8
Multidrug resistance in leukaemia PAUL BAINES PETER CUMBER ROSE ANN PADUA
A major problem in the treatment of leukaemia is the cross-resistance of some primary and many recurrent tumours to chemotherapeutic agents. Tumours can evade chemotherapy in several ways. At the population level, residual surviving cells may adapt to the toxicity, or resistant cells may have already existed. Frequently the malignant population, as a whole, appears insensitive even in the absence of prior exposure to cytotoxic drugs (Sato et al, 1990a). The individual cell can use several mechanisms when faced with a toxic insult. These include decreased drug accumulation, increased intracellular detoxification, amplification of genes coding for detoxifying enzymes, increased repair of DNA and activation of oncogenes . Recently, however, attention has focused on the roles of efflux pumps and topoisomerases because these mechanisms confer the phenomenon of multidrug resistance, where cells become insensitive to a broad spectrum of dissimilar chemotherapeutic agents (Biedler and Riehm, 1970). MDR P·GLYCOPROTEINS
The M DR1 gene which codes for a 170kDa transmembrane P-glycoprotein (P.170) has been extensively investigated in association with multidrug resistance (Hamada and Tsuruo, 1986, 1988; Willingham et al, 1987). P-170 is an energy-dependent efflux pump, and increased levels of the MDR1 transcript and protein have been detected in drug-resistant cells (J uliano and Ling, 1976; Beck et ai, 1979; Bradley et ai, 1988). The MDR1 gene has been shown to confer drug resistance when introduced into sensitive mammalian cells in vitro (Gras et al, 1986; Veda et ai, 1987a; Choi et al, 1988; Pastan et ai, 1988) and in vivo (Mickisch et ai, 1991). The MDR1 p.l70 is a member of the ATP-binding cassette family of transporters which are responsible for the uptake of metabolites in bacteria (Chen et al, 1986; Gras et ai, 1986), the secretion of pheromone in yeast (McGrath and Varshavsky, 1989) and pigment uptake into retinal cells in Drosophila (Dreesen et al, 1988). The protein includes 12 membranespanning hydrophobic regions and two ATP-binding sites (Figure 1). P-170 Bailli~re's ClinicalHaematologyVol. 5. No.4. October 1992 ISBN0-7020-1691-8
943 Copyright© 1992.by BailliereTindall All rightsofreproduction in any formreserved
944
P. BAINES ET AL
OUT
IN
COOH
Figure 1. Schematicdiagram of P-glycoprotein (P-170) in the cell membrane. The molecule comprisesa short, hydrophilic N-terminal region,a longhydrophobic region,which contains12 segmentslikelyto be closely associatedwiththe membrane, and a long, hydrophilic C·terminal region. The nucleotide-binding sites (marked N) are the probable sites of ATP binding.
is also closely related to the cystic fibrosis transmembrane regulator and is similar in its function as an ATP-dependent chloride channel (Valverde et aI, 1992). If the chloride and drug export transport mechanisms are linked. they do not appear to form part of a chloride-drug countertransport mechanism. Much of our knowledge of P·170 comes from cell lines selected for resistance by gradually increasing exposure to cytotoxic drugs in vitro. Characteristically, P-170confers resistance to drugs other than the selecting agent. This phenomenon of multidrug resistance is limited in vitro to certain categories of drugs such as the vinca alkaloids (mitotic spindle poisons). epipodophyllotoxins (topoisomerase II inhibitors) and anthracyclines (DNA-interacting agents), and does not extend to the alkylating agents (chlorambucil, cyclophosphamide) or the antimetabolites (cytarabine, methotrexate, 5-fluorouracil), all of which are important in the treatment of leukaemia. It seems unlikely that P·170 can have a high affinity for all the diverse substrates with which it can evidently interact and which are selected for their affinity to their intracellular target molecules. This has led to the proposal that P-170 must intercept the drug before it enters the cell (Gros et aI, 1986) (Figure 2). If this is true, the ability of P-170 to mediate resistance against broad categories of some drugs, but not others, may reflect the way in which a drug enters the cell. This would explain why multidrug resistant cells are resistant to trimetrexate, which enters the cell via passive diffusion. but not to methotrexate, which enters via the folate receptor (Klohs et al, 1986). Despite their undoubted contribution to our understanding of multidrug resistance, the mechanisms employed by drug-selected cell lines may not always be those used by malignant populations in vivo. This may well be true for mutations in the MDRI gene at codon 185 which give rise to specific patterns of cross-resistance (Choi et aI, 1988), and for gene amplification (Veda et aI, 1987a). Both these abnormalities may be artefacts of the drug selection conditions imposed on emergent resistant lines and are unlikely to
945
MULTIDRUG RESISTANCE IN LEUKAEMIA
P-glycoprotein
FITe J MRK16 •
Figure 2. Schematic diagram of the P-glycoprotein (P·170) resistance mechanism and levels of control. The dark oval represents P·170. which is encoded by the MDRI gene and can be detected using the MRK16 antibody against an extraceUular epitope. P-170can use ATP as an energy source from within the cell to pump cytotoxics out (thick arrows) , which reduces the amount of drug reaching the nucleus (thin arrow) . F1TC, fluorescein isothiocyanate.
prevail in vivo, as some resistant tumours have had no exposure to drugs at all and have elevated levels of MDR1 expression. Much of the gene amplification occurs on unstable double minutes and homogeneously staining regions (Baskin et al, 1981; Biedler et ai, 1983;Tsuruo et ai, 1986), features rarely seen in fresh tumours. Amplification of the MDR1 gene has rarely been observed in clinical haematological samples (Holmes et al, 1989; Ito et al, 1989) and does not correlate with resistance (Sugimoto et ai, 1987). Altered patterns of cross-resistance have been found associated with particular point mutations at codon 185of the MDR1 gene in human carcinoma cell lines (KB) selected in vinblastine or colchicine (Choi et al, 1988). Cells selected in, and with a preferential resistance to colchicine, had a two-base pair change at nu~leotide pos~tio? 554 (g~anine. to thyt;nine) an~ position 555 (adenine to thymine], resulting In a glycine (vinblastine-specific sequence) to a valine (colchicine-specific) change at codon 185. Cells transfected with mutant MDR1 cDNA clones had increased relative resistance to colchicine compared with wild-type transfectants. It has been proposed that position 185 is the site of drug binding, and mutations at this position may alter the affinity of the P-170 !or .different ~rugs, allow.ing cells to become drug resistant due to a quahta tive change in the protem. However, mutations at amino acid position 185 have not been found in DNA from leukaemic cell lines and acute lymphoblastic leukaemia (ALL) patients (Gekeler et aI,
946
P. BAINES ET AL
1991), and in chronic lymphocytic leukaemia (CLL) and acute myelogenous leukaemia (AML) demonstrating a range of MDR1 mRNA expression (Holmes et ai, 1992). A further divergence between in vitro and in vivo models was highlighted by the data of Siapak et al (1990). Cell lines exposed to increasing toxic stress became partially resistant before a noticeable increase in drug efflux was discernible. At higher doses, increased drug efflux preceded the upregulation of MDR1, which was increased only at the highest drug concentrations despite being one of the first mechanisms recruited in de novo malignancy. MDR1 mRNA levels (4.5 kb transcript) are raised in resistant acute non-lymphoblastic leukaemia (ANLL) (Goldstein et ai, 1989; Holmes et ai, 1989; Sato et aI, 1990a; Herweijer et aI, 1990; Nooter et aI, 1990a), but are low in sensitive ALL and ANLL (Fojo et ai, 1987a; Goldstein et aI, 1989). Consequently, cell lines selected for lower levels of resistance, without gene amplification but with increased MDR1 mRNA expression, may be better models for the in vivo situation (Shen et al, 1986). Increased levels of MDR1 mRNA and protein have been observed in B-cell CLL (Groulx et ai, 1988; Perri et al, 1989; Herweijer et al, 1990; Holmes et ai, 1990a; Cumber et ai, 1991) and in relapsed chronic myelogenous leukaemia (CML) (Kuwazuru et ai, 1990). In CML, overexpression of MDR1 has been detected in both chronic and blast crisis phases of the disease (Goldstein et ai, 1989; Herweijer et al, 1990; Weide et al, 1990). In contrast, P-170 has been reported to be present on CML blasts but absent on chronic phase cells (Kuwazuru et al, 1990;Sato et al, 1990b). No correlation with clinical resistance was made in all these studies. P-170-positive cell numbers correlated with in vitro resistance to doxorubicin in a study of myeloma, lymphoma and breast cancer (Salmon et al, 1989), but again no clinical data were available for these patients. High levels of P-170 and increased daunorubicin efflux were reported in blasts from a case of resistant ALL (Redner et ai, 1990). Increased MDR1 mRNA and P-170 levels were found in relapsed and refractory AML patients (Holmes et ai, 1989; P.M. Cumber, H. Limaye, T. Hoy, A. Al Sabah, J. Whittaker and R.A. Padua, unpublished data) and in relapsed ALL (Rothenberg et ai, 1989). There is little published data on MDR1 mRNA levels in the same patient before and after chemotherapy, although the incidence of raised MDR1 mRNA is higher in secondary AML patients following previous exposure to cytotoxic drugs, in myelodysplastic marrow (Ma et ai, 1987; Holmes et al, 1989) and in B-cell CLL following chlorambucil or cyclophosphamide treatment (Holmes et al, 1990a). The mechanism by which MDR1 mRNA transcription might be increased is uncertain, but it is possible that promoter usage may playa role (Veda et ai, 1987b). Sequential analyses of AML, ALL and B-cell CLL patients did not reveal consistent increases of either MDR1 transcripts or protein (Ito et al, 1989; Shustik et aI, 1991). A preliminary study of P-170 in over 200 AML patients did not find a correlation between protein levels and prognosis (Ball et al, 1990). In contrast, a recent comprehensive study clearly demonstrated that complete remission rates
MULTIDRUG RESISTANCE IN LEUKAEMIA
947
were significantly lower in P-17D-positive ANLL than in P-17D-negative ANLL (Campos et al, 1992). Transcription of MDR3, which shows close homology to MDR1 and is similarly located on chromosome 7 (Chen et al, 1986; Roninson et aI, 1986; Van der Bliek et aI, 1987, 1988), is increased in prolymphocytic leukaemia, and there is evidence that this may also be associated with an efflux pump (Nooter et aI, 1990b). Surprisingly, unlike MDR1, MDR3 does not confer drug resistance to sensitive cells in in vitro transfection experiments (Van der Bliek et al, 1988) and its role in conferring drug resistance is unclear. In mice, there is a third member of this gene family, M DR2 (Croop et al, 1989). However, there is no human homologue of this gene (Van der Bliek et aI, 1988). MDRI IN NORMAL TISSUES
Cells of some normal tissues (adrenal, kidney) contain as much or more MDR1 mRNA as cell lines several hundred-fold resistant to drugs (Shen et al, 1986; Fojo et al, 1987a). This distribution is probably determined by the normal chloride channel function of the MDRI P-17D in the apical regions of the secretory epithelia common to these regions (Valverde et al, 1992). Colon, rectum, liver and kidney also exhibit elevated levels of P-17D. It has been suggested that the normal physiological role of P-17D is to export toxins from these organs. Primitive haematopoietic stem cells also possess high levels of P-170, and it was in this situation that the common association of P-170, the presence of the CD34 sialomucin and drug resistance were first noted (Chaudhary and Roninson, 1991; Geller et aI, 1991; List et aI, 1991; Campos et ai, 1992). Marrow lymphocytes also exhibit elevated P-170. If this is a general feature of the lymphoid lineage, this would explain the resistance of peripheral blood lymphocytes to cytotoxic drugs (Kaspers et al, 1991). However, Holmes et al (1990a) reported that MDR1 mRNA levels were low in these cells, whereas total peripheral blood RNA from these normal individuals had high levels of transcripts, presumably from the myeloid compartment. Why cells of some lineages should express more P-170 than others is far from clear. This may be related to cell cycle state and previous exposure to toxic agents, but the answer may well lie in the requirement for a chloride channel network by the cell. DETECTION OF MULTIDRUG RESISTANCE Apart from the measureme~t of intracellular mRNA levels by Northern/slot blotting and RNAse protection (Bradley et ai, 1988), P-170 positive cells can be detected using the MRK16 antibody to an extracellular epitope (Hamada and Tsuruo, 1986) by indirect immunofluorescence. However abnormal glycosylation may mask the epitope re~ognized by MRK16, anda neuraminidase incubation should precede antibody labelling (Cumber et aI, 1990;
948
P. BAINES ET AL
1991), this also means that considerable care needs to be exercised when interpreting the results of earlier studies. Another parameter of multidrug resistance is the reduction in intracellular accumulation of drug. The reduced autofluorescence of daunorubicin due to increased efflux in multidrug resistant cells can be detected at 488 nm (Herweijer et al, 1989; Cumber et al, 1991) using a fluorescence-activated cell sorter. Methods for determining the sensitivity of cell populations to cytotoxic drugs in vitro have been recently reviewed by Veerman and Pieters (1990). Clonogenic assays have the advantage of measuring the effects of drugs on leukaemic stem cells, but the effect on resting clonogenic cells may be missed and, in any case, many samples fail to grow. The differential staining cytotoxicity assay, or DISC assay (Weisenthal et ai, 1983), has the advantages of being rapid (2-4 days), of being applicable to most haematological tumours, and of detecting the morphology of viable and dead cells, which is useful when non-tumour cells contaminate the sample. A modification of the MIT assay (Hansen et al, 1989), in which a soluble tetrazolium salt, 3-4, 5-dimethylthiazol-2-5-diphenyl tetrazolium bromide, is added to cells following 2-4 days of culture in decreasing drug concentrations, can also be used (Santini et al, 1989; Sargent and Taylor, 1989; Pieters et al, 1991). Viable cells convert the salt into a purple formazan precipitate, which can be dissolved with a dimethylformamide/sodium dodecyl sulphate (SDS) solution and the extinction read at 570 nm on a microplate reader. The dose of drug required to kill 50% of cells can be calculated from the resulting curve. Unlike the DISC assay, the MIT technique only gives an overall result for the entire population, so contaminating maturing cells may influence the 100
~
~
It)
1G
~
J
80
60 40
20 0
.008
.03
.13
.5
Daunorublcln cone • JJg'm1 Figure 3. Decreasing cell viability, measured by decreasing reduction of MIT. in resistant VLB (0 and .) and sensitive CEM (0 and .) human leukaemic T-cell lines, with increasing
concentrations ofdaunorubicin alone (open symbols) orwith both daunorubicin and1.25l'-glml of the 'modifier' cyclosporin A (closed symbols) added over 48 h of culture.
MULTIDRUG RESISTANCE IN LEUKAEMIA
949
results. Nevertheless, this is a quick procedure not open to subjective interpretation and is highly reproducible when tested on multidrug resistant cell lines (Figure 3). Furthermore, resistance, as determined by MIT, can distinguish between responders and non-responders in de novo AML (Santini et al, 1989; Sargent and Taylor, 1989) and has been found to be predictive of remission duration induced by thioguanine, daunorubicin and prednisolone in childhood ALL (Pieters et ai, 1991). REVERSAL OF MULTIDRUG RESISTANCE A typical feature of P-170-mediated multidrug resistance is that it can be reversed by a variety of 'modifiers' which also inhibit its chloride channel activity (Valverde et al, 1992). These are diverse and have been recently reviewed by Ford and Heit (1990), but include calcium channel blockers (verapamil), cyclosporins, vinca alkaloid and anthracycline analogues, steroids, tamoxifen and metabolic poisons. The analogues of vinca alkaloids and anthracyclines seem likely to compete for P-170 binding, and photoactive analogues of verapamil show that this drug can also bind P-170 directly (Safa, 1988). Cyclosporins may also act in this way. The mechanisms though are far from clear, with some workers (Hamada and Tsuruo, 1988) claiming that the ATPase associated with P-170 is the target of verapamil, in which case phosphorylation of P-170 could affect drug binding via a conformational change (Hamada et aI, 1987). Cyclosporins can also bind calmodulin and cyclophilins, which in turn inhibits the translocation of transcription factors to the nucleus (Flanagan et al, 1991; Liu et al, 1991); this could restrict MDRl expression. Consistent with the reversal of resistance, verapamil (Tsuruo et ai, 1981, 1982) increases intracellular vincristine or doxorubicin levels in multidrug resistant cells by inhibiting their efflux, and this results in an increase in the sensitivity of these cells to these drugs. Verapamil can increase daunorubicin accumulation in AML blasts in vitro (Maruyama et al, 1989), and cyclosporin A can restore daunorubicin accumulation in AML and B-cell CLL (Nooter et al, 1990a; Cumber et al, 1991). The change in drug accumulation mediated by cyclosporin A treatment appears to be correlated with levels of P-170 (Cumber et al, 1991). Cyclosporin A has already been used to increase the sensitivity of resistant AML to daunorubicin therapy (Sonneveld and Nooter, 1990). Multidrug resistant cells are quite often more sensitive than their parent lines to the modifier alone. Verapamil, alone, is particularly toxic to several multidrug resistant lines (Twentyman et aI, 1986; Warr et al, 1986; CanoGauci and Riordan, 1987), but this effect is also seen for less-specific agents such as detergents (Bech-Hansen et al, 1976), shear forces (Riordan and Ling, 1979) and osmotic lysis'. These less-specific effects may reflect lipid
changes in the mem~rane, ~hlCh are not regarded as instrumental in the development of.multldrug reslsta!1ce (Montaudan et al, 1986) but which may mediate the action of some modifiers.
Modifiers may do more than reverse P-170-mediated resistance.
950
P. BAINES ET AL
Doxorubicin-selected HL60 lines did not express P-170 but showed decreased drug accumulation, reversible with verapamil (McGrath and Center, 1987; McGrath et aI, 1989). This suggests other efflux pumps may exist which can be blocked by modifiers. TOPOISOMERASES Although increased P-170-mediated drug efflux is a common response of cells to toxic insult, some drugs such as VP-16, VM-26 and mitoxantrone select lines which exhibit non-P-170 forms of multidrug resistance (Odaimi et aI, 1986; Beran and Andersson, 1987; Yalowich et al, 1987; Dalton et al, 1988; Harker et aI, 1989). In addition, these ceIl lines are cross-resistant to most of the anticancer drugs but, unusually, not to the vinca alkaloids (Beck et aI, 1987; Danks et aI, 1987)-a situation termed 'atypical multidrug resistance'. Many of these drugs, in fact, appear to exert their pharmacological activity via interaction with DNA topoisomerase II (Topo II) (Lock and Ross, 1987; Sullivan et al, 1987; D'Arpa and Liu, 1989). Together with its family member, topoisomerase I, Topo II is an enzyme responsible for the breakage and religation of coiled DNA (see Wang 1985, 1989 for reviews). Both of these enzymes are associated with highly transcribed genes (Fleischman et aI, 1984; Riou et aI, 1989). A number of drugs appear to stabilize the complex between Topo-II and DNA (Osheroff, 1989), increasing single- and double-strand breaks (Ross et al, 1984; Gewirtz, 1991). This can lead to G2 arrest and inhibition of p34 cdc2 kinase activity (Dive and Hickman, 1991). Atypical multidrug resistance is associated with a reduced level of Topo II activity (Danks et aI, 1988; Cole et aI, 1991; Sinha and Eliot, 1991). The Topo II gene has been cloned and localized to chromosome 17 (TsaiPflugfelder et aI, 1988). Undetectable levels have been observed in the lymphocytes of CLL patients, whereas detectable levels of expression were observed in ALL and non-Hodgkin's lymphoma (Potmesil et al, 1988). These data have not been correlated with clinical drug resistance. Point mutations in Topo II have been observed in drug-resistant cell lines (Bugg et aI, 1991; Hinds et aI, 1991), but such mutations have not yet been detected in vivo. Intercalating agents can also stabilize Topo II-DNA complexes (Tewey et aI, 1984; Fox and Smith, 1990), which would explain why this category of drugs can select ceIl lines exhibiting both P-170-mediated and atypical multidrug resistance (Caprinico et aI, 1986; Zijlstra et aI, 1987; Sinha et al, 1988; Friche et aI, 1991). These two forms of multidrug resistance may interact within resistant cells. Modifiers, such as verapamil, potentiate etoposide-induced single-strand breaks associated with increased etoposide accumulation (Yalowich and Ross, 1984), which probably results from verapamil inhibition of P-170 activity. Even greater complexity arises in lines where verapamil appears to inhibit repair of single-strand breaks directly (Harker et al, 1986). Immunological detection of Topo II is now possible (Smith and Makinson, 1989) and assays for its activity and DNA
MULTIDRUG RESISTANCE IN LEUKAEMIA
951
cleavage are becoming available (Andrea et al, 1991), but the number of strand breaks induced by some drugs (anthracyclines) does not always correlate well with toxicity (Gerwirtz, 1991). This may mean that damage to specific genes is more important or it may reflect the variable presence of other resistance mechanisms. GLUTATHIONE-S-TRANSFERASES The glutathione-S-transferases (GSTs) are a family of enzymes involved in drug detoxification (Hayes and Wolf, 1988). The cytosolic GST isoenzymes conjugate electrophilic drugs, toxins and carcinogens to reduced glutathione (GSH) before elimination from the body. The GST a isoenzymes have glutathione peroxidase activity and prevent oxidative damage (Ketterer et ai, 1986). GST .... reduces peroxidized DNA, which may have a role in DNA repair (Tan et al, 1988). Neither GST a nor I-L seem to be expressed in haematopoietic cells (Hall et al, 1990a,b). Increased expression of GST 1T has been demonstrated in cell lines resistant to alkylating agents (Wang and Tew, 1985; Batist et al, 1986; Kramer et al, 1988) and doxorubicin (Robson et ai, 1986; Samuels et ai, 1991). In chemically induced hepatocarcinogenesis, and in non-small cell lung carcinoma, both GST1T and MDRI P-170 are overexpressed (Kitahara et ai, 1984; Thorgeirsson et ai, 1987; Volm et al, 1991). The expression of GST 1T is increased in a number of haematological malignancies (McQuaid et al, 1989; Holmes et al, 1990b; Schisselbauer et al, 1990) and may be higher on relapse (Moscow et al, 1989a). However, sequential studies have not found modulation of GST 1T expression in CLL patients treated with chlorambucil (Holmes et ai, 1990b). In AML patients sampled at presentation and subsequent relapse, the levels were reduced or remained constant following chemotherapy. There was no coordinate expression of GST 1T and MDRI. While GST is an important detoxifying enzyme in the liver and expression of GST 1T or a in yeast confers drug resistance (Black et ai, 1990), the evidence to support a role for the GSH redox system in the resistance expressed by haematological neoplasms is uncertain (Holmes et ai, 1990b; Begleiter et ai, 1991). Furthermore, introducing GST 1T into MCF-7 breast cancer cells by transfection did not confer a multidrug resistant phenotype (Moscow et ai, 1989b). OTHER DRUG RESISTANCE MECHANISMS Little work has been carried out on other known detoxification pathways such as cytochrome P450 (Porter and Coon, 1991) and carbonyl reductase (Forrest et ai, 1991). Oncogene activation has been linked to drug resistance since transformation with c-MYC, c-RAS, v-ras or v-rafcan increase P-170 and GSTI GSH levels in NIH 31'3 cells (S~lar, 1988; Niimi et ai, 1990) and in rat liver epithelial cells transformed with v-Ha-RAS or v-raf (Burt et ai, 1988).
952
P. BAINES ET AL
Oncogene levels increased following etoposide, but this was concluded to reflect a first step in cell death (Rubin et aI, 1991). Finally, there are a number of ill-defined proteins whose levels are elevated in resistant lines (Bhalla et aI, 1985; Danks et aI, 1985;Tsuruoet al, 1986).
THERAPEUTIC OPTIONS Although tumour drug resistance in vitro, and its parameters, do not take account of the overall pharmacology of cytotoxic drugs in vivo, these assays can be of diagnostic and prognostic value. For example, in de novo AML, responders showed lower in vitro resistance than non-responders (Santini et aI, 1989; Sargent and Taylor, 1989). In childhood ALL, in vitro resistance to thioguanine, daunorubicin and prednisolone was associated with poorer outlook (Pieters et al, 1991). Poor remission rates correlated with high MDRI expression in a varietyofleukaemias (Marie et aI, 1991) and in AML (Sato et aI, 1990a). Remission rates in ANLL were significantly lower in leukaemias with high numbers of P-170-positive cells (Campos et al, 1992), and residual P-170-positive cells during remission may underlie early relapse (Musto et aI, 1991). Strategies for the reversal of drug resistance which have emerged from research into drug resistance mechanisms include inhibition of the promoter of the dihydrofolate receptor gene by rnithramycin, which restores the efficacy of methotrexate in breast carcinoma lines with multiple copies of the dihydrofolate reductase gene. Such a molecular approach might be extended to MDRI transcription, perhaps via anti-sense oligonucleotide therapy. More immediate attention has focused on the reversal of multidrug resistance by modifiers. Verapamil has been studied in most detail and initial results in clinical trials have been encouraging. Dalton et al (1989) studied patients with resistant multiple myeloma and non-Hodgkin's lymphoma. Three out of seven patients with P·170-positive tumours, previously refractory to vincristine, doxorubicin and dexamethazone, showed improvement upon the addition of verapamil during their chemotherapy regimens. The dose-limiting factor was cardiotoxicity with hypotension, first degree atrioventricular block and junctional rhythms. The use of the racemic form D-verapamil to decrease cardiotoxicity has been disappointing (Dalton et aI, 1989). Cyclosporin A has been used to treat a patient with refractory AML, with a transient elimination of an MDRI positive clone and a short-lived clinical response (Sonneveld and Nooter, 1990). The major toxicities associated with cyclosporin A are renal impairment and immunosuppression. The development of cyclosporin A analogues which are far less immunosuppressive and nephrotoxic may circumvent these problems. These agents have shown promise in vitro and are shortly to be used in clinical trials (Twentyman, 1988; Gaveriaux et aI, 1989; Boesch et al, 1991). Quinine, which has been shown to increase anthracycline accumulation in vitro, is less cardiotoxic than verapamil and may be useful (Chauffert et al, 1990). Tamoxifen has also been shown to be effective in the reversal of multidrug
MULTIDRUG RESISTANCE IN LEUKAEMIA
953
resistance in cell lines which appear to be using non-P-170 mechanisms (Chatterjee and Harris, 1990). The development of these rational, therapeutic options from a better understanding of the mechanisms involved in clinical drug resistance is encouraging and can be seen as an important step towards specific therapy designed for the individual patient. SUMMARY Multidrug resistance hampers successful chemotherapy in many haernatological neoplasms and is mediated by several cellular proteins. In some cases, the genes encoding these proteins have been shown to confer resistance on transfer to drug-sensitive cell lines. This is true for the efflux pump product of the MDRI gene, P-170. Upregulation of enzymes such as GST has been observed, although the contribution of this enzyme in drug resistance expressed by malignant haematopoietic cells is still uncertain. Cells also appear to be able to downregulate enzymes which are drug targets. Examples include the decrease in Topo II which accompanies the resistance shown by cells to VP·16 and VM-26. Although many reports include both presentation and relapsed patients, there are few data on samples drawn from the same patients before and after chemotherapy. While P-170 and GST appear to be raised more often in cells from resistant and relapsed disease, it is quite clear that such mechanisms can be active in de novo malignancy and do not necessarily emerge as a consequence of prior chemotherapy.
Methods of detecting drug resistance are reviewed here; these include in vitro cellular assays for drug toxicity, and molecular, immunological and functional detection of P-170 or Topo II. The clinical evaluation of such assays is only just beginning and some of the data are contradictory. To some extent, this may reflect the complex way in which the various resistance mechanisms may interact. Nevertheless, there are some encouraging early signs that the application of these assays to clinical material will yield valuable data on the relative contributions of these mechanisms and on ways in which they may be overcome. At present, much attention has focused on the potential of agents which prevent the P-170 efflux pump from exporting cytotoxics from the cell. This is likely to be only the first of new therapies arising from an improved understanding of multidrug resistance. More immediately, assays for multidrug resistance and its parameters may find their place as routine diagnostic and prognostic tools in the laboratory. REFERENCES Andrea JE, Adachi K & Morgan AR (1991) F1uorometric assays for DNA topo isomerases and topoisomerase-targeted drugs: quantitation of catalytic activity and DNA cleavage . Motecular Pharmacology 40: 495-501. Ball ED, Lawrence D, Malnar Met al (1990) Correlation of CD34 and multi-drug resistance P·l7D with FAB and cytogenetics but not prognosis in acute myeloid leukemia (AML) . Blood 76(i): 2513.
954
P. BAINES ET AL
Baskin F, Rosenberg RN & Dev V (1981) Correlation of double minute chromosomes with unstable multidrug resistance in uptake mutants of neuroblastoma cells. Proceedings ofthe National Academy of Sciences of the USA 78: 3654-3658. Batist G, Tulpule A, Sinha BK et al (1986) Overexpression of a novel anionic glutathione transferase in multi-drug-resistant human breast cancer cells. Journal of Biological Chemistry 261: 15544-15549. Bech-Hansen NT, Till JE & Ling V (1976) Pleiotropic phenotype of colchicine resistant CHO cells: cross resistance and collateral sensitivity. Journal of Cellular Physiology 88: 23-32. Beck WT, Mueller TJ & Tanzer LR (1979) Altered surface membrane glycoproteins in vinca alkaloid-resistant human leukemic Iymphoblasts. Cancer Research 39: 207(}..2076. Beck WT, Cirtain MC, Danks MK et al (1987) Pharmacologic, molecular and cytogenetic analysis of 'atypical' multidrug resistant human leukemic cells. Cancer Research 47: 5455-5460. Begleiter A, Goldenberg GJ, Anhalt CD et al (1991) Mechanisms of resistance to chlorambucil in chronic lymphocytic leukemia. Leukemia Research IS: 1019-1027. Beran M & Andersson BS (1987) Development and characterisation of a human myelogenous leukemia cell line resistant to 4-(9'acridinyl amino )-3-methanesulfon-M-anisidide. Cancer Research 47: 1897-1904. Bhalla K, Hindenburg A, Taub RN & Grant S (1985) Isolation and characterisation of an anthracycline-resistant human leukemic cell line, Cancer Research 45: 3657-3662. Biedler JL & Riehm JI (1970) Cellular resistance to actinomycin D in Chinese hamster ceils in vitro: cross-resistance, radioautographic and cytogenetic studies. Cancer Research 30: 1174-1184. Biedler JL, Chang T, Meyers MB et al (1983) Drug resistance in Chinese hamster lung and mouse tumour cells. Cancer Treatment Reports 67: 859-867. Black SM, Beggs JO, Hayes JO et al (1990) Expression of human glutathione S-transferases in Saccharomyces cerevislae confers resistance to the anticancer drugs adriamycin and chlorambucil. Biochemical Journal 268: 309-315. Boesch D, Gaveriaux C, Jachez B et al (1991) In vivo circumvention of P-glycoproteinmediated multidrug resistance of tumour cells with SDZ PSC 833. Cancer Research 51: 4226-4233. Bradley G, Juranka PF & Ling V (1988) Mechanisms of multidrug resistance. Biochimica et Biophysica Acta 948: 87-128. Bugg BY, Danks MK, Beck WT & Suttle DP (1991) Expression of a mutant DNA topoisomerase II in CCRF-CEM human leukemic cells selected for resistance to teniposide. Proceedings of the National Academy of Sciences of the USA 88: 7654-7658. Burt RK, Garfield S, Johnson K & Thorgeirsson SS (1988) Transformation of rat liver epithelial cells with v·JI-ras or v-ra] causes expression of mdr·l, glutathione-Sxransferase-P and increased resistance to cytotoxic chemicals. Carcinogenesis 9: 2329-2332. Campos L, Guyotat D, Archimbaud E et al (1992) Clinical significance of multidrug resistance P-glycoprotein expression on acute nonlymphoblastic leukemia cells at diagnosis. Blood 79: 473-476. Cano-Gauci DF & Riordan JR (1987) Action of calcium antagonists on multidrug-resistant cells: specific cytotoxicity independent of increased cancer drug accumulation. Biochemical Pharmacology 36: 2115-2123. Caprinico G, Dasdia T & Zunino F (1986) Comparison of doxorubicin-induced DNA damage in doxorubicin-sensitive and resistant P388 murine leukemia cells. International Journal of Cancer 37: 227-231. Chatterjee M & Harris AL (1990) Reversal of acquired resistance to adriamycin in CJlO cells by tamoxifen and 4·hydroxy tamoxifen: role of drug interaction with alpha 1 acid glycoprotein. British Journal of Cancer 62: 712-717. Chaudhary PM & Roninson IB (1991) Expression and activity of P-glycoprotein, a multidrug efflux pump, in human hematopoietic stem cells. Cell 66: 85-94. Chauffert B, Pelletier JI, Corda C et al (1990). Potential usefulness of quinine to circumvent the anthracycline resistance in clinical practice. British Journal of Cancer 62: 395-397. Chen C·J, Chin JE, Veda K er al (1986) Internal duplication and homology with bacterial transport proteins in the mdr-l (P-glycoprotein) gene from multidrug resistant human cells. Cell 47: 381-389. Choi K, Chen CJ, Kriegler M & Roninson IB (1988) An altered pattern of cross-resistance in
MULl1DRUG RESISTANCE IN LEUKAEMIA
955
multidrug-resistant human cells results from spontaneous mutations in the mdr1 (Pglycoprotein) gene. Cell 53: 519-529. Cole SPC, Chanda ER, Dicke FP et al (1991) Non P-glycoprolein-mediated multidrug resistance in a small cell lung cancer cell line: evidence for decreased susceptibility to drug-induced DNA damage and reduced levels of topoisomerase II . Cancer Research 51: 3345-3352. Croop JM, Raymond M, Haber 0 et al (1989) The three mouse multi-drug resistance (mdr) genes are expressed in a tissue specific manner in normal mouse tissues. Molecular and Cellular Biology 9: 1346-1350. Cumber PM, Jacobs A, Hoy T et al (1990) Expression of the multidrug res istance gene (mdr1) and epitope masking in chronic lymphatic leukaemia. British Journal ofl/aematology 76: 226-230. Cumber PM, Jacobs A, Hoy T et al (1991) Increased drug accumulation ex vivo with cyclesporin in chronic lymphatic leukemia and its relationship to epitope masking of Pglycoprotein. Leukemia 5: 1050-1053. Dalton WS, Cress AE, Alberts DS & Trent JM (1988) Cytogenetic and phenotypic analysis of a human colon carcinoma cell line resistant to mitoxantrone. Cancer Research 48: 18821888. Dalton WS, Grogan TM, Meltzer PS et al (1989) Drug resistance 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: 415-424. Danks MK, Metzger OW, Ashmun RA & Beck WT (1985) Monoclonal antibodies to glycoproteins of vinca alkaloid-resistant human leukemic cells. Cancer Research 45: 3220-3224. Danks MK, Yalowich JC & Beck WT (1987) Atypical multidrug resistance in a human leukemic cell line selected for resistance to teniposide (VM-26). Cancer Research 47: 1297-1301. Danks MK. Schmidt CA. Cirtain MC et al (1988) Altered catalytic activ ity of and DNA cleavage by DNA topoisomerase II from human leukemic cells selected for resistance to VM-26. Biochemistry 27: 8861-8869. D'Arpa P & Liu LF (1989) Topoisomerase-targeting anti-tumour drugs. Blochimica et Biophysica Acta 989: 163-177. D ive C & Hickman JA (1991) Drug-target interactions: only the first step in the commitment to a programmed cell death? British Journal of Cancer 64: 192-196. Dreesen TO, Johnson DH & HenikoffS (1988) The brown protein of Drosophila melanogaster is similar to the white protein and to components of active transport complexes. Molecular and Cellular Biology 8: 5206-5215. Flanagan WM, Carthesy B, Bram RJ & Crabtree GR (1991) Nuclear association of a T-cell transcription factor blocked by FK-506 and cyclosporin A. Nature 352: 80H07. Fleischman G. Pftugfelder G, Steiner EK et al (1984) Drosophila DNA topoisomerase I is associated with transcriptionally active regions of the genome. Proceedings ofthe National Academy ofSciences of the USA 81: 6958-
Multidrug resistance in leukaemia PAUL BAINES PETER CUMBER ROSE ANN PADUA
A major problem in the treatment of leukaemia is the cross-resistance of some primary and many recurrent tumours to chemotherapeutic agents. Tumours can evade chemotherapy in several ways. At the population level, residual surviving cells may adapt to the toxicity, or resistant cells may have already existed. Frequently the malignant population, as a whole, appears insensitive even in the absence of prior exposure to cytotoxic drugs (Sato et al, 1990a). The individual cell can use several mechanisms when faced with a toxic insult. These include decreased drug accumulation, increased intracellular detoxification, amplification of genes coding for detoxifying enzymes, increased repair of DNA and activation of oncogenes . Recently, however, attention has focused on the roles of efflux pumps and topoisomerases because these mechanisms confer the phenomenon of multidrug resistance, where cells become insensitive to a broad spectrum of dissimilar chemotherapeutic agents (Biedler and Riehm, 1970). MDR P·GLYCOPROTEINS
The M DR1 gene which codes for a 170kDa transmembrane P-glycoprotein (P.170) has been extensively investigated in association with multidrug resistance (Hamada and Tsuruo, 1986, 1988; Willingham et al, 1987). P-170 is an energy-dependent efflux pump, and increased levels of the MDR1 transcript and protein have been detected in drug-resistant cells (J uliano and Ling, 1976; Beck et ai, 1979; Bradley et ai, 1988). The MDR1 gene has been shown to confer drug resistance when introduced into sensitive mammalian cells in vitro (Gras et al, 1986; Veda et ai, 1987a; Choi et al, 1988; Pastan et ai, 1988) and in vivo (Mickisch et ai, 1991). The MDR1 p.l70 is a member of the ATP-binding cassette family of transporters which are responsible for the uptake of metabolites in bacteria (Chen et al, 1986; Gras et ai, 1986), the secretion of pheromone in yeast (McGrath and Varshavsky, 1989) and pigment uptake into retinal cells in Drosophila (Dreesen et al, 1988). The protein includes 12 membranespanning hydrophobic regions and two ATP-binding sites (Figure 1). P-170 Bailli~re's ClinicalHaematologyVol. 5. No.4. October 1992 ISBN0-7020-1691-8
943 Copyright© 1992.by BailliereTindall All rightsofreproduction in any formreserved
944
P. BAINES ET AL
OUT
IN
COOH
Figure 1. Schematicdiagram of P-glycoprotein (P-170) in the cell membrane. The molecule comprisesa short, hydrophilic N-terminal region,a longhydrophobic region,which contains12 segmentslikelyto be closely associatedwiththe membrane, and a long, hydrophilic C·terminal region. The nucleotide-binding sites (marked N) are the probable sites of ATP binding.
is also closely related to the cystic fibrosis transmembrane regulator and is similar in its function as an ATP-dependent chloride channel (Valverde et aI, 1992). If the chloride and drug export transport mechanisms are linked. they do not appear to form part of a chloride-drug countertransport mechanism. Much of our knowledge of P·170 comes from cell lines selected for resistance by gradually increasing exposure to cytotoxic drugs in vitro. Characteristically, P-170confers resistance to drugs other than the selecting agent. This phenomenon of multidrug resistance is limited in vitro to certain categories of drugs such as the vinca alkaloids (mitotic spindle poisons). epipodophyllotoxins (topoisomerase II inhibitors) and anthracyclines (DNA-interacting agents), and does not extend to the alkylating agents (chlorambucil, cyclophosphamide) or the antimetabolites (cytarabine, methotrexate, 5-fluorouracil), all of which are important in the treatment of leukaemia. It seems unlikely that P·170 can have a high affinity for all the diverse substrates with which it can evidently interact and which are selected for their affinity to their intracellular target molecules. This has led to the proposal that P-170 must intercept the drug before it enters the cell (Gros et aI, 1986) (Figure 2). If this is true, the ability of P-170 to mediate resistance against broad categories of some drugs, but not others, may reflect the way in which a drug enters the cell. This would explain why multidrug resistant cells are resistant to trimetrexate, which enters the cell via passive diffusion. but not to methotrexate, which enters via the folate receptor (Klohs et al, 1986). Despite their undoubted contribution to our understanding of multidrug resistance, the mechanisms employed by drug-selected cell lines may not always be those used by malignant populations in vivo. This may well be true for mutations in the MDRI gene at codon 185 which give rise to specific patterns of cross-resistance (Choi et aI, 1988), and for gene amplification (Veda et aI, 1987a). Both these abnormalities may be artefacts of the drug selection conditions imposed on emergent resistant lines and are unlikely to
945
MULTIDRUG RESISTANCE IN LEUKAEMIA
P-glycoprotein
FITe J MRK16 •
Figure 2. Schematic diagram of the P-glycoprotein (P·170) resistance mechanism and levels of control. The dark oval represents P·170. which is encoded by the MDRI gene and can be detected using the MRK16 antibody against an extraceUular epitope. P-170can use ATP as an energy source from within the cell to pump cytotoxics out (thick arrows) , which reduces the amount of drug reaching the nucleus (thin arrow) . F1TC, fluorescein isothiocyanate.
prevail in vivo, as some resistant tumours have had no exposure to drugs at all and have elevated levels of MDR1 expression. Much of the gene amplification occurs on unstable double minutes and homogeneously staining regions (Baskin et al, 1981; Biedler et ai, 1983;Tsuruo et ai, 1986), features rarely seen in fresh tumours. Amplification of the MDR1 gene has rarely been observed in clinical haematological samples (Holmes et al, 1989; Ito et al, 1989) and does not correlate with resistance (Sugimoto et ai, 1987). Altered patterns of cross-resistance have been found associated with particular point mutations at codon 185of the MDR1 gene in human carcinoma cell lines (KB) selected in vinblastine or colchicine (Choi et al, 1988). Cells selected in, and with a preferential resistance to colchicine, had a two-base pair change at nu~leotide pos~tio? 554 (g~anine. to thyt;nine) an~ position 555 (adenine to thymine], resulting In a glycine (vinblastine-specific sequence) to a valine (colchicine-specific) change at codon 185. Cells transfected with mutant MDR1 cDNA clones had increased relative resistance to colchicine compared with wild-type transfectants. It has been proposed that position 185 is the site of drug binding, and mutations at this position may alter the affinity of the P-170 !or .different ~rugs, allow.ing cells to become drug resistant due to a quahta tive change in the protem. However, mutations at amino acid position 185 have not been found in DNA from leukaemic cell lines and acute lymphoblastic leukaemia (ALL) patients (Gekeler et aI,
946
P. BAINES ET AL
1991), and in chronic lymphocytic leukaemia (CLL) and acute myelogenous leukaemia (AML) demonstrating a range of MDR1 mRNA expression (Holmes et ai, 1992). A further divergence between in vitro and in vivo models was highlighted by the data of Siapak et al (1990). Cell lines exposed to increasing toxic stress became partially resistant before a noticeable increase in drug efflux was discernible. At higher doses, increased drug efflux preceded the upregulation of MDR1, which was increased only at the highest drug concentrations despite being one of the first mechanisms recruited in de novo malignancy. MDR1 mRNA levels (4.5 kb transcript) are raised in resistant acute non-lymphoblastic leukaemia (ANLL) (Goldstein et ai, 1989; Holmes et ai, 1989; Sato et aI, 1990a; Herweijer et aI, 1990; Nooter et aI, 1990a), but are low in sensitive ALL and ANLL (Fojo et ai, 1987a; Goldstein et aI, 1989). Consequently, cell lines selected for lower levels of resistance, without gene amplification but with increased MDR1 mRNA expression, may be better models for the in vivo situation (Shen et al, 1986). Increased levels of MDR1 mRNA and protein have been observed in B-cell CLL (Groulx et ai, 1988; Perri et al, 1989; Herweijer et al, 1990; Holmes et ai, 1990a; Cumber et ai, 1991) and in relapsed chronic myelogenous leukaemia (CML) (Kuwazuru et ai, 1990). In CML, overexpression of MDR1 has been detected in both chronic and blast crisis phases of the disease (Goldstein et ai, 1989; Herweijer et al, 1990; Weide et al, 1990). In contrast, P-170 has been reported to be present on CML blasts but absent on chronic phase cells (Kuwazuru et al, 1990;Sato et al, 1990b). No correlation with clinical resistance was made in all these studies. P-170-positive cell numbers correlated with in vitro resistance to doxorubicin in a study of myeloma, lymphoma and breast cancer (Salmon et al, 1989), but again no clinical data were available for these patients. High levels of P-170 and increased daunorubicin efflux were reported in blasts from a case of resistant ALL (Redner et ai, 1990). Increased MDR1 mRNA and P-170 levels were found in relapsed and refractory AML patients (Holmes et ai, 1989; P.M. Cumber, H. Limaye, T. Hoy, A. Al Sabah, J. Whittaker and R.A. Padua, unpublished data) and in relapsed ALL (Rothenberg et ai, 1989). There is little published data on MDR1 mRNA levels in the same patient before and after chemotherapy, although the incidence of raised MDR1 mRNA is higher in secondary AML patients following previous exposure to cytotoxic drugs, in myelodysplastic marrow (Ma et ai, 1987; Holmes et al, 1989) and in B-cell CLL following chlorambucil or cyclophosphamide treatment (Holmes et al, 1990a). The mechanism by which MDR1 mRNA transcription might be increased is uncertain, but it is possible that promoter usage may playa role (Veda et ai, 1987b). Sequential analyses of AML, ALL and B-cell CLL patients did not reveal consistent increases of either MDR1 transcripts or protein (Ito et al, 1989; Shustik et aI, 1991). A preliminary study of P-170 in over 200 AML patients did not find a correlation between protein levels and prognosis (Ball et al, 1990). In contrast, a recent comprehensive study clearly demonstrated that complete remission rates
MULTIDRUG RESISTANCE IN LEUKAEMIA
947
were significantly lower in P-17D-positive ANLL than in P-17D-negative ANLL (Campos et al, 1992). Transcription of MDR3, which shows close homology to MDR1 and is similarly located on chromosome 7 (Chen et al, 1986; Roninson et aI, 1986; Van der Bliek et aI, 1987, 1988), is increased in prolymphocytic leukaemia, and there is evidence that this may also be associated with an efflux pump (Nooter et aI, 1990b). Surprisingly, unlike MDR1, MDR3 does not confer drug resistance to sensitive cells in in vitro transfection experiments (Van der Bliek et al, 1988) and its role in conferring drug resistance is unclear. In mice, there is a third member of this gene family, M DR2 (Croop et al, 1989). However, there is no human homologue of this gene (Van der Bliek et aI, 1988). MDRI IN NORMAL TISSUES
Cells of some normal tissues (adrenal, kidney) contain as much or more MDR1 mRNA as cell lines several hundred-fold resistant to drugs (Shen et al, 1986; Fojo et al, 1987a). This distribution is probably determined by the normal chloride channel function of the MDRI P-17D in the apical regions of the secretory epithelia common to these regions (Valverde et al, 1992). Colon, rectum, liver and kidney also exhibit elevated levels of P-17D. It has been suggested that the normal physiological role of P-17D is to export toxins from these organs. Primitive haematopoietic stem cells also possess high levels of P-170, and it was in this situation that the common association of P-170, the presence of the CD34 sialomucin and drug resistance were first noted (Chaudhary and Roninson, 1991; Geller et aI, 1991; List et aI, 1991; Campos et ai, 1992). Marrow lymphocytes also exhibit elevated P-170. If this is a general feature of the lymphoid lineage, this would explain the resistance of peripheral blood lymphocytes to cytotoxic drugs (Kaspers et al, 1991). However, Holmes et al (1990a) reported that MDR1 mRNA levels were low in these cells, whereas total peripheral blood RNA from these normal individuals had high levels of transcripts, presumably from the myeloid compartment. Why cells of some lineages should express more P-170 than others is far from clear. This may be related to cell cycle state and previous exposure to toxic agents, but the answer may well lie in the requirement for a chloride channel network by the cell. DETECTION OF MULTIDRUG RESISTANCE Apart from the measureme~t of intracellular mRNA levels by Northern/slot blotting and RNAse protection (Bradley et ai, 1988), P-170 positive cells can be detected using the MRK16 antibody to an extracellular epitope (Hamada and Tsuruo, 1986) by indirect immunofluorescence. However abnormal glycosylation may mask the epitope re~ognized by MRK16, anda neuraminidase incubation should precede antibody labelling (Cumber et aI, 1990;
948
P. BAINES ET AL
1991), this also means that considerable care needs to be exercised when interpreting the results of earlier studies. Another parameter of multidrug resistance is the reduction in intracellular accumulation of drug. The reduced autofluorescence of daunorubicin due to increased efflux in multidrug resistant cells can be detected at 488 nm (Herweijer et al, 1989; Cumber et al, 1991) using a fluorescence-activated cell sorter. Methods for determining the sensitivity of cell populations to cytotoxic drugs in vitro have been recently reviewed by Veerman and Pieters (1990). Clonogenic assays have the advantage of measuring the effects of drugs on leukaemic stem cells, but the effect on resting clonogenic cells may be missed and, in any case, many samples fail to grow. The differential staining cytotoxicity assay, or DISC assay (Weisenthal et ai, 1983), has the advantages of being rapid (2-4 days), of being applicable to most haematological tumours, and of detecting the morphology of viable and dead cells, which is useful when non-tumour cells contaminate the sample. A modification of the MIT assay (Hansen et al, 1989), in which a soluble tetrazolium salt, 3-4, 5-dimethylthiazol-2-5-diphenyl tetrazolium bromide, is added to cells following 2-4 days of culture in decreasing drug concentrations, can also be used (Santini et al, 1989; Sargent and Taylor, 1989; Pieters et al, 1991). Viable cells convert the salt into a purple formazan precipitate, which can be dissolved with a dimethylformamide/sodium dodecyl sulphate (SDS) solution and the extinction read at 570 nm on a microplate reader. The dose of drug required to kill 50% of cells can be calculated from the resulting curve. Unlike the DISC assay, the MIT technique only gives an overall result for the entire population, so contaminating maturing cells may influence the 100
~
~
It)
1G
~
J
80
60 40
20 0
.008
.03
.13
.5
Daunorublcln cone • JJg'm1 Figure 3. Decreasing cell viability, measured by decreasing reduction of MIT. in resistant VLB (0 and .) and sensitive CEM (0 and .) human leukaemic T-cell lines, with increasing
concentrations ofdaunorubicin alone (open symbols) orwith both daunorubicin and1.25l'-glml of the 'modifier' cyclosporin A (closed symbols) added over 48 h of culture.
MULTIDRUG RESISTANCE IN LEUKAEMIA
949
results. Nevertheless, this is a quick procedure not open to subjective interpretation and is highly reproducible when tested on multidrug resistant cell lines (Figure 3). Furthermore, resistance, as determined by MIT, can distinguish between responders and non-responders in de novo AML (Santini et al, 1989; Sargent and Taylor, 1989) and has been found to be predictive of remission duration induced by thioguanine, daunorubicin and prednisolone in childhood ALL (Pieters et ai, 1991). REVERSAL OF MULTIDRUG RESISTANCE A typical feature of P-170-mediated multidrug resistance is that it can be reversed by a variety of 'modifiers' which also inhibit its chloride channel activity (Valverde et al, 1992). These are diverse and have been recently reviewed by Ford and Heit (1990), but include calcium channel blockers (verapamil), cyclosporins, vinca alkaloid and anthracycline analogues, steroids, tamoxifen and metabolic poisons. The analogues of vinca alkaloids and anthracyclines seem likely to compete for P-170 binding, and photoactive analogues of verapamil show that this drug can also bind P-170 directly (Safa, 1988). Cyclosporins may also act in this way. The mechanisms though are far from clear, with some workers (Hamada and Tsuruo, 1988) claiming that the ATPase associated with P-170 is the target of verapamil, in which case phosphorylation of P-170 could affect drug binding via a conformational change (Hamada et aI, 1987). Cyclosporins can also bind calmodulin and cyclophilins, which in turn inhibits the translocation of transcription factors to the nucleus (Flanagan et al, 1991; Liu et al, 1991); this could restrict MDRl expression. Consistent with the reversal of resistance, verapamil (Tsuruo et ai, 1981, 1982) increases intracellular vincristine or doxorubicin levels in multidrug resistant cells by inhibiting their efflux, and this results in an increase in the sensitivity of these cells to these drugs. Verapamil can increase daunorubicin accumulation in AML blasts in vitro (Maruyama et al, 1989), and cyclosporin A can restore daunorubicin accumulation in AML and B-cell CLL (Nooter et al, 1990a; Cumber et al, 1991). The change in drug accumulation mediated by cyclosporin A treatment appears to be correlated with levels of P-170 (Cumber et al, 1991). Cyclosporin A has already been used to increase the sensitivity of resistant AML to daunorubicin therapy (Sonneveld and Nooter, 1990). Multidrug resistant cells are quite often more sensitive than their parent lines to the modifier alone. Verapamil, alone, is particularly toxic to several multidrug resistant lines (Twentyman et aI, 1986; Warr et al, 1986; CanoGauci and Riordan, 1987), but this effect is also seen for less-specific agents such as detergents (Bech-Hansen et al, 1976), shear forces (Riordan and Ling, 1979) and osmotic lysis'. These less-specific effects may reflect lipid
changes in the mem~rane, ~hlCh are not regarded as instrumental in the development of.multldrug reslsta!1ce (Montaudan et al, 1986) but which may mediate the action of some modifiers.
Modifiers may do more than reverse P-170-mediated resistance.
950
P. BAINES ET AL
Doxorubicin-selected HL60 lines did not express P-170 but showed decreased drug accumulation, reversible with verapamil (McGrath and Center, 1987; McGrath et aI, 1989). This suggests other efflux pumps may exist which can be blocked by modifiers. TOPOISOMERASES Although increased P-170-mediated drug efflux is a common response of cells to toxic insult, some drugs such as VP-16, VM-26 and mitoxantrone select lines which exhibit non-P-170 forms of multidrug resistance (Odaimi et aI, 1986; Beran and Andersson, 1987; Yalowich et al, 1987; Dalton et al, 1988; Harker et aI, 1989). In addition, these ceIl lines are cross-resistant to most of the anticancer drugs but, unusually, not to the vinca alkaloids (Beck et aI, 1987; Danks et aI, 1987)-a situation termed 'atypical multidrug resistance'. Many of these drugs, in fact, appear to exert their pharmacological activity via interaction with DNA topoisomerase II (Topo II) (Lock and Ross, 1987; Sullivan et al, 1987; D'Arpa and Liu, 1989). Together with its family member, topoisomerase I, Topo II is an enzyme responsible for the breakage and religation of coiled DNA (see Wang 1985, 1989 for reviews). Both of these enzymes are associated with highly transcribed genes (Fleischman et aI, 1984; Riou et aI, 1989). A number of drugs appear to stabilize the complex between Topo-II and DNA (Osheroff, 1989), increasing single- and double-strand breaks (Ross et al, 1984; Gewirtz, 1991). This can lead to G2 arrest and inhibition of p34 cdc2 kinase activity (Dive and Hickman, 1991). Atypical multidrug resistance is associated with a reduced level of Topo II activity (Danks et aI, 1988; Cole et aI, 1991; Sinha and Eliot, 1991). The Topo II gene has been cloned and localized to chromosome 17 (TsaiPflugfelder et aI, 1988). Undetectable levels have been observed in the lymphocytes of CLL patients, whereas detectable levels of expression were observed in ALL and non-Hodgkin's lymphoma (Potmesil et al, 1988). These data have not been correlated with clinical drug resistance. Point mutations in Topo II have been observed in drug-resistant cell lines (Bugg et aI, 1991; Hinds et aI, 1991), but such mutations have not yet been detected in vivo. Intercalating agents can also stabilize Topo II-DNA complexes (Tewey et aI, 1984; Fox and Smith, 1990), which would explain why this category of drugs can select ceIl lines exhibiting both P-170-mediated and atypical multidrug resistance (Caprinico et aI, 1986; Zijlstra et aI, 1987; Sinha et al, 1988; Friche et aI, 1991). These two forms of multidrug resistance may interact within resistant cells. Modifiers, such as verapamil, potentiate etoposide-induced single-strand breaks associated with increased etoposide accumulation (Yalowich and Ross, 1984), which probably results from verapamil inhibition of P-170 activity. Even greater complexity arises in lines where verapamil appears to inhibit repair of single-strand breaks directly (Harker et al, 1986). Immunological detection of Topo II is now possible (Smith and Makinson, 1989) and assays for its activity and DNA
MULTIDRUG RESISTANCE IN LEUKAEMIA
951
cleavage are becoming available (Andrea et al, 1991), but the number of strand breaks induced by some drugs (anthracyclines) does not always correlate well with toxicity (Gerwirtz, 1991). This may mean that damage to specific genes is more important or it may reflect the variable presence of other resistance mechanisms. GLUTATHIONE-S-TRANSFERASES The glutathione-S-transferases (GSTs) are a family of enzymes involved in drug detoxification (Hayes and Wolf, 1988). The cytosolic GST isoenzymes conjugate electrophilic drugs, toxins and carcinogens to reduced glutathione (GSH) before elimination from the body. The GST a isoenzymes have glutathione peroxidase activity and prevent oxidative damage (Ketterer et ai, 1986). GST .... reduces peroxidized DNA, which may have a role in DNA repair (Tan et al, 1988). Neither GST a nor I-L seem to be expressed in haematopoietic cells (Hall et al, 1990a,b). Increased expression of GST 1T has been demonstrated in cell lines resistant to alkylating agents (Wang and Tew, 1985; Batist et al, 1986; Kramer et al, 1988) and doxorubicin (Robson et ai, 1986; Samuels et ai, 1991). In chemically induced hepatocarcinogenesis, and in non-small cell lung carcinoma, both GST1T and MDRI P-170 are overexpressed (Kitahara et ai, 1984; Thorgeirsson et ai, 1987; Volm et al, 1991). The expression of GST 1T is increased in a number of haematological malignancies (McQuaid et al, 1989; Holmes et al, 1990b; Schisselbauer et al, 1990) and may be higher on relapse (Moscow et al, 1989a). However, sequential studies have not found modulation of GST 1T expression in CLL patients treated with chlorambucil (Holmes et ai, 1990b). In AML patients sampled at presentation and subsequent relapse, the levels were reduced or remained constant following chemotherapy. There was no coordinate expression of GST 1T and MDRI. While GST is an important detoxifying enzyme in the liver and expression of GST 1T or a in yeast confers drug resistance (Black et ai, 1990), the evidence to support a role for the GSH redox system in the resistance expressed by haematological neoplasms is uncertain (Holmes et ai, 1990b; Begleiter et ai, 1991). Furthermore, introducing GST 1T into MCF-7 breast cancer cells by transfection did not confer a multidrug resistant phenotype (Moscow et ai, 1989b). OTHER DRUG RESISTANCE MECHANISMS Little work has been carried out on other known detoxification pathways such as cytochrome P450 (Porter and Coon, 1991) and carbonyl reductase (Forrest et ai, 1991). Oncogene activation has been linked to drug resistance since transformation with c-MYC, c-RAS, v-ras or v-rafcan increase P-170 and GSTI GSH levels in NIH 31'3 cells (S~lar, 1988; Niimi et ai, 1990) and in rat liver epithelial cells transformed with v-Ha-RAS or v-raf (Burt et ai, 1988).
952
P. BAINES ET AL
Oncogene levels increased following etoposide, but this was concluded to reflect a first step in cell death (Rubin et aI, 1991). Finally, there are a number of ill-defined proteins whose levels are elevated in resistant lines (Bhalla et aI, 1985; Danks et aI, 1985;Tsuruoet al, 1986).
THERAPEUTIC OPTIONS Although tumour drug resistance in vitro, and its parameters, do not take account of the overall pharmacology of cytotoxic drugs in vivo, these assays can be of diagnostic and prognostic value. For example, in de novo AML, responders showed lower in vitro resistance than non-responders (Santini et aI, 1989; Sargent and Taylor, 1989). In childhood ALL, in vitro resistance to thioguanine, daunorubicin and prednisolone was associated with poorer outlook (Pieters et al, 1991). Poor remission rates correlated with high MDRI expression in a varietyofleukaemias (Marie et aI, 1991) and in AML (Sato et aI, 1990a). Remission rates in ANLL were significantly lower in leukaemias with high numbers of P-170-positive cells (Campos et al, 1992), and residual P-170-positive cells during remission may underlie early relapse (Musto et aI, 1991). Strategies for the reversal of drug resistance which have emerged from research into drug resistance mechanisms include inhibition of the promoter of the dihydrofolate receptor gene by rnithramycin, which restores the efficacy of methotrexate in breast carcinoma lines with multiple copies of the dihydrofolate reductase gene. Such a molecular approach might be extended to MDRI transcription, perhaps via anti-sense oligonucleotide therapy. More immediate attention has focused on the reversal of multidrug resistance by modifiers. Verapamil has been studied in most detail and initial results in clinical trials have been encouraging. Dalton et al (1989) studied patients with resistant multiple myeloma and non-Hodgkin's lymphoma. Three out of seven patients with P·170-positive tumours, previously refractory to vincristine, doxorubicin and dexamethazone, showed improvement upon the addition of verapamil during their chemotherapy regimens. The dose-limiting factor was cardiotoxicity with hypotension, first degree atrioventricular block and junctional rhythms. The use of the racemic form D-verapamil to decrease cardiotoxicity has been disappointing (Dalton et aI, 1989). Cyclosporin A has been used to treat a patient with refractory AML, with a transient elimination of an MDRI positive clone and a short-lived clinical response (Sonneveld and Nooter, 1990). The major toxicities associated with cyclosporin A are renal impairment and immunosuppression. The development of cyclosporin A analogues which are far less immunosuppressive and nephrotoxic may circumvent these problems. These agents have shown promise in vitro and are shortly to be used in clinical trials (Twentyman, 1988; Gaveriaux et aI, 1989; Boesch et al, 1991). Quinine, which has been shown to increase anthracycline accumulation in vitro, is less cardiotoxic than verapamil and may be useful (Chauffert et al, 1990). Tamoxifen has also been shown to be effective in the reversal of multidrug
MULTIDRUG RESISTANCE IN LEUKAEMIA
953
resistance in cell lines which appear to be using non-P-170 mechanisms (Chatterjee and Harris, 1990). The development of these rational, therapeutic options from a better understanding of the mechanisms involved in clinical drug resistance is encouraging and can be seen as an important step towards specific therapy designed for the individual patient. SUMMARY Multidrug resistance hampers successful chemotherapy in many haernatological neoplasms and is mediated by several cellular proteins. In some cases, the genes encoding these proteins have been shown to confer resistance on transfer to drug-sensitive cell lines. This is true for the efflux pump product of the MDRI gene, P-170. Upregulation of enzymes such as GST has been observed, although the contribution of this enzyme in drug resistance expressed by malignant haematopoietic cells is still uncertain. Cells also appear to be able to downregulate enzymes which are drug targets. Examples include the decrease in Topo II which accompanies the resistance shown by cells to VP·16 and VM-26. Although many reports include both presentation and relapsed patients, there are few data on samples drawn from the same patients before and after chemotherapy. While P-170 and GST appear to be raised more often in cells from resistant and relapsed disease, it is quite clear that such mechanisms can be active in de novo malignancy and do not necessarily emerge as a consequence of prior chemotherapy.
Methods of detecting drug resistance are reviewed here; these include in vitro cellular assays for drug toxicity, and molecular, immunological and functional detection of P-170 or Topo II. The clinical evaluation of such assays is only just beginning and some of the data are contradictory. To some extent, this may reflect the complex way in which the various resistance mechanisms may interact. Nevertheless, there are some encouraging early signs that the application of these assays to clinical material will yield valuable data on the relative contributions of these mechanisms and on ways in which they may be overcome. At present, much attention has focused on the potential of agents which prevent the P-170 efflux pump from exporting cytotoxics from the cell. This is likely to be only the first of new therapies arising from an improved understanding of multidrug resistance. More immediately, assays for multidrug resistance and its parameters may find their place as routine diagnostic and prognostic tools in the laboratory. REFERENCES Andrea JE, Adachi K & Morgan AR (1991) F1uorometric assays for DNA topo isomerases and topoisomerase-targeted drugs: quantitation of catalytic activity and DNA cleavage . Motecular Pharmacology 40: 495-501. Ball ED, Lawrence D, Malnar Met al (1990) Correlation of CD34 and multi-drug resistance P·l7D with FAB and cytogenetics but not prognosis in acute myeloid leukemia (AML) . Blood 76(i): 2513.
954
P. BAINES ET AL
Baskin F, Rosenberg RN & Dev V (1981) Correlation of double minute chromosomes with unstable multidrug resistance in uptake mutants of neuroblastoma cells. Proceedings ofthe National Academy of Sciences of the USA 78: 3654-3658. Batist G, Tulpule A, Sinha BK et al (1986) Overexpression of a novel anionic glutathione transferase in multi-drug-resistant human breast cancer cells. Journal of Biological Chemistry 261: 15544-15549. Bech-Hansen NT, Till JE & Ling V (1976) Pleiotropic phenotype of colchicine resistant CHO cells: cross resistance and collateral sensitivity. Journal of Cellular Physiology 88: 23-32. Beck WT, Mueller TJ & Tanzer LR (1979) Altered surface membrane glycoproteins in vinca alkaloid-resistant human leukemic Iymphoblasts. Cancer Research 39: 207(}..2076. Beck WT, Cirtain MC, Danks MK et al (1987) Pharmacologic, molecular and cytogenetic analysis of 'atypical' multidrug resistant human leukemic cells. Cancer Research 47: 5455-5460. Begleiter A, Goldenberg GJ, Anhalt CD et al (1991) Mechanisms of resistance to chlorambucil in chronic lymphocytic leukemia. Leukemia Research IS: 1019-1027. Beran M & Andersson BS (1987) Development and characterisation of a human myelogenous leukemia cell line resistant to 4-(9'acridinyl amino )-3-methanesulfon-M-anisidide. Cancer Research 47: 1897-1904. Bhalla K, Hindenburg A, Taub RN & Grant S (1985) Isolation and characterisation of an anthracycline-resistant human leukemic cell line, Cancer Research 45: 3657-3662. Biedler JL & Riehm JI (1970) Cellular resistance to actinomycin D in Chinese hamster ceils in vitro: cross-resistance, radioautographic and cytogenetic studies. Cancer Research 30: 1174-1184. Biedler JL, Chang T, Meyers MB et al (1983) Drug resistance in Chinese hamster lung and mouse tumour cells. Cancer Treatment Reports 67: 859-867. Black SM, Beggs JO, Hayes JO et al (1990) Expression of human glutathione S-transferases in Saccharomyces cerevislae confers resistance to the anticancer drugs adriamycin and chlorambucil. Biochemical Journal 268: 309-315. Boesch D, Gaveriaux C, Jachez B et al (1991) In vivo circumvention of P-glycoproteinmediated multidrug resistance of tumour cells with SDZ PSC 833. Cancer Research 51: 4226-4233. Bradley G, Juranka PF & Ling V (1988) Mechanisms of multidrug resistance. Biochimica et Biophysica Acta 948: 87-128. Bugg BY, Danks MK, Beck WT & Suttle DP (1991) Expression of a mutant DNA topoisomerase II in CCRF-CEM human leukemic cells selected for resistance to teniposide. Proceedings of the National Academy of Sciences of the USA 88: 7654-7658. Burt RK, Garfield S, Johnson K & Thorgeirsson SS (1988) Transformation of rat liver epithelial cells with v·JI-ras or v-ra] causes expression of mdr·l, glutathione-Sxransferase-P and increased resistance to cytotoxic chemicals. Carcinogenesis 9: 2329-2332. Campos L, Guyotat D, Archimbaud E et al (1992) Clinical significance of multidrug resistance P-glycoprotein expression on acute nonlymphoblastic leukemia cells at diagnosis. Blood 79: 473-476. Cano-Gauci DF & Riordan JR (1987) Action of calcium antagonists on multidrug-resistant cells: specific cytotoxicity independent of increased cancer drug accumulation. Biochemical Pharmacology 36: 2115-2123. Caprinico G, Dasdia T & Zunino F (1986) Comparison of doxorubicin-induced DNA damage in doxorubicin-sensitive and resistant P388 murine leukemia cells. International Journal of Cancer 37: 227-231. Chatterjee M & Harris AL (1990) Reversal of acquired resistance to adriamycin in CJlO cells by tamoxifen and 4·hydroxy tamoxifen: role of drug interaction with alpha 1 acid glycoprotein. British Journal of Cancer 62: 712-717. Chaudhary PM & Roninson IB (1991) Expression and activity of P-glycoprotein, a multidrug efflux pump, in human hematopoietic stem cells. Cell 66: 85-94. Chauffert B, Pelletier JI, Corda C et al (1990). Potential usefulness of quinine to circumvent the anthracycline resistance in clinical practice. British Journal of Cancer 62: 395-397. Chen C·J, Chin JE, Veda K er al (1986) Internal duplication and homology with bacterial transport proteins in the mdr-l (P-glycoprotein) gene from multidrug resistant human cells. Cell 47: 381-389. Choi K, Chen CJ, Kriegler M & Roninson IB (1988) An altered pattern of cross-resistance in
MULl1DRUG RESISTANCE IN LEUKAEMIA
955
multidrug-resistant human cells results from spontaneous mutations in the mdr1 (Pglycoprotein) gene. Cell 53: 519-529. Cole SPC, Chanda ER, Dicke FP et al (1991) Non P-glycoprolein-mediated multidrug resistance in a small cell lung cancer cell line: evidence for decreased susceptibility to drug-induced DNA damage and reduced levels of topoisomerase II . Cancer Research 51: 3345-3352. Croop JM, Raymond M, Haber 0 et al (1989) The three mouse multi-drug resistance (mdr) genes are expressed in a tissue specific manner in normal mouse tissues. Molecular and Cellular Biology 9: 1346-1350. Cumber PM, Jacobs A, Hoy T et al (1990) Expression of the multidrug res istance gene (mdr1) and epitope masking in chronic lymphatic leukaemia. British Journal ofl/aematology 76: 226-230. Cumber PM, Jacobs A, Hoy T et al (1991) Increased drug accumulation ex vivo with cyclesporin in chronic lymphatic leukemia and its relationship to epitope masking of Pglycoprotein. Leukemia 5: 1050-1053. Dalton WS, Cress AE, Alberts DS & Trent JM (1988) Cytogenetic and phenotypic analysis of a human colon carcinoma cell line resistant to mitoxantrone. Cancer Research 48: 18821888. Dalton WS, Grogan TM, Meltzer PS et al (1989) Drug resistance 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: 415-424. Danks MK, Metzger OW, Ashmun RA & Beck WT (1985) Monoclonal antibodies to glycoproteins of vinca alkaloid-resistant human leukemic cells. Cancer Research 45: 3220-3224. Danks MK, Yalowich JC & Beck WT (1987) Atypical multidrug resistance in a human leukemic cell line selected for resistance to teniposide (VM-26). Cancer Research 47: 1297-1301. Danks MK. Schmidt CA. Cirtain MC et al (1988) Altered catalytic activ ity of and DNA cleavage by DNA topoisomerase II from human leukemic cells selected for resistance to VM-26. Biochemistry 27: 8861-8869. D'Arpa P & Liu LF (1989) Topoisomerase-targeting anti-tumour drugs. Blochimica et Biophysica Acta 989: 163-177. D ive C & Hickman JA (1991) Drug-target interactions: only the first step in the commitment to a programmed cell death? British Journal of Cancer 64: 192-196. Dreesen TO, Johnson DH & HenikoffS (1988) The brown protein of Drosophila melanogaster is similar to the white protein and to components of active transport complexes. Molecular and Cellular Biology 8: 5206-5215. Flanagan WM, Carthesy B, Bram RJ & Crabtree GR (1991) Nuclear association of a T-cell transcription factor blocked by FK-506 and cyclosporin A. Nature 352: 80H07. Fleischman G. Pftugfelder G, Steiner EK et al (1984) Drosophila DNA topoisomerase I is associated with transcriptionally active regions of the genome. Proceedings ofthe National Academy ofSciences of the USA 81: 6958-
