Effects of steric hindrance and electron density of ester prodrugs on controlling the metabolic activation by human carboxylesterase

Article abstract of DOI:10.1016/j.dmpk.2021.100391

Carboxylesterase (CES) plays an important role in the hydrolysis metabolism of ester–type drugs and prodrugs. In this study, we investigated the change in the hydrolysis rate of hCE1 by focusing on the steric hindrance of the ester structure and the elect

Full text of DOI:10.1016/j.dmpk.2021.100391

Drug Metabolism and Pharmacokinetics 38 (2021) 100391
Contents lists available at ScienceDirect
Drug Metabolism and Pharmacokinetics
journal homepage: http://www.journals.elsevier.com/drug-metabolism-and-
pharmacokinetics
Effects of steric hindrance and electron density of ester prodrugs on
controlling the metabolic activation by human carboxylesterase
Masato Takahashi*, Ibuki Hirota, Tomoyuki Nakano, Tomoyuki Kotani, Daisuke Takani,
Kana Shiratori, Yura Choi, Masami Haba, Masakiyo Hosokawa
Faculty of Pharmacy, Chiba Institute of Science, 15-8, Shiomi-cho, Choshi, Chiba, 288-0025, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Carboxylesterase (CES) plays an important role in the hydrolysis metabolism of esteretype drugs and
prodrugs. In this study, we investigated the change in the hydrolysis rate of hCE1 by focusing on the
steric hindrance of the ester structure and the electron density. For 26 kinds of synthesized indomethacin
prodrugs, the hydrolytic rate was measured in the presence of human liver microsomes (HLM), human
small intestine microsomes (HIM), hCE1 and hCE2. The synthesized prodrugs were classified into three
types: an alkyl ester type that is specifically metabolized by hCE1, a phenyl ester type that is more easily
metabolized by hCE1 than by hCE2, and a carbonate ester type that is easily metabolized by both hCE1
and hCE2. The hydrolytic rate of 1-methylpentyl (hexane2eyl) ester was 10etimes lower than that of 4
emethylpentyl ester in hCE1 solution. hCE2 was susceptible to electron density of the substrate, and
there was a difference in the hydrolysis rate of up to 3.5etimes between p-bromophenyl ester and p-
acetylphenyl ester. By changing the steric hindrance and electron density of the alkoxy group, the factors
that change the hydrolysis rate by CES were elucidated.
Received 3 December 2020
Received in revised form
9 March 2021
Accepted 15 March 2021
Available online 25 March 2021
Keywords:
Carboxylesterase
Ester
Drug metabolism
Prodrug
Substrate specificity
© 2021 The Japanese Society for the Study of Xenobiotics. Published by Elsevier Ltd. All rights reserved.
1. Introduction
and metabolic activation of commercially available compounds have
demonstrated that hCE1 and hCE2 recognize different substrates
Carboxylesterase (CES) is an enzyme that catalyzes a hydrolysis
reaction that belongs to the phase I reaction. CES catalyzes the hy-
drolysis reactions of carboxylic acid derivatives to produce metabo-
lites such as carboxylic acid, alcohol, amine and thiol, which are
targets of the phase II reaction. CES is known to have a wide range of
substrate recognition capabilities and is involved in the detoxifica-
tion of pesticides, environmental chemicals and foods as well as
drugs [1,2]. In humans, human CES1 (hCE1), which is mainly
expressed in the liver and lung [3], and human CES2 (hCE2), which is
mainly expressed in the small intestine, liver and kidney [4],
contribute significantly to drug metabolism. Studies on metabolism
depending on the sizes of the acyl and alkoxy groups. Among the
commercially available compounds, methylphenidate (a psychotro-
pic drug) [5], meperidine (an analgesic drug) [6] and lidocaine (a
local anesthetic) [7], which have relatively small alkoxy groups, are
mainly metabolized by hCE1. On the other hand, heroin (a narcotic)
[8] and procaine (a local anesthetic) [9], which have relatively small
acyl groups, are mainly metabolized by hCE2. In a prodrug for which
the parent compound is a carboxylic acid compound, the alkoxy
group moiety as the modifying group is smaller than the acyl group.
Such substrates with small alkoxy groups tend to be easily recog-
nized by hCE1. Examples of hCE1 substrates are carboxylic acid-
based prodrugs such as temocapril, imidapril, alacepril (ACE inhibi-
tor) [10], oseltamivir (influenza drug) [11], indomethacin ester (anti-
inflammatory agent) [12], atorvastatin ester [13] and clopidogrel
(antiplatelet agent) [14]. On the other hand, in a prodrug for which
the parent compound is an alcohol compound, the acyl group as a
modifying group is smaller than the alkoxy group. Such substrates
with small acyl groups tend to be easily recognized by hCE2. Ex-
amples of hCE2 substrates are alcohol (or phenol)-based prodrugs
such as CPT-11 (anticancer drug) [15,16], prasugrel (antiplatelet drug)
[17] and haloperidol ester (antipsychotic drug) [18]. Thus, the
Abbreviations: ACE, angiotensin-converting-enzyme; BPHB, butyl p-hydrox-
ybenzoate; CES, carboxylesterase; DCC, dicyclohexylcarbodiimide; DMAP, dime-
thylaminopyridine; DMSO, dimethylsulfoxide; EDC, 1-(3-dimethylaminopropyl)-3-
ethylcarbodiimide; hCE, human carboxylesterase; HEPES, 4-(2-hydroxyethyl)-1-
piperazineethanesulfonic acid; HIM, human intestine microsomes; HLM, human
liver microsomes; HPLC, higheperformance liquid chromatography; IR, infrared
spectroscopy; NMR, nuclear magnetic resonance; ND, not detected; PNPA, p-
nitrophenyl acetate.
* Corresponding author.
E-mail address: matakahashi@cis.ac.jp (M. Takahashi).
https://doi.org/10.1016/j.dmpk.2021.100391
1347-4367/© 2021 The Japanese Society for the Study of Xenobiotics. Published by Elsevier Ltd. All rights reserved.

M. Takahashi, I. Hirota, T. Nakano et al.
Drug Metabolism and Pharmacokinetics 38 (2021) 100391
substrate recognition ability of hCE1 and hCE2 is characterized in
some substrates. However, it has not yet been investigated how
much the hydrolysis rate changes due to the difference in more
detailed structure and the difference in electron density of the ester.
In this study, we investigated the change in the hydrolysis rate of
hCE1 by focusing on the steric hindrance of the ester structure and
the electron density. Using indomethacin as a substrate, we syn-
thesized ester prodrugs with steric hindrance and alkoxy groups
with different electron densities, and we clarified the relationship
between these structures and the rate of hydrolysis.
2.3.2. General procedure for synthesis of 2ge2o, 3f and 4f
A mixture of indomethacin (1) (0.5 mmol), alcohol reagent
(0.75 mmol), EDC (0.75 mmol) and DMAP (0.05 mmol) in CH2Cl2
(5.0 mL) was stirred at room temperature under an argon atmo-
sphere for 1 day. After addition of 10% citric acid solution (5.0 mL),
the mixture was stirred for 5 min at the same temperature, and the
mixture was extracted with CH2Cl2 (20 mL). The organic solution
was dried over MgSO4 and evaporated under reduced pressure. The
residue was purified by column chromatography on SiO2 (ethyl
acetate/nehexane) to give an ester derivative 2.
2.3.3. General procedure for synthesis of 2t-2x
2. Materials and methods
A mixture of indomethacin (1) (0.5 mmol), alkyl halide reagent
(0.75 mmol), and Na2CO3 (1.5 mmol) in DMF (2.0 mL) was stirred at
room temperature under an argon atmosphere for 1 day. After
addition of H2O (20 mL), the mixture was extracted with hexane/
AcOEt (20 mL, 1/1). The organic solution was washed with H2O
(10 mL ꢀ 2) and brine (10 mL), dried over MgSO4 and evaporated
under reduced pressure. The residue was purified by column
chromatography on SiO2 (ethyl acetate/nehexane) to give an ester
derivative 2.
2.1. Reagents
Indomethacin was purchased from Tokyo Chemical Industry Co.,
Ltd. (Tokyo, Japan). Ester, amide and thioester derivatives were
synthesized by a condensing reaction using indomethacin (1). Butyl
p-hydroxybenzoate (BPHB) as the internal standard, alcohol re-
agents (2a: pentanol, 2b: hexan-2-ol, 2c: 2-methylpentanol, 2d: 3-
methylpentanol, 2e: 4-methylpentanol, 2f: phenol, 2g: 2-
methoxyphenol, 2h: 3-methoxyphenol, 2i: 4-methoxyphenol, 2j:
2-bromophenol, 2k: 3-bromophenol, 2l: 4-bromophenol, 2m: 20-
hydroxyacetphenone, 2n: 30-hydroxyacetphenone, 2o: 40-hydrox-
yacetphenone, 2p: 2-fluorophenol, 2q: 3-fluorophenol, 2r: 4-
fluorophenol, 2s: ethanol, 3f: aniline, 4f: thiophenol), alkyl halide
reagents (2t: chloromethyl pivalate, 2u: 3-bromophthalide, 2v: 1-
chloroethyl isopropyl carbonate, 2w: 1-chloroethyl cyclohexyl
carbonate, 2x: 4-chloromethyl-1,5-dimethyl-1,3-dioxol-2-one),
dicyclohexylcarbodiimide (DCC), 1-(3-dimethylaminopropyl)-3-
ethylcarbodiimide (EDC) and dimethylaminopyridine (DMAP)
were purchased from FUJIFILM Wako Pure Chemical Corporation
(Osaka, Japan). Human intestine microsomes (HIM) (7 doners,
mixed gender, 24e69 years old, Caucasian and African American),
human carboxylesterase 1b (hCES1b, hCE1) and hCE2 were pur-
chased from Corning inc. (MA, USA).
2.4. Enzyme reaction
2.4.1. p-Nitrophenyl acetate hydrolase activity
A typical incubation mixture (final volume of 1.0 mL) contained
10 mM Na,K phosphate buffer (pH 7.4) and 1 mM p-nitrophenyl
acetate (PNPA) (acetonitrile final concentration of 1.0% in the in-
cubation mixture) at 30 ꢁC. An enzyme source (0.05 mg/mL) was
added, and the PNPA hydrolase activity was measured spectro-
photometrically (detection at 405 nm) regarding the formation of
4-nitrophenol (ε ¼ 16400 Mꢂ1 cmꢂ1) at 30 ꢁC for 3 min, with data
obtained at 5-s intervals, using an Ultrospec 6300 pro (GE
Healthcare, formerly Amersham Bioscience, NJ, USA) [19].
2.4.2. General procedure for hydrolysis reaction
Solutions of indomethacin prodrugs (0.3125, 0.625, 1.25, 2.5, 5.0,
10 and 50 mM) were prepared in DMSO. A typical incubation
mixture (final volume of 100 mL) in 100 mM HEPES (pH 7.4) con-
2.2. Preparation of HLM samples
tained an enzyme (0.25 mg/mL) and 0.003125e0.5 mM of indo-
methacin prodrug (DMSO final concentration of 1.0% in the reaction
mixture). After 30emin incubation at 37 ꢁC, the reaction was
terminated by the addition of 0.3 mM BPHB in acetonitrile solution
Human liver microsomes (HLM) (single donor, male, 51 years
old, Caucasian) were prepared from human livers that were ob-
tained from the Human and Animal Bridging Research Organization
(HAB, Chiba, Japan), which is in partnership with the National
Disease Research Interchange (NDRI, Philadelphia, PA, USA). The
use of human livers was approved by the Ethics Committee of Chiba
Institute of Sciences (No. 22-1). The livers were homogenized and
centrifuged at 9,000 ꢀ G for 20 min at 4 ꢁC, and the supernatant
was ultra-centrifuged at 105,000 ꢀ G for 60 min at 4 ꢁC. The
microsome fraction was suspended in sucrose-EDTA-Tris (SET)
buffer (pH 7.4).
(100 mL) on an ice bath at 10 min. The mixture was centrifuged at
21,600 ꢀ G for 15 min at 4 ꢁC to precipitate the protein, and the
supernatant was subjected to HPLC analysis to determine
indomethacin.
2.5. Data analysis
2.5.1. Procedure for HPLC analysis
HPLC analyses were carried out using an LC solution (Shimadzu,
Kyoto, Japan) chromatographic system equipped with an LC-10AT
pump, SIL-20A auto sampler, CTO-10AS VP column oven, SPD-20A
UV/VIS detector, and SCL-10A VP system controller. Separations
2.3. Synthesis
2.3.1. General procedure for synthesis of 2ae2f, 2pe2r and 2s
A mixture of indomethacin (1) (0.5 mmol), alcohol reagent
(0.75 mmol), DCC (0.75 mmol) and DMAP (0.05 mmol) in CH2Cl2
(5.0 mL) was stirred at room temperature under an argon atmo-
sphere for 1 day. After addition of 10% citric acid solution (5.0 mL),
the mixture was stirred for 5 min at the same temperature, and the
mixture was extracted with CH2Cl2 (20 mL). The organic solution
was dried over MgSO4 and evaporated under reduced pressure. The
residue was purified by column chromatography on SiO2 (ethyl
acetate/nehexane) to give an ester derivative 2.
were performed on a column of Mightysil RP-18 GP 150-4.6 (5 mm)
(Cica-Reagent, Tokyo, Japan) at 30 ꢁC with detection at 254 nm.
Analytical HPLC was performed with elution at 1.0 mL/min using
MeOH with 0.1% H3PO4 aqueous solution ramped up from 65% to
90% over 25 min and then returned to 65% over 10 min. Analytical
HPLC was performed with BPHB as an internal standard
(tR ¼ 7.2 min). The enzyme activity was calculated from the area
ratio of indomethacin (tR ¼ 11.7 min) and BPHB. The Km and Vmax
values were calculated by the Michaelis-Menten equation using
nonlinear regression analysis with software (GraphPad Prism 7).
2

M. Takahashi, I. Hirota, T. Nakano et al.
Drug Metabolism and Pharmacokinetics 38 (2021) 100391
2.5.2. Statistical analysis
P < 0.01 was considered statistically significant in the Pearson
correlation coefficient test.
ester 2a; however, the hydrolysis rate increased as the methyl
group moved away from the ester, and a difference between the
hydrolysis rates of nepentyl ester 2a and 4emethylpentyl ester 2e
was not found. Among the aryl esters (2fe2r), phenyl ester 2f had
the lowest hydrolytic rate. The hydrolytic rate of m-acetylphenyl
ester 2n was 3.4etimes higher than that of phenyl ester 2f and
hydrolytic rate of o-fluorophenyl ester 2p was 4.3etimes higher
than that of phenyl ester 2f. Alkyl diesters (2t and 2u) showed
higher hydrolysis rates than those of alkyl monoesters (2ae2e, 2s).
Next, the hydrolytic rates of the prodrugs were investigated in HIM
solution (Fig. 3B). Aryl esters (2fe2r), alkyl diesters (2t and 2u) and
carbonate esters (2ve2x) had relatively high hydrolytic rates, but
alkyl esters (2ae2e, 2s) were hardly hydrolyzed in HIM solution.
Among the aryl esters (2fe2r), p-acetylphenyl ester 2o had a hy-
drolytic rate about 2.8etimes higher than that of phenyl ester 2f.
Based on the results shown in Fig. 3A and B, the HLM/HIM hydro-
lytic rate ratios were calculated (Fig. 3C). Alkyl esters (2ae2e, 2f),
which were hardly hydrolyzed in the HIM solution, had HLM/HIM
hydrolytic rate ratios of 140etimes or more, indicating that they
were HLMeselective. Aryl esters (2fe2r) also had high HLM/HIM
hydrolytic rate ratios as a whole, and only carbonate ester 2w had
an HLM/HIM ratio of less than 1.0.
3. Results
3.1. Synthesis
A total of 26 indomethacin prodrugs were synthesized. To
confirm the effects of steric hindrance, pentyl ester 2a and methyl
groupesubstituted pentyl esters (2be2e) were synthesized. To
confirm the effect of electron density, phenyl ester 2f and phenyl
ester in which the methoxy group, bromo group, acetyl group and
fluoro group were substituted at the oe, me, and pepositions,
respectively, were synthesized. Indomethacin prodrugs (2se2x)
were also synthesized with a modifying group frequently used in
commercially available prodrugs. Alkyl esters (2ae2e, 2s) and flu-
orophenyl esters (2pe2r) were synthesized using DCC, and aryl
esters (2fe2r), amide 3f and thioester 4f were synthesized using
EDC by reacting with the corresponding alcohol reagents (Fig. 1).
Pivoxil ester 2t, proxetyl ester 2v and medoxomil ester 2x were
synthesized using alkyl halides by the previously reported method
with slight modification [20] (Fig. 2). The other esters (2u and 2w)
were also synthesized using alkyl halides. The structures of the
synthesized compounds were determined by 1H-NMR, 13C-NMR,
and IR spectrum data.
3.3. Hydrolysis reaction in hCE1 and hCE2
Next, the hydrolytic rates of the indomethacin prodrugs were
investigated in hCE1 solution (Fig. 4A). Similar to the hydrolysis
reactions in HLM solution, all alkyl esters (2ae2e, 2se2u), aryl
esters (2fe2r) and carbonate esters (2ve2x) showed metabolic
activation to indomethacin. No hydrolysis was observed with
amide 3f, and the hydrolytic rate of thioester 4f was low. The hy-
drolytic rates of alkyl diesters (2t and 2u), aryl esters (2fe2r) and
carbonate esters (2ve2x) were high, and the hydrolysis rate of
carbonate ester 2v was the highest of all synthesized prodrugs. The
hydrolytic rate of 1-methylpentyl (hexane2eyl) ester 2b was
7.8etimes lower than that of n-pentyl ester 2a; however, that of 4-
methylpentyl ester 2e was 1.3etimes higher than that of n-pentyl
ester 2a. In aryl esters (2fe2r), the hydrolytic rates of methox-
yphenyl esters (2ge2i) and bromophenyl esters (2je2l) were lower
3.2. Hydrolysis reaction in HLM and HIM
First, the hydrolytic rates of the synthesized prodrugs were
investigated in HLM solution (Fig. 3A). Metabolic activation to
indomethacin was observed in all of alkyl esters (2ae2e, 2se2u),
aryl esters (2fe2r), and carbonate esters (2ve2x) in HLM solution.
The hydrolytic rates of pivoxil ester 2t and proxetyl ester 2v were
high: 23.6 and 23.4 nmol/mg protein/min, respectively. No hydro-
lysis was observed for the amide 3f, and the hydrolysis rate of the
thioester 4f was 38etimes lower than that of the ester 2f. In the
pentyl esters (2ae2e), the hydrolysis rate of 1emethylpentyl
(hexan-2-yl) ester 2b was 9.8etimes lower than that of nepentyl
Fig. 1. Synthesis of indomethacin ester prodrugs (2ae2s, 3f and 4f) using alcohol reagents.
3

M. Takahashi, I. Hirota, T. Nakano et al.
Drug Metabolism and Pharmacokinetics 38 (2021) 100391
Fig. 2. Synthesis of indomethacin ester prodrugs (2te2x) using alkyl halide reagents.
Fig. 4. Hydrolytic rates of indomethacin prodrugs (2ae2x, 3f, 4f) in the presence of
hCE1 (A) and hCE2 (B), and their hCE1/hCE2 ratios (C). Concentrations: prodrugs
(0.1 mM), hCE1 (0.25 mg/mL), hCE2 (0.25 mg/mL); reaction time: 30 min; tempera-
ture: 37 ꢁC; values are means S.D. (n ¼ 3). ND: not detected.
Fig. 3. Hydrolytic rates of indomethacin prodrugs (2ae2x, 3f, 4f) in the presence of
HLM (A) and HIM (B), and their HLM/HIM ratios (C). Concentrations: prodrugs
(0.5 mM), HLM (0.25 mg/mL), HIM (0.25 mg/mL); reaction time: 30 min; temperature:
37 ꢁC; values are means S.D. (n ¼ 3). ND: not detected.
1-methylpentyl ester 2b and amide 3f, which had low hydrolysis
rates in other enzyme solutions. In aryl esters (2fe2r), the hydro-
lytic rates of me and pemethoxyphenyl esters (2h and 2i) and me
and pebromophenyl esters (2k and 2l) were lower than that of the
phenyl ester 2f, and the hydrolytic rates of me and peacetylphenyl
esters (2n and 2o) and me and pefluorophenyl esters (2q and 2r)
were higher than that of the phenyl ester 2f. However, all
oesubstituted aryl esters (2g, 2j, 2m and 2p) had lower hydrolytic
rates than those of the other positional isomers. The hydrolytic rate
than that of the phenyl ester 2f, and the hydrolysis rates of ace-
tylphenyl esters (2ne2o) excluding the o-form 2m and fluo-
rophenyl esters (2pe2r) were higher than that of the phenyl ester
2f. Next, the hydrolytic rates of the indomethacin prodrugs were
investigated in hCE2 solution (Fig. 4B). As in HIM solution, the
hydrolytic rate of the alkyl ester was low and the hydrolytic rates of
the alkyl diesters (2t and 2u), aryl esters (2fe2r), and carbonate
esters (2ve2x) were relatively high. No hydrolysis was observed for
4

M. Takahashi, I. Hirota, T. Nakano et al.
Drug Metabolism and Pharmacokinetics 38 (2021) 100391
of oeacetylphenyl ester 2m was 7.3etimes lower than that of
peacetylphenyl ester 2o. Based on the results shown in Fig. 4A and
B, the hCE1/hCE2 hydrolytic rate ratios were calculated (Fig. 4C).
Monoalkyl esters (2ae2e, 2s) were hydrolyzed with high hCE1
selectively. Alkyl diesters (2t and 2u) and aryl esters (2fe2r) had
hCE1/hCE2 ratios of 1.0 or higher. In addition, the o-substituted aryl
esters (2g, 2j, 2m and 2p) had higher hCE1 selectivity than the m-
substituted and p-substituted aryl esters. Carbonate esters (2ve2x)
had hCE1/hCE2 ratios of 1.0 or less.
analysis with GraphPad Prism 7. The hydrolytic rate measured in
0.1 mM solution (Fig. 4A) and the Vmax value were correlated with
most substrates, but the Vmax values of 2f, 2t and 2v with high Km
values were different. Substrates with a large Vmax value also had a
large Km value. Next, the hydrolysis parameters of hCE2 were
calculated using prodrugs (2se2x) having a modifying group
frequently used in commercially available prodrugs. The Vmax
values of alkyl ester 2s and alkyl diesters (2t and 2u) were low and
those of carbonate esters (2ve2x) were as high as the Vmax values of
hCE1. Substrates with a large Vmax value also had a large Km value,
but alkyl ester 2s had a considerably large Km value despite its low
Vmax value (see Table 2).
3.4. Correlations among hydrolytic activities of the enzymes
There are several hydrolases other than hCE1 and hCE2 in HLM
and HIM [21]. In order to confirm whether the hydrolysis rate in
HLM or HIM solution is related to the hydrolysis rate in hCE1 or
hCE2 solution, the values shown in Figs. 3 and 4 were used for
investigation (Fig. 5). The data for hCE1 and HLM, which are highly
expressed in the human liver, were highly correlated (Fig. 5A),
suggesting that the hydrolysis reaction of the indomethacin pro-
drug in HLM solution is mainly promoted by hCE1. The correlation
between hCE2 and HLM was not higher than that of hCE1, because
the expression level of hCE2 was low in the human liver (Fig. 5B). A
correlation between HIM and hCE2 was also observed because
hCE2 is the main hydrolase expressed in the human small intestine
(Fig. 5C). Unfortunately, the results showed that there is a corre-
lation between HIM and hCE1, which is not expressed in the small
intestine (Fig. 5D). A correlation between hCE1 and hCE2 was also
confirmed, and it is thought that this result was obtained because
the overall reactivity of the synthesized indomethacin prodrugs
was similar between hCE1 and hCE2.
4. Discussion
As shown in the introduction, indomethacin is a carboxylic
acidecontaining drug, and indomethacin prodrugs have a structure
in which the acyl group is larger than the alkoxy group and are
easily hydrolyzed by hCE1. Therefore, it is thought that the reac-
tivity with hCE1 was high for all of the synthesized prodrugs. The
results showed that there are three types of synthesized prodrugs:
alkyl esters (2ae2e, 2se2u) that are specifically metabolized by
hCE1, aryl esters (2fe2r) that are more easily metabolized by hCE1
than by hCE2, and carbonate esters (2ve2x) that are easily
metabolized by both hCE1 and hCE2. Alkyl esters are further
divided into two types, monoalkyl esters (2ae2e, 2s) and alkyl di-
esters (2t and 2u), and it is the monoalkyl esters (2ae2e, 2s) that
are selectively hydrolyzed by hCE1. The electron densities of the
carbonyl carbons of the monoalkyl esters (2ae2e, 2s) are relatively
high, and it is thought that the hydrolytic rate changes due to a
simple difference in steric hindrance. It is thought that 1-
methylpentyl (hexan-2-yl) ester 2b is bulky near the ester and
that it is difficult for 2b to approach CESs. CESs are typical serine
hydrolases, and hydrolysis proceeds by a pingepong bibietype
reaction mechanism [22]. The first step is the formation of a
CESesubstrate complex by nucleophilic attack of serine on the
ester, and the second step is hydrolysis of the CESesubstrate
3.5. Hydrolytic parameters
The hydrolytic parameters of hCE1 were investigated using alkyl
esters (2ae2e, 2se2u), some aryl esters (2fe2o), and carbonate
esters (2ve2w) (Table 1). The Vmax and Km values were calculated
by the MichaeliseMenten equation using nonlinear regression
Fig. 5. Correlations among hydrolytic activities of the enzymes that were examined. A: HLM and hCE1, B: HLM and hCE2, C: HLM and hCE2, D: Pearson's correlation coefficients
among the studied enzymes. **P < 0.01.
5

M. Takahashi, I. Hirota, T. Nakano et al.
Drug Metabolism and Pharmacokinetics 38 (2021) 100391
Table 1
Kinetic parameters of indomethacin prodrugs (2ae2x) in hCE1 solutions. Prodrugs (2pe2r, 3f and 4f) were not tested. Concentrations: prodrugs (0.1, 0.05, 0.025, 0.0125,
0.00625, 0.003125 mM), hCE1 (0.25 mg/mL); reaction time: 30 min; temperature: 37 ꢁC; values are means S.E. (n ¼ 3).
Compound
Vmax (nmol/mg
protein/min)
Km
(m
M)
Compound
Vmax (nmol/mg
protein/min)
Km (mM)
2a
2b
2c
2d
2e
2f
2g
2h
2i
1.08 0.0297
0.146 0.0066
0.408 0.00607
0.976 0.0510
1.38 0.0540
6.70 0.469
3.85 0.248
4.51 0.363
4.20 0.212
2.58 0.121
2.60 0.100
4.23 0.526
6.54 1.15
4.42 0.293
8.71 1.62
7.29 1.08
70.9 9.20
20.2 3.61
20.6 4.56
18.9 2.69
13.3 1.96
12.7 1.56
2l
2.00 0.0410
2.29 0.125
6.10 0.313
5.44 0.363
2.37 0.137
6.65 0.829
4.42 0.204
12.8 2.15
11.1 0.754
13.9 2.35
29.4 3.73
24.2 4.23
12.5 2.32
45.7 12.2
30.5 3.44
81.8 24.4
32.4 4.85
26.4 6.60
2m
2n
2o
2s
2t
2u
2v
2w
2x
3.76 0.235
3.75 0.368
2j
2k
Table 2
Kinetic parameters of indomethacin prodrugs (2se2x) in hCE2 solutions. Prodrugs (2a-2r, 3f and 4f) were not tested. Concentrations: prodrugs (0.1, 0.05, 0.025, 0.0125,
0.00625, 0.003125 mM), hCE2 (0.25 mg/mL); reaction time: 30 min; temperature: 37 ꢁC; values are means S.E. (n ¼ 3).
Compound
Vmax (nmol/mg
protein/min)
Km
(
mM)
Compound
Vmax (nmol/mg
protein/min)
Km (mM)
2s
2t
2u
0.00955 0.00180
0.801 0.0281
1.55 0.0670
22.3 11.3
7.12 0.947
12.4 1.72
2v
2w
2x
14.3 1.46
4.24 0.239
4.87 0.267
101 17.1
25.8 3.75
27.2 3.77
complex by H2O. It is thought that the reaction rate of an ester with
a large steric hindrance is greatly reduced because the nucleophilic
attack of serine in the first step is difficult to proceed. Since the
hydrolysis rate increased as the methyl group moved away from the
ester (2be2e), it is possible that the steric hindrance caused by the
substituent at the 4eposition of the alkoxy group had almost no
effect. Alkyl diesters (2t and 2u) are known to have a different
mechanism of metabolic activation than that of monoalkyl esters
(2ae2e, 2s) [23,24]. After the ester on the modifying group side is
hydrolyzed, the resulting hemiacetal is autolyzed to produce an
indomethacin, corresponding to aldehyde (formaldehyde or acet-
aldehyde) and carboxylic acid (see Supplementary Fig. S1). There-
fore, it is thought that the selectivity of CES is controlled by the
ester structure on the modifying group side. The pivoxyl ester 2t
having high selectivity has a structure that is easily recognized by
hCE1 because the acyl group side is a bulky pivaloyl group and the
alkoxy group is an oxymethylene group with less steric hindrance.
However, the electron nucleophilicity of the oxygen atom near the
pivaloyl ester enhances the nucleophilicity of the ester. As a result,
it is thought that the reactivity of alkyl diesters (2t and 2u) was
higher than that of monoalkyl esters (2ae2e, 2s), and the selectivity
of hCE1/hCE2 was lowered. Therefore, it is thought that adding an
electron withdrawing group to an alkyl ester has the effect of
increasing the hydrolytic rate, while it may reduce the selectivity of
hCE1/hCE2. Aryl esters (2fe2r) have been found to have lower
nucleophilicity of phenoxide ions, which are hydrolysis products,
than that of alkoxide ions, which are hydrolysis products of alkyl
esters (2ae2e, 2se2u). This fact suggests that the reverse reaction
of the first step in the hydrolysis mechanism of CES is less likely to
occur. As a result, the hydrolysis rate of aryl esters (2fe2r) is
increased in CES solution. It is thought that the overall reactivity
was higher than that of alkyl esters (2ae2e, 2se2u) and the
selectivity of hCE1 was lowered. The effects of steric hindrance on
aryl esters have also been confirmed. In the hydrolysis reaction in
hCE1 solution, the hydrolytic rate of o-acetylphenyl ester 2m is
more than 2.5etimes lower than the hydrolytic rates of meform 2n
and peform 2o, and the hydrolytic rates of all oesubstituted esters
(2g, 2j, 2m and 2p) are lower than those of m-form and p-form in
hCE2 solution. It is thought that hCE2 is susceptible to steric hin-
drance of the substituent at the o-position, but hCE1 is also thought
to be affected when it becomes a large substituent such as an acetyl
group. It was found that the electron density of the ester affects the
hydrolysis rate of both hCE1 and hCE2. It is thought that hCE2 is
susceptible to electronic influence because the hydrolysis rates of
acetylphenyl esters (2me2o), which have strong electron with-
drawing properties, increased. Carbonate esters (2ve2x) are known
to have the same metabolic activation mechanism as that of alkyl
diesters [25] (see Supplementary Fig. S1). Since carbonate esters
cannot be classified into acyl groups and alkoxy groups, it is
thought that hCE1 and hCE2 cannot selectively recognize carbonate
esters (2ve2x). In general, carbonate esters (2ve2x) are known to
be less reactive for hydrolysis because they have a higher electron
density than that of esters. Nonetheless, carbonate esters (2ve2x)
had high hydrolytic rates in hCE1 and hCE2 solutions. It is thought
that this is because the hydrolytic rate of CES is greatly affected by
steric hindrance and the steric hindrance of carbonyl carbons of
carbonate esters (2ve2x) are lower than that of ester. Since the
hydrolysis rate of cyclic cilexetil 2w having a bulky cyclohexyl
group and medoxomil 2x decreased, it is thought that the hydro-
lysis rate is likely to decrease due to steric factors.
Summarizing the above, the following three points are impor-
tant new findings. First, even small alkyl groups such as methyl
groups have been found to be effective means of limiting the rate of
metabolism of hCE1 and hCE2. In particular, as is clear from the
results of compound (2be2e), it can be determined that it is
possible to efficiently limit the metabolic rate by adjusting the
steric hindrance near the ester. Second, the findings on the prop-
erties of alkyl and aryl esters have been clarified. For substrates
with relatively large acyl groups, such as indomethacin esters, the
use of small alkyl esters may be favorable as a prodrug that is not
metabolized in human small intestine or hCE2 but metabolized in
human liver or hCE1. Conversely, as long as the substrate has a large
acyl group, it is considered difficult to form an hCE2eselective ester
structure. Aryl esters reduce the selectivity of hCE1/hCE2 but in-
crease the reactivity. Aryl esters metabolize faster than alkyl esters,
but are unsuitable for maintaining hCE1/hCE2 selectivity. Third, it
6

M. Takahashi, I. Hirota, T. Nakano et al.
Drug Metabolism and Pharmacokinetics 38 (2021) 100391
J
Pharmacol Exp Therapeut 2004;310:469e76. https://doi.org/10.1124/
has been shown that the metabolic reaction by CES is not signifi-
cantly affected by the electron density of the ester. Carbonate esters
are chemically less reactive than carboxylic acid esters. However,
the metabolic rate of CES tends to be higher than that of carboxylic
acid esters. In addition, electron-withdrawing substitution ester
(2me2r) has a slightly higher hydrolysis rate, and the ortho-
substitution product is affected by steric hindrance and the meta-
bolic rate decreases. In predicting the rate of metabolism by CES, it
is necessary to consider the factor of steric hindrance rather than
the factor of electron dencity.
In this study, 26 kinds of indomethacin prodrugs were used to
elucidate the substrate recognition ability of carboxylesterases 1
and 2, which play the most important role in hydrolytic metabolism
in humans. By changing the steric hindrance and electron density of
the alkoxy group, the factors that change the hydrolysis rate by CES
were elucidated. Based on the results of this study, it is hoped that a
theoretical ester prodrug design will be carried out in consideration
of the substrate recognition ability of the enzyme.
jpet.104.067116.
[6] Zhang J, Burnell JC, Dumaual N, Bosron WF. Binding and hydrolysis of
meperidine by human liver carboxylesterase hCE-1. J Pharmacol Exp Ther-
apeut 1999;290:314e8.
[7] Higuchi R, Fukami T, Nakajima M, Yokoi T. Prilocaine- and Lidocaine-induced
methemoglobinemia is caused by human carboxylesterase-, CYP2E1-, and
CYP3A4-mediated metabolic activation. Drug Metab Dispos 2013;41:
1220e30. https://doi.org/10.1124/dmd.113.051714.
[8] Kamendulis LM, Brzezinski MR, Pindel EV, Bosron WF, Dean RA. Metabolism of
cocaine and heroin is catalyzed by the same human liver carboxylesterases.
J Pharmacol Exp Therapeut 1996;279:713e7.
[9] Jewell C, Ackermann C, Payne NA, Fate G, Voorman R, Williams FM. Specificity
of procaine and ester hydrolysis by human, minipig, and rat skin and liver.
Drug
dmd.107.015727.
Metab
Dispos
2007;35:2015e22.
https://doi.org/10.1124/
[10] Geshi E, Kimura T, Yoshimura M, Suzuki H. A single nucleotide polymorphism
in the carboxylesterase gene is associated with the responsiveness to imi-
dapril medication. Hypertens Res 2005;28:719e25. https://doi.org/10.1291/
hypres.28.719.
[11] Shi D, Yang J, Yang D, LeCluyse EL, Black C, You L, et al. Anti-influenza prodrug
oseltamivir is activated by carboxylesterase human carboxylesterase 1, and
the activation is inhibited by antiplatelet agent clopidogrel. J Pharmacol Exp
Therapeut 2006;319:1477e84. https://doi.org/10.1124/jpet.106.111807.
[12] Takahashi M, Ogawa T, Kashiwagi H, Fukushima F, Yoshitsugu M, Haba M,
et al. Chemical synthesis of an indomethacin ester prodrug and its metabolic
activation by human carboxylesterase 1. Bioorg Med Chem Lett 2018;28:
997e1000. https://doi.org/10.1016/j.bmcl.2018.02.035.
Author contributions
Conception and design of the study: M. Takahashi, M. Hosokawa.
Conducted experiments: M. Takahashi, I. Hirota, T. Nakano, T.
Kotani, D. Takani, K. Shiratori, Y. Choi.
[13] Mizoi K, Takahashi M, Sakai S, Ogihara T, Haba M, Hosokawa M. Structure-
activity relationship of atorvastatin derivatives for metabolic activation by
hydrolases.
Xenobiotica
2020;50:261e9.
https://doi.org/10.1080/
00498254.2019.1625083.
Analysis and/or interpretation of data: M. Takahashi, I. Hirota, T.
Nakano, T. Kotani, D. Takani, K. Shiratori, Y. Choi.
Drafting the manuscript: M. takahashi, M. Haba, M. Hosokawa.
Approval of the version of the manuscript to be published: All
authors.
[14] Kristensen KE, Zhu HJ, Wang X, Gislason GH, Torp-Pedersen C, Rasmussen HB,
et al. Clopidogrel bioactivation and risk of bleeding in patients cotreated with
angiotensin-converting enzyme inhibitors after myocardial infarction:
a
proof-of-concept study. Clin Pharmacol Ther 2014;96:713e22. https://
doi.org/10.1038/clpt.2014.183.
[15] Satoh T, Hosokawa M, Atsumi R, Suzuki W, Hakusui H, Nagai E. Metabolic
activation of CPT-11, 7-ethyl-10-[4-(1-piperidino)-1- piperidino]carbon-
yloxycamptothecin, a novel antitumor agent, by carboxylesterase. Biol Pharm
Bull 1994;17:662e4.
[16] Hatfield MJ, Tsurkan L, Garrett M, Shaver TM, Hyatt JL, Edwards CC, et al.
Organ-specific carboxylesterase profiling identifies the small intestine and
kidney as major contributors of activation of the anticancer prodrug CPT-11.
Declaration of competing interest
The authors declare that there is no conflict of interest.
Biochem
Pharmacol
2011;81:24e31.
https://doi.org/10.1016/
j.bcp.2010.09.001.
Acknowledgments
[17] Williams ET, Ponsler GD, Lowery SM, Perkins EJ, Wrighton SA, Ruterbories KJ,
et al. The biotransformation of prasugrel, a new thienopyridine prodrug, by
the human carboxylesterases 1 and 2. Drug Metab Dispos 2008;36:1227e32.
https://doi.org/10.1124/dmd.107.020248.
[18] Takahashi M, Uehara T, Nonaka M, Minagawa Y, Yamazaki R, Haba M, et al.
Synthesis and evaluation of haloperidol ester prodrugs metabolically acti-
vated by human carboxylesterase. Eur J Pharmaceut Sci 2019;132:125e31.
https://doi.org/10.1016/j.ejps.2019.03.009.
[19] Hosokawa M, Maki T, Satoh T. Multiplicity and regulation of hepatic micro-
somal carboxylesterases in rats. Mol Pharmacol 1987;31:579e84.
[20] Bandgar BP, Sarangdhar RJ, Viswakarma S, Ahamed FA. Synthesis and bio-
logical evaluation of orally active prodrugs of indomethacin. J Med Chem
2011;54:1191e201. https://doi.org/10.1021/jm101085j.
We thank the HAB Research Organization for providing human
livers. This work was supported by JSPS KAKENHI Grant Number
19K16421.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.dmpk.2021.100391.
[21] Fukami T, Yokoi T. The emerging role of human esterases. Drug Metabol
Pharmacokinet 2012;27:466e77. https://doi.org/10.2133/dmpk.DMPK-12-
RV-042.
References
~
[22] Aranda J, Cerqueira NMFSA, Fernandes PA, Roca M, Tunon I, Ramos MJ. The
[1] Hosokawa M. Structure and catalytic properties of carboxylesterase isozymes
involved in metabolic activation of prodrugs. Molecules 2008;13:412e31.
https://doi.org/10.3390/molecules13020412.
[2] Di L. The impact of carboxylesterases in drug metabolism and pharmacoki-
netics. Curr Drug Metabol 2019;20:91e102. https://doi.org/10.2174/
1389200219666180821094502.
[3] Imai T. Human carboxylesterase isozymes: catalytic properties and rational
drug design. Drug Metabol Pharmacokinet 2006;21:173e85. https://doi.org/
10.2133/dmpk.21.173.
[4] Xu G, Zhang W, Ma MK, McLeod HL. Human carboxylesterase 2 is commonly
expressed in tumor tissue and is correlated with activation of irinotecan. Clin
Canc Res 2002;8:2605e11.
catalytic mechanism of carboxylesterases: a computational study. Biochem-
istry 2014;53:5820e9. https://doi.org/10.1021/bi500934j.
[23] Sun X, Zeckner DJ, Current WL, Boyer R, McMillian C, Yumibe N, et al. N-
Acyloxymethyl carbamate linked prodrugs of pseudomycins are novel anti-
fungal agents. Bioorg Med Chem Lett 2001;11:1875e9. https://doi.org/
10.1016/S0960-894X(01)00333-X.
[24] Falch E. Salicylamide N-mannich bases as double prodrug forms for amines L-
7 1986;29:19e28.
[25] Huang B, Liu X, Tian Y, Kang D, Zhou Z, Daelemans D, et al. First discovery of a
potential carbonate prodrug of NNRTI drug candidate RDEA427 with sub-
micromolar inhibitory activity against HIV-1 K103N/Y181C double mutant
strain. Bioorg Med Chem Lett 2018;28:1348e51. https://doi.org/10.1016/
j.bmcl.2018.03.012.
[5] Sun Z, Murry DJ, Sanghani SP, Davis WI, Kedishvili NY, Zou Q, et al. Methyl-
phenidate is stereoselectively hydrolyzed by human carboxylesterase CES1A1.
7