Abstract
Cynomolgus monkeys are widely used in preclinical studies during drug development because of their evolutionary closeness to humans, including their cytochrome P450s (P450s). Most cynomolgus monkey P450s are almost identical (≥90%) to human P450s; however, CYP2C76 has low sequence identity (approximately 80%) to any human CYP2Cs. Although CYP2C76 has no ortholog in humans and is partly responsible for species differences in drug metabolism between cynomolgus monkeys and humans, a broad evaluation of potential substrates for CYP2C76 has not yet been conducted. In this study, a screening of 89 marketed compounds, including human CYP2C and non-CYP2C substrates or inhibitors, was conducted to find potential CYP2C76 substrates. Among the compounds screened, 19 chemicals were identified as substrates for CYP2C76, including substrates for human CYP1A2 (7-ethoxyresorufin), CYP2B6 (bupropion), CYP2D6 (dextromethorphan), and CYP3A4/5 (dextromethorphan and nifedipine), and inhibitors for CYP2B6 (sertraline, clopidogrel, and ticlopidine), CYP2C8 (quercetin), CYP2C19 (ticlopidine and nootkatone), and CYP3A4/5 (troleandomycin). CYP2C76 metabolized a wide variety of the compounds with diverse structures. Among them, bupropion and nifedipine showed high selectivity to CYP2C76. As for nifedipine, CYP2C76 formed methylhydroxylated nifedipine, which was not produced by monkey CYP2C9, CYP2C19, or CYP3A4, as identified by mass spectrometry and estimated by a molecular docking simulation. This unique oxidative metabolite formation of nifedipine could be one of the selective marker reactions of CYP2C76 among the major CYP2Cs and CYP3As tested. These results suggest that monkey CYP2C76 contributes to bupropion hydroxylation and formation of different nifedipine oxidative metabolites as a result of its relatively large substrate cavity.
Introduction
Because of their evolutionary closeness and physiologic resemblance to humans, cynomolgus monkeys have been widely used in various biomedical research studies, including neuroscience and reproduction physiology studies. Age-related alterations of physiologic parameters in monkeys are in agreement with clinical observations in humans, suggesting that monkeys might be a suitable animal model for prediction of age-related changes in pharmacokinetics in humans (Koyanagi et al., 2014a,b). Although monkeys are frequently used in preclinical studies for drug development, differences in the pharmacokinetic profile of monkeys and humans are occasionally seen.
The cytochrome P450 (P450) superfamily consists of a large number of drug-metabolizing enzyme genes. P450s play a critical role in the metabolism of drugs, steroids, fatty acids, and environmental pollutants. In humans, the CYP1-3 families are important for drug metabolism, whereas the CYP4 family is mainly involved in metabolism of endogenous substrates and, to a lesser extent, drugs. To date, more than 20 P450 isoforms have been identified in cynomolgus monkeys (Uno et al., 2011a). It should be noted that P450s of primate species, such as cynomolgus monkeys, are evolutionarily much closer to those of humans than to those of dogs, rats, or mice, which are commonly used in drug metabolism studies and preclinical toxicity tests.
Most cynomolgus monkey P450 cDNAs have high sequence identities (≥90%) to human P450 cDNAs; however, CYP2C76 cDNA has only 75–78% sequence identity to human CYP2C cDNAs (Uno et al., 2006). In addition, CYP2C76 has no ortholog in humans (Uno et al., 2006, 2010), raising the possibility that CYP2C76 is involved in species differences in drug metabolism between cynomolgus monkeys and humans. CYP2C76 protein is expressed in the liver and catalyzes tolbutamide 4-hydroxylation and testosterone 2α-/16α-hydroxylation but not paclitaxel 6α-hydroxylation or S-mephenytoin 4′-hydroxylation, showing substrate specificity differences from those of other cynomolgus monkey and human CYP2Cs (Uno et al., 2006). CYP2C76 also metabolizes non-CYP2C substrates in humans, such as 7-ethoxyresorufin and bufuralol (Uno et al., 2011b). Pitavastatin is metabolized largely by monkey CYP2C19 and CYP2C76, but CYP2C76 catalyzes the same reaction catalyzed by CYP3A in humans and is also involved in the formation of metabolites not found in humans (Uno et al., 2007). These facts suggest that CYP2C76 is responsible for the differences in drug metabolism between monkeys and humans. However, a broad evaluation of potential substrates for CYP2C76 has not yet been conducted to understand the substrate specificity of this isoform.
In this study, 89 marketed compounds, including human CYP2C and non-CYP2C substrates or inhibitors (Rendic, 2002), also found in Food and Drug Administration Drug-Drug Interaction Draft Guidance, 2006 (http://www.fda.gov/cder/guidance/6695dft.htm) and European Medicines Agency guidelines (http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2012/07/WC500129606.pdf), were screened as potential CYP2C76 substrates. From the compounds screened, 19 chemicals were identified as substrates for CYP2C76, and these structures had wide diversity, from relatively small-molecule (e.g., nootkatone) to larger-molecule (e.g., troleandomycin) compounds. Among the newly identified substrates, CYP2C76 metabolized nifedipine, forming a metabolite unique to cynomolgus monkeys, and thus substrate could be a selective marker reaction of CYP2C76 among a variety of monkey CYP2C and CYP3A mediated reactions. We report herein that CYP2C76 has relatively wide substrate specificity and ability to form unique metabolites, which could result in species differences in drug oxidation between monkeys and humans.
Materials and Methods
Nifedipine and bupropion were purchased from Wako (Osaka, Japan). Pitavastatin lactone and dehydronifedipine were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The other drugs in Table 1 were obtained from one of the following sources: Wako, Sigma-Aldrich (St. Louis, MO), Cosmo Bio (Tokyo, Japan), or Nacalai Tesque (Kyoto, Japan). Cynomolgus monkey P450 recombinant enzymes CYP2C76, CYP2C9, CYP2C19, CYP2B6, and CYP3A4 were expressed in Escherichia coli and subsequently purified as described previously (Iwata et al., 1998; Daigo et al., 2002). Other reagents used in this study were of the highest quality commercially available.
Oxidations of Nifedipine, Bupropion, and Other Possible Monkey P450 Substrates.
The substrate was 1 or 20 μM dissolved in final concentrations of 0.01–0.02% dimethylsulfoxide or 1% methanol. Incubation mixtures contained substrate, 10–25 pmol/ml recombinant cynomolgus monkey P450, 0.25 mM β-NADP+, 2.5 mM glucose 6-phosphate, 0.025 U glucose-6-phosphate dehydrogenase, and 30 mM magnesium chloride in a final volume of 100 μl of 50 mM potassium phosphate buffer, pH 7.4. After incubation at 37°C for 0–30 minutes, reactions were terminated by addition of 200 μl of ice-cold methanol containing 5 nM terbinafine or 300 nM mefenamic acid as internal standards. Samples were then centrifuged at 21,600g for 10 minutes, and the supernatants were analyzed by liquid chromatography–tandem microscopy (LC-MS/MS). All incubations were performed in duplicate.
LC-MS Analytical Methods.
The LC system consisted of a pump and a Nanospace SI-2 autosampler (Shiseido, Tokyo, Japan) using an analytic C18 reversed-phase column (CAPCELL CORE C18, 2.1 × 50 mm, 2.7 μm; Shiseido) with a TSQ Quantum Ultra Triple-Quadrupole mass spectrometer (Thermo Fisher Scientific, Waltham, MA). The injection volume was 2–10 μl, and LC conditions were as follows: solution A contained 0.1% formic acid (v/v) in water, and solution B contained 0.1% formic acid (v/v) in methanol. The following gradient program was used at a flow rate of 0.40 ml/minute: 0–0.1 minute, linear gradient from 98% A to 0% A (v/v); 0.1–1.6 minutes, hold at 0% A; 1.6–1.7 minutes, linear gradient from 0% A to 98% A (v/v); 1.7–4.2 minutes, hold at 98% A. After electrospray ionization, compounds were analyzed by selected reaction-monitoring mode. The peak area ratio of the analyte to the internal standard was determined for each injection and used to measure substrate depletion. In the chromatographic analysis of nifedipine and bupropion metabolites, CAPCELL PAC C18 ACR column (2.0 × 150 mm, 5 μm; Shiseido), and the following gradient program were used at a flow rate of 0.40 ml/minute: 0–3 minutes, hold at 98% A; 3–50 minutes, linear gradient from 98% A to 0% A (v/v); 50–55 minutes, hold at 0% A; 55–55.1 minutes, linear gradient from 0% A to 98% A (v/v); 55.1–60 minutes, hold at 98% A. Single MS scan was used to monitor the metabolites.
Data Analysis.
Residual ratios at 30 minutes after incubation of each substrate were calculated and converted to substrate disappearance ratios by the following equation (Eq. 1):
(1)In vitro intrinsic clearance (CLint) was determined by the following equation, where k is the slope of the linear regression from the log concentration versus incubation time (for 0, 15, and 30 minutes) relationships (Eq. 2):
(2)Docking Simulation.
Cynomolgus monkey CYP2C76 primary sequence was aligned with a crystal structure of human CYP2C9 (Protein Data Bank code 1R9O) (Wester et al., 2004) using Molecular Operating Environment software (version 2013.10, Computing Group, Montreal, QC, Canada) for modeling of the three-dimensional structure. Before docking simulation, the energy of the P450 structures was minimized using the CHARMM22 force field. Docking simulation was carried out for nifedipine binding to P450 enzymes using the MMFF 94x distributed in the ASE Dock software (Ryoka Systems, Tokyo, Japan). Solutions were generated for each docking experiment and ranked according to the total interaction energy (U value).
Results
CYP2C76 Metabolism of 89 Compounds.
A total of 89 compounds (Table 1) were evaluated for their potential to undergo metabolism by CYP2C76 and other monkey P450s (CYP2C9, CYP2C19, CYP3A4, and CYP2B6). Pitavastatin lactone served as positive control because of the important role of CYP2C76 for monkey-specific metabolism of pitavastatin lactone (Uno et al., 2007). From the results of substrate depletion assay, 19 compounds showed relatively rapid metabolism with CYP2C76 (substrate disappearance ratios >20%) (Fig. 1A). Most of these compounds were also metabolized by monkey CYP2C9 and/or CYP2C19; however, nifedipine and bupropion exhibited higher selectivity to CYP2C76 (Fig. 1B). Because nifedipine (Guengerich et al., 1986) and bupropion (Hesse et al., 2000) are known as substrates for CYP3A4 and CYP2B6 in humans, respectively, metabolic activity toward the corresponding monkey P450s were also evaluated. CLint values are shown in Table 2. In monkeys, nifedipine was also metabolized by CYP3A4, and the CLint value was higher than that of CYP2C76. For bupropion, which was also metabolized by monkey CYP2B6, the CLint value was lower than that of CYP2C76. Compared with pitavastatin lactone, the CLint values of nifedipine and bupropion metabolized by CYP2C76 were 10-fold higher and similar, respectively.
Metabolite Profiling of Nifedipine and Bupropion.
Because nifedipine and bupropion showed relatively higher selectivity to CYP2C76, their metabolites were further investigated by LC-MS. Figure 2 shows the extracted ion chromatograms of the metabolites observed for nifedipine and bupropion. For nifedipine, [M − H]− ion of a CYP2C76-specific metabolite was detected at m/z 361 (Fig. 2A). In addition, [M + H]+ ion of another nifedipine metabolite at m/z 345 was observed with monkey CYP3A4, and this metabolite was also detected with CYP2C76 (Fig. 2A). For bupropion, [M + H]+ ion at m/z 256 was formed by both CYP2C76 and monkey CYP2B6 (Fig. 2B). An [M − H]− ion of new nifedipine metabolite at m/z 361 and [M + H]+ ion of bupropion metabolite at m/z 256 were 16 Da greater than the parent compounds, suggesting that these were oxidative metabolites of the parent compounds. An [M + H]+ ion of known nifedipine metabolite at m/z 345 was 2 Da lower than the parent compound, suggesting that this metabolite was formed by dehydrogenation of the 1,4-dehydropyridine ring of nifedipine. To obtain structural information of a new nifedipine metabolite, mass spectrum analysis was also conducted by LC-MS in product ion scan mode. Nifedipine and its oxidative metabolite generated the same fragment ion of m/z 122, indicating inclusion of a nitrobenzene unit in their structures (Fig. 3A). The product ions of m/z 222 and m/z 238 observed in nifedipine and its oxidative metabolite under the present conditions, respectively, indicated that the other unit of nifedipine (1,4-dehydropyridine form) was oxidized by CYP2C76 (Fig. 3A). An [M + H]+ ion of another nifedipine metabolite at m/z 345 was also subjected to product ion scan. The spectral pattern of fragment ions was identical to that of the dehydronifedipine standard, indicating that this metabolite was the known dehydronifedipine (Fig. 3B). Bupropion and its metabolite showed different spectral patterns of product ions. The product ion of m/z 184 in Fig. 3C (a) was considered to be generated by elimination of a t-butyl group of bupropion. The fragment ion of m/z 238 in Fig. 3C (b) was considered to be generated by neutral loss of water from the bupropion metabolite. The latter metabolite has been known to be generated in other species such as humans and rats, and the obtained spectral data were consistent with the previous report (Yeniceli et al., 2011), suggesting that the same metabolite is also generated in monkeys.
Docking Simulation.
Molecular dockings of nifedipine and bupropion to the modeled CYP2C76 enzyme were investigated (Fig. 4). The methyl carbon that branched from the 1,4-dihydropyridine ring of nifedipine was positioned toward the active site of CYP2C76; the ligand-P450 interaction energies (U values) and distance from center of the heme were found to be −36.1 kcal/mol and 4.6 Å, respectively (Fig. 4A). This finding suggested that this position of methyl carbon that branched from the 1,4-dihydropyridine ring of nifedipine could become a metabolic site of CYP2C76, consistent with LC-MS/MS analysis shown in Fig. 3A. In bupropion, the methyl carbon of the t-butyl group was positioned toward the active site of CYP2C76; U values and distance from the center of the heme were −29.2 kcal/mol and 6.5 Å, respectively (Fig. 4B), indicating this position to be a better metabolic site of CYP2C76. These simulation results were consistent with the known structure of bupropion metabolites reported previously (Yeniceli et al., 2011). The estimated structures of nifedipine and bupropion metabolites are shown in Fig. 5.
Discussion
Several compounds, including pitavastatin lactone, testosterone, and 7-ethoxyresorufin, have been identified as substrates for CYP2C76; however, a broad evaluation of potential substrates for CYP2C76 has not yet been conducted to understand the substrate specificity of this isoform. In this study, 89 marketed compounds, including human CYP2C and non-CYP2C substrates or inhibitors, were screened as potential substrates for CYP2C76. From the compounds screened, 19 chemicals were identified as substrates for CYP2C76 (Fig. 1A). These structures consisted of the substrates or inhibitors for human P450s listed in the Food and Drug Administration Drug-Drug Interaction Draft Guidance of 2006 and European Medicines Agency guideline; substrates for CYP1A2 (7-ethoxyresorufin), CYP2B6 (bupropion), CYP2D6 (dextromethorphan), and CYP3A4/5 (dextromethorphan and nifedipine) and inhibitors for CYP2B6 (sertraline, clopidogrel, and ticlopidine), CYP2C8 (quercetin), CYP2C19 (ticlopidine and nootkatone), and CYP3A4/5 (troleandomycin). Of course, limitations exist in the use of the substrate depletion method to identify potential CYP2C76 substrates (i.e., the inability to detect slowly metabolized compounds) because of the loss of protein activity and the difficulty of ascertaining true metabolism from experimental noise. As an example, bufuralol was not detected as a substrate for CYP2C76 in this study, but it has been identified to be a substrate by examining metabolite formation (Uno et al., 2011b). The intent of this work was not to identify all potential CYP2C76 substrates from our list of 89 compounds, and some low clearance compounds may have been overlooked; however, several structurally diverse compounds were newly identified in this study.
In this study, monkey CYP2C76 was shown to participate in bupropion hydroxylation and different drug oxidative metabolite(s) formation of nifedipine, presumably due to its relatively large substrate cavity with suitable substrate orientations. To compare the active site cavities for CYP2C76 and human CYP3A4, the CYP2C76 homology model was built from human CYP2C9 crystal structure. With binding of nifedipine as a ligand, the cavity size of CYP2C76 was comparable to human CYP3A4 (data not shown). The ability of CYP2C76 to metabolize a wide variety of diverse structures from relatively small molecules (e.g., nootkatone) to larger molecules (e.g., troleandomycin) may be attributed to its wide binding pocket of the active site with suitable substrate orientations. Human CYP3A4 has been reported to have a hydrophobic cluster consisting of Phe-213, Phe-214, Phe-219, and Phe-220 in terms of ligand-binding for catalytic function (Ekroos and Sjögren, 2006). Monkey CYP2C76 also had Phe-201, Phe-205, and Phe-294 over the heme around the active site area, similar to human CYP3A4 (data not shown). The total interaction energies (U value) of nifedipine with modeled CYP2C76 enzyme were −36.1 kcal/mol (Fig. 4), which was comparable to those with modeled monkey CYP3A4 (−42.9 kcal/mol) and human CYP3A4 (−59.9 kcal/mol). These structural similarities between human CYP3A4 and monkey CYP2C76 might be one of the determinant factors for CYP2C76 to be involved in nifedipine oxidation being flexible to accommodate different substrate orientations.
Although most of the CYP2C76 substrates also disappeared in the presence of other monkey CYP2Cs (CYP2C9 and CYP2C19), bupropion and nifedipine showed high selectivity to CYP2C76 (Fig. 1B). Bupropion is oxidized by CYP2B6 in humans and by CYP2B1 in rats and converted to hydroxybupropion whose molecular weight is 16 Da larger than the parent compound (Coles and Kharasch, 2007; Yeniceli et al., 2011; Pekthong et al., 2012). In monkeys, bupropion was predominantly metabolized by CYP2C76 and CYP2B6 (Table 2), whereas, nifedipine is dehydrogenated by CYP3A4 in humans and by CYP2C11 in rats, and converted to a pyridine form whose molecular weight is 2 Da smaller than nifedipine (Guengerich et al., 1986; Shimada et al., 1997; Chovan et al., 2007). In monkeys, nifedipine was predominantly metabolized by CYP3A4 and CYP2C76 (Table 2). These differences of P450 isoforms may result in species differences in pharmacokinetic profiles of some drugs.
An important result was that CYP2C76 formed a unique nifedipine metabolite whose molecular weight was 16 Da larger than the parent compound, suggesting the addition of a single oxygen atom. This metabolite was not produced by monkey CYP2C9, CYP2C19, or CYP3A4 (Fig. 2A) tested in the present study. Therefore, formation of this unique nifedipine metabolite could be a selective marker reaction of CYP2C76, although more detailed experiments with other monkey P450 isoforms or liver microsomal fractions are needed in the future. To our knowledge, this unique nifedipine metabolite has not been reported in humans. The lines of results indicated that CYP2C76 was able to produce some monkey-specific metabolite(s) of new chemical entities in cynomolgus monkeys. It would be beneficial to consider the presence of CYP2C76 when cynomolgus monkeys are used in preclinical studies. For example, in case some lethal toxicity or other severe problems were observed in monkeys, which were accounted for by CYP2C76 function, one would not have to withdraw the compound from development.
Monkeys are widely used in preclinical studies to predict bioavailability (BA) in humans; however, they occasionally show a poorer BA than humans (Chiou and Buehler, 2002; Takahashi et al., 2009; Akabane et al., 2010). Among the reported poor-BA compounds were amitriptyline, nifedipine, and imipramine, which were found as CYP2C76 substrates in this study (Fig. 1). BA is a product of the fraction absorbed from the gastrointestinal tract (Fa*Fg) and fraction escaping hepatic elimination. In the case of amitriptyline, low BA can be attributed to first-pass intestinal metabolism, resulting in low Fa*Fg (Akabane et al., 2010). Lately, P450 of small intestine was quantified in cynomolgus monkeys by immunoblotting; the content of CYP3A was most abundant (about 80% of immunoquantified total P450 content), similar to humans (Paine et al., 2006; Uehara et al., 2014). Because CYP2C76 was not detected in monkey small intestine (Uehara et al., 2014), CYP2C76 seems to be irrelevant to low Fa*Fg of amitriptyline. On the other hand, nifedipine and imipramine were reported to show low hepatic elimination (Takahashi et al., 2009). Considering that CYP2C76 is expressed in the liver of cynomolgus monkeys (4%), albeit at lower content than CYP3A4 (26%) (Uehara et al., 2011), CYP2C76 might contribute to high hepatic extraction of nifedipine and imipramine.
In conclusion, 19 structurally diverse substrates for CYP2C76 were identified among the 89 substrates evaluated. CYP2C76 has the ability to metabolize a wide variety of diverse structures. Among the newly identified substrates, CYP2C76 formed the unique metabolite of nifedipine, which is possibly a selective marker reaction of CYP2C76. The results revealed that CYP2C76 had relatively wide substrate specificity and the ability to form unique metabolite, which might result in species differences in drug metabolism between monkeys and humans. Our findings about substrate specificity of CYP2C76 should help to gain a better understanding of drug metabolism in monkeys and a better interpretation of preclinical data.
Acknowledgments
The authors thank Lance Bell for writing assistance.
Authorship Contributions
Participated in research design: Uno, Yamazaki.
Conducted experiments: Hosaka, Murayama. Uehara, Fujino.
Contributed new reagents or analytic tools: Satsukawa.
Performed data analysis: Hosaka, Shimizu, Iwasaki, Iwano, Uno.
Wrote or contributed to the writing of the manuscript: Hosaka, Iwano, Uno, Yamazaki.
Footnotes
- Received September 19, 2014.
- Accepted October 15, 2014.
Abbreviations
- BA
- bioavailability
- CLint
- intrinsic clearance
- Fa*Fg
- fraction absorbed from the gastrointestinal tract
- LC
- liquid chromatography
- MS/MS
- tandem mass spectrometry
- P450
- cytochrome P450
- Copyright © 2014 by The American Society for Pharmacology and Experimental Therapeutics