Prediction of Hepatic Clearance of Stereoselective Metabolism of Carvedilol in Liver Microsomes and Hepatocytes of Sprague-Dawley and Cytochrome P450 2D-Deficient Dark Agouti Rats

- Hepatic clearance ( CL h ) of carvedilol (CAR), which is eliminated via stereoselective metabolism by the CYP2D subfamily of cytochromes P450 (CYPs), was predicted using liver microsomes and hepatocytes from Sprague-Dawley (SD) rats and CYP2D-deficient Dark Agouti (DA) rats to determine the usefulness of prediction method. Plasma concentrations of CAR following intravenous injection to DA rats were higher than those in SD rats. The volume of distribution at steady state and total clearance ( CL tot ) of S -CAR were approximately two times greater than those of R -CAR in both strains. CL h predicted from in vitro studies using DA rat liver microsomes was different from that obtained from in vivo studies. In contrast, in vitro CL h prediction using DA rat hepatocytes was nearly identical to the CL h observed in DA rats in vivo , and was lower than that in SD rats. The predicted CL h in vitro using hepatocytes correlated well with the observed CL tot in vivo , which is expected to be nearly the same as CL h . These results suggest that in vitro metabolic studies using hepatocytes are more relevant with regard to stereoselectively predicting CL h of CAR than those using liver microsomes. _______________________________________________________________________________________


PURPOSE
Carvedilol (CAR), a third-generation βadrenoceptor antagonist that also blocks the β1 adrenoreceptor (vasodilatation),(1) is used clinically as a racemic mixture of R-and S-CAR, and exhibits enantioselective pharmacokinetic and pharmacodynamic properties. (2)(3)(4)(5) Both enantiomers have equivalent -blocking activity, but the β-blocking activity of S-CAR is stronger than that of R-CAR. (5,6) In humans, clearance of R-CAR is lower than the clearance of S-CAR, (3,4) and is reduced further in poor metabolizers (PM) with cytochrome P450 (CYP) 2D6 deficiency.(2,3) R-CAR is metabolized to form predominantly 4′-and 5′-hydroxy-CAR by polymorphic CYP2D6. (7) For new drugs in early stages of development for which metabolic processes are not completely elucidated, older drugs for which metabolic profiles have not been reported, drugs for which metabolic pathways are multiple and complex, or drugs for which metabolites are not commercially available, the disappearance rate of a parent compound from an incubation medium is useful for identification of CYP isoform(s) responsible for metabolism and for assessment of linearity of pharmacokinetics. (8)(9)(10)(11) We recently quantitatively characterized stereoselective biotransformation of CAR by human CYP isoforms by measuring disappearance rates of R-CAR and S-CAR in human liver microsomes and recombinant human CYPs. We showed that CYP2D2 predominantly catalyzed R-CAR metabolism, whereas CYP2D2 and CYP3A1/2 catalyzed S-CAR metabolism in Sprague Dawley (SD) rats. (12) Dark Agouti (DA) rats exhibit lower mRNA levels of CYP2D1 and CYP2D2 than SD rats. Furthermore, DA rats have low metabolic activities toward CYP2D6 substrates. These findings suggest that DA and SD rats may be suitable animal models for PM and extensive metabolizers (EM) of CYP2D6 substrates, respectively. (13,14) However, previous studies in our laboratories have demonstrated a discrepancy between in vivo pharmacokinetic profiles in female DA rats and in vitro predicted profiles using liver microsomes for propranolol but not metoprolol.(15) In contrast, rat hepatocytes are useful for the prediction of in vivo pharmacokinetic profiles of some drugs metabolized _________________________________________ by phase Ⅱ enzymes. (16,17) Prediction of hepatic clearance (CLh) in vitro using liver microsomes should use a multi-dimensional approach, including metabolic experiments evaluating CYP and/or phase Ⅱ enzymes such as UDP-glucuronosyltransferase (UGT) and protein binding to plasma and microsomal proteins. However, in vitro metabolic studies using hepatocytes cultured with plasma do not require estimation of protein binding for the prediction of CLh values. Furthermore, as CAR is metabolized by CYPs and UGT in humans, dogs, and mice, (18,19) in vitro metabolic studies using liver microsomes should evaluate both CYPs and UGT, suggesting that in vitro metabolic studies using hepatocytes may be a more convenient approach. However, as UGT contributes less to CAR elimination than to CYP in rats, (19,20) the characteristic feature of CAR metabolism in rats have advantage in comparison of utility of liver microsomes and hepatocytes with plasma for evaluation of CAR elimination. It remains unclear whether liver microsomes and hepatocytes would be useful for the prediction of CLh,in vivo of CAR. Thus, we compared the predicted CLh of CAR using liver microsomes and hepatocytes from SD and CYP2Ddeficient DA rats as a model for human CYP2D6 PM. We also determined whether in vitro studies using liver microsomes and hepatocytes were useful for the prediction of stereospecific CLh.

Materials
R-(+)-and S-(−)-CAR were obtained from Toronto Research Chemicals (Toronto, Canada). Propranolol hydrochloride, used as an internal standard for highperformance liquid chromatography (HPLC) analysis, and Hanks' Balanced Salt Solution were obtained from Sigma Aldrich (St. Louis, MO, USA). NADP + , glucose-6-phosphate (G6P), and glucose-6phosphate dehydrogenase (G6PDH) were purchased from Oriental Yeast (Tokyo, Japan). All other chemicals and solvents were of HPLC grade or the highest purity available.

Animal studies
Male SD rats, 8 weeks of age and weighing between 250 and 290 g, and male DA rats, 8 weeks of age and weighing between 140 and 180 g, were obtained from Kiwa Laboratory Animals Co., Ltd. (Wakayama, Japan) and Shimizu Laboratory Supplies Co., Ltd. (Kyoto, Japan), respectively. Animals were housed in a humidity-controlled room at 24±2°C and had free access to water and a standard diet (MF, Oriental Yeast Co., Ltd., Tokyo). Animals were fasted overnight prior to experiments. CAR was dissolved in saline and administered as a single intravenous bolus at a dose of 4 mg/kg. Blood samples (250 µl for each sample) were collected predose and at 5, 15, 30, 60, and 90 min after the dose via the jugular vein using syringes, placed in heparinized tubes, and immediately centrifuged at 6700 g for 10 min to obtain plasma. Animal experiments were conducted according to the Committee for the Care and Use of Laboratory Animals of the Faculty of Pharmacy of Kindai University (Osaka, Japan).
Preparation of rat liver microsomes and hepatocytes Liver microsomes from SD and DA rats were prepared according to previously described methods. (21) Protein concentrations were determined using a BCA protein assay kit (Pierce Biotechnology, Inc., Rockford, IL, USA).
Hepatocytes were isolated from SD and DA rats by in situ perfusion of the liver with collagenase as reported by Seglen(22) and suspended in Hanks' Balanced Salt Solution. (21) Preparations with greater than 75% cell viability as determined by Trypan blue dye exclusion were used.
Incubation of CAR enantiomers with rat liver microsomes and hepatocytes Incubation of rat liver microsomes or rat primary hepatocytes with R-and S-CAR was conducted under the following conditions. The incubation mixture (total volume, 1 ml) contained 1 mg/ml of rat liver microsomes or 1  10 6 cells/ml of hepatocytes in suspension, 100 mM potassium phosphate buffer (pH 7.4), an NADPH-generating system (1 mM NADP + , 10 mM glucose-6-phosphate, 1 IU/ml glucose-6-phosphate dehydrogenase, and 10 mM magnesium chloride), and 0.05-10 µM R-or S-CAR. The reaction was initiated by adding R-or S-CAR solution following a 5-min pre-incubation at 37°C. At each time point, an aliquot of the reaction mixture (100 µl) in rat liver microsomes or rat primary hepatocytes was placed in a centrifuge tube and the reaction was stopped by addition of 1 M NaOH (100 µl). Seven milliliters of diethyl ether and 20 µl of 2.5 µM propranolol hydrochloride as an internal standard (20 µl) were added and the mixture was shaken for 10 min. After centrifugation at 1200 × g for 10 min, the organic phase was removed and evaporated by vacuum evaporation, and the residue was dissolved immediately in 100 µl of methanol for HPLC analysis.
Determination of plasma protein and microsomal binding and blood to plasma concentration ratio Unbound fraction of plasma binding (fu,plasma) and microsomal binding (fu,microsomes) were determined by equilibrium dialysis as reported previously. (21) Plasma or an in vitro incubation mixture containing rat liver microsomes and CAR (10 µg/ml for plasma and 10 µM for in vitro incubation mixture) without an NADPH-generating system were added to one chamber, and dialysis was performed at 37°C for 8 h using Spectra/Por MWCO 12-14,000 (Spectrum Japan, Ritto, Japan). In preliminary experiments, we confirmed that equilibrium was reached within this incubation period.
Whole blood to plasma concentration ratio (Rb) was determined by comparing the CAR concentration in plasma obtained from blood samples after a single oral dose of CAR (20 mg/kg). For samples obtained from in vitro experiments, chromatographic separation was achieved using an Inertsil ODS-3 column (4.6  100 mm, 4 µm, GL Sciences Inc., Tokyo, Japan) maintained at 40°C. The mobile phase consisted of 100 mM potassium phosphate buffer (pH 6.8):methanol (40:60, v/v), and the flow rate was maintained at 1.0 ml/min. The excitation and emission wavelengths were set at 280 nm and 350 nm, respectively. Standard curves for Rand S-CAR were linear over the concentration range of 0.01-10 µM.

Data Analysis
In vivo pharmacokinetic parameters were calculated by non-compartmental analysis using WinNonlin (Pharsight Co., Cary, NC, USA). Elimination halflife (t1/2) was calculated using ln 2/ke, where ke is the elimination rate constant estimated by least-squares regression analysis of the plasma concentration-time curve. Area under the CAR concentration-time curve from time 0 to infinity (AUC0-∞) was calculated by the trapezoidal rule using the determined values, which were extrapolated to infinity by dividing the last measured plasma concentration (Clast) by ke. Total clearance (CLtot) was estimated by dose/AUC0-∞. Volume of distribution at steady state (Vdss) was calculated by multiplying the mean residence time (MRT), which was estimated by dividing the area under the first moment curve (AUMC0-∞) by AUC0-∞, by CLtot. Enzyme kinetic parameters were estimated using the substrate depletion assay in rat microsomes and hepatocytes as previously described. (11,21) Briefly, the time course of the percentage of remaining CAR was fitted to the first elimination function to estimate the initial substrate depletion rate constant (Kdep). The first elimination function was determined by fitting the values at initial concentrations to the equation (1) using a non-linear least-squares regression analysis using WinNonlin: CLh was estimated from CLint in hepatocytes with (3) or without (4) plasma as follows: Qh, fu,inc, and fb are hepatic blood flow, unbound fraction in the incubation mixture, and unbound fraction in the blood calculated by fu,plasma / Rb, respectively.
The difference in the mean between the groups was analyzed statistically using Student's t-test.

Pharmacokinetics of CAR after intravenous dosing
Plasma concentrations of R-and S-CAR after intravenous injection to SD and DA rats are shown in Figure 1, and the pharmacokinetic parameters are summarized in Table 1. Plasma concentrations of both CAR enantiomers were higher in DA rats than in SD rats. R-CAR concentrations were higher than S-CAR concentrations in both strains. Vdss and CLtot of S-CAR were approximately two times greater than those of R-CAR in both strains.
Enzymatic kinetics of CAR in rat liver microsomes and hepatocytes Depletion profiles of R-and S-CAR in liver microsomes of SD and DA rats are shown in Figure  2. The relationship between initial substrate concentration and Kdep is shown in Figure 5a. Estimated kinetic parameters are summarized in Table 2. Although the calculated CLint for S-CAR was slightly higher than that for R-CAR in both strains, CLint of both enantiomers in DA rats was 42-43% of that in SD rats. Depletion profiles of CAR enantiomers in hepatocytes of SD and DA rats and the relationship between initial substrate concentration and Kdep are shown in Figures 3 and 5b, respectively, with kinetic parameters summarized in Table 3. Kdep values obtained using hepatocytes from SD rats were strongly concentration-dependent, as shown using liver microsomes. However, Kdep values obtained from DA rat hepatocytes, especially for R-CAR, were only slightly concentration-dependent. No marked differences were observed between the calculated CLint for S-CAR and R-CAR in SD rats. However, CLint for S-CAR in DA rats was 2.5 times higher than that for R-CAR. Thus, CLint of R-and S-CAR in DA rats was 9.3% and 20%, respectively, of CLint in SD rats.
Rat plasma was added to hepatocytes to investigate the effects of plasma protein binding. Depletion profiles of R-and S-CAR in SD and DA rat hepatocytes in the presence of rat plasma, and the relationship between initial substrate concentration and Kdep are shown in Figures 4 and 5c, respectively. Kinetic parameters are summarized in Table 4. CLint values in hepatocytes in the presence of plasma were 3-6% of those in the absence of plasma (Table 4) for both enantiomers and both strains, suggesting that plasma protein binding of CAR markedly decreased the free concentration of CAR.    Plasma protein and microsomal binding, and blood to plasma concentration ratio The fu values of R-and S-CAR were 2.4-4.5% in both strains (Table 5). These values were comparable with the decrease in CLint values in hepatocytes in the presence of plasma (3-6%, Table  4). The fu value of S-CAR was 1.3 times greater than that of R-CAR in both strains. The fu values in DA rats were 1.5 times greater than those in SD rats for both enantiomers. There were no differences in Rb and fu in incubation mixtures containing rat liver microsomes between SD rats and DA rats.

Estimation of hepatic clearance
CLh values were calculated from estimated CLint in rat liver microsomes and hepatocytes, as shown in Tables 2-4 by using the well-stirred model. CLh values estimated from experiments using rat liver microsomes were similar for both enantiomers and between SD and DA rats. Thus, CLh calculated from in vitro experiments with liver microsomes from DA rats was different from estimates obtained using in vivo experiments (Table 1). In contrast, CLh values predicted from in vitro experiments with DA rat hepatocytes were almost identical to those observed in vivo and were lower than CLh values in SD rats, reflecting polymorphic pharmacokinetics in vivo. The relationship between predicted CLh values from in vitro hepatocyte experiments with plasma correlated well with observed CLtot in vivo, which is expected to be very similar to CLh (R 2 = 0.7734) ( Figure 6). The predicted CLh values were within the range of unity ± 33%.  Table 1 and the horizontal line shows CLh,in vitro of hepatocytes with plasma in Table 5. The solid and dashed lines represent unity and unity ± 33%, respectively.

DISCUSSION
In our previous study, the quantitative contribution of different CYP isoforms to enantioselective CAR metabolism was evaluated using a substrate depletion assay in humans. We demonstrated that CYP2D6 showed the highest affinity for R-CAR metabolism, followed by CYP3A4, CYP1A2, and CYP2C9. In contrast, metabolism of S-CAR is mainly mediated by CYP1A2, followed by CYP2D6 and CYP3A4.(24) In rats, R-CAR was metabolized primarily by CYP2D2, followed by CYP3A2 and CYP3A1, whereas CYP2D1/2 and CYP3A1/2 catalyzed S-CAR metabolism.(12) Thus, the CYP2D subfamily is the main mediator of enzymatic R-CAR metabolism in both rats and humans. In contrast, S-CAR is metabolized by CYP2D1/2 and CYP3A1/2 in rats and CYP1A2 in humans.
Plasma concentrations of R-and S-CAR after intravenous injection of CYP2D-deficient DA rats were higher than those in SD rats, and the concentrations of R-CAR were higher than S-CAR in both strains. Vdss and CLtot were greater for S-CAR than for R-CAR. These results were similar to previous reports of human pharmacokinetic characteristics of CAR enantiomers,(2,3) confirming the possibility that DA and SD rats are suitable PM and EM animal models with respect to CYP2D6 activity. (13,14) Although liver microsomes are commonly used for evaluation of CYP isoform(s) and for prediction of in vivo pharmacokinetics, (8)(9)(10)(11) reports have shown that in vivo pharmacokinetic profiles of propranolol, but not metoprolol, in female DA rats were different from the profiles predicted in vitro using liver microsomes. (11) Previous studies in our laboratories have shown that rat hepatocytes are accurate predictors of in vivo pharmacokinetic profiles. (16,17) Therefore, in the present study, prediction of CLh of CAR, which is eliminated via plasma protein binding-sensitive stereoselective metabolism by the CYP2D subfamily, using liver microsomes and hepatocytes from SD and DA rats was evaluated to investigate which model is most useful for development of new candidate drugs. When predicted from SD and DA rat depletion profiles of R-and S-CAR in rat liver microsomes, the calculated CLint for S-CAR was slightly higher than that for R-CAR in both strains. However, CLint in DA rats was 42-43% of CLint in SD rats for both enantiomers. In contrast, estimation of CLint using SD and DA rat hepatocytes showed no difference between S-CAR and R-CAR in SD rats. However, CLint for S-CAR in DA rats was 2.5 times higher than that for R-CAR. In addition, CLint of R-and S-CAR in DA rats was estimated at 9.3% and 20% of R-CAR and S-CAR in SD rats, respectively. Furthermore, plasma protein binding affected uptake from blood to hepatic cells. The fu values were comparable to the decrease in CLint values observed in hepatocytes in the presence of plasma. The fu value of S-CAR was higher than that of R-CAR in both strains, and the values in DA rats were higher than those in SD rats for both enantiomers. Thus, the predicted CLh values calculated from CLint in rat liver microsomes were similar both among enantiomers and between SD and DA rats. As such, CLh calculated in vitro using liver microsomes from DA rats was different from in vivo results. In contrast, predicted CLh values calculated from CLint in DA rat hepatocytes were similar to those observed in vivo and were lower than CLint values in SD rats, reflecting polymorphic pharmacokinetics in vivo. In addition, predicted CLint values from in vitro hepatocyte studies correlated with the observed CLtot in vivo, which is expected to be very similar to CLh. CAR is metabolized by UGT, including UGT2B7, UGT2B4, and UGT1A1, following hydroxylation in humans.(28,29) Furthermore, CAR is metabolized by UGT in addition to CYPs in humans, dogs, and mice, but not rats. (18,19) Although the mechanisms for stereoselective CAR metabolism in hepatocytes but not liver microsomes are unknown, possible explanations are as follows: (1) hydroxylated CAR metabolite(s), which inhibit CYP activities, might be eliminated via glucuronidation in rat hepatocytes, but not liver microsomes, and/or (2) CAR might be metabolized by UGT in rats as reported by Ishida et al. (30) that the glucuronidation of R-and S-CAR was observed in rat liver microsomes. There are other possibilities that the differences of the accessibility of CAR to metabolic enzymes or CAR uptake between liver microsomes and hepatocytes could affect the CAR metabolism. Further detailed studies evaluating metabolism in rat hepatocytes are necessary. The relatively high CLint of CAR could contribute to a good relationship between predicted CLh values from in vitro CLint from hepatocytes. A different result in the drug with low hepatic extraction ratio may be obtained, although the in vivo CLint of a low hepatic extraction ratio such as warfarin could be predicted from in vitro CLint from hepatocytes (31). It would be useful to further investigate whether the drugs with low hepatic extraction ratio as well as CAR show good relationships between predicted CLh values from in vitro hepatocyte experiments with plasma and those with observed CLtot in vivo. In addition, in vitro metabolic studies using hepatocytes are expected to be more convenient than using liver microsomes depending on the species differences of metabolism. In vitro metabolic studies using human hepatocytes would be of particular interest and value.
In conclusion, substrate depletion assay using hepatocytes rather than liver microsomes was useful for the prediction of CLh,in vivo of CAR, leading the possibility that this approach could be apply the prediction of CLh,in vitro of the other drugs metabolized by CYPs and phase Ⅱ metabolic enzymes. The data represent the estimate ± standard error obtained from MULTI program. Table 1. Kinetic parameters of R-and S-CAR after intravenous injection to SD and DA rats.
Parameters SD DA R S R S C0 (mg/ml) 5.5 ± 0.9 2.9 ± 0.5 14.8 ± 1.1 6.7 ± 0.4 Table 4. Physiological parameters of R-and S-CAR in SD and DA rats.  The data represent the estimate ± standard error obtained from MULTI program.