenotype is dominant in somatic cell hybrids and all drug resistances cosegregate when chromosomes are lost in these hybrids4. In experiments using DNA extracted from multidrug-resistant KB cells, the complete phenotype can be transferred using DNA fragments approximately 150 kb in length6. Characteristic6 of multidrugresistant cells The human multidrug-resistant KB mutants were found to share characteristics with mutants isolated from rodent and other human cell lines in the laboratories of Ling, Biedler, Dano, Beck, Tsuruo and many othersr. Mutant cells show decreased drug accumulation which correlates with the extent of their resistance7f8. This decreased accumulation results from increased drug efflux, as opposed to decreased uptake, Drug efflux can be blocked with inhibitors of energy metabolism, such as 2-deoxyglucose and azide, suggesting that multidrug resistance is due to an energydependent effiux pumpe. In tissue culture, the multidrug resistance phenotype can be completely suppressed by a variety of drugs including the Ca*+ channel blocker verapamil (Fig, 1) and non-Ca*+ channel blockers such as quini-











Fig. 1. Structural formulas of several drugs used in these studies. Mullidtug.resistant cell lines are cross-rtuistant to vinblastirte,daunorubicin, actinomycin 0 and colchiclne, among many other drugs. Verapamil and other agents phenotyplcally reverse mulldrug mslstance. [‘s&lMSVoan be used to photoamnnylabel P~giycoptutein, the putatiw drug fransport pump in multidrug-resistant

dine and reserpine. These drugs all appear to make multidrugresistant cells sensitive to cytotoxic agents by increasing the accumulation of drugs in the cellse~9. A variety of protein changes in multidrug-resistant cell lines has been described, but the only consistent biochemical alteration shared by all multidrug-resistant cdl lines is an increase in a i70 kDA membrane glycoprotein, first described by Juliano and Ling in Chinese hamster cell lines”, and termed P170 or Pglycoprotein (P for permeability) by these workers. Properties of the P-glycoprotein Because of its increased expression in multidrug-resistant

cells, P-glycoprotein was a good candidate to be a drug efflux pump. Using monoclonal antibodies to P-glycoprotein”, Willingham et al. showed that Pglycoprotein in human multidrugresistant cells is primarily a membrane proteinT2. plasma Comwell et al. used a photoaffinity analo of vinblastine, N(p-azido-[3 -lz~I]salicyl)-N’-(~aminoethyl)vindesine ([‘251]NASV) (Fig* 1) to label membrane proteins in drug-resistant and drugThis reagent sensitive cells. specifically labels P-glycoprotein in membranes from drug-resistant cells13. Labeling is inhibited by vinblastine, vincristine and to some extent by daunorubicin, but not by colchicine or actinomycin D. Agents which reverse drz;g resist-

TIPS - Februa y 1988 [Vol. 91


Fig. 2. Model of the human P-glycoprotein based on its amino acid sequence. (Reproducedwith permissionfrom Ref. 35.)

ante, including

the Ca*+ channel blockers verapamil and diltiazem, blockers and non-Ca + channel such as quinidine and reserpine, are also effective inhibitors of [1251]NASV labeling. There is no correlation between ability to block Ca*+ channels and ability to reverse drug resistance or inhibit [1251]NASVlabeling. Several of the reversing agents bind to membranes from multidrug-resistant cells, but not to membranes from sensitive cells14, suggesting that they block [‘*‘I]NASV labeling by binding to P-glycoprotein. The Ca*+ channel blocker azidopine, which reverses drug resistance, has recently been shown to bind directly to P-glycoprotein”. These results show that P-glycoprotein binds both drugs and agents which reverse drug resistance, and hence is a good candidate for the energy-dependent &=lLlX pump in multidrug-resistant cell lines. Cloning and characterization of the human MDRl (P-glycopiotein) gene To develop molecular probes for human studies, it was necessary

to clone the gene responsible for the multidrug resistance phenotype from th;B hucT;; multidrug-resistant Karyotypic analysis of chromosomes from highly multidrugresistant cells showed the presence of multiple minute and double minute chromosomes, as would be expected if a gene or genes were amplified in these resistant cell lines16. In collaboration with I. Roninson at the University of Illinois, who had used a novel in-gel renaturation technique to isolate an amplified genomic fragment from multidrug-resistant hamster cell linesl’, homoiogous human amplified DNA segments were isolated from multidrug-resistant (MDR) cell lines18. An amplified DNA fragment derived from a colchicineselected human KB cell line (pMDRl)*, was found to hybridize to an mRNA of 4.5 kb whose levels correlated with drug resistance in all human and rodent multidrug-resistant cell lines tested”. In one series of human *MDRl is the human gene responsible for multidrug resistunce; mdr is the non-human counkrpart of the MDRl gene.

cells, an multidrug-resistant increase in levels of this MDRl mRNA preceded amplification of the MDRl gene, demonstrating that gene amplification was not the only means by which cell lines could become multidrugresistant”. Based on increased expression of RN& in multiclrugresistant hamster and mouse cell lines, similar mdr probes have also been is&ted in other laboratories20*21. There appears to be a small family of m&-related genes, and at least one other linked gene, termed MDR2, has been found in multidrug-resistant KB cells”. Whether any of these other MDR genes has any role in multidrug resistance of human cells remains to be determined. Ueda et al. used the genomic pMDR1 probe to isolate an overlapping set of MDRl cDNAs from a library prepared from a human multidrug-resistant cell line=. These MDRl cDNAs crosshybridize with cDNAs for the P-glycoprotein gene23,24, indicating that the human MDRl gene encodes P-glycoprotein. Sequence analysis, carried out in collaboration with Roninson’s laboratory, showed that the MDRl cDNA encodes a 1280 amino acid protein with 12 transmembrane-spanning domains” (shown schematically in Fig. 2). similar sequences have been obtained from mouse26 and hamste?’ mdr or P-glycoprotein cDNAs. I”ne P-glycoprotein gene appears to have arisen from the tandem duplication of an ancestral gene. The protein contains two regions homologous to nucleotide binding domains fn.:m energytransducing subunits of multicomponent bacterial transport systems (i.e. hisP, ma/K, oppD,

Fig. 3. Cartoon showing how P-glycoproteinmust use the energy of ATP Io act as an efflux pump for mulfiple drugs. Upper pand: P-gWoProlein Vansportingdrug. Lowor panel: how verapamil might act to inhibit drug-binding and transport by inferacling dimMy with P-glycoprotein.

TIPS - February 1988 [Vol. 91 pstB and ~IYB)~~~‘. Biochemical analysis shows that, as predicted, P-glycoprotein binds ATP2a. The sequence analysis showing that Pglycoprotein has multiple transmembrane domains and energy transducing sites, taken together with localization studies demonstrating that P-glycoprotein is in the plasma membrane, and biochemical studies showing that Pglycoprotein binds drugs and ATP, indicates that P-glycoprotein is the energy-dependent drug efflux pump in multidrugresistant cell lines. A model incorporating these features is shown in Fig. 3, along with a modification to explain how agents such as verapamil, which reverses drug resistance, might work by inhibiting binding of toxic drugs to the P-glycoprotein eftlw pump. Cloned full-length cDNAs for the human MD1911 gene from a resistant cell line and a mouse mdr gene from a drug-sensitive cell line have been inserted into eukaryotic expression vectors29,30. These cloned genes can confer the complete multidrug resistance phenotype on recipient drugsensitive cells. This experiment proves that expression of Pglycoprotein at high levels is sufficient to confer drug resistance, in support of the hypothesis that Pglycoprotein is a drug efflux pump. It is possible that other cellular proteins and factors are needed for highly efficient expression of this system, but Pglycoprotein appears from these results to be the limiting protein in the multidrug resistance sys= tern. Definitive confirmation that P-glycoprotein is a drug efflux pump will require its purification to homogeneity and demonstration of its ability to transport drugs in membrane reconstitution experiments, Expression of the MDRI gene in normal human tissues and tumors Expression of the MDRl gene is regulated in human tissues, with some tissues (adrenal gland, kidney, liver, colon and small intesexpressing substantially tine) higher levels of MDRl RNA than other tissues31. In addition, during chemical hepatocarcinogenesis and following hepatectomy in the rat, mdr RNA levels increase dramatica!!y32. To study regulation of the MDRl gene, Ueda et n1.22 have mapped the sites of



fig. 4. immunological localization (red lines) of P-glycoproteh in normal &sues. (Reproduced with permission from Ref. 33.)

initiation of the MDRl transcript in drug-resistant cell lines and in some normal tissues. In colchicine-selected KB multidrugresistant cells, transcription of MDRl RNA initiates at both an upstream and a downstream promoter. In most other multidrug-resistant cells and in normal only the downstream tissues, promoter appears to be active. This downstream promoter sequence has ben isolated and studies on the factors which affect its expression are in progress. In normal tissues, the location of P-glycoprotein determined by immunohistochemical techniques is consistent with its role as a transport protein. In liver, it is found in the bile canalicular surface of hepatocytes; in kidney, it is on the lumenal surface of the brush border of proximal tubules; and in the intestine, Pglycoprotein is found on the mucosal surface of both jejunum and co!on33. In the human adrenal gland, P-glycoprotein is found diffusely distributed on the sur-

face of cells of the cortex and medulla. Although the function of P-glycoprotein in the human adrenal gland is uncertain, in the kidney, iiver and well situated to be an efflux pump for hydrophobic natural compounds either ingested in the diet (i.e. products of plants and microorganisms) or produced in the body. A schematic illustration of the immunocytochemical localization of P-glycoprotein is shown in Fig. 4. Not surprisingly, MDRl RNA expression is found in tumors derived from normal tissues which have elevated levels of Pi.e. cancers of the glycoproteir?‘, colon, adrenal gland, kidney and liver. We speculate that the intrinsic resistance of such tumors to chemotherapy might be related to this MDRI RNA expression. In a few instances, such as childhood lymphocytic leukemia, acute neuroblastoma and pheochromocytoma, increased expression of MDRl RNA has been seen in association with the emergence of boWe!p



TilJS - Februartj 1988

58 drug-resistant tumors3*. These tumors appear to have subverted a normal transport system as a defense against chemotherapeutic agents. However, many tumors with acquired resistance and some with intrinsic resistance do not show elevated levels of MDA1 RNA, implying that there may be by which other mechanisms tumors become resistant to multiple chemotherapeutic agents. Clinical implications of studies on multidrug resistance The information obtained from analysis of the molecular basis of multidrug resistance in human tissue culture cells may be used clinically in several ways. Patterns of drug resistance in human cancer are different from multidrug resistance in tissue culture cells and sometimes include resistance to additional agents such as nucleotide analogs, aliylating agents and antimetabolites. However, if MDR? RNA expression is shown to be responsible for clinical resistance to some drugs (such as Vinca alkaloids and doxorubicin), cloned cDNA probes for the MDXl gene could be used to help design rational chemotherapy protocols for treating individual cancers. Non-toxic drugs able to reverse the multidrug resistance phenotype might be useful in converting drug-resistant tumors to drug-responsive tumors. FitzGerald et al. have recently shown that an antibody against Pglycoprotein linked to Pseudomonas toxin will specifically kill multidrug-resistant Kl3cell lines34. The use of sUch immunotoxins to kill drug-resistant cells or prevent their emergence during chemotherapy is possible. Extensive clinical analysis will be needed to determine whether MDRl RNA expression is responsible for multidrug resistance in human cancer and whether reversing this resistance can help make tumors more susceptible to chemotherapy.

1 Riordan, J. R. and Ling, V. (1985) Pharmacol. Thu. 28, 51-75 2 Biedler. 1. C. and Peterson. H. F. 11981) in Mole&w

Aclions aud’ Targets for Cancer Ckemo~lrempcalicAgtwk (Sartor-

elli, A. C., Bertino, J. R. and Lazo, J. S., eds). pp. 453-482, Academica Press 3 Beck, W. T., Mueller, T. J. and Tanzer, L. R. (1979)Carrcer Res. 39, 207&2076

4 Akiyama, S. I., Fojo, A., Hanover, J. A., Pastan, I. and Gottesman, M. M. (1985) Somatic Cell Mol. Genet. 11, 117-126 5 Shen, P-W., Cardarelli, C., Hwang, J., Cornwell, M., Richert, N., Ishii, S., Pastan, I. and Got&man, M. M. (1986) 1. Biol. Chem. 261, 7762-7770 6 Shen, D-W., Fojo, A., Roninson, 1. B., Soffir, R., Pastan, I. and Gottesman, M. M. (1986)Mol. Cell. Biol. 6,4039-4045 7 Fojo, A. T., Akiyama, S. I., Gottesman, M. M. and Pastan, I. (1985) Cancer Res. 45,30023007 8 Willingham. M. C., Comwell, M. M., Cardarelli, C. 0.. Gottesman, M. M. and Pastan, I. (1986) Cancer Res. 46, 59415946 9 Tsuruo, T., Iida, H., Tsukagosti, S. and Sakurai, Y. (1981) Cancer Res. 41, 1967-

Gottesman, M.M.

20 21 22




11 Hamada, H. and Tsuruo, T. (1986) Proc. Nat1 Acad. Sci. USA 83,770$_7789


2X6-2170 15 Safa, A. R., Glover, C. J,, Sewell, J. L., Meyers, M. B., Biedler, J. L. and Felsted, R. L. (1987)J. Biol. Chem. 262,7884-7888 16 Fojo, A. T., Whang-Peng, J., Gottesman, M. M. and Pastan, I. (1985) Droc. Nat/ Acad. Sci. USA 82,7661-7665 17 Roninson, I. B., Abelson, H. T., Housman, D. E., Howell, N. and Varshavsky,

A. (1954) Nature 309. 626-628 18 Ronins& 1. B., Chin, J. E., Choi, K., Gras, I’., Housman, D. E., Fojo, A., Shen, D-W., Gottesman, M. M. and Pastan, I. (1986) Prw. Nat1 Acad. Sci. USA 83,453%4542 19 Shen, D-W., Foojo,A., Chin, J. E., Ronin-

son, 1. 8.. Richert,

N., Pastan,

1. and

(1986) Science 292, 643-645 Van der Bliek, A. M., Van der VeldeKoerts, T., Ling, V. and Borst, I’. (1986) Mol. Cell. Biol. 6, 1671-1678 Scotto, K. W., Biedler, J. L. and Melera, P. W. (1986) Science 232,751-755 Ueda, K., Clark, D. I’., Chen, C-J., Roninson, 1. B., Gottesman, M. M. and Pastan, I. (19871J. Biol. Chem. 262, 505508 Riordan, J. R., Deuchars, K., Kartner, N., Alon, N., Trent, J. and Ling, V. 119851 .-- -, Nature X6.817-819

24 Ueda, K., Comw&ll, M. M., Gottesman,

1972 10 Juliano, R. L. and Ling, V. (1976) Biochim. Biophys. Acta 455,152-162 12 Willinaham, M. C., Richert, N. D.: Comw~lI, M. M., Tsuruo, T., Hamada, H., Cottesman, M. M. and Pastan, I. (1987)J. Histochem. Cytochem. 35,14511456 13 Cornwell, M. M., Safa, A. R., Felsted, R. L., Gottesman, M. M. and Pastan, I. (1986) Proc. Natf Acad. Sci. USA 83, 3847-3850 14 Cornwell, M. M., Pastan, I. and Gottesman, M. M. (1987) J. Biol. Chem. 262,

I Vol.91



M. M., Pastan, I., Roninson, I. B., Ling, V. and Riordan, J. R. (1986) Eiochek Biophys. Res. Commun. 141,956-962 Chen, C-J., Chin, J. E,, Ueda, K., Clark, D. P., Pastan, I., Gottesman, M. M. and Roninson, I. (1986) Cell 47,381389 Gras, P., Croop, J, and Housman, D. (1986) Cell 371-380 Gerlach. 1. H.. Endicott. 1. A. and Juranka; P: F. (1486) Nat&J% 485-489 Cornwell, M. M., Tsuruo, T., Gottesman, M. M. and Pastan, I. (1987)FASEB J. 1,51-54 Gms, P., Ben Neriah, Y., Croop, J. M. and Housman, D. E. (1986) Nature 323, 728-731

30 Ueda, K., Cardarelli, C., Gottesman, M. M. and Pastan, I. (1987) Proc. Nat1 Acad. Sci. USA 84.3004-3008 31 Fojo, A. T., Ueda, K., Slamon, D. J.,

Poplack, D. G., Gottesman, M. M. and Pastan, I. (1987) Proc. Nat1 Acad. Sci.

USA 84,265-269 32 Thorgeirsson, S. S., Huber, B. E., Bori rell, S., Fojo, A., Pastan, I. and Gottesman, M. M. (1987) Science 236, 11201122 33 Thiebaut, F., Tsumo, T., Hamada, H., Gottesman, M. M., Pastan, 1. and WilIiilgham, M. C. (1987) Proc. Nat1 Acad. Sci. USA 84.7735-7738 34 F&Gerald, ‘D. J. P., Williqham, M. C..

Cardarelli, C. O., Tsuruo, T., Hamada, H., Gottesman, M. M. and Pastan, I. (1987) Proc. Nat/ Acad. Sci. USA 84,

42%==4292 35 Marx, J. (1986) Science 234, 818-820

Muscarinic Receptor Subtypes III Single copies of the supplement Muscarinic Receptor Subtypes frl, distributed with this month’s issue of BPS, may be purchased

for ~9.SOiUS$l?,SO from:

Carol Smith, Trends in Pharmacological Sciences, Elsevier Publications Cambridge, 66 Hills Road, Cambridge CB2 1LA, UK,or from Elsevier Science Publishing a., Journal Information Center, 52 Vanderbilt Avenue, New York NY 1OOlZ USA. Note:Payment must accompany order. You may pay by E cheque, f: Eurocheque or $ check drawn on a US bank. We also accept AmEx, Mastercard and Visa credit cards (please give your card number, expiry date, issuing bank cardholder’s name-, address and signature).