Cleavage of tertiary amidomethyl ester prodrugs of carboxylic acids by rat liver homogenates

Article abstract of DOI:10.1016/S0928-0987(99)00057-3

The hydrolysis of tertiary amidomethyl ester prodrugs of carboxylic acids by rat liver homogenates is reported. Amidomethyl esters are rapidly and quantitatively converted to the corresponding acid and secondary amide. Reactivity is inversely dependent upon the molar refractivity and lipophilicity of the ester, as well as with steric bulk in the carboxylic acid moiety. In contrast to chemical and plasma hydrolyses, no dependence upon the pK(a) of the carboxylate leaving group was observed, nor was there any dependence upon the amide N-substituent. The rate of decomposition was inhibited by the carboxylesterase inhibitor eserine but not by the cytochrome P450 inhibitor SKF-525A, indicating the involvement of esterases in the hydrolysis reaction. These results indicate that amidomethyl esters may be expected to be readily cleaved in vivo. Copyright (C) 1999 Elsevier Science B.V.

Full text of DOI:10.1016/S0928-0987(99)00057-3

European Journal of Pharmaceutical Sciences 9 (1999) 201–205
www.elsevier.nl/locate/ejps
Cleavage of tertiary amidomethyl ester prodrugs of carboxylic acids by rat
liver homogenates
a
b ,
b
b
*
Jim Iley , Eduarda Mendes , Rui Moreira , Sofia Souza
a
Chemistry Department, The Open University, Milton Keynes, MK7 6AA, UK
b
CECF, Faculdade de Farm a´ cia, Avenida das For c¸ as Armadas, Universidade de Lisboa, 1600 Lisbon, Portugal
Received 22 April 1999; received in revised form 13 July 1999; accepted 20 July 1999
Abstract
The hydrolysis of tertiary amidomethyl ester prodrugs of carboxylic acids by rat liver homogenates is reported. Amidomethyl esters are
rapidly and quantitatively converted to the corresponding acid and secondary amide. Reactivity is inversely dependent upon the molar
refractivity and lipophilicity of the ester, as well as with steric bulk in the carboxylic acid moiety. In contrast to chemical and plasma
hydrolyses, no dependence upon the pK of the carboxylate leaving group was observed, nor was there any dependence upon the amide
a
N-substituent. The rate of decomposition was inhibited by the carboxylesterase inhibitor eserine but not by the cytochrome P450 inhibitor
SKF-525A, indicating the involvement of esterases in the hydrolysis reaction. These results indicate that amidomethyl esters may be
expected to be readily cleaved in vivo.
 1999 Published by Elsevier Science B .V . All rights reserved.
Keywords: Amidomethyl esters; Prodrugs; Carboxylic acids; Hydrolysis; Rat liver homogenates; Structure–metabolism relationships
1
. Introduction
medium is dependent upon (1) the pK of the carboxylate
a
leaving group, (2) the size and number of the a-sub-
stituents in the ester moiety and (3) the lipophilicity of the
compounds (Iley et al., 1997). However, plasma activation
is not particularly efficient for those compounds involving
sterically-hindered carboxylate moieties. For example, for
Many drugs containing a free carboxylic acid group
have their therapeutic effectiveness reduced upon oral
administration as a result of either physicochemical limita-
tions (e.g. several b-lactam antibiotics (Durckheimer et al.,
1
2
3
1
988), and leukotriene (Brooks and Summers, 1996) and
compound
1
(R 5Ph, R 5CH CO Et and R 5
2 2
fibrinogen (Stilz et al., 1996) antagonists) or serious side-
effects (e.g. cyclooxygenase-1 inhibitors (Laufer et al.,
(CH CH CH ) CH) the half-lives in human plasma and
3
2
2 2
phosphate buffer, pH 7.4, are 22 and 82 h, respectively.
This raises the problem of efficient systemic delivery of
the carboxylic acid parent drug from esters 1. Ideally, a
prodrug aimed for oral administration should be chemical-
ly stable in the gastrointestinal tract and be rapidly
activated after traversing the gastrointestinal barrier (Bund-
gaard, 1987).
1
994). The tertiary amidomethyl esters, 1, are potentially
useful pro-drugs of both the carboxylic acid (Moreira et
al., 1996; Iley et al., 1997) and secondary amide (Bund-
gaard and Rasmussen, 1991) moieties. In particular, those
1
compounds based on the ethyl hippurate carrier (R 5Ph,
2
R 5CH CO Et) and derived from carboxylic acids with
2
2
pK .4 and are very stable in aqueous buffers (Iley et al.,
Given the known ability of liver homogenates to cleave
esters and amides (Yoshikawa et al., 1995) the present
study was undertaken to assess the hydrolysis of
amidomethyl esters 1 in these systems and thereby to be
better able to predict the ability of these prodrugs to be
cleaved in vivo. The compounds used were chosen on the
a
1
997). In 80% human plasma the amidomethyl esters 1
hydrolyse with vastly different rates, and reactivity in this
*
Corresponding author. Tel.: 1351-1-794-6416; fax: 1351-1-796-470.
E-mail address: ermendes@ff.ul.pt (E. Mendes)
0
928-0987/99/$ – see front matter
 1999 Published by Elsevier Science B .V . All rights reserved.
PII: S0928-0987(99)00057-3

2
02
J. Iley et al. / European Journal of Pharmaceutical Sciences 9 (1999) 201–205
withdrawn, added to acetonitrile (500 ml) at 48C and
centrifuged at 13 000 rpm for 5 min. The supernatant was
analysed by HPLC for the presence of the substrate and
parent drug using the following system: a Shimdazu LC-9
isochratic HPLC system coupled to a Shimadzu SPD-M6A
diode-array detector; a Merck LiChrospher 100 RP-8 5 mm
1
2534 mm column; mobile phase, acetonitrile–water
containing 0.2 M sodium acetate buffer (55:45 to 70:30,
2
1
v/v); flow-rate, 1.0 ml min . Quantitation of the prodrug
and corresponding parent drug was obtained from mea-
surements of the peak areas in relation to those of
corresponding standards chromatographed under the same
conditions.
Fig. 1. Structures of carboxylic acids used in the present study.
2
.4. Effect of SKF-525A and eserine on decomposition
Each inhibitor was incubated for 30 min at 378C with
basis of the varying structure and pK values of the parent
a
carboxylic acid drug moieties (Fig. 1).
the liver homogenate complemented with the NADPH-
generating system prior to the addition of substrate.
Substrates were injected into the incubation mixture to
2
4
give an initial concentration of 2310 M. The remainder
of the experiment was carried out as described above.
2
.5. Computation
Molecular refractivities (MR) were calculated using the
2
. Experimental
software CHEMSKETCH v.3.0 (Advanced Chemistry Develop-
ment). Multiple linear regression was performed using the
software AJUSTE LINEAR MULTIPLO from Cecul (University of
Lisbon).
2
.1. Chemicals
Prodrugs were synthesised by reaction of the corre-
sponding carboxylic acid with the appropriate chloro-
methylamides (Moreira et al., 1994; Iley et al., 1997).
SKF-525A was a kind gift from SmithKline–Beecham
3
. Results and discussion
(
Portugal) and eserine was purchased from Sigma.
3
.1. Kinetics of hydrolysis
2
.2. Preparation of liver homogenates
For all the amidomethyl esters 1, hydrolysis by rat liver
homogenates gives rise to quantitative formation of the
parent carboxylic acid and corresponding secondary amide.
Reactions follow first-order kinetics (Fig. 2) and hydrol-
ysis is rapid with half-lives ranging from 4 min to ¯1 h.
This represents a substantial increase in reactivity when
compared to hydrolysis in human plasma (Iley et al.,
Four male Sprague–Dawley rats of 180–200 g body
weight were dosed with phenobarbital (80 mg/kg/day) for
days. On day 5, the animals were killed and their livers
rapidly removed, minced, homogenised with isotonic
phosphate buffer, pH 7.4, centrifuged at 700 g for 20 min
and the supernatant then pooled and stored at 2708C.
3
1
997). For example, the valproic acid esters 1d and 1l
hydrolyse, respectively, 120 and 50 times more rapidly in
the rat liver homogenate, while the ibuprofen and clofibric
acid derivatives 1a and 1e display a 20-fold increase in
reactivity. It is likely that such differences in prodrug
activation rates might reflect, to some extent, interspecies
differences. However, similar differences in hydrolytic
2
.3. Kinetics
An NADPH generating system (120 ml), comprising
.25 mM glucose-6-phosphate, 1.25 mM NADP , 6 mM
1
6
MgCl2 and 2.5 U/ml of glucose-6-phosphate dehydro-
genase, was added to a suspension of liver homogenate
1
reactivity have been reported previously for the N - and
3
(
300 ml) and the resulting mixture incubated at 378C for 5
N -pivaloyloxymethyl derivatives of cimetidine, 2 and 3,
1
min prior to the addition of substrate. Substrates (80 ml of
respectively, as well as for the N -pivaloyloxymethyl
a stock solution in acetonitrile) were added to give an
initial concentration of 2310 M. At appropriate inter-
vals, aliquots (50 ml) of the incubation mixtures were
derivative of 5-fluorouracil, 4, which exhibit efficient
hydrolysis in rat liver homogenate but slow hydrolysis in
human plasma (Buur et al., 1985; Buur and Bundgaard,
2
4

J. Iley et al. / European Journal of Pharmaceutical Sciences 9 (1999) 201–205
203
Fig. 2. Rate of hydrolysis of 1d in rat liver homogenate (j), in rat liver homogenate pre-treated with 2 mM eserine (d), and rat liver homogenate
pre-treated with 10 mM eserine (m), at 378C.
1
991). This behaviour contrasts with the high rates of
ibuprofen derivatives 1a and 1j all imply that the hippurate
hydrolysis in human plasma of the analogous N-acylox-
ymethyl esters derived from compounds such hydantoins
and phenyltetrazoles (Bundgaard, 1987; Alexander et al.,
carrier is not crucial to achieve rapid liver activation.
1
994), although for these cases no liver homogenate data is
available. This difference might reflect the fact that the
parent compounds of the nitrogen-containing fragment of
esters 1, 4 (a secondary amide) and 2/3 (imidazole) have
pKa values (for the formation of the anion) close to
¯
13–15 (Grimmett, 1979; Perrin et al., 1981), whereas
tetrazoles, imides, and related NH-compounds have pKa
values of ¯5–8 (Craig, 1990). In contrast, derivatives
1
i–k, which contain an N-methyl group, hydrolyse with
3.2. Structure–metabolism relationships
rates slower than in pH 7.4 buffers. Finally, the data for the
benzoate derivatives 1g, 1h and 1i, the valproate deriva-
tives 1d and 1l, the naproxen derivatives 1b and 1k and the
The rate data in Table 1 were analysed for quantitative
relationships with several physicochemical parameters,
Table 1
1
Lipophilicity and molar refractivity data and half-lives for the hydrolysis of amidomethyl prodrugs 1a–l (R 5Ph) in rat liver homogenates at 378C
2
3
a
3
2
21
Compound
R
R CO H
Log P
MR (cm )
10 k (min
)
t1 / 2 (min)
2
1
1
1
1
1
1
1
1
1
1
1
1
a
b
c
d
e
f
g
h
i
CH CO Et
Ibuprofen
Naproxen
3.86
3.39
4.17
3.85
2.73
3.00
1.63
3.62b
20.82
3.25
119.24
124.99
135.00
99.1
111.72
132.06
91.65
96.87
75.97
103.57
109.31
83.43
1.8960.30
1.7060.07
1.1860.27
6.1160.18
10.261.54
2.2960.11
20.560.14
17.060.42
14.760.20
2.1160.03
36.4
40.6
58.9
11.3
6.76
30.1
3.36
4.06
4.69
32.7
46.9
7.56
2
2
CH CO Et
2
2
CH CO Et
Diclofenac
Valproic acid
Clofibric acid
Probenecid
Benzoic acid
Benzoic acid
Benzoic acid
Ibuprofen
2
2
CH CO Et
2
2
CH CO Et
2
2
CH CO Et
2
2
CH CO Et
2
2
C H
6
5
Me
Me
Me
Me
j
k
l
Naproxen
Valproic acid
2.95
3.00
1.4710.10
9.1261.42
a
From Iley et al. (1997).
b
Calculated from log P50.711.9 log k9 (Iley et al., 1997), using a log k9 value of 20.80.

2
04
J. Iley et al. / European Journal of Pharmaceutical Sciences 9 (1999) 201–205
Table 2
Effect of the NADPH-generating system, eserine and SKF-525A on the reactivity of amidomethyl esters 1a and 1d in rat liver homogenates
2
21
Compound
Reaction
10 k (min
)
t1 / 2 (min)
1a
1a
1d
1d
1d
1d
1d
1d
Complete liver homogenate
Liver homogenate without NADPH-generating system
Complete liver homogenate
Liver homogenate without NADPH-generating system
Complete liver homogenate with 0.5 mM SKF525A
Complete liver homogenate with 1 mM eserine
Complete liver homogenate with 2 mM eserine
Complete liver homogenate with 10 mM eserine
1.89
1.84
6.11
5.04
4.30
2.58
a1.52
36.4
37.5
11.0
13.7
16.0
38.8
a45.6
a
No decomposition after 48 h.
including the molar refractivity (MR) of the prodrug, the
pK of the carboxylate leaving group and the log P values
log t1 / 2 5 (0.030460.0045)MR 2 (1.6560.25)
(
2)
2
a
n 5 4 s 5 0.016 r 5 1.00
(
Iley et al., 1997). No clear correlation between the rate
For compounds 1a–h, inclusion of the log P parameter
gave a slightly improved Eq. (3) (the MR and log P
parameters for the compounds in question are essentially
data encompassing all amidomethyl esters 1 and the
corresponding physicochemical parameters emerged. How-
ever, one marked difference between the hydrolysis of
compounds 1 by rat liver homogenate and either the
pH-independent hydrolysis or the decomposition in human
plasma is that the reactivity of 1 in liver homogenates
2
orthogonal, with a correlation coefficient of r 50.19).
(
Similar inclusion of log P into Eq. (2) is unwarranted
because (a) the data set is too small, and (b) there is
2
significant inter-correlation (r 50.52) between MR and
appears not to be dependent on the pK of the carboxylate
a
log P for these compounds).
leaving group. Indeed, while the clofibric acid ester 1e
(
pK 3) displays an 11-fold increase in reactivity over the
a
log t1 / 25(0.023560.0049)MR 1 (0.1560.10)log P 2 (1.9960.51)
valproic acid ester (pK 5) in human plasma, such differ-
(3)
a
2
n58 s50.194 r 50.89
ence is drastically reduced to a factor of ¯1.6 in liver
homogenate. A second notable difference between the liver
homogenate hydrolysis and the plasma hydrolysis is the
lack of any dependence of the liver homogenate reactions
on the N-substituent. Thus, compounds 1g–i all have
similar hydrolysis rates, as do the pairs of compounds 1a
and 1j, 1b and 1k, and 1d and 1l, although the N-
CH CO Et group and the N-CH group have different
Eqs. (1)–(3) indicate that the substrate volume, as mea-
sured by MR, plays a major role in the reactivity of
amidomethyl esters 1 in liver homogenates. The reduction
in reactivity with increasing MR can be ascribed to an
increase in steric hindrance in the ester moiety of 1. Eq.
(
3) also suggests that lipophilicity plays a role in the
2
2
3
hydrolysis of esters 1 by liver homogenates, though
marginal when compared to plasma hydrolysis. This is
consistent with our finding that cytochrome P450 enzymes
are probably not involved in the hydrolysis of 1 because
P450 usually requires lipophilic substrates (Iley and Con-
stantino, 1994). The differences in structure–metabolism
relationships between rat liver homogenates and human
plasma might be ascribed to different esterases catalysing
the ester cleavage.
electronic and steric properties (the Taft’s s* values are
.8 for N-CH CO Et and 0.0 for N-CH , while the
0
2
2
3
Charton’s n values are 0.58 and 0.52, respectively). In
plasma, rate differences vary by as much as a factor of 5
between compounds bearing an N-Me group and their
N-CH CO Et counterpart and by a factor of 3 beween the
N-Ph and N-CH CO Et analogues (Iley et al., 1997).
However, comparison of data presented in Table 1 with the
2
2
2
2
1
data for hydrolysis in aqueous buffers (Iley et al., 1997)
suggest that compounds 1 can be separated into two
classes. Esters 1a–h are hydrolysed in liver homogenates
much more rapidly than in pH 7.4 buffer, while the
N-methyl compounds 1i–l include derivatives that are
hydrolysed more slowly in liver homogenates than in pH
3.3. Effect of the NADPH-generating system, and of
esterase and cytochrome P450 inhibitors
The rate-enhancement observed with liver homogenates,
compared to hydrolysis in human plasma, in part could be
due to the action of cytochrome P450 enzymes
(Guengerich, 1987). This was tested by incubating com-
pounds 1a and 1d in the absence of the NADPH-generat-
ing system (which is an obligatory co-factor of P450). As
indicated in Table 2, there is a slight increase in the
observed t1 / 2 for 1a and 1d, but not enough to accommo-
date the possibility that cytochrome P450 enzymes are
7
.4 buffer, except for the valproic acid ester, 1l. Eqs. (1)
and (2) were obtained for the groups 1a–h and 1i–l,
respectively.
log t1 / 2 5 (0.026860.0049)MR 2 (1.8760.56)
(
1)
2
n 5 8 s 5 0.215 r 5 0.83
1
t1 / 2 for 1b in phosphate buffer, pH 7.4, at 378C is 3 min (Mendes, 1997).

J. Iley et al. / European Journal of Pharmaceutical Sciences 9 (1999) 201–205
205
Bundgaard, H., Rasmussen, G.J., 1991. Prodrugs of peptides. 11. Chemi-
cal and enzymatic hydrolysis kinetics of N-acyloxymethyl derivatives
of a peptide-like bond. Pharm. Res. 8, 1238–1242.
Buur, A., Bundgaard, H., Falch, E., 1985. Prodrugs of 5-fluorouracil. IV.
Hydrolysis kinetics, bioactivation and physicochemical properties of
various N-acyloxymethyl derivatives of 5-fluorouracil. Int. J. Pharm.
24, 43–60.
Buur, A., Bundgaard, H., 1991. Prodrugs of cimetidine with increased
lipophilicity: N-acyloxymethyl and N-alkoxycarbonyl derivatives.
Acta Pharm. Nord. 3, 51–56.
responsible for the bulk of the hydrolysis reactions.
Moreover, incubation of 1d with the cytochrome P450
inhibitor SKF-525A only led to a relatively small increase
in the t1 / 2 value to 16 min. To assess whether the rapid
ester activation can be ascribed to esterases, the effect of
eserine on the rate of hydrolysis was also studied. Eserine
is a known inhibitor of liver carboxylesterase (Junge,
1
984). A 4-fold increase in the half-life was observed upon
incubation with 2 mM eserine, and a complete inhibition
of hydrolysis was observed with 10 mM eserine (Fig. 2).
This strongly suggests that rat liver esterases catalyse the
hydrolysis of the amidomethyl esters 1. An identical
reduction in reactivity has been reported for the liver
esterase-catalysed hydrolysis of menahydroquinone-4 es-
ters in the presence of 1 mM eserine (Takata et al., 1995).
Craig, P.N., 1990. Drug compendium. In: Hansch, C. (Ed.), Comprehen-
sive Medicinal Chemistry. The Rational Design, Mechanistic Study
and Therapeutic Application of Chemical Compounds, Vol. 6, Per-
gamon Press, Oxford, pp. 237–991.
Durckheimer, W., Adam, F., Fischer, G., Kirrstetter, R., 1988. Recent
developments in the field of cephem antibiotics. In: Testa, B. (Ed.),
Advances in Drug Research, Vol. 17, Academic Press, London, pp.
6
1–634.
Grimmett, M.R., 1979. Diazoles, triazoles, tetrazoles, and their benzo-
analogues. In: Sammes, P.G. (Ed.), Comprehensive Organic Chemis-
try. The Synthesis and Reactions of Organic Compounds, Vol. 4,
Pergamon Press, Oxford, pp. 357–410.
Guengerich, F.P., 1987. Oxidative cleavage of carboxylic esters by
cytochrome P450. J. Biol. Chem. 262, 8459–8462.
3
.4. Implications for prodrug design
The present study indicates that rapid liver homogenate
hydrolysis of amidomethyl esters 1 can be obtained,
suggesting that these prodrugs may be expected to be
readily cleaved in vivo. Indeed, rapid liver activation, as
compared with plasma activation, suggests that it should
be possible to achieve systemic site-specific delivery of the
parent carboxylic acid. Rates of parent carboxylic acid
delivery are higher for smaller substrates and decrease with
increasing steric hindrance in the carboxylic acid moiety
and lipophilicity of the ester.
Iley, J., Constantino, L., 1994. The microsomal dealkylation of N,N-
dialkylbenzamides. Biochem. Pharmacol. 42, 275.
Iley, J., Moreira, R., Calheiros, T., Mendes, E., 1997. Acyloxymethyl as a
drug protecting group: Part 4. The hydrolysis of tertiary amidomethyl
ester prodrugs of carboxylic acid agents. Pharm. Res. 14, 1634–1639.
Junge, W., 1984. Carboxylesterase. In: Bergmeyer, H.U. (Ed.), Methods
of Enzymatic Analysis, Vol. IV, Verlag, Weinheim, pp. 2–8.
Laufer, S., Augustin, J., Dannhardt, G., Werner, K., 1994. (6,7-
Diaryldihydropyrrolidizin-5-yl)acetic acids, a novel class of potent
dual inhibitors of both cyclooxygenase and 5-lipoxygenase. J. Med.
Chem. 37, 1894–1897.
Mendes, E., 1997. Desenvolvimento de potenciais pr o´ -f a´ rmacos de a´ cidos
carbox ´ı licos, Ph.D. Thesis, University of Lisbon.
Acknowledgements
Moreira, R., Mendes, E., Calheiros, T., Bacelo, M.J., Iley, J., 1994. A
new direct synthesis of tertiary N-acyloxymethylamide prodrugs of
carboxylic acid drugs. Tetrahed. Lett. 35, 7107–7110.
We are grateful to Professor Luis Constantino for
supplying the rat liver homogenates. This work was
supported by JNICT (Portugal) under the contract PBIC/
SAU/1546/92.
Moreira, R., Calheiros, T., Cabrita, J., Mendes, E., Pimentel, M., Iley, J.,
1
996. Acyloxymethyl as a drug protecting group. Part 3. Tertiary
O-amidomethyl esters of penicillin G: chemical hydrolysis and anti-
bacterial activity. Pharm. Res. 13, 70–75.
Perrin, D.D., Dempsey, B., Serjeant, E.P., 1981. In: pKa Prediction for
Organic Acids and Bases, Chapman and Hall, London, p. 127, see
Appendix 2.
References
Alexander, J., Renyer, M.L., Rork, G.S., 1994. Investigation of N-
Stilz, H.U., Jablonka, B., Just, M., Knolle, J., Paulus, E.F., Zoller, G.,
1996. Discovery of an orally active non-peptide fibrinogen receptor
antagonist. J. Med. Chem. 39, 2118–2122.
[(acyloxy)alkyl] ester as a prodrug model for drugs containing the
phenyltetrazole moiety. J. Pharm. Sci. 83, 893–897.
Brooks, C.D.W., Summers, J.B., 1996. Modulators of leukotriene bio-
synthesis and receptor activation. J. Med. Chem. 39, 2629–2654.
Bundgaard, H., 1987. Design of bioreversible drug derivatives and utility
of the double prodrug concept. In: Roche, E.B. (Ed.), Bioreversible
Carriers in Drug Design: Theory and Application, Pergamon Press,
New York, pp. 13–94, and references cited therein.
Takata, J., Karube, Y., Hanada, M., Matsunaga, K., Matsushima, Y.,
Sendo, T., Aoyama, T., 1995. Vitamin K prodrugs: 1. Synthesis of
amino acid esters of menahydroquinone-4 and enzymatic reconversion
to an active form. Pharm. Res. 12, 18–23.
Yoshikawa, M., Nishiyama, S., Takaiti, O., 1995. Metabolism of dopa-
mine prodrug Docarpamine. Hypertens. Res. 18, 211–213.