Cocktail-substrate Approach-based High-throughput Assay for Evaluation of Direct and Time-dependent Inhibition of Multiple Cytochrome P450 Isoforms

Transcription

1 Drug Metab. Pharmacokinet. 29 (2): (2014). Regular Article Cocktail-substrate Approach-based High-throughput Assay for Evaluation of Direct and Time-dependent Inhibition of Multiple Cytochrome P450 Isoforms Kazumasa KOZAKAI 1, *,YasuhiroYAMADA 1,MotojiOSHIKATA 2,TaijiKAWASE 2, Etsuko SUZUKI 2,YukariHARAMAKI 2 and Hideki TANIGUCHI 1 1 Department of Regenerative Medicine, Yokohama City University School of Medicine, Yokohama, Japan 2 Solutions Center, Nihon Waters K.K., Tokyo, Japan Full text of this paper is available at Summary: Avoiding drug-drug interactions (DDIs) mediated through inhibition of cytochrome P450 (CYP) activity is highly desirable. Direct inhibition (DI) of CYP through new chemical entities (NCEs) or timedependent inhibition (TDI) through reactive metabolites should be elucidated at an early stage of drug discovery research. In particular, TDI of CYP occurring through reactive metabolites may be irreversible and even sustained, causing far more serious DDIs for TDIs than for DIs. Furthermore, it is important to ascertain whether an NCE inhibits multiple CYP isoforms. Hence, using a cocktail-substrate approach that we previously established (in which the activity of 8 CYP isoforms is simultaneously evaluated in a single run), we evaluated the IC 50 values of direct inhibitors and TDI parameters (k obs,shiftedic 50, K I and k inact ) of time-dependent inhibitors that affect multiple CYP isoforms. The IC 50 values for 8 CYP isoforms obtained using the cocktail-substrate approach were nearly identical to values previously reported. The TDI parameters for CYP1A2, 2C9, 2C19, 2D6, and CYP3A4/5 obtained using the cocktail-substrate approach were also nearly identical to those obtained using a single-substrate approach. Thus, the cocktail-substrate approach is useful for evaluating DI and TDI in the early stages of drug discovery and development processes. Keywords: CYP inhibition; direct inhibition; time-dependent inhibition; IC 50 -shift; UPLC-MS/MS; cocktail approach; drug-drug interaction Introduction A cytochrome P450 (CYP) enzyme is a heme protein involved in the metabolism of a number of xenobiotic and endogenous substances. 1) CYP enzymes have more than 400 isoforms, and the major isoforms involved in drug metabolism in the liver are CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, and CYP3A5. Enzyme activity can be inhibited or induced by a number of drugs, and may be altered by coadministered drugs via drug-drug interactions (DDIs). 2 4) Therefore, it is important to evaluate potential DDIs for the development of safe and effective new drugs. A potent DDI that inhibits CYP activity would cause termination of clinical development of a drug candidate or withdrawal of a new marketed drug. 5,6) Thus, it is vital for researchers engaged in drug discovery to understand the type of CYP isoform that may be inhibited by a new chemical entity (NCE) and the exact mechanism for this inhibitory action. Because a number of drugs are known to significantly inhibit multiple CYP isoforms, it would be valuable to simultaneously examine their inhibition profiles. Furthermore, it is critical to investigate structure-activity relationships (SARs) and evaluate the inhibition potentials for multiple CYP isoforms of NCEs in the early stages of drug discovery. Therefore, we previously established a high-throughput assay system for rapid evaluation of multiple CYP inhibition. 7) This assay system used a cocktail of probe substrates selective for eight CYP isoforms (CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, and CYP3A4/5) to evaluate the inhibition potential of NCEs against each isoform simultaneously in a single run. In the present study, the measurement of CYP2E1 activity was not Received August 25, 2013; Accepted October 18, 2013 J-STAGE Advance Published Date: October 29, 2013, doi: /dmpk.dmpk-13-rg-093 *To whom correspondence should be addressed: Kazumasa KOZAKAI, Department of Regenerative Medicine, Yokohama City University School of Medicine, 3-9 Fuku-ura, Kanazawa-ku, Yokohama , Japan. Tel , Fax , kozakai. kazumasa@gmail.com 198

2 Rapid Screening for CYP Inhibition with Substrate Cocktail 199 executed for the following reasons. Earlier evidence based on in vivo CYP cocktail studies suggest that chlorzoxazone can significantly inhibit midazolam hydroxylation (CYP3A activity) in healthy human subjects. 8) In addition, chlorzoxazone was also reported to interact with CYP1A2 substrates, 8,9) which may be responsible for the slower rate of metabolism of phenacetin (CYP1A2) in this study. In order to avoid such potential intersubstrate interaction, the CYP2E1 probe, chlorzoxazone, was excluded from the CYP cocktail substrate mixtures. This assay provided results comparable to those provided by a single substrate assay system. In the present study, we extended our initial investigation using this cocktail approach. The mechanism of CYP inhibition can be roughly divided into two categories. 10) One category is the inhibition of CYP activity mediated by the parent compound itself (direct inhibition, DI), and the other category is the inhibition mediated by the metabolites derived from the parent compound (time-dependent inhibition, TDI). In most cases, the inhibition caused by the former (DI) is reversible, while that caused by the latter (TDI) is irreversible. Consequently, in clinical situations, the potency of a direct inhibitor will decrease upon the elimination of the parent compound from its site of action. In contrast, even when the concentration of the parent compound decreases, the reactive metabolites may remain so that TDI may lead to sustained inhibition. Often this results in far more serious DDIs for TDI than those for DI. Moreover, irreversible TDI, which results from the irreversible binding of reactive metabolites formed during metabolic reactions with CYP proteins, may result in hapten formation and serious adverse effects such as idiosyncratic toxicity due to autoimmune reaction, as well as CYP inhibition. 11) Hence, postmarketing withdrawal occurs more often for a drug with TDI potency than for a drug with DI potency. 10) In order to avoid such serious failures, both DI and TDI need to be evaluated early in the process of drug discovery. Further, the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) require in vitro evaluations of the DI and TDI potential for a candidate drug in human liver microsomes (HLM). 12,13) Therefore, we also examined the feasibility of evaluating TDI by using the cocktail-substrate approach in our high-throughput CYP inhibition assay. In order to fully characterize TDI potency, it is necessary to obtain inhibition parameters (k inact and K I values) from a complex in vitro study. However, conducting such assays for a drug with weak TDI potential or without TDI potential would be a waste of time, labor, and money. Thus, the pharmaceutical industry has developed a stepwise strategy for evaluation: (1) assessment of the presence or absence of TDI potential based on a shift in the IC 50 value of a test drug by co-incubation of the test drug with NADPH, and then (2) measurement of TDI parameters only when an IC 50 value shifts leftward to a lower concentration. In following this strategy, evaluating test compounds with TDI by using multiple CYP isoforms simultaneously in a high-throughput assay would be highly useful. Therefore, in the present study, we examined whether screening with a cocktail of multiple substrates could provide useful results for selection of candidate compounds. Materials and Methods Chemicals and materials: Coumarin, 7-hydroxycoumarin, phenacetin, and testosterone were purchased from Nacalai Tesque (Kyoto, Japan). Acetaminophen, amodiaquine, bupropion, erythromycin, fluvoxamine maleate, fluconazole, furafylline, itraconazole, and ticlopidine hydrochloride were purchased from Sigma-Aldrich (St. Louis, MO). Acetaminophen-d 4, bufuralol, hydroxybupropion, 1A-hydroxybufuralol, 1A-hydroxybufuralol-d 9, 4A-hydroxydiclofenac- 13 C 6, («)-4A-hydroxymephenytoin-d 3, 1Ahydroxymidazolam, 6 -hydroxytestosterone, N-desethylamodiaquine, N-desethylamodiaquine-d 5, (S)-mephenytoin, and paroxetine hydrochloride were purchased from Toronto Research Chemicals (North York, Canada). Diclofenac and («)-4A-hydroxymephenytoin were purchased from Ultrafine (Manchester, UK). Midazolam was purchased from Wako Pure Chemicals Industries (Osaka, Japan). 4A-Hydroxydiclofenac, 7-hydroxycoumarin-d 5, hydroxybupropion-d 6, 6 -hydroxytestosterone-d 7, and tienilic acid were purchased from BD Gentest (Woburn, MA). Fluconazole and 1-hydroxymidazolam-d 4 were purchased from MP Biomedicals (Irvine, CA) and Cerilliant Corporation (Round Rock, TX), respectively. Pooled HLMs from 50 donors were provided by XenoTech, LLC (Lenexa, KS) and stored at ¹80 C until use. The microsome characterization (including microsomal protein concentration, total CYP content, and enzyme activity of each CYP isoform) was performed by the manufacturer. All other reagents and solvents were of analytical grade and commercially available. Determination of IC 50 for DI: All determinations of IC 50 values for DI were carried out as previously described, and all incubations were performed under conditions shown to be linear with respect to time and protein concentration. 7) Assays were performed using test tubes in triplicate in a reaction medium (250 µl) containing 0.1 M potassium phosphate (ph 7.4), 1 mm EDTA, 5 mm MgCl 2, an NADPH regeneration system (1 mm NADP +, 10 mm glucose 6-phosphate, and 1 IU glucose 6-phosphate dehydrogenase), 0.2 mg protein/ml of liver microsomes, a cocktail of 9 probe substrates for 8 CYP isoforms (Table 1), and a CYP direct inhibitor. The concentrations of the substrate were used at near or below their respective Michaelis-Menten constant (K m ) values for these marker reactions under the incubation conditions. Itraconazole (concentration range in incubation medium: µm), fluconazole ( µm), and fluvoxamine ( µm) at nine concentrations were used as direct inhibitors for multiple CYP isoforms. Inhibition curves were generated using nine concentrations of these inhibitors, and the IC 50 values were calculated. In all of the experiments, the substrates were dissolved in methanol (final concentration, 1.0%) to the required concentration. After preincubation at 37 C for 5 min, the reactions were initiated by the addition of the NADPH regeneration system. Following incubation at 37 C for 5 min, the reactions were terminated by the addition of 250 µl of an ice-cold mixture of acetonitrile and methanol (1:1, v/v) and the internal standards (ISs, Table 1). Stable-isotope compounds for CYP-specific probe substrate metabolites were used as ISs to avoid interference by ion suppression in liquid chromatography-tandem mass spectrometry (LC-MS/MS) quantification. After the reactions were quenched, aliquots of the solutions were analyzed by LC-MS/MS. 7) Comparison of TDI parameters between single substrates and a substrate cocktail: Furafylline (concentration range in initial incubation medium: µm), tienilic acid ( µm), ticlopidine ( µm), paroxetine ( µm), and erythromycin ( µm) at nine concentrations were used as prototypical time-dependent inhibitors for CYP1A2, 2C9, 2C19, 2D6, and 3A4/5, respectively. The following TDI parameters were determined using either single substrates or the substrate cocktail:

3 200 Kazumasa KOZAKAI, et al. Enzyme Table 1. Experimental conditions for measuring microsomal CYP activity in enzyme inhibition studies CYP probe substrate Concentration Metabolite monitored Internal standard (µm)* (Mass transition, m/z) (Mass transition, m/z) CYP1A2 Phenacetin 20 Acetaminophen (152 > 110) Acetaminophen-d 4 (156 > 114) CYP2A6 Coumarin 2 7-Hydroxycoumarin (163 > 107) 7-Hydroxycoumarin-d 5 (168 > 112) CYP2B6 Bupropion 5 Hydroxybupropion (256 > 238) Hydroxybupropion-d 6 (262 > 244) CYP2C8 Amodiaquine 0.1 N-Desethylamodiaquine (328 > 283) N-Desethylamodiaquine-d 5 (333 > 283) CYP2C9 Diclofenac 1 4A-Hydroxydiclofenac (312 > 230) 4A-Hydroxydiclofenac- 13 C 6 (318 > 236) CYP2C19 (S)-Mephenytoin 40 4A-Hydroxymephenytoin (235 > 150) («)-4A-Hydroxymephenytoin-d 3 (238 > 150) CYP2D6 Bufuralol 5 1A-Hydroxybufuralol (278 > 186) 1A-Hydroxybufuralol-d 9 (267 > 186) CYP3A4/5 Midazolam 2 1A-Hydroxymidazolam (342 > 203) 1A-Hydroxymidazolam-d 4 (346 > 203) Testosterone Hydroxytestosterone (305 > 269) 6 -Hydroxytestosterone-d 7 (312 > 276) *Concentration in the final incubation medium. the inhibitor concentration required for half-maximal inactivation (K I ), the maximum inactivation rate constant (k inact ), and the apparent inactivation rate constant (k obs ). All the initial incubations were performed in triplicate in 250 µl of reaction medium containing 2 mg protein/ml of liver microsomes and well-characterized, time-dependent inhibitors with 1 mm -NADPH at 37 C for 0, 3, 6, and 12 min. After the initial incubation, aliquots (12.5 µl) were sequentially transferred to a pre-warmed (37 C) secondary incubation mixture (total volume: µl) containing 1 mm - NADPH and the single substrate or substrate cocktail (Table 1). Following secondary incubation at 37 C for 5 min, all reactions were terminated by the addition of 250 µl of an ice-cold mixture (including ISs) of acetonitrile and methanol (1:1, v/v). After this initial incubation, a 20-fold dilution was performed for secondary incubation. The aliquots were analyzed by LC-MS/MS. 7) Assessment of TDI potencies by the IC 50 -shift approach (shifted IC 50 ): The effects of 5 prototypical time-dependent inhibitors mentioned above were compared and evaluated using an IC 50 -shift approach with and without an initial incubation step, and with or without a 20-fold dilution. The time-dependent inhibitors were pre-dissolved in methanol (2 mm) and stored at ¹80 C until use and diluted to 7 concentrations at the time of use. The concentration ranges in the secondary incubation medium for furafylline, tienilic acid, ticlopidine, paroxetine, and erythromycin were 16 1,000 nm, 1.5 1,500 nm, 5 5,000 nm, 1.5 1,500 nm, and 15 15,000 nm, respectively. The present experiments were assessed in a 96-well plate by using the substrate cocktail (Table 1). The IC 50 values obtained from only the secondary incubation step without an initial incubation step depict the DI potency, while the shifted IC 50 values that depict TDI potency were obtained from the two steps of the initial incubation and subsequent secondary incubation. Dilution method A two-step incubation scheme of initial and secondary incubation was used, and the experiments were designed such that the two-step incubation could be assayed in one 96-well plate. For the initial incubation step, all the incubations were performed in triplicate in a reaction medium (250 µl) containing 0.1 M potassium phosphate (ph 7.4), 1 mm EDTA, 5 mm MgCl 2,2mg protein/ml of liver microsomes, and each concentration of the time-dependent inhibitors with 1 mm -NADPH at 37 C for 0 or 12 min. The reactions were initiated by the addition of 1 mm -NADPH. For the secondary incubation step, aliquots (12.5 µl) of the initial incubation mixture were sequentially transferred to a pre-warmed (37 C) secondary incubation mixture (total volume: µl) containing 100 mm potassium phosphate buffer (ph 7.4), 1 mm EDTA, 5 mm MgCl 2,1mM -NADPH, and a single specific substrate for each CYP or the substrate cocktail. The concentrations of probe substrate were near the K m values reported in the literature. 14) Following incubation at 37 C for 5 min, the secondary incubations were terminated by the addition of 250 µl of an ice-cold mixture (including ISs) of acetonitrile and methanol (v/v, 1:1). The aliquots were analyzed by LC-MS/MS. 7) Non-dilution method Similar to the dilution method, a two-step incubation scheme was used, and the experiments were designed such that the twostep incubation could be assayed in one 96-well plate. Each compound was tested at least in triplicate. All incubation solutions (250 µl) contained 0.1 mg of human microsomal protein/ml, 1mM -NADPH, 1 mm EDTA, and 5 mm MgCl 2 in 0.1 M potassium phosphate buffer (ph 7.4). Each concentration of timedependent inhibitors was added to the incubation performed at 37 C either 30 min prior to the addition of the probe substrate cocktail (initial incubation step) or simultaneously with the probe substrate cocktail (co-incubation or secondary incubation step), both in the presence of NADPH. The concentrations of the probe substrate were near their K m values. The addition of substrate did not significantly dilute the incubation mixture (i.e., 5% dilution). Following the addition of the substrate cocktail, the incubation continued for an additional 5 min. At the end of this incubation, an equal volume of an ice-cold mixture including ISs (Table 1) of acetonitrile and methanol (1:1, v/v) was added to the incubation, and the mixtures were centrifuged and analyzed by LC-MS/MS. 7) LC-MS/MS analysis: The simultaneous quantification of enzyme activities for 8 CYP isoforms using a cocktail of 9 probe substrates (two substrates were used only for CYP3A4/5) was performed by using LC-MS/MS with multiple reaction monitoring (MRM) as described previously. 7) Data analysis: Calculation of IC 50 and shifted IC 50 values Control samples were assayed without an inhibitor in each analytical run. The amount of metabolite in each sample (relative to the control samples) was plotted versus the concentration of the inhibitor present. The inhibition rates were fitted to a sigmoidshaped curve and IC 50 values were calculated using Prism software package version 5.02 (GraphPad Software, Inc., La Jolla, CA). The difference between the IC 50 value obtained without initial incubation and the IC 50 (shifted IC 50 ) value obtained with initial incubation was compared, and the fold-shift value (shift ratio of IC 50 values with and without initial incubation) was determined as described in Eq. (1).

4 Rapid Screening for CYP Inhibition with Substrate Cocktail 201 Table 2. Performance of the cocktail-substrate approach with DI for 8 CYP isoforms Enzyme Substrate IC 50 values (µm) of CYP inhibitors in cocktail Itraconazole Fluvoxamine Fluconazole CYP1A2 Phenacetin > >100 CYP2A6 Coumarin >10 > CYP2B6 Bupropion >100 CYP2C8 Amodiaquine >10 >30 >100 CYP2C9 Diclofenac > CYP2C19 (S)-Mephenytoin > CYP2D6 Bufuralol >10 >30 >100 CYP3A4/5 Midazolam Testosterone > IC 50 fold-shift ¼ shifted IC 50 =IC 50 ð1þ Calculation of TDI parameters (k obs,k I, and k inact ) To determine the inactivation kinetic constants, the natural logarithm of the remaining CYP activity in each sample was plotted against the incubation time. The k obs was determined from the slope of the initial linear phase. The value of k obs was plotted against the inhibitor concentrations, and the TDI kinetic parameters (K I and k inact ) were determined by the nonlinear least-squares method (Prism) by using Eq. (2). k obs ¼ k inact ½IŠ=ðK I þ½išþ ð2þ where [I] represents the inhibitor concentration. The efficiency of the inactivator as an inhibitor can be described by the quotient of the parameters, i.e., K I /k inact, and was used to compare the efficiency of inactivation among the various time-dependent inhibitors. Results Performance of the cocktail substrate approach on compounds that directly inhibit multiple CYP isoforms: The IC 50 values for itraconazole, fluvoxamine, and fluconazole for inhibition of each of the 8 CYP isoforms by using the cocktail substrate approach are shown in Table 2. These direct inhibitors showed selectivity for each corresponding CYP isoform. Itraconazole exhibited strong inhibition for CYP2B6 and CYP3A4/5 in this approach, whereas negligible inhibitory effects for itraconazole (IC 50 > 10 µm) were found for the other CYP isoforms. Fluvoxamine exhibited strong inhibition for CYP1A2 and CYP2C19, and moderate inhibition for CYP2B6, CYP2C9, and CYP3A4/5 using midazolam as a substrate, while barely inhibiting the other isoforms, including CYP3A4/5, using testosterone as a substrate (IC 50 > 30 µm). For isoforms other than CYP1A2, CYP2B6, CYP2C8, and CYP2D6, fluconazole displayed only marginal inhibition (IC 50 = µm). These results are consistent with those of previous reports ) A comparison of TDI parameters (k obs, K I and k inact ) obtained with single- and cocktail-substrate approaches: The TDI parameters for the prototypical time-dependent inhibitors furafylline, tienilic acid, ticlopidine, paroxetine, and erythromycin with CYP1A2, CYP2C9, CYP2C19, CYP2D6, and CYP3A4/5, respectively, were obtained using the single- and cocktail-substrate approaches (Table 3). The TDI parameters for each time-dependent inhibitor obtained by using the cocktail-substrate approach were approximately comparable to those obtained using the singlesubstrate approach. The difference between the k inact /K I ratios (an index of TDI potency) for both strategies ranged from to fold. In addition, the correlation between the cocktail-substrate and the single-substrate approach for the k obs values obtained at 10 µm and k inact /K I ratios was excellent, with correlation coefficients of r 2 = and , respectively (Fig. 1). The k inact /K I ratios were calculated using the cocktail-substrate method for inhibition of the eight CYP isoforms by the prototypical time-dependent inhibitors (with midazolam and testosterone used as probe substrates for CYP3A4/5) (Table 4). For furafylline, tienilic acid, paroxetine, and erythromycin, TDI potency was observed only against CYP1A2, CYP2C9, CYP2D6, and CYP3A4/5, respectively, and their k inact /K I ratios were 56, 88, 39, 1.6 (midazolam), and 1.4 (testosterone) ml/min/µmol. Little TDI potency was observed for these inhibitors with the other CYP isoforms, and their k inact /K I ratios were undetectable. However, ticlopidine demonstrated potent TDI activity not only for CYP2C19 but also for CYP2B6, with k inact /K I ratios of 176 and 16 ml/min/µmol, respectively; little TDI activity was observed for this inhibitor against the other CYP isoforms, and their k inact /K I ratios were undetectable. Evaluation of TDI potency with IC 50 -shift approaches (shifted IC 50 ): IC 50 -shift assays with our cocktail substrate or with a single substrate were used to evaluate the TDI potencies of the prototypical time-dependent inhibitors furafylline, tienilic acid, ticlopidine, paroxetine, and erythromycin by using dilution and non-dilution methods. Dilution method An inhibition curve (shifted-ic 50 values for TDI) obtained from the initial incubation responses of test drugs in the presence of NADPH was compared with an inhibition curve (IC 50 values for DI) obtained from responses without initial incubation to evaluate TDI potency based on the fold-shift of IC 50 values. A shift profile for the inhibition curves obtained using the cocktail-substrate approach is shown in Figure 2. This shift profile was in approximate agreement with that obtained using the single-substrate approach (data not shown). Following the initial incubation with NADPH, IC 50 shifts were observed only for furafylline, tienilic acid, paroxetine, and erythromycin to inhibit CYP1A2, CYP2C9, CYP2D6, and CYP3A4/5, respectively, with the corresponding fold-shifts of >15.5, 11.7, >12.0, and >2.6. Moreover, ticlopidine demonstrated an IC 50 shift for CYP2B6 and for CYP2C19 (fold-shifts of 4.2 and 4.1, respectively). The shifted-ic 50 values obtained from these timedependent inhibitors in single-substrate and cocktail-substrate approaches are shown in Table 3. The shifted-ic 50 values obtained in these approaches demonstrated an excellent correlation (r 2 = 0.998) (data not shown). Figures 3a (single substrate) and 3b (cocktail substrate) represent the correlations between k inact /K I ratios (indices of TDI potency) and k obs values for test drugs at 10 µm. Figures 3c (single substrate) and 3d (cocktail substrate) represent the correlation between k inact /K I ratios and shifted-ic 50 values. Each comparison demonstrated good correlation. Non-dilution method The shifted-ic 50 values for TDI obtained with the cocktailsubstrate approach were compared with IC 50 values for DI, to evaluate TDI activity based on fold-shift obtained from their IC 50 values using the non-dilution method (Fig. 4). As with the dilution method, IC 50 shifts were observed only for furafylline, tienilic acid, paroxetine, and erythromycin to inhibit CYP1A2, CYP2C9, CYP2D6 and CYP3A4/5, respectively, with the corresponding

5 202 Kazumasa KOZAKAI, et al. Table 3. Comparison of shifted IC 50 values and TDI parameters obtained using cocktail- and single-substrate incubations with the dilution method Enzyme Substrate Time-dependent inhibitor Inactivation kinetic value Single substrate Cocktail substrate CYP1A2 Phenacetin Furafylline shifted IC 50 (nm) k obs at 10 µm (min ¹1 ) K I (µm) k inact (min ¹1 ) k inact /K I (ml/min/µmol) (0.96)* CYP2B6 Bupropion Ticlopidine shifted IC 50 (nm) NT 32 k obs at 10 µm (min ¹1 ) NT 0.17 K I (µm) NT 4.2 k inact (min ¹1 ) NT k inact /K I (ml/min/µmol) NT 46 CYP2C9 Diclofenac Tienilic acid shifted IC 50 (nm) k obs at 10 µm (min ¹1 ) K I (µm) k inact (min ¹1 ) k inact /K I (ml/min/µmol) (1.09) CYP2C19 (S)-Mephenytoin Ticlopidine shifted IC 50 (nm) k obs at 10 µm (min ¹1 ) K I (µm) k inact (min ¹1 ) k inact /K I (ml/min/µmol) (0.69) CYP2D6 Bufuralol Paroxetine shifted IC 50 (nm) k obs at 10 µm (min ¹1 ) K I (µm) k inact (min ¹1 ) k inact /K I (ml/min/µmol) (1.10) CYP3A Midazolam Erythromycin shifted IC 50 (nm) 4,355 5,662 k obs at 10 µm (min ¹1 ) K I (µm) k inact (min ¹1 ) k inact /K I (ml/min/µmol) (1.88) Testosterone Erythromycin shifted IC 50 (nm) NT 2,339 k obs at 10 µm (min ¹1 ) NT 0.01 K I (µm) NT 30 k inact (min ¹1 ) NT k inact /K I (ml/min/µmol) NT 1.4 *The value in the parentheses expresses the ratio of the value obtained from the single and the cocktail substrate approach. NT: not tested. Measurement not conducted (because CYP2B6 TDI cannot be evaluated using the single substrate approach). Only TDI parameters for midazolam were calculated for the single substrate approach since CYP3A4/5 TDI is reportedly not substrate dependent. 31) fold-shifts of >6.5, 10.7, 24.2 and 2.9. In addition, ticlopidine demonstrated an IC 50 shift for CYP2B6 and CYP2C19, with foldshifts of 2.7 and 6.3, respectively. Discussion The aim in this study was to validate application of the cocktailsubstrate approach, which we had previously established, to evaluation of compounds that inhibit multiple CYP isoforms through DI or TDI. To characterize the potential of obtaining predictive information about TDI from shifted IC 50 or k obs values (available relatively quickly) without calculating the TDI parameters, we used our cocktail-substrate approach in TDI assays with an IC 50 -shift strategy. Thus, correlations among values of TDI parameters (k inact and K I ), shifted IC 50, and k obs at 10 µm were examined. Furthermore, to improve the efficiency of this IC 50 -shift approach, we conducted assays using dilution and non-dilution methods and compared the resulting shifted IC 50 values. In our previous report, we validated the use of a cocktail substrate in an inhibition assay system by using specific inhibitors for each CYP isoform. 7) However, studies in practical drug discovery have occasionally found that NCEs simultaneously inhibit multiple CYP isoforms. Therefore, to extend the validation of our assay, in the present study, we used compounds known to inhibit multiple CYP isoforms directly and derived IC 50 values for each CYP isoform after employing the substrate-cocktail approach (Table 2). Definitive IC 50 values across a wide range ( µm) were determined for itraconazole (CYP2B6 and CYP3A4/5), fluvoxamine (CYP1A2, CYP2B6, CYP2C9, CYP2C19, and CYP3A4/5), and fluconazole (CYP2A6, CYP2C9, CYP2C19, and CYP3A4/5), and were found to be approximately comparable to values reported in the literature ) These results indicate that just a single run of the assay using a cocktail substrate could properly evaluate the inhibition potential of a compound across multiple CYP isoforms. Recently, more emphasis has been placed on TDI than on DI in

6 Rapid Screening for CYP Inhibition with Substrate Cocktail 203 clinical settings, leading to an increasing need for a TDI assay earlier in the drug discovery process. However, traditional TDI assays using a single substrate are laborious and costly, making it difficult to evaluate a substantial number of NCEs early in drug discovery. If our cocktail substrate strategy was also a viable option for the TDI assays, then the labor and cost per assay would be significantly reduced. Thus, we used prototypical time-dependent inhibitors of major CYPs (CYP1A2, CYP2C9, CYP2C19, CYP2D6, and CYP3A4/5) to compare the TDI parameters (K I and k inact ) for the cocktail- and single-substrate approaches. In our previous report, 7) the evaluation method of DI for 8 CYP isoforms by the cocktail-substrate approach was validated using 8 prototypical direct inhibitors. In the present study, although the TDI of 8 CYP isoforms also could have been validated using 8 prototypical time-dependent inhibitors, we could obtain only 5 time-dependent inhibitors for the 5 most important CYP isoforms in the human liver. However, because it was confirmed via the cocktail substrate approach that ticlopidine achieves TDI not only for CYP2C19 but also for CYP2B6, the TDI for 6 CYP isoforms was validated Fig. 1. Correlations between the cocktail- and single-substrate approaches for (A) k obs at 10 µm, (B) K inact, (C) K I, and (D) K inact /K I using our cocktail substrate approach in the present study. CYP2B6 could not be evaluated using the single substrate approach, and so, 5 CYP isoforms were validated using the single substrate approach. A minor difference (within two-fold) between the two approaches was found for each TDI parameter (Table 3), with a good correlation between the two approaches (Fig. 1). For furafylline, tienilic acid, paroxetine, and erythromycin, TDI parameters were determined only for CYP1A2, CYP2C9, CYP2D6, and CYP3A4/5, respectively (Table 4). However, for ticlopidine, TDI parameters were calculated for CYP2B6 as well as for CYP2C19, indicating that the cocktail substrate approach is a viable assay option for compounds with TDI across multiple CYP isoforms (Table 4). Although FDA guidance 12) and EMA guidelines 13) recommend the prediction of clinical DDIs from TDI parameters (k inact /K I ), this requires a lot of labor and time. Our study found that the cocktail substrate approach is viable for determining parameters not only for DI but also for TDI for multiple CYP isoforms simultaneously, and represents a highly useful strategy in terms of time and cost early in drug discovery. The pharmaceutical industry uses a stepwise approach to effectively evaluate the profiles, including the TDI parameters (k inact and K I values), from a number of NCEs synthesized by medicinal chemists to reduce the tremendous amount of effort, time, and cost. For example, the consensus on CYP inhibition provided recently from Pharmaceutical Research and Manufacturers of America recommends utilization of the IC 50 -shift approach as the first step for the TDI assay. 24) This method involves the use of HLM to evaluate whether the potency of the NCE to inhibit CYP increases when the candidate compound is preincubated in the presence of NADPH. Obach et al. used this IC 50 -shift approach in an evaluation of TDI potential and illustrated a progressive flow strategy in which compounds are selected for further evaluation (to obtain the TDI parameters). 25) Therefore, we also used this IC 50 - shift approach for typical compounds showing TDI in our study to validate the cocktail-substrate approach. This approach was used with either the dilution or non-dilution method. In the dilution method, test compounds were incubated with a high concentration of liver microsomes (initial incubation), followed by dilution of the reaction system, and then incubation with probe substrates (secondary incubation). The non-dilution method followed the same process, except there was no dilution step before the second incubation. When we used our cocktail-substrate approach with the dilution method, we found that the preincubation caused the inhibition curve for each CYP isoform to shift leftward toward lower concentrations (Fig. 2). This leftward shift was also observed with the single-substrate approach (data not shown). In Table 4. Performance of the cocktail-substrate approach with selective TDI for each CYP isoform k inact /K I (ml/min/µmol) UD: undetectable. Time-dependent inhibitor Furafylline Tienilic acid Ticlopidine Paroxetine Erythromycin CYP1A2 (Phenacetin) 56 UD UD UD UD CYP2A6 (Coumarin) UD UD UD UD UD CYP2B6 (Bupropion) UD UD 46 UD UD CYP2C8 (Amodiaquine) UD UD UD UD UD CYP2C9 (Diclofenac) UD 88 UD UD UD CYP2C19 ((S)-Mephenytoin) UD UD 16 UD UD CYP2D6 (Bufuralol) UD UD UD 39 UD CYP3A4/5 (Midazolam) UD UD UD UD 1.6 (Testosterone) UD UD UD UD 1.4

7 204 Kazumasa KOZAKAI, et al. Fig. 2. Inhibition curves obtained using a substrate cocktail with the dilution method Each CYP-specific inactivator was incubated in a separate experiment without pre-incubation (, ---) or with pre-incubation (, ) as described in the Materials and Methods experimental section. The activity is expressed as the percentage of activity remaining compared with that of a control sample containing no inactivators. The results are the means of triplicate experiments. (A) Inhibition of phenacetin O-deethylation by furafylline; (B) Inhibition of bupropion hydroxylation by ticlopidine; (C) Inhibition of diclofenac 4A-hydroxylation by tienilic acid; (D) Inhibition of (S)-mephenytoin 4A-hydroxylation by ticlopidine; (E) Inhibition of bufuralol 1A-hydroxylation by paroxetine; (F) Inhibition of midazolam 1A-hydroxylation by erythromycin; (G) Inhibition of testosterone 6 -hydroxylation by erythromycin. addition, the observed shifted IC 50 values were comparable in both the cocktail- and single-substrate approaches within a 2-fold range (Table 3), and a good correlation was observed between the values for both approaches (data not shown). Although many pharmaceutical industries are adopting this IC 50 -shift approach as the first step in the TDI evaluation, currently a definitive criterion for the degree in the fold shift has not been indicated for NCEs that show positive TDI. Many pharmaceutical companies have been applying the fold-shift criterion of 1.2- to 3-fold, 24) and Berry et al. 26) also concluded that an IC 50 -shift of >1.5 indicated significant TDI potency. Thus, the criterion from the present study appears consistent with these previous reports. Although TDI potency is generally evaluated with full TDI parameters (k inact /K I ), as stated above, obtaining these parameters requires a lot of labor and time. We thus examined whether we could reduce the time and labor by using only a portion of the data necessary to obtain full TDI parameters, the k obs values at 10 µm in our assay, with both the cocktail- and single-substrate approaches. We observed a good correlation between the values of k obs at 10 µm or of the shifted-ic 50 and the full TDI parameters (k inact /K I ) with no differences in results between the cocktail- and singlesubstrate approaches (Fig. 3). Although TDI parameters are required for the prediction of clinical DDIs, these correlations suggest that the values of shifted-ic 50 and k obs at 10 µm could offer limited clinical TDI predictability. Indeed, recent reports have also demonstrated that the values of shifted-ic 50 and k obs at 10 µm correlate with k inact /K I ) In the present study, we also showed that the cocktail-substrate approach was a viable option for TDI evaluation, comparable to the options in these published reports. Therefore, we suggest that for TDI assays, the IC 50 -shift approach using a cocktail substrate is useful early in drug discovery to reduce time, labor, and cost. As described above, IC 50 -shift approaches have been performed using either the dilution or the non-dilution method, and most researchers adopt the dilution method. However, Parkinson et al. recommended that a true increase in CYP inactivation and IC 50 shift can be achieved by assessing TDI using a non-dilution method with a low HLM concentration, since a high HLM con-

8 Rapid Screening for CYP Inhibition with Substrate Cocktail 205 Fig. 3. Comparison of the relationships between k inact /K I and k obs at 10 µm (A, B) or k inact /K I and shifted IC 50 (C, D) using single- and cocktailsubstrates approaches centration leads to considerably less P450 inactivation due to inhibitor depletion and/or binding of the inhibitor to microsomes. 30) Moreover, the non-dilution method is less labor intensive and less expensive than the dilution method. Therefore, we examined whether the cocktail-substrate approach we established could also be applied to the IC 50 -shift assay with the non-dilution method for TDI evaluation. A comparison of fold-shift values for both approaches determined that the values for CYP2C9 and CYP3A were similar, but those for CYP1A2 and CYP2D6 were different. This difference in fold-shift values could have been caused by a difference in the microsomal protein concentration in the initial incubation. Using the dilution method, Parkinson 30) et al. has shown that inhibitor depletion, binding of the inhibitor to microsomes, and/or excessive generation of TDI can occur. Although differences in fold-shift values were found for some typical inhibitors, our results for the cocktail-substrate approach indicated that preincubation shifted the IC 50 values for the inhibitors in both the non-dilution and dilution methods, and that the non-dilution method is a viable option comparable to the dilution method for evaluation of TDI potential (Fig. 4). It is noteworthy that the nondilution method provides a more efficient assay system than the dilution method, since the non-dilution method uses a lower microsome concentration and does not require 2-step reactions, leading to highly efficient TDI evaluation when also using our cocktail-substrate approach. However, a limitation of the nondilution method is a lack of an exclusion process through reversible inhibition obtainable only in the dilution method. Thus, the results Fig. 4. Inhibition curves obtained using the substrate-cocktail approach with the non-dilution method Each CYP-specific inactivator was incubated in a separate experiment without pre-incubation (, ---) or with pre-incubation (, ) as described in the Materials and Methods experimental section. The activity is expressed as the percentage of activity remaining compared with that of a control sample containing no inactivators. The data are expressed as the means of triplicate experiments. (A) Inhibition of phenacetin O-deethylation by furafylline; (B) Inhibition of bupropion hydroxylation by ticlopidine; (C) Inhibition of diclofenac 4A-hydroxylation by tienilic acid; (D) Inhibition of (S)-mephenytoin 4A-hydroxylation by ticlopidine; (E) Inhibition of bufuralol 1A-hydroxylation by paroxetine; (F) Inhibition of midazolam 1A-hydroxylation by erythromycin.

9 206 Kazumasa KOZAKAI, et al. obtained by using the non-dilution approach are included for the TDI (irreversible inhibition) potency as well as the DI (reversible inhibition) potency. Two approaches have been used in CYP inhibition studies in drug discovery at many pharmaceutical companies: (1) a single- or cocktail-substrate (drug probe) approach, and (2) a fluorescent- or luminescent-substrate approach. Each approach has features that make it particularly valuable at different stages of drug discovery. In the initial discovery stage (lead compound discovery), it is most appropriate to use the fluorescent- or luminescent-substrate approach because this provides the highest throughput, and it will be necessary to evaluate a substantial number of compounds for SAR analyses and to exclude NCEs with potent CYP inhibition. Then, in the next stage (lead compound optimization), the cocktail substrate approach is reliable for more accurate evaluation of moderate inhibition, since more attention should be focused on selection of prospective NCEs without clinically relevant CYP inhibitory effects among candidate compounds. Finally, in the development stage after a preclinical study (in which candidate compounds are confirmed), the most reliable single-substrate approach should be applied, since at this stage, reliable rather than rapid results are desired, and accurate prediction of clinical DDIs is critical. Thus, more efficient and reliable evaluation of CYP inhibition will be obtained when an appropriate method is selected for each stage in drug discovery by considering the trade-offs between throughput and reliability. Because this cocktail substrate approach can measure the activity of various kinds of drug-metabolizing enzymes with high sensitivity simultaneously in a single run, it can be utilized not only for CYP inhibition and induction evaluations during the drug discovery stage but also for measuring the activity of drug-metabolizing enzymes in rare and expensive samples [e.g., differentiated hepatocyte-like cells from stem cells such as human induced pluripotent stem cells (ips) or embryonic stem cells (ES)]. In conclusion, we demonstrated that the cocktail-substrate approach that we had previously established has comparable reliability with the traditional single-substrate approach, and allows for the determination of IC 50 values for compounds with DI of 8 CYP isoforms and for determination of k inact and K I values for compounds with TDI of 6 CYP isoforms simultaneously in a single assay. We also demonstrated that the cocktail-substrate approach could be applied to determine k obs at 10 µm and IC 50 -shifts (in dilution and non-dilution methods) for simplification of the TDI assay early in drug discovery. That is, our cocktail-substrate approach was shown to be the most rapid and appropriate highthroughput assay for the evaluation of CYP inhibition, which is required in early drug discovery. It can be applied not only to CYP inhibition assays conducted in HLM but also to CYP inhibition and induction studies conducted in human hepatocytes. References 1) Soars, M. G., McGinnity, D. F., Grime, K. and Riley, R. J.: The pivotal role of hepatocytes in drug discovery. Chem. Biol. Interact., 168: 2 15 (2007). 2) Bachmann, K. A. and Lewis, J. D.: Predicting inhibitory drug-drug interactions and evaluating drug interaction reports using inhibition constants. Ann. Pharmacother., 39: (2005). 3) Wienkers, L. C. and Heath, T. G.: Predicting in vivo drug interactions from in vitro drug discovery data. Nat. Rev. Drug Discov., 4: (2005). 4) Zhou, S. F.: Drugs behave as substrates, inhibitors and inducers of human cytochrome P450 3A4. Curr. Drug Metab., 9: (2008). 5) Friedman, M. A., Woodcock, J., Lumpkin, M. M., Shuren, J. E., Hass, A. E. and Thompson, L. J.: The safety of newly approved medicines: do recent market removals mean there is a problem? JAMA, 281: (1999). 6) Lasser, K. E., Allen, P. D., Woolhandler, S. J., Himmelstein, D. U., Wolfe, S. M. and Bor, D. H.: Timing of new black box warnings and withdrawals for prescription medications. JAMA, 287: (2000). 7) Kozakai, K., Yamada, Y., Oshikata, M., Kawase, T., Suzuki, E., Haramaki, Y. and Taniguchi, H.: Reliable high-throughput method for inhibition assay of 8 cytochrome P450 isoforms using cocktail of probe substrates and stable isotope-labeled internal standards. Drug Metab. Pharmacokinet., 27: (2012). 8) Palmer, J. L., Scott, R. J., Gibson, A., Dickins, M. and Pleasance, S.: An interaction between the cytochrome P450 probe substrates chlorzoxazone (CYP2E1) and midazolam (CYP3A). Br. J. Clin. Pharmacol., 52: (2001). 9) Ono, S., Hatanaka, T., Hotta, H., Tsutsui, M., Satoh, T. and Gonzalez, F. J.: Chlorzoxazone is metabolized by human CYP1A2 as well as by human CYP2E1. Pharmacogenetics, 5: (1995). 10) White, R. E.: High-throughput screening in drug metabolism and pharmacokinetic support of drug discovery. Annu. Rev. Pharmacol. Toxicol., 40: (2000). 11) Riley, R. J., Grime, K. and Weaver, R.: Time-dependent CYP inhibition. Expert Opin. Drug Metab. Toxicol., 3: (2007). 12) US FDA: Guidance for Industry: Drug Interaction Studies-Study Design, Data Analysis, and Implications for Dosing and Labeling Recommendations. Rockville, Food and Drug Administration, 2012, pp ) EMA: Guideline on the investigation of drug interactions. London, UK, European Medicines Agency, 2012, pp ) US FDA: Guidance for Industry: Drug Interaction Studies-Study Design, Data Analysis, and Implications for Dosing and Labeling. Rockville, Food and Drug Administration, 2006, pp ) Walsky, R. L., Astuccio, A. V. and Obach, R. S.: Evaluation of 227 drugs for in vitro inhibition of cytochrome P450 2B6. J. Clin. Pharmacol., 46: (2006). 16) Peng, C. C., Templeton, I., Thummel, K. E., Davis, C., Kunze, K. L. and Isoherranen, N.: Evaluation of 6 -hydroxycortisol, 6 -hydroxycortisone, and a combination of the two as endogenous probes for inhibition of CYP3A4 in vivo. Clin. Pharmacol. Ther., 89: (2011). 17) Lukkari, E., Taavitsainen, P., Juhakoski, A. and Pelkonen, O.: Cytochrome P450 specificity of metabolism and interactions of oxybutynin in human liver microsomes. Pharmacol. Toxicol., 82: (1998). 18) Hesse, L. M., Venkatakrishnan, K., Court, M. H., von Moltke, L. L., Duan, S. X., Shader, R. I. and Greenblatt, D. J.: CYP2B6 mediates the in vitro hydroxylation of bupropion: potential drug interactions with other antidepressants. Drug Metab. Dispos., 28: (2000). 19) von Moltke, L. L., Greenblatt, D. J., Duan, S. X., Schmider, J., Wright, C. E., Harmatz, J. S. and Shader, R. I.: Human cytochromes mediating N- demethylation of fluoxetine in vitro. Psychopharmacology (Berl.), 132: (1997). 20) Sudsakorn, S., Skell, J., Williams, D. A., O Shea, T. J. and Liu, H.: Evaluation of 3-O-methylfluorescein as a selective fluorometric substrate for CYP2C19 in human liver microsomes. Drug Metab. Dispos., 35: (2007). 21) Taavitsainen, P., Anttila, M., Nyman, L., Karnani, H., Salonen, J. S. and Pelkonen, O.: Selegiline metabolism and cytochrome P450 enzymes: in vitro study in human liver microsomes. Pharmacol. Toxicol., 86: (2000). 22) Draper, A. J., Madan, A. and Parkinson, A.: Inhibition of coumarin 7- hydroxylase activity in human liver microsome. Arch. Biochem. Biophys., 341: (1997). 23) Niwa, T., Shiraga, T. and Takagi, A.: Effect of antifungal drugs on cytochrome P450 (CYP) 2C9, CYP2C19, and CYP3A4 activities in human liver microsomes. Biol. Pharm. Bull., 28: (2005). 24) Grimm, S. W., Einolf, H. J., Hall, S. D., He, K., Lim, H. K., Ling, K. H., Lu, C., Nomeir, A. A., Seibert, E., Skordos, K. W., Tonn, G. R., Van Horn, R., Wang, R. W., Wong, Y. N., Yang, T. J. and Obach, R. S.: The conduct of in vitro studies to address time-dependent inhibition of drug-metabolizing enzymes: a perspective of the pharmaceutical research and manufacturers of America. Drug Metab. Dispos., 37: (2009). 25) Obach, R. S., Walsky, R. L. and Venkatakrishnan, K.: Mechanism-based inactivation of human cytochrome p450 enzymes and the prediction of drug-drug interactions. Drug Metab. Dispos., 35: (2007). 26) Berry, L. M. and Zhao, Z.: An examination of IC50 and IC50-shift experiments in assessing time-dependent inhibition of CYP3A4,

10 Rapid Screening for CYP Inhibition with Substrate Cocktail 207 CYP2D6, and CYP2C9 in human liver microsomes. Drug Metab. Lett., 2: (2008). 27) Mori, K., Hashimoto, H., Takatsu, H., Tsuda-Tsukimoto, M. and Kume, T.: Cocktail-substrate assay system for mechanism-based inhibition of CYP2C9, CYP2D6, and CYP3A using human liver microsomes at an early stage of drug development. Xenobiotica, 39: (2009). 28) Grime, K. H., Bird, J., Ferguson, D. and Riley, R. J.: Mechanism-based inhibition of cytochrome P450 enzymes: an evaluation of early decision making in vitro approaches and drug-drug interaction prediction methods. Eur. J. Pharm. Sci., 36: (2009). 29) Burt, H. J., Galetin, A. and Houston, J. B.: IC50-based approaches as an alternative method for assessment of time-dependent inhibition of CYP3A4. Xenobiotica, 40: (2010). 30) Parkinson, A., Kazmi, F., Buckley, D. B., Yerino, P., Paris, B. L., Holsapple, J., Toren, P., Otradovec, S. M. and Ogilvie, B. W.: An evaluation of the dilution method for identifying metabolism-dependent inhibitors of cytochrome P450 enzymes. Drug Metab. Dispos., 39: (2011). 31) Watanabe, A., Nakamura, K., Okudaira, N., Okazaki, O. and Sudo, K.: Risk assessment for drug-drug interaction caused by metabolism-based inhibition of CYP3A using automated in vitro assay systems and its application in the early drug discovery process. Drug Metab. Dispos., 35: (2007).