Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The cloning and heterologous expression of the drug-metabolizing human CYPs¹ has resulted in the commercial availability of cDNA-expressed CYPs for use as reagents for in vitro quantitative phenotyping, and bridging the gap between cDNA-expressed cytochromes and human liver microsomes has been the subject of several recent reviews (Remmel and Burchell, 1993; Crespi, 1995; Crespi and Penman, 1997; Rodrigues, 1999).

The turnover number of CYPs in human liver microsomes is affected by several factors that are biochemically distinct from the CYP enzyme itself. Examples include the accessory electron-transfer proteins NADPH cytochrome P450-oxidoreductase (OR) and cytochrome b₅, membrane lipid composition, and ionic strength of the in vitro incubation matrix (Remmel and Burchell, 1993). If the goal of in vitro studies on cDNA-expressed CYPs is to extrapolate the results to human liver microsomal drug metabolism, the experimental conditions with respect to such accessory non-CYP factors should be parallel in the two systems (Remmel and Burchell, 1993). Although factors like ionic strength are easily controlled to ensure similar conditions in both systems, accessory protein concentrations similar to human liver microsomes may not be achievable and thus may well differ between microsomes containing cDNA-expressed CYPs and human liver microsomes. Moreover, heterologously expressed systems contain a single CYP interacting with the accessory proteins, whereas in liver microsomes multiple CYPs may compete for interaction with the accessory proteins. These considerations introduce additional complexity in bridging the gap between expression systems and human liver microsomes, thereby complicating in vitro-in vivo scaling of metabolic rates and relative contributions of the various isoforms.

The substrate concentration-velocity relationship for a hepatic drug biotransformation catalyzed by multiple CYPs can in principle be described as a linear combination of velocity functions for each CYP isoform [v_i(s)] weighted for the relative hepatic abundance of the respective isoforms (A_i):

(1)

The relative contribution of each isoform can then be determined as follows:

(2)

Enzyme kinetic studies using cDNA-expressed human CYPs are being increasingly used to define the functions v_i(s) for whatever individual CYPs biotransform a drug. Such analyses indicate the relative affinities and capacities of the CYPs contributing to a biotransformation. However, the relative contribution of each enzyme cannot be estimated unless the results incorporate the scaling factors (A_i).

Although immunoquantified levels of CYP isoforms in human liver microsomes (Shimada et al., 1994; Code et al., 1997; Ekins et al., 1998; Lasker et al., 1998; Stresser and Kupfer, 1999) are used as estimates of A_i (Yamazaki et al., 1997; Rodrigues, 1999; Shimada et al., 1999), the capacity (turnover number) of an enzyme determined from kinetic studies on cDNA-expressed CYP isoforms may be different from that of the liver microsomal counterpart (Crespi, 1995; Crespi and Miller, 1997; Crespi and Penman, 1997). Thus, A_i should not only reflect the relative abundance of isoform i in human liver microsomes but also account for turnover number differences between liver microsomes and the heterologous expression system. One such estimate of the scaling factor A_i is the relative activity factor (RAF) (Crespi, 1995; Crespi and Penman, 1997).

RAFs are determined for specific CYPs by comparing the rate of an isoform-specific index reaction at saturating substrate concentrations in human liver microsomes (V_max for liver microsomes) to the rate of the same reaction catalyzed by the specific cDNA-expressed CYP under identical conditions (V_max for cDNA-expressed enzyme):

(3)

When the numerator is expressed in picomoles metabolite per milligram of protein per minute and the denominator as picomoles of metabolite per picomole of CYP per minute, the units of RAF are picomoles of CYP per milligram of protein, which can be compared with the immunoquantified CYP contents. RAFs have been used as estimates of A_i in reaction phenotyping (Kobayashi et al., 1997; Nakajima et al., 1999; Roy et al., 1999; von Moltke et al., 1999; Huang et al., 2000), but their quantitative difference from immunoquantified CYP levels is not fully established. In addition, it is not clear whether RAFs determined for a given isoform are independent of the choice of index substrate.

The objectives of the present study were to investigate the biochemical bases of scaling from heterologously expressed CYPs to human liver microsomes and to compare RAF estimates with immunoquantified CYP content for multiple CYP isoforms using lymphoblast-expressed enzymes. We have also tested the substrate-independence assumption of the RAF approach using multiple index substrates of CYP1A2, CYP3A4, and CYP2D6. In addition, we have compared the OR:CYP and cytochrome b₅:CYP ratios in human liver microsomes to those in lymphoblast microsomes containing cDNA-expressed CYPs to provide an explanation for differences between RAFs and immunoquantified protein levels.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Human Liver Microsomes and Heterologously Expressed CYP Isoforms. Liver samples, obtained from the International Institute for the Advancement of Medicine (Exton, PA) or the Liver Tissue Procurement and Distribution Service (University of Minnesota), were from 12 different transplant donors (L1-L12) with no known liver disease. The donor population (median age 27 years, range 3-50 years; 5 females and 7 males) was balanced with respect to age, sex, race, smoking habits, and alcohol consumption. The tissue was partitioned and kept at 80°C until the time of microsome preparation as described previously (von Moltke et al., 1993, 1994).

Antibodies and Quantitative Western Blotting. Concentrations of CYP1A2, 2B6, 2C19, 2D6, and 3A4/5 in human liver microsomal preparations from livers L1 to L12 were determined by quantitative Western blotting (Perloff et al., 1999). Microsomal protein (varying amounts of lymphoblast-expressed CYP standards and an optimal amount of liver microsomal protein) was denatured for 5 min at 100°C in 100 mM Tris buffer (pH 6.8) containing 10% glycerol, 2% -mercaptoethanol, 2% SDS, and 5 µg/ml pyronin Y. Protein was separated by SDS-polyacrylamide gel electrophoresis in precast 7.5% polyacrylamide gels (Bio-Rad Laboratories, Hercules, CA) in 25 mM Tris/0.192 M glycine/0.1% SDS running buffer (pH 8.3) and transferred to Immobilon-P paper (0.45-µm pore size; Millipore, Bedford, MA) by electroblotting at 100 V for 1 h in 25 mM Tris/0.192 M glycine/20% methanol transfer buffer. Blots were blocked, incubated with primary antibody for 1 h, washed, reblocked (for CYPs 3A4 and 1A2 blots only), incubated with HRP-labeled secondary antibody for 1 h, washed again, and the bound HRP signal was activated by enhanced chemiluminescence using the Super Signal Cl-HRP substrate system (Pierce, Rockford, IL). All postantibody washings were done three times (5 min each) in TBS (0.15 M NaCl, 0.04 M Tris Cl, pH 7.7) containing 0.06% Tween 20 (TBS-Tween), with the exception of the final two washes before activation of the HRP signal, which were done in TBS alone. Blots were exposed to film, developed, and quantified by computer-aided densitometry (IMAGE 1.62 image analysis software, National Institutes of Health). A calibration curve of integrated band intensity (the product of band area and band intensity; Y) versus the quantity of CYP standard in picomoles (X) was generated and fit to an empirically determined equation (Y = mX + c for CYPs 1A2 and 2C19; Y = m ln X + c for CYPs 2B6 and 3A4; Y = mX + c or Y = AX^B for CYP2D6, where m and c are slope and intercept terms, respectively). Integrated band densities of liver microsomal samples were used to determine the concentration of CYP per milligram of microsomal protein relative to the calibration curve. Specific conditions of the Western blot assays (electrophoretic conditions, blocking conditions, and antibody dilutions) for each CYP isoform are given in Table 1. All antibodies recognized the target CYP isoform with minimal cross-reactivity, based on data from their manufacturers for the CYP2D6, 3A, and 1A antibodies and specificity studies performed in our laboratory for CYP2B6 and 2C19 antibodies. Although the CYP2C19 antibody showed a minor cross-reactivity with CYP2A6, this did not affect the analysis because the two proteins were electrophoretically resolved under the conditions used.

In Vitro Metabolic Incubations and Metabolite Analysis. Table 2 lists the incubation conditions and metabolite analysis methodology for index reactions used for determination of RAF estimates for CYPs 1A2, 2B6, 2C19, 3A, and 2D6.

Enzyme Kinetic Methods and Determination of RAF Estimates. Standard nonlinear regression-based enzyme kinetic methods were used in the determination of V_max values for ethoxyresorufin O-deethylation (EROD), methoxyresorufin O-demethylation (MROD), S-mephenytoin 4-hydroxylation, and diazepam 3-hydroxylation, both for human liver microsomes (L1-L12) and for the respective lymphoblast-expressed isoforms. RAF estimates were then calculated using eq. 3. The determination of CYP2C19 RAFs for this panel of 12 livers (Venkatakrishnan et al., 1998c) and the evidence for CYP2C19 poor metabolizing phenotype of livers L11 and L12 (Venkatakrishnan et al., 1998a,c) have been described previously. Although a Michaelis-Menten model best described the kinetics of EROD and MROD, the kinetics of diazepam 3-hydroxylation by human liver microsomes and lymphoblast-expressed CYP3A4 were fitted to a one enzyme Hill model (based on convex Eadie-Hofstee plots indicative of autoactivation by substrate). Phenacetin O-deethylation (POD) (von Moltke et al., 1996a) and nortriptyline E-10 hydroxylation (Venkatakrishnan et al., 1999b) displayed biphasic kinetics in human liver microsomes and were described by a two-enzyme model with a high-affinity Michaelis-Menten component and a linear approximation of the low affinity component. The V_max of the high-affinity components of POD and nortriptyline E-10 hydroxylation [previously validated as indices of CYP1A2 (Venkatakrishnan et al., 1998b) and CYP2D6 (Venkatakrishnan et al., 1999b) activity, respectively] were used in the calculation of the CYP1A2 and CYP2D6 RAFs, respectively. In addition to the above-described kinetic method, the CYP1A2 V_max of POD in human liver microsomes was also determined independently by an inhibition method, as the -naphthoflavone-sensitive component. For this, the difference between human liver microsomal POD rate at 1000 µM phenacetin and the rate in the presence of 2 µM -naphthoflavone [validated to specifically and completely inhibit CYP1A2 (Venkatakrishnan et al., 1999a)] was used as the V_max of the CYP1A2 component of POD, for determination of the POD RAF. For determination of CYP2D6 RAFs with bufuralol and desipramine as index substrates, the quinidine (5 µM)-inhibitable components of reaction rate were used, at substrate concentrations of 160 and 100 µM, respectively. For determination of CYP3A RAFs using trazodone (Rotzinger et al., 1998; Zalma et al., 2000) and triazolam (von Moltke et al., 1996b) as index substrates, and for determination of CYP2B6 RAFs using bupropion (Hesse et al., 2000), reaction rates were measured at saturating substrate concentrations (1000, 1000, and 250 µM, respectively).

Measurement of OR Concentrations. Concentrations of OR in human liver microsomes and in microsomes from lymphoblastoid cells expressing CYPs 1A2, 2B6, 2C19, 2D6, and 3A4 were determined using an activity assay that spectrophotometrically measures the rate of NADPH-mediated reduction of cytochrome c (Phillips and Langdon, 1962; Pearce et al., 1996). The 1-ml sample and reference cuvettes each contained 330 mM KH₂PO₄ (pH 7.6), 1 mM KCN, 50 µM cytochrome c (Sigma, St. Louis, MO) and microsomal protein (25-100 µg). Reactions were started by the addition of 10 µl of a 4.2 mM solution of -NADPH (Sigma) to the sample cuvette. The rate of reduction of cytochrome c was determined as the slope of the initial linear part of the curve describing the time-dependent increase in absorbance at 550 nm. Spectra were recorded at room temperature with a dual beam spectrophotometer (Uvikon, Kontron, Zurich). Human liver and lymphoblast microsomal OR concentrations were determined in reference to a calibration curve generated using 0.8, 1.6, 3.2, and 6.4 pmol of purified recombinant human OR (Oxford Biomedical Research, Oxford, MI).

Measurement of Cytochrome b₅ Concentrations. The concentration of cytochrome b₅ in human liver microsomes and microsomes from lymphoblastoid cells expressing CYPs 1A2, 2B6, 2C19, 2D6, and 3A4 were determined from the difference spectrum between NADH-reduced and oxidized microsomes (Estabrook and Werringloer, 1978; Pearce et al., 1996). Microsomes were diluted to 1 mg/ml in 100 mM KH₂PO₄, pH 7.4, and divided between two 1-ml cuvettes. After a baseline of equal light absorbance was obtained between 400 and 500 nm, 5 µl of 20 mM -NADH (Sigma) was added to the sample cuvette, and the difference spectrum was recorded. The concentration of cytochrome b₅ was determined from the absorbance difference between 410 nm (trough) and 425 nm (peak), based on an extinction coefficient of 185 mM¹cm¹.

Reconstitution of CYP1A2 with OR. The effects of increasing molar ratios of OR to CYP1A2 on the rates of EROD and POD were studied by reconstitution experiments (Shimada and Yamazaki, 1998) with purified Escherichia coli-expressed human CYP1A2 (Panvera, Madison, WI) and purified recombinant human OR (Oxford Biomedical Research). CYP1A2 (6.25 pmol) was mixed with varying amounts of OR (to yield molar OR:CYP ratios of 0.05, 0.1, 0.5, 1, 2, 5, and 10) and 5 µg of L--dilauroyl-sn-glycero-3-phosphocholine (added from a 1 mg/ml solution, freshly prepared in 10 mM KH₂PO₄, pH 7.4, by sonication at 4°C) and incubated at 25°C for 5 min with shaking, in a total volume of 55 µl. Reconstituted CYP1A2-OR-lipid mixtures were added to reaction tubes containing substrate (1000 µM phenacetin or 2500 nM ethoxyresorufin) and the prewarmed NADPH regenerating system and incubated with shaking at 37°C. EROD reactions using reconstituted CYP1A2 were performed for 1 min to ensure linearity of product formation rate with time and to avoid direct one-electron reduction of the resorufins to the nonfluorescent semiquinone imine form by OR (Schmalix et al., 1996). Incubations were processed for analysis as described earlier for microsomal incubations.

Statistical Analyses. Multiple linear regression and nonparametric Spearman rank order correlation analyses were performed using the SigmaStat software (SPSS Inc., Chicago, IL). Hypotheses regarding population median ratios were tested using one-sample Wilcoxon signed rank tests.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Immunoquantified CYP Contents, Lymphoblast RAF Estimates, and RAF:CYP Content Ratios. Representative immunoblots and calibration curves are shown in Fig. 1. Median values of immunoreactive content and RAF estimates are provided in Table 3. The median rather than arithmetic mean was chosen as a measure of central tendency due to the skewed nature of the distributions and wide range of values in the sample of 12 livers, especially for CYPs 2B6 and 3A. The ratio of lymphoblast RAF estimate to the immunoquantified level of each CYP isoform for the sample of 12 livers is presented in Fig. 2.

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Quantitative Western blot analysis of CYPs 1A2, 2B6, 2C19, and 3A in human liver microsomes. Representative blots are shown on the left and a representative calibration curve on the right. Only liver microsomal band intensities in the range of the calibration curve were used for quantification. CYP1A2 blot: Lanes 1 through 4 are calibration standards, with 1, 0.5, 0.25, and 0.125 pmol of lymphoblast-expressed CYP1A2 loaded per lane, respectively. Lanes 5 through 10 are varying amounts (6-24 µg) of microsomal protein from livers L1 through L6, respectively. Protein loading was normalized for CYP1A2 activity (570 pmol of ethoxyresorufin O-deethylated per minute at a substrate concentration of 2500 nM). Note the variable band intensity in lanes 5 through 10 despite activity-normalized loading of microsomal protein, indicating interindividual variability in turnover number of human liver microsomal CYP1A2. CYP2B6 blots: For blot 1 (lanes 1-9), lanes 1 though 5 are calibration standards, with 0.1, 0.25, 0.5, 1, and 2.5 pmol of lymphoblast-expressed CYP2B6 loaded per lane, respectively. Lanes 6, 7, 8, and 9 were loaded with 10 µg of microsomal protein from livers L1, L2, L3, and L10, respectively. For blot 2 (lanes 1'-10'), lanes 5' through 10' are calibration standards, with 2.5, 1, 0.5, 0.25, 0.1, and 0.05 pmol of lymphoblast-expressed CYP2B6 loaded per lane, respectively. Lanes 1', 2', 3', and 4' were loaded with 50 µg of microsomal protein from livers L8, L7, L6, and L5, respectively. CYP2C19 blot: Lanes 7 through 11 are calibration standards, with 0.03125, 0.0625, 0.125, 0.25, and 0.5 pmol of lymphoblast-expressed CYP2C19 loaded per lane, respectively. Lanes 1 through 6 were loaded with 15 µg of microsomal protein from livers L12, L11, L4, L3, L2, and L1, respectively. Note the lack of immunodetectable CYP2C19 in livers L12 and L11 (poor metabolizers of S-mephenytoin). CYP3A blot: Lanes 1 through 4 are calibration standards, with 4, 2, 1, and 0.5 pmol of lymphoblast-expressed CYP3A4 loaded per lane, respectively. Lanes 5 through 10 are varying amounts (2-21 µg) of microsomal protein from livers L1 through L6, respectively. Protein loading was normalized for CYP3A activity (13 pmol of triazolam 1-hydroxylated per minute at a substrate concentration of 1000 µM). Note the variable band intensity in lanes 5 through 10 despite activity-normalized loading of microsomal protein, indicating interindividual variability in turnover number of human liver microsomal CYP3A. CYP2D6 blot: Lanes 7 through 11 are calibration standards, with 0.5, 0.25, 0.125, 0.0625, and 0.03125 pmol of lymphoblast-expressed CYP2D6 loaded per lane, respectively. Lanes 1 through 6 were loaded with 5 to 25 µg of microsomal protein from livers L12, L11, L10, L9, L8, and L7, respectively. Note the lack of immunodetectable CYP2D6 in liver L8 (poor metabolizer phenotype).

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Ratio of RAF estimate to immunoquantified CYP content for CYPs 1A2, 2B6, 2C19, and 3A. Multiple index reactions were used in the determination of CYP1A2 and -3A RAFs. BUP, bupropion; SMH, S-mephenytoin 4-hydroxylation. Hollow circles represent individual human livers, with the heavy dashed line denoting the sample median. The solid line is the line of identity between RAF estimates and immunoquantified protein levels, and the dashed upper and lower control lines denote RAF:CYP content ratios of 3 and 0.33, respectively. Note that the RAF:CYP content ratio for CYPs 1A2 and 2B6 lie above the upper control line, whereas the ratios for CYPs 2C19, 3A, and 2D6 lie between the lower and upper control lines, for the majority of the livers.

Concentrations of Accessory Proteins in Human Liver Microsomes and Lymphoblast Microsomes Containing Heterologously Expressed Cytochromes. The concentration of OR in human liver microsomes was 69.8 ± 20 pmol/mg of protein (mean ± S.D., n = 12 livers; range 47-103 pmol/mg of protein). In microsomes from human lymphoblastoid cells coexpressing CYP3A4 and OR, the OR concentration was 35 pmol/mg of protein. In microsomes from human lymphoblastoid cells coexpressing CYP2D6 and OR, the OR concentration was 62.5 pmol/mg of protein. In contrast, microsomes from lymphoblastoid cells expressing CYPs 1A2, 2B6, and 2C19 had OR concentrations that were an order of magnitude lower (3.4, 3.7, and 5 pmol/mg of protein, respectively). The molar ratios of OR to CYP in lymphoblast-expressed CYP1A2, -2B6, and -2C19 were 92-, 29-, and 24-fold lower than their respective values (medians) in human liver microsomes, whereas the OR:CYP3A ratio was similar in the two systems (Table 4 and top panels of Fig. 3). The OR:CYP2D6 ratio in lymphoblast microsomes containing cDNA-expressed CYP2D6 was 3.7-fold lower than the median value in human liver microsomes; however, these values are within the same order of magnitude.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Accessory protein:CYP ratios in human liver microsomes and lymphoblast microsomes containing cDNA-expressed CYP isoforms. Top panels show OR:CYP ratios; lower panels show cytochrome b5:CYP ratios. HLM, human liver microsomes; LYM, lymphoblast microsomes. Hollow circles represent individual human livers, with the dashed line denoting the sample median ratio. Solid dots represent ratios in lymphoblast microsomes.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Representative cytochrome b5 difference spectra showing a representative human liver (with cytochrome b5 content close to the sample median) and lymphoblast microsomes containing CYP3A4. Equal microsomal protein concentrations (1 mg/ml) were used for both preparations. Note the approximately 5-fold lower absorbance difference trough (410 nm) to peak (425 nm) (indicative of cytochrome b5 content) in lymphoblast microsomes compared with human liver microsomes.

Association of Human Liver Microsomal Accessory Protein:CYP Content Ratios with RAF:CYP Content Ratios: Correlation Studies. To test the hypothesis that differences in accessory protein:CYP content ratios between liver microsomes and lymphoblast microsomes may account for the departure of RAFs from immunochemically determined abundance estimates, a statistical approach was used, involving the examination of correlations between OR:CYP ratios or cytochrome b₅:CYP ratios (X), and the ratio of RAF to immunoquantified levels of a particular CYP isoform (Y) in the panel of human liver microsomes studied. Due to the skewed nature of the distributions and the relatively small sample size, a nonparametric Spearman rank order correlation analysis was used.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Representative scatter plots showing the association between OR:CYP ratios and RAF:immunoquantified CYP content ratios in human liver microsomes, for CYPs 1A2 (A), 2B6 (B), 2C19 (C), and 3A (D). Spearman correlation coefficients (rs values) are provided.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Representative scatter plots showing the association between cytochrome b5:CYP ratios and RAF:immunoquantified CYP content ratios in human liver microsomes, for CYPs 1A2 (A), 2B6 (B), 2C19 (C), and 3A (D). Spearman correlation coefficients (rs values) are provided.

View larger version (10K): [in this window] [in a new window]   Fig. 7.   Reconstitution of CYP1A2 with varying amounts of OR. A and B, the concentration-dependent effect of increasing molar ratios of OR and CYP1A2 on EROD and POD activity, respectively. C, the effect of the OR:CYP ratio on the ratio of relative POD activity to relative EROD activity. Arrows, typical OR:CYP1A2 ratios in lymphoblast microsomes containing CYP1A2 and in human liver microsomes (0.04 and 3.5, respectively). Note a modest increase in the ratio of relative activities with increasing OR:CYP1A2 ratio, indicating a small degree of substrate dependence in OR requirement.

Reconstitution of CYP1A2 Activity: Effect of OR:CYP Molar Ratio. Rates of both CYP1A2-catalyzed EROD (Fig. 7A) and POD (Fig. 7B) increased with increasing concentrations of OR, at a fixed concentration of CYP1A2. At OR:CYP1A2 ratios of 0.05 (similar to those in lymphoblast microsomes containing CYP1A2), rates of EROD and POD were near the limit of detection. Reaction rates could be reliably measured only at OR:CYP ratios of 0.1 and higher. With both substrates, the metabolic rate (indicative of CYP1A2 turnover number) at OR:CYP ratios of 1 to 10 (parallel to those in human liver microsomes) were 10- to 20-fold higher than the rate at OR:CYP ratios of 0.1 (close to the ratio in lymphoblast microsomes containing CYP1A2).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Scaling from heterologously expressed CYPs to human liver microsomes requires the use of appropriate scaling factors (A_i values in eqs. 1 and 2). An accurate scaling factor should reflect both the relative abundance of a particular CYP isoform in human liver as well as the differences in turnover number between the cDNA-expressed CYP and its liver microsomal counterpart. This is the conceptual basis of the RAF scaling approach. Thus, the RAF estimate for a CYP will equal its immunoquantified content in liver microsomes only if the turnover number of the enzyme in heterologously expressed form equals that of its liver microsomal counterpart. The turnover numbers of CYP enzymes is known to depend on the concentrations of accessory electron-transfer proteins, which may well differ between cDNA-expressed systems and human liver. However, there has not yet been a systematic quantitative investigation of differences in CYP turnover number between the two systems. Immunoquantified CYP levels continue to be used as scaling factors (Rodrigues, 1999; Shimada et al., 1999), but they may not represent accurate estimates of A_I (Störmer et al., 2000) because they do not incorporate turnover number differences.

The present study has demonstrated that accessory protein:CYP content ratios in some cDNA-expressed CYP preparations are in fact different from those in human liver microsomes and that lymphoblast RAF estimates are significantly different from immunoquantified CYP concentrations for some CYP isoforms. These findings have important implications in bridging the gap between cDNA-expressed CYPs and human liver microsomes and, ultimately, in the accurate determination of relative contributions of individual CYP isoforms to overall clearance of a drug.

Lymphoblast cell lines expressing CYPs 2A6, 2C9, 2D6, 2E1, and 3A4 are engineered to coexpress OR. However, CYPs 1A2, 2B6, and 2C19 have not been successfully coexpressed with OR and are thus dependent on reductase endogenous to the cell line for metabolic activity. Considerable effort was devoted to isolating lymphoblasts coexpressing 1A2, 2B6, and 2C19 with OR; however, stable cell lines with activities higher than the cell lines without OR were not obtained. For the enzymes that could be expressed with OR, differences in CYP expression efficiency also influence the final OR:CYP ratio.

CYP1A2 and 2B6 could not be coexpressed with OR in the lymphoblast system, and the OR:CYP ratios in lymphoblast microsomes containing these cytochromes were over an order of magnitude lower than their respective values in human liver microsomes. Consistent with these observations, the RAF values for CYPs 1A2 and 2B6 were significantly higher than their immunoquantified protein levels. For CYP1A2, it was further established using reconstitution experiments that the OR:CYP ratio differences between liver and lymphoblast microsomes account for the deviation of RAF estimates from immunoquantified CYP1A2 levels. Although cytochrome b₅:CYP1A2 ratios in lymphoblast microsomes were also significantly lower than those in human liver microsomes, this should not influence the turnover number of the enzyme because it is established that cytochrome b₅ does not stimulate CYP1A2 activity (Yamazaki et al., 1997; Shimada and Yamazaki, 1998).

For CYP2C19, accessory protein:CYP ratios were significantly lower in lymphoblast microsomes than in human liver microsomes, but this was reflected as only a 1.5-fold higher RAF in comparison to immunoquantified protein levels. This suggests that the interaction of CYP2C19 with the flavoproteins is saturated at lower accessory protein:CYP ratios.

OR:CYP2D6 ratios in lymphoblast microsomes and human liver microsomes were within the same order of magnitude. This was consistent with similar values of CYP2D6 RAFs and immunochemical content. As with CYP1A2, cytochrome b₅:CYP2D6 ratios in lymphoblast microsomes were significantly lower than those in human liver microsomes. However, this did not affect the CYP2D6 turnover number, a finding consistent with the lack of stimulation of CYP2D6 activity by cytochrome b₅ in reconstituted systems (Yamazaki et al., 1997).

Median values of CYP3A RAFs were similar to immunoquantified CYP3A concentrations. This is consistent with the observation that the median values of both OR:CYP3A and cytochrome b₅:CYP3A ratios in lymphoblast microsomes were similar to those in human liver microsomes.

A large interindividual variability was noted in the RAF:CYP content ratio in the panel of 12 livers for CYPs 1A2, 2B6, and 3A. This indicates that the turnover number for a specific CYP isoform can be different between individuals. This is clearly evident upon examination of the CYP1A2 and -3A immunoblots in Fig. 1. Protein loading for the liver samples was activity-normalized, yet the immunoreactive band intensities show a large variation. Livers with lower turnover number had lower accessory protein:CYP ratios. These interindividual differences can confound the interpretation of CYP content-based results unless pools of human liver microsomes are used (and such differences are thus averaged) but are taken into account by the nature of the RAF approach. This finding suggests that the interindividual variability in hepatic drug metabolism is attributable to differences in the levels of expression of individual CYPs as well as differences in the levels of accessory electron-transfer proteins in relation to CYP content. The in vivo pharmacokinetic implications of this finding require further investigation.

Although the RAFs are expected, in principle, to be quantitatively different from immunoquantified CYP isoform content (Crespi, 1995; Crespi and Penman, 1997) due to potential differences in turnover number of heterologously expressed CYP isoforms from those of their human liver microsomal counterparts, this has not been demonstrated previously. The present study has demonstrated using lymphoblast-expressed CYP isoforms that some RAFs were identical to immunochemically determined CYP content, whereas others were not.

Unlike immunoquantified levels that only indicate hepatic abundance, RAFs also incorporate differences in turnover number between the cDNA-expressed and liver microsomal CYPs. RAFs, rather than immunoquantified CYP content, are thus better initial estimates of scaling factors (A_i values in eqs. 1 and 2) for bridging the gap between cDNA-expressed CYPs and human liver microsomes (Störmer et al., 2000). Although RAF estimation is somewhat dependent on the choice of index substrate (possibly due to substrate-dependent differences in accessory protein requirement), this factor introduces only a 2- to 4-fold difference in the estimated RAF. The 2- to 4-fold difference is similar to differences in mean specific CYP contents that have been reported in the literature. In contrast, much greater differences between RAFs and immunoquantified protein levels (up to an order of magnitude or higher) are demonstrated here using lymphoblast-expressed CYPs as the model system. Use of immunoquantified CYP levels (in place of RAFs) in conjunction with the in vitro kinetic parameters of cDNA-expressed CYPs may yield inaccurate predictions of relative contributions of the various CYPs to overall human liver microsomal metabolic rate, and thus drug clearance in vivo. In the present study, we identified an underestimation of the contribution of CYP isoforms such as CYP1A2 and 2B6 whose turnover numbers in the lymphoblast expression system are significantly lower than those of their human liver microsomal counterparts. Although the present study was restricted to CYPs expressed in the lymphoblast system, the effects observed should be applicable regardless of the expression system used. Indeed, no system may be capable of achieving physiological levels of accessory proteins across the spectrum of xenobiotic metabolizing CYPs. Therefore, consideration of relative enzymatic activities at some level may be unavoidable.