Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Intrinsic clearance (CL_int) is a valuable parameter to describe the inherent metabolic stability of a drug in the absence of other rate-determining processes that may potentially influence its clearance in vivo (Wilkinson and Shand, 1976). For compounds that exhibit classic Michaelis-Menten kinetics this parameter can be readily estimated in vitro from the ratio of V_max and K_m values and scaled to the whole liver to allow in vivo clearance predictions (Houston, 1994). This approach has been successful for the prediction of in vivo CL_int of a diverse range of cytochrome P-450 (CYP) substrates using in vitro data from both freshly isolated hepatocytes and hepatic microsomes (Houston and Carlile, 1997). The parameter CL_int describes the rate of metabolism under first order conditions; this is believed to be of most relevance as therapeutic concentrations rarely approach the K_m values commonly determined in vitro. However, for a drug displaying sigmoidal kinetics due to substrate activation, there is no first order region in the kinetic profile; this precludes the use of the standard approach for scaling the clearance of these compounds to the in vivo situation. The extent of activation can be substantial, thus the consequences of selecting an arbitrary substrate concentration and ignoring the phenomenon may result in a major underestimation of the clearance process. A number of CYP3A substrates have been reported to exhibit sigmoidal kinetics in microsomes (Ueng et al., 1997) but this distinctive activation phenomenon has yet to be demonstrated with other in vitro systems. CYP3A4 is the most abundant CYP present in the human liver (Shimada et al., 1994) and is responsible for the metabolism of at least 50% of the drugs currently in use (Benet et al., 1996). It is therefore important to devise and evaluate a procedure to allow the scaling of in vitro data for drugs showing sigmoidal kinetics and to provide a prediction of in vivo clearance.

Dextromethorphan (DEM) is an antitussive drug that undergoes two oxidative demethylations. The O-demethylation yielding dextrorphan (DOR) is mainly catalyzed by enzymes in the CYP2D subfamily and the N-demethylation to methoxymorphinan (MEM) is mainly catalyzed by members of the CYP3A subfamily in both humans (Kupfer et al., 1986; Coutts et al., 1994) and rats (Zysset et al., 1988; Kerry et al., 1993). Both primary metabolites can be further demethylated to yield 3-hydroxymorphinan (Fig. 1). In the rat in vivo, DOR is the main urinary metabolite (20%) and HOM and MEM represent 17% and 0.5% of dose, respectively (Bochner et al., 1994). Because Bochner et al. (1994) determined that 3-hydroxymorphinan was almost exclusively formed by the O-demethylation of MEM rather than the N-demethylation of DOR, the ratio of the two primary pathways is approximately one. However, this contrasts with reports using microsomal preparations (Kerry et al., 1993) in which DOR is the dominant pathway representing 95% of total clearance.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   DEM metabolic pathway.

The initial aim of this study is to determine whether sigmoidal kinetics are unique to microsomes by studying DEM metabolism in both hepatic microsomes and freshly isolated hepatocytes in suspension from rat. Furthermore the term CL_max, defined as the clearance at maximal autoactivation of the enzyme, is evaluated as a suitable parameter for scaling data that show sigmoidal metabolite kinetics to predict in vivo CL_int. For the purpose of scaling in vitro data, it would seem prudent to assume that full activation is the condition operating under in vivo conditions despite the lack of specific in vivo studies to confirm or refute this stance. We have evaluated our predictions from in vitro data using in vivo clearance values determined by i.v. infusion of DEM into the hepatic portal vein of rats with appropriate correction for plasma protein binding and red blood cell binding. These investigations also allow examination of the inconsistencies between the importance of the two demethylation pathways in vitro and in vivo.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. DEM, MEM, DOR, and 3-hydroxymorphinan were supplied by Hoffmann-La Roche, Inc. (Basel, Switzerland). All other chemicals were purchased either from the Sigma Chemical Company (Poole, Dorset, UK) or BDH (Lutterworth, Leicester, UK).

Animals and Treatment. Male Sprague-Dawley rats (240-260 g) were obtained from the University of Manchester Biological Services Unit. They were housed two to four per cage on a bedding of sawdust in rooms maintained at a temperature of 20 ± 2°C and a humidity of 45 to 55% with a 12-h light/dark cycle. They had free access to water and Chow Rat Mouse (Agro Foods, Peniston, UK) rat diet.

Hepatic Microsomal Studies. Rats (n = 4) were sacrificed by cervical dislocation and washed microsomes were prepared as described previously (Hayes et al., 1995). Pretreatment of animals with dexamethasone and subsequent preparation of microsomes was carried out as described previously by Carlile et al. (1997).

Hepatocyte Studies. Untreated rats (n = 4) were administered heparin (500 U i.p.) 5 min before being sacrificed by cervical dislocation. The isolated hepatocytes were prepared by collagenase end of lobe perfusion as described by Hayes et al. (1995) with the hepatocytes being finally resuspended in albumin-free William's media E.

(1)

(2)

(3)

(4)

(5)

In Vivo Studies. Untreated rats (n = 9) with previously inserted cannulae in the lineal vein and carotid artery (Harms and Ojeda, 1974) received a loading dose of DEM (9.9, 14.8, and 29.7 mg/kg in saline) into the hepatic portal vein via the lineal vein followed immediately by an infusion of DEM (in polyethylene glycol/ethanol/saline, 2:1:1, v/v) via the same route at a rate of either 8.8, 11.7, or 23.4 mg/h/kg (0.15 ml/h for 3 h). Loading doses and infusion rates were calculated from clearance and volume of distribution values (Ashforth et al., 1995) determined from preliminary experiments (260 ± 32 ml/min/kg and 19.8 ± 2.5 l/kg, respectively). Blood samples were taken at 30-min intervals from the carotid artery up to 3 h to determine steady-state plasma concentration (C_ss). After centrifugation to obtain plasma, the samples were stored at 20°C until analysis. In vivo clearance was calculated by dividing the infusion rate (mg/h/kg) by C_ss (mg/liter).

Assay Procedure for Analysis of DEM, DOR, and MEM. After termination of the reaction of microsomal incubations, the incubate was centrifuged and the supernatant (100 µl) injected onto the HPLC. Hepatocyte incubations (1 ml for 0.5-2.5 µM samples and 0.5 ml for 5-500 µM samples) were incubated with 0.5 ml -glucuronidase with sulfatase activity (200 U/ml in sodium acetate, pH 4.5) for 1 h in a shaking water bath at 37°C to hydrolyze any conjugates. After deconjugation, 10 M sodium hydroxide (10 µl) was added along with codeine (10 µl of 1 mg/ml solution in methanol) as internal standard. The sample was extracted with tert-butyl-methyl ether (5 ml) by rotary mixing for 20 min. The samples were then centrifuged (2500 rpm, 10 min) and the organic layer was removed and evaporated to dryness under a stream of oxygen-free nitrogen at 50°C. The extraction efficiencies were 95.2 ± 5.3% (n = 10), 97.6 ± 2.9%, and 92.1 ± 6.3% for DOR, MEM, and DEM, respectively. The sample was reconstituted in mobile phase (300 µl). Plasma samples were extracted as for hepatocytes with the exception of deconjugation. Standard curves were constructed over the range of 0.01 to 10 nmol for the two metabolites and 0.02 to 10 nmol for DEM; the amount of compound was quantified by interpolation from the standard curve for microsomes or by a peak area ratio method with respect to the internal standard codeine for hepatocyte and plasma samples.

Determination of Plasma Protein Binding and Blood-to-Plasma Ratio. Rat plasma (2.5 ml) was preincubated with DEM (final concentration 0.5-500 µM) in a shaking water bath at 37°C for 10 min. An aliquot (1 ml) was placed in an equilibrium dialysis cell (Dianorm, Munich, Germany) separated by semipermeable membrane (molecular weight cut-off 5000, Dianorm) from a cell containing 1 ml of 0.1 M phosphate buffer, pH 7.4. All incubations were carried out in duplicate. Preliminary experiments had determined that equilibrium was reached after 2 h and after 4 h the buffer and plasma were removed from the cells into clean test tubes and analyzed as described previously. The fraction unbound of DEM in rat plasma was calculated using eq. 6.

(6)

Blood (2 ml) was spiked with DEM (final concentration range 1-10 µM) and preincubated in a shaking water bath at 37°C for 10 min. An aliquot (200 µl) was analyzed as described previously in duplicate. The sample was then spun (2500 rpm, 10 min) and an aliquot of the plasma (200 µl) was analyzed in duplicate. The blood-plasma ratio was calculated from the amount of DEM per milliliter of blood and plasma respectively.

Data Analysis. In vitro data were analyzed using least-squares nonlinear regression with a weighting of 1/y (Siphar, version 3.3, Simed, Creteil, France). Log average substrate concentrations were used for calculation of kinetic parameters thereby correcting for loss of substrate from the incubation due to metabolism. MEM production in both microsomes and hepatocytes and DOR production in hepatocytes was best described by the Hill eq. 7.

(7)

where v and V_max are the observed and maximal rates of metabolism, S₅₀ is the concentration at half V_max, n is the Hill coefficient, and S the substrate concentration.

(8)

(9)

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Kinetics of Metabolite Formation in Hepatocytes and Hepatic Microsomes. MEM production, catalyzed by CYP3A isoforms, displays sigmoidal kinetics in both in vitro systems; this can be clearly seen from the characteristic convex Eadie-Hofstee plots (Fig. 2). Although the production of DOR is best described by a two-site Michaelis-Menten model consisting of a high-affinity, low-capacity site and a low-affinity, high-capacity site in microsomes, it displays sigmoidal kinetics in hepatocytes (Fig. 3). The kinetic parameters obtained by nonlinear regression of the rate data for both pathways in the two in vitro systems are given in Table 1. The S₅₀ value for MEM production is lower in hepatocytes than that in microsomes and there is an approximately 2-fold higher n value in the former system. Very similar S₅₀ values are obtained for both pathways in hepatocytes but the n value is greater for MEM than for DOR. For both in vitro systems, V_max is larger for MEM than DOR production. The secondary metabolite, 3-hydroxymorphinan, was not detected in either the microsomal or hepatocyte incubations.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Typical sets of data describing MEM production in hepatic microsomes (A; 0.5 mg protein/ml, 10-min incubation; ) and isolated hepatocytes (B; 0.75 × 106 cells/ml, 10-min incubation) as a function of total substrate concentration (B; ) or unbound substrate concentration () shown as Eadie-Hofstee plots.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Typical sets of data describing DOR production in hepatic microsomes (A; 0.5 mg protein/ml, 10-min incubation; ) and isolated hepatocytes (B; 0.75 × 106 cells/ml, 10-min incubation) as a function of total substrate concentration () or unbound substrate concentration () shown as Eadie-Hofstee plots.

Substrate Binding within In Vitro Matrices. The extent of binding to hepatic microsomes was measured over the substrate concentration range used for the kinetic studies. As can be seen from Table 2 there is minimal binding of DEM to the microsomes and no concentration dependence is apparent over the 1000-fold range studied.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Hepatocyte-to-media partition coefficient for DEM. Cellular concentrations of DEM calculated using eq. 2.

In Vivo Studies. The binding of DEM to rat plasma proteins was independent of substrate concentration (0.5-500 µM) and the fraction unbound was determined as 0.45 ± 0.06 (mean ± S.D., four concentrations in duplicate). The blood-to-plasma ratio was calculated to be 1.76 ± 0.37 (mean ± S.D., four concentrations in duplicate) over the concentration range studied (1-10 µM).

Scaled In Vitro Parameters. Scaled clearances based on the rate of metabolite formation divided by the substrate concentration for microsomes and hepatocytes are shown as a function of incubation concentration in Fig. 5. This graphical representation shows the progressive increase in the magnitude of the clearance until maximal activation is achieved and, subsequently, saturation of the enzyme system is evident. The value for CL_max was selected to provide a prediction of in vivo CL_int for the pathways displaying sigmoidal kinetics. For DOR production in microsomes CL decreases monotonically from the CL_int, which is the sum of the V_max/K_m values for the two sites. Table 4 shows the scaled CL_int and V_max values for DOR and MEM pathways for both in vitro systems. For both pathways the scaled V_max values are similar in the two in vitro systems, whereas the Cl_int values differ significantly. CL_int (calculated from CL_max) is 5-fold higher in hepatocytes than microsomes for MEM. Conversely, DOR CL_int is 8-fold higher in microsomes (calculated from the sum of the V_max/K_m ratios) than in cells (a CL_max estimate).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Typical sets of data showing clearance (v/S) of DEM by the DOR () and MEM () pathways and as total metabolism (×) in hepatic microsomes (A) and isolated hepatocytes (B). The solid lines show the best lines of fit from using v/S = (Vmax1/Km1 + S) + (Vmax2/Km2 + S) for DOR in microsomes and v/S = Vmax. Sn 1/S50n + Sn for MEM in microsomes and both metabolites in hepatocytes.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Despite the many known complexities of the CYP-450 enzyme system (Lewis, 1996) the kinetics of metabolite formation can frequently be described by simple Michaelis-Menten type models. An exception to this general trend is the kinetic behavior of the most abundant CYP isoform in human liver, CYP3A4. Sigmoidal kinetics, indicative of substrate activation, have been reported in both human liver microsomal and cDNA-expressed systems for substrates of CYP3A4 (Andersons et al., 1994; Kerr et al., 1994; Schmider et al., 1995; Ueng et al., 1997). Activation of this isoform by -naphoflavone can result in the loss of sigmoidal kinetics for several of these substrates (Ueng et al., 1997). It would appear that this particular subfamily of the CYP system has multiple binding sites. Recent investigations (Korzekwa et al., 1998) have indicated that other isoforms also display sigmoidal kinetics. To date sigmoidal kinetics have been observed only in microsomal preparations.

DEM was selected for study due to the simplicity of the metabolite pathway (see Fig. 1) and the known isoform involvement in these pathways. We have demonstrated that MEM production exhibits sigmoidal kinetics in both hepatic microsomes and isolated hepatocytes and the degree of sigmoidicity, as measured by the Hill coefficient, is greater in hepatocytes than in microsomes. Other investigations of MEM formation in rat hepatic microsomes have not observed sigmoidal kinetics, however, two of these studies were carried out in female rats (Zysset et al., 1988; Kerry et al., 1993) and the low CYP3A content in this gender is well known (Cooper et al., 1993). Roos et al. (1991) carried out microsomal incubations using the same strain and gender as our study but the substrate concentration range investigated was not reported. Activation is a phenomenon that occurs at low substrate concentrations and if the substrate concentration range does not extend to low concentrations (<10 µM) then the sigmoidicity may not be detected.

DOR production is best described in microsomes by a two-site Michaelis-Menten model consisting of a high-affinity, low-capacity site and a low-affinity, high-capacity site, as reported earlier by Kerry et al. (1993) and Zysset et al. (1988). Although DOR formation is regarded as a CYP2D pathway, our incubations with dexamethasone-induced microsomes demonstrated induction of DOR, as well as MEM, production providing good evidence for other isoform involvement (e.g., CYP3A) in the former pathway. We also report that the formation of DOR in hepatocytes shows sigmoidal kinetics albeit to a lesser degree than MEM, which may be a consequence of the contribution of the two isoform subfamilies to DOR production. With dexamethasone-induced microsomes, DOR production was best described by a two-site Michaelis-Menten model as was seen for untreated microsomes. Surprisingly, sigmoidicity was not evident for MEM production in dexamethasone-induced microsomes, in contrast to untreated microsomes, and the former data were best described by classical Michaelis-Menten kinetics. This may be due to a difference in the main CYP3A isoforms present in untreated and dexamethasone-induced animals. In the former, CYP3A2 represents 4.9% of total CYP (40 pmol/mg protein) and CYP3A1 is not detectable, whereas in the latter CYP3A1 (660 pmol/mg protein) and CYP3A2 (360 pmol/mg protein) are 30% and 16% of the total CYP content of the liver, respectively (Cooper et al., 1993). It may be speculated that CYP3A2 is the isoform responsible for the sigmoidal kinetic profile for MEM. Interestingly, CYP3A2-mediated acetaminophen oxidation has also been identified as being particularly sensitive to activation by caffeine and the mechanism implicated involves enhancement of electron transfer (Lee et al., 1991, 1997).

As sigmoidal kinetics have not been reported previously for isolated hepatocyte preparations, it is important to establish that this phenomenon is enzyme-related. For example, binding within the cell could result in lower concentrations available to the enzyme than would be anticipated from the initial media concentration, and if this effect were saturable then sigmoidal kinetics could artifactually arise. To determine the importance of saturable binding of DEM within the cell, the uptake of DEM into hepatocytes was investigated and extensive accumulation was shown. However, if it is assumed that this accumulation results solely from cellular binding, and the substrate concentration is expressed as an unbound concentration within the incubation, the rates of production of MEM and DOR in hepatocytes maintain their sigmoidal characteristics (Figs. 2B and 3B). Furthermore, although there is saturation of cellular uptake, this occurs at a relatively high concentration (100 µM; see Fig. 4). Conversely, if a carrier-mediated transport system were responsible for the extensive accumulation of DEM, then the substrate concentrations available for metabolism in the cell would be higher than the initial media concentration. When cell concentrations are used to calculate clearance, a low value is obtained, which is independent of substrate concentration. Because these clearance characteristics were inconsistent with all other data, the possible role of hepatic transporters was not pursued.

For simple Michaelis-Menten models, the use of the V_max/ K_m ratio, as a CL_int parameter has proved to be valuable for scaling in vitro data to a whole liver basis to allow both predictions of in vivo kinetics and comparison between different in vitro systems (Houston and Carlile, 1997). When sigmoidal kinetics is evident, an alternative approach to the scaling of in vitro data is required. Sigmoidal kinetics can be identified by a number of classic graphical plots including the Eadie-Hofstee. Figure 5 shows an alternative v/S (i.e., clearance) versus S plot that has a characteristic maximum. We propose the use of this maximum, CL_max, for scaling in vitro data in a similar manner to CL_int for drugs whose kinetics is characterized by Michaelis-Menten models.

Predicted CL_int were calculated by summing the clearance terms for both pathways. For the two in vitro systems the mean predictions are quite similar, 304 ml/min/SRW and 126 ml/min/SRW for microsomes and hepatocytes, respectively. The range of predictions from both in vitro systems and the range of in vivo values observed are compared in Fig. 6. Although microsomal predictions tend to provide higher predictions than hepatocytes, there is good overlap between the predicted CL_int values from both in vitro systems with the observed in vivo CL_int values.

View larger version (10K): [in this window] [in a new window]   Fig. 6.   Comparison of predicted CLint from hepatic microsomes and isolated hepatocytes with the observed in vivo CLint. Data shown for individual animals in vivo (n = 9) and individual microsomal (n = 4) and isolated hepatocyte (n = 4) preparations.

Consideration of the kinetic parameters describing the individual pathways in the two in vitro systems (Table 4) shows that the scaled V_max values are very similar for both pathways yet there is a large difference in the scaled CL_int values (9-fold and 5-fold for DOR and MEM, respectively). This observation emphasizes that although both in vitro systems are similar in their capacities to produce the two metabolites, the concentration-response relationships (for activation and/or metabolism) differ between the in vitro systems resulting in different clearances. This difference between the systems is particularly evident in Fig. 5, which illustrates the relative clearance for both pathways. Bochner et al. (1994) demonstrated that both pathways are of equal importance in vivo (see introduction) using a dose of 20 mg/kg DEM to ensure plasma concentrations less than 4 µM. It can be seen from Fig. 5A that DOR is the predominant pathway in microsomes and at a substrate concentration of 4 µM the clearance ratio of DOR to MEM is 7.1 ± 2.6. In contrast, the clearance ratio of DOR to MEM in hepatocytes is 0.96 ± 0.17 at 4 µM (Fig. 5B). Also if the relative importance of the pathways is assessed assuming full autoactivation (by the use of Cl_max) then the ratio of DOR to MEM is markedly overestimated (18 ± 5) for microsomes whereas for hepatocytes there is reasonable concordance (0.4 ± 0.1) in vivo. Thus, overall hepatocytes are a better reflector of the relative importance of the metabolic pathways of DEM in vivo than microsomes, although both in vitro systems provide accurate prediction of the total metabolic clearance of DEM.

In conclusion, we report that sigmoidal kinetics is not unique to microsomes and can be demonstrated in isolated hepatocytes. This implies that the activator present in vivo is washed out from both in vitro systems during their preparation. We were unable to extend the range of our in vivo studies to establish whether activation can be observed in vivo. To investigate this would require detailed and judiciously planned studies. However, the currently available information would support the assumption of full activation in vivo is not unreasonable. The maximal clearance in vitro would appear to be a useful term for scaling sigmoidal data to the in vivo situation but its general applicability and the validity of the underlying assumption of full activation of the enzymes under in vivo conditions will require further investigation.