Abstract
Irreversible inhibition, characterized as mechanism-based inhibition (MBI), of cytochrome P450 in drugs has to be avoided for their safe use. A comprehensive assessment of drug-drug interaction (DDI) potential is important during the drug discovery process. In the present study, we evaluated the effects of macrolide antibiotics, erythromycin (ERM), clarithromycin (CAM), and azithromycin (AZM), which are mechanism-based inhibitors of CYP3A, on biotransformation of midazolam (MDZ) in monkeys. These macrolides inhibited the formation of 1′-hydroxymidazolam in monkey microsomes as functions of incubation time and macrolide concentration. Furthermore, the inactivation potentials of macrolides (kinact/KI: CAM ≅ ERM > AZM) were as effective as that observed in human samples. In in vivo studies, MDZ was administered orally (1 mg/kg) without or with multiple oral dosing of macrolides (15 mg/kg, twice a day on days 1–3). On day 3, the area under the plasma concentration-time curve (AUC) of MDZ increased 7.0-, 9.9-, and 2.0-fold with ERM, CAM, and AZM, respectively, compared with MDZ alone. Furthermore, the effects of ERM and CAM on the pharmacokinetics of MDZ were also observed on the day (day 4) after completion of macrolide treatments (AUC changes: 7.3- and 7.3-fold, respectively). Because the plasma concentrations of macrolides immediately before MDZ administration on day 4 were much lower than the IC50 values for reversible CYP3A inhibition, the persistent effects may be predominantly caused by CYP3A inactivation. These results suggest that the monkey might be a suitable animal model to predict DDIs caused by MBI of CYP3A.
- DDI, drug-drug interaction
- P450, cytochrome P450
- MBI, mechanism-based inhibition
- MDZ, midazolam
- ERM, erythromycin
- CAM, clarithromycin
- AZM, azithromycin
- HPLC, high-performance liquid chromatography
- kobs, apparent inactivation rate constant
- KI, apparent concentration required for half-maximum inactivation
- kinact, maximum inactivation rate constant
- Cmax, maximum plasma concentration
- Tmax, time to reach Cmax
- AUC, area under the plasma concentration-time curve
- CL/F, apparent oral clearance.
Drug-drug interactions (DDIs) arising from inhibition of cytochrome P450 (P450) may lead to adverse effects, and some drugs with marked DDIs have been withdrawn from the market. The CYP3A family, which is the most abundant P450 isoform expressed in both the liver and intestine, is susceptible to reversible and irreversible inhibition because of its broad substrate specificity. Irreversible inhibition, also referred to as mechanism-based inhibition (MBI), causes long-lasting impairment of P450s even after elimination of the inhibitor from blood or tissues, and it may induce serious drug toxicity through increased exposure to coadministered drugs (Rogers and Prpic, 1998; Schmassmann-Suhijar et al., 1998; Zhou et al., 2005). MBI can be characterized as loss of enzyme activity that is dependent on inhibitor concentration and time (Silverman, 1988). Consequently, the potential of drug candidates to cause MBI of CYP3A is assessed in in vitro experiments evaluating the degree of time-dependent inhibition in the early stage of drug development process. To date, some researchers have reported in vitro assay systems applicable to high-throughput screening (Lim et al., 2005; Watanabe et al., 2007; Mori et al., 2009). A number of methods for quantitative prediction of in vivo DDI potential by time-dependent inhibition have been described previously (Mayhew et al., 2000; Ito et al., 2003; Wang et al., 2004; Galetin et al., 2006; Polasek and Miners, 2006; Obach et al., 2007). In these methods, accurate estimation of an inhibitor concentration, fraction of the victim drug clearance mediated by CYP3A, and degradation rate constant of CYP3A are needed to accurately predict DDIs in humans. In particular, the impact of degradation rate constant of CYP3A on prediction accuracy and precision is significant for substrates of which CYP3A contributes more than 50% to the overall elimination (Galetin et al., 2006).
When compared with evaluation based on in vitro experiments, less effort has been directed toward predicting DDI potential in in vivo animal models because of substantial species differences in substrate specificities and inhibitor sensitivities of drug-metabolizing enzymes. Species-related differences in inhibitor sensitivities of metabolic reactions associated with CYP3A isozymes were demonstrated in in vitro and in vivo studies using rats (Ishigami et al., 2001; Kotegawa et al., 2002). More recently, there has been increasing interest in the use of monkeys as an animal model to predict the clinical DDI potential of drug candidates, presumably because of the similarities in functional activities and amino acid sequences of drug-metabolizing enzymes between monkeys and humans (Komori et al., 1992; Mankowski et al., 1999; Ward and Smith, 2004; Prueksaritanont et al., 2006; Komura and Iwaki, 2008). We previously reported that the DDIs of ketoconazole, an azole antifungal agent, with midazolam (MDZ), fexofenadine, and simvastatin in humans are reproduced in cynomolgus monkeys (Ogasawara et al., 2007, 2009). Regarding monkey CYP3A isozymes, Uno et al. (2007) have reported the expression of CYP3A5 and CYP3A8, which exhibit more than 90% amino acid sequence homology to human CYP3A5 and CYP3A4, respectively, in the monkey liver and intestine. To further assess the suitability of monkeys for DDI studies, we selected cynomolgus monkeys as an animal model to predict MBI potency.
Macrolide antibiotics are active against Gram-positive bacteria and are widely used for the treatment of infections. These drugs are often concomitantly administered with other drugs, and it is well known that some macrolides can be potential sources of clinically severe DDIs, such as interactions with carbamazepine, MDZ, triazolam, and cyclosporine A (Barzaghi et al., 1987; Freeman et al., 1987; Greenblatt et al., 1998). Both erythromycin (ERM) and clarithromycin (CAM) are potent mechanism-based inhibitors of CYP3A, and pretreatment with these macrolides results in significant increases in exposure to coadministered drugs metabolized by CYP3A (Olkkola et al., 1993; Zimmermann et al., 1996; Gorski et al., 1998; Okudaira et al., 2007). In contrast, a relatively small increase has been reported in the case of azithromycin (AZM) (Zimmermann et al., 1996). A proposed mechanism for drug interactions involving macrolides is CYP3A-mediated N-demethylation of macrolides to nitrosoalkane, which forms a stable, inactive complex with CYP3A (Pessayre et al., 1985; Periti et al., 1992).
The aim of the present study was to assess whether MBI-induced clinical DDIs could be reproduced in cynomolgus monkeys. The macrolides ERM, CAM, and AZM were chosen as examples of well established mechanism-based inhibitors of human CYP3A. First, we estimated the IC50 values for reversible CYP3A inhibition and kinetic parameters for CYP3A inactivation of ERM, CAM, and AZM in microsomes from the monkey liver and intestine. We then evaluated the effects of ERM, CAM, and AZM on the pharmacokinetics of MDZ, a marker substrate for CYP3A activity, after oral administration to monkeys to determine whether monkey is a suitable animal model to predict DDIs caused by MBI.
Materials and Methods
Chemicals.
MDZ and CAM were purchased from Wako Pure Chemical Industries (Osaka, Japan). ERM, alprazolam, and methyl cellulose (400 cps) were obtained from Sigma-Aldrich (St. Louis, MO). AZM and 1′-hydroxymidazolam were obtained from LKT Laboratories (St. Paul, MN) and Toronto Research Chemicals Inc. (North York, ON, Canada), respectively. Pooled liver and intestinal microsomes of human (liver microsomes: pooled from 50 donors, intestinal microsomes: pooled from 11 donors) and monkey (liver microsomes: pooled from 5 animals, intestinal microsomes: pooled from 7 animals) were provided by XenoTech, LLC (Lenexa, KS). All other reagents and solvents were of analytical grade and were commercially available.
Animals.
Male cynomolgus monkeys, 4 to 7 years old, were supplied by the Japan Laboratory Animals (Tokyo, Japan). Animals were housed in a temperature- and humidity-controlled room with a 12-h light/dark cycle. Animals were fed a standard animal diet (Laboratory Animal Diet PS; Oriental Yeast Co., Tokyo, Japan); food was provided ad libitum, with the exception of the overnight periods before dosing. Whenever overnight fasting was used before dosing, food was provided at 5 h after MDZ dosing. All procedures for the animal experiments were approved by the Animal Ethics Committee of Mitsubishi Tanabe Pharma Corporation.
Reversible Inhibition of 1′-Hydroxymidazolam Formation in Human and Monkey by Macrolide Antibiotics in Vitro.
All incubations were performed in duplicate. A typical incubation contained 0.08 M potassium phosphate (pH7.4), 1 mM EDTA, 5 mM MgCl2, 1 mM NADPH, 1 μM MDZ, and 0.1 mg/ml microsomal protein in a final volume of 100 μl. The incubation mixture for the determination of IC50 values contained macrolides. The final concentrations of macrolides were 0, 10, 20, 40, 80, and 160 μM. MDZ and the inhibitors were dissolved in acetonitrile. The final concentration of organic solvent in the incubation mixture was 1% (v/v). After preincubation at 37°C for 5 min, the reaction was initiated by addition of NADPH solution. After incubation at 37°C for 5 min (human samples) or 3 min (monkey samples), the reaction was terminated by addition of 100-μl ice-cold acetonitrile containing an internal standard (20 ng/ml alprazolam). After vortex mixing and centrifugation, the supernatant was transferred to a high-performance liquid chromatography (HPLC) vial for HPLC-mass spectrometry analysis.
Time-Dependent Inhibition of 1′-Hydroxymidazolam Formation in Human and Monkey by Macrolide Antibiotics in Vitro.
All incubations were performed in duplicate. Human or monkey microsomal proteins (2 mg/ml) were incubated with macrolides in the reaction mixture containing 0.08 M potassium phosphate buffer (pH7.4), 1 mM EDTA, 5 mM MgCl2, and 1 mM NADPH at 37°C for various time durations. Twenty microliters of incubation mixture was transferred to a tube containing 0.08 M potassium phosphate buffer (pH7.4), 1 mM EDTA, 5 mM MgCl2, 20 μM MDZ, and 1 mM NADPH (total volume of 380 μl). A saturating concentration of MDZ was chosen to measure the remaining CYP3A activity. The MDZ 1′-hydroxylation activity was determined at 37°C for 5 min (human sample) or 3 min (monkey sample), and the reaction was terminated by adding 0.4 ml of ice-cold acetonitrile. The final concentration of organic solvent in the incubation mixture was 0.5% (v/v). After the termination of the reaction, the sample was spiked with 80 μl of internal standard (100 ng/ml alprazolam). After vortex mixing and centrifugation, the supernatant was transferred to the HPLC vial for HPLC-mass spectrometry analysis.
In Vivo Studies.
The oral dosing solution of MDZ was prepared in 0.5% (w/v) methylcellulose aqueous vehicle. The macrolides were suspended in 0.5% (w/v) methylcellulose aqueous vehicle for the oral dosing solutions. The study was performed in the following three phases using four monkeys for each macrolide. In the first phase, to obtain the control values of the pharmacokinetic parameters for MDZ in individual animals, MDZ was administered orally concomitant with the vehicle for macrolide at a dose of 1 mg/kg before exposure to macrolides (day 7). In the second phase, just 1 week after MDZ dosing, the macrolide was administered orally (15 mg/kg) twice a day for 3 days. On days 1 and 3, MDZ was coadministered orally (1 mg/kg) on the first dosing of macrolide in each day. The third phase was designed to evaluate the prolonged effects of macrolide on the pharmacokinetic parameters of MDZ and the reversibility of CYP3A activity. On day 4, the day after completion of macrolide treatment, MDZ was administered orally (1 mg/kg) with concomitant oral dosing of the vehicle for macrolide. MDZ was administered orally (1 mg/kg) again with the vehicle for macrolide on 6 days (day 9) after completion of macrolide treatment. For all studies, blood was collected from the femoral vein at predose, at 0.5, 1, 2, 3, 5, 7, and 24 h after MDZ administration. Plasma was separated from blood by centrifugation.
Analytical Procedure.
Determination of 1′-hydroxymidazolam concentrations.
The concentration of 1′-hydroxymidazolam in the microsomal reaction mixture was measured by using an Agilent 1100 MSD system (Agilent Technologies, Palo Alto, CA) operating in the positive ion electrospray mode with selected ion monitoring. Chromatography was performed by using a Capcell Pak C18 MGII column (3-μm particle size, 2.0 × 100 mm; Shiseido, Tokyo, Japan). The mobile phase consisted of 10 mM acetate ammonium buffer and acetonitrile. The flow rate was 0.3 ml/min, and the initial mobile phase was 95% 10 mM acetate ammonium and 5% acetonitrile. The acetonitrile concentration was increased linearly to 60% over 10 min, and after being increased linearly up to 90% for 1 min, it was maintained for 2 min. Selected ion monitoring was performed at m/z 342 for 1′-hydroxymidazolam and m/z 309 for the internal standard. For the mass spectrometer, the drying gas flow rate was set at 12 l/min, nebulizer pressure was set at 20 psig, gas temperature was set at 330°C, and capillary voltage was set at 3000 V. The fragmentor voltages for 1′-hydroxymidazolam and the internal standard were set at 180 and 160 V, respectively. The concentrations of 1′-hydroxymidazolam were determined by its peak area ratio relative to an internal standard with reference to a standard curve using the ChemStation version 10.02 software system (Agilent).
Determination of MDZ and macrolide concentrations in plasma sample.
Twenty microliters of acetonitrile, 10 μl of internal standard solution containing 100 ng/ml alprazolam, and 450 μl of dipotassium hydrogen phosphate solution (1 mol/l) were added to a 50-μl aliquot of the plasma sample. The mixture was extracted with 5 ml of tert-butyl methyl ether. The organic layer was separated, evaporated to dryness under a stream of nitrogen, and reconstituted in 10 mM acetate ammonium/acetonitrile [75:25 (v/v)]. The concentrations of MDZ and macrolides were determined using an Agilent 1100 MSD system (Agilent Technologies). Chromatography was performed by using a Capcell Pak C18 MGII column (3-μm particle size, 2.0 × 100 mm). The mobile phase consisted of 10 mM ammonium formate buffer and acetonitrile for MDZ and CAM, 0.1% acetate and acetonitrile for ERM, and 0.1% formate and acetonitrile for AZM, respectively. The mobile phase was delivered at a flow rate of 0.3 ml/min. The gradient started at 10% acetonitrile (for ERM and AZM) or 25% acetonitrile (for MDZ and CAM) and changed linearly to 90% acetonitrile in 10 min, and after being maintained at 90% acetonitrile for 3 min, it then returned to the initial condition. Selected ion monitoring was performed at m/z 326 for MDZ, m/z 734 for ERM, m/z 748 for CAM, m/z 749 for AZM, and m/z 309 for the internal standard. For the mass spectrometer, the drying gas flow rate was set at 12 l/min, nebulizer pressure was set at 20 psig, gas temperature was set at 330°C, and capillary voltage was set at 3000 V. The fragmentor voltages for MDZ, ERM, CAM, AZM, and the internal standard were set at 160, 160, 160, 220, and 160 V, respectively. The concentrations of MDZ, ERM, CAM, and AZM were determined by its peak area ratio relative to an internal standard with reference to a standard curve using the ChemStation version 10.02 software system (Agilent).
Data Analysis.
The percent activity remaining of MDZ 1′-hydroxy activity was plotted against the range of inhibitor concentrations on a semilog scale. The IC50 values were determined by nonlinear regression analysis with Prism software package (version 3.02; GraphPad Software Inc., San Diego, CA). To determine the inactivation kinetic constants for CYP3A, the natural logarithm of the remaining MDZ 1′-hydroxy activity is plotted against the incubation time. The apparent inactivation rate constant (kobs) is determined from the slope of the initial linear phase. Furthermore, the value of kobs is plotted against the inhibitor concentrations, and the inactivation kinetic parameters (KI and kinact) were determined by the nonlinear least-squares method (Prism software package version 3.02; GraphPad Software Inc.) by using the following equation: kobs = kinact · [I]/(KI + [I]) + kd, where KI, kinact, and kd represent the apparent concentration required for half-maximum inactivation, the maximum inactivation rate constant, and the spontaneous in vitro degradation rate constant of the enzyme in the absence of the inhibitor, respectively.
The in vivo pharmacokinetic parameters for MDZ were calculated in individual animals by noncompartmental analysis by using WinNonlin Professional (version 4.0.1; Pharsight, Mountain View, CA). The maximum plasma concentrations (Cmax) and time to reach Cmax (Tmax) were recorded directly from experimental observations. Estimation of the terminal elimination rate constant (λ) was performed by log-linear regression of the last three concentration-time points. The area under the plasma concentration-time curve (AUC) was calculated using the trapezoidal rule up to the last measurable concentration and extrapolated concentration to infinity using the λ value. The apparent oral clearance (CL/F) was calculated by dividing the administered dose by the AUC.
Statistical Analysis.
The pharmacokinetic parameters, with the exception of Tmax, were logarithmically transformed before the statistical analysis. All statistical tests were performed by using SAS 9.1.3 (SAS Institute, Cary, NC). Two-way analysis of variance with one observation in each individual following Dunnett's multiple comparison test was used to analyze the differences between the pharmacokinetic parameters of MDZ obtained in control (before macrolide treatment) and those during and after macrolide treatment. In all cases, differences were considered statistically significant when p was <0.05.
Results
Reversible Inhibition of 1′-Hydroxymidazolam Formation in Human and Monkey by Macrolide Antibiotics in Vitro.
The effects of macrolides on 1′-hydroxymidazolam formation were evaluated in microsomes obtained from the liver and intestine of both human and monkey. To avoid the influence of time-dependent (irreversible) inhibition as much as possible, all incubations were performed within 5 min after addition of NADPH. Both ERM and CAM inhibited the formation of 1′-hydroxymidazolam in a concentration-dependent manner, and the IC50 values are summarized in Table 1. The IC50 values of ERM and CAM in monkey liver microsomes were 18 and 34 μM, respectively. These values were similar to those obtained with monkey intestinal microsomes. The IC50 values of ERM and CAM in human microsomes were approximately 1.6-fold higher compared with those in respective monkey microsomes. In contrast, the inhibition rates of AZM in monkey and human microsomes were approximately 20% at 160 μM, making it a less potent inhibitor compared with ERM and CAM. The rank order of inhibition potency (ERM > CAM > AZM) for the inhibition of MDZ 1′-hydroxylation was the same between monkey and human microsomes.
Time-Dependent Inhibition of 1′-Hydroxymidazolam Formation in Human and Monkey by Macrolide Antibiotics in Vitro.
The loss of CYP3A activity as functions of macrolide concentration and incubation time was determined with both liver and intestinal microsomes obtained from human as well as monkey. Figures 1, 2, and 3 show the profiles of MDZ 1′-hydroxylation activity remaining in the microsomes after incubation without and with various concentrations of ERM, CAM, and AZM, respectively. The kinetic parameters calculated for CYP3A inactivation are summarized in Table 2. The values of KI and kinact of CAM determined in monkey liver microsomes were similar to those determined in human liver microsomes. The value of kinact for ERM in monkey liver microsomes was equivalent to that in human sample (0.0352 min−1 versus 0.0325 min−1); however, the KI value in monkey was approximately one-half compared with that in human. For each macrolide, the KI values in monkey intestinal microsomes were lower than those in human intestinal microsomes, especially for ERM (5.1 versus 26.5 μM) and CAM (3.5 versus 32.6 μM). The kinact/KI ratios for ERM (2.4 × 10−3 min−1 μM−1) and CAM (8.3 × 10−3 min−1 μM−1) in monkey intestinal microsomes were approximately 3-fold and 10-fold, respectively, higher than those in human intestinal microsomes. The kinact/KI ratios of AZM in monkey microsomes were extremely low compared with those of ERM and CAM, as observed in human samples.
Effects of Multiple Oral Dosing with Macrolide Antibiotics on Pharmacokinetics of MDZ.
To estimate the effects of macrolides on the pharmacokinetics of CYP3A substrates, we determined MDZ concentrations in plasma after oral administration of MDZ to monkeys treated or not treated with macrolides. Figure 4 shows the plasma concentration-time profiles of MDZ administered to monkeys. The pharmacokinetic parameters of MDZ in combination with and in the absence of ERM, CAM, and AZM administered orally are summarized in Tables 3, 4, and 5, respectively. When MDZ was administered orally as well as concomitantly with a single dose of ERM on day 1, the plasma concentration-time profiles of MDZ were significantly affected. The Cmax and AUC of MDZ increased 7.7- and 9.1-fold, respectively, resulting in 82% reduction in CL/F compared with the control (day −7). After multiple oral dosing with ERM, the Cmax and AUC of MDZ increased on day 3 (7.3- and 7.0-fold, respectively), and the effect of ERM on the pharmacokinetics of MDZ was maintained on the day after completion of ERM treatment (day 4). Thereafter, the pharmacokinetic parameters of MDZ obtained on 6 days after completion of ERM treatment (day 9) were almost equivalent to those of the control. A single dose of CAM coadministered with MDZ resulted in a significant reduction (82%) in CL/F for MDZ and corresponding significant increases in Cmax and AUC (6.2- and 6.0-fold, respectively), compared with the control values. Multiple oral dosing with CAM elicited an additional increase in AUC. Furthermore, the significant change in AUC remained on the day after completion of CAM treatment, as observed in the ERM treatment. The effect of CAM on the pharmacokinetics of MDZ disappeared 6 days after completion of CAM treatment. In contrast to the observations from ERM and CAM treatments, a single dose of AZM coadministered with MDZ resulted in a significant increase only in the Cmax of MDZ. Furthermore, a significant difference in the AUC of MDZ was observed on day 3, but not on the day after completion of AZM treatment. Figure 5 presents the plasma concentration-time profiles of ERM, CAM, and AZM coadministered with MDZ on day 3. The plasma concentrations of ERM, CAM, and AZM reached Cmax of 834 ng/ml, 1708 ng/ml, and 719 ng/ml, respectively, on day 3. Plasma concentrations of ERM, CAM, and AZM immediately before MDZ administration on day 4 were 7 ng/ml, 318 ng/ml, and 127 ng/ml, respectively, as indicated at 24 h in Fig. 5. The AUCs of ERM, CAM, and AZM up to 7 h after first dosing on day 3 were 1400 ng · h/ml, 8278 ng · h/ml, and 1681 ng · h/ml, respectively.
Discussion
The potential for drug candidates to be subjected to MBI is typically investigated by in vitro studies using recombinant P450 isozymes or pooled human liver microsomes. In vivo models using experimental animals have not been commonly exploited in evaluation of MBI, possibly because of the interspecies variation in substrate specificities and inhibitor sensitivities of P450 isozymes. However, we considered that the evaluation of MBI in animal models, in combination with in vitro inhibition studies using human enzymes, is more efficient for the selection of drug candidates and would help to make clinical studies safer and more effective. In the present study, we examined the suitability of the cynomolgus monkey as an animal model for evaluation of MBI by using the macrolides ERM, CAM, and AZM, which have different inactivation potentials in human. Because it has been reported that MDZ is metabolized predominantly by CYP3A both in humans and monkeys (Gorski et al., 1994; Kanazu et al., 2004), we used MDZ as a probe substrate for CYP3A. First, we evaluated the reversible inhibition of macrolides for CYP3A in liver and intestinal microsomes of both human and monkey. Even though the IC50 values of ERM and CAM in monkey liver and intestinal microsomes tended to be slightly lower than those in respective human microsomes, there was no substantial difference between monkey and human in the reversible inhibition of CYP3A by macrolides. We then demonstrated that the formation of 1′-hydroxymidazolam in monkey liver and intestinal microsomes was inhibited by these macrolides with different inactivation potentials (kinact/KI: CAM ≅ ERM > AZM) as functions of incubation time and macrolide concentration, as effectively as observed in human samples. The inactivation kinetic parameters, kinact and KI, of each macrolide obtained from human liver microsomes agreed well with those reported previously (Ito et al., 2003). The values of kinact and KI for CAM estimated from the results of in vitro studies using monkey liver microsomes were similar to those estimated from corresponding human samples. However, with the exception of CAM in liver microsomes, the kinact/KI ratios of macrolides in monkey microsomes tended to be higher than those in human samples. In particular, the KI values of macrolides in monkey intestinal microsomes were lower than those in human intestinal microsomes. Uno et al. (2007) identified monkey CYP3A5 and CYP3A8 as P450 isoforms corresponding to human CYP3A5 and CYP3A4, respectively. These authors have also reported the preferential expression of CYP3A5 compared with that of CYP3A8 in the monkey intestine, which is not true for CYP3A4 and CYP3A5 in the human intestine. These differences might lead to species-related differences in MBI potency, and further studies, including evaluation of inactivation potency of macrolides toward the respective CYP3A isoforms, are needed to establish the reasons for species differences observed in intestinal microsomes.
There are remarkable differences in the substrate specificities and inhibitor sensitivities toward P450s between experimental animals and humans. A few reports describing MBI for CYP3A isozymes in experimental animals have been published. Sekiguchi et al. (2008) demonstrated that mibefradil showed marked MBI of MDZ metabolism in in vitro and in vivo experiments using rats. However, they also demonstrated that ERM and CAM did not cause MBI of MDZ metabolism in rat liver microsomes. Takedomi et al. (2001) reported that incubation of rat liver microsomes with ERM for 20 min did not affect the maximal rate and Km value for MDZ metabolism, and that formation of a metabolic intermediate complex leading to irreversible inhibition of CYP3A was not detected in microsomes. In the present study, the inactivation kinetic parameters of macrolides in monkey liver and intestinal microsomes were qualitatively similar to those in corresponding human samples. Taking these results together, we expected that the cynomolgus monkey might be an appropriate in vivo model for evaluation of DDIs caused by MBI of CYP3A.
Some macrolides are most noteworthy for producing clinically significant DDIs with drugs solely metabolized by CYP3A. There is some variation in the severity of DDIs among macrolides (Periti et al., 1992). In the present study, we sought to determine whether these changes in the pharmacokinetics of MDZ were also produced by macrolides in cynomolgus monkeys. As shown in Fig. 4, the plasma concentration-time profiles of MDZ were significantly affected by macrolide treatments, which produced increases in Cmax and AUC and a decrease in CL/F. The AUC of MDZ on day 3 increased 7.0-, 9.9-, and 2.0-fold after multiple dosing with ERM, CAM, and AZM, respectively, compared with the values for MDZ alone. The effect of each macrolide on the pharmacokinetics of MDZ reflected the respective kinact/KI ratio (CAM ≅ ERM > AZM) and systemic exposure of macrolides (CAM > ERM ≅ AZM). The plasma concentrations of MDZ were significantly elevated on the first day of ERM and CAM treatment, and the CL/F of MDZ progressively decreased during repeated dosing with ERM and CAM, indicating that the extent of inhibition of CYP3A by ERM and CAM depends on the duration of administration. Furthermore, the effects of ERM and CAM were maintained on the day after completion of a 3-day macrolide treatment (AUC changes were 7.3 and 7.3-fold with ERM and CAM, respectively). Based on the IC50 value of each macrolide for reversible CYP3A inhibition (Table 1) and the plasma concentrations of macrolides immediately before administration of MDZ on day 4 (ERM, 0.01 μM; CAM, 0.4 μM), the persistent effects of macrolides may be predominantly caused by inactivation of CYP3A.
In clinical DDI study of ERM with MDZ, it has been reported that the AUC of orally administered MDZ was increased 3.4-fold after ERM treatment for 4 days (200 mg, q.i.d) (Okudaira et al., 2007). Gorski et al. (1998) reported that the increase in AUC for orally administered MDZ was 7.0-fold after CAM treatment (500 mg b.i.d.) for 7 days. In contrast, AZM has a smaller potential for MBI of CYP3A compared with other macrolides; it increased the AUC of orally administered MDZ 1.2-fold after multiple oral dosing (500 mg q.d. for 3 days) (Zimmermann et al., 1996). Compared with the dosing regimen in clinical DDI studies, the dosage, 30 mg/kg/day, used in this study is slightly higher than those in human studies, assuming that human body weight is 60 kg (ERM, 13 mg/kg/day; CAM, 17 mg/kg/day; AZM, 8 mg/kg/day). In addition, treatment periods with macrolides, 3 days, is somewhat shorter than those in human. The plasma concentrations of macrolides in monkey during the treatment of macrolides were estimated to be roughly corresponding to those in each clinical study with ERM, CAM, and AZM; Cmax and AUC in human were 0.5 μg/ml and 2.2 μg · h/ml, 2.2 μg/ml, and 10.6 μg · h/ml, and 0.6 μg/ml and 3.2 μg · h/ml, respectively (Periti et al., 1989; Jain and Danziger, 2004; Traunmüller et al., 2007). As a result, the extent and the rank order of AUC changes caused by macrolide treatments were almost identical between monkeys and humans. For drug interaction with ERM, the quantity of ERM in each dose and the treatment period has been shown to be important (Okudaira et al., 2007), and additional studies using monkey with the same dosage regimen in clinical DDI studies are needed to clarify the similarities between monkey and human.
There is increased interest in the MBI of orally administered inactivators at the level of the gut wall, and the MBI at the gut wall further complicates the prediction of DDIs caused by MBI. In a clinical setting, Gorski et al. (1998) have reported more pronounced intestinal inhibition of MDZ metabolism after multiple oral dosing of CAM; the AUC change after an oral dose of MDZ was approximately 2.6-fold greater in comparison to that of i.v. dosing. In the present study, the effects of macrolides on intestinal CYP3A were not separately evaluated in vivo. We previously demonstrated the significant contribution of the intestine to the first-pass metabolism of orally administered MDZ in the monkeys (intestinal availability 2.3%, hepatic availability 69.6%) (Ogasawara et al., 2007). We also indicated the significant inhibition of intestinal CYP3A by orally administered ketoconazole, reversible inhibitor of CYP3A. Taking these findings together, it is possible that the increased Cmax and AUC of MDZ caused by macrolide treatments can be attributed predominantly to inactivation of intestinal CYP3A. A further investigation in which MDZ is administered intravenously as well as concurrently with oral dosing of macrolides is necessary to distinguish the effects of macrolides on hepatic and intestinal CYP3A in monkeys.
In conclusion, we investigated the effects of the macrolides ERM, CAM, and AZM on MDZ metabolism in cynomolgus monkeys in vitro and in vivo. When monkey liver and intestinal microsomes were incubated with macrolides, 1′-hydroxymidazolam formation was inhibited as functions of incubation time and macrolide concentration, as observed in human samples. The plasma concentrations of orally administered MDZ were significantly elevated by macrolide treatment, and the extent and the rank order of AUC changes caused by macrolide treatments were almost identical between monkeys and humans. Furthermore, the effects of macrolides were sustained on the day after completion of macrolides treatments, probably because of CYP3A inactivation. These results suggest that the cynomolgus monkey might be a suitable animal model for the prediction of DDIs caused by MBI of CYP3A.
Acknowledgments.
We wish to express our deep gratitude to Kazuyuki Hirakoso, General Manager of DMPK Research Laboratory, for his interest and encouragement.
Footnotes
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
doi:10.1124/dmd.109.028969
- Received June 12, 2009.
- Accepted August 18, 2009.
- Copyright © 2009 by The American Society for Pharmacology and Experimental Therapeutics