Vol. 26, Issue 12, 1202-1212, December 1998

Applications and Limitations of Interspecies Scaling and In Vitro Extrapolation in Pharmacokinetics

Jiunn H. Lin

Department of Drug Metabolism, Merck Research Laboratories

	Abstract

Top Abstract Introduction Interspecies pharmacokinetic... In vitro extrapolation Conclusion References

The search for new drugs is an extremely time-consuming and costly endeavor. Much of the time and cost are expended on generating data that support the efficacy and safety profiles of the drug. Because of ethical constraints, relevant pharmacological and toxicological assessments must be made in laboratory animals and in in vitro systems before human testing can begin. In support of the efficacy and safety evaluation during drug development, two fundamental challenges facing industrial drug metabolism scientists are (1) how to "scale-up" the pharmacokinetic data from animals to humans and (2) how to extrapolate the in vitro data to the in vivo situation. This review examines the applications and limitations of interspecies scaling and in vitro extrapolation in pharmacokinetics.

	Introduction

Top Abstract Introduction Interspecies pharmacokinetic... In vitro extrapolation Conclusion References

The ultimate goal of pharmaceutical companies is to develop novel therapeutic agents for the treatment of diseases. The search for new drugs is an extremely time-consuming and costly endeavor. On average, 8 years will have elapsed and $400 million will have been spent between the discovery of a new drug and the final Food and Drug Administration approval. Much of the time and cost are expended on generating data that support the efficacy and safety profiles of the drug. Because of ethical constraints, relevant pharmacological and toxicological assessments must be made in laboratory animals and in in vitro systems before the human testing can begin. In support of the efficacy and safety evaluation during drug development, two fundamental challenges facing industrial drug metabolism scientists are (1) how to "scale-up" the pharmacokinetic data from animals to humans and (2) how to extrapolate the in vitro data to the in vivo situation. Although it is generally believed that data from animal and in vitro studies can be extrapolated reasonably well to humans by using appropriate pharmacokinetic principles, the extrapolation is far from straightforward (Lin, 1995; Lin and Lu, 1997). The difficulty in extrapolation lies in the many intrinsic differences between animals and humans, as well as the complexity of the whole body, with a great number of interdependent factors.

With the breakthroughs in molecular biology and biochemistry, our knowledge of drug-metabolizing enzyme systems and drug transport systems has advanced greatly in recent years. These advances significantly improve our understanding of the factors that exhibit species differences in pharmacokinetics and of the controlling factors of drug metabolism and transport. This brief review will examine the applications and limitations of interspecies scaling and in vitro extrapolation in pharmacokinetics.

	Interspecies Pharmacokinetic Scaling

Top Abstract Introduction Interspecies pharmacokinetic... In vitro extrapolation Conclusion References

One of the primary objectives of preclinical pharmacokinetics is to generate information describing the absorption, distribution, metabolism, and excretion (ADME¹) processes in animals that can then be used for the extrapolation to human ADME processes. The many intrinsic differences in the ADME processes between animals and humans make extrapolation of animal data very difficult. This section will address some aspects of these processes that can be used legitimately for the extrapolation of animal data as well as for situations in which the extrapolation cannot be justified.

Absorption. Drug absorption is influenced by many physiological factors, but it depends also on the physicochemical characteristics of the drug itself. The physiological factors include gastric and intestinal transit time, blood flow rate, gastrointestinal pH, and first-pass metabolism, while the physicochemical factors are the drug's intrinsic properties, such as (-logKa, or the ionization constant), molecular size, solubility, and lipophilicity. Whereas the physiological factors are subject to species variation, the physicochemical factors are species-independent.

The oral bioavailability of a drug is defined as the fraction of an oral dose that actually reaches the systemic circulation. Because the entire blood supply of the upper gastrointestinal tract passes through the gut wall and the liver before reaching the systemic circulation, the drug may be metabolized by the liver and gut wall during the first passage of absorption. Kinetically, the oral bioavailability (F) can be described as:

(1)

where f_abs is the fraction of dose absorbed from the gastrointestinal lumen and f_g and f_h are the fractions of drug metabolized by the gut wall and liver, respectively, during the first passage of drug absorption (Lin and Lu, 1997). The f_abs of a drug is determined mainly by its permeability across the biomembrane, which is similar among species. Therefore, the f_abs of a drug is expected to be similar among species. However, the f_g and f_h of a drug may differ substantially from one species to another because of their intrinsic differences in metabolism, resulting in differences in bioavailability across species.

This was best exemplified by the observation of marked interspecies differences in the bioavailability of indinavir (MK-639) when the drug was given orally as a solution in 0.05 M citric acid. The bioavailability varied from 72% in dogs to 24% in rats and 19% in monkeys (Lin et al., 1996a). Ultimately, it was discovered that the low bioavailability observed in rats and monkeys was due to extensive hepatic first-pass metabolism. By comparing the drug concentration in the systemic circulation during portal or femoral vein infusion, hepatic first-pass extraction (f_h) was estimated to be 68% in rats. However, in situ studies with isolated intestinal loop preparations in anesthetized rats showed that intestinal first-pass metabolism (f_g) was minimal (<8%). Consistent with the in vivo and in situ studies, in vitro hepatic and intestinal first-pass extraction (f_h and f_g) for rats were estimated to be 55% and 5%, respectively, using the liver and intestinal microsomal V_max/K_M data. Although in vivo hepatic first-pass extraction was not determined for dogs and monkeys, the in vitro values were estimated to be 17% and 65%, respectively, using dog and monkey liver microsomes (Lin et al., 1996a). From eq. 1, and taking the f_g and f_h into account, the extent (f_abs) of indinavir absorbed from the gastrointestinal lumen was quite similar among species (55%-80%). Thus the observed species differences in the bioavailability of indinavir were due mainly to the differences in the magnitude of hepatic first-pass metabolism. When human intestinal and hepatic microsomes were used, the extent of intestinal and hepatic first-pass metabolism of indinavir in humans was estimated to be 5% and 26%, respectively (Chiba et al., 1997). With the extent of absorption (55%-80%) obtained from animal studies, we predicted that the bioavailability of indinavir in patients would be 40%-60%. As predicted, when clinical data became available, the bioavailability of indinavir was found to be approximately 60% (Yeh et al., 1998).

L-365,260, a potent CCK_B (cholecystokinin) receptor antagonist, is another example that shows species similarity in the fraction of drug absorbed (f_abs). The bioavailability of L-365,260 was 14% for rats and 9% for dogs when given orally as a suspension in 0.5% methylcellulose (Lin et al., 1996b). Because the extent of hepatic first-pass metabolism was low and estimated to be 30% for rats and 14% for dogs (Lin et al., 1996b), the limited bioavailability was attributed mainly to poor absorption as a result of low aqueous solubility (<2 µg/ml). When L-365,260 was given as a solution in PEG 600, the bioavailability increased to 50% in rats and 70% in dogs. Taking hepatic first-pass metabolism into consideration, the extent (f_abs) of L-365,260 absorbed from the gastrointestinal lumen was similar between rats and dogs at approximately 80%. With this information at hand, L-365,260 was administered in capsules containing PEG 600 in the subsequent clinical studies. As expected, the formulation resulted in good absorption of L-365,260 in humans. The peak concentration (C_max) and AUC were, respectively, 2.3 µg/ml and 450 µg · min/ml for dogs and 0.5 µg/ml and 148 µg · min/ml for normal human subjects when the same dose (50 mg) of L-365,260 in PEG capsules was given orally to dogs (12 kg) and normal volunteers (70 kg).² The C_max and AUC values were comparable in dogs and humans when compared on a weight-normalized dose basis.

The examples of indinavir and L-365,260 suggest that drug absorption in humans can be extrapolated reasonably well from animal data when information on first-pass metabolism is also available. Indeed, Clark and Smith (1984) reported in a survey that the fraction of dose absorbed (f_abs) from the gastrointestinal lumen for a large variety of drugs is remarkably consistent between animal species and humans. The bioavailability, however, differs substantially among species, presumably as a result of species differences in the magnitude of first-pass metabolism.

Distribution. The rate of distribution of a drug to the organs or tissues is determined by the blood flow perfusing the tissues and the ease with which the drug molecules cross the capillary wall and penetrate the cells. Like most physiological parameters, blood flow and circulation time can be extrapolated across species by use of an allometric equation (Edwards, 1975; Boxenbaum, 1980):

(2)

where y is blood flow or circulation time, w is body weight, a is the allometric coefficient, and b is the allometric exponent. The allometric relationship between blood circulation time (sec) and total body weight (kg) is 21 w^0.21 (Stahl, 1967). The blood circulation time is 15 sec in a 250-g rat and 50 sec in a 70-kg human. This implies that a drug molecule circulates the body four times per min in the rat and only once per min in humans. Similarly, both hepatic and renal blood flow (expressed as ml/min/kg body weight) decrease as the animal size increases. The hepatic blood flow is approximately 70 ml/min/kg for the rat and 20 ml/min/kg for humans (Boxenbaum, 1980), and the renal blood flow is 55 ml/min/kg for the rat and 15 ml/min/kg for humans (Edwards, 1975). Clearly, the smaller animal species deliver drugs faster and more frequently to the organs of elimination: namely, the liver and kidneys. Thus it is expected that smaller animal species would eliminate drugs more rapidly than would humans, particularly for those drugs with high clearance, when compared on a weight-normalized basis.

It is generally believed that only the unbound drug can diffuse across membranes that restrict distribution of a drug from the vascular compartment into tissues and vice versa. Therefore, drug protein binding in plasma and tissues can affect the distribution of drugs in the body. It is well known that the extent of binding of drugs to plasma proteins differs considerably among species. The observed species differences in plasma protein binding may reflect differences in the affinity or the number of binding sites on the protein molecule. Albumin, the major drug binding protein in plasma, is composed of a single polypeptide chain of ~590 amino acids. Although structural and functional homologies of albumin exist among species, there are small differences in the amino acid sequences between humans and animal species (Callan and Sunderman, 1973; Kragh-Hansen, 1981), leading to differences in the binding affinity and sites.

Diflunisal, a nonsteroidal anti-inflammatory drug, is bound extensively to plasma protein and eliminated mainly by conjugation as ester and ether glucuronides in humans and rats (Lin et al., 1985). Although diflunisal exhibits concentration-dependent pharmacokinetics in both rats and humans, the kinetics of the drug are different in these two species. The plasma clearance of diflunisal in rats remains relatively constant over the therapeutic concentration range of 50-150 µg/ml (150-450 µM) (Lin et al., 1985), while the clearance in humans decreases with concentration over this same range (Meffin et al., 1983). In contrast, in vitro binding studies have demonstrated that diflunisal shows nonlinear plasma protein binding in rats over the therapeutic concentration range, but the unbound fraction of diflunisal in plasma remains unchanged in humans over this range (Lin, 1989). A detailed kinetic study has demonstrated that lack of changes in plasma clearance in rats is a consequence of the opposing effects of saturable metabolism and saturable plasma protein binding (Lin et al., 1985), whereas the nonlinear clearance in humans is attributed more likely to the saturable metabolism alone. Further studies with serum albumin revealed that the number of binding sites for diflunisal was different between human and rat albumin, whereas the affinity for albumin in both species was comparable. The number of binding sites, which was 3 for human serum albumin and 1 for rat serum albumin, respectively, varied, while the respective association constants were similar (approximately 4.5 × 10⁵ M¹) (Lin, 1989). Consistent with these observations, displacement studies with binding markers specific to albumin (¹⁴C-diazepam, ¹⁴C-warfarin, and ³H-digitoxin) suggested that diflunisal bound to three discrete binding sites on human serum albumin but only one site on rat serum albumin (Lin, 1989). Although the albumin concentration in plasma (500-600 µM) is similar in humans and rats, these results strongly suggest that a much higher drug concentration is required to saturate plasma protein binding capacity of diflunisal in humans because of the larger number of binding sites.

The volume of distribution, a measure of the extent of drug distribution that is determined by the binding of drugs to tissue as well as plasma proteins, is an important determinant of half-life. It is, therefore, desirable if the volume of distribution in humans can be predicted from that in animals. Kinetically, the simplest quantitative expression relating the volume of distribution (V_d) to plasma and tissue binding (Lin, 1995) is given as:

(3)

where V_p is the plasma volume, V_t is the tissue volume, and f_p and f_t are the fractions of unbound drug in plasma and tissue, respectively. From this relationship, it is evident that the V_d increases when f_p is increased and decreases when f_t is increased.

Rearrangement of eq. 3 yields:

(4)

where V_f is defined as the volume of distribution of unbound drugs. From this equation, it is clear that a change in f_t has a greater impact than f_p on V_f because V_t is much greater than V_p.

Although it is easy to determine the plasma protein binding of drugs, the study of tissue binding is hampered by methodological problems. The technical difficulties associated with the determination of drug binding to tissues are reflected by the very limited amount of published information on the subject. Fichtl et al. (1991) reported that there were striking species differences in plasma protein binding and the V_d of propranolol. The values for V_d varied by more than 20-fold, being lowest in monkeys and highest in rabbits. However, when the V_d was corrected for f_p, the volume of distribution of unbound propranolol, V_f, was virtually the same for all species. Consistent with this, Sawada et al. (1984a) reported that the V_f values of ten basic drugs were quite similar among species, including humans. Based on these results, Fichtl et al. (1991) proposed that the V_f of drugs should be similar in humans and other species. The authors suggested that with knowledge of the V_f from laboratory animals and of f_p from human plasma protein binding determined in vitro, one can predict the V_d (V_f × f_p) in humans before the initial clinical studies are initiated. Unfortunately, this approach is not valid for all drugs. Boxenbaum (1982) compared the pharmacokinetic parameters for 12 benzodiazepines in dogs and humans. Eight of the 12 benzodiazepines had quite different V_f values between the dogs and humans, the differences being as much as sevenfold for lorazepam. The large species differences in the V_f values were also reported for -lactam antibiotics (Sawada et al., 1984b). Thus the species similarity in the V_f of propranolol observed by Fichtl et al. (1991) might simply be fortuitous. In conclusion, the results from the aforementioned examples suggest that the binding to plasma and tissue protein and V_d of drugs in humans cannot be readily extrapolated from animal data.

Metabolism. From an evolutionary standpoint, all mammals are similar because they originate from a common ancestor, yet they have differentiated as a result of their dissimilar environmental adaptions. Biochemistry provides countless examples of similarities and differences between species, the most instructive of which is the structure of cytochrome P450s. Cytochrome P450s appear to have evolved from a single ancestral gene over a period of 1.36 billion years. To date, at least 14 P450 gene families have been identified in mammals (Nelson et al., 1996). Although all of the members of this superfamily possess highly conserved regions of amino acid sequences, there are considerable variations in the primary sequences across species. Profound differences in substrate specificity, however, can arise even with a small change in the amino acid sequences. As a result of the species differences in the amino acid sequences of the isoforms, both the rate of drug metabolism and the metabolite pattern may differ significantly among animal species.

In addition to the species differences in amino acid sequences and substrate specificity, the levels of P450 isoforms may also differ across species. For example, the hepatic enzyme levels of CYP1A, CYP2C, and CYP3A isoforms in rats are approximately 28, 638, and 165 pmol/mg microsomal protein, respectively (De Waziers et al., 1990), and the corresponding values for humans are 37, 55, and 87 pmol/mg microsomal protein (Guengerich, 1995). Antipyrine is the most widely studied probe used for assessing in vivo metabolic functions in animals and humans. The major metabolic pathways of antipyrine in humans are N-demethylation, 4-hydroxylation, and 3-methyl-hydroxylation (Eichelbaum et al., 1982). At least four human cytochrome P450 isoforms (CYP1A2, CYP2C9, CYP2C18, and CYP3A4) have been identified as involved in the metabolism of antipyrine (Engel et al., 1996). Knowing the complexity of antipyrine metabolism and the interspecies differences in the levels of P450 isoforms, it is not surprising that Boxenbaum (1980) failed to extrapolate the intrinsic clearance (V_max/K_M) of antipyrine from animal data to humans when using the allometric approach.

Stevens et al. (1993) compared Phase I and Phase II hepatic drug metabolism activities, using human and monkey liver microsomes. Of the eight P450-dependent activities measured, only N-nitrosodimethylamine N-demethylase activity was not significantly different between the two species. Coumarin 7-hydroxylase activity was higher in humans than in monkeys. In contrast, erythromycin N-demethylase, benzphetamine N-demethylase, pentoxyresorufin O-dealkylase, ethoxycoumarin O-deethylase, and ethoxyresorufin O-deethylase activities were significantly greater in monkey microsomes than those in human microsomes. Of the seven microsomal and cytosolic Phase II activities measured, only 17-ethynyl estradiol glucuronidation was significantly higher in humans. These results are in contradiction with the popular belief that monkey metabolism is comparable to human metabolism.

Similar to cytochrome P450s, uridine diphosphate glycosyltransferases (UGTs) also show species differences. At least ten rat UGTs and eight human UGTs have been defined and characterized to date by cDNA cloning (Clarke and Burchell, 1994). Comparison of the amino acid sequences of all UGTs indicates that they share a common C-terminal domain, but that the N-terminal half of these isoforms is quite variable. Examination of each of the UGT isoforms has revealed that there are interspecies differences in UGT activities, both in quantitative and qualitative aspects.

Zidovudine (AZT), an HIV reverse transcriptase inhibitor, is extensively metabolized in humans but not in rats. Approximately 75% of an oral dose was recovered in human urine as the 5'-O-glucuronide, and 15% was recovered as unchanged drug (Blum et al., 1988). On the other hand, only 2% of an oral dose was recovered as AZT glucuronide in rat urine, whereas approximately 78% of the dose was excreted as unchanged drug (Good et al., 1986). Consistent with the in vivo data, in vitro studies confirmed that human liver UGT catalyzed the glucuronidation of 0.1 mM AZT 10- to 25-fold faster than did rat liver UGT (Resetar and Spector, 1989).

In a recent study, hepatic and intestinal UGT activities in rats and rabbits were investigated by measuring the glucuronidation of 1-naphthol, 2-methylumbelliferone, 4-nitrophenol, 2-hydroxybiphenyl, and 4-hydroxybiphenyl (Vargas and Franklin, 1997). Generally, intestinal UGT activities were higher in rabbits when compared with those of rats, while hepatic activities were much higher in rats than in rabbits. In rats, the activities (nmol/min/mg microsomal protein) in the small intestinal mucosa were much lower than those in liver, with the activities in the intestine representing 5%-15% of hepatic levels. In contrast, the intestinal activities were comparable (70%-100%) to the hepatic activities for most aglycones in rabbits.

In addition to the species differences in catalytic activities of drug-metabolizing enzymes, interspecies differences also exist in enzyme inhibition and induction. Various mechanisms are known to underlie enzyme inhibition, including competition for the catalytic site of the enzyme, noncompetitive (allosteric) interaction with the enzyme, suicide destruction of the enzyme, and competition for cofactors. Among these mechanisms, competitive inhibition is probably the most common. If enzyme inhibition occurs by the interaction of two substrates competing for the same enzyme, the competitive nature of the inhibition will depend on the K_M value of the substrate and the inhibitory constant of an inhibitor (K_i) value of the inhibitor as well as their concentrations at the site of enzyme. Because K_M and K_i values can be different between species, it is expected that the degree of enzyme inhibition would be species-dependent.

Isoforms of the CYP2D subfamily have been isolated from rats and humans and have been shown to have similar substrate specificities. Debrisoquine 4-hydroxylation is specifically catalyzed by these isozymes. The inhibition kinetics of debrisoquine 4-hydroxylase activity by quinidine and one of its diastereoisomers, quinine, have been compared in human and rat liver microsomes (Kobayashi et al., 1989). Both quinidine and quinine are potent competitive inhibitors of debrisoquine 4-hydroxylation. However, quinidine is a more potent inhibitor of this activity in humans than in rats, whereas the reverse is true for quinine. The K_i values of quinidine for debrisoquine 4-hydroxylation in humans and rats were 0.6 and 50 µM, respectively, whereas with quinine, the values were 13 and 1.7 µM, respectively. Similarly, furafylline exhibits species-dependent inhibition of phenacetin O-deethylase activity of liver microsomes (Sesardic et al., 1990). Furafylline, a mechanism-based inhibitor of CYP1A2, is more potent in inhibiting phenacetin O-deethylation in humans than in rats, despite the fact that phenacetin O-deethylation is catalyzed exclusively by CYP1A2 in both species.

Although the fundamental mechanisms of CYP1A induction are qualitatively similar in different species, including mice, rats, rabbits, and humans (McDonnell et al., 1992), there are important quantitative differences in the effectiveness of inducer-receptor coupling. For example, the gastric acid-suppressing drug omeprazole is a CYP1A2 enzyme inducer in humans but has little inductive effect in mice or rabbits (McDonnell et al., 1992; Diaz et al., 1990). Important species differences also exist in the response of other inducible subfamilies of cytochrome P450s. Phenobarbital induces predominately the members of the CYP2B subfamily in rats, whereas in humans it appears that the major form induced belongs to the CYP3A subfamily (Rice et al., 1992). Furthermore, members of the CYP3A subfamily in rats are inducible by the steroidal agent pregnenolone-16-carbonitrile but not by the antibiotic rifampin. The opposite is true in rabbits and humans (Strolin Benedetti and Dostert, 1994; Nebert and Gonzalez, 1990). Thus it should not be assumed that drugs that do not induce P450 enzymes in animals do not have enzyme-inducing capacity in humans, and vice versa. Despite well-known species differences in response to P450 inducers, mice and rats have been used routinely in most pharmaceutical companies to assess the risk of potential drug induction in humans. This type of risk assessment may be of little direct relevance for certain drugs.

In summary, these examples clearly demonstrate that extrapolation of drug metabolism from animals to humans often is fairly difficult, if not impossible, both in the qualitative and quantitative aspects. As will be discussed later, however, reliable extrapolation of drug metabolism can be made from in vitro experiments.

Excretion. Drugs and their metabolites are usually eliminated from the body via urine or bile or, sometimes, both. The relative contribution of biliary and urinary excretion to the overall elimination of drugs depends on the nature of the drugs and the animal species. Generally, biliary excretion predominates in drugs with relatively large molecular weights (>500). One striking feature of many drugs excreted in bile is that their structures are amphipathic in character (i.e. they contain both polar and nonpolar groups). Many lipophilic compounds are excreted into the bile at a higher rate after conjugation with glutathione or glucuronic acid, presumably because the reaction not only increases molecular weight but also adds a polar group.

The amount of an organic chemical that is excreted in bile varies widely among species. In general, mice, rats, and dogs are good biliary excreters, while rabbits, guinea pigs, monkeys, and humans are relatively poor biliary excreters. The species differences in biliary excretion become less marked when the molecular size of the drug being excreted exceeds 700 daltons. The underlying mechanism for the species differences is at present unclear. Species differences in hepatic blood flow and bile flow do not seem to correlate with the biliary excretion of compounds (Smith, 1971 and 1973). Thus it is difficult to predict the biliary excretion of drugs in humans from animal data.

Besides biliary excretion, many drugs are excreted mainly as unchanged drugs by the kidneys. The rate of renal excretion (renal clearance) is dependent on renal blood flow, glomerular filtration rate (GFR), and tubular secretion and reabsorption. The GFR values vary considerably among species because of species differences in the number of nephrons. The GFR values are 10, 8.7, 4.8, 4.0, and 1.8 ml/min/kg for mice, rats, rabbits, dogs, and humans, respectively, and corresponding values of nephron number (per kg body weight) are 5.0 × 10⁵, 2.9 × 10⁵, 1.6 × 10⁵, 0.9 × 10⁵, and 0.29 × 10⁵ (Renkin and Gilmore, 1973). Both the GFR and number of nephrons show a good allometric relationship. Like the GFR, the renal excretion of drugs also shows a good allometric relationship across species. Thus the renal clearance of drugs in humans can be extrapolated from animal data by use of the allometric approach. Ceftizoxime and methotrexate are good examples. Both drugs are excreted mainly as unchanged drugs in the urine. The plasma clearance of ceftizoxime and methotrexate in humans has been extrapolated successfully from animal data (Chapell and Mordenti, 1991).

Although the renal clearance of a drug in humans can be predicted reasonably well by use of the allometric approach, this approach requires at least four or five animal species in order to obtain a proper allometric relationship, thus limiting its practical value in drug development. A more simplistic yet useful alternative to predict human renal clearance is to use the ratio of GFR between rats and humans. As shown in table 1, the ratios of renal clearance of various drugs between rats and humans is roughly equal to the ratio of GFR between these two species (Lin, 1995). These results suggest that with knowledge of the GFR ratio and the renal clearance of a drug in rats, the renal clearance of the drug in humans can be estimated roughly.

View this table: [in this window] [in a new window]   TABLE 1 GFR and renal clearancea

	In Vitro Extrapolation

Top Abstract Introduction Interspecies pharmacokinetic... In vitro extrapolation Conclusion References

In recent years, there has been a large expansion in both the range and use of in vitro systems to study drug metabolism. Because of the simplicity of in vitro systems, they are very useful in studying the factors that influence pharmacokinetics and drug metabolism. A trickier task is to use these in vitro systems to predict in vivo pharmacokinetics and drug metabolism quantitatively. The difficulty in extrapolating in vitro to in vivo data lies in the complexity of the interdependent biological processes and their dynamic nature. Therefore, it is important to carefully set up the in vitro experimental conditions that simulate the in vivo situations and to understand the interdependent factors between biological components that affect each. In addition, a good understanding of pharmacokinetic principles is necessary for the in vitro/in vivo extrapolation.

Metabolite Profiles. In drug development, early information on human metabolism of a new drug is critical in predicting potential clinical drug-drug interactions and in selecting the appropriate animal species for the toxicity studies. For human risk assessment, regulatory agencies require that the systemic exposure of an unchanged drug and its major metabolites in the animal species used in the toxicity study exceeds that expected in humans to ensure a safety margin. It is important, therefore, to select animal species that have metabolite profiles similar to those of humans. However, the in vivo human drug metabolism study normally is not carried out until the later stages of drug development, which is often too late for animal selection. Fortunately, the increased availability of human tissues and advances in bioanalytical and biochemical technologies have provided opportunities for in vitro studies of human metabolism at the early stage of drug development before the toxicity studies are initiated (Wrighton et al., 1993).

In general, the metabolite profile of a drug obtained in vitro quite accurately reflects the in vivo metabolite pattern, although it is limited to qualitative aspects. From the physiological and biochemical points of view, precision-cut liver slices are especially useful when the complete in vitro metabolite profile of a drug is being obtained. This system retains the physiological conditions of enzymes and cofactors of both Phase I and Phase II reactions and, therefore, better simulates the in vivo situation (Dogterom, 1993). The metabolism of indinavir (MK-0639) illustrates this point. The major metabolic pathways of indinavir in humans have been identified as the following: (a) glucuronidation at the pyridine nitrogen to yield a quaternary ammonium conjugate, (b) pyridine N-oxidation, (c) para-hydroxylation of the phenylmethyl group, (d) 3'-hydroxylation of the indan, and (e) N-depyridomethylation. The metabolite profile of indinavir obtained from human liver slices accurately reflects the in vivo human metabolite pattern (Balani et al., 1995). Although all of the oxidative metabolites of indinavir were also found in human liver microsomes, the N-glucuronide was not detected when indinavir was incubated with native or Triton X-100-treated human liver microsomes in the presence of 10 mM uridinediphosphoglucuronic acid (Lin et al., 1996a). The reason for the inability of human liver microsomes to form the N-glucuronide is not clear. Nevertheless, these results suggest that the liver slice is a better in vitro model for the study of the metabolic pathways of drugs.

Although liver slices are valuable in identifying metabolic pathways, their use in obtaining kinetic parameters may be limited. Worboys et al. (1996) showed that the values of CL_int (V_max/K_M) of a series of drugs in slices are consistently less than those in hepatocytes by a factor ranging from 2 to 20. These results strongly suggest that a distribution equilibrium is not achieved between all of the cells within the slice and the incubation medium because of the slice thickness (~260 µm).

Isolated and cultured hepatocytes also are used often as in vitro models for identifying the metabolic pathway of drugs. In vitro metabolism of ketotifen, an antiasthmatic drug, by cultured rat, rabbit, and human hepatocytes was consistent with the in vivo metabolic pathways: namely, oxidation in rat hepatocytes; oxidation, glucuronidation and sulfation in rabbit hepatocytes; and reduction and glucuronidation in human hepatocytes (Le Bigot et al., 1987). However, the results obtained from hepatocytes should also be interpreted with caution when quantitative comparison is the purpose, since many enzyme activities decline spontaneously during hepatocyte isolation or culture.

One final consideration in metabolite profiling is the choice of drug concentrations for in vitro studies. The major metabolic pathway may be shifted, depending on the drug concentration used. The clinical studies indicated that N-demethylation is the major metabolic pathway of diazepam in humans. However, in vitro studies in human liver microsomes showed that 3-hydroxylation was the major metabolic pathway of diazepam metabolism when a high (100 µM) drug concentration was incubated (Inaba et al., 1988). This in vitro and in vivo discrepancy is likely a result of differences in the substrate concentration used. Indeed, when an in vivo relevant substrate concentration (2-4 µM) is used (Yasumori et al., 1993), the major metabolic pathway of diazepam is N-demethylation in human liver microsomes.

It is clear that each in vitro system has its advantages and disadvantages. As long as their limitations are recognized and appropriate cautions and considerations are taken in the design of the studies, in vitro systems can aid in the selection of the animal species for toxicity studies as well as provide preliminary profiles of human metabolism.

Drug-Drug Interactions. Whenever two or more drugs are administered over similar or overlapping time periods, the possibility for drug interactions exists. Because of the potential of adverse effects, drug interactions have always been an important aspect to consider during the development of new drugs. In the past, such drug-interaction studies were primarily conducted at a relatively late stage during Phase II and III clinical studies. With the availability of human tissues and recombinant human enzymes, in vitro systems have been used in recent years as screening tools to predict in vivo drug interactions at a much earlier stage before the drug is selected for the development. Because oxidative metabolism represents a major route of elimination for many drugs, inhibition of cytochrome P450s is one of the main reasons for drug interactions.

Drug metabolism is a complex process, very often involving several pathways and various enzyme systems. In some cases, all of the metabolic reactions of a drug are catalyzed by a single enzyme, while in other cases a single metabolic reaction may involve multiple isoforms or different enzyme systems. The metabolism of indinavir exemplifies the first scenario, in which a single isoform of P450, CYP3A4, catalyzes four oxidative metabolic reactionsN-oxidation, N-dealkylation, indan hydroxylation, and phenyl hydroxylationto produce six metabolites in human liver microsomes (Chiba et al., 1996). On the other hand, the S-oxidation of 10-(N,N-dimethylaminoalkyl) phenothiazines in human liver microsomes is catalyzed by numerous P450 isoforms, including CYP2A6, CYP2C8, and CYP2D6 (Cashman et al., 1993). Similarly, amitriptyline is N-demethylated to nortriptyline in humans by numerous P450 isoforms, including CYP1A2, CYP2C9, CYP2C19, CYP2D6, and CYP3A4 (Venkatakrishnan et al., 1998). Therefore, definitive identification of the P450 isoforms responsible for drug metabolism is essential in predicting potential for drug interactions.

Over the last 10 years, a great deal of information on human cytochrome P450s at the molecular level has become available. This information, along with available antibodies and chemical inhibitors, has made it possible to determine easily the P450 isoforms responsible for the metabolism of a drug. In addition to the identification of the P450 isoforms, it is also important to evaluate the relative contributions of the metabolic pathways being inhibited to the overall elimination of the drug. With the advent of commercial liquid chromatography/mass spectrometry instrumentation and the development of high-field nuclear magnetic resonance as well as liquid chromatography/nuclear magnetic resonance techniques, the relative contribution of the metabolic pathways can be readily obtained. It should be noted that a significant interaction occurs only when drugs compete for the same enzyme system and when the metabolic reaction is a major elimination pathway. Rowland and Matin (1973) developed a kinetic model to evaluate the relative contribution of the metabolic fraction on the degree of drug interaction. They concluded that a significant drug interaction occurs only when the metabolic fraction of a particular pathway being inhibited is greater than 50% of the total clearance.

Although the identification of P450 isoforms is relatively straightforward, the experimental design and interpretation of in vitro interaction studies can be complicated and tricky. One of the important criteria in in vitro drug interaction studies is the use of clinically relevant concentrations of inhibitor and substrate. The use of supratherapeutic drug concentrations may produce a drug interaction in vitro but not in vivo. In addition, the major metabolic pathway may be shifted, depending on the drug concentration used. As mentioned earlier, in vitro studies in human liver microsomes showed that 3-hydroxylation was the major pathway of diazepam when a high drug concentration (100 µM) was utilized (Inaba et al., 1988), while the major metabolic pathway of diazepam was N-demethylation in human liver microsomes when a clinically relevant drug concentration (2 µM) was used (Yasumori et al., 1993). It should be noted that the N-demethylation of diazepam is mainly catalyzed by CYP2C19, and 3-hydroxylation is mediated by CYP3A4. This example illustrates the importance of the use of drug concentration in in vitro drug interaction studies in order to accurately define the involved P450 isoforms and predict the in vivo situation.

Another important criterion in in vitro drug interaction studies is the concentration of microsomal protein used. The K_i values of an inhibitor may be overestimated in the presence of a high microsomal protein concentration because of the depletion of the inhibitor by nonspecific binding to the microsomal proteins and microsomal metabolism. The K_i values for ketoconazole in human liver microsomes were estimated to be approximately 8 µM when a high microsomal protein concentration (1.5 mg/ml) was used (Lampen et al., 1995), while the estimated K_i values were approximately 0.03 µM when a low microsomal protein concentration (0.25 mg/ml) was used (Bourrie et al., 1996; Von Moltke et al., 1996). A sixfold increase in the microsomal protein concentration resulted in a 270-fold increase in the estimated K_i values.

The choice of in vitro enzyme systems, such as liver microsomes, cDNA-based vector systems, and liver slices, is also an important factor in in vitro drug-interaction studies. For instance, the apparent K_M characterizing the hydroxylation of ritonavir, a potent HIV protease inhibitor, in -lymphoblastoid-derived microsomes was similar to that obtained with human liver microsomes, whereas the apparent K_M characterizing the N-dealkylation and decarbamoylation of ritonavir was 30- to 300-fold lower in -lymphoblastoid-derived microsomes than those obtained with human liver microsomes (Kumar et al., 1996). The reason for this discrepancy is unknown; however, it is clear that we should be cautious in interpreting the kinetic parameters obtained from different in vitro systems.

Furthermore, an understanding of the mechanism involved in enzyme inhibition is critical in providing a rational basis for designing experimental conditions and interpreting drug-interaction data. For example, a compound that irreversibly inactivates an enzyme will result in a decrease in the V_max but has no effect on the K_M. The pattern of the kinetic data is similar to that of a reversible noncompetitive inhibitor, which also causes a decrease in the V_max but not the K_M. Thus an irreversible inhibitor can be incorrectly referred to as a reversible noncompetitive inhibitor. The experimental results reported by Franklin (1977) are a good example. Depending on the experimental conditions, SKF-525A acts as a competitive inhibitor or metabolic intermediate (MI) complexation inducing agent. As shown in table 2, SKF-525A increased the K_M values of substrates but had little effect on the V_max values when incubated with substrates without preincubation of the inhibitor. In contrast, SKF-525A decreased the V_max values of substrates and had little effect on the K_M values when SKF-525A was preincubated prior to substrate addition. Thus preincubation of SKF-525A changed the kinetics of inhibition from the reversible competitive type to irreversible MI complexation.

View this table: [in this window] [in a new window]   TABLE 2 Inhibition of rat hepatic microsomal mono-oxygenase activity by SKF-525A with or without preincubation prior to substrate additiona

Although it is relatively easy to assess in vitro drug interaction, it must be emphasized that the correct prediction and extrapolation of in vitro interaction data to the in vivo situation requires a good understanding of pharmacokinetic principles. The following discussion will provide a few of the basic tenets of enzyme inhibition on the pharmacokinetics. If a drug is mainly metabolized by the liver, the total clearance is approximately equal to the hepatic clearance (CL_H) that can be expressed as eq. 5 (Wilkinson, 1987):

(5)

where Q_h is the hepatic blood flow, E is the hepatic extraction ratio, f_b is the unbound fraction of drug in blood, and CL_int, the intrinsic clearance, is a measure of the drug metabolizing activity (V_max/K_M) in the liver. Depending on the underlying mechanism of the inhibitor, the V_max value of a drug can be decreased or the K_M value can be increased. Regardless of the mechanism, enzyme inhibition always results in a decrease in the intrinsic clearance (V_max/K_M). Therefore, the concept of intrinsic clearance is the cornerstone for the extrapolation of in vitro data to the in vivo situation.

Kinetically, drugs can be classified by whether their hepatic clearance is "enzyme-limited" (low) or "flow-limited" (high). When the intrinsic clearance of a drug is very small relative to the hepatic blood flow (Q_h f_b · CL_int), the hepatic clearance is low and the CL_H is directly related to f_b and CL_int, as shown in eq. 6:

(6)

Thus a decrease in the intrinsic clearance (CL_int) caused by inhibition will result in an almost proportional change in the clearance of "low clearance" drugs. On the other hand, if the intrinsic clearance is so high that f_b · CL_int Q_h, then the hepatic clearance is limited by the hepatic blood flow, as shown in eq. 7:

(7)

Thus a decrease in the intrinsic clearance caused by inhibition has little effect on the hepatic clearance of "high clearance" drugs.

Because the hepatic first-pass effect reflects the hepatic CL_int, hepatic bioavailability (F) can be expressed as:

(8)

and the area under the curve (AUC) after oral dosing can be described as:

(9)

As shown in eq. 9, a decrease in the CL_int caused by enzyme inhibition will yield an almost proportional increase in the AUC after oral dosing, regardless of whether it is a low- or high-clearance drug. In contrast, after iv administration, a significant decrease in the CL_int only affects the clearance and AUC of low-clearance drugs because the CL_H is independent of the CL_int for high-clearance drugs, as indicated in eq. 7.

The indinavir-ketoconazole interaction is a good example of an in vivo drug-drug interaction that is route- and drug-dependent (low- or high-clearance drug). Indinavir is a high-clearance drug that has a blood clearance of 80-90 ml/min/kg in rats and 15-17 ml/min/kg in AIDS patients (Lin, 1997). These values are greater than rat hepatic blood flow (60-70 ml/min/kg) or close to human hepatic blood flow (20 ml/min/kg). Indinavir is eliminated exclusively by CYP3A-mediated biotransformation in both rats and humans (Chiba et al., 1996; Lin et al., 1996a). In vitro studies with rat and human liver microsomes indicated that ketoconazole competitively inhibited the metabolism of indinavir, with a K_i value of approximately 0.25 µM for both rat and human liver microsomes (Lin, 1996). Coadministration of ketoconazole (25 mg/kg po) had little inhibitory effect on the clearance of indinavir and its AUC after the iv administration of indinavir (10 mg/kg iv) in rats. The clearance rate decreased from 87 ml/min/kg in control rats to 83 ml/min/kg in ketoconazole-coadministered rats. However, ketoconazole significantly increased the bioavailability of indinavir and its AUC after oral dosing. The bioavailability increased from approximately 20% in control rats to 89% in ketoconazole-coadministered rats (Lin, 1996). Ketoconazole, on the other hand, is a low-clearance drug with a clearance rate of 8-10 ml/min/kg in rats. In vitro studies with rat liver microsomes revealed that indinavir also competitively inhibited the metabolism of ketoconazole, with a K_i value of 4.5 µM. As expected, coadministration of indinavir (20 mg/kg po) in rats significantly increased the AUCs of ketoconazole twofold after both iv and oral administration of ketoconazole (Lin, 1996). The clearance of ketoconazole in rats decreased from 8.5 ml/min/kg when given alone to 4.5 ml/min/kg when coadministered with indinavir.

The ultimate goal of an in vitro drug interaction study is to predict the quantitative effect of drug inhibition in vivo. For competitive inhibition, the per cent of inhibition can be described as in eq. 10:

(10)

where CL_int,o and CL_int,i are the intrinsic clearances in the absence and presence of inhibitor, respectively. K_M is the Michaelis constant of the substrate, K_i is the inhibitory constant of the inhibitor, and [S] and [I] are the substrate and inhibitor concentrations, respectively.

As shown in eq. 10, the degree of inhibition depends not only on the K_M and K_i values of substrate and inhibitor but also on their concentrations ([S] and [I]). Both [S] and [I] continue to change as a function of time in vivo after drug administration unless they are maintained under steady-state conditions. Thus appropriate pharmacokinetic models are needed in order to obtain an accurate in vitro/in vivo extrapolation. Lin et al. (1984) successfully applied a physiologically based pharmacokinetic model to predict product inhibition. This model incorporated the K_M and K_i values together with the kinetic parameters of the plasma profiles of the parent drug and its metabolite to predict the quantitative effect of product inhibition of salicylamide on the elimination of ethoxybenzamide in rabbits after a single dose. Although the physiologically based pharmacokinetic approach can provide an accurate prediction of drug interaction, this approach is very costly and time-consuming when all of the parameters needed are being obtained. A closer examination of the literature reveals that in most cases, in vitro interaction studies are generally carried out to assess the potential of drug interaction, more or less in a qualitative sense, by comparing the relative affinities of the substrate (K_M) and inhibitor (K_i) with their concentration ranges in clinical studies. One of the most common approaches is the use of in vitro K_i values together with in vivo values of the peak plasma concentration of inhibitor to forecast the possibility of drug-drug interactions in vivo.

Even for the qualitative prediction, the in vitro/in vivo extrapolation of drug-drug interaction appears to be difficult and controversial. One of the controversies is whether the total (bound + unbound) or unbound plasma concentration of the inhibitor should be used to predict the in vivo drug interaction. A basic tenet of pharmacokinetics is that only unbound drug can diffuse across hepatocytes, and that unbound drug concentration in the blood is in equilibrium with that in the hepatocytes. Thus it is generally believed that only unbound inhibitor can compete with the substrate for the enzymes. However, there are reports that contradict this basic tenet. For example, instead of unbound inhibitor concentration, total plasma concentration of ketoconazole gave a good in vitro/in vivo extrapolation of the terfenadine-ketoconazole interaction (Von Moltke et al., 1994). Similarly, Tran et al. (1997) reported that the in vivo K_i values of stiripentol on the metabolism of carbamazepine were more consistent with the in vitro K_i values when total plasma concentrations of stiripentol, but not unbound concentrations, were used to estimate the in vivo K_i values. These authors speculated that the stiripentol concentration at the enzyme site was much higher than its unbound concentration in the blood because of a high liver/plasma partition. A similar observation has been reported for selective serotonin reuptake inhibitors (Von Moltke et al., 1998). A good in vitro/in vivo extrapolation of drug-drug interaction by selective serotonin inhibitors was obtained only when a liver/plasma partition ratio was taken into account. The issue of intrahepatic exposure of enzyme to inhibitor or substrate and its relationship with plasma concentration requires further investigation.

Sometimes, the failure of in vitro/in vivo extrapolation may originate from the nature and design of the in vitro experiments. Cimetidine, an H₂-receptor antagonist, has been well-documented to inhibit cytochrome P450-mediated drug metabolism in humans in vivo. However, the concentration of cimetidine required for in vitro inhibition of a cytochrome P450-mediated reaction is typically 100 to 1000 times greater than the plasma concentration of cimetidine associated with the inhibition of drug metabolism in patients (Knodell et al., 1991). Clearly, the in vitro data will falsely predict the potential in vivo drug interaction. Although the reason for the in vitro and in vivo discrepancy is not fully understood, studies by Chang et al. (1991a, 1991b) have suggested that cimetidine may be a mechanism-based inhibitor. This may explain the in vitro/in vivo discrepancy. In vitro studies with rat liver microsomes revealed that cimetidine inhibited the activities of CYP2C11, CYP2B1/2, and CYP3A1/2, with 50% inhibitory concentration (IC₅₀) values in the range of 1.0 to 7.4 mM (Knodell et al., 1991). Preincubation of rat liver microsomes with a low concentration (0.05 mM) of cimetidine in the presence of NADPH resulted in a substantial decrease in the enzyme activities, suggesting that a mechanism-based inactivation is involved (Chang et al., 1991a). It is possible that cimetidine acts as an irreversible inhibitor in vivo but as a reversible inhibitor in vitro. Therefore, an understanding of the underlying mechanism involved in drug interaction is very important in order to provide a rational basis for the design of experimental conditions.

Prediction of In Vivo Metabolic Clearance. One of the main objectives of in vitro metabolism studies is the quantitative prediction of in vivo metabolic clearance from the in vitro data. The prediction of metabolic clearance from in vitro systems, however, is difficult and highly controversial. Some scientists believe that in vitro/in vivo extrapolation is possible, whereas others are less optimistic and believe that it is extremely difficult to predict in vivo clearance from in vitro metabolism data. Each group can cite examples from the literature to support its views (Sugiyama et al., 1989; Pang and Chiba, 1994; Houston, 1994; Gillette, 1984). Knowing that in vitro extrapolation is an approximation, we believe that quantitative in vitro metabolic data can be extrapolated reasonably well to the in vivo situation with a good understanding of the interdependent factors that are involved and the application of appropriate pharmacokinetic principles.

There are many examples of good quantitative in vitro and in vivo correlation. Ethoxybenzamide, an antipyretic agent, is exclusively metabolized to salicylamide by rat liver microsomes. The in vitro V_max and K_M values (3.46 µmol/min/kg and 0.378 mM) obtained from rat liver microsomes are in good agreement with those obtained in vivo by application of a two-compartment model (3.77 µmol/min/kg and 0.192 mM) (Lin et al., 1978). Indinavir, a potent HIV protease inhibitor, exhibited marked species differences in hepatic clearance. This drug was metabolized mainly by isoforms of the CYP3A subfamily to form oxidative metabolites in all species examined (Lin et al., 1996a). The in vitro hepatic clearance obtained from incubations of rat, dog, and monkey liver microsomal preparations was in good agreement with the corresponding in vivo hepatic clearance of indinavir. The in vitro hepatic clearances were 31, 25, and 7.8 ml/min/kg for rats, monkeys, and dogs, respectively, while the corresponding in vivo hepatic clearances were 43, 36, and 11 ml/min/kg (Lin et al., 1996a). Chiba et al. (1990) successfully predicted the steady-state concentration of imipramine and its active metabolite, desipramine, in rats by using the V_max and K_M values obtained from in vitro microsomal studies. Felodipine, a calcium channel blocker, is primarily metabolized to its pyridine analog in rats, dogs, and humans. The hepatic clearances of this drug obtained from in vitro studies with hepatic microsomes were 16 liters/hr for rats, 39 liters/hr for dogs, and 259 liters/hr for humans and agreed reasonably well with those clearances observed in vivo (6.2 liters/hr, 88 liters/hr, and 321 liters/hr; Baarnhielm et al., 1986). Similarly, a good in vitro and in vivo correlation of the clearance of cytarabine hydrochloride was reported by Dedrick et al. (1972). Furthermore, Iwatsubo et al. (1997) successfully predicted the in vivo clearance and bioavailability of YM796, a central nervous system drug for the treatment of Alzheimer's disease, by using a recombinant system of human CYP3A4 together with knowledge of the content of this isoform in human liver microsomes. Recently, Houston and Carlile (1997) showed an excellent correlation between in vivo and hepatocyte intrinsic clearance for 21 drugs in rats. These examples clearly show that the in vivo metabolic clearance can be approximated from in vitro metabolic data if appropriate pharmacokinetic principles are utilized.

As mentioned earlier, although allometric scaling has successfully been used to predict the renal clearance of drugs in humans from animal data, the scaling usually failed to predict the metabolic clearance of drugs (Boxenbaum, 1980). To improve the prediction of metabolic clearance in humans, empirical correction factors, such as brain weight or maximum life span, have been proposed and used (Boxenbaum, 1984). Recently, Lave et al. (1997) proposed a new allometric approach that integrated in vitro metabolic data to improve the predictability of hepatic clearance of drugs in humans. For ten extensively metabolized compounds, correction of the in vivo clearance in the different animal species for in vitro metabolic clearance significantly improved the predictions in humans, compared with the conventional approaches in which clearance is extrapolated directly using body weight or correcting for brain weight.

A literature survey revealed that in some cases, in vitro metabolic data failed to predict in vivo clearance. Sources of inaccuracy in predicting the in vivo metabolic clearance may include the nature and design of in vitro experiments, presence of extrahepatic metabolism, and active transport in the liver. Unfortunately, the reason for the lack of in vitro/in vivo correlation rarely has been understood or explained.

Recently, we carried out a study in our laboratory to examine the inductive effect of dexamethasone on the intestinal and hepatic first-pass metabolism of indinavir in rats. Pretreatment with dexamethasone (40 mg/kg/day po for 3 days) resulted in three- and tenfold increases in the V_max values in the intestinal and hepatic microsomes, while no effect was observed on the K_M values in either microsomal preparation. By using the intestinal and hepatic intrinsic clearance obtained from in vitro V_max/K_M values and the mucosal and hepatic blood flow values, the in vitro intestinal and hepatic first-pass metabolism extraction ratios of indinavir were estimated to be 0.009 and 0.45 in control rats and 0.036 and 0.88 in dexamethasone-treated rats, respectively. For comparison, the in vivo intestinal and hepatic first-pass metabolism (extraction ratio) were measured in control and dexamethasone-treated rats. The intestinal first-pass metabolism was determined using the in situ intestinal loop technique, while the hepatic first-pass metabolism was estimated by comparing the indinavir concentrations in the systemic circulation during portal vein or femoral vein infusion of the drug. The detailed experimental procedures for the intestinal and hepatic first-pass metabolism studies have been described elsewhere (Lin et al., 1996a). As shown in table 3, there is a reasonably good correlation between in vitro and in vivo hepatic first-pass metabolism extraction ratios both before and after dexamethasone induction, whereas a significant discrepancy between in vitro and in vivo intestinal first-pass metabolism extraction ratios was observed. The predicted in vitro intestinal first-pass metabolism is much lower than that determined in vivo by approximately sixfold in control rats and tenfold in dexamethasone-treated rats (table 3). The reason for the observed discrepancy between in vitro and in vivo intestinal first-pass metabolism is not clear at the present time.

View this table: [in this window] [in a new window]   TABLE 3 Prediction of intestinal and hepatic first-pass metabolism rate of indinavir in control and dexamethasone-treated rats (mean ± SD)

One possible explanation for the discrepancy is the involvement of P-glycoprotein in the intestinal first-pass metabolism of indinavir. P-Glycoprotein, located in the apical brush border membrane of enterocytes of the small intestine, can act as an efflux transporter that extrudes a drug from inside the enterocytes into the intestinal lumen as the drug is being absorbed across the epithelial cells. A portion of the extruded drugs can be reabsorbed into the enterocytes. Consequently, P-glycoprotein increases the exposure of drugs to drug-metabolizing enzymes and hence enhances the intestinal metabolism of drugs by prolonging the intracellular residence time through the repetitive processes of extrusion and reabsorption. The effect of P-glycoprotein on intracellular residence time and intestinal metabolism in Caco-2 cells was reported by Gan et al. (1996), who used cyclosporin A as a model compound. Cyclosporin A was metabolized to a greater extent when the drug was dosed to the apical side than to the basolateral side. Recently, indinavir has been shown to be a substrate of P-glycoprotein (Kim et al., 1998). Thus it is possible that the effect of P-glycoprotein may be to increase the intestinal metabolism of indinavir by prolonging the intracellular residence time of the drug. If P-glycoprotein plays a role in intestinal metabolism, a more complex model must be developed in order to predict in vivo intestinal first-pass metabolism.

	Conclusion

Top Abstract Introduction Interspecies pharmacokinetic... In vitro extrapolation Conclusion References

The use of animal models to predict pharmacological and toxic effects in humans has a long history, and it comes from the belief in the "unity life" of mammals. For years, scientists believed that the central physiological functions of circulation, respiration, and autonomic regulation were common in all mammals and that the major routes of endogenous metabolism, such as the citric acid cycle and oxidative phosphorylation, were similar in mammals. However, in the past 30 years, pharmacokineticists have failed to find an animal species in which the ADME processes of drugs are consistently the same as those in humans. In fact, it can be presumed that such an animal species will never be found.

With the recent breakthroughs in molecular biology, it is possible that in the very near future we can use transgenic (humanized) animals for studying human ADME. A number of genetically modified animals have been established as models for human genetic diseases, but the transgenic approach for evaluating the metabolism and transport of drugs has not been utilized yet to any large degree (Liggitt et al., 1992; Burki and Lederman, 1995). Using standard techniques, it may not be difficult to develop transgenic animals that express genes coding for both human CYP3A4 and P-glycoprotein in the liver and intestine. This type of transgenic animal would certainly provide valuable means for evaluating the intestinal and hepatic first-pass metabolism of drugs. Although it is ambitious, the dream of using transgenic animals in evaluating human ADME may become true within 10 years!

The concept that the properties of the whole are the sum of the properties of the parts has had a profound impact on the manner of conducting science. In their recent review article entitled "Complexity and Emergence in Drug Research," Kier and Testa (1995) rightly pointed out that this concept has guided scientists in all different scientific fields to the belief that the route of understanding nature is through the dissection of a system into its parts, followed by the study of these parts. Once a system has been dissected into its parts, scientists attempt to reassemble the information about the parts to understand the interdependence of all controlling factors of the whole system. Drug metabolism scientists also apply this fashionable concept by using a broad spectrum of methods, from subcellular (microsomes) to cellular (hepatocytes) to organ levels (isolated perfused liver) to study drug metabolism. However, extrapolation of in vitro metabolic data to the in vivo situation is not always straightforward. As seen with the example of the intestinal first-pass metabolism of indinavir, the failure of in vitro extrapolation may be a result of the involvement of P-glycoprotein. Thus it is important to understand fully the interdependent factors that influence intestinal metabolism before a complex predictive model in which all controlling factors are incorporated can be developed.

Although there are many limitations in the applications of interspecies scaling and in vitro extrapolation of ADME data to humans, we believe, through our experience, that a good prediction of certain ADME processes in humans can be made when in vitro data are integrated with in vivo animal data. The experience with indinavir is a good example. Metabolism and pharmacokinetic studies with indinavir in animals, combined with data from human tissue preparations in vitro, allowed accurate predictions of the oral bioavailability in human subjects, while recognition of the CYP3A4-inhibitory properties of indinavir provided insight into the potential drug-drug interactions with indinavir and other drugs.

	Footnotes

² Lin JH, Chen I-W and deLuna FA, unpublished data, 1996.

Send reprint requests to: Dr. Jiunn H. Lin, Drug Metabolism, Merck Research Laboratories, WP 42-2, West Point, PA 19486. e-mail: jiunn_lin{at}merck.com

	Abbreviations

Abbreviations used are: ADME, absorption, distribution, metabolism and excretion; F, bioavailability; f_abs, the fraction of dose absorbed from the gastrointestinal lumen; f_g, the fraction of drug metabolized by the gut wall; f_h, the fraction of drug metabolized by the liver; V_d, volume of distribution of total (bound + unbound) drug; V_f, volume of distribution of unbound drug; V_max, the maximum velocity of metabolite formation; K_M, Michaelis constant; K_i, inhibitory constant; AUC, area under plasma concentration curve; P450, cytochrome P450; UGT, uridine diphosphate glycosyltransferase; AZT, zidovudine; HIV, human immunodeficiency virus; GFR, glomerular filtration rate; CL_int, intrinsic clearance; CL_H, hepatic clearance.

	References

Top Abstract Introduction Interspecies pharmacokinetic... In vitro extrapolation Conclusion References

• Baarnheilm C, Dahlback H and Skanberg I (1986) In vivo pharmacokinetics of felodipine predicted from in vitro studies in rat, dog and man. Acta Pharmacol Toxicol  59: 113-122[Medline].
• Balani SK, Arison BH, Mathai L, Kauffman LR, Miller RR, Stearns RA, Chen I-W and Lin JH (1995) Metabolites of L-735,524, a potent HIV-1 protease inhibitor, in human urine. Drug Metab Dispos  23: 266-270[Abstract].
• Bensoussan C, Delaforge M and Mansuy D (1995) Particular ability of cytochromes P450 3A to form inhibitory P450-iron-metabolite complexes upon metabolic oxidation of amino drugs. Biochem Pharmacol  49: 591-602[Medline].
• Blum MR, Liao SHT, Good SS and deMiranda P (1988) Pharmacokinetics and bioavailability of zidovudine in man. Am J Med  85: 189-196[Medline].
• Bourrie M, Meunier V, Berger Y and Fabre G (1996) Cytochrome P450 isoform inhibitors as a tool for the investigation of metabolic reactions catalyzed by human liver microsomes. J Pharmacol Exp Ther  277: 321-332[Abstract/Free Full Text].
• Boxenbaum H (1980) Interspecies variation in liver weight, hepatic blood flow, and antipyrine intrinsic clearance in extrapolation of benzodiazepines and phenytoin. J Pharmacokinet Biopharm  8: 165-176[Medline].
• Boxenbaum H (1982) Comparative pharmacokinetics of benzodiazepines in dog and man. J Pharmacokinet Biopharm  10: 411-426[Medline].
• Boxenbaum H (1984) Interspecies pharmacokinetic scaling and the evolutionary-comparative paradigm. Drug Metab Rev  15: 1071-1121[Medline].
• Burki K and Lederman B (1995) Transgenic animals as pharmacological tools, in Advances in Drug Research (Testa B and Mayer UA eds) pp 143-177, Academic Press, New York.
• Callan WM and Sunderman FW, Jr (1973) Species variations in binding of ⁶³N (II) by serum albumin. Res Commun Chem Pathol Pharmacol  5: 459-472[Medline].
• Cashman JR, Young L and Wrighton SA (1993) Stereo- and regio-selective N- and S-oxidation of tertiary amines and sulfides in the presence of adult human liver microsomes. Drug Metab Dispos  21: 492-501[Abstract].
• Chang T, Levine M, Bandiera SM and Bellward GD (1991a) Selective inhibition of rat hepatic microsomal cytochrome P450: I. Effect of the in vivo administration of cimetidine. J Pharmacol Exp Ther  260: 1441-1449[Abstract/Free Full Text].
• Chang T, Levine M and Bellward GD (1991b) Selective inhibition of rat hepatic microsomal cytochrome P450: II. Effect of the in vitro administration of cimetidine. J Pharmacol Exp Ther  260: 1450-1455[Abstract/Free Full Text].
• Chapell WR and Mordenti J (1991) Extrapolation of toxicological and pharmacological data from animals to humans, in Advances in Drug Research (Testa B ed) vol 20, pp 1-16, Academic Press, London, UK.
• Chiba M, Fujita S and Suzuki T (1990) Pharmacokinetics correlation between in vitro hepatic microsomal enzymes kinetics and in vivo metabolism of imipramine and desipramine in rats. J Pharm Sci  79: 281-287[Medline].
• Chiba M, Hensleigh M and Lin JH (1997) Hepatic and intestinal metabolism of indinavir, a potent HIV protease inhibitor, in rat and human microsomes. Biochem Pharmacol  53: 1183-1185.
• Chiba M, Hensleigh M, Nishime JA, Balani SK and Lin JH (1996) Role of cytochrome P450 3A4 in human metabolism of MK-639, a potent human immunodeficiency virus protease inhibitor. Drug Metab Dispos  24: 307-314[Abstract].
• Clark B and Smith DA (1984) Pharmacokinetics and toxicity testing. Crit Rev Toxicol  12: 343-385[Medline].
• Clarke DJ and Burchell B (1994) The uridine diphosphate glucuronosyltransferase multigene family: Function and regulation, in Conjugation-Deconjugation Reactions in Drug Metabolism and Toxicity (Kauffman FC ed) pp 3-43, Springer-Verlag, Berlin, Germany.
• Dedrick RL, Forrester DD and Ho DHW (1972) In vitro-in vivo correlation of drug metabolism: Deamination of 1--D-arabino-furanosylcytosine. Biochem Pharmacol  21: 1-16[Medline].
• De Waziers I, Cugenec PH, Yang CS, Lerous JP and Beaune PH (1990) Cytochrome P450 isoenzymes, epoxide hydrolase and glutathione transferase in rat and human hepatic and extrahepatic tissues. J Pharmacol Exp Ther  253: 387-394[Abstract/Free Full Text].
• Diaz D, Fabre I, Daujat L, Saint Aubert B, Bories P, Michel H and Maurel P (1990) Omeprazole is an aryl hydrocarbon-like inducer of human hepatic cytochrome P450. Gastroenterology  99: 737-747[Medline].
• Dogterom P (1993) Development of a simple incubation system for metabolism studies with precision-cut liver slices. Drug Metab Dispos  21: 699-704[Abstract].
• Edwards NA (1975) Scaling of renal function in mammals. Comp Biochem Physiol  52 (A): 63-66.
• Eichelbaum M, Ochs HR and Robertz GM (1982) Pharmacokinetics and metabolism of antipyrine (phenazone) after intravenous and oral administration. Arzneimittel Forschung  32: 575-578[Medline].
• Engel G, Hofmann U and Heidemann H (1996) Antipyrine as a probe for human oxidative drug metabolism: Part I. Identification of the cytochrome P450 enzymes catalyzing 4-hydroxy, 3-hydroxymethyl- and norantipyrine formation. Clin Pharmacol Ther  59: 613-623[Medline].
• Fichtl B, Nieciecki AV and Walter K (1991) Tissue binding versus plasma binding of drugs: General principles and pharmacokinetics consequence, in Advances in Drug Research (Testa B ed) vol 20, pp 117-166, Academic Press, New York.
• Franklin MR (1977) Inhibition of mixed-function oxidations by substrates forming reduced cytochrome P450 metabolic-intermediate-complexes. Pharmacol Ther  2: 227-245.
• Gan S, Moseley MA, Khosla B, Augustijins PF, Bradshaw TP, Hendren RW and Thakker DR (1996) CYP3A-like cytochrome P450-mediated metabolism and polarized efflux of cyclosporin A in Caco-2 cells. Drug Metab Dispos  24: 344-349[Abstract].
• Gillette JR (1984) Problems in correlating in vitro and in vivo studies of drug metabolism, in Pharmacokinetics: A Modern View (Benet LZ, Levy G and Ferraiolo BL eds) pp 235-252, Plenum Press, New York.
• Good SS, Durack DT and DeMiranda P (1986) Biotransformation in various species and humans of 3'-azido-3'-deoxythymidine, a potential agent for the treatment of AIDS. Fed Proc  45: 44-52.
• Guengerich FP: Human cytochrome P450 enzymes, in Cytochrome P450: Structure, Mechanism and Biochemistry (Ortiz de Montellano PR ed.), pp 473-535, Plenum Press, New York.
• Houston JB (1994) Utility of in vitro drug metabolism data in predicting in vivo metabolic clearance. Biochem Pharmacol  47: 1469-1479[Medline].
• Houston JB and Carlile DJ (1997) Prediction of hepatic clearance from microsomes, hepatocytes and liver slices. Drug Metab Rev  29: 891-922[Medline].
• Inaba T, Tait A, Nakano M, Mahon WA and Kalow W (1988) Metabolism of diazepam in vitro by human liver independent variability of N-demethylation and C3-hydroxylation. Drug Metab Dispos  16: 605-608[Abstract].
• Iwatsubo T, Hirota N, Ooie T, Suzuki K, Shimada N, Chiba K, Ishizak T, Green CE, Tyson CA and Sugiyama Y (1997) Prediction of in vivo drug metabolism in the human liver from in vitro metabolism data. Pharmacol Ther  73: 147-171[Medline].
• Kier LB and Testa B (1995) Complexity and emergence in drug research, in Advances in Drug Research (Testa B and Meyer UA eds) pp 1-43, Academic Press, New York.
• Kim RB, Fromm MF, Wandel C, Leeake B, Wood AJJ, Roder DM and Wilkinson GR (1998) The drug transporter P-glycoprotein limits oral absorption and brain entry of HIV-1 protease inhibitors. J Clin Invest  101: 289-294[Medline].
• Knodell RG, Browne DG, Gwozdz GP, Brain WR and Guengerich FP (1991) Differential inhibition of individual human liver cytochromes P450 by cimetidine. Gastroenterology  101: 1680-1691[Medline].
• Kobayashi S, Murray S, Watson D, Sesardic D, Davies DS and Boobis AB (1989) The specificity of inhibition of debrisoquine 4-hydroxylase activity in quinidine and quinine in rat is the reverse of that in man. Biochem Pharmacol  38: 2795-2799[Medline].
• Kragh-Hansen U (1981) Molecular aspects of ligand binding to serum albumin. Pharmacol Rev  33: 17-53[Medline].
• Kumar GN, Rodrigues AD, Buko AM and Denissen JF (1996) Cytochrome P450-mediated metabolism of HIV-1 protease inhibitor ritonavir (ABT-538) in human liver microsomes. J Pharmacol Exp Ther  277: 423-431[Abstract/Free Full Text].
• Lampen A, Christians U, Guengerich FP, Watkin PB, Kolars JC, Bader A, Gonschior A-K, Dralle H, Hackbarth I and Sewing K-F (1995) Metabolism of immunosuppressant tacrolimus in the small intestine: Cytochrome P450, drug interaction, and interindividual variability. Drug Metab Dispos  23: 1315-1324[Abstract].
• Lave T, Dupin S, Schmitt C, Chou RC, Jaeck D and Cuassolo P (1997) Integration of in vitro data into allometric scaling to predict hepatic clearance in man: Application to 10 extensively metabolized drugs. J Pharm Sci  86: 584-590[Medline], 1997.
• Le Bigot JF, Begue JM, Kiechel JR and Guillouzo A (1987) Species differences in metabolism of ketotofen in rat, rabbit and human: Demonstration of similar pathways in vivo and in cultured hepatocytes. Life Sci  40: 883-890[Medline].
• Liggitt HD, Omiecinski CJ and Quaife CJ (1992) Potential applications of transgenic animals in drug development, in Progress in Drug Metabolism (Gibson GG ed) pp 1-33, Taylor & Francis, London.
• Lin JH (1989) Species differences in protein binding of diflunisal. Drug Metab Dispos  17: 221-223[Medline].
• Lin JH (1995) Species similarities and differences in pharmacokinetics. Drug Metab Dispos  23: 1008-1021[Abstract].
• Lin JH (1996) In vitro and in vivo interaction with indinavir. Symposium on Recent Advances in Drug-Drug Interaction. Washington, DC, Proceedings, International Business Communications, December 10-11, 1996.
• Lin JH (1997) HIV protease inhibitors: From drug design to clinical studies. Adv Drug Delivery Rev  27: 215-233.
• Lin JH, Chiba M, Balani SK, Chen I-W, Kwei GY-S, Vastag KJ and Nishine JA (1996a) Species differences in the pharmacokinetics and metabolism of indinavir, a potent HIV protease inhibitor. Drug Metab Dispos  24: 1111-1120[Abstract].
• Lin JH, Hayashi M, Awazu S and Hanano M. (1978) Correlation between in vitro and in vivo drug metabolism rate: Oxidation of ethoxybenzamide in rat. J Pharmacokinet Biopharm  6: 327-337[Medline].
• Lin JH, Hooke KF, Yeh KC and Duggen DE (1985) Dose-dependent pharmacokinetics of diflunisal in rats: Dual effects of protein binding and metabolism. J Pharmacol Exp Ther  235: 402-406[Abstract/Free Full Text].
• Lin JH and Lu AYH (1997) Role of pharmacokinetics and metabolism in drug discovery and development. Pharmacol Rev  49: 403-449[Abstract/Free Full Text].
• Lin JH, Storey DE, Chen I-W and Xu X (1996b) Improved absorption of L-365,260, a poorly soluble drug. Biopharm Drug Dispos  17: 1-16[Medline].
• Lin JH, Sugiyama Y, Hanano M and Awazu S (1984) Effect of product inhibition on elimination kinetics of ethoxybenzamide in rabbits. Drug Metab Dispos  12: 253-256[Abstract].
• McDonnell WM, Scheiman JM and Traber PG (1992) Induction of cytochrome P450 1A gene (CYP1A) by omeprazole in the human alimentary tract. Gastroenterology  103: 1509-1516[Medline].
• Meffin PJ, Brooks DM, Bertouch J, Veenendaal JR and Harrington BJ (1983) Diflunisal disposition and hypouricemic response in osteoarthritis. Clin Pharmacol Ther  33: 813-821[Medline] .
• Nebert DW and Gonzalez FL (1990) The P450 gene superfamily, in Principles, Mechanisms and Biological Consequences of Induction (Ruckpaul K and Rein H eds) pp 35-61, Taylor & Francis, London, UK.
• Nelson DR, Koymans L, Kamataki T, Stegeman JJ, Feyereisen RRW, Waxman DJ, Waterman MR, Gotoh O, Coon MJ, Estabrook RW, Gunsalus IC and Nebert DW (1996) P450 superfamily, update on new sequences gene mapping, accession numbers and nomenclature. Pharmacogenetics  6: 1-42[Medline].
• Pang KS and Chiba M (1994) Metabolism scaling-up from in vitro to organ and whole body, in Handbook of Experimental Pharmacology: Pharmacokinetics of Drugs (Welling PG and Balant LP eds) vol 110, pp 101-187, Springer-Verlag, Berlin, Germany.
• Renkin EM and Gilmore JP (1973) Glomerular filtration, in Handbook of PhysiologyRenal Physiology (Orlogg J and Berliner RW eds), chapter 9, pp 185-248, Williams & Wilkins, Baltimore, Maryland.
• Resetar A and Spector T (1989) Glucuronidation of 3'-azido-3'-deoxythymidine: Human and rat enzyme specificity. Biochem Pharmacol  38: 1389-1393[Medline].
• Rice JM, Diwan BA, Ward JM, Nimes RW and Lubet RA (1992) Phenobarbital and related compounds: Approaches to interspecies extrapolation. Prog Clin Biol Res  374: 231-249[Medline].
• Rowland M and Matin SB (1973) Kinetics of drug-drug interactions. J Pharmacokinet Biopharm  1: 553-567.
• Sawada Y, Hanano M, Sugiyama Y and Iga T (1984a) Prediction of the volumes of distribution of basic drugs in humans based on data from animals. J Pharmacokinet Biopharm  12: 587-596[Medline].
• Sawada Y, Hanano M, Sugiyama Y and Iga T (1984b) Prediction of -lactam antibiotics in humans from pharmacokinetic parameters in animals. J Pharmacokinet Biopharm  12: 241-261[Medline].
• Sesardic D, Boobis AR, Murray P, Murray S, Segura J, De La Torre R and Davies DS (1990) Furafylline is a potent and selective inhibitor of cytochrome P450 1A2 in man. Br J Clin Pharmacol  29: 651-663[Medline].
• Smith RL (1971) Excretion of drugs in bile, in Handbook of Experimental Pharmacology (Brodie BB and Gillette JR eds) vol XXVIII: Concepts in Biochemical Pharmacology , pp 354-389, Springer-Verlag, Berlin, Germany.
• Smith RL (1973) The Excretory Function of Bile: The Elimination of Drugs and Toxic Substances in Bile Chapman and Hall, London, United Kingdom.
• Stahl WR (1967) Scaling of respiratory variables in mammals. J Appl Physiol  22: 453-460[Free Full Text].
• Stevens JC, Shipley LA, Cashman JR, Vandenbranden M and Wrighton SA (1993) Comparison of human and rhesus monkey in vitro phase I and phase II hepatic drug metabolism activities. Drug Metab Dispos  21: 753-760[Abstract].
• Strolin Benedetti M and Dostert P (1994) Induction and autoinduction properties of rifamycin derivatives: A review of animal and human studies. Environ Health Perspect  102 (Suppl 9): 101-105 .
• Sugiyama Y, Sawada Y, Iga T and Hanano M (1989) Reconstruction of in vivo metabolism from in vitro data, in Xenobiotic Metabolism and Disposition (Kato R, Estabrook RW and Cayen MN eds) pp 225-235, Taylor & Francis, London, United Kingdom.
• Tran A, Rey E, Pons G, Rousseau M, d'Athis P, Olive G, Mather GG, Bishop FE, Wurden CJ, Labroo R, Trager WF, Kunze KL, Thummel KE, Vincent JC, Gillardin J-M, Lepage F and Levy RH (1997) Influence of stiripentol on cytochrome P450 mediated metabolic pathways in humans: In vitro and in vivo comparison and calculation of in vivo inhibition constants. Clin Pharmacol Ther  62: 490-504[Medline].
• Vargas M and Franklin MR (1997) Intestinal UDP-glucuronosyl-transferase activities in rats and rabbits. Xenobiotica  37: 413-421.
• Venkatakrishnan K, Greenblatt DJ, Von Moltke LL, Schmider J, Harmatz JS and Shader RI (1998) Five distinct human cytochromes mediate amitriptyline N-demethylation in vitro: Dominance of CYP2C19 and 3A4. J Clin Pharmacol  38: 112-121[Abstract].
• Von Moltke LL, Greenblatt DJ, Duan SX, Harmatz JS, Wright CE and Shader R (1994) In vitro prediction of the terfenadine ketoconazole pharmacokinetic interaction. J Clin Pharmacol  34: 1222-1227[Abstract].
• Von Moltke LL, Greenblatt DJ, Duan SX, Harmatz JS, Wright CE and Shader RI (1996) Inhibition of terfenadine metabolism in vitro by azole antifungal agents and by selective serotonin reuptake inhibitor antidepressants: Relation to pharmacokinetic interactions in vivo. J Clin Psychopharmacol  16: 104-112[Medline].
• Von Moltke LL, Greenblatt DJ, Duan SX, Harmatz JS, Wright CE and Shader R (1998) In vitro approaches to predicting drug interactions in vivo. Biochem Pharmacol  55: 113-122[Medline].
• Wilkinson GR (1987) Clearance approaches in pharmacology. Pharmacol Rev  39: 1-47[Medline].
• Worboys PD, Bradbury A and Houston B (1996) Kinetics of drug metabolism in rat liver slices: II. Comparison of clearance by liver slices and freshly isolated hepatocytes. Drug Metab Dispos  24: 676-681[Abstract].
• Wrighton SA, Vandenbraden M, Stevens JC, Shipley LA, Ring BJ, Rettie AE and Cashman JR (1993) In vitro methods for assessing human hepatic drug metabolism: Their use in drug development. Drug Metab Rev  25: 453-484[Medline].
• Yasumori T, Nagata K, Yang SK, Chen L-S, Murayama N, Yamazoe Y and Kato R (1993) Cytochrome P450 mediated metabolism of diazepam in human and rat: Involvement of human CYP 2C in N-demethylation in the substrate concentration-dependent manner. Pharmacogenetics  3: 291-301[Medline].
• Yeh KE, Deutsch PJ, Haddix H, Hesney M, Hoagland V, Ju WD, Justice SJ, Osborne B, Sterrett AT, Stone J, Wodf E and Waldman S (1998) Single-dose pharmacokinetics of indinavir and the effect of food. Antimicrob Agents Chemother  42: 332-338[Abstract/Free Full Text].