Multidrug Resistance: Clinical Relevance in Haematological Malignancies
S. B. Kaye, D. J. Kerr SUMMA R Y. Multidrug resistance describes an experimental observation which appears to explain cross-resistance to certain structurally unrelated cytotoxic agents, including anthracyclines, vinca alkaloids and podophyllotoxins. It is now clear that a major factor responsible for its development is increased expression of a membrane glycoprotein-P-glycoprotein, which functions as an energy-dependent efflux pump. Recent data, particularly in haematological malignancies such as acute non-lymphocytic leukaemia, myeloma and non-Hodgkin’s lymphoma, indicate that Pglycoprotein may be involved in the development of clinical drug resistance. The potential therefore exists for new therapeutic studies aimed at circumventing resistance which develops through this mechanism, by using modulators, such as verapamil, quinidine and several others, which prevent cellular drug efflux by competitive binding to P-glycoprotein.
Although the past 20 years have seen enormous advances in the management of haematological malignancies with chemotherapy, certain obstacles remain. Among the most important is the development of resistance to the agents used for treatment, whereby an initial response is eventually followed by a relapse which proves to be fatal. Acute non-lymphocytic leukaemia provides an excellent illustration of this phenomenon, which is all too familiar to oncologists dealing with solid tumours such as small cell lung cancer, breast cancer and ovarian cancer. In recent years, scientists in several laboratories have pieced together experimental data from drug resistant tumour models which have given grounds for fresh optimism to clinicians grappling with resistance, in haematological as well as other malignancies. These data refer to the phenomenon known as ‘multidrug resistance’, which describes experimental cross-resistance to a group of structurally unrelated
S. B. Kaye, D. J. Kerr, CRC Department of Medical Oncology, University of Glasgow, Alexander Stone Building, Garscube Estate, Switchback Road, Glasgow G61 IBD, UK. Blood Reviews (1991) 5.038-041 0 1991 Longman Group UK Ltd
agents linked by their common origin as natural products, from plants or micro-organisms.’ The study of these multidrug resistant cell lines, (lines selected for simultaneous resistance to vinca alkaloids, anthracyclines, actinomycin D, colchicine, etoposide and other related natural products) has demonstrated that resistant cell lines have decreased intracellular drug accumulation because the cells express a membrane bound 170-kilodalton glycoprotein (P-glycoprotein), which is thought to function as an energy dependent drug efflux pump’ (see Fig. 1). Moreover this pump may be inhibited by simultaneous exposure to membrane active non-cytotoxic agents. Interestingly these observations were first made in cell lines derived from human haemopoietic tumours, and one of the most noteworthy early observations was the remarkable structural similarity between P-glycoprotein and bacterial transport proteins, particularly those transporting haemolysins.3 Recent studies have also indicated that P-glycoprotein is in fact a component of the cell membrane in certain normal tissues, including the gastrointestinal tract and renal tubules. Although its function is unclear, its disposition on epithelial surfaceq4 indi-
BLOOD REVIEWS
COOH
I
HiP Fig.1 Model of P-glycoprotein, regions and 2 intracytoplasmic
I
NiP
showing 12 transmembrane nucleotide-binding sites.
cates that it probably has a protective and/or secretory role. A unifying hypothesis is that P-glycoprotein provides normal tissue protection against environmental toxins, and resistant tumour cells may hyperexpress this property when facing certain cytotoxic agents which are natural toxic products. There are two general areas of applied clinical research which have arisen following demonstration of the mechanistic relevance of P-glycoprotein: 1. Can the P-glycoprotein be identified in human tumour cells as well as normal tissues, and if so, does the presence of multidrug resistant tumour cells, expressing P-glycoprotein, alter or predict the patient’s response to chemotherapy?; 2. Can the agents which have been used in vitro to overcome the multidrug resistant phenotype, such as verapamil, quinidine and cyclosporin, be used in the clinical situation to modulate this drug resistant pathway with clinical benefit? Detection of P-Glycoprotein and Relationship to Therapy Various techniques are available to detect the presence of P-gycoprotein. Firstly, immunocytochemical studies (using anti-P-glycoprotein monoclonal antibodies, MRK-16 and C219) may be used. These have the advantage of providing information on precise cellular localisation of P-glycoprotein in a mixed cell population, but problems of cross-reactivity with other epitopes remain to be resolved.5 Semi quantitative analysis of P-glycoprotein can also be made using antibodies and flow cytometry.‘j Secondly, assays for the mRNA coding for the protein have been widely used, since a full length cDNA for the multidrug resistance (mdr-1) gene has been sequenced and is available as a probe for Northern blots, slot blots and the polymerase chain reaction. These methods have the advantage of specificity, but the disadvantage of failure to localise gene expression to particular cells within a mixed population of tumour cells and stromal cells. It is possible to detect the P-glycoprotein mRNA transcript using in situ hybridisation techniques, but these are laborious and rather difficult to perform. Using either of these diagnostic approaches, there are at least 6 reports of apparent correlations between the presence of P-glycoprotein hyperexpression in haematological malignancies and the timing of therapy, but only 1 of these documents serial changes in an individual patient.
39
Salmon et al, 19897 used an immunocytochemical method to demonstrate tumour cellular P-glycoprotein staining on bone marrow aspirates from myeloma patients and lymph node biopsies from patients with lymphoma. A correlation was then made with the sensitivity of the tumour cells to doxorubicin, assessed by clonogenic assay. The results were as follows: Number of patients
Number sensitive to doxorubicin
Number resistant to doxorubicin
Negative Myeloma Lymphoma
3 3
3 2
0
Positive Myeloma Lymphoma
4 3
0 0
4 3
P-glycoprotein staining
1
This preliminary work suggests that tumour cells staining positive for P-glycoprotein tend to be resistant to doxorubicin. Goldstein et al 19898 measured levels of mdr-1 mRNA in a range of tumour types (including acute lymphocytic leukaemia (ALL), acute non-lymphocytic leukaemia (ANLL), low grade non-Hodgkin’s lymphoma and chronic myeloid leukaemia (CML) in blast crisis) by a slot-blot method using mRNA levels in a multidrug resistant cell line as control. These workers found that mdr-1 mRNA levels were significantly elevated in 3 out of 3 cases of CML in blast crisis, but were not detectable in 3 patients in the chronic phase of CML (all patients untreated). mdr- 1 RNA levels were occasionally elevated in untreated cancers including ALL (13%) and ANLL (13%) in adults. mdr-1 mRNA levels were commonly elevated in ALL, ANLL and low grade non-Hodgkin’s lymphoma at relapse after initial therapy with doxorubitin containing regimens. Similarly, Brophy et al 1989’ found increased levels of mdr- 1 mRNA (using slot blot analysis) in cell lines derived from 9 out of 11 patients (81%) with previously treated lymphoma and ANLL, compared to 7 out of 14 (50%) previously untreated cases. Priesler et al 1989,” using Northern blot analysis, established a correlation between mdr gene expression and duration of remission in 36 newly diagnosed patients with acute non-lymphocytic leukaemia. Of 9 cases expressing the gene, 5 (65%) entered CR, with a median duration of 90 days. For the remaining nonexpressers, the CR rate was 76%, with a median duration of over 400 days. A correlation was also seen between mdr gene expression and response duration in 23 patients with relapsed disease. The greater the degree of P-glycoprotein expression, the shorter the duration of response. Sehested, Friche and Ersboll, 1989l’ from the Finsen Institute, Copenhagen, took bone marrow
40 MULTIDRUG RESISTANCE aspirates in 34 cases from patients with ANLL after the first cycle of chemotherapy, at which point it was possible to make a clinical assessment of drug sensitivity or resistance on the basis of the blast response to the first course of treatment. The presence of Pglycoprotein in this study was assessed immunocytochemically (using the C219 antibody), and a strong correlation between the presence of P-glycoprotein and clinical response was apparent. Of the 16 clinically resistant cases (defined by blast count reduction), 12 (75%) had detectable P-glycoprotein in marrow aspirate, compared to 3 of 18 clinically sensitive cases. Finally, Ma et al 198712 have shown, on sequential sampling from 2 patients with acute non-lymphocytic leukaemia, that the emergence of a population of Pglycoprotein-positive cells appears to correlate strongly with treatment failure, albeit at rather a late stage in disease progression. In some respects sequential data of this kind, implying the selection of a specific population of resistant cells, are the most persuasive, and further similar studies are clearly warranted. Taken together, these data imply tht mdr-1-mRNA, or P-glycoprotein itself, are commonly elevated in a range of haematological malignancies on relapse. In addition, there are some preliminary in vitro drug sensitivity data which suggest that detection of Pglycoprotein may predict for cytotoxic drug resistance. Reversal of multidrug resistance-a
clinical reality?
In 198 1, Tsuruo and co-workers first reported that the calcium channel blocker verapamil could overcome vincristine resistance by inhibiting active drug efflux. ’ 3 More recently the same group have confirmed that verapamil acts by direct competitive binding with P-glycoprotein.i4 In the interim a large body of evidence has accumulated to show that a range of other non-cytotoxic modulating agents serve as inhibitors of the process of multidrug resistance, presumably by direct binding at the same or related binding sites on P-glycoprotein, thereby increasing intracellular concentrations of the antineoplastic agents. But can these experimental observations using modulators be translated into clinical efficacy? Over the past few years a small number of clinical trials have been conducted in a number of tumour types, but without data on Pglycoprotein expression these trials have been of limited value. However, Dalton et al 1989i5 detected P-glycoprotein by immunohistochemical staining on tumour cells from 6 of 8 patients with clinically drug resistant disease (7 patients with multiple myeloma and 1 patient with intermediate grade non-Hodgkin’s lymphoma). All patients developed progressive disease while receiving regimens containing vincristine and doxorubicin. A continuous infusion of verapamil was added to therapy when progression was noted and 3 of the 8 patients responded (albeit
transiently) to verapamil plus vincristine/doxorubitin. The authors concluded that P-glycoprotein expression occurs in drug-refractory B cell neoplasms and that co-treatment with verapamil, a known inhibitor of the multidrug efflux pump, may partially circumvent drug resistance clinically. Similar clinical results in a small number of patients have been reported by Gore et al 1988.i6 A total of 11 patients with drug-resistant myeloma were retreated with a combination of vincristine, adriamycin and methyl-prednisolone, together with verapamil at the dose of only 1Omg daily by continuous infusion for 5 days. Clinical evidence of response was seen in 6, including a complete response in 1. These data are provocative because the plasma levels of verapamil achieved are probably lower than those which are most effective at causing modulation in vitro, and this raises the question of whether the clinical results seen are due to an effect of verapamil on immunoglobulin secretion, rather than on Pglycoprotein-mediated drug transport. A further possibility is that verapamil may interact with hepatic mono-oxygenase enzymes and alter the pharmacokinetics of doxorubicin, increasing the area under the plasma concentration-time curve by delaying plasma clearance. In a clinical study in Glasgow in patients with small cell lung cancer, such an effect was observed in 6 patients, who received chemotherapy in successive cycles with and without verapamil.” In a subsequent largescale randomised trial, comparing conventional chemotherapy with chemotherapy plus verapamil, a trend towards an increased complete response rate (44% versus 28%) was observed, together with an increased degree of myelosuppression, in those patients receiving verapamil.” It is conceivable that the effects seen related more to the pharmacokinetic interaction than to an impact on cellular drug efflux. The concentrations of verapamil achieved in this lung cancer study (approximately 1 yM) were lower than the most effective in vitro concentrations (6 uM), although it is noteworthy that the major metabolite, norverapamil, is present in equimolar concentrations and has comparable modulating activity in vitro.” Verapamil is a racemic mixture of D- and Lisomers, and clinical data indicate that D-verapamil has only a minor role in respect of the negative dromotropic effects on the heart.20 It may therefore be possible to use higher doses of D-verapamil than of the racemic mixture for the purpose of resistance modulation. This may improve the clinical potential for verapamil since our recent in vitro data confirm that D-verapamil has similar efficacy to the racemic mixture in terms of resistance modulation.21 Nevertheless, alternative modulators may be more effective, among which is quinidine, which has been shown to be active at clinically achievable concentrations (approximately 6 ~.LM).~~ In a recent pilot study in patients with breast cancer, we have confirmed the feasibility of combining quinidine (250mg bd) with
BLOOD REVIEWS
single-agent epirubicin ( 100mg/m2). No increase in acute or delayed toxicity was noted, and quinidine was generally very well tolerated, at concentrations which may have a biological impact on anthracycline resistance.23 This has allowed us to embark on a placebo-controlled trial in patients with breast cancer, which represents another tumour type in which increased levels of mdr-1 mRNA are frequently observed.24 Other modulators are also available, including analogues of cyclosporin,25 newer cephalosporins26 and the new antioestrogen, toremifene.27 These all carry the advantage over verapamil in that clinically achievable concentrations are active in in vitro models, and clinical trials with these and other agents appear justified. Overall, these preclinical data and early clinical studies in solid tumours may provide guidelines for similar studies in certain haematological malignancies. The data previously quoted on P-glycoprotein expression in these cancers suggest that such an approach may prove particularly fruitful. Acknowledgements The authors would like to thank the Cancer Research Campaign for financial support and Mrs Margaret Jenkins for typing the manuscript. References 1. Kaye S B 1988 The multidrug resistance phenotype. British Journal of Cancer 58: 691-94 2. Kartner N, Riordan J R, Ling V 1983 Cell surface Pglycoprotein is associated with multidrug resistance in mammalian cell lines. Science 221: 1285-88 3. Gerlach J H, Endicot A, Juranka P F et al 1986 Homology between P-glycoprotein and a bacterial haemolysin transport protein suggests a model for multidrug resistance. Nature 324: 485-89 4. Fojo A T, Ueda K, Slamon D J, Poplack D G, Gottesman M M, Pastan I 1987 Expression of a multidrug resistance gene in human tumours and tissues. Proceedings of the National Academy of Sciences, USA 84: 265-69 5. Thiebaut F. Tsuruo T. Hamada H, Gottesman M M, Pastan I Willingham M C 1989 Immunohistochemical localisation in normal tissues of different epitopes in the multidrug transport protein P170: Evidence for localisation in brain capillaries and cross-reactivity of one antibody with a muscle protein. Journal of Histochemistry and Cytochemistry 37: 159-164 6. Miller L P, Boner K, Ronnson I, Tsuruo T 1989 Detection of P-glycoprotein by flow cytometry in paediatric acute nonlymphoblastic leukaemia. Proceedings of the American Society of Clinical Oncology 8: 218 _ 7. Salmon S E. Groean T M. Miller T. Schemer R. Dalton W S 1989 Prediction 0’: doxorubicin resistance ‘in vitro in myeloma, lymphoma, and breast cancer by P-glycoprotein staining. Journal of the National Cancer Institute 81: 696-701 8. Goldstein L J, Galski H, Fojo A et al 1989 Expression of multidrug resistance gene in human cancers. Journal of the National Cancer Institute 81: 116-24
41
9. Brophy N A, Marie J-P, Berry J M et aI 1989 Expression of the mdrl gene in leukaemia and lymphoma cells. Proceedings of the American Society for Clinical Oncology 8: 55 10. Preisler H, Gottesman M, Raza A, Pastan I, Day R, Buffalo N Y 1989 The clinical sianificance of exnression of the MDR gene in acute non-lymphicytic leukaemi’a. Proceedings of the American Society of Clinical Oncology 8: 201 11. Sehested M, Friche E, Ersboll J 1989 Detection of multidrug resistance (MDR) associated glycoprotein P- 170 in human haematological malignancies. Investigational New Drugs 7: 425 12. Ma D, Davey R, Harman D et al 1987 Detection of a multidrug resistant phenotype in acute non-lymphoblastic leukaemia. Lancet 1: 135-37 13. Tsuruo T, Iida H, Tsukagoslin S, Sakwai Y 1981 Overcoming of vincristine resistance in P388 leukaemia through enhanced cytotoxicity by verapamil. Cancer Research 41: 1967-72 14. Yusa K, Tsuruo T 1989 Reversal mechanism of multidrug resistance by verapamil: direct binding of verapamil to Pglycoprotein on specific sites, and transport of verapamil outward across the plasma membrane of K562/ADM cells. Cancer Research 49: 5002206 15. Dalton W S, Grogan T M, Meltzer P S et al 1989 Drugresistance in multiple myeloma and non-Hodgkin’s lymphoma: detection of P-glycoprotein and potential circumvention by addition of verapamil to chemotherapy. Journal of Clinical Oncology 7: 415-24 16. Gore M, Selby P, Millar B et al 1988 The use of verapamil to overcome drug resistance in myeloma. Proceedings of the American Society of Clinical Oncology 7: 228 17. Kerr D J, Graham J. Cummings J et al 1986 The effect of verapamil on the pharmacokinetics of adriamycin. Cancer Chemotherapy and Pharmacology 1% 239-42 18. Milroy R, Connery L, Banham S et al 1988 A randomised trial of verapamil in addition to chemotherapy in small cell lung cancer. Lung Cancer 4: supp. A 101 19 Merry S. Flanigan P, Schlick E, Freshney R I, Kaye S B 1989 Inherent adriamycin resistance in a murine tumour line: circumvention with verapamil and norverapamil. British Journal of Cancer 59: 895-97 20 Echizen H. Brecht T, Niedergan S, Vogelgesang B. Eichelbaum M 1985 The effect of dextro-levo and racemic verapamil on A-V conduction in humans. American Heart Journal 109: 210-17 21. Plumb J A, Milroy R, Kaye S B 1990 The activity of verapamil as a resistance modifier in vitro in drug resistant human tumour cell lines is not stereospecific. Biochemical Pharmacology 39: 787-92 22. Tsuruo T. Lida H, Kitatani Y 1984 Effects of quinidine and related compounds on cytotoxicity and cellular accumulation of vincristine and adriamycin in drug-resistant tumour cells. Cancer Research 44: 4303-07 23 Jones R D, Rankin E M, Habeshaw T et al 1989 A pilot study of epirubicin and quinidine in advanced breast cancer. British Journal of Cancer 60: 451 24. Brown R, Keith N. Stallard S, Kaye S B 1989 Expression of mdrl and gst-pi in breast turnours: Correlations with chemoresponsiveness in vitro. Proceedings of the American Association of Cancer Research 30: 517 25. Twentman P R, Fox N E, White D J G 1987 Cyclosporin A and its analogues as modifiers of Adriamycin and vincristine resistance in a multidrug resistant human lung cancer cell line. British Journal of Cancer 56: 55-57 26. Gosland M P, Lum B L, Sikie B I 1989 Reversal by cefoperazone of resistance to etoposide, doxorubicin and vinblastine in multi-drug resistant human sarcoma cells. Cancer Research 49: 690 l-6905 27. DeGregorio M W, Ford J M, Benz C C, Wiebe V J 1989 Toremifene: Pharmacologic and phannacokinetic basis of reversing multidrug resistance. Journal of Clinical Oncology 7: 1359-64
S. B. Kaye, D. J. Kerr SUMMA R Y. Multidrug resistance describes an experimental observation which appears to explain cross-resistance to certain structurally unrelated cytotoxic agents, including anthracyclines, vinca alkaloids and podophyllotoxins. It is now clear that a major factor responsible for its development is increased expression of a membrane glycoprotein-P-glycoprotein, which functions as an energy-dependent efflux pump. Recent data, particularly in haematological malignancies such as acute non-lymphocytic leukaemia, myeloma and non-Hodgkin’s lymphoma, indicate that Pglycoprotein may be involved in the development of clinical drug resistance. The potential therefore exists for new therapeutic studies aimed at circumventing resistance which develops through this mechanism, by using modulators, such as verapamil, quinidine and several others, which prevent cellular drug efflux by competitive binding to P-glycoprotein.
Although the past 20 years have seen enormous advances in the management of haematological malignancies with chemotherapy, certain obstacles remain. Among the most important is the development of resistance to the agents used for treatment, whereby an initial response is eventually followed by a relapse which proves to be fatal. Acute non-lymphocytic leukaemia provides an excellent illustration of this phenomenon, which is all too familiar to oncologists dealing with solid tumours such as small cell lung cancer, breast cancer and ovarian cancer. In recent years, scientists in several laboratories have pieced together experimental data from drug resistant tumour models which have given grounds for fresh optimism to clinicians grappling with resistance, in haematological as well as other malignancies. These data refer to the phenomenon known as ‘multidrug resistance’, which describes experimental cross-resistance to a group of structurally unrelated
S. B. Kaye, D. J. Kerr, CRC Department of Medical Oncology, University of Glasgow, Alexander Stone Building, Garscube Estate, Switchback Road, Glasgow G61 IBD, UK. Blood Reviews (1991) 5.038-041 0 1991 Longman Group UK Ltd
agents linked by their common origin as natural products, from plants or micro-organisms.’ The study of these multidrug resistant cell lines, (lines selected for simultaneous resistance to vinca alkaloids, anthracyclines, actinomycin D, colchicine, etoposide and other related natural products) has demonstrated that resistant cell lines have decreased intracellular drug accumulation because the cells express a membrane bound 170-kilodalton glycoprotein (P-glycoprotein), which is thought to function as an energy dependent drug efflux pump’ (see Fig. 1). Moreover this pump may be inhibited by simultaneous exposure to membrane active non-cytotoxic agents. Interestingly these observations were first made in cell lines derived from human haemopoietic tumours, and one of the most noteworthy early observations was the remarkable structural similarity between P-glycoprotein and bacterial transport proteins, particularly those transporting haemolysins.3 Recent studies have also indicated that P-glycoprotein is in fact a component of the cell membrane in certain normal tissues, including the gastrointestinal tract and renal tubules. Although its function is unclear, its disposition on epithelial surfaceq4 indi-
BLOOD REVIEWS
COOH
I
HiP Fig.1 Model of P-glycoprotein, regions and 2 intracytoplasmic
I
NiP
showing 12 transmembrane nucleotide-binding sites.
cates that it probably has a protective and/or secretory role. A unifying hypothesis is that P-glycoprotein provides normal tissue protection against environmental toxins, and resistant tumour cells may hyperexpress this property when facing certain cytotoxic agents which are natural toxic products. There are two general areas of applied clinical research which have arisen following demonstration of the mechanistic relevance of P-glycoprotein: 1. Can the P-glycoprotein be identified in human tumour cells as well as normal tissues, and if so, does the presence of multidrug resistant tumour cells, expressing P-glycoprotein, alter or predict the patient’s response to chemotherapy?; 2. Can the agents which have been used in vitro to overcome the multidrug resistant phenotype, such as verapamil, quinidine and cyclosporin, be used in the clinical situation to modulate this drug resistant pathway with clinical benefit? Detection of P-Glycoprotein and Relationship to Therapy Various techniques are available to detect the presence of P-gycoprotein. Firstly, immunocytochemical studies (using anti-P-glycoprotein monoclonal antibodies, MRK-16 and C219) may be used. These have the advantage of providing information on precise cellular localisation of P-glycoprotein in a mixed cell population, but problems of cross-reactivity with other epitopes remain to be resolved.5 Semi quantitative analysis of P-glycoprotein can also be made using antibodies and flow cytometry.‘j Secondly, assays for the mRNA coding for the protein have been widely used, since a full length cDNA for the multidrug resistance (mdr-1) gene has been sequenced and is available as a probe for Northern blots, slot blots and the polymerase chain reaction. These methods have the advantage of specificity, but the disadvantage of failure to localise gene expression to particular cells within a mixed population of tumour cells and stromal cells. It is possible to detect the P-glycoprotein mRNA transcript using in situ hybridisation techniques, but these are laborious and rather difficult to perform. Using either of these diagnostic approaches, there are at least 6 reports of apparent correlations between the presence of P-glycoprotein hyperexpression in haematological malignancies and the timing of therapy, but only 1 of these documents serial changes in an individual patient.
39
Salmon et al, 19897 used an immunocytochemical method to demonstrate tumour cellular P-glycoprotein staining on bone marrow aspirates from myeloma patients and lymph node biopsies from patients with lymphoma. A correlation was then made with the sensitivity of the tumour cells to doxorubicin, assessed by clonogenic assay. The results were as follows: Number of patients
Number sensitive to doxorubicin
Number resistant to doxorubicin
Negative Myeloma Lymphoma
3 3
3 2
0
Positive Myeloma Lymphoma
4 3
0 0
4 3
P-glycoprotein staining
1
This preliminary work suggests that tumour cells staining positive for P-glycoprotein tend to be resistant to doxorubicin. Goldstein et al 19898 measured levels of mdr-1 mRNA in a range of tumour types (including acute lymphocytic leukaemia (ALL), acute non-lymphocytic leukaemia (ANLL), low grade non-Hodgkin’s lymphoma and chronic myeloid leukaemia (CML) in blast crisis) by a slot-blot method using mRNA levels in a multidrug resistant cell line as control. These workers found that mdr-1 mRNA levels were significantly elevated in 3 out of 3 cases of CML in blast crisis, but were not detectable in 3 patients in the chronic phase of CML (all patients untreated). mdr- 1 RNA levels were occasionally elevated in untreated cancers including ALL (13%) and ANLL (13%) in adults. mdr-1 mRNA levels were commonly elevated in ALL, ANLL and low grade non-Hodgkin’s lymphoma at relapse after initial therapy with doxorubitin containing regimens. Similarly, Brophy et al 1989’ found increased levels of mdr- 1 mRNA (using slot blot analysis) in cell lines derived from 9 out of 11 patients (81%) with previously treated lymphoma and ANLL, compared to 7 out of 14 (50%) previously untreated cases. Priesler et al 1989,” using Northern blot analysis, established a correlation between mdr gene expression and duration of remission in 36 newly diagnosed patients with acute non-lymphocytic leukaemia. Of 9 cases expressing the gene, 5 (65%) entered CR, with a median duration of 90 days. For the remaining nonexpressers, the CR rate was 76%, with a median duration of over 400 days. A correlation was also seen between mdr gene expression and response duration in 23 patients with relapsed disease. The greater the degree of P-glycoprotein expression, the shorter the duration of response. Sehested, Friche and Ersboll, 1989l’ from the Finsen Institute, Copenhagen, took bone marrow
40 MULTIDRUG RESISTANCE aspirates in 34 cases from patients with ANLL after the first cycle of chemotherapy, at which point it was possible to make a clinical assessment of drug sensitivity or resistance on the basis of the blast response to the first course of treatment. The presence of Pglycoprotein in this study was assessed immunocytochemically (using the C219 antibody), and a strong correlation between the presence of P-glycoprotein and clinical response was apparent. Of the 16 clinically resistant cases (defined by blast count reduction), 12 (75%) had detectable P-glycoprotein in marrow aspirate, compared to 3 of 18 clinically sensitive cases. Finally, Ma et al 198712 have shown, on sequential sampling from 2 patients with acute non-lymphocytic leukaemia, that the emergence of a population of Pglycoprotein-positive cells appears to correlate strongly with treatment failure, albeit at rather a late stage in disease progression. In some respects sequential data of this kind, implying the selection of a specific population of resistant cells, are the most persuasive, and further similar studies are clearly warranted. Taken together, these data imply tht mdr-1-mRNA, or P-glycoprotein itself, are commonly elevated in a range of haematological malignancies on relapse. In addition, there are some preliminary in vitro drug sensitivity data which suggest that detection of Pglycoprotein may predict for cytotoxic drug resistance. Reversal of multidrug resistance-a
clinical reality?
In 198 1, Tsuruo and co-workers first reported that the calcium channel blocker verapamil could overcome vincristine resistance by inhibiting active drug efflux. ’ 3 More recently the same group have confirmed that verapamil acts by direct competitive binding with P-glycoprotein.i4 In the interim a large body of evidence has accumulated to show that a range of other non-cytotoxic modulating agents serve as inhibitors of the process of multidrug resistance, presumably by direct binding at the same or related binding sites on P-glycoprotein, thereby increasing intracellular concentrations of the antineoplastic agents. But can these experimental observations using modulators be translated into clinical efficacy? Over the past few years a small number of clinical trials have been conducted in a number of tumour types, but without data on Pglycoprotein expression these trials have been of limited value. However, Dalton et al 1989i5 detected P-glycoprotein by immunohistochemical staining on tumour cells from 6 of 8 patients with clinically drug resistant disease (7 patients with multiple myeloma and 1 patient with intermediate grade non-Hodgkin’s lymphoma). All patients developed progressive disease while receiving regimens containing vincristine and doxorubicin. A continuous infusion of verapamil was added to therapy when progression was noted and 3 of the 8 patients responded (albeit
transiently) to verapamil plus vincristine/doxorubitin. The authors concluded that P-glycoprotein expression occurs in drug-refractory B cell neoplasms and that co-treatment with verapamil, a known inhibitor of the multidrug efflux pump, may partially circumvent drug resistance clinically. Similar clinical results in a small number of patients have been reported by Gore et al 1988.i6 A total of 11 patients with drug-resistant myeloma were retreated with a combination of vincristine, adriamycin and methyl-prednisolone, together with verapamil at the dose of only 1Omg daily by continuous infusion for 5 days. Clinical evidence of response was seen in 6, including a complete response in 1. These data are provocative because the plasma levels of verapamil achieved are probably lower than those which are most effective at causing modulation in vitro, and this raises the question of whether the clinical results seen are due to an effect of verapamil on immunoglobulin secretion, rather than on Pglycoprotein-mediated drug transport. A further possibility is that verapamil may interact with hepatic mono-oxygenase enzymes and alter the pharmacokinetics of doxorubicin, increasing the area under the plasma concentration-time curve by delaying plasma clearance. In a clinical study in Glasgow in patients with small cell lung cancer, such an effect was observed in 6 patients, who received chemotherapy in successive cycles with and without verapamil.” In a subsequent largescale randomised trial, comparing conventional chemotherapy with chemotherapy plus verapamil, a trend towards an increased complete response rate (44% versus 28%) was observed, together with an increased degree of myelosuppression, in those patients receiving verapamil.” It is conceivable that the effects seen related more to the pharmacokinetic interaction than to an impact on cellular drug efflux. The concentrations of verapamil achieved in this lung cancer study (approximately 1 yM) were lower than the most effective in vitro concentrations (6 uM), although it is noteworthy that the major metabolite, norverapamil, is present in equimolar concentrations and has comparable modulating activity in vitro.” Verapamil is a racemic mixture of D- and Lisomers, and clinical data indicate that D-verapamil has only a minor role in respect of the negative dromotropic effects on the heart.20 It may therefore be possible to use higher doses of D-verapamil than of the racemic mixture for the purpose of resistance modulation. This may improve the clinical potential for verapamil since our recent in vitro data confirm that D-verapamil has similar efficacy to the racemic mixture in terms of resistance modulation.21 Nevertheless, alternative modulators may be more effective, among which is quinidine, which has been shown to be active at clinically achievable concentrations (approximately 6 ~.LM).~~ In a recent pilot study in patients with breast cancer, we have confirmed the feasibility of combining quinidine (250mg bd) with
BLOOD REVIEWS
single-agent epirubicin ( 100mg/m2). No increase in acute or delayed toxicity was noted, and quinidine was generally very well tolerated, at concentrations which may have a biological impact on anthracycline resistance.23 This has allowed us to embark on a placebo-controlled trial in patients with breast cancer, which represents another tumour type in which increased levels of mdr-1 mRNA are frequently observed.24 Other modulators are also available, including analogues of cyclosporin,25 newer cephalosporins26 and the new antioestrogen, toremifene.27 These all carry the advantage over verapamil in that clinically achievable concentrations are active in in vitro models, and clinical trials with these and other agents appear justified. Overall, these preclinical data and early clinical studies in solid tumours may provide guidelines for similar studies in certain haematological malignancies. The data previously quoted on P-glycoprotein expression in these cancers suggest that such an approach may prove particularly fruitful. Acknowledgements The authors would like to thank the Cancer Research Campaign for financial support and Mrs Margaret Jenkins for typing the manuscript. References 1. Kaye S B 1988 The multidrug resistance phenotype. British Journal of Cancer 58: 691-94 2. Kartner N, Riordan J R, Ling V 1983 Cell surface Pglycoprotein is associated with multidrug resistance in mammalian cell lines. Science 221: 1285-88 3. Gerlach J H, Endicot A, Juranka P F et al 1986 Homology between P-glycoprotein and a bacterial haemolysin transport protein suggests a model for multidrug resistance. Nature 324: 485-89 4. Fojo A T, Ueda K, Slamon D J, Poplack D G, Gottesman M M, Pastan I 1987 Expression of a multidrug resistance gene in human tumours and tissues. Proceedings of the National Academy of Sciences, USA 84: 265-69 5. Thiebaut F. Tsuruo T. Hamada H, Gottesman M M, Pastan I Willingham M C 1989 Immunohistochemical localisation in normal tissues of different epitopes in the multidrug transport protein P170: Evidence for localisation in brain capillaries and cross-reactivity of one antibody with a muscle protein. Journal of Histochemistry and Cytochemistry 37: 159-164 6. Miller L P, Boner K, Ronnson I, Tsuruo T 1989 Detection of P-glycoprotein by flow cytometry in paediatric acute nonlymphoblastic leukaemia. Proceedings of the American Society of Clinical Oncology 8: 218 _ 7. Salmon S E. Groean T M. Miller T. Schemer R. Dalton W S 1989 Prediction 0’: doxorubicin resistance ‘in vitro in myeloma, lymphoma, and breast cancer by P-glycoprotein staining. Journal of the National Cancer Institute 81: 696-701 8. Goldstein L J, Galski H, Fojo A et al 1989 Expression of multidrug resistance gene in human cancers. Journal of the National Cancer Institute 81: 116-24
41
9. Brophy N A, Marie J-P, Berry J M et aI 1989 Expression of the mdrl gene in leukaemia and lymphoma cells. Proceedings of the American Society for Clinical Oncology 8: 55 10. Preisler H, Gottesman M, Raza A, Pastan I, Day R, Buffalo N Y 1989 The clinical sianificance of exnression of the MDR gene in acute non-lymphicytic leukaemi’a. Proceedings of the American Society of Clinical Oncology 8: 201 11. Sehested M, Friche E, Ersboll J 1989 Detection of multidrug resistance (MDR) associated glycoprotein P- 170 in human haematological malignancies. Investigational New Drugs 7: 425 12. Ma D, Davey R, Harman D et al 1987 Detection of a multidrug resistant phenotype in acute non-lymphoblastic leukaemia. Lancet 1: 135-37 13. Tsuruo T, Iida H, Tsukagoslin S, Sakwai Y 1981 Overcoming of vincristine resistance in P388 leukaemia through enhanced cytotoxicity by verapamil. Cancer Research 41: 1967-72 14. Yusa K, Tsuruo T 1989 Reversal mechanism of multidrug resistance by verapamil: direct binding of verapamil to Pglycoprotein on specific sites, and transport of verapamil outward across the plasma membrane of K562/ADM cells. Cancer Research 49: 5002206 15. Dalton W S, Grogan T M, Meltzer P S et al 1989 Drugresistance in multiple myeloma and non-Hodgkin’s lymphoma: detection of P-glycoprotein and potential circumvention by addition of verapamil to chemotherapy. Journal of Clinical Oncology 7: 415-24 16. Gore M, Selby P, Millar B et al 1988 The use of verapamil to overcome drug resistance in myeloma. Proceedings of the American Society of Clinical Oncology 7: 228 17. Kerr D J, Graham J. Cummings J et al 1986 The effect of verapamil on the pharmacokinetics of adriamycin. Cancer Chemotherapy and Pharmacology 1% 239-42 18. Milroy R, Connery L, Banham S et al 1988 A randomised trial of verapamil in addition to chemotherapy in small cell lung cancer. Lung Cancer 4: supp. A 101 19 Merry S. Flanigan P, Schlick E, Freshney R I, Kaye S B 1989 Inherent adriamycin resistance in a murine tumour line: circumvention with verapamil and norverapamil. British Journal of Cancer 59: 895-97 20 Echizen H. Brecht T, Niedergan S, Vogelgesang B. Eichelbaum M 1985 The effect of dextro-levo and racemic verapamil on A-V conduction in humans. American Heart Journal 109: 210-17 21. Plumb J A, Milroy R, Kaye S B 1990 The activity of verapamil as a resistance modifier in vitro in drug resistant human tumour cell lines is not stereospecific. Biochemical Pharmacology 39: 787-92 22. Tsuruo T. Lida H, Kitatani Y 1984 Effects of quinidine and related compounds on cytotoxicity and cellular accumulation of vincristine and adriamycin in drug-resistant tumour cells. Cancer Research 44: 4303-07 23 Jones R D, Rankin E M, Habeshaw T et al 1989 A pilot study of epirubicin and quinidine in advanced breast cancer. British Journal of Cancer 60: 451 24. Brown R, Keith N. Stallard S, Kaye S B 1989 Expression of mdrl and gst-pi in breast turnours: Correlations with chemoresponsiveness in vitro. Proceedings of the American Association of Cancer Research 30: 517 25. Twentman P R, Fox N E, White D J G 1987 Cyclosporin A and its analogues as modifiers of Adriamycin and vincristine resistance in a multidrug resistant human lung cancer cell line. British Journal of Cancer 56: 55-57 26. Gosland M P, Lum B L, Sikie B I 1989 Reversal by cefoperazone of resistance to etoposide, doxorubicin and vinblastine in multi-drug resistant human sarcoma cells. Cancer Research 49: 690 l-6905 27. DeGregorio M W, Ford J M, Benz C C, Wiebe V J 1989 Toremifene: Pharmacologic and phannacokinetic basis of reversing multidrug resistance. Journal of Clinical Oncology 7: 1359-64
