Structure-activity relationship study of human liver microsomes-catalyzed hydrolysis rate of ester prodrugs of MENT by comparative molecular field analysis (CoMFA)

Article abstract of DOI:10.1016/S0039-128X(02)00186-1

A series of MENT esters (3-71) was designed, prepared and tested to study the structure-activity relationship (SAR) of the hydrolysis rate with human liver microsomes of these prodrugs. Compounds were obtained covering a wide range of metabolic stability. The results are useful for the proper selection of prodrugs for different pharmaceutical formulations to deliver the potent and prostate-sparing androgen MENT. The MENT esters can especially be administered for male hormone replacement therapy and male contraception. Comparative molecular field analysis (CoMFA) was applied to a dataset of 28 esters, for which ED50 values could be obtained. The CoMFA model where the electrostatic and H-bond molecular fields were combined turned out to be most predictive. Despite the limited size of the dataset, CoMFA can help to rationalize the SAR of the ester hydrolysis rate of ester prodrugs of MENT.

Full text of DOI:10.1016/S0039-128X(02)00186-1

Steroids 68 (2003) 213–220
Structure–activity relationship study of human liver
microsomes-catalyzed hydrolysis rate of ester prodrugs of MENT by
comparative molecular field analysis (CoMFA)
Roberta Bursi, Arijan Grootenhuis, Jaap van der Louw, Jos Verhagen,
∗
Marcel de Gooyer, Peter Jacobs, Dirk Leysen
N.V. Organon, Research & Development Laboratories, Molenstraat 110, PO Box 20, 5340 BH Oss, The Netherlands
Received 12 October 2001; received in revised form 18 July 2002; accepted 23 July 2002
Abstract
A series of MENT esters (3–71) was designed, prepared and tested to study the structure–activity relationship (SAR) of the hydrolysis
rate with human liver microsomes of these prodrugs. Compounds were obtained covering a wide range of metabolic stability. The results
are useful for the proper selection of prodrugs for different pharmaceutical formulations to deliver the potent and prostate-sparing androgen
MENT. The MENT esters can especially be administered for male hormone replacement therapy and male contraception. Comparative
molecular field analysis (CoMFA) was applied to a dataset of 28 esters, for which ED50 values could be obtained. The CoMFA model
where the electrostatic and H-bond molecular fields were combined turned out to be most predictive. Despite the limited size of the dataset,
CoMFA can help to rationalize the SAR of the ester hydrolysis rate of ester prodrugs of MENT.
©
2002 Elsevier Science Inc. All rights reserved.
Keywords: Steroid; MENT; Ester hydrolysis; 3D-QSAR; CoMFA
1
. Introduction
of spermatogenesis. Although in clinical male contraceptive
practice, complete suppression of spermatogenesis is not
possible with the currently available androgen formulations
alone [3–5], this can be achieved by the combination of an
androgen with a progestagen [6,7]. Finally, new indications
for androgens in men are emerging [8], whereas potential
applications in women are also receiving more and more
attention [9].
Despite these promising clinical results, the current andro-
gen regimes need improvement both in terms of the pharma-
cokinetic profile of the active drugs, as in terms of the avail-
able pharmaceutical formulations. Indeed, the fluctuations
in the serum levels obtained with most injectable androgens
should be avoided as these are experienced by patients as
being unpleasant due to the resulting alterations in mood
and behavior. In addition, different routes of administration
are required for optimal patient compliance. One of the ap-
proaches to obtain androgens, that combine high potency
with desired pharmacokinetic profiles, is by using prodrugs.
This is based on the fact that the physicochemical parameters
of a prodrug mainly determine the pharmacokinetic profile of
the active drug, by influencing the amount and rate of absorp-
tion of the prodrug, as well as the release rate of the active
drug. By properly adjusting these parameters, these prodrugs
The major clinical indications for androgens are male
hormone replacement therapy and male contraception. Ag-
ing in male is associated with a gradual decline in endocrine
function and, as a result, a steady decline in serum levels
of testosterone. On average, 20–25% of men above the age
of 60 have bioavailable testosterone levels below the lower
normal limit in young adults [1]. These men are classified as
being hypogonadal. Especially in hypogonadal men, testo-
sterone supplementation has shown favorable effects on
many of the symptoms co-determined by the age-associated
decrease in plasma testosterone levels [2]. In addition, an-
drogens can be adequately used for male contraception.
This hormonal contraceptive method is based upon the fact
that the gametogenic function of the testes is dependent on
a proper functioning of the hypothalamo–pituitary–testis
axis. Administration of slightly supraphysiological doses
of androgens results in a dose-dependent suppression of
pituitary gonadotropins and consequently, a decrease in in-
tratesticular testosterone levels and a reversible inhibition
∗
Corresponding author. Tel.: +31-412-661479; fax: +31-412-662546.
E-mail address: dirk.leysen@organon.com (D. Leysen).
0
039-128X/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved.
doi:10.1016/S0039-128X(02)00186-1

2
14
R. Bursi et al. / Steroids 68 (2003) 213–220
Table 1
◦
Structures of MENT esters 3–71 with percentage hydrolysis at 37 C (mean ± S.E.M. of N = 2) after in vitro incubation with human liver microsomes
at 15, 60 and 240 min and calculated ED50 in minutes based on T = 0, 15, 60 and 240 min
1
Ester
R
Percent hydrolysis
at 15 min
Percent hydrolysis
at 60 min
Percent hydrolysis
at 240 min
ED50 (min)
∗
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
[11]
–CH3
–CH2Ph
–CH2OPh
–CH2OCH2Ph
–CH2CH3
–CH2CH2Ph
–(E)–CH=CHPh
–(CH2)2CH3
–CH(CH3)2
–(E)–CH=CHCH3
–C(CH3)=CH2
–(CH2)3CH3
–CH2CH(CH3)2
–C(CH3)3
–(CH2)2CH=CH2
–CH=C(CH3)2
–(CH2)4CH3
–CH(CH2CH3)2
–CH2–t–Bu
-Cyclopentyl
–(CH2)5CH3
-Cyclohexyl
–(CH2)6CH3
–(CH2)2–cyclopentyl
–(CH2)7CH3
–(CH2)8CH3
–(E)–CH=CH(CH2)6CH3
–(CH2)9CH3
–(CH2)8CH=CH2
–CH(CH3)(CH2)7CH3
-trans-4-Butylcyclohexyl
-Adamantyl
–(CH2)10CH3
–(CH2)12CH3
–(CH2)14CH3
–(CH2)15CH3
–(CH2)16CH3
–(Z)-(CH2)7CH=CH(CH2)7CH3
Phenyl
2-Methylphenyl
3-Methylphenyl
4-Methylphenyl
4-Ethylphenyl
4-Propylphenyl
4-Butylphenyl
4-tert-Butylphenyl
4-Pentylphenyl
4-Hexylphenyl
4-Heptylphenyl
2-Methoxyphenyl
3-Methoxyphenyl
4-Methoxyphenyl
4-(Ethyloxy)phenyl
4-(Butyloxy)phenyl
4-(Hexyloxy)phenyl
4-(Heptyloxy)phenyl
2-Fluorophenyl
82.1 ± 19.9
90.7 ± 4.9
11.2 ± 0.7
87.7 ± 1.1
67.4 ± 1.3
74.6 ± 4.4
23.8 ± 1.5
5.6 ± 0.4
98.0 ± 3.7
30.7 ± 3.7
107.8 ± 3.3
102.8 ± 5.7
78.7 ± 3.3
63.8 ± 2.7
9.0 ± 0.3
2
528
21
35
<1
152
∗
∗
∗
∗
∗
32.8 ± 0.4
21.2 ± 3.3
72.7 ± 4.2
5.8 ± 1.4
∗
∗
∗
∗
∗
∗
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
62.1 ± 1.3
95.8 ± 3.0
8.4 ± 0.4
11.2 ± 0.3
24.1 ± 9.6
5.7 ± 0.1
96.2 ± 2.6
102.7 ± 3.5
14.9 ± 0.8
39.5 ± 1.4
80.7 ± 4.8
9.6 ± 0.3
102.0 ± 0.0
106.4 ± 2.4
33.3 ± 1.4
80.2 ± 1.1
101.3 ± 9.9
27.8 ± 1.3
10.2 ± 0.0
109.4 ± 0.3
28.5 ± 0.4
102.6 ± 11.3
6.7 ± 0.4
12
<1
680
82
28
782
1.4 ± 1.9
∗
∗
∗
56.8 ± 0.8
7.1 ± 0.4
18.0 ± 0.4
101.9 ± 6.2
17.3 ± 1.2
62.9 ± 4.3
2.3 ± 0.4
15
1284
40
5.8 ± 0.1
∗
∗
∗
∗
∗
∗
∗
∗
9.7 ± 0.3
11.0 ± 0.4
4.6 ± 0.2
9.4 ± 0.3
11.4 ± 1.1
9.8 ± 0.7
8.8 ± 0.8
6.4 ± 0.3
8.8 ± 0.4
10.0 ± 0.0
5.4 ± 1.2
21.6 ± 0.3
46.6 ± 1.0
13.6 ± 0.1
38.6 ± 0.3
39.5 ± 1.5
33.7 ± 0.8
25.1 ± 3.6
17.7 ± 0.2
20.0 ± 0.4
31.4 ± 2.0
10.4 ± 1.0
4.2 ± 0.0
6.1 ± 0.4
10.8 ± 0.1
5.2 ± 0.7
43.1 ± 1.1
86.1 ± 3.0
31.8 ± 0.4
84.6 ± 1.4
83.8 ± 5.2
76.5 ± 2.2
66.1 ± 3.3
45.9 ± 1.2
52.7 ± 0.1
77.2 ± 4.8
22.2 ± 0.6
8.6 ± 0.9
354
66
626
79
77
98
139
286
218
100
1879
∗
0 [12]
∗
1
2
3 [13]
4
5
11.9 ± 0.1
28.1 ± 0.5
9.5 ± 0.3
∗
5.0 ± 0.2
793
6
7
8
9
6.6 ± 0.8
4.6 ± 0.1
3.9 ± 0.7
1.7 ± 2.2
6.5 ± 0.2
6.2 ± 0.2
7.1 ± 0.3
0
6.5 ± 0.3
1 [14]
3.4 ± 0.1
18.6 ± 0.6
6.4 ± 0.3
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
4.4 ± 0.4
4.9 ± 1.3
4.2 ± 0.1
7.9 ± 0.1
9.5 ± 0.3
8.3 ± 0.2
6.5 ± 0.1
5.0 ± 1.0
4.5 ± 0.5
10.7 ± 1.1
8.3 ± 0.1
6.0 ± 0.0
3.9 ± 1.6
4.2 ± 0.8
7.0 ± 0.2
7.1 ± 1.0
4.7 ± 0.9
3.5 ± 0.0
4.5 ± 0.2
10.0 ± 0.7
7.5 ± 0.1
7.7 ± 0.8
6.3 ± 1.0
7.1 ± 0.2
5.5 ± 0.1
28.9 ± 1.5
4.2 ± 0.1
4.6 ± 0.5
12.3 ± 1.6
∗
9
6.5 ± 0.4
866

R. Bursi et al. / Steroids 68 (2003) 213–220
215
Table 1 (Continued )
Ester
R1
Percent hydrolysis
at 15 min
Percent hydrolysis
at 60 min
Percent hydrolysis
at 240 min
ED50 (min)
60
61
62
63
64
65
66
67
68
69
70
71
3-Fluorophenyl
4-Fluorophenyl
2-Chlorophenyl
3-Chlorophenyl
4-Chlorophenyl
2-(Trifluormethyl)phenyl
3,5-Bis(trifluoromethyl)phenyl
3-Cyanophenyl
3.0 ± 4.0
7.1 ± 0.2
6.1 ± 2.1
15.7 ± 1.5
12.7 ± 0.3
7.8 ± 1.0
10.6 ± 0.8
11.1 ± 0.3
7.7 ± 0.1
8.0 ± 0.6
14.2 ± 0.6
25.5 ± 1.3
6.3 ± 1.1
5.4 ± 1.0
5.1 ± 0.0
5.3 ± 0.5
6.7 ± 0.1
2.3 ± 3.1
5.2 ± 0.6
6.9 ± 0.3
9.7 ± 0.5
3.2 ± 0.3
∗
4-Cyanophenyl
4-Phenylphenyl
1-Naphthyl
2.0 ± 2.6
796
2-Naphthyl
Compounds marked with an asterisk were included in the CoMFA model.
Scheme 1.
should be suitable for different pharmaceutical formulations
to be used for oral-, sublingual-, nasal- and transdermal ap-
plications, as well as long acting injectables and implants.
The synthetic androgen 7␣-methyl-19-nortestosterone
measurements (HRMS) were carried out with electro-
spray ionisation using a Sciex QSTAR Pulsar quadrupole
time-of-flight (Q-TOF) mass spectrometer using reserpine
and testosterone as calibrant.
(MENT, 2) is a promising compound. MENT is at least 10
times more potent than testosterone, 1. In addition, it is not
potentiated by 5␣-reduction, suggesting prostate-sparing ef-
fects [10]. Unfortunately, the pharmacokinetic properties of
MENT are not compatible with once daily oral administra-
tion, nor with long-acting injectables. As part of a program
to develop potent and safe androgens with improved phar-
macokinetic properties, we examined a series of 17-O-esters
of MENT (3–71). In this study, we report the hydrolysis
rate with human liver microsomes of MENT esters 3–71,
given in Table 1, and describe the resulting SAR by means
of CoMFA (Scheme 1).
2.1.2. General procedure for the preparation of MENT
esters 3–71
MENT 2 (0.288 g, 1.0 mmol) was dissolved in pyridine
(6–7 ml) and an excess of acid chloride (1.5–2.0 mmol; neat
in case of solid acid chlorides, or dissolved in pyridine
(1–2 ml) in case of liquid acid chlorides) was added in one
portion. The reaction mixture was stirred at room tempera-
ture overnight. Ice was added and stirring was continued for
1 h. The product was extracted into ethyl acetate (50 ml) and
the organic phase was washed with 2 M HCl (two times),
water, 1 M NaOH (cold), and brine, dried over sodium sul-
fate and concentrated under reduced pressure. The crude
products were purified by column chromatography to pro-
duce the desired esters with a purity of at least 97%. The
structures of MENT esters 3–71 are collected in Table 1.
2
2
2
. Materials and methods
.1. Chemistry
2.2. Enzyme preparation
.1.1. General
Melting points (m.p.) were determined on a Büchi 535
A mixed pool of human liver microsomes (pool from
melting point apparatus and are uncorrected. NMR spectra
were recorded in CDCl3 on a Varian Gemini 200 spectrome-
ter (200 MHz), a Varian VXR-300 spectrometer (300 MHz),
or on a Bruker Avance 400 spectrometer (400 MHz) us-
ing tetramethylsilane as internal standard. Accurate mass
15 different subjects, In vitro Technology (Santa Barbara,
USA)) were used for the in vitro incubation with the dif-
ferent MENT esters. The stock solution (20 mg protein/ml,
◦
−70 C) of microsomes were thawed on melting ice, diluted
to 75 ␮g/ml with assay buffer. The assay buffer was prepared

2
16
R. Bursi et al. / Steroids 68 (2003) 213–220
by the following procedure from a pool of human plasma
EDTA). Plasma was incubated for 1 h at room tempera-
ture with an excess of dihydrotestosterone and hydrocor-
tisone (for the saturation of sex hormone-binding globulin
correspond to the lowest energy conformation. The com-
pound structures were first minimized by means of the
Tripos force field [16] and at the equilibrium geometries
semi-empirical AM1 [17] partial charges were calculated to
account for electrostatics. The well-known alignment prob-
lem of grid-based QSAR techniques [18–20] is to a great
extent reduced when compounds like steroids are employed
because of the reasonably rigid structure of the steroidal
skeleton. In this study, the B and C rings were used as
template for steroid superposition. Standard steric and elec-
trostatic CoMFA fields as well as other field classes such as
H-bond [21], indicator and parabolic fields were obtained.
The indicator and parabolic fields however yielded no sig-
nificant statistical models and, therefore, they have not been
further included in this study. Choosing 1 Å grid distance
ensured translational and rotational invariance of statistical
results. No smoothing and a distance-dependent dielectric
constant were applied to field values. Smooth transitions
were chosen between the cut-off (30 kcal/mol) plateaus for
steric and electrostatic interaction calculations.
(
(SHBG)), and cleared with a solution (1:1) of dextran-coated
charcoal (15 min at room temperature followed by 5 min
5
000 × g). The supernatant (50% plasma) was diluted with
phosphate-buffered saline (5 mM with 130 mM NaCl con-
taining 0.05% (v/v) Tween 20) resulting in an assay buffer
containing 25% human serum in PBS buffer with 0.025%
Tween 20.
The use of an assay buffer containing 25% (v/v) human
serum proteins containing 0.025% (v/v) Tween 20 was the
best buffer for the solubilization of the different esters (with
a large range of clogP). All compounds dissolved well, after
centrifugation the same amount of MENT ester was present
in the supernatant.
2
.3. Determination of ester hydrolysis
In vitro ester hydrolysis was determined in duplicate by
Several CoMFA models were derived first by using the
steric, electrostatic and H-bond fields alone or in combina-
tion for the whole dataset of 28 compounds. Further, the
dataset was split in training and test sets consisting of 25
and 3 compounds, respectively. The latter compounds were
59 (ED50 = 866 min), 26 (ED50 = 77 min) and 3 (ED50 =
2 min). The only condition the test set was required to fulfill
was some spread in ED50 values. The attempt of reducing
the training set to 20 compounds to enlarge the test set to 8
compounds did not yield statistically significant models for
any field or combination of fields.
incubation of 5 ␮l of 2 mM MENT esters (in 100% ethanol)
◦
with 500 ␮l human liver microsomes at 37 C. At T = 0, 15,
6
0 and 240 min, respectively, 100 ␮l sample was removed,
hydrolysis was stopped with 500 ␮l ice-cold ethanol and
centrifuged at 3500 × g for 10 min. The supernatant was
dried in a speed vac dryer, the pellet was dissolved in 100 ␮l
00% ethanol, of which 10 ␮l was injected on a straight
phase HPLC column (Lichrosphere 100 CN 5 ␮M, length
50 mm, internal diameter 4.6 mm) and separated with hep-
1
1
tane:ethanol (85:15% v/v) with a flowrate of 1.5 ml/min. The
separated MENT ester and MENT were detected at 240 nM,
the amount of released MENT was calculated on the basis
of integrated peak area on a standard curve of MENT.
The percentage hydrolysis at the various time points was
calculated by comparing the hydrolysis of the different
MENT esters (20 ␮M) with the sample of 20 ␮M MENT
itself. If applicable with the 4 time points an ED50 (time
in minutes at which 50% of the MENT ester is hydrolyzed)
were calculated with 4-parameter fit model. Without added
microsomes no hydrolysis occurred up to 240 min of incu-
Intrinsic correlations’ chances were checked in all mod-
els by repeating cross-validation on “randomized” sets of
2
activities. In no case, the resulting q values—although with
large fluctuations—were comparable with the “correct” one.
3. Results and discussion
3.1. Rate of ester hydrolysis
◦
bation at 37 C.
3.1.1. MENT esters derived from aliphatic carboxylic acids
2
.4. CoMFA model
3.1.1.1. Chain length. Stability of MENT esters derived
from linear aliphatic carboxylic acids increases stepwise
with increasing chain length. A sharp inflection point was
found between C11 and C14, from 52.7 to 9.5% hydrolysis
at T = 240 min (see Fig. 1).
Comparative molecular field analysis [15] was applied to
a dataset of 28 MENT esters, for which ED50 values were
determined. Compounds included in the analysis are marked
in Table 1. Experimental data of this set of compounds are
spread over more than 3 units (0 ≥ ED50 (min) ≤ 1284).
By the set-up of the database, care was taken that the long
and flexible substituents on position 17 were consistently
built. Energy minima searches were performed to identify
the lowest energy conformation for the flexible chains.
The conformation where the keto-group of the ester moi-
ety points downward the steroidal skeleton turned out to
3.1.1.2. Unsaturation. Unsaturation in the ␣-position gen-
erally results in higher stability: compare, e.g. esters 12 with
10, 18 with 15, and 29 with 28 (MENT esters derived from
aliphatic carboxylic acids), and 41 with 24 (MENT ben-
zoate versus MENT cyclohexyl carboxylate). Unsaturation
at other positions seems to decrease stability: compare esters
17 with 14, and 31 with 30.

R. Bursi et al. / Steroids 68 (2003) 213–220
217
Fig. 1. Effect of increasing chain length of ester on percent hydrolyses after 240 min incubation and calculated ED50 in minutes after incubation with
human liver microsomes.
3
.1.1.3. Substitution. Alkyl substitution at the ␣-position
increases stability: compare, e.g. esters 16 with 11 or 7,
0 with 10, and 32 with 28. The same applies for alkyl
substitution at other positions: compare 15 with 10, and
1 with 19. In both cases gain in stability cannot be ex-
ortho-Fluorobenzoate and para-cyanobenzoate 59 and 68,
respectively, are less stable than MENT benzoate.
1- and 2-Naphthyl carboxylates 70 and 71 are more stable
than MENT benzoate. There is little difference in stability
between these isomers.
2
2
plained by homologation alone. ␣-Substitution of esters de-
rived from unsaturated carboxylic acids seems to reduce sta-
bility: compare 13 with 12. Introduction of a phenyl group
in the ␤-position seems to increase stability; compare 8
with 26.
3
.2. CoMFA model
The results of the CoMFA analyses are given in Table 2.
When looking at the whole dataset, all models yielded good
cross-validated correlation coefficients, q , and the best re-
sults were obtained when fields were used in combination.
However, the three-fields combination did not result in
the best statistics. This might well be due to some intrinsic
correlation between fields.
2
3
.1.1.4. Introduction of oxygen atoms. Introduction of
oxygen atoms into the chain of the carboxylic acids reduces
stability: compare phenoxy- and benzyloxy-substituted ac-
etates 5 and 6, respectively, with phenylacetate 4.
3
.1.1.5. Reproducibility of hydrolysis. The reproducibility
of the hydrolysis rate was assessed by the hydrolysis of
MENT undecanoate (N = 4) which was run as control on
separate days of analysis.
Table 2
CoMFA results on the complete dataset (28 compounds) and the training
set (25 compounds)
The average ED50 of MENT undecanoate is 200 min with
a coefficient of variation (CV) of 5.0%.
q2
#LV r2
2
pred
Fields
s
s
r
28 compounds
ster
elec
H-bond
0.465 0.692
0.448 0.674
0.432 0.698
0.472 0.687
0.607 0.607
0.642 0.609
4
2
3
4
5
5
5
0.952 0.207
0.841 0.362
0.829 0.384
0.970 0.165
0.982 0.129
0.981 0.132
0.984 0.124
–
–
–
–
–
–
–
3
.1.2. MENT esters derived from aromatic carboxylic acids
In general MENT esters derived from aromatic car-
boxylic acids are much more stable towards hydrolysis than
those derived from aliphatic acids. Benzoate esters in which
the phenyl ring is substituted with alkyl, alkoxy, chloro,
trifluoromethyl, and phenyl are more stable than MENT
benzoate itself. With methyl- or chloro-substituted ben-
zoates there is almost no difference among ortho-, meta-,
or para-substituted derivatives, which is in contrast to that
observed with fluoro- and cyano-substituted benzoates.
Homologation of alkyl or alkoxy in para-position has lit-
tle influence on stability. Stability of ortho-trifluoromethyl
benzoate 65 is comparable with ortho-methyl-substituted
ester 42. Most stable is MENT 2-methoxybenzoate 52.
ster + elec
ster + H-bond
elec + H-bond
ster + elec + H-bond 0.609 0.605
2
5 compounds
ster
elec
H-bond
0.249 0.796
0.314 0.742
0.385 0.703
0.327 0.735
0.447 0.667
0.491 0.655
4
3
3
3
3
4
3
0.957 0.190 0.987
0.952 0.196 0.974
0.850 0.348 0.728
0.926 0.243 0.901
0.864 0.331 0.699
0.964 0.175 0.929
0.893 0.293 0.799
ster + elec
ster + H-bond
elec + H-bond
ster + elec + H-bond 0.480 0.646
Grid distances of 1 Å and AM1 partial charges were employed.

2
18
R. Bursi et al. / Steroids 68 (2003) 213–220
An objective assessment of the predictive capability of
all kinds, e.g. from microorganisms to vertebrates [22]. They
are also widely expressed in humans: stomach, small intes-
tine, colon, liver, plasma, brain, macrophages and mono-
cytes [23]. Carboxylic ester hydrolases (EC 3.1.1) efficiently
catalyze the hydrolysis of a variety of esters into their re-
spective free acids. These enzymes exhibit broad and over-
lapping substrate specificity and the same substrate is often
hydrolyzed by more than one enzyme [24]. Especially the
carboxylesterases (EC 3.1.1.1) are major determinants of the
pharmacokinetic behavior of a vast number of structurally
diverse drugs, containing ester or amide bonds. Recently, it
was proposed [23] that the mechanism for hydrolysis by car-
a model is established via validation by external datasets.
Therefore, the whole dataset was split into training and test
sets. Three compounds covering small, intermediate and
large ED50 values were chosen to represent the test set. New
models were derived. Statistics on the training set changed
significantly for all models, clearly due to the relatively lim-
ited number of the compounds present. Combinations of
fields still yielded the best models, with the combination of
steric and electrostatic fields being the only exception. The
2
best final r values were obtained with the individual steric
and electrostatic fields and combinations of steric and elec-
trostatic, and electrostatic and H-bond fields.
boxylesterases involves Ser2 , Glu
03
335
and His
448
residues
Further, although within this particular study the test set
was limited in number, predictive r values were calculated.
in a catalytic triad. This complex facilitates the acylation of
Ser by a general base mechanism, whereas Gly –Gly
2
203
123
124
When looking only at these values, the best models were
provided by individual steric and electrostatic fields and the
combinations of steric and electrostatic and electrostatic and
H-bond fields.
form an oxyanion hole, where hydrogen bonds with the car-
bonyl oxygen stabilize the tetrahedral intermediate.
If we turn to our study, from the CoMFA contours of
the steric fields displayed in Fig. 2, it is clear that around
the ester group there is a strong restriction for steric bulk.
This is in line with the mechanism cited above [22]. More-
over, at a distance of 4–5 Å from the carbonyl group, steric
bulk seems to be tolerated, whereas the length of the R sub-
stituent in compounds 3–71 is limited. The contours of the
model obtained with the electrostatic field, given in Fig. 3,
strongly suggest dipole–dipole interactions with the enzyme.
The electrostatic field seems to be more descriptive for the
chain and, in particular, the type of chain adjacent to the
ester group. In addition, a strong dipole is present close
to the ester group. Finally, the model obtained with the
H-bond field is also in line with the proposed mechanism.
The contours (Fig. 4) clearly suggest the possibility of form-
ing several H-bonds (acceptor-field type in the ligands and
2
When the whole statistical profile of these models, e.g. q ,
2
2
r and r , is considered against the CoMFA model, where
pred
the electrostatic and H-bond fields were used in combination,
2
2
resulted to be the best one (q = 0.491, #LV = 4, r =
0
.964, r2 = 0.929).
pred
By looking at these results, it must be borne in mind
that useful statistical models could still be derived on such
small-sized dataset because of the application of a powerful
approach (CoMFA) on a highly congeneric series.
3
.3. General discussion
Many proteins capable of hydrolyzing esters are widely
spread in nature. They are present in biological material of
Fig. 2. CoMFA contours around MENT ester 18. The model includes the whole dataset of 28 compounds. The yellow and green contours describe the
favorable and unfavorable steric interactions, respectively.

R. Bursi et al. / Steroids 68 (2003) 213–220
219
Fig. 3. CoMFA contours around MENT ester 18. The model includes the whole dataset of 28 compounds. The red and blue contours describe the
electrostatic regions, which are favorable and unfavorable to a negative charge, respectively.
Fig. 4. CoMFA contours around MENT ester 18. The model includes the whole dataset of 28 compounds. Only H-bond acceptor-type field contributions
(green/favorable and yellow/unfavorable) do contribute to the analysis.
complementary donor-field type in the putative enzyme)
around the ester function. In the model, two of these are lo-
cated around the carbonyl oxygen, approximately located at
the end of the lone pairs of this oxygen. These most prob-
plication in combination with the desired pharmacokinetic
profile.
The CoMFA models showed that the steric, electrostatic
and H-bond fields can significantly describe the rate of ester
hydrolysis of our dataset of 28 compounds. The contours
of these models provide valuable information on the active
site of the carboxylesterases involved. This information is
in line with the recently proposed mechanism for this class
of enzymes.
ably represent the H-bonds formed with Gly1 –Gly
23
124
in
the oxyanion hole of the enzyme.
Certainly more information on the active site of the car-
boxylic ester hydrolases, especially the carboxylesterases,
will become available in the near future. At that time, it will
be possible to validate the observations suggested by these
results, while for the time being the data described above—
although approximated—could be useful to further refine
the models of the active site of these enzymes.
Acknowledgments
We thank Syncom BV, Groningen, The Netherlands, for
preparing the MENT esters 3–71.
4
. Conclusions
References
A series of MENT esters covering a wide range of physi-
cochemical properties and metabolic stabilities towards hu-
man liver microsomes was obtained. These data allow the
selection of the proper prodrug for different routes of ap-
[
[
1] Vermeulen A. Andropauze. Maturitas 2000;34:5–15.
2] Tenover JL. Testosterone replacement therapy in older adult men.
Int J Androl 1999;22:300–6.

2
20
R. Bursi et al. / Steroids 68 (2003) 213–220
[
[
[
[
3] Anonymous. Contraceptive efficacy of testosterone-induced [13] Leysen D, Van der Voort HAA, Van der Louw J, inventors;
azoospermia in normal men. World Health Organization Task
Force on methods for the regulation of male fertility. Lancet
Akzo Nobel N.V., The Netherlands, assignee. Preparation of
cycloalkyl-carboxylic acid esters of 7␣-methylestr-4-en-3-one 17␤-ol
(19-nor-7␣-methyltestosterone). EP 1998-202051; June 19, 1998.
[14] Bolt CC, De Jongh HP, Siegmann CM, Van Vliet NP, Zeelen FJ.
C17-Substituted 5- and 5(10)-estrenes and their 7␣-methyl homologs.
Recl Trav Chim Pays-Bas 1969;88:1061–9.
[15] Cramer RD, Patterson DE, Bunce JD. Comparative molecular field
analysis (CoMFA). 1. Effect of shape on binding of steroids to carrier
proteins. J Am Chem Soc 1988;110:5959–67.
1990;336:955–9.
4] Anonymous. Comparison of two androgens plus depot-medroxy-
progesterone acetate for suppression to azoospermia in Indonesian
men. World Health Organization Task Force on methods for the
regulation of male fertility. Fertil Steril 1993;60:1062–8.
5] Handelsman DJ, Farley TM, Peregoudov A, Waites GGGM. Factors
in nonuniform induction of azoospermia by testosterone enanthate
in normal men. World Health Organization Task Force on methods
for the regulation of male fertility. Fertil Steril 1995;63:125–33.
6] Wu FCW, Balasubramanian R, Mulders TMT, Coelingh-Bennink
HJT. Oral progestogen combined with testosterone as a potential male
contraceptive: additive effects between desogestrel and testosterone
enanthate in suppression of spermatogenesis, pituitary–testicular axis,
and lipid metabolism. J Clin Endocrin Metab 1999;84:112–22.
7] Martin CW, Riley SC, Everington D, Groome NP, Riemersma RA,
Baird DT, Anderson RA. Dose-finding study of oral desogestrel with
testosterone pellets for suppression of the pituitary–testicular axis in
normal men. Hum Reprod 2000;15:1515–24.
[16] SYBYL 6.7; Tripos Assoc, 1699 S. Hanley Rd, St. Louis, MO 63144.
[17] Stewart JJP. MOPAC 6.0. Quantum Chemical Program Exchange
455; 1990.
[18] Kubinyi H. 3D QSAR in drug design: theory, methods and
applications. ESCOM: Leiden; 1993.
[19] Kubinyi H, Folkers G, Martin YC. 3D QSAR in drug design: recent
advances. ESCOM (Kluwer): Leiden; 1997.
[
[20] Bursi R, Grootenhuis PDJ. Comparative molecular field analysis
and energy interaction studies of thrombin-inhibitor complexes. J
Comput-Aided Mol Des 1999;13:221–32.
[21] Bohacek RS, McMartin C. Definition and display of steric,
hydrophobic, and hydrogen-bonding properties of ligand binding sites
in proteins using Lee and Richards accessible surface: validation
of a high-resolution graphical tool for drug design. J Med Chem
1992;35:1671–84.
[
8] Gambineri A, Pasqualini R. Testosterone therapy in men: clinical and
pharmacological perspectives. J Endocrinol Invest 2000;23:196–214.
9] Davis S. Androgen replacement in women: a commentary. J Clin
Endocrin Metab 1999;84:1886–91.
[
[
10] Kumar N, Didolkar AK, Monder C, Bardin CW, Sundaram K. The
biological activity of 7␣-methyl-19-nortestosterone is not amplified
in male reproductive tract as is that of testosterone. Endocrinology
[22] Aldridge WN. The esterases: perspectives and problems. Chem-Biol
Interact 1993;87:5–13.
[23] Satoh T, Hosokawa M. The mammalian carboxylesterases: from
molecules to functions. Annu Rev Pharmacol Toxicol 1998;38:257–
88.
[24] Buchwald P. Structure–metabolism relationships: steric effects and
the enzymatic hydrolysis of carboxylic esters. Mini Rev Med Chem
2001;1:101–11.
1
992;130:3677–83.
11] Babcock JC, Campbell JA, inventors; Upjohn Co., US, assignee.
-Methyltestosterones. US 334155; September 12, 1967.
[
[
7
12] Leysen D, Van der Voort HAA, inventors; Akzo Nobel N.V.,
The Netherlands, assignee. Preparation of (7␣,17␤)-7-methyl-17-[(1-
oxoundecyl)oxy]estr-4-en-3-one. EP 1998-202052; June 19, 1998.