The ability to predict human pharmacokinetics is an important yet elusive goal in drug discovery and preclinical development. Methods involving animal models, extrapolation of animal data (Chappell and Mordenti, 1991), or in vitro methodologies (Houston and Carlile, 1997; Obach et al., 1997) are utilized with varying degrees of success. Nonhuman primates, especially great apes, are closer phylogenetically to humans, which would suggest better fidelity in their disposition of and response to new chemical entities. Well controlled studies confirming this hypothesis, however, are sparse.

Chimpanzees (Pan troglodytes) are the closest living relatives to humans, being 98.77% genetically similar (Wildman, 2002). Their remarkable similarity to humans in terms of their genetics, anatomy, physiology, and endocrinology, has led to the use of chimpanzees in the discovery and development of drugs. In particular, chimpanzees have served as a model of human disease (Nath et al., 2000) and have been used in the area of drug safety (Mueller et al., 1985). For similar reasons, chimpanzees have also been used as a pharmacokinetic model for the selection of drug candidates for clinical development (Table 1).

In general, the published data suggest that the chimpanzee may be a good pharmacokinetic model for humans. However, the published data are far from comprehensive, particularly for compounds eliminated primarily by hepatic clearance, since many of the literature examples display either significant renal elimination (i.e., carbapenems, cilastatin, ceftriaxone) or are metabolized by non-cytochrome P450 (P450) pathways (i.e., caspofungin; Groll and Walsh, 2001). Because of their size, many physiological relationships in chimpanzees that affect drug clearance, such as organ blood flow, should be similar to those clinical parameters in humans. However, published data comparing other important parameters influencing drug disposition such as drug-metabolizing enzyme activities in the liver, protein binding, and intestinal permeability are lacking. A careful characterization of the factors affecting drug disposition in the chimpanzee is necessary so that rational extrapolations can be made to humans.

This article describes a retrospective analysis of what has been learned from utilizing the chimpanzee as a pharmacokinetic surrogate for humans during the period from 1998 to 2001. Included are studies examining the disposition of well characterized, marketed drugs (i.e., propranolol, verapamil, and theophylline; Fig. 1) as well as a comparison of P450 activities in chimpanzee liver versus human liver. Also presented is a compilation of chimpanzee and human pharmacokinetic data for research and development candidates from the Bristol-Myers Squibb Company (formerly DuPont Pharmaceuticals Company). These comparisons illustrate the similarities and differences in drug disposition for these two species, so that chimpanzees can be utilized more appropriately in the drug discovery and development process.

Materials and Methods

Chimpanzee in Vivo Pharmacokinetic Studies with Propranolol, Verapamil, and Theophylline. The in vivo chimpanzee studies described in this article were conducted according to protocols approved by the Animal Care and Use Committee at the New Iberia Research Center (New Iberia, LA). Studies performed were either oral or i.v. studies. Blood and urine samples were collected at specific intervals. Blood was drawn from either the cephalic, femoral, or saphenous veins and collected in tubes containing either sodium citrate or EDTA. After collection, blood samples were centrifuged, and the plasma supernatant was separated and stored frozen at -20°C. When required, urine samples were obtained via a collection pan placed under the cage that drained into a vessel maintained at 4°C by wet ice. Urine samples were stored frozen at -20°C. Plasma and urine samples were analyzed using HPLC-fluorescence, HPLC-UV, or LC/MS/MS. All intravenous studies were performed on chimpanzees sedated using ketamine. Oral studies were performed on alert chimpanzees whenever possible. Chimpanzees used in these studies ranged from 59.0 to 78.6 kg and average weight was 64.6 ± 6.6 kg. Specifics of each study are detailed below.

Verapamil. One chimpanzee (male, 60.4 kg) was administered a 0.5 mg/kg intravenous dose of verapamil HCl (Sigma-Aldrich, St. Louis, MO) in isotonic saline via a 30-min infusion. Two additional chimpanzees (two males, 61.7 kg and 78.6 kg) were administered a 5 mg/kg oral dose of verapamil HCl in 100% Tang orange drink (Kraft Foods, Northfield, IL). Blood (∼5 ml) was collected at predose, 15, 30, 35, and 45 min, and 1, 1.5, 4, 8, 12, 16, and 24 h postinfusion for the i.v. study; and at predose, 2, 4, 8, 12, 16, and 24 h postdose for the p.o. study. Cumulative urine samples were collected for 24 h.

Propranolol. One chimpanzee (male, 60.6 kg) was administered a 0.5 mg/kg intravenous dose of propranolol HCl (Sigma-Aldrich) in isotonic saline via a 10-min infusion. Two chimpanzees (two males, 60.7 kg and 61.0 kg) were administered a 2 mg/kg oral dose of propranolol HCl in 100% Tang orange drink (Kraft Foods). Blood (∼5 ml) was collected at predose, 5, 15, 30, and 60 min, and 2, 4, 8, 12, and 24 h postinfusion for the i.v. study; and at predose, 4, 8, 12, 16, and 24 h postdose for the p.o. study. Cumulative urine samples were collected for 24 h.

Theophylline (as Aminophylline). One chimpanzee (male, 68.3 kg) was administered a 0.5 mg/kg intravenous dose of aminophylline (Sigma-Aldrich) in isotonic saline via a 30-min infusion. Two additional chimpanzees (males, 59.0 kg and 70.9 kg) were administered a 1 mg/kg oral dose of aminophylline in 100% Tang orange drink (Kraft Foods). Blood (∼5 ml) was collected at predose, 30, 45, and 60 min, and 2, 4, 6, 8, 10, 12, 16, and 24 h postinfusion for the i.v. study; and at predose, 4, 8, 12, 16, and 24 h postdose for the p.o. study.

Chimpanzee in Vivo Pharmacokinetic Studies with Proprietary Compounds. In oral experiments, chimpanzees (n = 1, 2, 3, or 4) were given oral doses of proprietary compounds (0.8-15 mg/kg) either as a solution or suspension. In the i.v. experiments, chimpanzees (n = 1 or 4) were given an intravenous dose of compound (0.07-10 mg/kg) via either a 30-min or 1-h infusion. The i.v. studies with DPC 423, DPC 333, and DPC 906 were performed as coadministration (cassette) studies in which these compounds were administered in combination with 4 to 10 other compounds at low doses (0.07-0.5 mg/kg). For all experiments, blood was collected at specified times after dosing in the manner described above. Chimpanzees used in these studies ranged in weight from 58 to 99 kg.

Pharmacokinetics Studies in Rats, Dogs, and Monkeys with Proprietary Compounds. Sprague-Dawley rats (n = 3-6 per compound), beagle dogs (n = 2-7 per compound), and cynomologus or rhesus monkeys (n = 2-8 per compound) were given doses of proprietary compounds either as a solution or a suspension by oral gavage. Compounds were given at oral doses of 5 to 30 mg/kg in rats, 5 to 10 mg/kg in dogs, and 5 to 45 mg/kg in monkeys. For all experiments, blood was collected at specified times after dosing and placed in tubes containing either sodium citrate or EDTA. After collection, blood samples were centrifuged, and the plasma supernatant was separated and stored frozen at -20°C. Plasma concentrations were determined by LC/MS/MS.

Human in Vivo Pharmacokinetic Studies with Proprietary Compounds. The data presented in this article are from phase I clinical trials in which development candidates were administered to healthy male volunteers as a single oral dose (Note: The only exception is for DPC 974, which was administered to female volunteers.) Briefly, blood was collected at specified intervals postdose; plasma was obtained by centrifugation and stored frozen. In instances where more than one dose level was administered, data from the dose level closest to the dose administered to the chimpanzee were chosen, and the dose was normalized.

Pharmacokinetic Analyses and Comparisons to Human Data. All pharmacokinetic parameters were calculated by noncompartmental methods as described in Gibaldi and Perrier (1982). Parameters in chimpanzees are presented as either an individual value (n = 1), a mean (n = 2), or a mean ± standard deviation (n = 3 or more). Due to the limited sample sizes, no attempts have been made to statistically compare the pharmacokinetic parameters from chimpanzees and humans for individual compounds.

Correlational analyses were performed on comparisons of oral clearances from humans with oral clearances from chimpanzees, Sprague-Dawley rats, beagle dogs, and monkeys using the SAS JMP 5.0.1a software package (SAS Institute, Cary, NC). Correlations were considered significant at p < 0.05.

In Vitro Comparison of P450 Activities. Liver (50-, 70-, or 125-g portions) was collected from three adult chimpanzees (one female, 10 years of age; two males, 9 and 10 years of age) by surgical resection under general anesthesia and immediately cut into small cubes; portions were flash frozen in liquid nitrogen. The liver tissue was processed into microsomes using standard methods as described by Lake (1987), and total protein concentrations were determined by the Lowry method (Lowry et al., 1951). After preparation, the microsomes were stored in aliquots in 0.5-ml cryovials at -80°C until use.

Chimpanzee microsomes were compared with pooled human liver microsomes (pool of 10) (In Vitro Technologies, Baltimore, MD) using substrates that have been demonstrated to be selective for human P450 isozymes. Specifically, the following substrate reactions were monitored for P450 activities according to the referenced methods: testosterone 6β-hydroxylation for 3A4 (Sonderfan et al., 1987), tolbutamide methylhydroxylation for 2C9 (Miners et al., 1988), mephenytoin 4-hydroxylation for 2C19 (Wrighton et al., 1993), phenacetin O-deethylation for 1A2 (Newton et al., 1995), testosterone 16β-hydroxylation for 2B6 (Sonderfan et al., 1987), and dextromethorphan O-demethylation for 2D6 (Kronbach, 1991). No marked differences were observed in the P450 enzyme activities for the female chimpanzee liver microsomes (n = 1) when compared with the male chimpanzee liver microsomes (n = 2). Therefore, substrate reaction rates are presented as a mean ± standard deviation for the chimpanzee liver microsomes (n = 3) and as a mean value for the pooled human liver microsomes.

In Vitro Comparison of Microsomal Intrinsic Clearance of Propranolol, Verapamil, and Theophylline. Estimates of intrinsic clearance were generated using pooled human liver (pool of 10 males) (BD Gentest, Woburn, MA) or chimpanzee liver microsomes (n = 2, male). Briefly, a 1 μM concentration of compound was incubated with 1 mM NADPH and 0.25 mg/ml of microsomal protein at 37°C. Incubations for each compound were performed in triplicate and were terminated at 0, 5, 10, 15, 20, and 30 min after the start of the incubations. The conditions were modified for theophylline due to its low turnover. Briefly, 50 μM theophylline was incubated with 1 mM NADPH and 2 mg/ml of microsomal protein at 37°C. Incubations were terminated at 0, 15, 30, and 40 min after the start of the incubations. Chimpanzee liver microsomes were pooled and performed in duplicate for the theophylline experiments to conserve microsomes. Although 50 μM is a significantly higher concentration than the 1 μM used for propranolol and verapamil, it is well below the K_m values reported for the metabolic pathways responsible for theophylline elimination (Tjia et al., 1996). Compound disappearance was monitored by LC/MS/MS or HPLC-UV. Intrinsic clearances were calculated and scaled to systemic clearance using the well stirred model (disregarding plasma protein binding) as described by Obach et al. (1997). For the chimpanzee, the liver weight used for scaling was assumed to be the same as for human (i.e., 1800 g), and chimpanzee hepatic blood flow was assumed to be 1.53 l/h/kg (see section on hepatic blood flow under Discussion). Results are presented as mean values of the triplicate determinations for propranolol and verapamil, and mean values of the duplicate determinations for theophylline.

Protein Binding Studies. Unbound or free fractions for all compounds presented were determined in vitro by either equilibrium dialysis or ultrafiltration. For the equilibrium dialysis experiments, compounds were added to pooled plasma or serum (human n = 6, chimpanzee n = 3; Bioreclamation Inc., Hicksville, NY) and equilibrated against isotonic phosphate buffer at 37°C using either Kel-F (San Diego Plastics Inc., National City, CA) (equilibrated for 24 h) or DiaNorm (DiaNorm, Munich, Germany) (equilibrated for 2-4 h) dialysis cells. After the equilibration period, concentrations of compound were measured and unbound fraction was calculated by dividing the buffer concentration by the plasma or serum concentration following dialysis.

For the ultrafiltration experiments (acetaminophen only), the unbound fraction of compound in pooled plasma or serum was determined by an ultrafiltration procedure at 1000g using Centrifree micropartition devices (Millipore Corporation, Bedford, MA). Unbound fractions were determined by dividing the compound concentration in the ultrafiltrate by the initial concentration of compound in the plasma or serum.

Compound concentrations in samples of plasma, serum, ultrafiltrate, and buffer were determined using either HPLC-fluorescence, HPLC-UV, or LC/MS/MS. All determinations were performed in triplicate. Since pooled plasma or serum was used in all determinations, bound and unbound fractions are presented as a mean of all determinations.

Western Blot Analysis. Pooled human liver microsomes (pool of 10 males) (BD Gentest) or chimpanzee liver microsomes (n = 2; male) were Western blotted as described by Ausubel et al. (1994). Briefly, microsomal proteins (20 μg used for anti-CYP3A4; 40 μg for all other antibodies) were separated by SDS-polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane, and reacted with antibodies recognizing human CYP3A4, CYP2D6, CYP2B6, CYP2C9, CYP1A2 (Xenotech, Lenexa, KS), or CYP2C19 (BD Gentest). Proteins were visualized using a WesternBreeze chromogenic Western blot immunodetection kit (Invitrogen, Carlsbad, CA).

Results

Chimpanzee in Vivo Pharmacokinetic Studies with Propranolol, Verapamil, and Theophylline. The pharmacokinetic parameters obtained from the chimpanzee in vivo studies with propranolol, verapamil, and theophylline (Fig. 1) are presented in Table 2. Systemic clearance, half-life, and V_dss (volume of distribution at steady state) for propranolol and verapamil in chimpanzees were comparable to literature values for humans. Propranolol was not detected in the plasma after oral administration. However, the presence of the drug in the urine demonstrates that it was absorbed. Based upon the limit of quantitation of the analytical assay, the bioavailability of propranolol was estimated to be <5% for both orally dosed chimps. Similarly, verapamil was not detected in the plasma from one of the two chimps that received the compound orally. The bioavailability of verapamil was estimated to be <1% for this chimp based on the limit of quantitation of the analytical assay. In the second orally dosed chimpanzee, the bioavailability of verapamil was estimated to be 4%. For both propranolol and verapamil, it appears that the bioavailability in the chimpanzee is less than that observed in humans. The contribution of renal excretion to the overall elimination of these two drugs is very low, accounting for less than 1% of the administered dose, and is similar to literature values for humans (Table 2).

In contrast, the systemic clearance for theophylline was ∼5-fold higher in the chimpanzee, resulting in an ∼5-fold shorter half-life than observed in humans (Table 2). The V_dss for theophylline was similar in two species (Table 2). Bioavailability was high in both species, being 78% and 96 ± 8% (Benet et al., 1996) in the chimpanzee and human, respectively.

Intrinsic Clearance Studies with Propranolol, Verapamil, and Theophylline in Human and Chimpanzee Liver Microsomes. Microsomal intrinsic clearances (CL_{int microsomal}) and predicted systemic clearances (CL_{s in vitro}), and bioavailabilities (F_{in vitro}) based upon human and chimpanzee liver microsome data are presented in Table 3. CL_{int microsomal} rates appeared to be higher in chimpanzee liver microsomes for propranolol (∼10-fold) and theophylline (∼5-fold) and similar for verapamil. CL_{s in vitro} for all three compounds agreed well with the in vivo systemic clearances reported in Table 2. Chimpanzee and human in vivo bioavailabilities for propranolol and theophylline were predicted well using the in vitro microsome data (Tables 2 and 3). For verapamil, F_{in vitro} determined with human liver microsomes was lower than the bioavailability observed clinically in humans (Tables 2 and 3). In contrast, F_{in vitro} for verapamil determined with chimpanzee liver microsomes was higher than the observed in vivo bioavailability in chimpanzees. In fact, the in vitro microsome data suggest that bioavailability should be similar between the two species (Table 3). This contrasts the in vivo situation, where verapamil bioavailability appears to be higher in humans (Table 2).

Protein Binding Studies with Theophylline, Propranolol, Verapamil, and Acetaminophen. The unbound fractions for theophylline, propranolol, verapamil, and acetaminophen in chimpanzee and human serum are presented in Table 4. The unbound fraction for all compounds was similar for both species.

In Vitro Comparison of P450 Activities. A comparison of the P450 enzyme activities in chimpanzee and human liver microsomes using substrates that have been demonstrated to be selective for human P450 isozymes is presented in Table 5. Rates of testosterone 6β-hydroxylation (selective for 3A4) and tolbutamide methylhydroxylation (selective for 2C9) (Rendic, 2002) were comparable in both species (Table 5). Dextromethorphan O-demethylation (selective for 2D6), phenacetin O-deethylation (selective for 1A2), mephenytoin 4-hydroxylation (selective for 2C19), and testosterone 16β-hydroxylation (selective for 2B6) activities (Rendic, 2002) in chimpanzee microsomes were greater than the respective activities observed in human liver microsomes by at least 4- to 10-fold. The selectivity of these substrates for chimpanzee isozymes has not been established. More extensive biochemical characterization of these activities with chemical inhibitors and inhibitory antibodies was not feasible due to the limited supply of nonrenewable chimpanzee microsomes.

Western Blot Analysis of Chimpanzee and Human Liver Microsomes.Figure 2 presents a Western blot analysis of chimpanzee and human liver microsomes using antibodies reactive against human CYP3A4, CYP2B6, CYP2C9, CYP2D6, CYP1A2, and CYP2C19. Since the antibodies used in these studies were raised against human P450s, the data should be considered more qualitative, rather than quantitative, in nature. As seen from Fig. 2, it appears that immunoreactive proteins are present in chimpanzee liver microsomes analogous to each of the human P450s investigated. In all cases, the antibodies reacted with a single protein in chimpanzee liver microsomes that migrated similarly to a corresponding human P450 during electrophoresis. The density of the immunoreactive protein bands in chimpanzee liver microsomes appeared to be higher when anti-CYP2B6, anti-CYP2D6, and anti-CYP1A2 were used.

Comparisons of in Vivo Pharmacokinetic Data for Proprietary Compounds in Humans and Chimpanzees.Table 6 contains chimpanzee pharmacokinetic data for 12 proprietary compounds (Fig. 3) with the corresponding human pharmacokinetic data from phase I clinical trials. In our comparison of chimpanzee and human oral pharmacokinetic parameters, differences less than 3-fold were considered to be within biological variability. Differences of ∼3- to 5-fold in hepatic 3A activities and midazolam i.v. clearance have been observed in healthy human subjects (Kinirons et al., 1999; Lee et al., 2002). Variability of midazolam oral clearance (CL_s/F) is in the higher end of this range, being approximately 5-fold (Kinirons et al., 1999). The pharmacokinetics following oral dosing is expected to be inherently more variable compared with intravenous administration, since differences in formulation, gastric motility, and transporter expression influence oral exposure. When taking into consideration both the magnitude of variability observed in humans and the route of administration of our studies, setting a biological variability of 3-fold or less is reasonable for our evaluation.

Of the 12 proprietary compounds studied, three compounds (DPC 423, DPC R1, and razaxaban) had oral clearances that differed more than 3-fold (Table 6; Fig. 4). Consistent with the higher oral clearance, DPC 423 also had a lower C_max, a smaller AUC, and a shorter t_1/2 in the chimpanzee (Table 6). For DPC R1 and razaxaban, AUC was the only other parameter aside from oral clearance that was greater than 3-fold different in the two species (Table 6). C_max for DPC 333 and t_1/2 for DMP 851 were also greater than 3-fold different when comparing the chimpanzee data to the human (Fig. 4). However, the oral clearances of these two compounds were less than 3-fold different.

Table 7 summarizes the pharmacokinetics for those proprietary compounds that were also dosed intravenously to chimpanzees. A wide range of systemic clearances was observed, from 0.016 l/h/kg (DPC 083) to 1.2 l/h/kg (DPC R1) (Table 7). All t_1/2 values estimated from the i.v. studies were similar to values obtained after oral dosing.

Species Comparisons of Oral Clearance of Proprietary Compounds.Figure 5 shows a comparison of human oral clearance versus chimpanzee, Sprague-Dawley rat, beagle dog, and monkey oral clearances for the proprietary compounds in Fig. 3. CL_s/F estimates were available for all 12 compounds in humans, chimpanzees, and Sprague-Dawley rats. In beagle dogs, CL_s/F estimates were compared for 10 of the 12 proprietary compounds since estimates were not available for DMP 961 and DPC 083 (Fig. 5C). Also, DMP 851, DPC 423, DPC 974, and DPC 681 were not dosed orally in monkeys; thus, the comparison was performed with only eight compounds (Fig. 5D). Oral clearances were significantly correlated (r = 0.68, p = 0.015) in the comparison between human CL_s/F and chimpanzee CL_s/F (Fig. 5A). Correlations were not significant for all other comparisons.

Protein Binding Studies with Proprietary Compounds. A summary of the physicochemical characteristics of nine proprietary compounds for which unbound fractions were obtained for both the chimpanzee and human is presented in Table 8. In general, the unbound fraction of all compounds tested was similar in both species (Table 8). The only exception was compound 1, where the unbound fraction in chimpanzee plasma (0.61%) was ∼2.5-fold higher than the corresponding value in human plasma (0.24%). However, this difference may be of no consequence considering the difficulty in the measurement of unbound fractions <1%.

Discussion

The chimpanzee represents a unique species for pharmacokinetics and drug disposition studies. Because of its similar size and phylogenetic relationship to humans, there is an inclination to view this species as an image of human drug metabolism and disposition. Few controlled pharmacokinetic studies exist in the literature to rigorously establish this view because there are few animals and sites that are capable of conducting such studies. In addition, most studies utilize very few animals, thus limiting the ability to compare pharmacokinetic parameters. The current work describes the first comprehensive effort to characterize this animal model using pharmacokinetic and biochemical probes.

Early evaluations of the chimpanzee as a preclinical safety model revealed similarities in metabolic pathways such as glutathione and glucuronide conjugation (Mueller et al., 1985). An examination of the excretion routes for the industrial solvent, trichloroethylene, and its metabolites in three primate species found that excretion patterns in chimpanzees were in close agreement with observations in humans (Mueller et al., 1982). As discussed and summarized in Table 1, the chimpanzee has been used as a pharmacokinetic model in drug discovery with relative success. Although the chimpanzee pharmacokinetics of compounds in Table 1 generally parallel that in humans, many of these compounds display significant renal elimination (i.e., carbapenems, cilastatin, ceftriaxone) or are metabolized by non-P450 pathways (i.e., caspofungin; Groll and Walsh, 2001), or there is insufficient information regarding the identity of the enzymes involved in any observed hepatic elimination. Because many drugs are eliminated primarily via hepatic metabolism, a comprehensive examination of species differences in factors affecting drug disposition for hepatically eliminated compounds is needed.

Two high metabolically cleared (propranolol, verapamil) drugs and one low metabolically cleared (theophylline) drug were initially chosen to be administered to chimpanzees to better characterize species differences in drug disposition (Fig. 1). Propranolol, a β-adrenergic receptor antagonist, is the prototypical high clearance compound whose metabolism in humans is mediated largely by oxidative metabolism and glucuronidation in the liver (Ward et al., 1989). Cytochrome P450 2D6, 1A2, and 2C19 appear to be involved in the formation of propranolol's main oxidative metabolites, 4-hydroxypropranolol (2D6) and naphthoxylacetic acid (1A2, 2C19) (Ward et al., 1989; Yoshimoto et al., 1995). Verapamil, a calcium channel antagonist, is pharmacokinetically similar to propranolol. However, its clearance in humans is largely via CYP3A4-mediated N-dealkylation to D-617 (2-(3,4-dimethoxyphenyl)-5-methylamino-2-isopropylvaleronitrile) and N-demethylation to norverapamil (Tracy et al., 1999). The systemic clearance of both these compounds approximates the reported estimate of hepatic blood flow in humans (∼1.2 l/h/kg; Davies and Morris, 1993) (Table 2). In contrast, the bronchodilator, theophylline, is a low clearance compound that is completely absorbed and undergoes negligible first-pass extraction by the liver in humans (Table 2). Theophylline is cleared in the liver via hydroxylation to 1,3-dimethyluric acid or N-demethylation to 3-methylxanthine or 1-methylxanthine. These reactions are catalyzed primarily by CYP1A2 (Ha et al., 1995).

Distinct species differences exist when examining the pharmacokinetics of propranolol, theophylline, and verapamil in the chimpanzee and human. The systemic clearance of propranolol and verapamil appears similar in the two species. However, the bioavailability of these compounds appears to be lower in the chimpanzee. The opposite situation appears true for theophylline. The systemic clearance of theophylline appears to be higher in the chimpanzee; however, bioavailability appears to be similar in the two species. To understand the nature of these differences, an examination of species differences in the parameters governing systemic clearance of hepatically cleared compounds (i.e., hepatic blood flow, protein binding, and metabolic enzyme activity) must be performed.

Hepatic Blood Flow. To date, there have been no experimental determinations of hepatic blood flow in the chimpanzee. However, similarities in hepatic blood flow are expected in the two species due to their similarities in size and organ weights. The in vivo pharmacokinetic data for propranolol support this assumption. The liver plays a significant role in the metabolism of propranolol in all species examined (Evans et al., 1973), and there are no reports of significant intestinal metabolism of propranolol. Its lack of significant renal elimination further supports the conclusion that hepatic clearance is the predominant route of elimination in chimpanzees (Table 2). The low oral bioavailability for propranolol suggests almost complete hepatic extraction in chimpanzees; therefore, the systemic clearance of propranolol should provide an estimate of hepatic blood flow. An estimate for hepatic blood flow of approximately 1.53 l/h/kg (Table 2) in the chimpanzee is similar to the reported hepatic blood flow estimate of 1.2 l/h/kg (Davies and Morris, 1993) in humans, and appears reasonable based upon the similarity in size to humans of the chimpanzee used in our study (i.e., 60.6 kg). However, if significant extrahepatic elimination of propranolol occurs in chimpanzees, the proposed hepatic blood flow approximation would be an overestimation.

Protein Binding. There are limited data comparing the unbound fractions of compounds between chimpanzees and humans, largely because the chimpanzee is not a commonly used preclinical species. All previous literature reports of compounds in which protein binding information is available for both species have shown similarities in the unbound fraction (Table 9). Our own data with propranolol, verapamil, theophylline, and acetaminophen, as well as the proprietary compounds, lead us to draw similar conclusions. Figure 6 shows the relationship between human versus chimpanzee unbound fraction for both the literature data presented in Table 9 and our own data presented in Tables 4 and 8. The slope of 0.93 for the regression line suggests an excellent agreement across a wide range of unbound fractions. These data imply that the major proteins involved in the binding of xenobiotics, albumin and α₁-acid glycoprotein are similar in both species.

P450 Enzyme Activity. The collection of liver tissue from healthy chimpanzees for biochemical characterization is rare. With respect to the major P450s involved in the metabolism of drugs, the information presented in Table 5 suggests that chimpanzees have levels of CYP3A (testosterone 6β-hydroxylation) and 2C9 (tolbutamide methylhydroxylation)-like enzyme activity similar to those of humans. However, levels of CYP2D (dextromethorphan O-demethylation) and 1A (phenacetin O-deethylation)-like enzyme activity appear to be higher (∼10-fold) in the chimpanzee. Interestingly, the Western blot analysis in Fig. 2 is consistent with the enzyme activity data presented in Table 5 in that the immunoreactive protein bands appear darker in chimpanzee liver microsomes probed with antibodies reactive against human CYP2D6 and 1A2. From our limited dataset, there is no evidence of sex differences in P450 enzyme activity in chimpanzees. As mentioned under Materials and Methods, the P450 enzyme activity observed in liver microsomes from the one female chimpanzee were similar to activities observed in liver microsomes from the two male chimpanzees. More extensive biochemical characterization of the chimpanzee liver microsomes with chemical inhibitors and inhibitory antibodies was not feasible due to the limited supply of these nonrenewable microsomes.

Previously, higher levels of 7-ethoxyresorufin-O-deethylase activity were observed in microsomes prepared using liver from one female chimpanzee when compared with microsomes from two human donors, whereas 7-benzyloxyresorufin O-dealkylase activity was similar (Rodrigues et al., 1995). Since 7-ethoxyresorufin-O-deethylase activity is mediated by the CYP1A subfamily and 7-benzyloxyresorufin O-dealkylase activity is mediated by the CYP3A and 2C subfamilies in humans (Rodrigues et al., 1995), these observations are consistent with our results using more selective substrates. Similarities in the rate of formation of CYP3A4-mediated indinavir metabolites in human and chimpanzee microsomes reported by Chiba et al. (2000) also support our data. Chimpanzees should be a useful model for compounds metabolized by CYP3A and 2C isozymes if substrate specificity is similar in the two species. These two subfamilies are responsible for the oxidative metabolism of approximately half of all marketed drugs (Lewis et al., 2002).

Similarities and Differences in the Pharmacokinetics of Propranolol, Verapamil, and Theophylline. The distinct species differences in the pharmacokinetics of propranolol, verapamil, and theophylline may be explained in part by the differences in P450 enzyme activity discussed previously. The similarities in the systemic clearance of propranolol and verapamil in chimpanzees and humans are expected since systemic clearance approximates hepatic blood flow for high clearance compounds metabolized primarily by the liver (Wilkinson and Shand, 1975). The species difference in the oral bioavailability of propranolol appears to be consistent with observed differences in hepatic enzyme activity.

Propranolol is mainly oxidized by CYP2D6, 1A2, and 2C19 in humans. If substrate specificity is similar in chimpanzees, the higher activity of the analogous isozymes in the chimpanzee would result in a slightly higher hepatic extraction in this species. If we assume that hepatic blood flow is similar in chimpanzees and humans, the higher hepatic extraction in chimpanzees would not cause significant species differences in systemic clearance. However, the higher hepatic extraction in the chimpanzee could result in a significant decrease in oral bioavailability in comparison to humans. The lack of detectable plasma concentrations in chimpanzees after oral administration is consistent with this scenario. The ∼10-fold higher CL_{int microsomal} and lower F_{in vitro} of propranolol, estimated using chimpanzee liver microsomes (Table 3), also supports the situation described above.

Unlike propranolol, the comparable estimates of verapamil CL_{int microsomal} from chimpanzee and human liver microsomes do not support the notion that verapamil's species difference in bioavailability results from a higher hepatic extraction in chimpanzees (Tables 2 and 3). Verapamil is metabolized primarily by CYP3A4 at clinically relevant concentrations in humans (Tracy et al., 1999). The similarity in verapamil CL_{int microsomal} is consistent with observed similarities in 3A-like activity (i.e., testosterone 6β-hydroxylation) in chimpanzee and human liver microsomes. The possibility that formulation differences would explain the lower bioavailability in chimpanzees is unlikely considering the reasonably good aqueous solubility of verapamil HCl (i.e., 83 mg/ml; Budavari, 1996). It has recently been demonstrated that intestinal first-pass metabolism plays a significant role in the oral absorption of verapamil (Von Richter et al., 2004). In addition, verapamil is a known competitive substrate/modulator of P-glycoprotein (Varma et al., 2003). Thus, it is possible that species differences in intestinal first-pass metabolism and/or transporter function may be responsible for the observed differences in verapamil's oral bioavailability.

In contrast to verapamil and propranolol, the systemic clearance of theophylline, a low clearance compound, is dependent on protein binding and metabolic enzyme activity (Wilkinson and Shand, 1975). Theophylline is primarily metabolized by 1A2 in humans. Assuming substrate specificity is similar in the two species, the higher systemic clearance and shorter half-life of this drug in the chimpanzee is consistent with higher hepatic CYP1A-like activity, as evidenced by higher microsomal phenacetin O-deethylase activity. Although systemic clearance appears to be ∼5-fold higher in the chimpanzee, theophylline would still be considered a low clearance compound in this species if hepatic blood flow is ∼1.53 l/h/kg. Therefore, substantial species differences in bioavailability are not predicted, despite the apparent increase in systemic clearance. This is consistent with both the in vivo and in vitro data for theophylline (Tables 2 and 3).

The volume of distribution appears to be similar in chimpanzees and humans for propranolol, verapamil, and theophylline, as well as for the literature compounds for which the volume of distribution was estimated (Tables 1 and 2). Since there appeared to be no gross species difference in protein binding for most compounds (Fig. 6), the observed similarities in V_dss suggest that tissue binding in chimpanzees and humans is also likely to be similar.

Oral Pharmacokinetics of Proprietary Compounds in Chimpanzees and Humans. The oral pharmacokinetics of 12 proprietary compounds of diverse chemical structure (Fig. 3) suggests that, in general, the chimpanzee is a reasonable pharmacokinetic model for the human when the P450 reaction phenotype is considered. The oral clearances (CL_s/F) in chimpanzees for nine of the compounds were less than 3-fold different from that observed in humans (Fig. 4A). These nine compounds, for the most part, had AUCs, C_max values, and t_1/2 values that were less than 3-fold different (Table 6; Fig. 4). Three of the 12 proprietary compounds (DPC R1, DPC 423, razaxaban), had oral clearances that differed more than 3-fold in biological variability (Fig. 4A).

The systemic clearance for DPC R1, presented in Table 7, categorizes DPC R1 as a high clearance compound in the chimpanzee. Based upon reaction phenotyping experiments, the oxidative metabolism of DPC R1 appears to be mediated by CYP3A4, 1A1, and 1A2 (unpublished results). If DPC R1 is metabolized by the corresponding orthologs in the chimpanzee, an over-expression of 1A-like activity in the chimpanzee liver could potentially increase the hepatic extraction of DPC R1 in this species. Thus, whereas systemic clearance would be unaffected by this change in hepatic extraction, oral bioavailability (F) would be decreased to a level similar to those effects observed for propranolol. A lower bioavailability would result in a higher oral clearance in chimpanzees when compared with humans, since oral clearance by definition is CL_s/F. This is consistent with estimates of DPC R1 oral clearance for chimpanzees and humans presented in Table 6.

In contrast to DPC R1, DPC 423 appears to be a low clearance compound in the chimpanzee based upon its systemic clearance (Table 7). In human liver microsomes, the main carboxylic acid metabolite as well as other metabolites of DPC 423 appear to be formed by CYP1A1, 2D6, and 3A4 (Mutlib et al., 2002). Higher activities of the CYP1A and 2D subfamilies in the chimpanzee could result in a higher systemic clearance in chimpanzees in comparison to humans. Since DPC 423 is a low clearance compound like theophylline, we would not expect the bioavailability of this compound to be very different between the species. This suggestion of higher systemic clearance, but similar bioavailability, is consistent with the higher oral clearance of this compound observed in the chimpanzee (Table 6).

Razaxaban and DPC 333 were two compounds whose oral clearances were approximately 3-fold different between chimpanzees and humans (Fig. 4), with the CL_s/F of razaxaban being slightly over 3-fold different, and the CL_s/F of DPC 333 being slightly under 3-fold different. Reaction phenotyping data suggest that the oxidative metabolism of these compounds is primarily mediated by CYP3A (unpublished results). The oral clearances of both compounds place them on the border of our 3-fold biological variability. The large difference in C_max for DPC 333 (∼4.5-fold; Fig. 4) could be due to experimental conditions, such as less frequent sampling times in the chimpanzee studies than in human clinical trials. Since the half-lives for both compounds were very similar between chimpanzees and humans (Fig. 4; Table 6), it is possible that the gastrointestinal absorption of these compounds may have been impacted by factors such as differences in formulation, intestinal first-pass metabolism, and/or transporter function.

Species Comparisons of Oral Clearance of Proprietary Compounds. Species comparisons of oral clearances of the proprietary compounds in Fig. 3 were performed to evaluate the chimpanzees as a pharmacokinetic model for humans relative to more common laboratory animal species (Fig. 5). Although CL_s/F estimates from chimpanzees and humans were directly compared in the preceding section, direct comparisons were not made in these analyses since hepatic blood flow in rats, dogs, and monkeys is higher than that observed in humans (Davies and Morris, 1993). Instead, correlational analyses were performed, so that all species examined could be evaluated on a more equal basis. The information in Fig. 5 suggests that the chimpanzee proved to be the animal species which most successfully rank ordered the oral clearance of proprietary compounds in humans.

The experience to date suggests that significant differences in major P450 activities exist between chimpanzees and humans. These differences appear to contribute to the discrepant pharmacokinetics observed for propranolol, theophylline, DPC R1, and DPC 423. However, conclusions must still be made cautiously because of the low number of chimpanzees used in the studies, combined with our lack of knowledge of the biological variability of P450 enzyme expression in this species. Differences in formulation may also have affected the oral absorption/oral pharmacokinetics of the proprietary compounds in chimpanzees and humans. Furthermore, possible species differences in P450 substrate specificity, differences in non-P450 drug-metabolizing enzymes, and transporters may exist and would contribute to compound-dependent differences in pharmacokinetics. Despite the differences observed in direct comparisons of pharmacokinetic parameters from chimpanzees and humans, the dataset as a whole suggests that the performance of the chimpanzee is at least equal, or arguably better, when compared with more common laboratory animal species for the assessment of human pharmacokinetics.

In summary, the data presented for propranolol, verapamil, theophylline, DPC R1, and DPC 423 emphasize the requirement to view the chimpanzee as yet another nonclinical species that should be used in drug discovery, with the appreciation of the biochemical differences between chimpanzees and humans. Although the chimpanzee proved to be useful as a pharmacokinetic model for humans, its use should be reserved for compounds whose clearance is governed by those enzymes which demonstrate species similarities in activity or by physiological processes that are similar in both species. Further characterization of chimpanzees should allow the more rational use of this species in the discovery and development of new medicines for important human diseases.