Functional expression and comparative characterization of four feline P450 cytochromes using fluorescent substrates

Article abstract of DOI:10.1080/00498254.2016.1257172

1.?Cytochrome P450s (CYP) are a major group of metabolizing enzymes for xenobiotics in humans and other mammals. The properties of CYP isoforms in the domestic cat, an obligate carnivore, are largely unknown at present. In this study, we studied relative expression in tissues and enzymatic properties of nine significant feline CYP isoforms. 2.?CYP2E2 transcript was most abundant in the feline liver, followed by CYP2A13 and 2E1. Transcripts of CYP3A131, 1A2 and 1A1 were also present in the liver, while CYP2D6 and 3A132 were only slightly expressed. CYP3A131 was a major transcript in the small intestine. 3.?Four major CYP isoforms in the feline liver and small intestine (CYP1A2, CYP2A13, CYP2E2 and CYP3A131) were heterologously expressed in Escherichia coli to generate functional monooxygenase systems. We carried out screenings of 17 test compounds known to be inhibitors of CYP isoforms in other mammals as well as two anticancer drugs to assess the activity modulation of feline CYP isoforms using fluorogenic substrates. These CYP isoforms showed similar selectivity to counterparts in other mammals against inhibitors as a whole but with many exceptions. 4.?The present study suggests the usefulness of the feline CYP recombinant system to obtain chemical affinity information and possible drug interactions in CYP metabolism of domestic cats.

Full text of DOI:10.1080/00498254.2016.1257172

Xenobiotica
the fate of foreign compounds in biological systems
ISSN: 0049-8254 (Print) 1366-5928 (Online) Journal homepage: http://www.tandfonline.com/loi/ixen20
FUNCTIONAL EXPRESSION AND COMPARATIVE
CHARACTERIZATION OF FOUR FELINE P450
CYTOCHROMES USING FLUORESCENT SUBSTRATES
Gaku Okamatsu, Kei Kawakami, Tetsuya Komatsu, Takio Kitazawa, Yasuhiro
Uno & Hiroki Teraoka
To cite this article: Gaku Okamatsu, Kei Kawakami, Tetsuya Komatsu, Takio Kitazawa,
Yasuhiro Uno & Hiroki Teraoka (2016): FUNCTIONAL EXPRESSION AND COMPARATIVE
CHARACTERIZATION OF FOUR FELINE P450 CYTOCHROMES USING FLUORESCENT SUBSTRATES,
Xenobiotica, DOI: 10.1080/00498254.2016.1257172
To link to this article: http://dx.doi.org/10.1080/00498254.2016.1257172
Accepted author version posted online: 14
Nov 2016.
Submit your article to this journal
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=ixen20
Download by: [Athabasca University]
Date: 15 November 2016, At: 06:51

FUNCTIONAL EXPRESSION AND COMPARATIVE
CHARACTERIZATION OF FOUR FELINE P450 CYTOCHROMES
USING FLUORESCENT SUBSTRATES
Gaku Okamatsu a, Kei Kawakami a, Tetsuya Komatsu a, Takio Kitazawa a, Yasuhiro
Uno b and Hiroki Teraoka a,*
a School of Veterinary Medicine, Rakuno Gakuen University, Ebetsu, Hokkaido 069-
8501, Japan
b Pharmacokinetics and Bioanalysis Center, Shin Nippon Biomedical Laboratories,
Ltd., Kainan, Wakayama 642-0017, Japan
Running title: Comparison of feline P450 cytochromes
MS. has 37 pages, 3 figures and 4 tables
*Corresponding author at School of Veterinary Medicine, Rakuno Gakuen University,
Ebetsu, Hokkaido 069-8501, Japan.
E-mail address: hteraoka@rakuno.ac.jp (H. Teraoka).

ABSTRACT
Cytochrome P450s (CYP) are a major group of metabolizing enzymes for xenobiotics in
humans and other mammals. The properties of CYP isoforms in the domestic cat, an
obligate carnivore, are largely unknown at present. In this study, we studied relative
expression in tissues and enzymatic properties of nine significant feline CYP isoforms.
CYP2E2 transcript was most abundant in the feline liver, followed by CYP2A13 and
2E1. Transcripts of CYP3A131, 1A2 and 1A1 were also present in the liver, while
CYP2D6 and 3A132 were only slightly expressed. CYP3A131 was a major transcript in
the small intestine. Four major CYP isoforms in the feline liver and small intestine
(CYP1A2, CYP2A13, CYP2E2 and CYP3A131) were heterologously expressed in
Escherichia coli to generate functional monooxygenase systems. We carried out
screenings of 17 test compounds known to be inhibitors of CYP isoforms in other
mammals as well as two anticancer drugs to assess activity modulation of feline CYP
isoforms using fluorogenic substrates. These CYP isoforms showed similar selectivity
to counterparts in other mammals against inhibitors as a whole but with many
exceptions. The present study suggests the usefulness of the feline CYP recombinant
system to obtain chemical affinity information and possible drug interactions in CYP
metabolism of domestic cats.
Keywords: carnivore; cytochrome P450; domestic cats; drug interaction; felid;
xenobiotics

1. Introduction
Drugs approved for use in humans can be prescribed for off-label use in veterinary
medicine at the discretion of a veterinarian even in developed countries (Animal
Medicinal Drug Use Clarification Act of 1994 in USA). Thus, drugs approved for
humans are often also used in companion animals. However, knowledge on the
biotransformation of drugs and other xenobiotics in cats is limited. The situation for cats
is quite different from that of dogs. Dogs are important experimental animals for the
purpose of development of human drugs and much more detailed information on drug
metabolism as well as pharmacological and toxicological effects are obtained,
especially in drug companies (Bailey et al., 2015). Domestic cats are the most common
companion animal and an obligate carnivore, which is dependent on a supply of at least
some animal-derived materials in its diet, unlike the dog, which has a relatively
developed carbohydrate digestive system (MacDonald et al., 1984). All felids,
including domestic cats, must eat meat or tissue of animals. It is well known that cats
are very sensitive to some xenobiotics, such as acetaminophen (Court, 2013). This is
partly explained by a deficiency of a major part of glucuronidation. For example, it is
believed that cats are more susceptible than dogs to acetaminophen toxicity due to lack
of UGT1A6 and UGT1A9 (Court and Greenblatt, 2000). Apart from these, significant
interspecies variations in drug metabolic activity have been recognized (Nebbia et al.,
2003; Baririan et al., 2006). These findings strongly suggest the insufficiency of
extrapolations solely based on body weight or body surface from dog and human to cat
in determination of drug treatments (Court, 2013).
Cytochrome P450 (CYP) monooxygenase is most important for metabolism of

xenobiotics to polar substances for excretion and to active intermediates, causing
toxicological responses in some cases. Among the numerous members of CYP
superfamily, CYP1A, 2B, 2C, 2D, 2E, and 3A families are particularly important for
xenobiotic metabolism (Lewis and Ito, 2008). However, knowledges on feline CYP are
limited. Until now, only CYP1A1/2 (Tanaka et al., 2006), 2A13 (Okamatsu et al.,
2015), 2D6 (Komatsu et al., 2010), 2E1/2 (Tanaka et al., 2005), and 3A131/132 (Honda
et al., 2011) have been reported to be present in the liver of cats. Characteristically,
CYP2B6 transcripts and protein were present in the lung and small intestine, but not in
the liver of cats (Okamatsu et al., 2016).
Using liver microsomes, CYP1A-, 2C-, 2D- and 3A-dependent enzymatic activities
of CYP families of domestic cats have been previously characterized (Chauret et al.,
1997; Pearce et al., 1992; Shah et al. 2007). Enzymatic activities of six CYP families of
cats were compared with those of dogs with fluorescent substrates specific for each
CYP isoform (van Beusekom et al., 2010). However; in previous studies with cat liver
microsomes, the substrate specific to the human orthologues of the enzymes have been
used despite the potential for interspecies differences in CYP activities.
The most straightforward method of defining substrate specificity is characterization
of CYP isoforms that are prepared by heterologous expression, because CYP isoforms
have overlapping metabolic activity for each substrate. E. coli is one of the most
extensively utilized host in the production of recombinant CYP enzymes. It was
reported that the same patterns of apparent allosteric effect are observed with proteins
derived from either the liver or from E. coli for human CYP3A4 (Ueng et al., 1995).
Appropriate N-terminal modifications do not appear to affect enzyme activity, and
sometimes introduce desirable characteristics that are likely to facilitate attempts to

crystallize the catalytic domains of microsomal P450s (Yun et al., 2006). Canine
CYP1A1, 2B11, 2C21, 2C41, 2D15, 3A12, and 3A26 with canine cytochrome P450
reductase has been heterologously expressed in E. coli to study basic metabolic
activities (Locuson et al., 2009).
Recently, we prepared recombinant of feline CYP2A13 (Okamatsu et al., 2015) and
CYP2B6 (Okamatsu et al., 2016) in Escherichia coli (E. coli). Although we showed
metabolic activity similar to those of human and canine CYP2As for coumarin, 7-
ethoxycoumarin, and nicotine by feline CYP2A13, additional properties including
interaction with other xenobiotics are largely unknown (Sasaki and Shimoda, 2015).
Recombinant CYP isoforms of mice and human heterologously expressed in E. coli has
been extensively studied for the metabolic activity and inhibition by xenobiotics using
fluorogenic substrates suitable for large-scale screening (McLaughlin et al., 2008;
Kajbaf et al., 2011). We previously confirmed metabolic activity of feline CYP2A13
and CYP2B6 using resorufin- and coumarin-based CYP substrates that are the most
extensively used substrates (Okamatsu et al., 2015, 2016).
It is important to know that the property of CYP isoforms from domestic cat that
shows very different feeding behavior as a strickt carnivore and lacks some major UGT
isoform. In this study, we characterized relative expression of nine feline CYPs in the
liver and small intestine. Then, we heterologously coexpressed four feline P450s with
human P450 oxidoreductase (POR) in E. coli, thus generating functional CYP systems
in vitro for enzyme characterization. For each CYP isoform, we identified an adequate
fluorogenic probe substrate with the apparent kinetic parameters. These probe substrates
were then used to determine the IC50 of 19 test compounds including inhibitors specific
for human and canine CYP isoforms and chemotherapeutic agents for each CYP,

allowing us to compare inhibitory potencies as well as affinities to various xenobiotics
for each CYP isoform.
2. Materials and Methods
2.1. Chemicals
The 7-benzyloxy-4-trifluoromethylcoumarin (BFC), 7-hydroxy-4-(trifluoromethyl)
coumarin (HFC) and 7-ethoxy-4-(trifluoromethyl)coumarin (EFC) were purchased from
BD Gentest (Woburn, MA, USA) and 7-ethoxyresorufin, 7-methoxyresorufin, and 7-
pentoxyresorufina were obtained from Molecular Probes Europe BV (Leiden, The
Netherlands). Coumarin, 7-hydroxycoumarin, resorufin, cyclophosphamide,
erythromycin and tamoxifen were purchased from Sigma-Aldrich (St. Louis, MO, USA)
and 7-benzoxyresorufin was purchased from Roche (Mannheim, Germany).
Atipamezole, medetomidine, methoxsalen, miconazole and ticlopidine were obtained
from Wako Pure Chemical (Osaka, Japan). 7-ethoxycoumarin and the other reagents use
as test compounds were purchased from Tokyo Kasei (Tokyo, Japan). All other reagents
used in this study were of analytical grade.

2.2. Quantitative real-time RT-PCR
Tissues of mongrel cats of both sexes (short-haired type; 5 male cats and 6 female
cats) collected in our previous study (Okamatsu et al., 2015) were also used in this study.
They are all adult (3 years old). Total RNA was extracted from the liver and small
intestine of domestic cats using a conventional acid guanidine-phenol-chloroform
method (Trizol; Invitrogen, Carlsbad, CA, USA) (Komatsu et al., 2010). We used
middle part of small intestine, because RNA degradation of feline small intestine
(particularly for upper part) was very rapid in the preliminary experiment. We
synthesized cDNA from total RNA using oligo dT primer and reverse transcriptase
(QuantiTecht, Qiagen, Hilden, Germany). Quantitative real-time RT-PCR (qRT-PCR)
analysis was performed with a real-time PCR detector (Chromo4; Bio-Rad, Hercules,
CA, USA) using an SYBR Green-based kit (Thunderbird qPCR Mix, Toyobo, Osaka,
Japan). The primer pairs used for CYP1A1, 1A2, 2E1, and 2E2 are listed in Table 1.
The sequences of primers for CYP2B6 (Okamatsu et al., 2016), 2A13 (Okamatsu et al.,
2015), 2D6 (Komatsu et al., 2010), and 3A131 and 3A132 (Honda et al., 2011) have
been previously reported. Standard curves were generated with serial dilutions of PCR
products. The concentration of PCR product was quantified with a microchip
electrophoresis system (MultiNA, Shimadzu, Kyoto, Japan). The primer sets used for

qRT-PCR were confirmed to produce a single peak in the melting curve as well as a
single band by agarose gel electrophoresis. Each CYP expression was normalized per
μg total RNA used for reverse transcription.
2.3. Heterologous expression in E. coli
Membrane fractions of feline CYP2A13 with human oxidoreductase (POR) was
prepared according to our previous report (Okamatsu et al., 2015). Protein expression in
Escherichia coli (E. coli) using the generated expression plasmids, membrane
preparations, and measurement of CYP2A13 content were performed as described
previously (Okamatsu et al., 2015).
Heterologous expression of CYP3A131 was carried out according to Okamatsu et
al. (2015, 2016). This was based on the method originally developed by Pritchard et al.
(2006). Feline full length CYP3A131 was amplified by RT-PCR with primers (3 and 4
in Table 2) and cDNA derived from the liver. To enhance functional protein expression
in E. coli, bacterial ompA leader sequences were fused to the 5’-end of the complete
coding region of feline full length CYP3A131 by PCR with genomic DNA of E. coli
and the primers 5 and 10 (first-round PCR). After being purified by agarose gel

extraction, resultant PCR product was used as forward primer in the second-round PCR
with the CYP3A131 3’-end specific reverse primer 11 and full length feline CYP3A131
as a template. The amplified fragments were subcloned into T-vector (pTAC-2:
BioDynamics, Tokyo, Japan) and the plasmid was obtained by transfection to
competent cells (DH5α: Frontier Science, Ishikari, Japan) and purified with a plasmid
mini kit (FastGene Plasmid Mini Kit, Nippon Genetics). After digestion of the products
with Nde I and Xba I of the plasmid were purified by gel elecrophoresis, resultant insert
fragment was subcloned, using a DNA Ligation Kit (Mighty Mix, Takara), into
pCWOri+ vector (Barnes, 1996), in which the human POR cDNA had already been
accommodated (Iwata et al., 1998). The resultant construct was used for transformation
with competent cells and CYP3A131 bactosomes were harvested as described
previously (Okamatsu et al., 2015). POR activity of bactosomes were measured as
reported by Iwata et al. (1998).
Feline CYP1A2 bactosomes were prepared according to Iwata et al. (1998), who
replaced the second to seventh N-terminal amino acid regions with bovine CYP17α to
facilitate protein expression in E. coli. Feline full length cDNA of CYP1A2 was
amplified by RT-PCR with liver-derived cDNA and primers (1 and 2 in Table 2). The
PCR product was used as a template for the PCR with primers 6 and 7 to add an Nde I

site in N-terminus and Xba I site in C-terminus. The following procedures are the same
as those for feline CYP3A131.
Heterologous expression of feline CYP2E2 in E. coli was carried out using the
expression vector that contained fragment of modified mouse 2C29 N-terminals (Uno et
al., 2006), because we failed to obtain functional CYP2E2 by the method described for
CYP2A13 and CYP3A131. RT-PCR for the gene encoding mouse CYP2C29 was
carried out with cDNA derived from the mouse liver and primer sets (12 and 13 in
Table 2). The amplified fragments were added with EcoRI site in N-terminus and Sal I
site in C-terminus by the nested PCR with primers (14 and 15) and the PCR product of
the previous PCR as a template. The resultant PCR product was further used as a
template for the PCR with primers 16 and 15 to extend and add Nde I site in N-terminus.
Purified digests with Nde I and Sal I of the plasmid obtained were subcloned into
pCWOri+ vector (Modified pCWori+ vector). Otherwise, feline CYP2E2 insert with
EcoRI and Sal I sites were obtained with primers (8 and 9) and cat liver cDNA. Purified
digests of the two restriction enzymes were subcloned into Modified pCWori+ vector.
The following procedures are the same as described above.

2.4. Western blotting
Four feline recombinant CYPs (0.5 pmol except CYP1A2 at 0.26 pmol)and liver
microsomes (30 μg) were run on 10% SDS polyacrylamide gels and transferred to
cellulose nitrate filters (pore size 0.45μm) (Advantec, Tokyo, Japan) according to our
previous study (Okamatsu et al., 2015, 2016). The filters were immunoblotted with
primary antibody specific for human CYP protein, followed by incubation with
biotinylated rabbit anti-goat IgG antibody or biotinylated goat anti-rabbit IgG antibody
(Vectastatin ABC-AP KIT, Vector Labs, Burlingame, CA, USA). The primary
antibodies used were rabbit anti-human CYP1A2 antibody (Abgent, San Diego, CA,
USA) (1:50), rabbit anti-human CYP2E1 antibody (Abcam, Cambridge, UK) (1:100),
and rabbit anti-human CYP3A43 antibody (1:100) (Abgent). After the avidin-biotin
reaction, a specific band was visualized using a DCIP/NBT substrate kit (Vector Labs).
2.5. Enzyme activity assay
Metabolic activity of feline recombinant CYPs with coumarin was determined
spectrofluorometrically as described previously (Okamatsu et al., 2015). In brief,

recombinant feline CYP proteins (2-8 pmol/100μL) were incubated with coumarin (0-
5,000 μM as a final concentration) at 37 °C for 15 min in potassium phosphate buffer
(pH 7.4) after 5 min preincubation. The fluorescence of 7-HC (excitation 370 nm and
emission 450 nm) was measured with a fluorescence spectrophotometer (650-10S,
Hitachi, Tokyo, Japan). The same reaction condition was also used for determining
metabolism of BFC (0-1,000 µM) and MFC (0-5,000 µM) . Fluorescence of HFC
converted (420 nm for excitation and 535 nm for emission) was determined in a reaction
buffer without extraction after the reaction was stopped by addition of 900 μL ice-cold
potassium buffer. The detection limit of HFC (3σ) was 0.45 pmol.
Alkoxyresorufin-O-dealkylase activity was determined as previously described
(Okamatsu et al., 2016). Alkoxyresorufins (benzyloxyresorufin, ethoxyresorufin,
methoxyresorufin, and pentoxyresorufin) (0-75 µM))were incubated with recombinant
feline CYP2A (2–8 pmol/100 mL) in potassium phosphate buffer (pH 7.4) at 37 oC.
Resorufin formed by the reaction was measured by an excitation wavelength of 535 nm
and emission wavelength of 595 nm. The detection limit of resorufin (3σ) was 0.14
pmol.

2.6. Fluorometric enzyme inhibition assays
Effects of 19 compounds were studied using fluorescent substrates selected from
the previous study (McLaughlin et al., 2008). Basically, all incubations were done under
the conditions described in the previous section. After test compounds and substrates in
0.1 M potassium phosphate buffer (pH 7.4) were preincubated for 5 min at 37 oC,
reactions were carried out for 20 min after addition of NADPH (final 1 mM).
Concentrations of substrates used for enzyme inhibition assays were around the Km
values;100µM BFC for CYP1A2, 1.5 µM coumarin for CYP2A13, 300 µM 7-EC for
CYP2E2, and 300 µM BFC for CYP3A131. The concentration of test compound as
inhibitors were in the range of 0.05 nM – 10 µM (the lower limit ) to 500 – 2,000 µM
(upper limit), dependent on the effectiveness of compounds in the preliminary
experiments. IC50 values were calculated in Microsoft Excel as previously described
(Okamatsu et al., 2016).

2.7. Statistical methods
Results are presented as means±SE. Significance of differences among groups was
determined by one-way ANOVA followed by the Tukey-Kramer test (p < 0.05). For
linear regression, Eadie-Hofstee plots were generated and data were analyzed in
Microsoft Excel using least-squares linear regression to generate the best fit line.
3. Results
3.1. Tissue expression of CYP isoforms
We characterized the relative expression of the CYP isoforms that we identified
previously (CYP2A13, 2B6, 2D6, 3A131, 3A132) (Komatsu et al., 2010; Honda et al.,
2011; Okamatsu et al., 2015, 2016) and other CYP isoforms (1A1, 1A2, 2E1, 2E2)
(Tanaka et al., 2005, 2006) in the liver and small intestine (Fig. 1). Significant sex-
dependent difference were not recognized in the amount of transcripts of any isoform in
the liver and small intestine. Thus, all data were shown as the mean of all the animals
analyzed. In the liver, CYP2E2 transcripts were most abundant, followed by 2A13, 2E1,
3A131 = 1A2 = 1A1, 2D6, and CYP3A132. CYP2B6 was not detected in the liver,
which is in agreement with our previous work (Okamatsu et al., 2016). CYP3A131 was
a major CYP transcript in the small intestine, while there were few transcripts for other

CYPs except CYP2B6 and 1A1 with slight expressions (Fig. 1B).
3.2. Heterologous coexpression of four major CYP isoforms in E. coli.
We focused on CYP2E2, CYP2A13 and CYP3A131 due to their abundance in the
liver and small intestine. Additionally, we prepared CYP1A2 recombinant protein,
because CYP1A2 metabolizes a variety of arylamines and heterocyclic arylamines in
therapeutic drugs (Zhou et al., 2009).
The reduced-CO difference spectra with these recombinant CYP proteins showed
clear absorption peaks at around 450 nm (Fig. 2). We have already reported
recombinant protein of CYP2A13 and CYP2B6 (Okamatsu et al., 2015, 2016).
Bactosomes of four feline CYP isoforms prepared in this study contained 0.02-0.46
nmoles/mg protein of P450 (CYP1A2 0.03 ± 0.01, 2A13 0.26 ± 0.11, 2E2 0.04 ± 0.02,
3A131 0.03 ± 0.01: n = 3) with 56.6-143.9 nmol/min/mg protein as activities of POR
(CYP1A2 82.2 ± 10.8, 2A13 117.1 ± 13.8, 2E2 111.8 ± 11.7, 3A131 83.0 ± 13.3: n = 3).
These recombinant CYP proteins were detected by immunoblotting with each anti-
human CYP polyclonal antibodies (human CYP1A2, 2E1 and 3A43) (Fig. 3). These
antibodies reacted with feline CYP counterparts but not other feline CYPs including
CYP2A13 and CYP2B6 (Fig. 3). Immunoreactivities of CYP1A, CYP2E and CYP3A
were detected in the liver microsomes by these antibodies, irrespective of sex (Fig. 3).

3.3. Metabolic activities of recombinant CYPs
We characterized metabolic activities of four recombinant feline CYPs prepared in
the previous chapter, along with CYP2A13. For high-throughput screening, fluorogenic
substrates were selected by reference to substrates for human CYP isoforms (Donato et
al., 2004). None of the recombinant CYP isoforms studied showed O-dealkylase activity
of 7-pentoxy resorufin.
Feline CYP1A2 heterologously expressed in E. coli caused concentration-
dependent hydroxylation of 7-benzoxyresorufin (BR), 7-ethoxyresorufin (ER), and 7-
methoxyresorufin (MR). Coumarin and 7-ethoxycoumarin were not studied. Values of
Km and Vmax obtained from the dose-response curves of hydroxylation of some
substrates by CYP1A2 together with other CYPs are given in Table 3. CYP1A2 showed
a high-affinity hydroxylation with BR and ER compared to that with MR with lower
maximal activity (Table 3). CYP1A2 also hydroxylated 7-benzyloxy-4-
(trifluoromethyl)-coumarin (BFC) with lower affinity but higher Vmax.
Feline CYP2A13 showed significant hydroxylation activity with ER, coumarin,
and 7-ethoxycoumarin (7-EC) (Table 3). Among these substrates, coumarin showed the
highest affinity, while 7-EC showed the highest activity but lowest affinity.
Feline CYP2E2 showed significant metabolic activity with MR, coumarin, and 7-
EC (Table 3). However, hydroxylation of coumarin by CYP2E2 needed a very high
concentration of substrates and a Vmax value was not obtained even at 5 mM. Only
very low activity was recognized with MR.
Feline CYP3A131 hydroxylated BFC into hydroxycoumarin, while it showed
significant but lower affinity in metabolic activity with coumarin and EFC (Table 3).

Thus, we selected BFC for CYP1A2 and CYP3A131, coumarin for CYP2A13 and
7-EC for CYP2E2 as fluorogenic substrates to test the effects of xenobiotics in the
following study.
3.4. Determination of ligand IC50
We selected 17 compounds that are reported to inhibit CYP activity in mice and
humans, including selective inhibitors for CYP isoforms. We also selected two
anticancer drugs as a pilot study. We used 5–7 concentrations of the substrates for
calculation of Km and Vmax values in triplicate measurements, after we tested 7–9
concentrations of substrates to determine an appropriate range in preliminary
experiments. The IC50 data obtained for each CYP isoform are indicated in Table 4.
Overall, there was no compound that showed activation of CYP isoforms studied
(Table 4). IC50 values of most chemicals, with exceptions, exceeded 100 µM on
CYP2E2 in cat. Other CYP isoforms, CYP1A2, 2A13, and 3A131 showed their
characteristic sensitivity to these 19 chemicals.
Three azole antifungals, clotrimazole, ketoconazole and miconazole showed very
strong inhibitory activity on feline CYP3A131 at around the submicromolar range.
These antifungals moderately inhibited CYP1A2, following CYP3A131. CYP2A13 was
relatively insensitive to these antifungals. The antifungals hardly affected CYP2E2
activity except miconazole that showed weak inhibition.
Fluvoxamine, a selective serotonin reuptake inhibitor (SSRI) inhibited activity of
feline CYP1A2 in the micromolar range. By contrast, the other SSRI, sertraline, showed
relatively strong inhibition on feline CYP3A131 much more than CYP1A2.

Two selective inhibitors for adenosine receptor (P2Y) that are used as antiplatelet
agents, clopidogrel and ticlopidine, were generally ineffective on these feline CYP
isoforms. Quinidine, an antiarrhythmic agent, showed some modulation to CYP3A131
only.
Methoxsalen that is used in treatment of various skin diseases inhibited feline
CYP2A13 activity at the submicromolar range. Methoxsalen also strongly inhibited
CYP1A2 and 3A131.
Two antiulcer agents, cimetidine, a H2 histamine receptor blocker and omeprazole,
a proton pump inhibitor, hardly affected feline CYP isoforms, although omeprazole
showed weak inhibition on CYP3A131. Erythromycin, a macrolide antibiotic agent
weakly inhibited feline CYP3A131.
Tryptamine, a monoamine alkaloid, showed weaker inhibition on feline CYP2A13
but relatively strong effect on feline CYP1A2. Quercetin, a flavonol found in many
plants, inhibited feline CYP1A2 and CYP3A131.
Medetomidine and atipamezole are frequently used for sedative hypnotics and the
antidote, respectively, in veterinary medicine. Both imidazole derivatives strongly
inhibited most feline CYP isoforms except CYP2E2. CYP2A13 was particularly
sensitive to these agents with an IC50 of 0.2 µM. Effects of dexamethasone were not
observed practically.
Finally, the effects of two anti-cancer agents, cyclophosphamide and tamoxifen,
were studied. The result was that both drugs showed no or only weak inhibition on
CYPs examined (Table 4).

4. Discussion
We determined relative expression of nine CYPs in the liver and small intestine of
cats. The results showed that CYP2E2, CYP2A13, and CYP2E1 are three major
transcripts in the feline liver. These three transcripts accounted for about 60% of nine
CYPs transcripts addressed. In our previous study, CYP2B subtype was not present in
the feline liver but present in the lung and small intestine as both transcript and
immunoreactivity of CYP2B6. The expression pattern of CYP isoforms in the feline
liver was very different to that of canine liver containing CYP2E1 (26%), CYP2A25,
CYP2D15 (25% for both CYP isoforms), CYP2C21 (14%), CYP3A12 (5%) and
CYP2B11 (2%) (Martinez et al., 2014). The expression pattern of CYP isoforms in the
feline liver was also different to that of the human liver containing CYP2E1 (c.a. 55%),
CYP3A4 (11%), 2C8 (9), 2C9 (8), 1A2 (6), 2A6 (3) (Bièche et al., 2007). Although
CYP2E1 was the most major transcript of CYP isoforms that were involved in
xenobiotics metabolism (c.a. 57%) and contribution of CYP3A4/5 was only about 12%
(Bièche et al., 2007), estimated CYP3A protein level (c.a. 29%) exceeded that of
CYP2E1 (c.a. 7%) in the human liver, as revealed by immunoblotting (Shimada et al.,
1994). This discrepancy was partially explained by negative regulation of CYP2E1
protein level by miR-378, a microRNA (Mohri et al., 2010). Thus, protein levels of
CYPs in the feline liver should be studied in the following study. Overall, however, it is
suggested that a few CYP isoforms are majorly expressed in the liver of cat compared to
those of dog and human at least in mRNA level.
CYP3A131 was the overwhelmingly predominant transcript of all CYP isoforms
studied in the feline small intestine in our data. In dog, CYP3A12 is the most abundant

CYP isoform in the intestine, using multiple reaction monitoring mass spectrometry
(Heikkinen et al., 2012). Of two CYP3A isoforms in dog, CYP3A26 is a major
transcript of the CYP3A isoform compared to CYP3A12 in the liver, but the reverse is
the case in the small intestine (Mealey et al., 2008). In cat, transcripts of CYP3A131 are
a hundred times as high as those of CYP3A132 in both the liver and small intestine
(Honda et al., 2011). Expression of CYP3A131 in the small intestine was comparable
with that in the liver of cat, suggesting the importance of CYP3A131 in oral drug
administration in cat (Honda et al., 2011). As the highest concentration of CYPs in the
small intestines of other mammals is found in the duodenum (Heikkinen et al., 2012),
more CYP3A131 transcripts could be detected in the upper part of the small intestine,
because w used middle part of the small intestine. Abundance and activity of both CYP
isoforms were higher in the liver than in the small intestine of dog (Mealey et al., 2008).
In human, the intrinsic activities of hepatic and intestinal CYP3A4 were comparable for
the 16 drug metabolism (Galetin and Houston, 2006), although abundance of CYP3A4,
a major isoform also in the small intestine, is lower than that in the liver based on
immunoblotting studies (Paine et al., 2006).
In this study, the properties of metabolism of nine fluorescent substrates by four
major feline CYP isoforms were characterized. Fluorescent substrates are useful
because of the easy and simple procedures adequate for high-throughput screening
(Crespi et al., 1997: van Beusekom et al., 2010). Although substrates strictly specific to
each CYP isoform should be explored, there is no information on gold standard
substrate in feline CYP isoform. Fluorescent substrates are useful because of the easy
and simple procedures adequate for high-throughput screening (Crespi et al., 1997: van
Beusekom et al., 2010). In this study, functional CYP1A2 could be prepared based on

the method by Iwata et al. (1998). Although feline functional CYP3A131 also could be
prepared with the same method (Iwata et al., 1998), we selected the method used by
Pritchard et al. (2006), because usage of OmpA leader sequences improved the amount
of recombinant CYP3A131 protein. Functional feline CYP2E2 was not obtained by the
same method for CYP1A2 and 2A13, but by the method according to Uno et al. (2006).
Overall, Km values of the four CYP isoforms to chemicals addressed were similar
to those of human and dog using recombinant systems; however, some differences were
also confirmed. Km values of metabolic activities of resorufin derivatives, BR and MR
by feline CYP1A2 were similar to those of human CYP1A2, while Km in metabolism
of BFC, a coumarin derivative of feline CYP1A2, is seven times as high as that by
human CYP1A2 (Stresser et al., 2002). Feline CYP3A131 showed ten times lower
affinity to BFC than that of human CYP3A4 (Stresser et al., 2002). Feline CYP2A13
catalyzed coumarin with similar affinity to human CYP2A6 and canine CYP2A13,
while canine CYP2A25 did not show significant metabolic activity of coumarin
(Stresser et al., 2002; Zhou et al, 2010). Information on fluorescent substrates
metabolism by CYP2E is very limited. Although human CYP2E1 catalyzed MFC in
both bactosome of E. coli and microsomes of immortalized human liver epithelial cells
individually expressing human CYP2E1 (Stresser et al, 2002; Donato et al., 2004),
feline CYP2E2 did not show any activity in this study. None of the four CYP isoforms
studied catalyzed MFC. This is in contrast to the study with feline liver microsomes,
showing significant metabolic activity of MFC (van Beusekom et al., 2010). As MFC is
known as a good substrate for CYP2C9 and CYP2E1 in human (Donato et al., 2004),
we should consider possible difference of substrate specificity of feline CYP isoform to
MFC as well as possible involvement of CYP isoform other than four CYP isoforms we

tested. Feline CPY2C has not been reported and recombinant feline CYP2E1 is not
obtained yet. Of these nine fluorescent substrates we studied, BFC for CYP1A2 and
CYP3A131, coumarin for CYP2A13, and 7-EC for CYP2E2, which can be used as
substrates for drug-drug interaction studies.
In this study, we developed fluorescence-based drug-drug interaction systems with
four major feline CYP isoforms and their fluorescent substrates. We confirmed possible
interactions of four major CYP orthologues with 19 chemicals, most of which were
reported to inhibit enzymatic activities of CYP isoforms in other mammals. Two thirds
of them are medical drugs used for feline patients (Papich, 2010; Plumb et al., 2011). In
vitro screening studies for inhibition on CYP isoforms heterologously expressed in E.
coli with fluorescent substrates were extensively carried out in human (Kajbaf et al.,
2011) and mice (McLaughlin et al., 2008). We selected twelve test chemicals from the
drugs previously examined in these two studies. IC50 values of these overlapping
chemicals in our feline experiments were compared with those of human and mouse in
Table 4. Kajbaf et al. (2011) tested the effects of a large number of drugs on
metabolism of some substrates by bactosome expressing human CYP1A2, 2C9, 2C19,
3A4 and human lymphoblastoid cells expressing human CYP3A4. Kajbaf et al. (2011)
confirmed fairly good correlation was confirmed between substrates with some
exceptions. They also showed a reasonable agreement between IC50 values obtained
with bactosomes and microsomes of human lymphoblastoid cells. Additionally, Donato
et al. (2004) also reported the same conclusion in primary human hepatocytes or in
human liver microsomes after comparison and the results obtained with fluorescent
probes and other specific probes. Furthermore, they also concluded that correlation
between bactosome and human in vivo data is even better than that between

lymphoblastoid preparation and in vivo data. It was reported that inhibitory effect of a
drug on enzymatic activity was dependent on the substrate employed in human
CYP3A4 (Kenworthy et al., 1999; Yuan et al., 2002). Additionally, we used human
POR not feline one, although the species of POR used in the single CYP systems may
be of some importance in regards to enzyme kinetics (Locuson et al., 2009). There is a
possibility that the system could be a good in vivo representation based on findings from
similar studies in humans but would require evaluation of feline microsomes and in vivo
data. Thus, we should be careful about the data obtained with limited substrates;
however, the data could be worth evaluating in the screening system.
In this study, it was CYP3A131 that was highly sensitive to the largest number of
drugs out of four feline CYP isoforms assayed. Human CYP3As are involved in
metabolism of a variety of drugs and are regarded as one of the most sensitive CYP
isoforms to chemicals, and thus the site of many drug-drug interaction. Human CYP3A
isoforms have the largest active site and is associated with the greatest magnitude of
diversity in substrates (Lewis et al., 1998). By contrast, CYP2E2 which was the most
abundant transcript of CYP isoforms studied in the feline liver, was insensitive to the
xenobiotics examined in this study. Feline CYP2A13 was very sensitive to a limited
number of xenobiotics with submicromolar IC50 values.
It is well known that azole fungicides exert potent inhibition on most CYP
isoforms in the human. Particularly, ketoconazole preferentially inhibits human
CYP3A4/5, while miconazole shows relative selectivity to human CYP2 C19 and
CYP3A4 (Zhang et al., 2002). Bactosome preparations of murine and human CYP3As
were similarly modulated by ketokonazole (IC50: CYP3A11 < CYP2A4 < CYP1A2 <<
CYP2E1 no inhibition) (McLaughline et al., 2008). Similar to the case in mouse, low

concentrations of ketoconazole also inhibited feline CYP3A131 preferentially. The
inhibitory effect of ketokonazole was moderate to feline CYP1A2 and weak for both
feline CYP2A13 and CYP2E2 in this experiment. We also confirmed similar selective
inhibition by miconazole and clotrimazole to feline CYP3A131 and CYP1A2 with
somewhat different potency. Serum concentration of ketoconazole could reach 10 µM
in adult human (Huang et al., 1986). Information on pharmacokinetic ketoconazole is
not available in cat. Quite recently, however, it was reported that the Cmax for
posaconazole in healthy cats that were orally administered (15 mg/kg) is approximately
1.7 µM (Mawby et al., 2016). These suggest that 3A131 inhibition is indeed a concern
by azole antifungals in cat.
It was reported that medetomidine, a substrate for canine CYP2B11 showed strong
inhibition on canine CYP2B11 that was heterologously expressed in E. coli (IC50:
0.0048 µM for racemic form) as well as other CYP isoforms (IC50: 1.1 μM for
CYP1A1 and CYP2D15) (Baratta et al., 2010). In this and previous studies (Table 4,
Okamatsu et al., 2016), medetomidine effectively inhibited activities of feline CYP1A2,
CYP2B6 (IC50: 0.024 µM), CYP2A13, and CYP3A131, although CYP2B6 was not
expressed in the feline liver (Okamatsu et al., 2015). Additionally, atipamezole, an
antidote for medetomidie also showed strong inhibition on activities of CYP1A2,
CYP2A13, CYP2B6 (IC50: 0.11 µM) (Okamatsu et al., 2016) and CYP3A131 but not
CYP2E2 (Table 4). According to company’s instructions, atipamezole is used in a five-
fold concentration of medetomidine to feline patients, sounding the alarm on the
possible drug-drug interactions that should be considered for many drugs in veterinary
medicine, although we do not have any information on serum and the exact-free hepatic
concentrations of medetomidine and atipamezole concentrations in cat.

Fluvoxamine, a specific inhibitor for human CYP1A2 and CYP2C19
(Madhusoodanan et al., 2014), showed reasonably comparable affinity to CYP1A2 and
CYP3A131 in cat, while it required about 300-fold concentrations to obtain IC50 on
CYP3A4 compared to CYP1A2 in human (Kajbaf et al., 2011). Methoxsalen, an
inhibitor specific to human CYP2A6 (Zhang et al., 2001) showed strong inhibition on
feline CYP2A13 and also on feline CYP1A2 and 3A131, while tryptamine, another
inhibitor for CYP2A6/13 (Stephens et al., 2012) showed weaker inhibition on feline
CYP2A13 but not CYP3A131. Tryptamine also inhibited feline CYP1A2 activity. We
already reported that clopidogrel and ticlopidine, inhibitors specific to human CYP2B6,
also inhibit feline CYP2B6 enzymatic activity in lower concentrations (IC50: 0.33 µM
for clopidogrel, 5.82 µM for ticlopidine) (Okamatsu et al., 2016). In this study, these
xenobiotics hardly affected activities of all feline CYP isoforms addressed, suggesting
the selective inhibition on feline CYP2B6. Sertraline, a SSRI, inhibits human CYP2B6
(IC50: 3.2 µM) (Walsky et al., 2006). Feline CYP2B6 also was very sensitive to this
drug (IC50: 0.49 µM) (Okamatsu et al., 2016). Sertraline exerted relatively strong
inhibition on feline CYP3A131 in this study. Recently, sertraline was reported to show
mechanism-based inhibition on human CYP3A4 (Masubuchi and Kawaguchi, 2013).
The proton pump inhibitor, omeprazole, is a metabolism-dependent reversible inhibitor
of human CYP2C19 and CYP3A4 (Ogilvie et al., 2011). Feline CYP3A131 showed
similar or moderate sensitivity to omeprazole compared to corresponding isoforms in
mouse and human in general, although different IC50 values were reported for human
CYP3A4 by substrates used (Table 4). Another group of antiulcer drug, cimetidine, is
well known as a weak inhibitor for various human CYP isoforms including CYP1A2,
2D6, and CYP3A4, at least with human liver microsomes (Martínez et al., 1999). On

the contrary, human CYP isoform-expressing bactosomes or lymphoblastoid cells are
only inhibited by cimetidine at the concentration of 100 µM or higher (Kajbaf et al.,
2011). In this study, cimetidine did not affect activity of feline CYP1A2 and showed
only weak inhibition on feline CYP3A131 with an IC50 of about 188 µM. Erythromycin,
a classical antibiotic, inhibits human CYP3A4 using bactosomes of E. coli and
microsomes of human β-lymphoblastoid cell lines (Table 4). We confirmed weaker
inhibitory effects of erythromycin in feline CYP3A131 in this study (Table 4).
CYP2C isoforms have not been reported yet in cat. We were not able to generate a
functional feline CYP2D6 recombinant preparation using the same strategies used for
the other five feline CYP isoforms in E. coli. Nevertheless, we also compared the
specificity of inhibitors for these CYP isoform with feline CYP orthologs we had.
Quercetin, a flavonoid, is known as an inhibitor specific to human CYP2C8 (Hosaka et
al., 2015). Quercetin showed strong inhibition on other CYP isoforms in mouse (IC50:
CYP1A2 < CYP3A11 < CYP2A4 = CYP2E1) (Table 4). The present results in feline
CYP isoforms showed a similar rank order of potencies as that in mouse, although
lower affinities. Quinidine, a well-known inhibitor specific to human CYP2D6, showed
only weak inhibition on feline CYP isoforms except CYP3A131. IC50 values for
quinidine are relatively variable with recombinant human CYP3A4 bactosomes of E.
coli ranging from 5.8 - 60 µM, dependent on substrates used (Table 4).
Dexamethasone is widely known as an inducer of human CYP3As, but this has not
been observed in cultured canine hepatocytes (Lu and Li, 2001). It was reported that
clinical oral doses of dexamethasone decreases intrinsic clearance of quinidine, a canine
CYP3A substrate possibly through reduction of α1-acid glycoprotein (AGP) and
inhibition of CYP3A activity (Zhang et al., 2006). Significant changes in the

pharmacokinetic parameters of oleandomycin were observed after intramuscular
injection of dexamethasone to cats but not after oral administration (Milanova and
Lashev, 2002). In this study, dexamethasone showed only weak inhibition on feline
CYP3A131 (IC50: 137.1 µM), similar to human CYP3A4 (Table 4).
Finally, we also tested two representative anticancer drugs in veterinary medicine
as these drugs are used on geriatric cats, generally possessing, lower metabolic ability,
over long time periods. In the results, cyclophosphamide showed slight inhibition on
CYP2A13 only, while tamoxifen slightly inhibited most CYP isoforms except
CYP2A13. Although cyclophosphamide exerts moderate inhibition on both mouse
CYP2A4 (IC50 87.1 µM) and CYP2C9 (IC50 60.7 µM) using bactosome systems (Table
4; McLaughlin et al., 2008), a low concentration of tamoxifen inhibits human CYP3A4
(IC50 1.6 µM) in contrast to feline CYP3A131 (Table 4). Tamoxifen strongly inhibits
mouse CYP1A1 (IC50 1.8 µM) and CYP2B20 (IC50 0.86 µM) and moderately inhibits
some CYP isoforms at 100 µM, while it showed similar affinity to mouse CYP3A11
(IC50 28 µM) (McLaughlin et al., 2008; Kajbaf et al., 2011). Further experiments with
other anticancer drugs should be undertaken to determine modulation of feline CYP
isoforms.
In conclusion, we established an in vitro screening system for possible drug-drug
interactions on four major feline CYP isoforms that were heterologously expressed in E.
coli. Four feline CYP isoforms showed similar selectivity to homologues in other
mammals against fluorescent substrates and inhibitors as a whole but with many
exceptions. A different expression pattern of CYP isoforms was confirmed in the liver
and small intestine in cat, compared to those in human and dog. Many drugs used in
veterinary practice are off-label human drugs and pharmacokinetic data in cats are

extremely limited at the present time. Therefore, prediction of drug-drug interactions
from in vitro data is of great value. Although we are not able to know how these
recombinant feline CYP isoforms reflect endogenous ones, the present study is a first
step to clarify this important problem.
Acknowledgments.
This study was supported by Grants-in-Aid for Scientific Research (MEXT/JSPS
KAKENHI 24580462 (C)) and a Grant-in-Aid to Cooperative Research from Rakuno
Gakuen University (H.T.). We sincerely thank Dr. Isogai and Dr. Takahashi for
providing animals and Ms. Tamura for technical assistance.

REFERENCES
Bailey J, Thew M, Balls, M. (2015). Predicting human drug toxicity and safety via
animal tests: can any one species predict drug toxicity in any other, and do
monkeys help? Altern Lab Anim 43:393-403.
Baratta MT, Zaya MJ, White JA, Locuson CW. (2010). Canine CYP2B11 metabolizes
and is inhibited by anesthetic agents often co-administered in dogs. J Vet
Pharmacol Ther 33:50-55.
Baririan N, Desager JP, Petit M, Horsmans Y. (2006). CYP3A4 activity in four different
animal species liver microsomes using 7-benzyloxyquinoline and
HPLC/spectrofluorometric determination. J Pharm Biomed Anal 40:211-214.
Barnes HJ. (1996). Maximizing expression of eukaryotic cytochrome P450s in
Escherichia coli. Methods Enzymol 272:3–14.
Bièche I, Narjoz C, Asselah T, Vacher S, Marcellin P, Lidereau R, Beaune P, de Waziers
I. (2007). Reverse transcriptase-PCR quantification of mRNA levels from
cytochrome (CYP)1, CYP2 and CYP3 families in 22 different human tissues.
Pharmacogenet Genomics 17:731-742.
Chauret N, Gauthier A, Martin J, Nicoll-Griffith DA. (1997). In vitro comparison of