Abstract
Several cancer treatments are shifting from traditional, time-limited, nonspecific cytotoxic chemotherapy cycles to continuous oral treatment with specific protein-targeted therapies. In this line, imatinib mesylate, a selective tyrosine kinases inhibitor (TKI), has excellent efficacy in the treatment of chronic myeloid leukemia. It has opened the way to the development of additional TKIs against chronic myeloid leukemia, including nilotinib and dasatinib. TKIs are prescribed for prolonged periods, often in patients with comorbidities. Therefore, they are regularly co-administered along with treatments at risk of drug-drug interactions. This aspect has been partially addressed so far, calling for a comprehensive review of the published data. We review here the available evidence and pharmacologic mechanisms of interactions between imatinib, dasatinib, and nilotinib and widely prescribed co-medications, including known inhibitors or inducers of cytochromes P450 or drug transporters. Information is mostly available for imatinib mesylate, well introduced in clinical practice. Several pharmacokinetic aspects yet remain insufficiently investigated for these drugs. Regular updates will be mandatory and so is the prospective reporting of unexpected clinical observations.
Introduction
Targeted cancer therapies have been designed to interact with particular proteins associated with tumor development or progression. Many of these agents are tyrosine kinases inhibitors (TKIs), targeting enzymes whose disregulated expression and activity are associated with various cancers.1 The pioneer small-molecule TKI imatinib has revolutionized the treatment and prognosis of chronic myeloid leukemia (CML). Imatinib inhibits the tyrosine kinase Bcr-Abl,2 a fusion oncoprotein resulting from the translocation t(9;22)(q34;q11),3 which is associated with the characteristic Philadelphia chromosome,2 a hallmark of chronic myeloid leukemia and of some acute lymphoblastic leukemias.4
However, some patients, especially those in the advanced phase of the disease, develop resistance to imatinib therapy, because of various mechanisms such as BCR-ABL gene amplification,5 low imatinib absorption, or more frequently point mutations into the oncoprotein sequence.6 Several new inhibitors have been developed with increased potency and a broader range of activity against imatinib-resistant mutants. In vitro studies have shown that nilotinib, an imatinib derivative, and dasatinib, structurally unrelated to imatinib, are, respectively, 20- and 300-fold more potent than imatinib against unmutated Abl7 and are active against many imatinib-resistant Bcr-Abl mutants.7
TKIs are extensively metabolized by cytochrome P450 enzymes (CYP), whose activities are characterized by a large degree of interindividual variability.8 Some TKIs are also substrates or inhibitors of the drug transporters P-glycoprotein (Pgp; coded by ABCB1) Breast Cancer Resistance Protein (BCRP; ABCG2) and the organic cation transporter 1 (hOCT1; SLC22A1).9–13 A standard regimen can therefore produce very different circulating and cell concentration profiles from one patient to another, thus favoring the selection of resistant cellular clones by subtherapeutic drug exposure or the occurrence of toxicity in case of overexposure.14,15 Identifying the best active and safe dosing schedule for individual patients to maximize therapeutic benefit has become a scientific and clinical challenge. Combination therapies have been investigated in various conditions, which certainly add a level of treatment complexity, because overlapping toxicities and pharmacokinetic interactions have to be taken into consideration.16,17
We review here systematically and present under an easy-consulting form (Table 1) the information available on pharmacologic interactions between imatinib, dasatinib, and nilotinib and drugs concomitantly prescribed to patients receiving TKIs. The drugs were selected on the basis of the information extracted from our database, used within the framework of Therapeutic Drug Monitoring of TKIs.15 Moreover, classical inhibitors or inducers of cytochromes P450 or drug transporters were also included in this review. We do not intend here to replace individualized medical evaluation, and the data presented here should be used in addition to thorough clinical judgment. Indeed, it may be that our searches still missed some interactions, and actually most interactions do not represent true contraindications but rather call for appropriate dosage adjustments and treatment monitoring measures.
Review of the literature
In addition to official monographs of the drugs,9 literature from Medline and Evidence-Based Medicine Reviews was systematically searched, using the following MeSH terms: “Drug interactions,” “Cytochrome P-450 Enzyme System,” “P-Glycoprotein,” “ABCG2 protein,” “organic cation transporter 1,” “Protein binding,” and the respective TKI and concomitant drugs names. In addition, 2 drug information databases (UpToDate online18 and Cancer Care Ontario19 ) were screened, and abstracts of international and national conferences, review articles, and references given in identified articles were also scanned.20–22 All relevant cited literature on pharmacokinetic or pharmacodynamic interactions was considered for inclusion in Table 1.
Drug interactions were either clinically documented or derived from mechanistic considerations on proven or putative metabolic pathways, protein binding, and transmembrane transport. When data on a particular combination were unavailable, potential interactions were extrapolated from the reported disposition mechanisms of the agents and of similar substrates.
Interaction with imatinib
Imatinib is metabolized mainly by CYP isoenzyme 3A4, whereas CYP1A2, CYP2C9, CYP2C19, CYP2D6, and CYP3A5 are reported to play a minor role in its metabolism.11 This TKI has also been shown to be a substrate of hOCT1, Pgp, and BCRP.9,23–25 However, a controversial report26 suggests that imatinib is an inhibitor rather than a substrate of BCRP, thus leaving uncertainty about the role of this pathway. The metabolites of imatinib are eliminated predominantly through biliary excretion. One metabolite, an N-demethylated piperizine derivative (CGP 74588) shows pharmacologic activity comparable to the parent drug, but the systemic exposure represents ∼ 15% of that for imatinib.13 The fecal-to-urinary excretion ratio is ∼ 5:1. Moreover, imatinib can competitively inhibit the metabolism of drugs that are CYP2C9, CYP2C19, CYP2D6, and CYP3A4 substrates.13 Imatinib is ∼ 95% bound to human plasma proteins, mainly albumin and α1-acid glycoprotein.11,27–29
Interactions should therefore be considered when administering inhibitors of the CYP3A family in combination with imatinib. Strong inhibition, such as achieved with ketoconazole, caused a 40% increase of imatinib exposure in healthy volunteers.30 Interactions are likely to occur with other inhibitors of CYP3A4, such as levothyroxine31,32 voriconazole,33 or amiodarone,34 leading to an increase in plasma concentrations of imatinib. Nevertheless, a study suggests that inhibition of CYP3A4 by the potent irreversible inhibitor ritonavir does not result in increased steady-state plasma concentrations of imatinib, possibly because of the induction of compensatory metabolism or transport mechanisms by ritonavir.35
Concomitant administration of imatinib with inhibitors of both CYP3A4 and Pgp increase not only plasma but also intracellular imatinib concentrations. Dual CYP3A4 and Pgp inhibitors such as verapamil,9 erythromycin,36 clarithromycin,36 ciclosporin,37,38 ketoconazole,30 fluconazole,9,18 and itraconazole9,18 increase intracellular concentrations of imatinib by inhibiting both its metabolism and its efflux by Pgp and might therefore increase its cellular toxicity.
Moreover, inhibition of Pgp by proton pump inhibitors such as pantoprazole was shown to increase brain penetration of imatinib.40 Conversely, another study reported that concomitant administration of a Mg2+-Al3+–based antacid is not associated with meaningful alterations in imatinib absorption.41
Concomitant administration of CYP3A4 inducers such as rifampicin or certain antiepileptics may lead to a reduction of as much as 74% in imatinib exposure.12,13,42 Moreover, the pharmacokinetic profile of imatinib was significantly altered by St John's wort, with reductions of 30% in the median area under the concentration-time curve (AUC).43,44 Concomitant use of enzyme inducers, including St John's wort, may thus necessitate an increase in imatinib dosages to maintain clinical effectiveness.43,44
Interactions with quinidine, ranitidine, or midazolam, known inhibitors of hOCT1, may paradoxically increase the circulating concentrations of imatinib but decrease the intracellular exposure of target cancer cells, known to express this carrier.9,25
With regard to all these mechanisms, it is worth recalling that plasma concentrations of imatinib appear correlated with efficacy and toxicity.29,45–47 A change in imatinib exposure because of a drug interaction might therefore definitely influence its therapeutic efficacy.
TKIs can also inhibit drug transporters and enzymes, leading to changes in the exposure of coadministered drugs. Imatinib enhances the intestinal absorption of ciclosporin, a CYP3A4 and Pgp substrate, and may increase the pharmacologic effects and possibly toxicity of ciclosporin.37,38 Moreover, the clearance of simvastatin (a CYP3A4 substrate) was reduced by 70% when associated with imatinib.13 Administration of imatinib together with metoprolol, a CYP2D6 substrate, resulted in an increase in metoprolol exposure by 23%.13
Data concerning interactions involving protein binding are poorly documented for imatinib. A study showed that St John's wort does not alter the protein binding of imatinib over a wide range of concentrations in vivo.43,44
Interactions of potential clinical relevance can occur with calcium channel blockers such as verapamil and diltiazem, substrates of CYP3A4, which circulating levels are increased when associated with imatinib.18,19 Interactions with simvastatin, amiodarone, and quinidine, involving the same P450 isoenzyme, may also be of clinical relevance.9,18,19,48 In patients taking imatinib, such drugs should be either tapered or avoided and replaced by safer alternatives (eg, pravastatin or sotalol).
Imatinib is also known to inhibit the O-glucuronidation of acetaminophen, possibly inducing hepatotoxicity and liver failure.9 The use of acetaminophen should be limited in patients taking imatinib. A limit has been suggested of 1300 mg acetaminophen per day.49 Liver function tests might be useful to monitor during prolonged treatment.50 Acenocoumarol and phenprocoumon, substrates of CYP2C9, show also increased concentrations; however, this interaction can be compensated by the monitoring of prothrombin time or international normalized ratio.9,18,51
Finally, physicians should be aware that patients with hypothyroid conditions who receive imatinib need increased levothyroxine doses.31,32 The suspected mechanism responsible for this phenomenon is an induction of non–deiodination clearance.31,32 The fraction of levothyroxine that is deiodinated into biologically active triiodothyronine is mainly subject to conjugation with glucuronates and sulfates.31,32 Although the liver primarily mediates glucuronidation and sulfation, these conjugations occur in extrahepatic sites such as the kidney and intestine as well.31,32 Therefore, induction of uridine diphosphate–glucuronyl transferases (UGTs) seems to be involved.31,32 A 2-fold increase in levothyroxine substitution therapy at initiation of imatinib treatment is recommended, along with close monitoring of thyroid function.31,32
Interaction with dasatinib
Dasatinib is metabolized in an active derivative and other inactive metabolites by the CYP3A4 isoenzyme and was also reported to be a substrate of BCRP and Pgp.9,18,52 The active metabolite appears to play a negligible role in therapeutic activity. Dasatinib has an inhibitory activity against CYP2C8 and CYP3A4. Plasma protein binding is ∼ 96% for dasatinib, mainly to albumin.53,54
In healthy subjects receiving ketoconazole, systemic exposure (AUC) to dasatinib was increased by 5-fold.39 Interactions may then occur between dasatinib and other inhibitors of CYP3A4, such as levothyroxine31,32 and voriconazole,33 leading to a marked increase in plasma concentrations of this TKI. Drugs that inhibit both BCRP and CYP3A4, such as verapamil,55 may lead to even larger increase in dasatinib exposure.
Inhibitors of both CYP3A4 and Pgp will increase not only plasma but also intracellular concentrations of dasatinib; this is expected for verapamil,9 erythromycin,9,18 clarithromycin,9,18 ciclosporin,38 ketoconazole,39 fluconazole,9,18 and itraconazole.9,18
Concomitant administration of the CYP3A4 inducer rifampicin leads to a reduction of 80% in dasatinib exposure.12,13,42 St John's wort, a CYP3A4 inducer, may also decrease dasatinib plasma concentrations and should be discouraged in patients receiving dasatinib.56 Antiepileptics (phenobarbital, phenytoin, carbamazepine) are expected to decrease dasatinib concentrations as well.
Moreover, the solubility of dasatinib appears to be pH dependent. Dasatinib exposure is reduced by 61% when famotidine is administered before dasatinib dosing.57 As a result, concomitant administration of agents that provide prolonged gastric acid suppression, such as H2 antagonists and proton pump inhibitors, is not recommended.42 In contrast, dasatinib exposure is unchanged when Mg2+-Al3+–based antacids are administered ≥ 2 hours before dasatinib; but coadministration reduced dasatinib exposure by 55%-58%.57
Dasatinib can also slightly inhibit drug transporters and enzymes, leading to changes in the exposure of coadministered drugs.9,18 The coingestion of dasatinib with simvastatin resulted in a 20% increased exposure to simvastatin.13 Concurrent use with calcium channel blockers such as verapamil and diltiazem, substrates of CYP3A4, should be avoided.18,51
Studies about interactions involving protein binding were unavailable for dasatinib.
In clinical trials, dasatinib treatment has been associated with prolongation of the QTc interval on electrocardiograms, and sudden cardiac deaths have occurred, which are probably related to ventricular repolarization abnormalities.58,59 Association of QT-prolonging drugs such as digoxin, quinolones, methadone, or several psychotropic medications, may increase the risk of such events by additive effect.9,19 Regular electrocardiographic controls (ECG) are strongly recommended in such situations.58,59
Interactions with nilotinib
Nilotinib undergoes metabolism by CYP3A4. It is also a substrate of the efflux transporter BCRP.9,23 Nilotinib is known to inhibit CYP2C8, CYP2C9, CYP2D6, CYP3A4, UGT1A1, and Pgp. In vitro studies suggest that nilotinib also induces CYP2B6 enzymes.19 Note that UGT1A1 inhibition has been associated with an increase in bilirubin levels (especially in patients homozygous for the UGT1A1*28 reduced-function variant).60 The determination of UGT1A1*28 is therefore approved by the Food and Drug Administration as a valid pharmacogenetic test for patients treated by nilotinib.61 This TKI is ∼ 98% bound to albumin and α1-acid glycoprotein.54
Nilotinib exposure is expected to increase under CYP3A4 inhibitors. For example, AUC of nilotinib was increased by a 3-fold factor in healthy subjects receiving ketoconazole.12 Moreover, a study showed that concurrent intake of 240 mL of grapefruit juice increased by 60% nilotinib AUC. Concomitant administration of nilotinib with grapefruit juice is therefore not recommended.62
Conversely, concomitant administration of CYP3A4 inducers such as rifampicin leads to a reduction by a 4.8 factor in nilotinib exposure.12,13,42
Literature about interactions involving protein binding were lacking for nilotinib.
The same potential clinically significant interactions with imatinib and dasatinib can occur with nilotinib. For example, acenocoumarol and phenprocoumon, substrates of CYP2C9, show increased concentrations, imposing careful monitoring of prothrombin time or international normalized ratio.9 Moreover, as with dasatinib, nilotinib has been associated with prolongation of the QTc interval, and cases of sudden cardiac death have been reported.58,59 Accordingly, nilotinib prescribing information includes a black box warning about the risk of QTc prolongation and sudden death and warns that nilotinib should not be used in patients with hypokalemia, hypomagnesemia, or long QT syndrome, either congenital or drug induced.58,59
Conclusions
Pharmacokinetics, drug interactions, and safety recommendations are best characterized for imatinib, which was the first TKI on the market. The other TKIs, just recently marketed, have so far only a limited documentation about clinically relevant interactions. Their concentration profile might be affected to a more dramatic degree by interactions than imatinib exposure.
The 3 TKIs reviewed are indeed substrates of several drug transporters and metabolizing enzymes. They are also capable of inhibiting drug transporters and enzymes, making their disposition and metabolism rather complex and difficult to predict.
Most of the available pharmacokinetic information is based on information obtained from in vitro experiments, animal studies, drug–drug interaction studies, and studies in healthy volunteers with a single dose of the aimed TKI. These results must be translated into treatment adjustment recommendations for the clinical oncology practice, where these drugs are administered on a daily basis in patients receiving various co-medications. The actual relevance of predicted drug interactions is thus still uncertain. Most of the interactions outlined in Table 1 (except those in boldface) are theoretical and have not been confirmed in clinical studies; therefore, they should only be considered indicative. Further interaction mechanisms may still be unknown at present.
We advise the reader to regularly monitor for updates about this topic. Therapeutic Drug Monitoring of TKIs63 should be considered if a drug interaction is suspected, or in case of toxicity, or lack of satisfactory clinical response. Finally, documenting unexpected observations and reporting them to the Pharmacovigilance network is of definite importance.
The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.
Authorship
Contribution: A.H., N.W., M.A.D., M.M., T.B., and L.A.D. are the sole authors of this review article. A.H. and N.W. wrote the manuscript, which was corrected and edited by M.A.D. and M.M. (for hematology and oncology aspects), and T.B. and L.A.D. (for clinical pharmacology aspects).
Conflict-of-interest disclosure: L.A.D. and T.B. have received unrestricted research grants from Novartis. A.H., N.W., T.B., and L.A.D. have received travel grants from Novartis for participating in international meetings on chronic myeloid leukemia and on Therapeutic Drug Monitoring of TKIs. Work on the cardiovascular drug interactions and TKIs has been done following a request from the Swiss GIST Patients Association (http://www.gastrointestinale-stromatumoren.com/) supported by Novartis.17 The remaining authors declare no competing financial interests.
Correspondence: Laurent A. Decosterd, Division of Clinical Pharmacology – Laboratory, BH18 – Lab 218-226, Centre Hospitalier Universitaire Vaudois and University of Lausanne, CH-1011 Lausanne, Switzerland; e-mail: laurentarthur.decosterd@chuv.ch.