Metabolism of P450 Probe Substrates by Cynomolgus Monkey CYP2C76

Authors


Author for correspondence: Yasuhiro Uno, Pharmacokinetics and Bioanalysis Center, Shin Nippon Biomedical Laboratories, Ltd., 16-1 Minami Akasaka, Kainan, Wakayama 642-0017, Japan (fax +81 73 483 7377, e-mail uno-yasuhiro@snbl.co.jp).

Cytochrome P450 (P450 or CYP) is a gene family consisting of 57 functional genes and 58 pseudogenes in human beings [1]. More than 20 P450 isoforms have been identified in cynomolgus monkey [2], a species frequently used in drug metabolism studies because of its evolutionary closeness and physiological similarities to human beings. In this study, all the cynomolgus P450s are named according to the P450 Nomenclature Committee [2]. Most of these cynomolgus P450s are highly identical to human P450s and generally show substrate specificity similar to human P450s in the same subfamily [2]. Among these P450s, CYP2C76 has the lowest sequence identity to any human P450, approximately 80% identity to human CYP2Cs. CYP2C76 protein was detected in liver by immunoblotting using anti-CYP2C76 antibody and metabolized tolbutamide and testosterone [3]. However, CYP2C76 is located at the end of the CYP2C gene cluster, which corresponds to the intergenic region in the human genome [3], suggesting that a CYP2C76 orthologue is not present in human beings.

The absence of a CYP2C76 orthologue in human beings raised the possibility that CYP2C76 was involved in species differences in drug metabolism between cynomolgus monkey and human beings. Indeed, the metabolic assays of pitavastatin metabolism (as a model case) showed that cynomolgus CYP2C76 was responsible for formation of a metabolite that is not formed in human liver and for the reaction that is mediated by CYP3A in human beings and rat [4]. Therefore, CYP2C76 is at least partly responsible for species differences in drug metabolism between monkeys and human beings. Importantly, the fact that CYP2C76 metabolizes non-CYP2C substrates in human beings indicates that the metabolic properties of CYP2C76 need to be assessed using non-CYP2C substrates, as well as CYP2C substrates.

In this study, the metabolic activity of CYP2C76, along with other P450s, was assessed by analysing 7-ethoxyresorufin O-deethylation (EROD), coumarin 7-hydroxylation, testosterone 16β-hydroxylation, bufuralol 1′-hydroxylation and midazolam 1′-hydroxylation, the typical reactions of human CYP1A, CYP2A, CYP2B, CYP2D and CYP3A, respectively. Kinetic analysis was also carried out for reactions catalysed by CYP2C76.

Materials and Methods

Materials.  Pooled hepatic microsomes from human beings and cynomolgus monkeys were purchased from XenoTech LLC (Lenexa, KA, USA) and BD Gentest (Woburn, MA, USA), respectively. Bufuralol, coumarin, 7-ethoxyresorufin, 7-hydroxycoumarin, testosterone and midazolam were purchased from Wako Pure Chemical Industries, Ltd., (Osaka, Japan), 14C-testosterone was purchased from Amersham Biosciences UK, Ltd. (Buckinghamshire, UK), and all other reagents were from Sigma-Aldrich (St. Louis, MO, USA) unless otherwise specified.

Expression and preparation of cynomolgus P450 proteins.  For cynomolgus P450 metabolic assays, expression plasmids were prepared using a pCW vector that contained human NADPH-P450 reductase cDNA. Protein expression was carried out in Escherichia coli; membranes were prepared, and P450 protein and reductase in each membrane were quantified, as described previously [3,5–7]. These P450s include CYP1A1, CYP1A2, CYP2A23, CYP2A24, CYP2B6, CYP2C76, CYP2D17, CYP2D44, CYP3A4, CYP3A5, CYP4A11, CYP4F2, CYP4F3v2 and CYP4F11.

Metabolic assays.  The drug-metabolizing ability of CYP2C76 was assessed by quantifying EROD, coumarin 7-hydroxylation, testosterone 16β-hydroxylation, bufuralol 1′-hydroxylation and midazolam 1′-hydroxylation. Testosterone 16β-hydroxylation was analysed as described previously [6]. EROD, coumarin 7-hydroxylation, bufuralol 1′-hydroxylation and midazolam 1′-hydroxylation were analysed as follows. Briefly, typical incubations (0.5 ml) contained recombinant protein (100 pmol/ml) or liver microsomes (97.6 pmol P450/ml), an NADPH-generating system (1.55 mM NADP+, 3.3 mM glucose 6-phosphate and 0.4 unit/ml glucose 6-phosphate dehydrogenase) and substrate in 100 mM potassium phosphate buffer (pH 7.4). The final concentrations were 2.4, 9.0, 20 and 24 μM for 7-ethoxyresorufin, coumarin, bufuralol and midazolam, respectively. Reactions were incubated at 37°C for 30 min. and terminated by addition of 0.25 ml acetonitrile. Incubates were centrifuged at 7500 × g for 10 min., and supernatants were analysed by high-performance liquid chromatography (HPLC; Shimadzu, Kyoto, Japan)–tandem mass spectrometry (LC/MS/MS; Applied Biosystems, Foster City, CA, USA). The HPLC and LC/MS/MS systems consisted of two pumps (LC-10A; Shimadzu), an autosampler (SIL-10A; Shimadzu), column oven (CTO-10A; Shimadzu), UV detector (SPD-10A; Shimadzu) and MS/MS detector (API4000; Applied Biosystems). The mobile phases for analysis of resorufin, 7-hydroxycoumarin, 1′-hydroxybufuralol and 1′-hydroxymidazolam were as follows: A: formic acid/water (0.1/99.9), B: formic acid/acetonitrile (0.1/99.9), and a linear gradient was used from 0 to 10 min. (20% B – 80% B). A Hydrosphere C18, 15 cm × 2.0 mm (YMC, Kyoto, Japan), analytical column was used, at ambient temperature. The detection of metabolites was performed using mass spectrometry. Q1/Q3 (m/z) of the metabolites was 214/186 for resorufin, 163/107 for 7-hydroxycoumarin, 278/141 for 1′-hydroxybufuralol and 342/203 for 1′-hydroxymidazolam. For EROD and bufuralol 1′-hydroxylation, kinetic parameters were determined by non-linear regression analysis to estimate the apparent Km and Vmax. Mixtures were incubated in duplicate for 10 min. using five concentrations of the substrates, 7-ethoxyresorufin and bufuralol, at 1–50 μM and 1–400 μM, respectively.

Results

To determine the involvement of CYP2C76 in the reactions mediated by non-CYP2C isoforms of human beings, metabolic assays were conducted using recombinant protein heterologously expressed in E. coli, including EROD, coumarin 7-hydroxylation, testosterone 16β-hydroxylation, bufuralol 1′-hydroxylation and midazolam 1′-hydroxylation, reactions typically catalysed, respectively, by human CYP1A, CYP2A, CYP2B, CYP2D, CYP2E and CYP3A isoforms. For comparison, cynomolgus CYP1A1, CYP1A2, CYP2A23, CYP2A24, CYP2B6, CYP2C76, CYP3A4, CYP3A5, CYP4A11, CYP4F2, CYP4F3v2 and CYP4F11 recombinant proteins were also analysed. The results showed that CYP2C76 was involved in EROD and bufuralol 1′-hydroxylation, which CYP1A and CYP2D isoforms typically catalyse, respectively (table 1). CYP2C76 also catalysed midazolam 1′-hydroxylation but at a much lower rate than cynomolgus CYP3A4 and CYP3A5, which efficiently catalysed this reaction (table 1).

Table 1. 
Metabolism of non-CYP2C substrates of human beings by cynomolgus CYP2C76.
 Ethoxyresorufin O-deethylationCoumarinTestosteroneBufuralolMidazolam
7-hydroxylation16β-hydroxylation1′-hydroxylation1′-hydroxylation
  1. nmol/min/nmol recombinant P450 or total P450

  2. Data from Uno et al. [5]1, [6]2 and Yamazaki et al. [19]3.

  3. Metabolic assays were carried out using substrate (2.4 μM 7-ethoxyresorufin, 9.0 μM coumarin, 100 μM testosterone, 20 μM bufuralol or 24 μM midazolam) and protein (5 pmole P450/ml for CYP2B6, 100 pmole P450/ml for other P450s or 97.6 pmole P450/ml for human liver microsomes) as described in Materials and Methods. Values are based on duplicate determinations.

CYP1A10.882<0.0233<0.1000.396<0.00467
CYP1A20.561<0.0233<0.1000.08010.00565
CYP2A230.003430.8831<0.1000.004190.0213
CYP2A24<0.001172.971<0.100<0.00467<0.00467
CYP2B6<0.00117<0.023320.92<0.00467<0.00467
CYP2C760.147<0.0233<0.1000.8300.00461
CYP2E10.0102<0.0233<0.100<0.00467<0.00467
CYP3A40.002940.0251<0.1000.05981.74
CYP3A50.004670.201<0.1000.2832.87
CYP4A11<0.00117<0.0233<0.100<0.00467<0.00467
CYP4F2<0.00117<0.0233<0.100<0.00467<0.00467
CYP4F3v2<0.00117<0.0233<0.100<0.00467<0.00467
CYP4F110.00169<0.0233<0.1000.115<0.00467
Human liver0.06361.740.00930.1311.37

Other P450s substantially catalysed reactions typical of the corresponding human P450s: EROD by CYP1A1/2, coumarin 7-hydroxylation by CYP2A23/24, testosterone 16β-hydroxylation by CYP2B6 and midazolam 1′-hydroxylation by CYP3A4/5 (table 1). Interestingly, some P450s apparently catalysed reactions, albeit weakly, other than those typically catalysed by the corresponding human P450s: EROD by CYP2A23, CYP2E1, CYP3A4, CYP3A5 and CYP4F11; coumarin 7-hydroxylation by CYP3A4 and CYP3A5; bufuralol 1′-hydroxylation by CYP1A1, CYP1A2, CYP2A23, CYP3A4, CYP3A5 and CYP4F11; and midazolam 1′-hydroxylation by CYP1A2 and CYP2A23 (table 1).

CYP2C76 substantially catalysed EROD and bufuralol 1′-hydroxylation; therefore, kinetic parameters of CYP2C76 were determined for these reactions. To assess the contribution of CYP2C76 to these reactions, EROD and bufuralol 1′-hydroxylation catalysed by CYP1A1/2 and CYP2D17/44, respectively, were also analysed. The analysis showed that the apparent Km was 1.2, 2.4 and 8.9 μM, and the estimated intrinsic clearance was 3.0, 1.2 and 0.13 ml/min/nmol P450, respectively, for CYP1A1, CYP1A2 and CYP2C76 catalysed EROD (table 2). For CYP2C76, CYP2D17 and CYP2D44 catalysed bufuralol 1′-hydroxylation, the apparent Km was 65, 32 and 21 μM, and the estimated intrinsic clearance was 0.08, 0.38 and 0.34 ml/min/nmol P450, respectively (table 2). The results indicate the contribution of CYP2C76 to EROD and bufuralol 1′-hydroxylation, albeit to a lower rate at a molecular basis, comparatively, than CYP1As and CYP2Ds, respectively.

Table 2. 
Kinetic parameters of CYP2C76-catalysed ethoxyresorufin O-deethylation (EROD) and bufuralol 1′-hydroxylation.
 ERODBufuralol 1′-hydroxylation
KmVmaxCLint (Vmax/Km)KmVmaxCLint (Vmax/Km)
μMnmol/min/nmol recombinant or total P450ml/min/nmol recombinant or total P450μMnmol/min/nmol recombinant or total P450ml/min/nmol recombinant or total P450
  1. N.D., not determined.

  2. Metabolic assays were performed using substrate (7-ethoxyresorufin or bufuralol) and protein (CYP2C76, CYP1A1/2, CYP2D17/44 or cynomolgus monkey liver microsomes) as described in Materials and Methods. Kinetic parameters were determined by non-linear regression analysis (mean ± SE). Values are based on duplicate determinations. Molar ratios of reductase to CYP2C76, 1A1, 1A2, 2D17 and 2D44 were 1.0, 0.42, 0.36, 2.2 and 2.3, respectively.

2C768.9 ± 5.41.2 ± 0.20.1365 ± 105.2 ± 0.30.08
1A11.2 ± 0.43.6 ± 0.23.0N.D.N.D.N.D.
1A22.4 ± 0.72.8 ± 0.21.2N.D.N.D.N.D.
2D17N.D.N.D.N.D.32 ± 712 ± 2.00.38
2D44N.D.N.D.N.D.21 ± 87.2 ± 1.00.34
Liver microsomes4.9 ± 1.00.073 ± 0.0070.2412 ± 20.53 ± 0.020.04

Discussion

Cynomolgus CYP2C76, which is not orthologous to any human P450, is responsible for species differences in the metabolism of pitavastatin [4] and possibly other drugs. However, the metabolic properties of CYP2C76 have not been investigated using non-CYP2C substrates. In this study, metabolic assays of EROD, coumarin 7-hydroxylation, testosterone 16β-hydroxylation, bufuralol 1′-hydroxylation and midazolam 1′-hydroxylation were carried out for CYP2C76, along with CYP1A1, CYP1A2, CYP2A23, CYP2A24, CYP2B6, CYP3A4, CYP3A5, CYP4A11, CYP4F2, CYP4F3v2 and CYP4F11. All these P450s, except for CYP2C76, substantially catalysed the reactions typical of the corresponding human P450s (table 1).

CYP2C76 catalysed bufuralol 1′-hydroxylation (table 1), a reaction typically catalysed by human CYP2D6. This and another CYP2D-dependent reaction, dextromethorphan O-demethylation, occur to a greater extent in cynomolgus monkey liver than in human liver [8,9], for which cynomolgus CYP2C76 could be partly responsible. However, the apparent Km was higher and the estimated intrinsic clearance was lower for CYP2C76-catalysed bufuralol 1′-hydroxylation as compared with CYP2D17 and CYP2D44 (table 2), which also catalyse this reaction in cynomolgus monkey [7,10]. This suggests that CYP2C76 catalyses bufuralol 1′-hydroxylation less efficiently, based on the molecular activity. Moreover, CYP2D protein appears to be more abundantly expressed in cynomolgus monkey liver than in human liver [9]. Therefore, higher bufuralol 1′-hydroxylation in cynomolgus monkey liver than in human liver is accounted for by the higher efficiency and/or more abundant hepatic expression of cynomolgus CYP2Ds, in addition to CYP2C76 involvement.

CYP2C76 also catalysed EROD (table 1), which CYP1A1 and CYP1A2 catalyse efficiently in cynomolgus monkey [11]. Hepatic EROD is generally greater in cynomolgus monkey than in human beings [8,9]. This difference might be partly accounted for by CYP2C76, although the apparent Km was higher and the estimated intrinsic clearance was lower for CYP2C76-catalysed EROD, as compared with CYP1A1 and CYP1A2 (table 2). The apparent Km for EROD and the estimated intrinsic clearance of 7-ethoxyresorufin was similar in cynomolgus monkey liver and human liver, but EROD was less inhibited by anti-CYP1A antibodies in cynomolgus monkey liver microsomes than in human liver microsomes [12]. Moreover, CYP1A expression in cynomolgus monkey liver is generally much lower than in human liver; CYP1A expression is minimal or undetectable in cynomolgus monkey liver by immunoblotting, unless induced by inducers such as β-naphthoflavone, 3-methylcholanthrene and 2,3,7,8-tetrachlorodibenzo-p-dioxin [13–15]. Therefore, a higher level of EROD in cynomolgus monkey liver than in human liver might be partly accounted for by non-CYP1A isoforms, such as CYP2C76 and CYP1D1. CYP1D1 is a newly identified cynomolgus monkey P450 that catalyses EROD [11]. Considering that CYP1As can be induced by diet, such confounding factors might also play a role in this difference in EROD.

Cynomolgus CYP2C76 substantially metabolized non-CYP2C substrates, 7-ethoxyresorufin and bufuralol (table 1). Similarly, in dogs, CYP2C21 metabolizes phenacetin, temazepam, bufuralol and midazolam [16], which CYP1A, CYP2B, CYP2D and CYP3A, respectively, metabolize in human beings. The difference in CYP2C substrate specificity is even noted in primate: marmoset CYP2C8 metabolizes tolbutamide (human CYP2C9 substrate) but not paclitaxel (human CYP2C8 substrate) [17]. In contrast, cynomolgus CYP2C8 metabolizes paclitaxel [3]. The substrate selectivity of CYP2C8 partly reflects the evolutionary closeness of cynomolgus monkey to human being, as compared to marmoset.

In this study, other cynomolgus P450s also catalysed reactions other than the typical reactions of their corresponding human P450s. For example, cynomolgus CYP3A5 metabolized bufuralol (table 1) as has been shown previously [18], contributing to bufuralol 1′-hydroxylation, together with CYP2D17/44 and CYP2C76. However, most cynomolgus P450s metabolized the probe substrates of their corresponding human P450s (table 1). Orthologous P450s are relatively well conserved between macaque (such as cynomolgus monkey) and human beings, based on genome analysis [2]. CYP2C76 metabolized non-CYP2C substrates 7-ethoxyresorufin and bufuralol, but less efficiently than CYP1A and CYP2D, respectively, at the same molar enzyme levels. These results partly suggest an overall similarity in P450 characteristics between cynomolgus monkey and human beings. Information on substrate specificity of cynomolgus P450s presented in this study is expected to help better understand drug metabolism in cynomolgus monkey.

Acknowledgements

We greatly thank Mr. Masahiro Utoh, Dr. Koichiro Fukuzaki and Dr. Ryoichi Nagata for their support for this work. We also appreciate Mr. Patrick Gray for reviewing the paper.

Ancillary