Combination therapy with imatinib mesylate (sti571): synopsis of in vitro studies

Chronic myelogenous leukaemia (CML) is the first malignancy in which knowledge about the pathomechanisms of the disease has led to the development of a specific and highly effective molecular therapy. The first clinical trials with the tyrosine kinase inhibitor imatinib mesylate showed high response rates with minimal toxicity and raised hope for a possible ‘cure’ of CML by drug therapy alone. However, not all patients show a complete response and even in responders relapses continue to occur, particularly in advanced phases of the disease. Combination therapy with imatinib and established or investigational antileukaemic agents may lead to an increase in the response rates and remission duration. This review summarizes relevant in vitro studies in the hope that translation of promising preclinical data into clinical trials will lead to further increases in the life expectancy of CML patients. A novel promising treatment modality for chronic myelogenous leukaemia (CML) is selective inhibition of the Bcr-Abl tyrosine kinase. The 2-phenylaminopyrimidine derivative imatinib mesylate (STI571, Glivec®) has been shown to inhibit c-Abl, Bcr-Abl, platelet-derived growth factor receptor β (PDGFRβ), c-Kit and Abl-related gene (ARG) tyrosine kinases (Buchdunger et al, 1996; Druker et al, 1996; Okuda et al, 2001). Imatinib selectively induces apoptosis in Bcr-Abl-positive cells (Gambacorti-Passerini et al, 1997), and suppresses their proliferation and colony formation (Kasper et al, 1999). High response rates to imatinib and sustained remissions have been reported in chronic phase CML (Druker et al, 2001a; Kantarjian et al, 2001). Haematological response rates of over 50% could be achieved in blast crisis patients. However, a significant proportion of patients in myeloid blast crisis (over 50%) and nearly all patients in lymphatic blast crisis and with Philadelphia chromosome-positive (Ph+) acute lymphoblastic leukaemia (ALL) eventually relapsed (Druker et al, 2001b). One way to increase the response rates and prevent the outgrowth of resistant clones may be the combination of imatinib with other antileukaemic agents. Although substances traditionally used for CML treatment, such as interferon-α (IFN), cytosine arabinoside (ara-C), hydroxyurea (HU) and busulphan should be considered preferentially, chemotherapeutic agents that are not part of the classic therapeutic repertoire in CML should also be tested, as concomitant treatment with imatinib may alter the chemosensitivity of leukaemic cells. Combinations of investigational drugs with antileukaemic activity, including homoharringtonine (HHT), As2O3, arginine butyrate, tumour necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL), 17-allylamino-17-demethoxygeldanamycin (17-AAG) and a series of signal transduction inhibitors other than imatinib also seem to be promising. For the choice of agents to be combined with imatinib, two major approaches can be followed. Firstly, substances that act synergistically in combination with imatinib may provide an increased antileukaemic effect. Secondly, drugs that are still active in imatinib-resistant cells are a particularly important approach for the treatment of CML in blastic transformation, in which both primary and acquired resistances to imatinib are a common problem (Druker et al, 2001b). As the therapeutic result is both a function of antileukaemic effect (remission rate, duration of remission, lifespan) and normal tissue damage (e.g. myelosuppression and other adverse effects), toxicity of the therapy to non-leukaemic cells should also be taken into consideration. Several pharmacological methods have been proposed to study drug interactions. When leukaemic cell lines were used as model systems, cytotoxicity was assessed using either variations of the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Thiesing et al, 2000; Tipping et al, 2000; Kano et al, 2001; Topaly et al, 2001a) or the 3H-thymidine incorporation assay (Sun et al, 2001) after 18–72 h of drug exposure. The effects on primary CML and normal progenitor cells were investigated using semisolid colony-forming assays with evaluation of colony-forming unit granulocyte–macrophage (CFU-GM) and burst-forming unit erythroid (BFU-E) colonies after 14 d of incubation (Thiesing et al, 2000). As imatinib is a competitive tyrosine kinase (TK) inhibitor, leukaemic cells were continuously exposed to imatinib in combination treatment settings in most experiments. A concept for analysing drug interactions that is now widely used was first proposed by Loewe & Muischnek (1926). It is based on an assumption that two drugs in a combination have a similar mode of action (e.g. cell growth inhibition) and each of them acts as a substitute of the other. As an agent cannot interact with itself, a predicted result of such combination is assumed to be additive. Combined effects that are greater than predicted are termed synergistic and those less than predicted are antagonistic. This can be illustrated graphically by an isobologram (Fig 1). Different analytical methods for assessment of synergy have been reviewed in detail by Greco et al (1995). Isobologram for X% effect (e.g. 50% growth inhibition of leukaemia cells). An additive interaction can be expressed mathematically as 1 = d1/D1 + d2/D2, where d1 and d2 are concentrations of two drugs in a mixture that yield X% effect, and D1 and D2 are the concentrations of the same drugs that produce X% effect alone. The right-hand side of this equation is equal to Combination Index (CI) of Chou & Talalay (1984) for the mutually exclusive case. This is illustrated by an isobologram with an additivity line depicting data points corresponding to CI = 1. Synergistic combinations yield CI values < 1 (data points below the additivity line) and antagonistic combinations CI values > 1 (data points above the additivity line). Methods for the analysis of drug interactions were not available from all reports analysed for this review. However, when given, drug interaction models, based on either modifications of the isobologram method, e.g. envelope of additivity (Steel & Peckham, 1979), or the median effect method (Chou & Talalay, 1984), were used. When the classic isobologram method was used, the level of synergism was assessed at single effect levels chosen between 50% and 80% of growth inhibition or IC50 and IC80 respectively (Thiesing et al, 2000; Kano et al, 2001). The median effect method of Chou & Talalay (1984) allows an assessment of drug interactions over the entire range of cell growth inhibition. For combinations with drugs that have no antileukaemic effect themselves, an analysis in terms of synergism or antagonism, according to Loewe & Muischnek (1926), is not applicable, sensitization (potentiation) or inhibition of the active agent, e.g. by modulating the IC50 value, should be assessed instead. The leukaemic cell lines used for cytotoxicity tests included the human CML cell lines BV173, EM-3 (Topaly et al, 2001a), KU812, K562, TCC-S, TCC-Y (Kano et al, 2001), KBM-5 (Scappini et al, 2000) and some others derived from patients with CML blast crisis. Further, the human megakaryoblastic cell line MO7e (Thiesing et al, 2000; Sun et al, 2001) and the murine cell lines Ba/F3 (Tipping et al, 2000), FDC-P1 and 32D (Sun et al, 2001) that were rendered growth factor independent by Bcr-Abl transduction were also used by some investigators. Some combination treatment data are available for imatinib-resistant cell lines (Tipping et al, 2000; Hoover et al, 2001) produced by long-term culturing of parental cell lines in imatinib-containing medium. For surrogate assessment of toxicity to normal tissue, either Bcr-Abl-negative leukaemic cell lines or normal haematopoietic progenitors were treated under the same test conditions (Topaly et al, 2001a). It should be noted that MTT-based cell proliferation assays do not distinguish between inhibition of cell proliferation and decrease in cell number due to induction of apoptosis. To confirm the apoptotic mechanism of the observed cytoreduction, e.g. annexin V/propidium iodide staining or terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick end labelling (TUNEL) were carried out by some investigators (Barteneva et al, 2000; Fang et al, 2000; O'Dwyer et al, 2000; Tabrizi et al, 2002; Topaly et al, 2002). More detailed studies on the signal transduction pathways responsible for the induction of apoptosis have been undertaken by several groups. However, assessment of single aspects of apoptosis does not allow inferences regarding the overall synergistic activity of combinations of several agents and is therefore not the subject of this review. Barteneva et al (1999) reported that imatinib pretreatment enhanced cytotoxic activity of IFN in K562 cells. IFN was later reported by the same group to upregulate expression of TRAIL, a cytotoxic cytokine that induces apoptosis in different cell systems, including K562 and KBM-5 CML cell lines. This mechanism possibly underlies the augmentation of imatinib-induced apoptosis by IFN (Barteneva et al, 2000); another possible mechanism is the downregulation of Bcr-Abl protein expression by IFN (Pane et al, 1999). Thiesing et al (2000) reported that the combination of imatinib plus IFN showed an additive effect in K562 and MO7p210 cells. They also observed a significant decrease in and CML colony-forming cells when IFN IFN concentrations between and 1 was to However, the effect of IFN on colony formation was not shown and the interaction was not analysed in terms of et al (2000) reported that the combination of imatinib plus IFN produced less than additive effects on from two patients with Kano et al reported that K562 cells were highly resistant to IFN and therefore could not be with this in the other cell lines TCC-S, imatinib in most synergistically with and in no significant additive effect, was observed when imatinib was combined with The study of et al (2000) no synergism in KBM-5 no effect of IFN in this cell line and concomitant treatment with IFN to the activity of Tipping et al (2000) the of the imatinib-resistant cell lines and to IFN and no significant to their data similar of imatinib in primary CML IFN may an effective therapy in imatinib-resistant patients. in terms of synergism or additivity reported by different may have from in conditions and systems, such as cells used, drug treatment assay methods used to and analytical methods for the effects of the drug combinations (e.g. evaluation at different levels of cell growth Bcr-Abl expression was to be in CML cells resistant to imatinib et al, 2000; & 2000; et al, 2001; et al, 2001). in was reported to levels in primary CML cells in vitro (Pane et al, which further a combination with To some the combination of IFN with imatinib was additive to synergistic in most the that such combinations the remission and This is by reported clinical trials et al, 2001; O'Dwyer et al, 2001). The combination of imatinib and was to be synergistic by several investigators. observed a synergism in and EM-3 cells. K562 cells were highly resistant to and a could not be derived for further evaluation (Topaly et al, 2001a). Thiesing et al (2000) reported synergism in K562 and MO7p210 cells. Imatinib in combination with also produced a significant decrease in leukaemic and colony Kano et al reported this combination to be additive in cell lines K562, TCC-S, using a et al (2000) imatinib + in the KBM-5 cell line and primary CML and reported the drug interaction to be at additive. Fang et al (2000) investigated the effects of this combination in K562 and cells using the annexin A of apoptotic cells was observed after with both agents with each agent alone. have not been analysed in terms of synergism or (Scappini et al, 2000). Tipping et al (2000) the imatinib-resistant cell lines and to be to and observed an enhanced cytotoxic effect when was combined with of into the high activity of each drug in all phases of CML and the interaction imatinib plus is in or considered for for both chronic and advanced phases of the disease (Druker et al, imatinib with not result in a synergistic activity when on Bcr-Abl-positive cell as reported by several investigators. this combination to be additive or antagonistic in BV173, K562 and EM-3 cells (Topaly et al, 2001a). Kano et al reported an additive Tipping et al (2000) an in the imatinib-resistant cell line Thiesing et al (2000) reported in K562 and MO7p210 cells. However, when analysed in colony-forming assays using a significant decrease of colony formation was by the This could be due to the in the assay systems, in or other is a of agents et al, et al, 1999). et al reported that imatinib cytotoxic effects on the CML cell and the CML which is in line with the of et al investigated the of progenitor cells from patients with Bcr-Abl-positive CML and a significant decrease in the in patients treated with The of CML cells by hydroxyurea may the and the imatinib-induced which may the of To combinations of imatinib and are not considered for clinical the less interaction of imatinib plus in cannot be out that CML patients from such a combination as both drugs have and are highly active in Some of the therapeutic by a combination of the two drugs can be even in of antagonism, both agents no toxicity to normal The combination imatinib plus busulphan of busulphan was produced an additive effect in and EM-3 cells at levels of cell growth inhibition. However, at levels in EM-3 cells was observed (Topaly et al, 2002). Combination with busulphan could be considered for CML to other established treatment or for for or cell Different levels of synergism have been reported for the combinations of imatinib with et al, 2001), (Topaly et al, et al, 2001) or (Kano et al, 2001) in Bcr-Abl-positive cell lines. The combination of imatinib plus was as additive by Thiesing et al (2000) in proliferation studies using the assay and by Tabrizi et al in an annexin apoptosis Imatinib plus was assessed to be additive by Kano et al an additive effect was observed when imatinib with or et al, 2001; Kano et al, 2001). Imatinib plus was reported to be antagonistic by Kano et al data this for levels of cell growth at effect additivity was observed et al, 2001). The combination of imatinib with was antagonistic as et al, 2001). Combinations of imatinib with chemotherapeutic drugs traditionally used for CML blast crisis such as and can possibly the therapy of this disease synergism when imatinib was for h after of cells. of toxicity was observed in Bcr-Abl-negative cells et al, 2001). Some augmentation of the antileukaemic effect therefore be for therapy with the of patients is considered to on the effect of the a tumour cell at the of of therapy for or with imatinib be a cytotoxic was from the to the of of protein in the of cells is considered to be the mechanism of action of It the of cells from phase into phase and from phase into phase et al, has been used in clinical trials of CML in chronic phase and a significant effect in CML patients have an IFN treatment et al, 2000). Kano et al the combination of imatinib and in KU812, K562, TCC-S, TCC-Y and reported an additive et al (2000) the combination in the Bcr-Abl-positive cell line KBM-5 and in a assay of primary CML cells. A additive effect was observed in both test Tipping et al (2000) investigated imatinib plus in imatinib-resistant cell lines and and a synergistic effect, even a of to of was observed Combinations using may relevant data on are a transferase inhibitor was reported to selectively inhibit haematopoietic colony formation of primary human CML cells and to with Ba/F3 cells et al, 2001). were used for the in human However, of have not been reported to in CML and the mechanism of action in CML is still the subject of et al, et al, 2000; 2000). Some may a Hoover et al reported that was to inhibit the proliferation of imatinib-resistant Bcr-Abl-positive cell lines and haematopoietic colony formation from CML patients to Further, imatinib-resistant cells were to imatinib by of the combination imatinib + was by et al on primary CML progenitor cells in As is an active agent with an toxicity et al, 2000), with imatinib may be a promising treatment to to group synergism between the and imatinib in EM-3 CML cells in the and K562 cell lines the same drug activity (Topaly et al, 2001b). A number of other e.g. et al, 2001), and et al, 2000), are investigated in phase or clinical studies and may for combination with et al reported that and both inhibitors of the kinase kinase interact synergistically with imatinib to apoptosis in Bcr-Abl-positive cells. The is by & and is considered to An with the has been reported et al, et al, can by on both a tyrosine and a & et al (2000) that single agent treatment with a specific inhibitor, apoptosis in Bcr-Abl-positive K562 cells. of the Bcr-Abl tyrosine kinase by imatinib in the downregulation of at e.g. at h et al, et al, 2002). However, at the activity was or even increased under imatinib treatment, a response a of under imatinib treatment by of the inhibitor in a of apoptosis in an K562 cell line and increased the of an imatinib-resistant cell line to imatinib et al, 2002). The that inhibitors the activity of imatinib and a number of other cytotoxic including et al, et al, and et al, 2000), of inhibitors for the treatment of human As should be considered and a of the toxicity of this in and of each agent in is et al a increase of apoptosis of cell lines K562 and when treated with imatinib and synergism between both agents. data on a novel with and have been reviewed in detail by a number of most kinases to a growth factor receptor protein kinase and Although the action of is on cytotoxic activity is not to that other may also an important preclinical was shown to activity in a number of human tumour cell lines and to act synergistically with a range of drugs 2001). effects were also observed with in a phase clinical in patients with including and as as et al, their in vitro et al used and concentrations of imatinib and which the of using this combination for CML other combinations with imatinib, Tipping et al (2000) combinations including the inhibitor and the inhibitor and are in pathways to inhibitors were to act synergistically with imatinib, which some effects of and that are independent of the Bcr-Abl However, a possible to was reported for the imatinib-resistant cell line Imatinib plus was also by Sun et al reported that at concentrations with no independent effect the cytotoxic activity of imatinib in cell lines and MO7p210 derived from 32D and MO7e Synergistic interactions between imatinib and inhibitors were by et al combinations using or in cell lines. An effective inhibition of growth of CML chronic phase and blast crisis cells by the combinations was also a has a in protein of a of which are in cell and gene inhibitors are to cytotoxic effects a range of human tumour cells in vitro and in by cell et al, 2000). inhibitors have been shown to Bcr-Abl expression and tyrosine kinase activity, which several of apoptosis et al, 1999). et al investigated the combination of imatinib and a and growth inhibition effects in Bcr-Abl-positive cell lines and at no with imatinib was observed in imatinib-resistant cell lines. treatment with imatinib and in antagonistic additive effects were observed when the drugs were imatinib was to for imatinib-resistant no sensitization to imatinib could be achieved either by or of has an and has phase clinical trials for the treatment of chronic leukaemia and a of & 2001). in the treatment of CML to be necrosis factor (TNF)-related apoptosis-inducing ligand or is a of the that can apoptosis in a of and cells. induces apoptosis of cells normal cells and is therefore an for clinical in patients et al, et al, 1999). O'Dwyer et al (2000) an additive interaction between imatinib and in the human CML cell line K562 with to induction of and annexin The additive effect was by with this et al reported that imatinib apoptosis in Bcr-Abl-positive leukaemia cells. Barteneva et al (2000) data for K562 and CML cell lines as assessed by annexin of imatinib due to combination with e.g. be of as are to the of et al, 2000; et al, 2001b). A downregulation of the Bcr-Abl protein was also observed in cells treated with et al, 2001) and arginine et al, has been reported by et al to act or even synergistically in combination with imatinib in K562 and MO7p210 cells. and are in phase clinical trials for leukaemia the are may for clinical combination treatment However, with the combination of imatinib plus an additive effect was observed in cells (Topaly et al, 2001b). toxicity in a phase clinical et al, 2000). of agents that are should be considered with is a that of at It is important for of the of the human and and is in human & most normal activity is or in of with each a that eventually the of to growth and when activity is proliferation a novel inhibitor et al, 2001), when combined with imatinib, was reported to produce an additive of of imatinib-resistant cells in of than those that growth of cells a effect on human normal progenitor a therapeutic et al, 2001). highly specific such as et al, 2001), are under cells increased levels of which in and the activity of imatinib, as shown by et al (2000) in an model with the Bcr-Abl-positive human leukaemic cell line with imatinib for and can in to imatinib (Gambacorti-Passerini et al, 2000). However, this concentrations of h in with concentrations of than that are to than are achieved treatment of in 2000). The of imatinib by was also by et al reported that of normal human at concentrations than to no significant on proliferation of Bcr-Abl-positive K562 cells treated in vitro with imatinib with imatinib treatment alone. derived from CML and with a different from normal to the effect of imatinib on cells in The of levels in CML patients further studies an inhibition of imatinib by can be considered in a clinical and with than in the kinase were reported as of imatinib in CML patients et al, 2001; et al, 2001). The knowledge about and of the tyrosine kinase has of tyrosine kinase inhibitors in the of with different in to by For according to et al some that tyrosine kinase activity of the Bcr-Abl protein to imatinib are not for a and tyrosine kinase possible e.g. et al, 2000), are also As of the tyrosine kinase are a of imatinib particularly in the advanced of the treatment with combinations of two different tyrosine kinase inhibitors are also are et al reported the development of a imatinib-resistant cell line which levels of and a of factor with the parental cell This may a novel mechanism of as factor is to expression of in is the in the of that to some As has been reported to the from the as a result of the effect of an on the imatinib-resistant cell line was with a decrease in the levels of and of and of the resistant cell of a specific inhibitor of also led to of the imatinib The effect of on the clinical of CML patients to should be The normal protein is to between and cell and to function as a of cell growth in the et al, The kinase can be in response to damage and an important in the apoptosis. of to in an of the protein in the and on cell growth from to and K562 CML & that the active Bcr-Abl protein is not into the However, after the Bcr-Abl tyrosine kinase was by imatinib, the of Bcr-Abl was with an as inhibitor of in of the Bcr-Abl The of the Bcr-Abl protein not the cell the of imatinib that led to of the function apoptosis. Bcr-Abl, the factor in the of can be into a of this disease by the Bcr-Abl protein in the carried out with murine cells by a Bcr-Abl a significant increase of selective of Bcr-Abl-positive cells was for the combination of imatinib plus as with each agent alone. to be in a phase clinical with and et al, This toxicity the of to e.g. for as proposed by et al in combination with imatinib, a to is and an in treatment with of the can be The of the antileukaemic of the combination treatment an inhibitor of is an for further at development or of similar less 2001). The studies reviewed in this show that a number of classic chemotherapeutic as as novel signal transduction inhibitors and act synergistically in combination with that the synergism is observed when imatinib is combined with apoptosis-inducing and at levels of growth inhibition (Fig inhibition of the Bcr-Abl tyrosine kinase may induction of apoptotic pathways by chemotherapeutic Combination (CI) as a function of the synergism of imatinib and at effect levels in cells. CI values 1 synergistic interactions were observed even in imatinib-resistant cells. substances with independent antileukaemic activity that are still active in imatinib-resistant disease are of Although in vitro data do not with long-term clinical the studies reported provide for based on the synergistic activity in vitro will into response and rates of CML patients similar combination therapy to be will show such combinations can patients with chronic myelogenous as by in the & 2001). are to and for the This was in part by a from the