Multidrug resistance in cancer

BMJ 1994;308:148-149 (15 January)

Editorials

Multidrug resistance in cancer

Malignant neoplasms vary in their response to cytotoxic drugs:some are sensitive, others resistant. Understanding of the biochemicalbasis of this resistance might lead to the development of markersthat would correlate with the clinical response to the drugs- or even lead to ways of overcoming the resistance.

Multidrug resistance was first described in 1970 after selectionof Chinese hamster ovarian cancer cells exposed to increasingconcentrations of actinomycin D.¹ Though the cells had beenselected by a single agent, they proved to be resistant to arange of clinically important anticancer drugs, including theanthracyclines (doxorubicin and daunomycin), the vinca alkaloids(vincristine, vinblastine, and vindesine), etoposide, and colchicine.Riordan and Ling went on to show that the multidrug resistantcells had lower concentrations of the drug - a drug accumulationdeficit - and that a membrane glycoprotein of 170 kDa was responsible.²At first the deficit was thought to be due to a fault in permeationand the glycoprotein was named P glycoprotein (P-gp).³ The genefor this, mdr-1, has been cloned and sequenced, and the aminoacid has been sequenced and the amino acid sequence derived.⁴P glycoprotein has 1280 amino acids, and its structure suggeststhat it arose from the fusion of two closely related genes;it uses ATP as a source of energy. The precise molecular mechanismwhereby P glycoprotein can transport such a wide range of structurallydiverse drugs is uncertain, and several models have been proposed.*RF5-7*

P glycoprotein belongs to a superfamily of ATP binding cassettetransporters, whose members include the cystic fibrosis transmembraneregulator, the chloroquine transporter of Plasmodium falciparum,and a yeast transporter.⁸ Recently a new member of this familywas cloned from a human small cell lung cancer cell line, H69AR.Though it was resistant to doxorubicin and vincristine, it didnot express P glycoprotein.⁹ This protein has 1522 amino acidsand has been termed multidrug resistance associated protein(MRP). Cells with this type of resistance may not have accumulationdeficits of doxorubicin. One possible explanation is that multidrugresistance associated protein facilitates the sequestrationof cytotoxic drugs into intracellular organelles.⁹ Whether thisprotein binds to the drugs against which it confers resistanceis not known, but the emergence of a potential new target formodulation, if it is widely expressed in human tumours, is clearlyan exciting development.

Several probes have been developed that can detect the P glycoproteingene, its mRNA, and its protein product.^10,11 Studies have shownthat the gene expressed in normal gastrointestinal mucosal cells,renal tubular cells, biliary canalicular cells, and adrenocorticalcells.¹¹ Tumours that are initially sensitive to chemotherapybut resistant on relapse commonly show increases in expressionof P glycoprotein. Leukaemias with high concentrations of Pglycoprotein tend to be resistant to inductive treatment withchemotherapy based on anthracycline.¹² Expression of P glycoproteinis an adverse factor in multivariate prognostic models for childhoodsarcoma and neuroblastoma.^13,14 Before long it may be practicableto "tissue type" tumours according to expression of P glycoproteinand, perhaps, to avoid the toxicity and reduced quality of lifeassociated with treatment with ineffective anticancer drugs.

After an initial observation by Tsuruo et al,¹⁵ several groupshave shown that transport defect mediated by P glycoproteincan be blocked by many non-cytotoxic drugs, including nifedipine,verapamil, quinine, chloroquine, progestogens, tamoxifen, cyclosporinA and its analogues, reserpine and tricyclic anti-depressants.¹⁶Clinical trials have assessed the combination of conventionalchemotherapy with modulators of P glycoprotein; in one studyof relapsed non-Hodgkin's lymphoma 13 of 18 patients respondedto infusional chemotherapy with high doses of verapamil.¹⁷ Unfortunately,the amounts of verapamil needed to reverse drug resistance producedcongestive cardiac failure and heart block. Research groupsare currently looking for potent, non-toxic modulators of Pglycoprotein to combine with conventional chemotherapy. CyclosporinA has been reported to reverse clinical multidrug resistancein myeloma.¹⁸ The combination of quinidine with the antineoplasticagent epirubicin was recently assessed in a prospective, placebocontrolled randomised study in patients with advanced breastcancer.¹⁹ Tumour response rates and survival were similar inthe two arms of the trial, a finding that probably reflectsthe low potency of quinidine to bind P glycoprotein.⁷

Another possible application of the genetics of P glycoproteinis in gene therapy. Transgenic mice that express human P glycoproteinin their bone marrow are resistant to chemotherapy,²⁰ so intheory if human bone marrow could be transfected with the genethat might protect it against myelosuppression from anticancerchemotherapy.⁸

Novel, improved modulators of P glycoprotein seem likely tobe developed in parallel with mechanistic research on the functionof the gene, and thus should lead to the rational design ofinhibitor. Future phase II clinical trials with new modulatorsof P glycoprotein must include pharmacokinetic studies sinceinteractions can occur with cytotoxic drugs - for example, etoposide'sactivity is increased by 80% when it is given in combinationwith cyclosporin A.²¹ These trials should focus on those cancersthat are known to express high concentrations of P glycoprotein(including renal and colorectal cancers and sarcomas). SinceP glycoprotein is expressed in normal tissues there is someconcern that its modulators might alter the cellular transportof physiological metabolites such as bilirubin and thereby delaythe clearance of cytotoxic drugs that undergo hepatobiliaryexcretion. These clinical trials should therefore be performedin units used to handling the toxicity associated with intensivechemotherapy.