Different hydrolases involved in bioactivation of prodrug-type angiotensin receptor blockers: Carboxymethylenebutenolidase and carboxylesterase 1

Article abstract of DOI:10.1124/dmd.113.053595

Olmesartan medoxomil (OM) is a prodrug-type angiotensin II type 1 receptor blocker (ARB). We recently identified carboxymethylenebutenolidase homolog (CMBL) as the responsible enzyme for OM bioactivation in humans. In the present study, we compared the bioactivating properties of OM with those of other prodrug-type ARBs, candesartan cilexetil (CC) and azilsartan medoxomil (AM), by focusing on interspecies differences and tissue specificity. In invitro experiments with pooled tissue subcellular fractions of mice, rats, monkeys, dogs, and humans, substantial OM-hydrolase activities were observed in cytosols of the liver, intestine, and kidney in all the species tested except for dog intestine, which showed negligible activity, whereas lung cytosols showed relatively low activities compared with the other tissues. AM-hydrolase activities were well correlated with the OM-hydrolase activities. In contrast, liver microsomes exhibited the highest CC-hydrolase activity among various tissue subcellular fractions in all the species tested. As a result of Western blot analysis with the tissue subcellular fractions, the band intensities stained with anti-human CMBL and carboxylesterase 1 (CES1) antibodies well reflected OM- and AM-hydrolase activities and CC-hydrolase activity, respectively, in animals and humans. Recombinant human CMBL and CES1 showed significant AM- and CC-hydrolase activities, respectively, whereas CC hydrolysis was hardly catalyzed with recombinant carboxylesterase 2 (CES2). In conclusion, OM is bioactivated mainly via intestinal and additionally hepatic CMBL not only in humans but also in mice, rats, and monkeys, while CC is bioactivated via hepatic CES1 rather than intestinal enzymes, including CES2. AM is a substrate for CMBL. Copyright

Full text of DOI:10.1124/dmd.113.053595

Supplemental material to this article can be found at:
http://dmd.aspetjournals.org/content/suppl/2013/08/14/dmd.113.053595.DC1.html
1
D
521-009X/41/11/1888–1895$25.00
RUG METABOLISM AND DISPOSITION
http://dx.doi.org/10.1124/dmd.113.053595
Drug Metab Dispos 41:1888–1895, November 2013
Copyright ª 2013 by The American Society for Pharmacology and Experimental Therapeutics
Different Hydrolases Involved in Bioactivation of Prodrug-Type
Angiotensin Receptor Blockers: Carboxymethylenebutenolidase
s
and Carboxylesterase 1
Tomoko Ishizuka, Yasushi Yoshigae, Nobuyuki Murayama, and Takashi Izumi
Drug Metabolism and Pharmacokinetics Research Laboratories, Daiichi Sankyo Co., Ltd., Tokyo, Japan
Received July 1, 2013; accepted August 13, 2013
ABSTRACT
Olmesartan medoxomil (OM) is a prodrug-type angiotensin II type 1
receptor blocker (ARB). We recently identified carboxymethylene- activity among various tissue subcellular fractions in all the species
contrast, liver microsomes exhibited the highest CC-hydrolase
butenolidase homolog (CMBL) as the responsible enzyme for OM
bioactivation in humans. In the present study, we compared the
bioactivating properties of OM with those of other prodrug-type
ARBs, candesartan cilexetil (CC) and azilsartan medoxomil (AM), by
focusing on interspecies differences and tissue specificity. In in-
tested. As a result of Western blot analysis with the tissue
subcellular fractions, the band intensities stained with anti-human
CMBL and carboxylesterase 1 (CES1) antibodies well reflected
OM- and AM-hydrolase activities and CC-hydrolase activity, re-
spectively, in animals and humans. Recombinant human CMBL and
vitro experiments with pooled tissue subcellular fractions of mice, CES1 showed significant AM- and CC-hydrolase activities, re-
rats, monkeys, dogs, and humans, substantial OM-hydrolase spectively, whereas CC hydrolysis was hardly catalyzed with
activities were observed in cytosols of the liver, intestine, and
kidney in all the species tested except for dog intestine, which
showed negligible activity, whereas lung cytosols showed relatively
low activities compared with the other tissues. AM-hydrolase
activities were well correlated with the OM-hydrolase activities. In
recombinant carboxylesterase 2 (CES2). In conclusion, OM is
bioactivated mainly via intestinal and additionally hepatic CMBL
not only in humans but also in mice, rats, and monkeys, while CC is
bioactivated via hepatic CES1 rather than intestinal enzymes,
including CES2. AM is a substrate for CMBL.
Introduction
and Yokoi, 2012), interspecies differences in tissue-specific esterase
expressions at times challenge extrapolation from animal data to
humans. For instance, many examples have shown higher hydrolase
activity in rodent species than humans, while there is minimal
hydrolase activity in dog small intestine (Liederer and Borchardt,
Historically, many ester-based prodrugs have been developed with
the aim of overcoming a number of barriers to drug-like properties
Liederer and Borchardt, 2006; Huttunen et al., 2011). Esterases,
which are involved in bioactivation of ester-based prodrugs and
inactivation of clinical drugs as well, are typically represented by
carboxylesterases (CES) (Satoh and Hosokawa, 2006; Hosokawa,
(
2
006; Satoh and Hosokawa, 2010). These observations suggest the
importance of selecting appropriate animal species that most closely
represent humans in drug development processes, based on compre-
hensive understanding of the species differences.
Angiotensin receptor blockers (ARBs) that directly inhibit the
binding of angiotensin II to angiotensin II type 1 receptor provide an
effective pharmacologic strategy in the management of hypertension,
cardiovascular disease, and renal disease (De Gasparo et al., 2000).
Among eight market-available ARBs, candesartan cilexetil (CC),
2
008), cholinesterases (Taylor, 1991; Taylor and Radic, 1994), and
paraoxonases (Draganov and La Du, 2004). In addition, recent works
revealed several valuable aspects of other esterases responsible for
drug hydrolysis: valacyclovirase as a bioactivating hydrolase for the
amino-acid ester prodrugs valacyclovir and valganciclovir (Kim et al.,
003; Lai et al., 2008); arylacetamide deacetylase as a hydrolase being
associated with organ toxicities of clinical drugs such as flutamide,
2
phenacetine, and rifamycins (Watanabe et al., 2009; Watanabe et al., olmesartan medoxomil (OM), and azilsartan medoxomil (AM) are
2
1
010; Nakajima et al., 2011); and a/b hydrolase domain-containing ester-based prodrugs that require hydrolysis to be converted into their
0 (ABHD10), which functions as the detoxification enzyme of respective active metabolites. As shown in Fig. 1, the prodrug moieties
mycophenolic acid acyl-glucuronide (Iwamura et al., 2012).
of OM and AM are identical, where the medoxomil moiety links to the
Although esterases are attractive targets for prodrug bioactivating carboxylic acid of the drug via an ester bond (Laeis et al., 2001;
enzymes due to their wide distribution in the blood, liver, intestine, Kawaguchi et al., 2013), while the CC is a cyclohexyloxycarbonyloxy
and many other biologic fluids and tissues (Berry et al., 2009; Fukami ethyl ester of the carboxylic acid of candesartan (Nishikawa et al.,
1
997). We recently reported that a novel human hydrolase,
dx.doi.org/10.1124/dmd.113.053595.
s This article has supplemental material available at dmd.aspetjournals.org.
carboxymethylenebutenolidase homolog (CMBL), is responsible for
the OM bioactivation (Ishizuka et al., 2010, 2012); however,
ABBREVIATIONS: AM, azilsartan medoxomil; ARB, angiotensin receptor blocker; BNPP, bis-p-nitrophenylphosphate; CC, candesartan cilexetil;
CES, carboxylesterase; CMBL, carboxymethylenebutenolidase homolog; K , Michaelis constant; OM, olmesartan medoxomil; RNH-6272, 2-butyl-
-(1-hydroxy-1-methylethyl)-1-[29-(1H-tetrazole-5-yl)-4-biphenylyl]-1H-imidazole-5-carboxylic acid; S9, 9,000g supernatant; Vmax maximum
velocity.
m
4
,
1888

Bioactivation of ARB Prodrugs by CMBL and CES1
1889
responsible enzymes for CC- and AM-bioactivations have not been Japan (Tokyo, Japan), respectively. Bis-p-nitrophenylphosphate (BNPP) was
purchased from Tokyo Chemical Industry (Tokyo, Japan). All the other
clearly identified thus far.
chemicals and reagents were of the highest grade available and were obtained
This study was designed to investigate the difference in the
from standard commercial sources.
bioactivating properties of the prodrug-type ARBs by comparing
CC- and AM-hydrolase activities with that of OM, using human and
animal tissue subcellular fractions and recombinant proteins of can-
Tissue Subcellular Fractions and Plasma. Pooled male CD1 mouse, male
Sprague-Dawley rat, male cynomolgus monkey, male beagle dog, and mixed
gender human tissue (liver, intestine, kidney, and lung) cytosolic and microsomal
didate human hydrolases. Also, this work provides new insights into
fractions were used in the study. Mouse intestinal, human kidney, rat, monkey,
interspecies differences in tissue-specific expressions and activities
dog, and human lung cytosolic fractions were separated from commercially
of the novel hydrolase CMBL between humans and other animal available 9000g supernatant (S9) fractions of respective tissues (XenoTech, LLC,
species.
Lenexa, KS) as supernatant by ultracentrifugation at 105,000g at 4°C for 1 hour.
Mouse lung microsomal fraction was also prepared from a mouse lung S9
fraction (XenoTech, LLC). The pellet obtained by ultracentrifugation of the S9
fraction at 105,000g at 4°C for 1 hour was washed and repelleted with 10 mM
potassium phosphate buffer (pH 7.4) containing 1.15% KCl by ultracentrifugation
with the same conditions as above. The obtained pellet was resuspended with
Materials and Methods
Olmesartan medoxomil (OM), olmesartan, candesartan, and RNH-6272 [2-
butyl-4-(1-hydroxy-1-methylethyl)-1-[29-(1H-tetrazole-5-yl)-4-biphenylyl]-1H-
imidazole-5-carboxylic acid] were synthesized in Daiichi Sankyo (Tokyo, 10 mM potassium phosphate buffer (pH 7.4) containing 30% glycerol for use as
Japan). RNH-6272 is a structural analog of olmesartan and was used as an
the microsomal fraction. The other cytosolic and microsomal fractions were
internal standard for determination of prodrug active metabolites (olmesartan, purchased from XenoTech, LLC. Pooled male CD1 mouse, male Sprague-
candesartan, and azilsartan). Candesartan cilexetil (CC) was synthesized at
Chemtech Laboratory, Inc. (Tokyo, Japan). Azilsartan medoxomil (AM) and
Dawley rat, male cynomolgus monkey, male beagle dog, and male human plasma
were used in the study. Blood was collected from three or more individuals and
azilsartan were purchased from Cosmo Bio (Tokyo, Japan) and Sigma-Aldrich centrifuged to separate plasma.
Fig. 1. Bioactivation of the prodrug-type ARBs. Hydrolytic reactions of olmesartan medoxomil (A), candesartan cilexetil (B), and azilsartan medoxomil (C) are shown.

1
890
Ishizuka et al.
Recombinant Human Carboxylesterase. To construct a destination vector was initiated by adding the substrate. The prodrugs were incubated in 50 mM
containing a FLAG epitope tag, the synthesized oligonucleotides encoding Tris-HCl buffer (pH 7.5, incubation volume of 0.2 ml) containing 5% (v/v)
FLAG-tag (amino acid sequence: DYKDDDDK) and reading frame A
Invitrogen, Carlsbad, CA) were inserted into pcDNA3.1(+) (Invitrogen) as
a NheI/HindIII and EcoRV fragment, respectively. The resulting vector
conditioned media from CES1- or CES2-overexpressed cells as an enzyme
source at 37°C for 2 minutes. The enzymatic activity was expressed as
a metabolite formation rate per reaction (nanomole per minute per reaction)
(
pcDNA3.1-CFLAG-GW was verified by sequencing. pDONR-human CES1 based on the production of respective metabolites.
and pDONR-human CES2 entry clones, which contain human CES1 and CES2
Hydrolysis by Recombinant Human CMBL. The OM-, CC- and AM-
genes without their native stop codon, respectively (produced by Invitrogen), hydrolase activities by recombinant human CMBL were measured with the
were recombined with the destination vector pcDNA3.1-CFLAG-GW using LR
10-mM prodrugs as substrates (final solvent concentration: 2.5% acetonitrile) in
clonase II (Invitrogen) to generate the expression vectors pcDNA3.1-CFLAG- triplicate. After 5-minute preincubation, the reaction was initiated by adding the
CES1 and pcDNA3.1-CFLAG-CES2, which encode human CES1 and CES2, substrate. The prodrugs were incubated in 100 mM HEPES buffer (pH 7.4,
respectively, with a FLAG tag at the C terminus. The recombinant proteins
were produced in mammalian FreeStyle 293-F Cells (Invitrogen). Gene-
transfected cells were cultured for 7 days in FreeStyle 293 Expression Medium
incubation volume of 0.2 ml) at 37°C for 5 minutes with 100 mg of the human
CMBL-transfected mammalian cell lysate constructed in-house (Ishizuka et al.,
2010) as an enzyme source. The enzymatic activity was expressed as
(
Invitrogen). The conditioned media from the cells overexpressing human
CES1 and CES2 were dialyzed against 20 mM Tris-HCl buffer (pH 8.0) and
filtered with a polyethersulfone membrane filter (0.45 mm; Thermo Fisher subtracted as nonenzymatic hydrolysis.
a metabolite formation rate (v
based on the production of azilsartan, from which the vector control was
0
: nanomole per min per milligram protein)
Scientific, Rockford, IL). The overexpressed FLAG-tagged protein in the
filtrate was purified with a two-step purification process: anion chromatography
with HiTrap Q-XL (GE Healthcare Japan, Tokyo, Japan) followed by anti-
In addition, the kinetic analysis for AM hydrolysis by recombinant human
CMBL was carried out with a substrate range of 0.781–100 mM. The kinetic
constant (K ) and maximum velocity (Vmax) for AM hydrolysis were estimated
m
FLAG M2 affinity gel (Sigma-Aldrich). The eluates were collected, desalted using WinNonlin Professional (version 1.4.1; Pharsight, Sunnyvale, CA) by
with PD MidiTrap G-25 (GE Healthcare Japan), and then the resultant protein
was stored frozen at –80°C until use.
Hydrolase Activity Measurement in Tissue Subcellular Fractions. The and substrate concentration, respectively.
hydrolase activities for the three prodrug-type ARBs, OM, CC, and AM, in
Western Blot Analysis of CMBL and CES1 Expression. The following
animal and human tissue subcellular fractions were measured with each 10-mM tissue subcellular fractions (liver, intestine, kidney, and lung cytosol and
a nonlinear least-squares regression analysis fitting to the Michaelis-Menten
equation, v + [S]), where v and [S] represent initial velocity
= Vmax  [S]/(K
0
m
0
prodrug as a substrate (final solvent concentration: 2.5% acetonitrile) in
duplicate. After 5-minute preincubation, the reaction was initiated by adding
microsomes) and diluted plasma were subjected to Western blot analysis:
mouse, rat and dog, 2 mg protein for each subcellular fraction and 10 ml of
the substrate. OM and AM were incubated in 100 mM HEPES buffer (pH 7.4, 100-fold diluted plasma; monkey and human, 0.5 mg protein for each
incubation volume of 0.2 ml) with 100 mg cytosolic or microsomal protein/ml, subcellular fraction and 5 ml of 200-fold diluted plasma. As positive controls,
and CC was incubated in 100 mM Tris-HCl buffer (pH 7.5, incubation volume
0.5 mg of human CMBL-overexpressed mammalian Freestyle 293-F cell lysate
of 0.2 ml) at 37°C for 1, 2, and 10 minutes with 10–100 mg cytosolic or and 10 ng of purified recombinant protein of human CES1 were also analyzed.
microsomal protein/ml. The reaction was terminated by adding a 4-fold volume
of ice-cold 87.5% acetonitrile containing RNH-6272 as an internal standard for
the determination of the respective active metabolite concentrations and 0.25%
The sample proteins were separated by SDS-PAGE using 12% sodium dodecyl
sulfate-polyacrylamide gel (Mini-PROTEAN TGX precast gel 12%; Bio-Rad,
Hercules, CA) and were transferred electrophoretically onto a polyvinylidene
formic acid for preventing the nonenzymatic degradation of the prodrugs. The difluoride membrane (Immun-Blot polyvinylidene difluoride membrane, 0.2 mm;
mixture, with a volume of ca. 200 ml, was filtered using a Captiva 96-well filter
Bio-Rad). The CMBL and CES1 proteins in animals and humans were stained
plate.
with affinity-purified rabbit polyclonal IgG against a human CMBL peptide (1:
For olmesartan measurement, the Captiva filtrate was mixed with 200 ml of 5000 dilution; IBL, Shizuoka, Japan) and affinity-purified rabbit polyclonal
0% methanol containing 1% formic acid, and was analyzed by liquid
IgG against a human CES1 peptide (ab52941, 1:1000 dilution; Abcam,
chromatography–tandem mass spectrometry consisting of a Prominence LC- Cambridge, MA) as primary antibodies, respectively, and ECL anti-rabbit IgG
0A system (Shimadzu, Kyoto, Japan) and an API3200 (Applied Biosystems/ horseradish peroxidase–linked, from donkey (GE Healthcare Japan, 1:10000
5
2
MDS SCIEX, Foster City, CA) as previously reported (Ishizuka et al., 2013).
For candesartan and azilsartan measurements, the filtrate was directly analyzed
dilution) as a secondary antibody. These immunoblots were visualized by
chemiluminescence with an ECL Advance Western Blotting Detection Kit (GE
by liquid chromatography–tandem mass spectrometry consisting of a Promi- Healthcare Japan). The immunoreactive signals were detected by a lumino-
nence CBM-20A system (Shimadzu) and an API4000 (Applied Biosystems/ image analyzer (LAS-4000UV mini; FujiFilm, Tokyo, Japan) and the signal
MDS SCIEX). The two metabolites were separated with a reversed-phase
intensities were semiquantified by Science Laboratory 2005 Multi Gauge
ODS column (Shim-pack XR-ODS, 2.0-mm  300-mm internal diameter; software (ver. 3.0; FujiFilm).
Shimadzu). The mobile phase total flow was set to 0.75 ml/min with binary
gradient elution, using solvent A (5% acetonitrile, 0.2% formic acid, 100 mM
ammonium acetate) and B (95% acetonitrile, 0.2% formic acid, 100 mM
ammonium acetate). The gradient started with 45% B for 0.5 minutes and was
increased to 100% B for 0.5 minutes. Elution was continued for 1 minute at
Protein Assay. Protein concentration was determined by the Bradford
method using bovine serum albumin as the reference standard.
Results
100% B, and then the total flow was increased to 5 ml/min to wash the column.
Prodrug Hydrolysis in Tissue Subcellular Fractions. The OM-
Candesartan and azilsartan were determined by monitoring the ion transition of
m/z 441 to m/z 263 and m/z 457 to m/z 233, respectively, with multiple reaction hydrolase activities with mouse, rat, monkey, dog, and human tissue
monitoring in the positive electrospray ionization mode.
cytosolic and microsomal fractions are shown in Fig. 2. The activity
The enzymatic activity was expressed as a metabolite formation rate (nanomole was generally higher in cytosols than in respective microsomes. Liver,
per minute per milligram protein) based on the production of respective active
intestine, and kidney cytosols exhibited substantial OM-hydrolase
metabolite for the reaction by each enzyme source, from which the buffer control as
activities in all the species tested except for dog intestine, whereas
nonenzymatic hydrolysis was subtracted. The incubation time and cytosolic/
lung cytosols showed relatively low activities. In rats, liver and kidney
microsomal protein concentration were selected to be in the linear range for the
microsomes showed comparable activities to those of respective
metabolite formation.
cytosols. In AM hydrolysis, similar patterns of the species- and tissue-
specific activity to OM hydrolysis were observed in both cytosols and
microsomes (data not shown). In contrast to OM hydrolysis, the
Hydrolysis by Recombinant Human CES1 and CES2. The OM-, CC- and
AM-hydrolase activities by recombinant human CES1 and CES2 were
measured with 10-mM prodrugs as substrates (final solvent concentration:
2
1
.5% acetonitrile) with or without 1 mM BNPP (final solvent concentration: CC-hydrolase activity was generally higher in microsomes than in
% dimethyl sulfoxide) in triplicate. After 5-minute preincubation, the reaction respective cytosols as shown in Fig. 3. In all the species tested, liver

Bioactivation of ARB Prodrugs by CMBL and CES1
1891
Fig. 3. CC-hydrolase activities with animal and human tissue subcellular fractions.
Prodrug CC was incubated with animal and human tissue cytosolic (A) and
microsomal (B) fractions, and the formed active metabolite candesartan was
measured by liquid chromatography–tandem mass spectrometry. Data represent the
means of duplicate determinations.
Fig. 2. OM-hydrolase activities with animal and human tissue subcellular fractions.
Prodrug OM was incubated with animal and human tissue cytosolic (A) and
microsomal (B) fractions, and the formed active metabolite olmesartan was
measured by liquid chromatography–tandem mass spectrometry. Data represent
the means of duplicate determinations.
microsomes exhibited the highest CC-hydrolase activity among different tissues (liver, intestinal, kidney, and lung) of five species
various tissue subcellular fractions. The CC hydrolysis was mostly (mouse, rat, monkey, dog, and human) were used. As shown in Fig. 4,
liver-specific in humans, whereas significant activities were observed no correlation was observed between the CC- and OM-hydrolase
in the kidney and lung along with liver in the other species. Minimal activities in both cytosolic and microsomal fractions, whereas the AM-
activity of CC hydrolysis was observed in the intestine in all the hydrolase activities were well correlated with the OM-hydrolase
species tested.
activities in both cytosolic and microsomal fractions, with correlation
Using the above activity data, the CC-hydrolase and AM-hydrolase coefficients of 0.929 and 0.832, respectively.
activities were plotted against the OM-hydrolase activity. Twenty data
Hydrolysis by Recombinant Human CES1 and CES2. To
points for each cytosolic and microsomal fraction prepared from four confirm the substrate specificities of the recombinant human CES1
Fig. 4. Correlation analysis of hydrolase activities for prodrug-type
ARBs with animal and human tissue subcellular fractions. CC-
hydrolase (A and B) and AM-hydrolase (C and D) activities are
plotted against OM-hydrolase activity. Twenty data points for each
cytosolic (A and C) and microsomal (B and D) fractions prepared
from four different tissues (liver, intestinal, kidney, and lung) of five
species (mouse, rat, monkey, dog, and human) were used. The
prodrugs were incubated with animal and human tissue cytosolic and
microsomal fractions and their respectively formed active metabolites
were measured by liquid chromatography–tandem mass spectrome-
try. Data represent the means of duplicate determinations.

1
892
Ishizuka et al.
and CES2 in conditioned media from the overexpressing cells, the
hydrolase activities of p-nitrophenyl acetate, clopidogrel, and CPT-11,
which are a general esterase substrate, CES1-specific and CES2-
specific substrate, respectively, were measured. The CES1- and CES2-
conditioned media showed comparable hydrolytic activity toward
p-nitrophenyl acetate, and that CES1 and CES2 specifically catalyzed
the clopidogrel hydrolysis, forming clopidogrel carboxylic acid and
the CPT-11 hydrolysis forming SN-38, respectively (Supplemental
Fig. 1).
As shown in Fig. 5, CC was rapidly de-esterified with the re-
combinant CES1 and produced its active metabolite candesartan at
a rate of 514 pmol/min per reaction, whereas the reaction was hardly
catalyzed with the recombinant CES2 (24.2 pmol/min per reaction).
Both activities were completely abolished by the addition of 1 mM human CES1- or CES2-overexpressed cell culture as an enzyme source under the
BNPP, a carboxylesterase inhibitor. Although the recombinant CES1
Fig. 5. Hydrolysis of prodrug-type ARBs by recombinant human CES1 and CES2.
Prodrugs OM, CC, and AM were incubated with 5% (v/v) conditioned media from
conditions with or without 1 mM BNPP, and the formed active metabolites
olmesartan, candesartan, and azilsartan were measured by liquid chromatography–
showed the slight OM- and AM-hydrolase activities as being higher
tandem mass spectrometry. Data represent the means
determinations.
6 S.D. of triplicate
than that of the mock control, the activities were much lower than that
for CC hydrolysis. The OM- and AM-hydrolase activities by the
recombinant CES2 were comparable with the mock control with or
without BNPP, suggesting that only nonenzymatic hydrolysis was
slightly observed.
Hydrolysis by Recombinant Human CMBL. AM, which is
another medoxomil-ester prodrug-type ARB, was substantially hydro-
lyzed by the recombinant human CMBL as well as OM (Fig. 6A) and
proteins, by probing against cynomolgus monkey, dog, and human
CES1 and CES2 enzymes (Williams et al., 2011). As previously
reported (Satoh and Hosokawa, 2010; Williams et al., 2011), CES1
proteins were highly expressed in the liver in the three species, and
were significantly expressed in monkey intestine and kidney, dog
kidney, lung, and plasma, and human lung as a single band with
a molecular mass approximately at 60 kDa. Regarding mouse and rat
samples, single or multiple bands were detected in all the tissues and
plasma.
exhibited simple Michaelis-Menten kinetics (Fig. 6B), with K
m
and
Vmax values of 14.8 61.3 mM and 5.40 6 0.16 nmol/min per milligram
protein (the mean 6 standard error of triplicate determinations). CC was
not hydrolyzed by the recombinant CMBL (Fig. 6A).
The protein band intensities of CMBL in Western blot analysis
mostly corresponded with the distribution of OM- and AM-hydrolase
activities, and those of CES1 mostly corresponded with the dis-
tribution of CC-hydrolase activities in the eight different tissue sub-
cellular fractions tested. The analysis plots for OM- and CC-hydrolase
activities toward CMBL and CES1 protein expressions in humans
were presented in Fig. 8 as exemplary cases. As shown in Table 1, the
correlation coefficients between CMBL protein expression levels and
OM- or AM-hydrolase activities were over 0.8 with one exception,
where that between rat CMBL protein and OM-hydrolase activity
was 0.477. The correlation coefficients between CES1 protein ex-
pression levels and CC-hydrolase activities were also over 0.8
without exception.
Western Blot Analysis of CMBL and CES1 Expression. The
immunoblots stained with polyclonal anti-human CMBL and anti-
human CES1 antibodies were shown in Fig. 7. In the human liver,
intestine, and kidney cytosols, the anti-human CMBL antibody
intensely stained a single band at approximately 25 kDa, which was
almost consistent with the theoretical mass 28 kDa, whereas it was
weakly stained in human lung cytosol. Compared with the cytosols,
the weaker CMBL expression was observed in the respective micro-
somes. Regarding the other animal species as well, liver, intestine, and
kidney cytosols demonstrated strongly stained bands with approxi-
mately 25-kDa molecular mass except for dog intestine, whereas lung
cytosols showed weak protein bands. The band intensity was generally
lower in microsomes than in respective cytosols in the animal species,
as was seen in humans. Collectively, these data indicate that the anti-
human CMBL antibody recognized the CMBL orthologs in the animal
species tested. No significant protein was detected in plasma samples
of all the species tested including humans.
Discussion
In the present study, we comparatively showed the difference
The anti-human CES1 antibody used in the present study was between the bioactivating properties of prodrugs OM and CC, which
reported to be CES1-specific, without cross-reactivity with CES2 belong to the same class of antihypertensive, ARB. We previously
Fig. 6. Hydrolysis of prodrug-type ARBs by recombinant human
CMBL. (A) OM-, CC-, and AM-hydrolase activities by recombinant
human CMBL. Prodrugs OM, CC, and AM were incubated with
recombinant CMBL-overexpressed cell lysate as an enzyme source,
and the formed active metabolites, olmesartan, candesartan, and
azilsartan, were measured by liquid chromatography–tandem mass
spectrometry. Data represent the means 6 S.D. of triplicate determi-
nations. (B) Kinetic analysis of AM hydrolysis. The solid line is the best
fit by nonlinear least-squares regression to the Michaelis-Menten
equation. Data represent the means 6 S.D. of triplicate determinations.
The respective Eadie-Hofstee plot is presented in the inset.

Bioactivation of ARB Prodrugs by CMBL and CES1
1893
and showed an increase of bioavailability as respective active metabolites
(Atacand, 2000; Laeis et al., 2001), meaning that OM and CC are both
good exemplary cases of absorbability improvement by esterification of
the carboxylic acid group.
Also, we presented for the first time species- and tissue-specific OM
hydrolysis catalyzed via CMBL—which was recently identified as an
OM-bioactivating hydrolase and had not received much attention until
then (Ishizuka et al., 2010)—using various tissue subcellular fractions
of animals and humans. The kidney and lung were selected in addition
to the liver and intestine, which were already demonstrated in humans
as bioactivating organs for the prodrug OM (Ishizuka et al., 2010).
Human kidney cytosol showed significant OM hydrolase activity
comparable to that of liver and intestinal cytosol, whereas lung cytosol
showed meaningfully lower activity. This tissue distribution pattern of
the activity, namely high in the liver, intestine, and kidney and low in
the lung, is consistent with the CMBL protein expression pattern
demonstrated by Western blotting in the present study and also with
the mRNA expression pattern previously reported (Ishizuka et al.,
2
010). The predominant expression of CMBL in the liver, intestine,
and kidney implies that CMBL has primarily functioned in de-
toxification of xenobiotic compounds through the evolutional process
from bacteria to humans. Bacterial CMBL, also called dienelactone
hydrolase, has been reported as a catabolic enzyme enabling the use of
chloroaromatic compounds (Schmidt and Knackmuss, 1980; Ngai
et al., 1987), which are generally toxic to organisms.
Mice, rats, and monkeys, all commonly used experimental animal
species, exhibited similar tissue distribution patterns to humans in the
OM-hydrolase activity and CMBL protein expression. Based on the
results, these animal species would provide dependable information to
predict human bioactivation of ester compounds hydrolyzed by
CMBL in the preclinical drug discovery and development settings.
Although relatively high activities in monkey liver and kidney, and
low activity in rat intestine appeared comparable with those in
humans, the in vitro study results would be helpful when considering
the quantitative difference of prodrug bioactivation between species.
As a notable exception to the tissue distribution of CMBL, in the dog
intestine slight OM-hydrolase activity or negligible CMBL protein
expression was observed, while intestinal cytosols in humans and the
other animal species exhibited significant activities and clear protein
expressions. The absence of dog intestinal CMBL was observed
despite highly conserved amino acid sequences with 87% homol-
ogy between humans and dogs (GenBank Accession Numbers
Fig. 7. Western blot analysis of CMBL and CES1 expressions in animals and
humans. The following tissue subcellular fractions (liver, intestine, kidney, and lung
cytosol and microsomes) and diluted plasma were subjected to Western blot
analysis: mouse, rat, and dog, 2 mg protein for each subcellular fraction and 10 ml of NP_620164.1 and XP_535793.1, respectively). This is a common
1
00-fold diluted plasma; monkey and human, 0.5 mg protein for each subcellular
feature of well-investigated esterases such as CES, paraoxonases,
and cholinesterases (Williams et al., 2011) and shows interesting
fraction and 5 ml of 200-fold diluted plasma. As positive controls, 0.5 mg of human
CMBL-overexpressed mammalian Freestyle 293-F cell lysate and 10 ng of purified
recombinant protein of human CES1 were also analyzed. The proteins were functional and evolutionary implications. Low capacity to detoxify
separated by SDS-PAGE, transferred electrophoretically onto a polyvinylidene
esters in dog intestine potentially suggests the minor importance of this
difluoride membrane, and stained by polyclonal anti-human CMBL and anti-human
ability in carnivores, which would have infrequent exposure to toxic
CES1 antibodies. The signal intensities of the protein bands in the subcellular
fractions at the mobility indicated with closed triangles were semiquantified. White substances in plants.
bars and open triangles represent molecular mass marker proteins.
Moreover, we also demonstrated that CMBL is a bioactivating
enzyme for recently launched prodrug-type ARB AM (azilsartan
medoxomil), whose prodrug moiety is identical to that of OM.
reported that the bioactivation of OM predominantly proceeds via Kawaguchi et al. (2013) recently reported a likely estimation that
intestinal CMBL in humans and that hepatic CMBL and plasma enzymes involved in the hydrolysis of AM would be the same as OM
paraoxonase 1 are additionally involved (Ishizuka et al., 2010, 2012). by considering the identical chemical structures of the prodrug moiety
On the contrary, present in vitro studies demonstrated that the between the two ARBs and in vitro metabolic study results using liver
contribution of intestinal enzymes to the CC bioactivation is relatively and intestinal S9 fractions and plasma. As shown in the results of our
small and that hepatic CES1 is mainly involved. Although the in vitro study using recombinant human CMBL, obvious AM-
responsible hydrolases and bioactivation sites were presented to be hydrolase activity clearly demonstrated the involvement of CMBL in
different between these two prodrugs, they were both reported to be the reaction. Furthermore, the clear positive correlation between OM-
rapidly and completely converted to their active metabolites through and AM-hydrolase activities suggests the significant contribution of
the first pass metabolism before they reach the systemic circulation CMBL, not only in humans but also other tested animal species, to

1
894
Ishizuka et al.
Fig. 8. Correlation between protein expression levels of hydrolases
and prodrug-hydrolase activities in humans. Data from eight
different human tissue subcellular fractions (liver, intestine, kidney,
and lung cytosolic and microsomal fractions) were employed to
show the correlations between the protein expressions and activities.
(A) CMBL versus OM-hydrolase activity, (B) CMBL versus CC-
hydrolase activity, (C) CES1 versus OM-hydrolase activity, (D)
CES1 versus CC-hydrolase activity.
hepatic and intestinal AM bioactivation. Only limited information on showed minimal CC-hydrolase activity in all the species tested
the substrate specificity of CMBL is available at present: It specifically including humans, indicating negligible involvement of CES2, which
catalyzes ring-opening reaction of cyclic esters such as faropenem had been reported to be expressed in the intestine in mice, rats,
medoxomil and lenampicillin (Ishizuka et al., 2010), as well as OM monkeys, and humans (Taketani et al., 2007; Satoh and Hosokawa,
and AM. What seems to be lacking, as a further interest, is to define 2010). Contrary to CC hydrolysis, the recombinant human CES1
the endogenous substrates and to elucidate the physiologic functions showed slight OM- and AM-hydrolase activities, suggesting that
of CMBL proteins.
human CES1 may partially catalyze OM- and AM-hydrolysis in the
Also, we have shown that human CES1 rather than CES2 catalyzed liver. However, higher activities were observed in cytosolic fraction of
the bioactivation of CC using recombinant proteins. It has often been the liver rather than the microsomal fraction in which human CES1 is
reported that CC is converted to candesartan by carboxylesterases via mostly distributed (Fig. 2 for OM hydrolysis). This fact indicates that
intestinal and hepatic first-pass metabolism (Miwa et al., 1998; the CES1 contribution in the liver to OM and AM hydrolysis is
Easthope and Jarvis, 2002); however, which isozymes of CES1 or limited. CES2 showed only mock-control-level OM- and AM-
CES2 are a major contributor remained unknown until now. Although hydrolase activities. Three prodrug-type ARBs, namely OM, CC,
Laizure et al. (2013) supposed that CC is a CES2 substrate because of and AM, similarly require hydrolytic bioactivation to exert their
CC’s large alcohol group of the prodrug moiety that CES2 prefers as antihypertensive effect; however, needless to say, the responsible
a substrate, in the present study, the recombinant CES2 exhibited only hydrolases depend on the chemical structures of their prodrug moieties,
low CC-hydrolase activity in contrast to rapid hydrolysis by the OM and AM being substrates for CMBL while CC is a substrate for
recombinant CES1. Moreover, the intestinal subcellular fractions CES1.
TABLE 1
Correlation coefficients between protein expression levels of hydrolases and prodrug-hydrolase activities
Data from protein expressions and activities of eight different subcellular fractions (liver, intestine, kidney, and lung cytosolic and
microsomal fractions) were employed for each species. The protein levels were semiquantified as signal intensities of the immunoreactive
bands at the mobility (indicated with closed triangles in Fig. 7).
Correlation Coefficient
Protein
Species
OM-Hydrolase Activity
CC-Hydrolase Activity
AM-Hydrolase Activity
CMBL
Mouse
Rat
Monkey
Dog
Human
Mouse
Rat
Monkey
Dog
0.986
0.477
0.988
0.995
0.842
20.221
0.443
20.002
0.048
20.236
20.221
0.012
20.019
20.200
0.970
0.827
0.985
0.979
0.959
0.864
0.942
0.990
0.998
0.913
0.171
0.021
0.010
0.005
20.097
CES1
Human
20.167

Bioactivation of ARB Prodrugs by CMBL and CES1
1895
Iwamura A, Fukami T, Higuchi R, Nakajima M, and Yokoi T (2012) Human a/b hydrolase
domain containing 10 (ABHD10) is responsible enzyme for deglucuronidation of mycophe-
nolic acid acyl-glucuronide in liver. J Biol Chem 287:9240–9249.
Kawaguchi N, Ebihara T, Takeuchi T, Morohashi A, Yamasaki H, Tagawa Y, Takahashi J,
Kondo T, and Asahi S (2013) Absorption of TAK-491, a new angiotensin II receptor antag-
onist, in animals. Xenobiotica 43:182–192.
In conclusion, we comparatively demonstrated the difference in the
bioactivating properties of prodrug-type ARBs: OM is bioactivated
mainly via intestinal and additionally hepatic CMBL, therefore the
newly identified CMBL substrate AM is likely to be similarly
bioactivated to OM, while CC is via hepatic CES1 rather than in-
testinal enzymes. Also, we presented species differences in hydrolytic
activity and protein expression of the uninvestigated hydrolase CMBL
in four major organs, namely liver, intestine, kidney, and lung, by
comparison between humans and other animal species.
Kim I, Chu XY, Kim S, Provoda CJ, Lee KD, and Amidon GL (2003) Identification of a human
valacyclovirase: biphenyl hydrolase-like protein as valacyclovir hydrolase. J Biol Chem 278:
2
5348–25356.
Laeis P, Puchler K, and Kirch W (2001) The pharmacokinetic and metabolic profile of olmesartan
medoxomil limits the risk of clinically relevant drug interaction. J Hypertens Suppl 19:
S21–S32.
Lai L, Xu Z, Zhou J, Lee KD, and Amidon GL (2008) Molecular basis of prodrug activation by
human valacyclovirase, an alpha-amino acid ester hydrolase. J Biol Chem 283:9318–9327.
Laizure SC, Herring V, Hu Z, Witbrodt K, and Parker RB (2013) The role of human carbox-
ylesterases in drug metabolism: have we overlooked their importance? Pharmacotherapy 33:
Acknowledgments
2
10–222.
The authors thank Kaori Nakahara and Dr. Yuji Ogura for kindly providing
the recombinant human CES1 and CES2 proteins.
Liederer BM and Borchardt RT (2006) Enzymes involved in the bioconversion of ester-based
prodrugs. J Pharm Sci 95:1177–1195.
Miwa K, Murakami H, Masaki K, Iwasaki M, Yoshimura Y, and Motohashi M (1998) In vitro
hydrolysis of candesartan cilexetil to candesartan. J Hypertens 16 (Suppl 2):130.
Nakajima A, Fukami T, Kobayashi Y, Watanabe A, Nakajima M, and Yokoi T (2011) Human
arylacetamide deacetylase is responsible for deacetylation of rifamycins: rifampicin, rifabutin,
and rifapentine. Biochem Pharmacol 82:1747–1756.
Authorship Contributions
Participated in research design: Ishizuka, Yoshigae, Murayama.
Conducted experiments: Ishizuka.
Ngai KL, Schlömann M, Knackmuss HJ, and Ornston LN (1987) Dienelactone hydrolase from
Performed data analysis: Ishizuka.
Wrote or contributed to the writing of the manuscript: Ishizuka, Yoshigae,
Murayama, Izumi.
Pseudomonas sp. strain B13. J Bacteriol 169:699–703.
Nishikawa K, Naka T, Chatani F, and Yoshimura Y (1997) Candesartan cilexetil: a review of its
preclinical pharmacology. J Hum Hypertens 11 (Suppl 2):S9–S17.
Satoh T and Hosokawa M (2006) Structure, function and regulation of carboxylesterases. Chem
Biol Interact 162:195–211.
Satoh T and Hosokawa M (2010) Carboxylesterases: structure, function and polymorphism in
mammals. J Pestic Sci 35:218–228.
References
Schmidt E and Knackmuss HJ (1980) Chemical structure and biodegradability of halogenated
aromatic compounds. Conversion of chlorinated muconic acids into maleoylacetic acid. Bio-
chem J 192:339–347.
Taketani M, Shii M, Ohura K, Ninomiya S, and Imai T (2007) Carboxylesterase in the liver and
small intestine of experimental animals and human. Life Sci 81:924–932.
Taylor P (1991) The cholinesterases. J Biol Chem 266:4025–4028.
Taylor P and Radi ꢀc Z (1994) The cholinesterases: from genes to proteins. Annu Rev Pharmacol
Toxicol 34:281–320.
Watanabe A, Fukami T, Nakajima M, Takamiya M, Aoki Y, and Yokoi T (2009) Human
arylacetamide deacetylase is a principal enzyme in flutamide hydrolysis. Drug Metab Dispos
Atacand. (2000) Prescribing information. AstraZeneca, Wilmington, DE.
Berry LM, Wollenberg L, and Zhao Z (2009) Esterase activities in the blood, liver and intestine of
several preclinical species and humans. Drug Metab Lett 3:70–77.
de Gasparo M, Catt KJ, Inagami T, Wright JW, and Unger T (2000) International union of
pharmacology. XXIII. The angiotensin II receptors. Pharmacol Rev 52:415–472.
Draganov DI and La Du BN (2004) Pharmacogenetics of paraoxonases: a brief review. Naunyn
Schmiedebergs Arch Pharmacol 369:78–88.
Easthope SE and Jarvis B (2002) Candesartan cilexetil: an update of its use in essential hyper-
tension. Drugs 62:1253–1287.
Fukami T and Yokoi T (2012) The emerging role of human esterases. Drug Metab Pharmaco-
kinet 27:466–477.
Hosokawa M (2008) Structure and catalytic properties of carboxylesterase isozymes involved in
metabolic activation of prodrugs. Molecules 13:412–431.
Huttunen KM, Raunio H, and Rautio J (2011) Prodrugs—from serendipity to rational design.
Pharmacol Rev 63:750–771.
Ishizuka T, Fujimori I, Kato M, Noji-Sakikawa C, Saito M, Yoshigae Y, Kubota K, Kurihara A,
Izumi T, and Ikeda T, et al. (2010) Human carboxymethylenebutenolidase as a bioactivat-
ing hydrolase of olmesartan medoxomil in liver and intestine. J Biol Chem 285:11892–
3
7:1513–1520.
Watanabe A, Fukami T, Takahashi S, Kobayashi Y, Nakagawa N, Nakajima M, and Yokoi T
2010) Arylacetamide deacetylase is a determinant enzyme for the difference in hydrolase
(
activities of phenacetin and acetaminophen. Drug Metab Dispos 38:1532–1537.
Williams ET, Bacon JA, Bender DM, Lowinger JJ, Guo WK, Ehsani ME, Wang X, Wang H,
Qian YW, and Ruterbories KJ, et al. (2011) Characterization of the expression and activity of
carboxylesterases 1 and 2 from the beagle dog, cynomolgus monkey, and human. Drug Metab
Dispos 39:2305–2313.
1
1902.
Ishizuka T, Fujimori I, Nishida A, Sakurai H, Yoshigae Y, Nakahara K, Kurihara A, Ikeda T,
and Izumi T (2012) Paraoxonase as major bioactivating hydrolase for olmesartan
medoxomil in human blood circulation: molecular identification and contribution to plasma
metabolism. Drug Metab Dispos 40:374–380.
Ishizuka T, Rozehnal V, Fischer T, Kato A, Endo S, Yoshigae Y, Kurihara A, and Izumi T (2013)
Interindividual variability of carboxymethylenebutenolidase homolog, a novel olmesartan
medoxomil hydrolase, in the human liver and intestine. Drug Metab Dispos 41:1156–1162.
1
a
Address correspondence to: Dr. Tomoko Ishizuka, Drug Metabolism and
Pharmacokinetics Research Laboratories, Daiichi Sankyo Co., Ltd., 1-2-58
Hiromachi, Shinagawa-ku, Tokyo 140-8710, Japan. E-mail: ishizuka.tomoko.
xx@daiichisankyo.co.jp