* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Enzyme Kinetics for Clinically Relevant CYP Inhibition
Discovery and development of direct thrombin inhibitors wikipedia , lookup
Neuropsychopharmacology wikipedia , lookup
Discovery and development of non-nucleoside reverse-transcriptase inhibitors wikipedia , lookup
Discovery and development of cyclooxygenase 2 inhibitors wikipedia , lookup
Pharmaceutical industry wikipedia , lookup
Neuropharmacology wikipedia , lookup
Prescription drug prices in the United States wikipedia , lookup
Prescription costs wikipedia , lookup
Drug discovery wikipedia , lookup
Theralizumab wikipedia , lookup
Discovery and development of proton pump inhibitors wikipedia , lookup
Discovery and development of integrase inhibitors wikipedia , lookup
Pharmacognosy wikipedia , lookup
Drug design wikipedia , lookup
Discovery and development of neuraminidase inhibitors wikipedia , lookup
Pharmacogenomics wikipedia , lookup
Plateau principle wikipedia , lookup
Discovery and development of ACE inhibitors wikipedia , lookup
Current Drug Metabolism, 2005, 6, 241-257 241 Enzyme Kinetics for Clinically Relevant CYP Inhibition Zhi-Yi Zhang* and Y. Nancy Wong Drug Disposition, Eisai Research Institute, Andover, MA 01810, USA Abstract: In vitro cytochrome P450 (CYP)-associated metabolic studies have been considered cost-effective for predicting the potential clinical drug-drug interactions (DDIs), one of the major attritions in drug development. The breakthroughs during the past decade in understanding the biochemistry of CYP-mediated biotransformation and molecular biology of CYP gene regulation in humans have provided the scientific bases for such endeavors in early drug development. In this review, the enzyme kinetics of CYP inhibitions is described, with the primary focus on the ones proven with clinical relevance, namely the competitive inhibition and mechanism-based inactivation (MBI). Competitive CYP inhibition, the most often detected reversible inhibition, is well understood and has been studied extensively both in vitro and in clinical setting. Recently, MBI has received increasing attention. It has been recognized that MBI could occur more often than anticipated, due in part to the redox cycling-allied enzymatic action of CYPs. As commonly as an irreversible inhibition, MBI would inactivate the target proteins, and thus would be generally considered of high potential for causing clinical DDI. Moreover, the reversible inhibitions other than the competitive, namely noncompetitive, uncompetitive and mixed, were also documented for the important drug-metabolizing CYP members, particularly CYP1A2 and CYP2C9. Finally, the unusual kinetic interactions, which did not follow the Michaelis-Menten (M-M) kinetics, were detected in vitro for the majority of drug-metabolizing CYP members, and manifested for CYP3A4. However, the clinical relevance of the interactions involving the unusual CYP kinetics has not yet been fully understood. Nonetheless, the reversibility and inhibitory potency should be considered as the major determinants of the clinical relevance, particularly in combination with the therapeutic exposure levels. With rapid expansion of knowledge and technology, the evaluation of the clinically relevant CYP-associated DDIs in vitro is not only desirable but also achievable. Key Words: Cytochrome P450, drug-drug interaction, enzyme inhibition, mechanism-based inactivation, atypical kinetics, in vitro-in vivo correlation, pharmacokinetics and clinic. INTRODUCTION The inhibition of metabolic enzymes, particularly cytochromes P450 (CYPs), has been recognized as the pivotal cause of the drug-drug interactions (DDIs) in clinic [1, 2]. With the increasing number of new therapeutic agents on the market and the therapeutic trend in drug combination, the DDI becomes one of the focal points for the drug development and clinical application [3, 4]. Fortunately, the breakthrough in understanding the mechanism of action for human drug-metabolizing enzymes, CYPs in particular, during the past decade, has shed some light on examining DDI in vitro to predict the DDI in clinic [5]. The methods for in vitro CYP-associated DDI studies have been well established, and the human enzyme preparations, from recombinant individual metabolic enzymes to freshly isolated primary hepatocytes, are commercially available. Thus the in vitro DDI studies, recently enforced by the US Food and Drug Administration (FDA) [6-8], are routinely performed during drug development in pharmaceutical companies. However, in vitro-in vivo correlation for DDI, though occasionally successful [9, 10], has not always been satisfactory, due in part to the intrinsic differences between the simplified non-physiological in vitro models and the *Address correspondence to this author at the Eisai Research Institute, Drug Disposition, 100 Research Drive, Wilmington, MA 01887; Tel: 978-6617242; Fax: 978-657-7715; E-mail: [email protected] 1389-2002/05 $50.00+.00 dynamic physiological human bodies, besides the ambiguity of in vivo CYP kinetic behaviors. Therefore, the reliability of in vitro DDI study is uncertain. On the other hand, some of the potential biases for the in vitro-in vivo correlations, as suggested by the recent studies [11, 12], are amendable, providing opportunities for further refinement of the predictive kinetic models. Although challenging, the improvement of the predictability of in vitro results by integrating the key in vivo factors such as pharmacokinetics (PK) characteristics, should be anticipated and encouraged. Therefore, we would like to, both from theoretic and practical prospective, provide our thoughts on the kinetic mechanisms of CYP inhibitionassociated DDI, which will hopefully enhance the predictability of the findings in DDI studies in vitro. The present paper is divided into two sections: the kinetics of CYP inhibitions in DDI and some practical concerns for the prediction of clinical DDI using in vitro data. A. Enzyme Kinetics in CYP-Mediated Biotransformation and CYP Inhibition Both Michaelis-Menten (M-M) kinetics and nonMichaelis-Menten kinetics (or atypical kinetics) have been detected for CYP-mediated drug metabolism and its inhibition. Though not yet convincingly proven, the mechanisms underlying these distinct kinetic behaviors are, at least empirically, believed to be due to the existence of divergent interactions between the drug molecules and the active © 2005 Bentham Science Publishers Ltd. 242 Current Drug Metabolism, 2005, Vol. 6, No. 3 Zhang and Wong site(s) of the enzymes. Thus, M-M kinetics would be simply elucidated using the model of single binding site, whereas atypical kinetics might be interpreted using the model of two or possibly more binding regions within the enzyme active site(s). 1. One-Binding Site (1) Substrate Alone One-binding site, or M-M hyperbolic kinetic model (Eq. 1), the most simple, however, the basis of the majority of the derived complex kinetic models later described, is historically applied to CYP-mediated metabolism. V = Vmax*S/(S + Km) = Vmax/(1 + Km/S) 1 Where V is reaction rate, and Vmax is the maximal, and Km is the substrate concentration at which the reaction rate is half of its maximum. S is substrate concentration. Interestingly, in contrast to the unequivocal definition, the calculation of Vmax may not always be consistent in literature [1316]. To avoid such a potential inconsistency, the classic enzyme kinetic concept is depicted here. E+S k1 k2 ES k3 E+P where E, S, ES and P are the concentrations of the enzymes, substrates, substrate-bound enzymes and enzyme products (metabolites), respectively. k1 is the association rate constant, and k2 and k3 are the dissociation rate constants from ES to E and S, and E and P, respectively. As shown in Scheme 1, 2 In steady state, k1*E*S = k2*ES + k3*ES, thus, ES = E*S*k1/(k2 + k3) 3 Michaelis constant (Km) is defined as Km = (k2 + k3)/k1 4 Eq. 3 becomes ES = E*S/Km, and Eq. 2 is converted to V = (k3/Km)*E*S As typical enzyme catalysis per se, ET is usually unchanged during the catalytic process, particularly in the reconstituted in vitro system (in the absence of mechanismbased inactivation). Therefore, differentiating the potential difference in defining Vmax may not be crucial, and both the connotations might be used interchangeably. However, for the prediction of metabolic enzyme-associated DDI in clinic in which the enzyme quantities may vary, the later connotation should be more applicable, as demonstrated by intrinsic clearance (CLint). CLint, the key parameter bridging enzyme kinetics and PK, is defined as Vmax/Km. In a well-stirred model with the assumptions of complete intestinal absorption and exclusive hepatic elimination, the intrinsic clearance can be directly associated to the area under the plasma-concentration-time curve of an orally administered drug (AUCpo), as shown in Eq. A (Equations for enzyme kinetics are listed in numerical order, while those for PK in alphabetic order.) [17]. Dosepo/AUCpo = fu*CLint or Scheme 1. The Classic Enzyme Catalytic Process. V = k3*ES thought to be the maximal conversion rate from substrate to product [16], then it should be regarded as kcat*ET (derived from Eq. 2, and the concentration of total active enzymes ET = ES + E. ET is approximately equal to ES when enzymes are saturated with substrates.). In that case, Vmax is no longer a constant, but the function of ET. 5 Therefore, the reaction rate (V) depends on the concentrations of both unbound enzyme (or the available active sites of enzymes) and substrate in steady state. Eq. 5 could be further converted to Eq. 1. When the enzymes are saturated by the substrates, the enzyme catalytic rate reaches the maximum (Vmax) and becomes substrate concentration-independent. The dissociation rate constant k3 would be viewed simply as kcat, the maximal catalytic rate constant. Therefore, if Vmax is considered the capacity of the enzyme catalysis, it could be interpreted as a parameter in a unit of enzymes, however, it might not be suitable in a unit of unpurified proteins, i.e., mg of proteins, as sometimes used in the literature. Thus, Vmax is, in this case, identical to kcat. On the other hand, if Vmax is AUCpo = Dosepo/(fu*CLint) A Where fu is the unbound drug fraction in the blood, and AUCpo is applicable for both that of time zero to infinity in a single dose and over a dosing interval in steady state in multiple doses. ET in vivo is controlled by protein synthesis and degradation. In steady state or equilibrium of these inverse processes, ET could be denoted as Ess, the concentration of total enzymes available for the metabolism in steady state. Thus, Vmax = kcat*Ess. However, such equilibrium or the level of Ess would be potentially modulated by a series of factors, particularly the exogenous effectors including enzyme inducers or inactivators. Therefore, in contrast to likely quasi-constant Vmax in vitro, Vmax and CLint in vivo (CLint = (kcat/Km)*Ess, and kcat/Km is defined as catalytic efficiency.) are subject to variation, which potentially affects AUCpo of drug substances in clinic (Eq. B). AUCpo = Km/kcat*(Dosepo/(fu*Ess)) B Therefore, the concept of Vmax as kcat*ET is used throughout the paper. CLint, or Vmax/Km, can be determined in vitro by applying the traditional biochemical plots, including M-M plot constructed by non-linear regression analyses, and those transformed from M-M plot such as Lineweaver-Burke or Eadie-Hofstee plot (Fig. 1) [18, 19], in the assays using different enzyme sources including recombinant human CYPs, human liver microsomal preparations (HLM) or even hepatocyte suspensions [20, 21]. Each of these enzyme sources has the intrinsic pros and cons, as discussed previously [22]. Thus, the selection of in vitro systems should facilitate the experimental goals, since different in vitro systems would provide, though still some overlapping, differential kinetic information. Enzyme Kinetics for Clinically Relevant CYP Inhibition Current Drug Metabolism, 2005, Vol. 6, No. 3 243 Fig. (1). Biochemical plots for the determination of kinetic parameters of CYP-mediated reactions. A, Lineweaver-Burk or double-reciprocal plot; B, Eadie-Hofstee plot. Double-reciprocal plot, though frequently used, may not be ideal, particularly compared to Eadie-Hofstee plot. The data points in doublereciprocal plot tend to be unevenly distributed thus potentially lead to the unreliable reciprocals of lower metabolic rates (1/V), which however dictate the linear regression curves therefore the apparent values of Km and Vmax. In contrast, the data points in Eadie-Hofstee plot are usually homogeneously distributed, and thus tend to be more accurate. Moreover, Eadie-Hofstee plot appears diagnostic of recognizing deviations from Michaelis-Menten kinetics, such as homotropic cooperation. (2) Substrate and Inhibitor In the presence of two different compounds interacting with the same enzymes, either compound would interfere with the binding, or more generally with the interactions of the other with the enzymes. In regard to metabolism, the resulting outcome is typically an enzyme, particularly CYP inhibition, i.e the essence of this review. The enzyme inhibitions in the arena of M-M kinetics are usually categorized into four different modes as shown in Scheme 2, namely competitive, uncompetitive, noncompetitive and 244 Current Drug Metabolism, 2005, Vol. 6, No. 3 Zhang and Wong mixed inhibition, with competitive inhibition being the prevalent and the uncompetitive, the rare [4]. E+S I 1 Ki EI Ks S Ks' 3 ES E+P I Ki' 2 ESI Finally, the mixed inhibition would be considered as the upshot of the inhibitor-enzyme binding with simultaneous two or more above mechanisms, particularly the combination of the competitive with the noncompetitive. To demonstrate the potential utilities of differentiating these M-M inhibition patterns for DDI studies, IC 50, the inhibitor concentration at which the reaction rates are suppressed by 50%, is further described here. The values of IC50 in different models could be calculated using Eqs. 10-13, as directly derived from Eqs. 1 & 6-9. Competitive: Scheme 2. Reversible Enzyme Inhibitions Follows MichaelisMenten Kinetics. Competitive inhibition, inhibitor binds only unbound enzyme (Pathway 1). Uncompetitive inhibition, inhibitor binds only substratebound enzyme (Pathway 2). Noncompetitive inhibition, inhibitor binds both unbound and substrate-bound enzyme at the site(s) different from the substrate-binding site (Pathways 1 and 2). Mixed inhibition , inhibitor binds, in a similar manner as the noncompetitive, but possibly at the site(s) overlapped with substrate-binding site, the unbound and the substratebound enzyme with potentially different affinities (Pathways 1, 2 and 3). The equations of reaction rate derived from these models are as follows: V = Vmax/(1+Km/S/(1+I/Ki)) 6 V = Vmax/(1+I/Ki+Km/S) 7 V = Vmax/(1+Ks/S)/(1+I/Ki) 8 V = Vmax/((1+I/Ki’)+(1+Ks/S)*(1+I/Ki)) 9 Eq. 6 is for the competitive, Eq. 7 for the uncompetitive, Eq. 8 for the noncompetitive, and Eq. 9 for the mixed inhibition model. I is the inhibitor concentration. Ki is the inhibition constant, while Ki' is the noncompetitive inhibition constant for the binding of inhibitor with the substrate-bound form of enzyme (ES) in the mixed inhibition model. K i and Ki' may simply be viewed as the indicators of the binding affinities between inhibitor and enzyme, similar to Km as that of the binding affinity between substrate and enzyme. Ks, the dissociation constant of the enzyme-substrate complex, is approximately equal to Km (k3 << k2). In competitive inhibition, the inhibitor usually binds to the enzyme at the same site as the substrate, or possibly different site(s), and subsequently blocks the substrate binding. In contrast to clearly defined competitive inhibition, uncompetitive and noncompetitive inhibitions appear occasionally ambiguous [23]. The uncompetitive inhibition might be viewed as that the inhibitor binds effectively to the enzyme only in the substrate-bound form; while in the noncompetitive manner, the inhibitor binds to the enzyme in both unbound and substrate-bound forms with the same binding affinity (Ki), or in other words, without apparent preference. The enzyme inhibition by the polyclonal antibodies could be often in a noncompetitive manner [24]. IC50 = Ki*(1 + S/Km) 10 Uncompetitive: IC50 = Ki*(1 + Km/S) 11 Noncompetitive: IC50 = Ki 12 Mixed: IC50 = Ki*(1 + S/Km)/ (1 + (Ki/Ki')*(S/Km)) 13 IC50 usually depends on the substrate concentrations, with the exception in the noncompetitive model, and likely higher than Ki as shown by Eqs. 10 & 11. When competitive inhibition occurs, IC50 rises. In other words, the degree of inhibition reduces if the inhibitor concentration is fixed, along with the increase of substrate concentration. However, such a scenario would be a reverse for uncompetitive inhibition, in that, the higher the substrate concentration, the lower the IC50 value, since the inhibitor effectively binds only to the substrate-bound form of the enzyme (ES). In contrast, IC50 in noncompetitive inhibition is a constant and equal to Ki (Eq. 12), because the inhibitor binds only to the enzyme at the sites different from the substrate-binding site. In other words, the enzyme-inhibitor binding is independent of the enzyme-substrate binding. Moreover, if substrate concentration is low (or S/Km << 1), as often occurred in vivo, IC50 values would be similar, when calculated using different inhibition models excluding the uncompetitive (Eqs.10, 12 & 13). The competitive model thus might be chosen for IC50 determination if the mechanism of a reversible inhibition is unclarified in vivo. However, the inhibition would be potentially more pronounced in a noncompetitive than that in a competitive manner, if the substrate concentration available to the enzymes is indeed substantial and comparable to Km, as likely the case when the substrate is actively extracted by the metabolic tissues [25]. Therefore, the clinical CYP-associated DDI could be potentially underestimated if noncompetitive inhibition is misconceived as competitive inhibition. On the other hand, uncompetitive inhibition, besides the rare occurrence, tends to be less significant since the inhibitor would exhibit diminished inhibitory effects because of much higher than Ki IC50 values at low substrate concentration (or Km/S >> 1) (Eq. 11). Therefore, noncompetitive inhibition, though unlikely common, might be possibly relevant in DDI in clinic. Interestingly, the mechanisms of reversible inhibitions by a given inhibitor may also elicit substrate dependence even if the substrates and inhibitor presumably bind to the same enzyme active site (but possibly at different regions and/or with different binding orientations within the active site). One such example is the differential inhibitions of CYP1A2mediated O-deethylation of 7-ethoxyresorufin (EROD) and 7-ethoxycoumarin (EOC) by α-naphthoflavone (αNF) [26]. The inhibitions of EROD and EOC by αNF are very potent Enzyme Kinetics for Clinically Relevant CYP Inhibition due to the tight-binding between CYP1A2 and αNF, with the apparent Ki values of 1.4 and 55 nM in reconstituted system containing the purified recombinant CYP1A2, respectively. However, EROD inhibition by αNF was delineated as the competitive, whereas EOC inhibition by αNF, the noncompetitive, as proposed by Cho et al. [26]. The equation for IC50 determination in the mixed model (Eq. 13) appears complex. It could be easily converted to the other IC50 equations following M-M kinetics (Eq. 10-12). For instance, if the competitive is the predominant component of a mixed inhibition, or Ki << Ki', (1 + (Ki/Ki')*(S/Km)) in Eq. 13 is approximately equal to 1, Eq. 13 becomes Eq. 10. If the noncompetitive is the predominant component of a mixed inhibition, or Ki = Ki', (1 + (Ki/Ki')*(S/Km)) is equal to (1 + S/Km), Eq. 13 becomes Eq. 12. Finally if the uncompetitive is the predominant mechanism, or Ki >> Ki', Eq. 13, to be better illustrated, should be first converted to IC50 = Ki'*(1 + S/Km)/(Ki'/Ki + S/Km), after both the numerator and denominator of the equation are multiplied by Ki'/Ki. Thus, the denominator or (Ki'/Ki + S/Km) in the transformed Eq. 13 is approximately equal to S/Km, and Eq. 13 becomes IC50 = Ki'*(1 + Km/S), or Eq. 11. Ki' here is the same as Ki in Eq. 11, since inhibitor in this case effectively binds only the substrate-bound enzyme [4]. Although numerous plot methods are available [15], it may not be straightforward to classify the enzyme kinetics of reversible inhibitions. Therefore, the computational simulation methods, such as simultaneous nonlinear regression (SNLR), should be considered as the alternative, and they Current Drug Metabolism, 2005, Vol. 6, No. 3 245 are being increasingly appreciated and applied [28, 29]. As demonstrated by Fig. 2, the experimental data could be fitted using SNLR with all possible kinetic models, thus the potential inhibitory mechanism assigned would be unlikely biased. Meanwhile, the traditional plots, such as the Dixon plot, are still often required for visually inspecting the inhibitory mechanism, particularly when more than one kinetic models are suggested using SNLR [29]. 2. Two-Binding Sites Non-M-M kinetics or atypical CYP kinetics, a topic in academia for quite sometime, has recently received some attention in pharmaceutical industry [30-34]. The mechanisms of such reversible kinetics, or so-called 'allosterism', have been substantially studied and generally associated with the assumed existence of two or more substrate binding regions in enzyme active site [32, 35-37]. Atypical kinetics for CYPs in vitro appeared more common than expected in the past when our knowledge in CYP actions was limited, thus the classic reversible kinetics, e.g. M-M kinetics, in most cases was assumed. CYP3A4, among the major drugmetabolizing CYP forms, is most known for its atypical kinetic behavior [38, 39]. However, besides some of the minor CYP members [40, 41], CYP2C9 and CYP1A2 were also proven to exhibit potential atypical kinetics [32, 42-44]. As drug metabolism per se, the three major CYP forms would likely be responsible for the metabolism of more than half of the therapeutics currently on the market [5]. The bodies of evidence for the potential two (or multiple) substrate-binding sites were mostly derived from the numerous enzyme kinetic studies, especially for CYP3A4 and CYP2C9 Fig. (2). 3D plots of simultaneous nonlinear regression (SNLR). A, by fitting the real data; B, by fitting the data applying a presumed competitive model. Apparently, the competitive model was quite appropriate in this case (p < 0.01). However, if two or more kinetic models appear applicable using SNLR, the standard biochemical plots such as Dixon plot should be considered since SNLR analysis provides little or limited information for the mechanism of enzyme inhibition. 246 Current Drug Metabolism, 2005, Vol. 6, No. 3 Zhang and Wong [32, 39, 42, 43]. Atypical CYP kinetics has recently been extensively reviewed [4, 45, 46]. Therefore, a brief description with the focus on clinical DDI is provided. (1) Substrate Alone Atypical CYP kinetics associated with substrates only could be simplified as biphasic kinetics owing to the assumption of two (or possibly more) substrate-binding sites, as suggested by an elegant work from Korzekwa et al. [37]. These types of kinetics, exhibiting characteristic concentration dependence, are most known for substrate inhibition and autoactivation. The substrate inhibition is not uncommon in vitro [47]. The mechanism for such phenomenon is, though still not clear, generally attributed again to two or more binding regions in CYP activity sites [47]. Between two binding regions, one site is presumed 'productive', and the other 'unproductive'. The binding at the unproductive site is thought to suppress the productivity of the enzyme, but unnecessarily due to the repression of the substrate binding at the productive site. The substrate could thus behave as an inhibitor upon the regions it binds. Consequently, the interactions among the molecules of substrates within the enzyme active site would be somewhat similar to the competitive or possibly uncompetitive inhibition following M-M kinetics (Eq. 14). However, such interactions would not likely resemble the noncompetitive one due to the presumed involvement of enzyme active sites. V = Vmax/(1+S/Ki+Km/S) 14 Several points are worth noting. (i) Ki is usually much greater than Km, as the substrate preferably binds to the productive region than the unproductive region. At low substrate concentrations (S <<Ki), the inhibitory effect is likely unobservable, thus the enzyme reaction is rather 'typical'. However, the inhibitory effect begins to elicit along with the rise of substrate concentration, and becomes apparent when the concentration approaches the level at which the activities start to be saturated if in a typical hyperbolic kinetics. (ii) The concentrations of substrate and 'inhibitor' are the same within the enzyme active sites. (iii) The largely different affinities for the binding of substrate to two different regions within the enzyme active site and the identical concentrations of substrate and 'inhibitor' would differentiate substrate inhibition from typical competitive inhibition following M-M kinetics. In competitive inhibition, inhibitor concentration is totally independent to substrate, thus even a weak inhibitor (Ki > Km) may markedly suppress the substrate-enzyme binding if the inhibitor concentration is sufficiently high. Therefore, the substrate inhibition would most likely resemble the uncompetitive in M-M kinetics, as shown by Eq. 14 (or Eq. 7). Misconception of substrate inhibition would potentially lead to underestimate the true values of Vmax and Km. However, the resulting biases for CLint (Vmax/Km), attenuated by the reductions of both Vmax and Km, may not be necessarily significant [47, 48]. Moreover, substrate inhibition would be unexpected in clinic, as the near enzyme-saturating levels of drug concentrations in patients are unfeasible. Autoactivation is somewhat opposite to substrate inhibition, which is often detected by a sigmoidal concentrationactivity plot (Fig. 3). Autoactivation is usually referred to the homotropic cooperation, in differentiating from the heterotropic cooperation that involves two (or more) different Fig. (3). Homotropic activation in standard biochemical plots. A, Michaelis-Menten plots; B, Eadie-Hofstee plots. Both plots would elicit the potential cooperation. However, homotropic cooperation would be more evidently shown by Eadie-Hofstee plots, which therefore should be selected for the detection of atypical kinetics, particularly for autoactivation. Enzyme Kinetics for Clinically Relevant CYP Inhibition Current Drug Metabolism, 2005, Vol. 6, No. 3 247 ligands. Again, the phenomenon could be simply interpreted by applying the model with two different substrate-binding sites, as demonstrated in Eq. 15 [37]. prediction of clinical DDI, the goal of this review. Therefore, atypical CYP kinetics beyond the aforementioned will not be discussed further. V = (Vmax2/(Km1*Km2) + Vmax1/(Km1*S))/(1/(Km1*Km2) + 1/(Km1*S) + 1/S2) 15 The bodies of evidence on atypical CYP kinetics, in general, are substantial in vitro, however, very limited in vivo, and if so, almost exclusively from experimental animals [53, 54]. Atypical kinetics is often concentration dependent, requiring a large concentration range, particularly high concentrations of substrates and/or modulators to be detected [33]. In this regard, although the heterotropic cooperation between CYP3A4 substrates and endogenous steroids is well acknowledged [55, 56], the extrapolation of such in vitro findings to that in human bodies should be cautious in that the concentrations and ranges of the drugs and particularly the modulators such as the endogenous steroids are much lower and narrower in vivo, compared to those studied in vitro, respectively [55, 57]. Therefore, even with the supporting ex vivo results [58, 59], atypical CYP kinetics and the potential relevance to DDI in clinic, though suggested by the recent study [42], are largely uncertain, which should be further investigated. The subscripts '1' and '2' of the parameters (Vmax and Km) are denoted to the individual binding sites (or singly- or doubly-ligated binding state). It should be worthwhile to show the relationship between Eq. 15 and Eq. 14, since both are derived from the same model, one substrate with two binding sites. After being carefully examined, Eq. 14 appears to be a special form of Eq. 15 in which Vmax2 = 0 and Km2 = Ki. Therefore, Eq. 15 could be written as V = (Vmax/(Km*S))/(1/(Km*Ki) + 1/(Km*S) + 1/S2), thus readily converted to V = Vmax/(S/Ki+ 1 + Km/S) or Eq. 14. Eq. 15, though considered as a simplified version of those for the two-site model, is still complex. Thus, if the individual kinetic parameters of the models are determined, the computational methods, such as SNLR, have to be employed [37, 48]. Nevertheless, like the majority of atypical kinetics, sigmoidal kinetics may be common in vitro, however uncertain in vivo [49]. (2) Substrate and Effector (Modulator) Heterotropic cooperation, often detected for CYP3A4mediated reactions in vitro [50], involves two different ligands, substrate S and activator B. The rate of biotransformation of the substrate is enhanced by the existence of activator [46]. Though more than one potential mechanism might be speculated, the putative two-binding site model appears prevailing [37, 51]. The mathematical expressions for heterotropic cooperation in two-binding site kinetic models are usually complex, as exemplified by Eq. 16. V = Vmax*(α*KB + β*B)/[(α*KB + B) + α*(KB + B)*(Km/S)] 16 Eq. 16 is directly converted from the previously reported equation as follows [43]. V = Vmax*S/{Km*(1 + B/KB)/[1 + (β*B)/(α*KB)] + S*[1 + B/(α*KB)]/[1 + (β*B)/(α*KB)]} 3. Irreversible CYP Inhibition or Mechanism-Based Inactivation (MBI) Three types of mechanism-based inactivation of CYP were reported: (i) inhibitor covalently binds to enzyme apoprotein, (ii) covalently binds to prosthetic heme, or (iii) chelates to the heme or tightly binds to the apoprotein. The third type of inhibition is, strictly speaking, still noncovalent and pseudo-irreversible [60]. A CYP inhibition has to meet a number of criteria to be classified as MBI [61], and some of the important ones are briefly described: (i) Time-dependent inactivation occurs at the condition supporting the enzyme reactions, e.g., in the presence of NADPH for CYP-mediated reactions. This is the key characteristic of MBI. (ii) Inactivation of CYP is practically irreversible, thus the removal of inhibitor by dialysis or filtration would not recover the enzyme activities. (iii) A conversion from inactivator to a reactive intermediate during the enzyme catalyses should take place, and thus be proposed. (iv) The rate of inactivation typically follows the hyperbolic kinetic pattern, similar to MM kinetics in enzyme catalyses. In Eq. 16, B is activator concentration, and KB the binding constant for the activator. α is the relative alteration of Km, and β the relative alteration of Vmax resulting from the activator binding, while α < 1 and β > 1 would be expected in heterotropic activation. CYP mechanism-based inactivators (MBI, which is used both for mechanism-based inactivation and the inactivator), though not yet been fully delineated structurally, often confer terminal unsaturated carbon atom, alkylamino, furan or furano-pyridinyl group, or sulphur-containing aromatic systems such as thiophene [60, 62-65]. The mathematical expressions would indeed become more sophisticated for some of the heterotropic interactions other than the activation, i.e. so-called differential kinetics among substrate, effector and enzyme [45, 51]. Nonetheless, despite the successes in interpreting the unusual CYP kinetic behaviors, the existing two-binding site models for atypical kinetics including heterotropic interactions and associated mathematical expressions tend to be elusive and static, with little or no concern of other possible origins, i.e., protein conformation or motion, CYP catalytic mechanism, and thus are still subject to further refinement [50, 52]. Moreover, due to the empirical nature, these models tend to be retrospective rather than prospective, thus with limited value for the In contrast to atypical reversible CYP inhibitions, MBI appears to occur both in vitro and in clinic [65-68]. All human hepatic drug-metabolizing CYP members, including the abundant forms such as CYP3A4, CYP2C9 and CYP1A2 [64, 69, 70], the important polymorphic forms, i.e., CYP2C19 and CYP2D6, and the minor members, i.e., CYP2A6, CYP2B6 and CYP2E1, are subject to MBI [62, 65, 71-73]. However, MBI, if not intentionally sought, could be easily overlooked, even in vitro. Reconstituted in vitro systems may not always be apt for detecting time-dependent events including MBI, due in part to the limited thermal stability of CYP activities particularly during an extended assay period. Moreover, MBI is the substrate, thus the competitive 248 Current Drug Metabolism, 2005, Vol. 6, No. 3 Zhang and Wong inhibitor for the other substrates of the target enzyme. The component of competitive inhibition may, if potent, obscure the component of MBI. Fortunately, in accordance with the explicit concept as described and schematically shown (Scheme 3), the methods for detecting MBI are readily available. E+I k1 k2 EI k3 k4 [EI] E-I zero-order, while the degradation is considered to be of the first-order, resulting in steady state, -dE(t)/dt = R – Kdeg*Ess = 0, or R = Kdeg*Ess Where R is rate of protein synthesis, Ess the concentration of active enzyme in steady state, and Kdeg the overall rate constant of protein degradation due to both the endogenous (kE) and exogenous forces such as MBI (kI). kI could be practically substituted with kobs determined in vitro, and the equation could be revised to R = (kE + kI)*Ess-I, or Ess-I = R/(kE + kI) k5 In the absence of MBI (kI = 0), Ess-0 = R/kE, thus E+P Ess-I/Ess-0 = [R/(kE + kI)]/(R/kE) = kE/(kE + kI) Scheme 3. Mechanism-based Inactivation. Where EI is the inactivator and enzyme at the initial noncovalent interaction stage. [EI] is metabolic intermediate complex (MIC). E-I is the inactivated enzyme covalently bound with the inactivator. The rate of enzyme inactivation by MBI (-dE(t)/dt) could be depicted as -dE(t)/dt = kobs*E(t) = kinact*I*E(t)/(I + KI) where kobs is the observed first order inactivation rate constant. kinact is the maximal inactivation rate constant, while KI irreversible inhibition constant that might be viewed as the inactivator concentration at which the inactivation rate is the half of the maximum. I is inactivator concentration. E(t) is the concentration of active enzymes at some time t. The relationship among k obs, KI, and kinact is illustrated in Eq. 17 [74]. kobs = kinact*I/(I + KI) 17 kobs ≈ (kinact/KI)*I, if I << KI, as often expected in the clinical setting. KI and kinact, according to the kinetic model in Scheme 3, could be described as KI = (k2 + k3)*(k4 + k5)/[k1*(k3 + k4 + k5)] KI ≈ (k2 + k3)/k1, if k4 + k5 >> k3 18 kinact = k3*k4/(k3 + k4 + k5) kinact ≈ k3, if k4 >> k3 + k5 19 As KI and kinact, shown in Eqs. 18 & 19, resemble Km (Eq. 4) and kcat in M-M enzyme kinetics, respectively, it would be reasonable to consider that MBI of CYP is the 'unproductive' version of CYP catalysis. Therefore, k inact/KI, the potency of MBI, might be simply viewed as the unproductive counterpart of kcat/Km, the catalytic efficiency in M-M kinetics. The important MBI constants (kobs, KI, and kinact) could be estimated applying the plotting methods shown in Fig. (4) [62, 64]. In contrast to continuous decreases of active enzyme quantities during MBI in vitro, the degradation of enzymes comprising that caused by MBI in vivo, is compensated by the protein synthesis, thus the equilibrium or steady state between enzyme degradation and production could be reached. Protein synthesis is usually considered to be of where Ess-I and Ess-0 are the steady sate concentrations of active enzymes in the presence and absence of MBI. Obviously, the declined level of active enzymes due to MBI in steady state would result in reduced overall metabolic capacity, thus altering the associated PK characteristics, as demonstrated by the elevation of the area under the plasma-time-concentration curve (AUC) in a well-stirred model with the assumptions of complete intestinal absorption and exclusive hepatic elimination (Eq. C). CLint = (kcat/Km)*Ess, and AUCpo = Dosepo/(fu*CLint), thus AUCpo = Dosepo*Km/(fu*kcat*Ess) AUCpo-I/AUCpo-0 = [Dosepo*Km/( fu*kcat*Ess-I)]/[Dosepo*Km/( fu*kcat*Ess-0)], or AUCpo-I/AUCpo-0 = Ess-0/Ess-I = (kE + kI)/kE, or AUCpo-I/AUCpo-0 = 1 + kI/kE C where AUCpo-I and AUCpo-0 are the AUCs of orally administered drugs in the presence and absence of MBI. As demonstrated by Eq. C, the increase of AUC due to MBI would largely depend on the ratio of MBI rate constant (kI) to endogenous protein degradation rate constant (kE). Both kI and kE would be possibly estimated. kE in humans has been suggested for a few CYP members including CYP3A4 [7579]. A kE of approximately 0.693/day (or 0.00048/min), corresponding to a half-life ( t1/2) of 1 day (averaged from 0.5 to 2 days), might be projected for CYP3A4 degradation. [7578]. In addition, the concentration of human hepatic CYP3A4 (Ess) was suggested to be about 5 nmol/g liver [79]. However, such data associated with CYP degradation in humans were somewhat indirect, and the rate of protein degradation is apparently conditional, i.e., increase during MBI. Therefore, the existing literature information of CYP degradation should be cautiously applied. Nevertheless, MBI might be one of the major causes for the clinical DDIs [80-82], which has been potentially overlooked in the past [12]. B. In vitro-In vivo Correlation: A Practical Consideration The applicability of the prediction of CYP-associated DDI in clinic using in vitro approaches has been demonstrated to be appreciable in some cases, but challenging for the others [3, 21, 83-85]. This is indeed not surprising since Enzyme Kinetics for Clinically Relevant CYP Inhibition Current Drug Metabolism, 2005, Vol. 6, No. 3 249 Fig. (4). Plots for kinetic characterization of CYP mechanism-based inactivation (MBI). A, for the estimation of kobs; B, for the determination of kinact and KI. It should be worth noting that enzyme inactivation is kinetically assumed in a standard hyperbolic manner. Therefore, the similarities between MBI and M-M kinetics are apparent, as kobs, kinact, and KI in MBI appears the counterparts of V/E (V per unit of enzyme), kcat, and Km in M-M kinetics, respectively. CYP inhibition in vivo is governed by multiple factors, some of which would not be easily envisioned or accurately evaluated in vitro, i.e. CYP induction following the inhibition, or simultaneous and/or sequential metabolism in different metabolic tissues. With the concerns of the in vivo variables, a variety of empirical and physiologybased models for the prediction of PK alterations due to CYP inhibition using in vitro data are emerging [12, 86, 87]. 250 Current Drug Metabolism, 2005, Vol. 6, No. 3 Zhang and Wong Therefore, only a few crucial in vivo variables directly affecting the predictability of in vitro data and the equations associated with these variables are presented. In addition, some factors, though potentially important, may also be excluded in the following context, as either too intricate to be rationalized and often evaluated on a case-by-case basis, or deviated from the scope of this review, e.g. CYP inhibitions by multiple inhibitory species including the metabolites [88-90] or involvements of metabolism-associated proteins other than CYPs such as active transporters [91, 92]. DDIs in clinic have been simply categorized into six common patterns, based mainly on the types of CYP interactions and the sequences of drug co-administration [93]. Moreover, as have recently summarized by Levy and his colleagues [2], both reversible and irreversible CYP inhibitions occur in an inhibitor dose (concentration)-dependent manner, in the clinical setting. Therefore, to quantitatively predict the clinical DDIs using in vitro approaches, though highly challenging, should be considered practically achievable. 1. Multiple Enzymes In vivo hyperbolic kinetic model with a fraction being competitively inhibited by the concomitant. Thus, it does not appear practically straightforward to ascertain the actual value of fm, particularly in vivo, apparently limiting the applicability of Rowland and Matin equation (Eq. D). Nevertheless, this equation has been, though rather infrequently, applied for in vitro-in vivo correlation of CYP inhibition, and proven of value, especially for competitive inhibition in a retrospective manner, as demonstrated by the clinical study of the effect of sulphaphenazole on tolbutamide metabolism [94, 95]. If the metabolic enzymes responsible for the drug biotransformation would be both hepatic and extrahepatic, the situation would become more complex. The extrahepatic tissues particularly intestine, kidney and lung are known to express quantifiable metabolic enzymes, including the ones tissue-specifically expressed [96-98]. However, other than the concerns for the intestinal CYP metabolism and/or inhibition as part of the first-pass effect [99, 100], DDI, CYP inhibition in particular, involving both hepatic and extrahepatic enzymes, has not been generally addressed due in part to the complexity of the processes and the lack of appropriate in vitro model systems. Where fm is the fraction of the dose eliminated by the enzyme being inhibited in the absence of inhibitor. I is the inhibitor concentration, while Ki inhibition constant. It is often desirable for the therapeutic agents to be converted by several enzymes in humans. Such a multiple enzymes-mediated metabolism tends to attenuate the ramification of enzyme inhibition, particularly in a competitive manner, in clinic [101, 102]. On the other hand, metabolism, in fact metabolic inhibition in the small intestines, could be substantial for orally administered drugs, particularly the substrates of the enzymes highly expressed in the enterocytes, i.e., CYP3A4 [103]. Oral co-administration with a CYP3A4 inhibitor may consequently enhance oral bioavailability and the potential of DDI for the drugs primarily metabolized by CYP3A4. Therefore, the enzyme suppression in gut provides the common ground for both risks and benefits, demonstrated by the severe elevation of DDI potential due to the co-administration of mibefradil in clinic [104] and the possible promise for the development of orally available paclitaxel [105]. Eq. D could be transformed to 2. CYP Inactivation In vivo AUCpo-i/AUCpo-0 = (1 + I/Ki)/(1 + I/Ki - fm*I/Ki) As mentioned earlier, the inhibitory effect in vivo would likely manifest if MBI is the major component of overall enzyme inhibition [106, 107]. This could be illustrated by the equation for the alteration of AUC (Eq. E) proposed by Hall and his colleagues [12, 108]. In the model that Eq. E was derived from, one of the responsible enzymes in the liver, was affected by MBI. The drugs, particularly those undergone both phase 1 and 2 enzymes-mediated, or even just CYPs-mediated conversions, are likely metabolized by several enzymes either simultaneously or sequentially in vivo. Meanwhile, the enzyme inhibition in vivo may involve one or a few of these responsible enzymes. Therefore, the impact on the overall metabolism in vivo due to CYP inhibition should be weighted, as described by Rowland and Matin (Eq. D) [94], with the assumptions of oral administration, linear (or firstorder) PK, complete absorption (or fa ≈ 1; fa, the fraction absorbed from gut to the portal vein), and prevailing hepatic metabolism with negligible non-hepatic clearance. AUCpo-i/AUCpo-0 = 1/[fm/(1 + I/Ki) + 1 – fm] D Apparently, the AUC alteration by the enzyme inhibition, as shown by the ratio, is largely controlled by the fraction of the metabolism being inhibited (fm), the inhibitor concentration (I), and the inhibition potency (Ki). The determining variable in Eq. D appears to be the inhibitor concentration I, since fm might be viewed as a pseudo-constant under certain circumstances, i.e., an oral drug with a fixed dose metabolized by a limited number of abundant CYP members. However, it is important to mention that fm often varies along with the drug concentrations (or drug doses) due to potentially different kinetic characteristics, such as substrateenzyme binding affinities (Km) or catalytic efficiencies (Kcat/Km), among the responsible enzymes, besides the interindividual variations in enzyme activities and quantities. Furthermore, due to the potential existence of atypical kinetics, particularly for the CYP3A4-mediated [56], the effect on drug metabolism by a concomitant would not always be appropriately estimated using an assumed AUCpo-I/AUCpo-0 = 1/{[fm/(1 + kI/kE)] + 1 – fm} E If both reversible inhibition and MBI occur simultaneously, the ramification could be additive. The changes of AUC due to reversible (usually the competitive) and MBI might be postulated using Eq. F derived from Eqs. D & E. AUCpo-iI/AUCpo-0 = 1/{[fm-i/(1 + I/Ki)] + [fm-I/(1 + kI/kE)] + 1 – fm-i – fm-I} or AUCpo-iI/AUCpo-0 = 1/{[fm-i/(1 + I/Ki)] + [fm-I/(1 + kinact*I/((I + KI)*kE)] + 1 – fm-i – fm-I} Enzyme Kinetics for Clinically Relevant CYP Inhibition If I << KI (and often I << Ki), as usually occurred in the clinical setting, AUCpo-iI/AUCpo-0 = 1/{[fm-i/(1 + I/Ki)] + [fm-I/(1 + (kinact/kE)*I/KI)] + 1 – fm-i – fm-I} F AUCpo-iI is the AUCpo, in the presence of both the components of reversible and irreversible inhibition. fm-i is the fraction of the dose being eliminated by the enzymes inhibited by the competitive inhibitor, and fm-I is the fraction of the dose being eliminated by the enzymes inactivated by MBI, in the absence of inhibitors. The terms other than fm-i and f m-I in Eq. F were previously described. However, Eq. F, though might be theoretically sound, clearly needs to be validated in the real case studies. When the competitive inhibition and MBI occur for the same enzyme, fm-i and fm-I should be identical. kE is again the rate of CYP degradation (kdeg) under normal physiological conditions. The values of kE, as mentioned earlier, are expected to be 0.5-2/day, or 0.00035-0.0014/min, for hepatic CYP3A4 in humans, and possibly some of the other CYP members [77, 109]. kE thus appears much smaller than kinact, as the latter, in the case of MBI of CYP3A4, are usually in the range of 0.05-1/min [12]. Therefore, in contrast to likely comparable Ki and KI, often in the order of micromolar concentrations, the values of kinact and kE tend to be markedly disparate, resulting in the ratio of kinact/kE around or often higher than 100, thus the much greater (kinact/kE)*I/KI than I/Ki in Eq. F. Therefore, though competitive inhibition and MBI should, in principle, take place simultaneously [110, 111], the change of AUCpo due to both the inhibitory mechanisms would be primarily attributed to MBI, the usual cases in the clinical setting. For example, CYP3A4 inhibition by erythromycin was earlier considered competitive, however, mechanism-based recently, since the PK alteration associated with the enzyme inhibition would not be reasonably explicated unless MBI is the predominant, if not the sole, mechanism [106, 107]. MBI appeared to occur more frequently than expected in the past, particularly for the drugs that form reactive intermediates in vitro such as tamoxifen, diltiazem and erythromycin [108, 111-114]. One of the potential indications for MBI, besides the nonlinear PK, would be the prolonged effect after the concurrent discontinuation or removal from the body. 3. Inhibitor Concentration In vivo The concentration of an inhibitor within the active site of enzyme, where DDI takes place, dictates the inhibitory effect, regardless of the inhibitory mechanism. However, such a concentration has been often assumed as the one in the systemic circulation. Besides the apparently higher exposure level of the cells in the gut wall during the oral administration of drugs, the intracellular drug concentration in the metabolic tissues could be quite different from the blood concentration, particularly if active cellular uptake or secretion is involved [115, 116]. The difference between the drug concentration in the tissues and that in the blood could be depicted by tissue-to-plasma (or tissue-to-blood) concentration ratio (Kp). For instance, high Kp values for azithromycin, a macrolide antibiotic, were suggested in humans, while determined in rats to be in the range from 40 to 80 in several tissues, including liver [117]. There are many controlling factors for Kp, mostly associated with the physiochemical properties of the drugs and the interactions Current Drug Metabolism, 2005, Vol. 6, No. 3 251 between the drugs and the host cells [25, 118]. Apparently, Kp of human tissues is usually unavailable and unlikely generalized for structurally diversified therapeutic agents. Therefore, in the absence of direct implication, i.e. the involvement of active cellular uptake or efflux in metabolic tissues, the concentration available to the liver metabolic enzymes might have to be simply assumed as the one in the circulation. However, the concentrations in the circulation and that in the metabolic organs, even without the involvement of active cellular processes, are unlikely identical, which could be in part elucidated by the model proposed by Ito and his colleagues. (Eq. G) [119]. Iin = I + ka*Dose*fa/Qh G Where Iin is the inhibitor concentration at the inlet of liver (or the hepatic artery), while I is the inhibitor plasma concentration after an oral administration. ka is the absorption rate constant, and fa, the fraction absorbed from gut to the portal vein. Qh is the hepatic blood flow. If the unbound fraction of inhibitor in the blood (fu) is known, the equation could be expressed as Iin, u = (I + ka*Dose*fa/Qh)*fu H Where Iin, u is the unbound fraction of Iin. ka, fa, and Qh are assumed to be 0.1 min-1, 1 (completely absorbed), and 1610 ml/min, respectively. In this model, the blood-to-plasma concentration ratio (RB) is assumed to be 1, and the inhibitor concentration available to the enzyme is presumed to be the one in the blood entering the liver (Iin or Iin, u). One of the utilities of the model is the prediction of clinical CYP inhibition-associated DDI, as Iin, u/Ki (or Iin/Ki), higher than I/Ki, the widely accepted term for DDI prediction [6, 119], would be relatively conservative, hence leads to a minimal false-negative prediction. The concentration available to the metabolic enzymes is very difficult to determine and unlikely generalized because the physiochemical properties of therapeutics are diverse, and so are the interactions of these chemicals with the hosts at the cellular and molecular level. Therefore, it may not be fully appropriate to employ the intracellular concentration in metabolic tissues, even if determined, as the drug concentration available to the metabolic enzymes due to the potential influences of cell physiology and functions such as organelle sequestration and/or transportation. For example, the metabolic enzymes are either cytosolic or intracellular membrane-associated, while the majority of monooxidases including CYPs and some of the conjugative enzymes such as UDP-glucuronosyltransferases (UGTs) are endoplasmic reticulum (ER)-bound. Moreover, the locations of the ERassociated metabolic enzymes (or their catalytic domains) could be distinctive, i.e., the cytosolic CYP versus the lumenal UGT active domains [120]. It has been suggested that the ER lumenal catalytic domains would be less accessible for their substrates/co-substrates/inhibitors, as supported by the well-known activity latency of UGTs [121]. Therefore, enzymes at different intracellular locations, such as CYPs and UGTs, may interact with drugs at different effective concentrations in vivo. It would be worthwhile to mention that the drug concentration in the in vitro metabolic systems, particularly 252 Current Drug Metabolism, 2005, Vol. 6, No. 3 Zhang and Wong that reconstituted without cellular matrices and functions, might not represent the enzyme systems in vivo. 4. Non-Specific Binding As one of many determinants for the drug effective concentration, the non-specific binding to proteins and lipids has been known for over 40 years [122], and often viewed as the variable to be taken into account for the prediction of DDI [123-125]. Interestingly, the effectiveness of considering non-specific binding in enzyme assays has been suggested depending on the drug physiochemical properties, thus the considerations of protein binding appeared helpful for the acidic, while fortuitous for the basic agents for the in vitro-in vivo correlation [126]. Moreover, the effect of nonspecific binding is thought to be concentration-dependent [124, 127], principally in a hyperbolic saturation manner (Eq. I) [124]: fu = Kd/(B + Kd) I where fu is the unbound fraction; Kd, dissociation constant; and B, the concentration of binding component such as CYP substrate or inhibitor. The effective concentration of the substrates or inhibitors available for CYPs is generally viewed as the unbound portion thus the fu*B being in steady state. The non-specific bindings would thus exert impacts on both biotransformation and enzyme inhibition [125, 127]. The apparent kinetic parameters determined in vitro are clearly affected by the non-specific bindings and could be rectified applying the fraction of free drugs (fu) if desired [128, 129]. One such example is the estimation of Km or Ki using the apparent value detected in vitro (Eqs. J & K). Km = Kmapp *fu Ki = Kiapp *fu J K where Km is Michaelis-Menten constant, and Kmapp the apparent Km detected in vitro. Similarly, Ki is inhibition constant, and Kiapp the apparent Ki detected in vitro. It would thus be beneficial to understand the non-specific bindings in vivo, particularly for the prediction of clinical DDIs using in vitro data. The potential divergences among the non-specific bindings of the concomitants may affect DDI outcomes. For instance, the potential of CYP inhibition because of the co-administration of a substrate of highly non-specific binding with an inhibitor of the low binding would be likely greater than the prediction using in vitro inhibitory results, in which the non-specific binding is not taken into account. Nevertheless, the relevance of nonspecific binding to the clinical DDI is also governed by other determinants, particularly pharmacokinetics. likely be limitedly affected by the hepatic flow for the orally administered drugs, particularly those with substantial firstpass effect, while significantly for the intravenously (i.v.) administered as illustrated by the following equations (Eqs. L & M): Clpo = Clint intestinal + Clint hepatic*fu L Cliv = Qh*Clint hepatic*fu/(Qh + Clint hepatic*fu) M Clpo is hepatic clearance of orally administered drugs, during the process of first-pass in particular. Cliv is hepatic clearance of i.v. administered drugs, which should also be applicable for the orally administered drugs with negligible presystemic elimination or already in systemic circulation. fu is the unbound fraction of drug and Qh is hepatic blood flow rate. Clint hepatic and Clint intestinal are the hepatic and intestinal intrinsic clearance that could be estimated in vitro. These intrinsic clearances should be viewed as the sums of the intrinsic clearances of individual responsible enzymes in the tissues. Therefore, for an orally administered drug, the clearance (Clpo) is likely governed both by intestinal and hepatic metabolism (Eq. L), thus the inhibition of the metabolic enzymes, particularly the ones highly expressed, e.g. CYP3A4, in one or both of these tissues, would directly suppress the clearance. Similarly, the likely reduction of clearance should also be expected if the metabolism is mediated by multiple CYP members, and the dose fraction metabolized by the enzymes being inhibited in the absence of inhibitor is known. In contrast, it may not be so straightforward to estimate the impact on the clearance (Cliv) from the enzyme inhibition by an i.v. concomitant, which is therefore elaborated as follows. Eq. M could be expressed as Cliv = Qh*Eh and Eh = 1/[1 + Qh/(Clint*fu)] N where Eh is hepatic extraction; Clint, intrinsic clearance by the hepatic enzymes (or Clint hepatic). If Cliv or Eh is high (e.g., Eh ≥ 0.8), Clint*fu >> Qh, thus Eq. N becomes Eh ≈ 1, and Eq. M becomes Cliv ≈ Qh O Thus, hepatic clearance (Cliv) appears blood flow (Qh)dependent, however, metabolism (Clint)- and non-specific binding (fu)-independent for the drugs with high hepatic extraction (Eh). If Cliv or Eh is low (e.g., Eh ≤ 0.2), Clint*fu << Qh, Eq. N becomes 5. Pharmacokinetics Eh ≈ Clint*fu/Qh, and Eq. M becomes Largely because of the apparent disparities between close and defined in vitro model systems and open and dynamic human bodies, pharmacokinetics (PK)-related variables, such as the methods of drug co-administration, would be difficult to assess in vitro, thus create hurdles for clinical DDI prediction. However, a few of PK variables may still be incorporated into the scheme for in vitro-in vivo correlation. Thus, hepatic clearance (Cliv) appears metabolism (Clint)and non-specific binding (fu)-dependent, however blood flow (Qh)-independent for the drugs with low hepatic extraction (Eh). Apparently, hepatic clearance would be controlled by both metabolism and blood flow if the extraction is moderate (0.3 ≤ Eh ≤ 0.7). The routes of drug administration may influence the in vivo outcome of the potential CYP inhibition, due in part to the first-pass metabolism. The hepatic clearance (Clh) would Cliv ≈ Clint*fu P Therefore, the potential scenarios of i.v. co-administration with the inhibitory concomitant might be expected as follows with the assumption of constant hepatic blood flow: Enzyme Kinetics for Clinically Relevant CYP Inhibition (i) If the drug is mainly metabolized by single enzyme, regardless of the catalytic efficiency of the enzyme, the clearance (Cliv) would likely be minimally affected if the drug is highly hepatic extracted, while markedly reduced if the drug is weakly hepatic extracted in the presence of the inhibitor. (ii) If the drug is metabolized by several enzymes of high catalytic efficiency and the suppression of one of these enzymes, particularly the meagerly expressed, would not largely affect the overall hepatic extraction, the alteration of hepatic clearance due to the enzyme inhibition by the concurrent would unlikely be markedly elicited (Eq. O). (iii) If the drug is metabolized by the enzymes of both high and low catalytic efficiencies, the effect of the concurrent would largely depend on the enzyme being inhibited. Apparently, if the enzyme(s) of low efficiency is involved thus overall high hepatic extraction unaffected, the reduction of clearance due to enzyme inhibition is expected to be, if any, unsubstantial. (iv) This may however not be the case if the enzyme(s) of high catalytic efficiency, particularly the abundantly expressed, is indeed suppressed. The inhibition of enzyme(s) of high efficiency may lead to the overall hepatic extraction from high to low, thus potentially switch the controlling factor of hepatic clearance from blood flow to liver metabolism. (v) Finally, if the drug would be metabolized by several enzymes with low catalytic efficiency, hepatic clearance, governed mainly by the intrinsic clearance of responsible enzymes (Eq. P), would be likely affected by enzyme inhibition if the dose fraction metabolized by the enzyme being inhibited is substantial. These aforementioned potential scenarios would also be projected for the clearance of orally administered drugs already in systemic circulation. The PK-related variables are numerous. Dosing sequence, frequency and duration of co-administration, and drug formulations including the racemic co-formulation, are just a few examples [130-133]. DDI from a PK standard point is generally beyond the scale of this paper, and should be thoroughly presented elsewhere. Regardless, the amalgamation of PK-related factors serves as one of many determinants for the final outcome of CYP inhibition in clinic. 6. Others (1) CYP up-regulation by the inhibitors or so-called autoinduction, sometimes regarded as the feedback of the enzyme suppression, occurs in clinic, and is frequently reported for the inducible drug-metabolizing CYP members, such as CYP1A2 and CYP3A4 [93, 134, 135]. Apparently, the clinical output of the combination of CYP inhibition and induction is intricate, depending largely on the net upshot of the inverse effects [135, 136]. Due to the transcriptional activation of genes, CYP induction tends to be gradual. Therefore, the temporal difference between the instantaneous inhibition and the delay-responded induction, besides the inducibility of the enzymes involved, would serve as one of the important indications for detecting auto-induction in clinic. In vitro, CYP induction at the protein level should be assessed in the intact cells, preferably the primary culture of human hepatocytes, for at least 2 days of exposure [137], in contrast to CYP inhibition in the reconstituted system containing one of possibly a variety of enzyme preparations for at the most 2 hours of incubation. CYP inhibition and induction, therefore, should be separately studied in vitro. Current Drug Metabolism, 2005, Vol. 6, No. 3 253 Clinical DDI due to enzyme induction (or possibly the combination of both induction and inhibition) might be demonstrated by the concurrence of tamoxifen and Letrozole. Letrozole, a non-steroidal aromatase inhibitor, is metabolized by CYP3A4 and CYP2A6 in humans [138]. While tamoxifen, an estrogen receptor antagonist widely used for the treatment of hormone-dependent breast cancer, is known to effectively inhibit CYP3A4 in vitro both in reversible (Ki ~ 6-22 µM) and irreversible manner (KI ~ 0.2 µM and kinact ~ 0.04/min) [110]. The steady state plasma concentrations of tamoxifen after conventional clinical doses (20-40 mg/day) in patients were usually in the range from 0.25 to 1.1 µM, though could be as high as 3.6 µM [139]. Therefore, these in vitro and in vivo data in combination suggested that tamoxifen might, particularly after an extended period of continuous dosing, reduce the hepatic clearance of other concurrent drugs that are mainly metabolized by CYP3A4. Surprisingly, it was not the case for the combination therapy of tamoxifen and Letrozole, as AUC of Letrozole was indeed reduced by 38% as compared to that detected in the mono-therapy of Letrozole, indicating the induction of one or both of CYP3A4 and CYP2A6 by tamoxifen [140]. (2) Interindividual variations in CYP-associated DDI are drastic due to the divergences of genetic predisposition and living environment [141]. Clearly, genetic makeup, exposure background, health condition, and other factors such as race, gender, age, body size and shape, etc. would potentially influence the response to CYP inhibition for a given individual or possibly a sub-population [142]. Genetic polymorphism, the pivotal CYP genetic variation associated with drug metabolism, was recognized as a risk factor for clinical DDI and studied extensively [143-145]. Virtually all of the important drug-metabolizing CYP members are phenotypically heterogeneous in population [146-148], the majority of which could be attributed to the corresponding genetic bases [149]. Among the polymorphic CYP forms, CYP2D6 and CYP2C19 are particularly noticeable because of the clear genetic origins, broad substrate spectra, and drastic phenotypic imprints resulting in poor (PM) and extensive metabolizers (EM) in the population [150-152]. The outcomes of CYP inhibition involving the polymorphic members in clinic are complex and highly variable for different individuals. The risk of DDI would be difficult to predict for a given individual without knowing CYP genetic makeup. Therefore, it would not be desired, in principle, to pursue a new chemical entity that is an inhibitor, particularly a MBI of the polymorphic CYP2C19 or CYP2D6. CYP polymorphismassociated interindividual variability is crucial for clinical DDI, and should be and has been solely discussed [141, 153]. (3) Some of diseases, particularly the chronic ones such as chronic inflammation and a certain type of diabetes, have been recently shown to modulate the expression of CYPs, to potentially act as a variable for the clinical DDI via the alteration of enzyme quantities or activities in the metabolic tissues [154, 155]. CYP modulation by the diseases appears bi-directional. Type II diabetes might be associated with the up-regulation of hepatic CYP2E1 [155], while chronic inflammation is speculated to down regulate the expression of several CYP members via the LPS-activated signaling 254 Current Drug Metabolism, 2005, Vol. 6, No. 3 Zhang and Wong pathways [156, 157]. However, the significance of diseases on CYP regulation and CYP-associated DDI should be further investigated, particularly in clinic. clinical DDI. Besides the common competitive CYP inhibition, the mechanism-based inactivation may occur more frequently than anticipated in the past, and is potentially one of the key underlying mechanisms for DDIs in clinic. On the other hand, atypical enzyme kinetics is still elusive with no concrete support by structural details [172-174]. Thus, the significance of atypical CYP kinetics and the value of proposed associated kinetic models should be investigated in clinic. In contrast to the rapid advance in CYP enzymology, the prediction of CYP inhibition in clinic, though being improved continuously, has not yet been generally appreciated, due in part to the uncertainty of CYP kinetic behaviors including atypical kinetics in vivo, which potentially depend on the substrates/inhibitors/effectors and their concentrations. The suitable kinetic model may therefore need to be established based on which drug and drug combination is being applied. Moreover, with the knowledge of CYP inhibition as the predominant DDI determinant, DDI associated with other metabolism- and toxicity-related proteins, such as UDP-glucuronosyltransferases (UGTs), P-glycoprotein (Pgp), and human ether-a-go-go related gene (hERG) potassium channel [175-180], should not be overlooked. UGT and P-gp inhibition could be the potential causes, while the inhibition of hERG channel is likely the severe consequence of clinical DDI. Similar to CYPs, UGTs, P-gp and hERG are membrane-bound, and promiscuous for the substrates/ inhibitors/ligands. Therefore, it is challenging to delineate the interactions between these proteins and therapeutic (4) Finally, assay condition is practically deemed the most important prerequisite for all in vitro, including DDI studies. The possible variations in the in vitro reaction systems are present but are not limited to buffers/matrices, co-factors/co-substrates, probe substrates, the concentration (or concentration ranges) of the substrates and inhibitors (or effectors), enzyme and protein quantities, vehicle or organic solvent selection and quantities, and assay procedures. The assay conditions, though being largely standardized, will be continuously refined with the expansion of our knowledge, and should be adjusted in accordance with the aims of the studies. Nevertheless, the current consensuses on performing CYP inhibition assays appear to be appropriate [6]. The substrate concentration-dependent CYP form-specific activities, one of the pivotal variables in DDI study in vitro, is highlighted in Table 1. The appropriate experimental condition and the criteria for the selection of probe substrates in detail should be referred to the consensus papers with a regulatory perspective and the review articles published previously [4, 6-8]. SUMMARY AND OUTLOOK In the article, we presented the enzyme kinetics for the study of CYP inhibition with a focus on the relevance to Table 1. Substrate Concentration-Dependent CYP Form-Specific Activities Human Drug-Metabolizing CYP Substrate Probea,b Phenacetin S-Warfarin O-DeEt 7-OH Conc. (µM) 50 500 1A2 2C9 + + + + + + + 50 500 2C19 S-Mephenytoin 4'-OH 50 500 + + Bufuralol 1'-OH 5 50 500 + + Dextromethorphan Chlorzoxazone TST/MDZ a Reactionc O-DeMe 6-OH 6β-OH/1'-OH 5 50 500 5 50 500 50 500 + + + + + + 2D6 2E1 3A4/5 + + + + + + + + + + + Recommended by FDA; TST, testosterone; MDZ, midazolam. References: Phenacetin, [158, 159]; S-Warfarin, [160]; S-Mephenytoin, [161]; Bufuralol, [162, 163]; Dextromethorphan, [164, 165]; Chlorzoxazone, [166, 167]; Testosterone, [168, 169]; Midazolam, [170, 171]. c O-DeEt, O-deethylation; O-DeMe, O-demethylation; OH, hydroxylation. +, responsible enzyme. b Enzyme Kinetics for Clinically Relevant CYP Inhibition agents at the molecular level, which will warrant future endeavors to thoroughly understand thus reliably predict clinical DDIs including both the CYP-associated and the CYP-unrelated. Current Drug Metabolism, 2005, Vol. 6, No. 3 [30] [31] [32] ACKNOWLEDGEMENTS We gratefully acknowledge Dr. E. Liang and Mr. R. Pelletier for the helpful discussions. [33] [34] [35] REFERENCES [36] [1] [37] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] Dresser, G.K.; Spence, J.D.; Bailey, D.G. (2000) Clin. Pharmacokinet., 38(1), 41-57. Levy, R.H.; Hachad, H.; Yao, C.; Ragueneau-Majlessi, I. (2003) Curr. Drug Metab., 4(5), 371-380. Kremers, P. (2002) Scientific World J., 2, 751-766. Venkatakrishnan, K.; von Moltke, L.L.; Obach, R.S.; Greenblatt, D.J. (2003) Curr. Drug Metab., 4(5), 423-459. Guengerich, F.P. (2003) Mol. Interv., 3(4), 194-204. Bjornsson, T.D.; Callaghan, J.T.; Einolf, H.J.; Fischer, V.; Gan, L.; Grimm, S.; Kao, J.; King, S.P.; Miwa, G.; Ni, L.; Kumar, G.; McLeod, J.; Obach, R.S.; Roberts, S.; Roe, A.; Shah, A.; Snikeris, F.; Sullivan, J.T.; Tweedie, D.; Vega, J.M.; Walsh, J.; Wrighton, S.A. (2003) Drug Metab. Dispos., 31(7), 815-832. Yuan, R.; Madani, S.; Wei, X.X.; Reynolds, K.; Huang, S.M. (2002) Drug Metab. Dispos., 30(12), 1311-1319. Tucker, G.T.; Houston, J.B.; Huang, S.M. (2001) Clin. Pharmacol. Ther., 70(2), 103-114. Li, X.Q.; Bjorkman, A.; Andersson, T.B.; Gustafsson, L.L.; Masimirembwa, C.M. (2003) Eur. J. Clin. Pharmacol., 59(5-6), 429-442. Obach, R.S. (2000) Drug Metab. Dispos., 28(9), 1069-1076. Ito, K.; Brown, H.S.; Houston, J.B. (2004) Br. J. Clin. Pharmacol., 57(4), 473-486. Wang, Y.H.; Jones, D.R.; Hall, S.D. (2004) Drug Metab. Dispos., 32(2), 259-266. Staack, R.F.; Theobald, D.S.; Paul, L.D.; Springer, D.; Kraemer, T.; Maurer, H.H. (2004) Xenobiotica, 34(2), 179-192. Murayama, N.; Soyama, A.; Saito, Y.; Nakajima, Y.; Komamura, K.; Ueno, K.; Kamakura, S.; Kitakaze, M.; Kimura, H.; Goto, Y.; Saitoh, O.; Katoh, M.; Ohnuma, T.; Kawai, M.; Sugai, K.; Ohtsuki, T.; Suzuki, C.; Minami, N.; Ozawa, S.; Sawada, J. (2004) J. Pharmacol. Exp. Ther., 308(1), 300-306. Dickinson, F.M. (1996) In Enzymology., LabFax, (Engel, P.C. Ed.), BIOS Scientific Publishers Limited, Oxford, UK, pp. 84-96. Stryer, L. (1988) In Biochemistry., 3rd Edition (Stryer, L. Ed.), W. H. Freeman and Company, New York, pp. 177-200. Wilkinson, G.R.; Shand, D.G. (1975) Clin. Pharmacol. Ther., 18(4), 377-390. Mor, G.; Eliza, M.; Song, J.; Wiita, B.; Chen, S.; Naftolin, F. (2001) J. Steroid Biochem. Mol. Biol., 79(1-5), 239-246. Yamamoto, T.; Hagima, N.; Nakamura, M.; Kohno, Y.; Nagata, K.; Yamazoe, Y. (2003) Drug Metab. Dispos., 31(1), 60-66. Sahi, J.; Black, C.B.; Hamilton, G.A.; Zheng, X.; Jolley, S.; Rose, K.A.; Gilbert, D.; LeCluyse, E.L.; Sinz, M.W. (2003) Drug Metab. Dispos., 31(4), 439-446. Jones, H.M.; Hallifax, D.; Houston, J.B. (2004) Drug Metab. Dispos., 32(5), 572-580. Donato, M.T.; Castell, J.V. (2003) Clin. Pharmacokinet., 42(2), 153-178. Fournet, N.; Weitsman, S.R.; Zachow, R.J.; Magoffin, D.A. (1996) Endocrinology, 137(1), 166-174. Suzuki, T.; Kishi, Y.; Totani, M.; Kagamiyama, H.; Murachi, T. (1987) Biotechnol. Appl. Biochem., 9(2), 170-180. Kim, R.B. (2003) Eur. J. Clin. Invest., 33(Suppl. 2), 1-5. Cho, U.S.; Park, E.Y.; Dong, M.S.; Park, B.S.; Kim, K.; Kim, K.H. (2003) Biochim. Biophys. Acta, 1648(1-2), 195-202. Tipton, K.F. (1996) In Enzymology., LabFax, (Engel, P.C. Ed.), BIOS Scientific Publishers Limited, Oxford, UK, pp. 115-136. Mei, Q.; Tang, C.; Lin, Y.; Rushmore, T.H.; Shou, M. (2002) Drug Metab. Dispos., 30(6), 701-708. Zhang, Z.Y.; King, B.M.; Mollova, N.N.; Wong, Y.N. (2002) Drug Metab. Dispos., 30(7), 805-813. [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] 255 Harlow, G.R.; Halpert, J.R. (1998) Proc. Natl. Acad. Sci. USA, 95(12), 6636-6641. Torimoto, N.; Ishii, I.; Hata, M.; Nakamura, H.; Imada, H.; Ariyoshi, N.; Ohmori, S.; Igarashi, T.; Kitada, M. (2003) Biochemistry, 42(51), 15068-15077. Miller, G.P.; Guengerich, F.P. (2001) Biochemistry, 40(24), 72627272. Tang, W.; Stearns R.A. (2001) Curr. Drug Metab., 2(2), 185-198. Atkins, W.M.; Lu, W.D.; Cook, D.L. (2002) J. Biol. Chem., 277(36), 33258-33266. Galetin, A.; Clarke, S.E.; Houston, J.B. (2003) Drug Metab. Dispos., 31(9), 1108-1116. Oda, Y.; Kharasch, E.D. (2001) J. Pharmacol. Exp. Ther., 297(1), 410-422. Korzekwa, K.R.; Krishnamachary, N.; Shou, M.; Ogai, A.; Parise, R.A.; Rettie, A.E.; Gonzalez, F.J.; Tracy, T.S. (1998) Biochemistry, 37(12), 4137-4147. Wrighton, S.A.; Schuetz, E.G.; Thummel, K.E.; Shen, D.D.; Korzekwa, K.R.; Watkins, P.B. (2000) Drug Metab. Rev., 32(3-4), 339-361. Ekins, S.; Stresser, D.M.; Williams, J.A. (2003) Trends Pharmacol. Sci., 24(4), 161-166. Ekins, S.; VandenBranden, M.; Ring, B.J.; Wrighton, S.A. (1997) Pharmacogenetics, 7(3), 165-179. Yamazaki, H.; Inui, Y.; Yun, C.H.; Guengerich, F.P.; Shimada, T. (1992) Carcinogenesis, 13(10), 1789-1794. Egnell, A.C.; Eriksson, C.; Albertson, N.; Houston, B.; Boyer, S.J. (2003) Pharmacol. Exp. Ther., 307(3), 878-887. Hutzler, J.M.; Kolwankar, D.; Hummel, M.A.; Tracy, T.S. (2002) Drug Metab. Dispos., 30(11), 1194-1200. Ekins, S.; Ring, B.J.; Binkley, S.N.; Hall, S.D.; Wrighton, S.A. (1998) Int. J. Clin. Pharmacol. Ther., 36(12), 642-651. Shou, M.; Lin, Y.; Lu, P.; Tang, C.; Mei, Q.; Cui, D.; Tang, W.; Ngui, J.S.; Lin, C.C.; Singh, R.; Wong, B.K.; Yergey, J.A.; Lin, J.H.; Pearson, P.G.; Baillie, T.A.; Rodrigues, A.D.; Rushmore, T.H. (2001) Curr. Drug Metab., 2(1),17-36. Tracy, T.S. (2003) Curr. Drug Metab., 4(5),341-346. Lin, Y.; Lu, P.; Tang, C.; Mei, Q.; Sandig, G.; Rodrigues, A.D.; Rushmore, T.H.; Shou, M. (2001) Drug Metab. Dispos., 29(4 Pt 1), 368-374. Tracy, T.S.; Hutzler, J.M.; Haining, R.L.; Rettie, A.E.; Hummel, M.A.; Dickmann, L.J. (2002) Drug Metab. Dispos., 30(4), 385-390. Hutzler, J.M.; Frye, R.F.; Korzekwa, K.R.; Branch, R.A.; Huang, S.M.; Tracy, T.S. (2001) Eur. J. Pharm. Sci., 14(1), 47-52. Ngui, J.S.; Tang, W.; Stearns, R.A.; Shou, M.; Miller, R.R.; Zhang, Y.; Lin, J.H.; Baillie, T.A. (2000) Drug Metab. Dispos., 28(9), 1043-1050. Shou, M.; Dai, R.; Cui, D.; Korzekwa, K.R.; Baillie, T.A.; Rushmore, T.H. (2001) J. Biol. Chem., 276(3), 2256-2262. Hutzler, J.M.; Powers, F.J.; Wynalda, M.A.; Wienkers, L.C. (2003) Arch. Biochem. Biophys., 417(2), 165-175. Lasker, J.M.; Huang, M.T.; Conney, A.H. (1982) Science, 216(4553), 1419-1421. Tang, W.; Stearns, R.A.; Kwei, G.Y.; Iliff, S.A.; Miller, R.R.; Egan, M.A.; Yu, N.X.; Dean, D.C.; Kumar, S.; Shou, M.; Lin, J.H.; Baillie, T.A. (1999) J. Pharmacol. Exp. Ther., 291(3), 1068-74. Torimoto, N.; Ishii, I.; Hata, M.; Nakamura, H.; Imada, H.; Ariyoshi, N.; Ohmori, S.; Igarashi, T.; Kitada, M. (2003) Biochemistry, 42(51), 15068-15077. Wang, R.W.; Newton, D.J.; Liu, N.; Atkins, W.M.; Lu, A.Y. (2000) Drug Metab. Dispos., 28(3), 360-366. Kenworthy, K.E.; Clarke, S.E.; Andrews, J.; Houston, J.B. (2001) Drug Metab. Dispos., 29(12), 1644-1651. Maenpaa, J.; Hall, S.D.; Ring, B.J.; Strom, S.C.; Wrighton, S.A. (1998) Pharmacogenetics, 8(2), 137-155. Witherow, L.E.; Houston, J.B. (1999) J. Pharmacol. Exp. Ther., 290(1), 58-65. Murray, M.; Murray, K. (2003) Xenobiotica, 33(10), 973-987. Silverman, R.B. (1995) Methods Enzymol., 249, 240-283. Fan, P.W.; Gu, C.; Marsh, S.A.; Stevens, J.C. (2003) Drug Metab. Dispos., 31(1), 28-36. Alvarez-Diez, T.M.; Zheng, J. (2004) Chem. Res. Toxicol., 17(2), 150-157. Lu, P.; Schrag, M.L.; Slaughter, D.E.; Raab, C.E.; Shou, M.; Rodrigues, A.D. (2003) Drug Metab. Dispos., 31(11), 1352-1360. 256 [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] Current Drug Metabolism, 2005, Vol. 6, No. 3 Ha-Duong, N.T.; Dijols, S.; Macherey, A.C.; Goldstein, J.A.; Dansette, P.M.; Mansuy, D. (2001) Biochemistry, 40(40), 1211212122. Racha, J.K.; Rettie, A.E.; Kunze, K.L. (1998) Biochemistry, 37(20), 7407-7419. Bailey, D.G.; Dresser, G.K.; Bend, J.R. (2003) Clin. Pharmacol. Ther., 73(6), 529-537. Backman, J.T.; Olkkola, K.T.; Aranko, K.; Himberg, J.J.; Neuvonen, P.J. (1994) Br. J. Clin. Pharmacol., 37(3), 221-225. He, K.; Woolf, T.F.; Hollenberg, P.F. (1999) J. Pharmacol. Exp. Ther., 288(2), 791-797. Koenigs, L.L.; Peter, R.M.; Hunter, A.P.; Haining, R.L.; Rettie, A.E.; Friedberg, T.; Pritchard, M.P.; Shou, M.; Rushmore, T.H.; Trager, W.F. (1999) Biochemistry, 38(8), 2312-2319. Bertelsen, K.M.; Venkatakrishnan, K.; Von Moltke, L.L.; Obach, R.S.; Greenblatt, D.J. (2003) Drug Metab. Dispos., 31(3), 289-293. Denton, T.T.; Zhang, X.; Cashman, J.R. (2004) Biochem. Pharmacol., 67(4), 751-756. Mikstacka, R.; Gnojkowski, J.; Baer-Dubowska, W. (2002) Acta Biochim. Pol., 49(4), 917-925. Tatsunami, S.; Yago, N.; Hosoe, M. (1981) Biochim. Biophys. Acta, 662(2), 226-235. Watkins, P.B.; Wrighton, S.A.; Schuetz, E.G.; Maurel, P.; Guzelian, P.S. (1986) J. Biol. Chem., 261(14), 6264-6271. Pichard, L.; Fabre, I.; Daujat, M.; Domergue, J.; Joyeux, H.; Maurel, P. (1002) Mol. Pharmacol., 41(6), 1047-1055. Correia, M.A. (1991) Methods Enzymol., 206, 315-325. Greenblatt, D.J.; von Moltke, L.L.; Harmatz, J.S.; Chen, G.; Weemhoff, J.L.; Jen, C.; Kelley, C.J.; LeDuc, B.W.; Zinny, M.A. (2003) Clin. Pharmacol. Ther., 74(2), 121-129. Iwatsubo, T.; Hirota, N.; Ooie, T.; Suzuki, H.; Shimada, N.; Chiba, K.; Ishizaki, T.; Green, C.E.; Tyson, C.A.; Sugiyama, Y. (1997) Pharmacol. Ther., 73(2), 147-1471. Tsao, S.C.; Dickinson, T.H.; Abernethy, D.R. (1990) Drug Metab. Dispos., 18(2), 180-182. Bartkowski, R.R.; Goldberg, M.E.; Larijani, G.E.; Boerner, T. (1989) Clin. Pharmacol. Ther., 46(1), 99-102. Mullins, M.E.; Horowitz, B.Z.; Linden, D.H.; Smith, G.W.; Norton, R.L.; Stump, J. (1998) JAMA, 280(2), 157-158. Fayer, J.L.; Zannikos, P.N.; Stevens, J.C.; Luo, Y.; Sidhu, R.; Kirkesseli, S. (2001) J. Clin. Pharmacol., 41(3), 305-316. Rodrigues, A.D.; Winchell, G.A.; Dobrinska, M.R. (2001) J. Clin. Pharmacol., 41(4), 368-373. Paine, M.F.; Davis, C.L.; Shen, D.D.; Marsh, C.L.; Raisys, V.A.; Thummel, K.E. (2000) Eur. J. Pharm. Sci., 12(1), 51-62. Chien, J.Y.; Mohutsky, M.A.; Wrighton, S.A. (2003) Curr. Drug Metab., 4(5), 347-356. Ito, K.; Ogihara, K.; Kanamitsu, S.; Itoh, T. (2003) Drug Metab. Dispos., 31(7), 945-954. von Moltke, L.L.; Greenblatt, D.J.; Schmider, J.; Harmatz, J.S.; Shader, R.I. (1995) Clin. Pharmacokinet., 29(Suppl. 1), 33-43. He, M.; Kunze, K.L.; Trager, W.F. (1995) Drug Metab. Dispos., 23(6), 659-663. Sutton, D.; Butler, A.M.; Nadin, L.; Murray, M. (1997) J. Pharmacol. Exp. Ther., 282(1), 294-300. Zhou, S.; Kestell, P.; Paxton, J.W. (2002) Drug Metab. Rev., 34(4), 751-90. Wandel, C.; Kim, R.B.; Kajiji, S.; Guengerich, P.; Wilkinson, G.R.; Wood, A.J. (1999) Cancer Res., 59(16), 3944-3948. Armstrong, S.C.; Cozza, K.L.; Sandson, N.B. (2003) Psychosomatics, 44(3), 255-258. Rowland, M.; Matin, S.B. (1973) J. Pharmacokinet. Biopharm., 1(6), 553-567. Christensen, L.K.; Hansen, J.M.; Kristensen, M. (1963) Lancet, 41, 1298-1301. Ding, X.; Kaminsky, L.S. (2003) Annu. Rev. Pharmacol. Toxicol., 43, 149-173. Hashizume, T.; Imaoka, S.; Mise, M.; Terauchi, Y.; Fujii, T.; Miyazaki, H.; Kamataki, T.; Funae, Y. (2002) J. Pharmacol. Exp. Ther., 300(1), 298-304. Anders, M.W.; Dekant, W. (1998) Annu. Rev. Pharmacol. Toxicol., 38, 501-537. Iatsimirskaia, E.; Tulebaev, S.; Storozhuk, E.; Utkin, I.; Smith, D.; Gerber, N.; Koudriakova, T. (1997) Clin. Pharmacol. Ther., 61(5), 554-562. Zhang and Wong [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] Obach, R.S.; Zhang, Q.Y.; Dunbar, D.; Kaminsky, L.S. (2001) Drug Metab. Dispos., 29(3), 347-352. Jamis-Dow, C.A.; Pearl, M.L.; Watkins, P.B.; Blake, D.S.; Klecker, R.W.; Collins, J.M. (1997) Am. J. Clin. Oncol., 20(6), 592-599. Pesco-Koplowitz, L.; Hassell, A.; Lee, P.; Zhou, H.; Hall, N.; Wiesinger, B.; Mechlinski, W.; Grover, M.; Hunt, T.; Smith, R.; Travers, S. (1999) J. Clin. Pharmacol., 39(1), 76-85. Gorski, J.C.; Jones, D.R.; Haehner-Daniels, B.D.; Hamman, M.A.; O'Mara, E.M. Jr.; Hall, S.D. (1998) Clin. Pharmacol. Ther. , 64(2), 133-143. Veronese, M.L.; Gillen, L.P.; Dorval, E.P.; Hauck, W.W.; Waldman, S.A.; Greenberg, H.E. (2003) J. Clin. Pharmacol., 43(10), 1091-1100. Kruijtzer, C.M.; Schellens, J.H.; Mezger. J.; Scheulen. M.E.; Keilholz, U.; Beijnen, J.H.; Rosing, H.; Mathot. R.A.; Marcus, S.; van Tinteren, H.; Baas, P. (2002) J. Clin. Oncol., 20(23), 45084516. Phillips, J.P.; Antal, E.J.; Smith, R.B. (1986) J. Clin. Psychopharmacol., 6(5), 297-299. Ito, K.; Iwatsubo, T.; Kanamitsu, S.; Ueda, K.; Suzuki, H.; Sugiyama, Y. (1998) Pharmacol. Rev., 50(3), 387-412. Mayhew, B.S.; Jones, D.R.; Hall, S.D. (2000) Drug Metab. Dispos., 28(9), 1031-1037. Malhotra, S.; Bailey, D.G.; Paine, M.F.; Watkins, P.B. (2001) Clin. Pharmacol. Ther., 69(1), 14-23. Zhao, X.J.; Jones, D.R.; Wang, Y.H.; Grimm, S.W.; Hall, S.D. (2002) Xenobiotica, 32(10), 863-878. Jones, D.R.; Gorski, J.C.; Hamman, M.A.; Mayhew, B.S.; Rider, S.; Hall, S.D. (1999) J. Pharmacol. Exp. Ther., 290(3), 1116-1125. Kanamitsu, S.; Ito, K.; Green, C.E.; Tyson, C.A.; Shimada, N.; Sugiyama, Y. (2000) Pharm. Res., 17(4), 419-426. Mani, C.; Pearce, R.; Parkinson, A.; Kupfer, D. (1994) Carcinogenesis, 15(12), 2715-2720. Periti, P.; Mazzei, T.; Mini, E.; Novelli, A. (1992) Clin. Pharmacokinet., 23(2), 106-131. Takedomi, S.; Matsuo, H.; Yamano, K.; Yamamoto, K.; Iga, T.; Sawada, Y. (1998) Drug Metab. Dispos., 26(4), 318-323. Yamano, K.; Yamamoto, K.; Kotaki, H.; Sawada, Y.; Iga, T. (1999) Drug Metab. Dispos., 27(3), 395-402. Sugie, M.; Asakura, E.; Zhao, Y.L.; Torita, S.; Nadai, M.; Baba, K.; Kitaichi, K.; Takagi, K.; Takagi, K.; Hasegawa, T. (2004) Antimicrob. Agents Chemother., 48(3), 809-814. Hatanaka, T. (2000) Clin. Pharmacokinet., 39(6), 397-412. Ito, K.; Chiba, K.; Horikawa, M.; Ishigami, M.; Mizuno, N.; Aoki, J.; Gotoh, Y.; Iwatsubo, T.; Kanamitsu, S.; Kato, M.; Kawahara, I.; Niinuma, K.; Nishino, A.; Sato, N.; Tsukamoto, Y.; Ueda, K.; Itoh, T.; Sugiyama, Y. (2002) AAPS Pharm. Sci., 4(4), E25. Radominska-Pandya, A.; Czernik, P.J.; Little, J.M.; Battaglia, E.; Mackenzie, P.I. (1999) Drug Metab. Rev., 31(4), 817-899. Bossuyt, X.; Blanckaert, N. (1997) Biochem. J., 323( Pt 3), 645648. Gillette, Jr. (1963) Adv. Enzyme Regul.,17, 215-223. Francesco, C.D.; Bickel, M.H. (1977) Chem. Biol. Interact. , 16(3), 335-346. Romer, J.; Bickel, M.H. (1979) J. Pharm. Pharmacol., 31(1), 7-11. Margolis, J.M.; Obach, R.S. (2003) Drug Metab. Dispos., 31(5), 606-611. Obach, R.S. (1999) Drug Metab. Dispos., 27(11), 1350-1359. Kalvass, J.C.; Tess, D.A.; Giragossian, C.; Linhares, M.C.; Maurer, T.S. (2001) Drug Metab. Dispos., 29(10), 1332-1336. Gibbs, M.A.; Kunze, K.L.; Howald, W.N.; Thummel, K.E. (1999) Drug Metab. Dispos., 27(5), 596-599. McLure, J.A.; Miners, J.O.; Birkett, D.J. (2000) Br. J. Clin. Pharmacol., 49(5), 453-461. Yang, J.; Kjellsson, M.; Rostami-Hodjegan, A.; Tucker, G.T. (2003) Eur. J. Pharm. Sci., 20(2), 223-232. Fang, J.; McKay, G.; Hubbard, J.W.; Hawes, E.M.; Midha, K.K. (2000) Biopharm. Drug Dispos., 21(7), 249-259. Mountfield, R.J.; Senepin, S.; Schleimer, M.; Walter, I.; Bittner, B. (2000) Int. J. Pharm., 211(1-2), 89-92. Karim, A.; Piergies, A. (1995) Clin. Pharmacol. Ther., 58(2), 174184. Wenk, M.; Todesco, L.; Krahenbuhl, S. (2004) Br. J. Clin. Pharmacol., 57(4), 495-499. Enzyme Kinetics for Clinically Relevant CYP Inhibition [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] Luo, G.; Lin, J.; Fiske, W.D.; Dai, R.; Yang, T.J.; Kim, S.; Sinz, M.; LeCluyse, E.; Solon, E.; Brennan, J.M.; Benedek, I.H.; Jolley, S.; Gilbert, D.; Wang, L.; Lee, F.W.; Gan, L.S. (2003) Drug Metab. Dispos., 31(9), 1170-1175. Hsu, A.; Granneman, G.R.; Witt, G.; Locke, C.; Denissen, J.; Molla, A.; Valdes, J.; Smith, J.; Erdman, K.; Lyons, N.; Niu, P.; Decourt, J.P.; Fourtillan, J.B.; Girault, J.; Leonard, J.M. (1997) Antimicrob. Agents Chemother., 41(5), 898-905. Li, A.P.; Maurel, P.; Gomez-Lechon, M.J.; Cheng, L.C.; JurimaRomet, M. (1997) Chem. Biol. Interact., 107(1-2), 5-16. Wire, B.; Valles, B.; Parkinson, A.; Madan, A.; Probst, A.; Zimmerlin, A.; Gut, J. (1996) ISSX Proceedings, 10, 359. Millward, M.J.; Cantwell, B.M.; Lien, E.A.; Carmichael, J.; Harris, A.L. (1992) Eur. J. Cancer, 28A(4-5), 805-810. Dowsett, M.; Pfister, C.; Johnston, S.R.; Miles, D.W.; Houston, S.J.; Verbeek, J.A.; Gundacker, H.; Sioufi, A.; Smith, I.E. (1999) Clin. Cancer Res., 5(9), 2338-2343. Lin, J.H.; Lu, A.Y. (2001) Annu. Rev. Pharmacol. Toxicol., 41, 535-567. Bachmann, K.A. (2002) Am. J. Ther., 9(4), 309-316. Yu, K.S.; Yim, D.S.; Cho, J.Y.; Park, S.S.; Park, J.Y.; Lee, K.H.; Jang, I.J.; Yi, S.Y.; Bae, K.S.; Shin, S.G. (2001) Clin. Pharmacol. Ther., 69(4), 266-273. Brosen, K. (1998) Int. Clin. Psychopharmacol.,13(Suppl. 5), S45S47. Meyer, U.A.; Amrein, R.; Balant, L.P.; Bertilsson, L.; Eichelbaum, M.; Guentert, T.W.; Henauer, S.; Jackson, P.; Laux, G.; Mikkelsen, H.; Peck, C.; Pollock, B.G.; Priest, R.; Sjoqvist, F.; Delini-Stula, A. (1996) Acta Psychiatr. Scand., 93(2), 71-79. Dahl, M.L. (2002) Clin. Pharmacokinet., 41(7), 453-470. Schwarz, U.I. (2003) Eur. J. Clin. Invest., 33(Suppl. 2), 23-30. Landi, M.T.; Sinha, R.; Lang, N.P.; Kadlubar, F.F. (1999) IARC Sci. Publ., 148, 173-195. Ingelman-Sundberg, M. (2004) Trends Pharmacol. Sci., 25(4),193200. Kosuge, K.; Jun, Y.; Watanabe, H.; Kimura, M.; Nishimoto, M.; Ishizaki, T.; Ohashi, K. (2001) Drug Metab. Dispos., 29(10), 12841289. Mendoza, R.; Wan, Y.J.; Poland, R.E.; Smith, M.; Zheng, Y.; Berman, N.; Lin, K.M. (2001) Clin. Pharmacol. Ther. , 70(6), 552560. Brosen, K.; Hansen, J.G.; Nielsen, K.K.; Sindrup, S.H.; Gram, L.F. (1993) Eur. J. Clin. Pharmacol., 44(4), 349-355. Tamasi, V.; Vereczkey, L.; Falus, A.; Monostory, K. (2003) Inflamm. Res., 52(8), 322-333. Orlando, R.; Piccoli, P.; De Martin, S.; Padrini, R.; Palatini, P. (2003) Br. J. Clin. Pharmacol., 55(1), 86-93. Wang, Z.; Hall, S.D.; Maya, J.F.; Li, L.; Asghar, A.; Gorski, J.C. (2003) Br. J. Clin. Pharmacol., 55(1), 77-85. Donato, M.T.; Guillen, M.I.; Jover, R.; Castell, J.V.; GomezLechon, M.J. (1997) J. Pharmacol. Exp. Ther., 281(1), 484-490. Current Drug Metabolism, 2005, Vol. 6, No. 3 [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] [171] [172] [173] [174] [175] [176] [177] [178] [179] [180] 257 Cheng, P.Y.; Wang, M.; Morgan, E.T. (2003) J. Pharmacol. Exp. Ther., 307(3), 1205-1212. Nakajima, M.; Kobayashi, K.; Oshima, K.; Shimada, N.; Tokudome, S.; Chiba, K.; Yokoi, T. (1999) Xenobiotica, 29(9), 885-898. Venkatakrishnan, K.; von Moltke, L.L.; Greenblatt, D.J. (1998) J. Pharm. Sci., 87(12), 1502-1507. Kaminsky, L.S.; Zhang, Z.Y. (1997) Pharmacol. Ther., 73(1), 6774. Goldstein, J.A.; de Morais, S.M. (1994) Pharmacogenetics, 4(6), 285-299. Gelboin, H.V.; Krausz, K.W.; Shou, M.; Gonzalez, F.J.; Yang, T.J. (1997) Pharmacogenetics, 7(6), 469-477. Mankowski, D.C. (1999) Drug Metab. Dispos., 27(9), 1024-1028. Ono, S.; Hatanaka, T.; Hotta, H.; Satoh, T.; Gonzalez, F.J.; Tsutsui, M. (1996) Xenobiotica, 26(7), 681-693. Haritos, V.S.; Ching, M.S.; Ghabrial, H.; Gross, A.S.; Taavitsainen, P.; Pelkonen, O.; Battaglia, S.E.; Smallwood, R.A.; Ahokas, J.T. (1998) Pharmacogenetics, 8(5), 423-432. Ono, S.; Hatanaka, T.; Hotta, H.; Tsutsui, M.; Satoh, T.; Gonzalez, F.J. (1995) Pharmacogenetics, 5(3), 143-150. Gorski, J.C.; Jones, D.R.; Wrighton, S.A.; Hall, S.D. (1997) Xenobiotica, 27(3), 243-256. Wrighton, S.A.; Ring, B.J.; Watkins, P.B.; VandenBranden, M. (1989) Mol. Pharmacol., 36(1), 97-105. Mei, Q.; Tang, C.; Assang, C.; Lin, Y.; Slaughter, D.; Rodrigues, A.D.; Baillie, T.A.; Rushmore, T.H.; Shou, M. (1999) J. Pharmacol. Exp. Ther., 291(2), 749-759. Kronbach, T.; Mathys, D.; Umeno, M.; Gonzalez, F.J.; Meyer, U.A. (1989) Mol. Pharmacol., 36(1), 89-96. Ghosal, A.; Satoh, H.; Thomas, P.E.; Bush, E.; Moore, D. (1996) Drug Metab. Dispos., 24(9), 940-947. Yano, J.K.; Wester, M.R.; Schoch, G.A.; Griffin, K.J.; Stout, C.D.; Johnson, E.F. (2004) J. Biol. Chem., 279(37), 38091-38094. Williams, P.A.; Cosme, J.; Ward, A.; Angove, H.C.; Matak Vinkovic, D.; Jhoti, H. (2003) Nature, 424(6947), 464-468. Schoch, G.A.; Yano, J.K.; Wester, M.R.; Griffin, K.J.; Stout, C.D.; Johnson, E.F. (2004) J. Biol. Chem., 279(10), 9497-9503. Lin, J.H.; Wong, B.K. (2002) Curr. Drug Metab., 3(6), 623-646. Stone, A.N.; Mackenzie, P.I.; Galetin, A.; Houston, J.B.; Miners, J.O. (2003) Drug Metab. Dispos., 31(9), 1086-1089. Penzotti, J.E.; Lamb, M.L.; Evensen, E.; Grootenhuis, P.D. (2002) J. Med. Chem., 45(9), 1737-1740. Pajeva, I.K.; Globisch, C.; Wiese, M. (2004) J. Med. Chem., 47(10), 2523-2533. Mitcheson, J.S.; Chen, J.; Lin, M.; Culberson, C.; Sanguinetti, M.C. (2000) Proc. Natl. Acad. Sci. USA, 97(22), 12329-12333. Cavalli, A.; Poluzzi, E.; De Ponti, F.; Recanatini, M. (2002) J. Med. Chem., 45(18), 3844-3853.