New pyrazolo[3,4-b]pyridones as selective A1 adenosine receptor antagonists: Synthesis, biological evaluation and molecular modelling studies

Article abstract of DOI:10.1039/b502831k

A series of ethyl 4-amino-1-(2-chloro-2- phenylethyl)-6-oxo-6,7-dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylates (5a-j) has been synthesized as potential A₁ adenosine receptor (A ₁ AR) ligands. Binding affinities of the new compound

Full text of DOI:10.1039/b502831k

View Article Online / Journal Homepage / Table of Contents for this issue
A R T I C L E
New pyrazolo[3,4-b]pyridones as selective A1 adenosine receptor
antagonists: synthesis, biological evaluation and molecular
modelling studies
Paola Fossa,a Marco Pestarino,a Giulia Menozzi,*a Luisa Mosti,a Silvia Schenone,a
Angelo Ranise,a Francesco Bondavalli,a M. Letizia Trincavelli,b Antonio Lucacchinib and
Claudia Martinib
a Dipartimento di Scienze Farmaceutiche, Universita` degli Studi di Genova, Viale Benedetto
XV 3, 16132, Genova, Italy. E-mail: menozzi@unige.it; Fax: +39 010 3538358;
Tel: +39 010 3538351
b Dipartimento di Psichiatria, Neurobiologia, Farmacologia e Biotecnologie, Universita` degli
Studi di Pisa, Via Bonanno Pisano 6, 56126, Pisa, Italy; Fax: +39 050 2219609;
Tel: +39 050 2219526
Received 24th February 2005, Accepted 28th April 2005
First published as an Advance Article on the web 19th May 2005
A series of ethyl 4-amino-1-(2-chloro-2-phenylethyl)-6-oxo-6,7-dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylates
(5a–j) has been synthesized as potential A1 adenosine receptor (A1 AR) ligands. Binding affinities of the new
compounds were determined for adenosine A1, A2A and A3 receptors. Compounds 5b and 5g showed good affinity
(Ki = 299 nM and 517 nM, respectively) and selectivity towards A1 AR, whereas 5f showed good affinity for A2A AR
(Ki = 290 nM), higher than towards A1 AR (Ki = 1000 nM). The only arylamino derivative of the series 5j displayed
high affinity (Ki = 4.6 nM) and selectivity for A3 AR. Molecular modelling and 3D-QSAR (CoMFA) studies carried
out on the most active compounds gave further support to the pharmacological results.
Introduction
Adenosine is a powerful and widespread natural neuromodu-
lator in the nervous system.1 It is also an intermediate of the
metabolic pathways responsible for adenine nucleotide salvage
and recycling that are critical for the maintenance of ATP (itself
an adenine nucleotide) levels in all types of cells, including
neurones. Adenosine has numerous actions but, in particular,
Fig. 1 Molecular formula of compounds 1.
in the central nervous system it powerfully suppresses glutamate
release from presynaptic terminals in hippocampus2 and also
in this class of compounds also. In parallel with the biological
directly hyperpolarises neurones via activation of G-protein
evaluation, a computational study with the theoretical model
coupled inwardly rectifying potassium (GIRK) channels at
of A1 AR, previously published by us,11 has been performed
the postsynaptic site.3 Its biological functions are exploited by
to better understand the pharmacological results and to
activation of G-protein coupled receptors classified into A1, A2A,
gain insight into the binding mode of some families of non-
A2B and A3 subtypes. Adenosine receptors (ARs) from different
classical A1 antagonists, in particular of the newly synthesized
pyrazolo[3,4-b]pyridones. The results of this latest approach
have been evaluated with a 3D-QSAR analysis, obtaining useful
suggestions for the rational design of new derivatives.
species show 82–93% amino acid sequence homology, the only
exception being the A3 subtype, which exhibits 74% primary
sequence homology between rat and human.4
In the last few years, much effort has been directed towards
the synthesis of selective AR antagonists since they are attractive
tools for pharmacological intervention in many pathophysio-
logical conditions.5 In particular, A1 AR selective antagonists
have been developed as antihypertensives and potassium-saving
diuretics,6 cognition enhancers5 and useful therapeutics for the
alleviation of the symptoms of Alzheimer’s disease.5,6 Among
these, a wide variety of nitrogen-containing heterocycles have
been synthesized as possible A1 selective antagonists and
pyrazolo[3,4-b]pyridine7 and pyrazolo[3,4-d]pyrimidine8 deriva-
tives were found to be selective ligands with antagonistic activity
for A1 ARs.
In the course of our studies directed to obtain new non-
classical adenosine ligands,9,10 we synthesized a series of ethyl
4-amino-1-(2-chloro-2-phenylethyl)-1H-pyrazolo[3,4-b]pyridine-
5-carboxylates 1 (see Fig. 1), some of which showed interesting
affinity and antagonistic activity towards A1 AR.9 As a
continuation of these studies we have then planned the synthesis
of the pyrazolo[3,4-b]pyridone derivatives 5a–j, analogues of
1, in order to further explore structure–activity relationships
Results and discussion
Chemistry
The synthetic route to the target pyrazolo[3,4-b]pyridones 5a–j
is depicted in Scheme 1. Starting compound, namely ethyl 5-
amino-1-(2-hydroxy-2-phenylethyl)-1H-pyrazole-4-carboxylate
2, was prepared according to the procedure reported in a
previous paper.9 Condensation of 2 with diethyl malonate gave
the cyclization to ethyl 4-hydroxy-1-(2-hydroxy-2-phenylethyl)-
6-oxo-6,7-dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylate
3. Treatment of 3 with a mixture of thionyl chloride and
dimethylformamide provided the dichloroderivative 4. Finally,
the aromatic displacement of the C4 chloro substituent of
the intermediate 4 with various amines gave the pyrazolo[3,4-
b]pyridones 5a–j.
The nucleophilic substitution was performed under mi-
crowave irradiation (CEM Discovery system), to enhance yields
and to reduce reaction times. Microwave technology, indeed, can
2 2 6 2
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 2 6 2 – 2 2 7 0
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
©

View Article Online
The A1, A2A and A3 AR binding affinities for compounds 5a–j
are expressed as Ki or the percentage of inhibition of binding (at
10 lM compound concentration) and are reported in Table 1.
The results of the binding tests of the newly synthesized
derivatives showed various affinity levels towards A1 or A2A or
A3 ARs, in most cases with poor selectivity.
If we compare the results of the binding tests of the new
pyrazolo[3,4-b]pyridone derivatives 5 with the data relevant to
pyrazolo[3,4-b]pyridine analogues 1,9 we can notice, generally,
a decrease in A1 AR affinity and a rise in A2A AR affinity. It
should be remarked that compounds 5a, 5d and 5f exhibited an
A2A AR affinity higher than A1 AR affinity. Compounds 5b and
5g showed the highest affinity, with some selectivity, towards A1
AR. The N6-cyclopentyl substituent is actually known to induce
high adenosine A1 receptor affinity and selectivity.14
The cyclopropylamino substituent, on the contrary, gave an
A2A AR affinity higher than A1 AR affinity (compound 5f) in
this series.
Scheme 1 Compounds: 5a, R = NH(CH2)2CH3; 5b, R = NHCH2C6H5;
5c,
Only the anilino derivative 5j exhibited high affinity and
selectivity towards A3 AR.
R
=
NH(CH2)2C6H5; 5d, R= NHCH(CH3)C6H5; 5e,
R
=
NH(CH2)2OC2H5; 5f, R = NH-cyclopropyl; 5g, R = NH-cyclopentyl;
5h, R = NH-cyclohexyl; 5i, R = 1-pyrrolidinyl; 5j, R = NHC6H5.
Reagents and conditions: (a) EtONa, absolute ethanol, reflux, 6 h; (b) dil.
HCl; (c) SOCl2, DMFA, chloroform, reflux, 3 h; (d) method A: amine,
CH3COOH, dioxane, microwave irradiation, 10 min; method B: amine,
chloroform, 60 ◦C, 2 h.
Structure–activity relationship (SAR) considerations
The biological results of the new pyrazolo-pyridones showed
that the variation of the nature of the amino substituent at the C4
position significantly influenced the ARs affinity in the following
ways: elongation, ramification, shortening of the benzyl chain of
5b, the most active A1 AR ligand of the series, gave derivatives
5c, 5d and 5j, with a lower affinity towards both A1 and A2A ARs,
in comparison with 5b; a short alkyl or cycloalkyl substituent
(n-propyl or cyclopropyl of compounds 5a and 5f, respectively)
was associated with an A2A affinity higher than A1 AR affinity;
in the case of cyclopropyl derivative 5f a higher affinity towards
both A1 and A2A ARs, together with a poor A2A AR selectivity,
was found; in comparison with the n-propyl derivative 5a, the
introduction of a longer alkoxyalkyl chain (5e) increased the
affinity towards both A1 and A2A ARs, with loss of selectivity;
enlargement of the aliphatic ring from three (5f) to five terms
(5g) produced an increase in affinity and selectivity towards A1
AR, while a further enlargement to six terms (5h) brought about
a reduction in both A1 and A2A ARs affinity; substitution of
a secondary amino group with a tertiary one (1-pyrrolidinyl
derivative 5i) maintained some affinity towards both A1 and A2A
ARs, with loss of selectivity; the anilino group produced high
affinity and selectivity towards A3 AR.
be successfully applied in many organic reactions12 and, in par-
ticular, it was demonstrated to be useful in nucleophilic aromatic
substitutions.13 Compounds 5c, 5f and 5j were also prepared by
conventional procedure, to compare the yields obtained from the
two methods. All the reactions were performed in closed vessels,
but employing different solvents and experimental conditions.
Microwave-assisted reactions were carried out in dioxane added
with acetic acid, under irradiation at 300 W (final temperature
150 ◦C, pressure 35 psi), for 10 min. Following the conventional
procedure, reactions were performed in chloroform, at the tem-
perature of 60 ◦C, for 2 h. Experimental results confirmed that
the conventional method, besides requiring a longer reaction
time, gives lower yield, in comparison with the microwave-
assisted technique (see Experimental).
Biological evaluation of compounds 5a–j
The biological activity of the compounds was tested by radioli-
gand competition experiments.
Table 1 Affinity of 5a–j derivatives towards bovine A1, A2A ARs and human A3 ARs
Ki/nM or inhibition (%)a
b
c
d
Compound
R
Bovine A1
Bovine A2A
Human A3
5a
5b
5c
5d
5e
5f
NH(CH2)2CH3
NHCH2C6H5
54%
299
40%
48%
1285
1000
517
2443
2173
46%
2341
2103
6%
1299
1492
290
2459
24%
2696
52%
21%
38%
13%
29%
30%
3663
31%
16%
18%
4.6
NH(CH2)2C6H5
NHCH(CH3)C6H5
NH(CH2)2OC2H5
NH-cyclopropyl
NH-cyclopentyl
NH-cyclohexyl
1-Pyrrolidinyl
NHC6H5
5g
5h
5i
5j
a The Ki values are means SEM of three separate assays, each performed in triplicate. b Displacement of specific [3H]DPCPX binding in bovine
cortical membranes or percentage of inhibition of specific binding at 10 lM concentration. c Displacement of specific [3H]CGS21680 binding in
bovine striatal membranes or percentage of inhibition of specific binding at 10 lM concentration. d Displacement of specific [125I] AB-MECA binding
in CHO cell membranes or percentage of inhibition of specific binding at 1 lM concentration.
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 2 6 2 – 2 2 7 0
2 2 6 3

View Article Online
Molecular modelling studies
points around the molecule by means of a suitable probe,
usually an sp3 carbon atom with a charge of +1. Partial Least
Squares (PLS) is used as the regression method to develop
the relationship between steric and electrostatic potentials and
biological activity. The graphical representation of CoMFA
model, in the form of coefficient isocontour maps, efficiently
locates the regions where the variation in steric and electrostatic
properties of different molecules in a data set is correlated with
the variation of biological activity. CoMFA isocontour maps
may thus provide useful indications to develop sound working
hypothesis on the nature of putative ligand–macromolecule
interactions.
In the present work, CoMFA models were developed using a
common structure alignment derived from the superimposition
of structures 14–16, 35, 46, 56, 64 and 67–71 as derived from
docking studies. Results for the best model obtained are reported
in Table 4. The model exhibited a cross-validated r2 (r2cv) of
0.540 with cancellation groups validation procedure. A non-
cross-validated r2 (r2ncv) of 0.937 was obtained.
In order to further rationalize the biological results obtained,
compounds displaying the most interesting activity on the A1
AR, 5b and 5g, were docked by means of FlexX15 into our A1
AR virtual model.11 In order to properly evaluate the validity
of such a theoretical approach, in the absence of any structural
data on the binding mode of any adenosine receptor ligand
and its biological target, a data set of twelve selective A1
AR antagonists, 14–16, 35, 46, 56, 64 and 67–71, Table 2,
representative for different chemical families, was docked into
A1 AR. The main molecular interactions between the A1AR
and compounds 14–16, 35, 46, 56, 64 and 67–71, plus 5b and
5g, are presented in Table 3. According to our calculations,
a binding area common to all the antagonists studied was
defined. Compounds 5b and 5g were also accommodated here.
In this area (Fig. 2 and Table 3), hydrogen bonding with residue
Thr91 seems necessary for the interaction of any ligand with
the receptor, while the affinity of the antagonist appears to
be determined by additional H-bonds with residue His251, in
agreement with results coming from site-directed mutagenesis
experiments.4 Near these hydrophilic interactions, hydrophobic
interactions with three binding pockets, characterized by a
different capacity of allocating hydrophobic portions of the
ligands,9 play a key-role in modulating the affinity of the
antagonist with the biomolecule. According to our docking
studies, pocket P1, defined by amino acid residues Ile67, Val
87, Leu88 and Thr91, is able to accept various substituents
(R or R’ on the molecular skeleton) even with a significant
steric hindrance. Pocket P2, defined by residues Trp188, Leu250,
His251 and Asp254, and pocket P3, defined by residues Ser94,
Ile95, Leu98, Thr277 and Ser281, could accommodate only
short alkyl or small cycloalkyl substituents without any further
branching. On the basis of these binding simulations, the affinity
of the antagonist towards the receptor is greatly determined by
its ability to properly occupy all the three pockets at the same
time.
In addition, almost equal contributions were observed for
steric (52%) and electrostatic (48%) fields, represented in Fig. 3
as 3D contour plots, suggesting a balanced model with slightly
prevalent steric effects.
Fig. 2 Compound 15 (KW3902), coloured by atom type, docked into
the binding site of A1 AR. For clarity, only the most important residues
are reported and labelled. The three hydrophobic pockets are coloured
as follows: P1 in yellow, P2 in orange and P3 in magenta. Two residues
not part of a pocket but important for the binding are coloured in cyan.
Fig. 3 CoMFA steric and electrostatic contour maps. Yellow denotes
regions where steric bulk is detrimental to activity and green denotes
regions where steric bulk enhances activity. Red denotes regions where
positive charge is detrimental to activity and blue denotes region where
positive charge enhances activity. The superimposed antagonists 6–71
are also shown.
In tandem with our molecular docking studies, a quantitative
evaluation of the structure–activity relationships inside A1
AR antagonists has been performed through a comparative
molecular field analysis (CoMFA)16 study, on compounds 6–
71 (Table 2). The CoMFA method is widely used as 3D-QSAR
technique to relate the biological activity of a series of molecules
with their steric and electrostatic fields sampled at several grid
The steric contour map (Fig. 3) showed yellow (steric
unfavourable interaction) polyhedra differently delimiting the
2 2 6 4
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 2 6 2 – 2 2 7 0

View Article Online
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 2 6 2 – 2 2 7 0
2 2 6 5

View Article Online
2 2 6 6
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 2 6 2 – 2 2 7 0

View Article Online
Table 3 Main protein–ligand interactions as displayed by the best docking orientation for each inhibitor
Compound
Amino acid residues involved in H-bonds
Amino acid residues involved lipophilic interactions
14
15
16
35
46
56
64
67
68
69
70
71
5b
5g
Thr91, Ser94, His251
Thr91, Ser94, His251
Thr91, Ser94, His251
Thr91, His251
Thr91, His251
Thr91, His251
Thr91, His251
Thr91
Thr91, His251
Thr91, His251
Thr91
Thr91, Trp188
Thr91, Ser94
Thr91
Ile67, Leu88, Ile95, Leu250, Ala273, Ile274
Ile67, Leu88, Ile95, Ile274
Leu88, Ile95, Leu250, Ile274
Ile67, Ile95, Ile274
Ile67, Val87, Ile95
Ile274
Ile95, Ile274
Ile95, Ala273, Ile274
Leu88, Ala273, Ile274
Ile67, Leu88, Ile95, Ile274
Ile67, Ile95, Ile274
Ile274
Ile67, Ile95
Ile67, Ile95
at N1 than pyrazolo[3,4-b]pyridine derivatives 19 previously
synthesized by us and endowed with a noteworthy A1 AR
antagonistic activity. Many of the newly synthesized compounds
showed an interesting affinity towards A1 and A2A ARs, but
poor selectivity. Compounds 5b and 5g turned out to be A1
AR antagonists with good affinity (Ki = 299 and 517 nM,
respectively) but a fair selectivity. Compound 5f, on the con-
trary, showed good affinity towards A2A AR (Ki = 290 nM)
and 5j exhibited high affinity (Ki = 4.6 nM) and selectivity
towards A3 AR. In conclusion, the biological data proved
that the substitution of the pyridine ring in the structure of
compounds 1 with the pyridone nucleus is not profitable for
the A1 AR antagonistic activity and selectivity. Unexpectedly,
the introduction of an anilino group on the C4 position of the
pyrazolopyridone scaffold (5j) provided high A3 AR affinity
and selectivity. Molecular modelling studies on A1 AR fully
support and contribute to explain the experimental results
obtained on A1 AR. In particular, the coordinated use of
CoMFA and docking approaches has lead to a significant and
complementary insight into the complex biological interactions
between antagonist and A1 receptor. Helpful suggestions for
the synthesis of new and more potent antagonists in the class
of pyrazolo[3,4-b]pyridines will be derived from the CoMFA
model.
Table 4 Summary of CoMFA–PLS results
No. of compounds
Opt. no. of components
Cross-validated r2
Std error of estimate
Non-cross-validated r2
F Values
65
5
0.540
0.229
0.937
141.867
0.510
0.490
0.014
0.960
Steric contribution
Electrostatic contribution
Std dev.
R2bs
borders of the three pockets P1, P2 and P3. In agreement with the
above reported docking studies, P1 is not strictly limited in size,
while P2 and P3 are more conditioned by surrounding yellow
regions. Some green (steric favourable interaction) polyhedra
in correspondence to one side of P2 and P3 pockets defined
spatial regions where a steric interaction between the ligand and
the receptor would enhance the strength of the binding. On the
basis of these findings, in the evaluation of steric contour maps
of all the newly synthesized compounds 5, derivatives 5b and
5g were the only ones able to occupy green regions avoiding
the occupancy of any yellow portion. On the contrary, for all
other compounds unfavourable steric interactions were observed
within the CoMFA model. The electrostatic contour map, shown
in Fig. 3, defined a blue region between P1 and P2 where
the presence of positive charges enhances activity (electrostatic
favourable interaction). This region corresponds mainly to
tertiary heterocyclic nitrogen in the antagonist molecule. A
small red region (electrostatic unfavourable interaction) where
the presence of positive charges is detrimental for activity was
instead defined at the border between P2 and P3 pockets.
Almost all of the studied ligands with their polar substituents
or polar groups did not occupy this area. According to these
data, compounds 5, characterized by a pyrazolo [3,4-b]pyridone
nucleus, showed a slightly unfavourable electrostatic interaction
with the counter part, positioning the carbonyl group in C-6 at
the borderline of the red region. On the contrary, analogues 1,9
characterized by a pyrazolo [3,4-b]pyridine core, do not present
any unfavourable electrostatic interaction with the counterpart.
Thus, our calculations, as a further support to biological data,
in the case of 5b and 5g suggest that the enhancement of activity
due to favourable steric interactions is only slightly reduced by an
unfavourable electrostatic interaction. For all others compounds
5 the sum of two detrimental contributions (mostly steric and
only partially electrostatic) could justify the low affinity values
experimentally determined towards A1 AR.
Experimental
Chemistry protocols
All chemicals and solvents used were commercially available
and of analytical grade or were prepared according to the
procedure described in the literature.9 All of the microwave-
assisted reactions were performed in a CEM Discover system
(CEM Corporation). All compounds were tested for purity by
thin-layer chromatography (TLC) on silica gel plates (60 F254
,
Merck; ethyl acetate–petroleum ether 1 : 1 as eluant), visualizing
with ultraviolet light. Melting points were determined with a
Fisher–Johns apparatus and are uncorrected. IR spectra were
registered on a Perkin-Elmer 398 spectrophotometer and are
expressed in cm−1. 1H NMR spectra were registered on a
Varian Gemini 200 (200 MHz) spectrometer; chemical shifts
are reported as d values (ppm) relative to TMS as internal
standard; coupling constants (J) are expressed in Hertz (Hz).
The following NMR abbreviations are used: br (broad), s
(singlet), d (doublet), t (triplet), q (quartet), m (multiplet) and
ex (exchangeable with D2O). Microanalyses for C, H, N were
performed using a Carlo Erba Elemental Analyzer Model EA
1110 and results agree within 0.4% with calculated values
Ethyl 4-hydroxy-1-(2-hydroxy-2-phenylethyl)-6-oxo-6,7-dihydro-
1H-pyrazolo[3,4-b]pyridine-5-carboxylate 3. Diethyl malonate
(9.61 g, 60 mmol) was slowly added to a solution of sodium
ethoxide, prepared from sodium (2.07 g, 90 mmol) in absolute
Conclusions
In this report, we described the synthesis of a series of
pyrazolo[3,4-b]pyridones 5a–j, having the same pharmacophore
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 2 6 2 – 2 2 7 0
2 2 6 7

View Article Online
ethanol (80 cm3). After stirring at rt for 15 min, ethyl 5-amino-1-
(2-hydroxy-2-phenylethyl)-1H-pyrazole-4-carboxylate 2 (8.26 g,
30 mmol) was added and the mixture obtained was refluxed
for 6 h. After cooling, the reaction mixture was concentrated
under a reduced pressure to half volume, diluted with water
(100 cm3), cooled in an ice bath and acidified to pH = 5 with
diluted HCl (6 N and 1 N, successively). The white solid which
separated was filtered and dried for 3 h in a vacuum oven at
100 ◦C. Recrystallization from glacial acetic acid afforded the
title compound 3 (5.7 g, 55%); mp 188–190 ◦C (dec); (found:
C, 59.43; H, 4.98; N, 12.31%. C17H17N3O5 requires C, 59.47; H,
4.99; N, 12.24%); mmax (KBr)/cm−1 3526, 1661, 1640 and 1615.
dH (200 MHz; DMSO-d6; Me4Si): 1.30 (3 H, t, J 7.0, CH2CH3),
4.31 (2 H, q, J 7.0, CH2CH3), 4.1–4.6 (2 H, m, CH2N), 4.95–5.05
(1 H, m, CHOH), 5.60 (1 H, br s, OH, ex), 7.2–7.6 (5 H, m, Ph),
7.99 (1 H, s, 3-H), 11.95 (1 H, s, 4-OH, ex), 13.45 (1 H, s, NH, ex).
(2 H, d, J 4.8, CH2Ph), 5.50–5.65 (1 H, m, CHCl), 7.25–7.60 (10
H, m, 2 Ph), 7.95 (1 H, s, 3-H), ∼8.9 (1 H, very br s, NH, ex),
∼12.80 (1 H, very br s, NH, ex).
Ethyl 1-(2-chloro-2-phenylethyl)-6-oxo-4-(2-phenylethyl)amino-
6,7-dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylate 5c. As
described for 5a, from 4 and 2-phenylethylamine (0.24 g,
2 mmol). Recrystallization from 95% ethanol gave 5c as a white
solid (0.51 g, 85%); mp 183–185 ◦C; (found: C, 64.54; H, 5.46; N,
11.98%. C25H25ClN4O3 requires C, 64.58; H, 5.42; N, 12.05%);
mmax (CHCl3)/cm−1 3396, 1644, 1610 and 1585. dH (200 MHz;
CDCl3; Me4Si): 1.24 (3 H, t, J 7.0, CH2CH3), 3.13 (2 H, t, J 7.0,
NHCH2CH2Ph), 3.98 (2 H, q, J 7.0 Hz, NHCH2CH2Ph), 4.33
(2 H, q, J 7.0, CH2CH3), 4.65–4.80 and 4.90–5.10 (2 H, 2 dd,
CH2N), 5.50–5.65 (1 H, m, CHCl), 7.20–7.60 (10 H, m, 2 Ph),
7.99 (1 H, s, 3-H), ∼8.40 (1 H, very br s, NH, ex), ∼12.80 (1 H,
very br s, NH, ex).
Ethyl
4-chloro-1-(2-chloro-2-phenylethyl)-6-oxo-6,7-dihydro-
Ethyl 1-(2-chloro-2-phenylethyl)-6-oxo-4-(1-phenylethyl)amino-
6,7-dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylate 5d. As
described for 5a, from 4 and 1-phenylethylamine (0.24 g,
2 mmol). Recrystallization from 95% ethanol gave 5d as a white
solid (0.42 g, 70%); mp 186–187 ◦C; (found: C, 64.29; H, 5.52; N,
11.94%. C25H25ClN4O3 requires C, 64.58; H, 5.42; N, 12.05%);
mmax (CHCl3)/cm−1 3395, 1644, 1613 and 1582. dH (200 MHz;
CDCl3; Me4Si): 1.42 (3 H, t, J 7.0, CH2CH3), 1.73 (3 H, d, J
6.8, NHCH(Ph)CH3), 4.48 (2 H, q, J 7.0, CH2CH3), 4.60–5.00
(2 H, m, CH2N), 5.16 (1 H, quintet, J 6.8, NHCH(Ph)CH3),
5.40–5.55 (1 H, m, CHCl), 7.25–7.50 (10 H, m, 2 Ph), 7.67 (1
H, s, 3-H), ∼9.18 (1 H, very br s, NH, ex).
1H-pyrazolo[3,4-b]pyridine-5-carboxylate 4. Ethyl 4-hydroxy-
1-(2-hydroxy-2-phenylethyl)-6-oxo-6,7-dihydro-1H-pyrazolo-
[3,4-b]-pyridine-5-carboxylate
3 (3.43 g, 10 mmol) was
dissolved in a mixture of thionyl chloride (7.3 cm3, 100 mmol),
dimethylformamide (1.1 cm3, 14 mmol) and chloroform
(40 cm3). The mixture was refluxed for 3 h, cooled and
evaporated to dryness in vacuo. The residue was dissolved in
chloroform (50 cm3); this solution was washed twice with water
(2 × 15 cm3), dried by magnesium sulfate and evaporated
in vacuo. The solid residue was purified by recrystallization
from ethyl acetate to yield the title compound 4 as a white
solid (2.05 g, 54%); mp 165–168 ◦C (dec); (found: C, 53.71;
H, 4.20; N, 11.03%. C17H15Cl2N3O3 requires C, 53.70; H, 3.98;
N, 11.05%); mmax (CHCl3)/cm−1 1732, 1642, 1611 and 1573. dH
(200 MHz; CDCl3; Me4Si): 1.43 (3H, t, J 7.0, CH2CH3), 4.47
(2 H, q, J 7.0, CH2CH3), 4.65–4.80 and 4.95–5.10 (2 H, 2 dd,
CH2N), 5.45–5.55 (1 H, m, CHCl), 7.25–7.55 (5 H, m, Ph), 8.02
(1 H, s, 3-H), ∼13.25 (1 H, very br s, NH, ex).
Ethyl
1-(2-chloro-2-phenylethyl)-4-(2-ethoxyethyl)amino-6-
oxo-6,7-dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylate 5e.
As described for 5a, from 4 and 2-ethoxyethylamine (0.18 g,
2 mmol). Recrystallization from 95% ethanol gave 5e as
a white solid (0.44 g, 80%); mp 154–155 ◦C; (found: C,
59.98; H, 6.21; N, 13.31%. C21H25ClN4O4 requires C, 59.93;
H, 5.99; N, 13.31%); mmax (CHCl3)/cm−1 3393, 1644, 1609
and 1586. dH (200 MHz; CDCl3; Me4Si): 1.29 (3H, t, J 7.0,
NHCH2CH2OCH2CH3), 1.48 (3 H, t, J 7.0, CH2CH3), 3.63
(2 H, q, J 7.0, NHCH2CH2OCH2CH3), 3.70–4.00 (4 H,
m, NHCH2CH2OCH2CH3), 4.51 (2 H, q, J 7.0, CH2CH3),
4.65–4.80 and 4.90–5.05 (2 H, 2 dd, CH2N), 5.50–5.65 (1 H, m,
CHCl), 7.25–7.60 (5 H, m, Ph), 7.95 (1 H, s, 3-H), ∼8.65 (1 H,
very br s, NH, ex), ∼12.80 (1 H, very br s, NH, ex).
Method A: general procedure for the preparation of pyrazolo-
[3,4-b]pyridones 5a–j. Representative preparation of ethyl 1-
(2-chloro-2-phenylethyl)-6-oxo-4-propylamino-6,7-dihydro-1H-
pyrazolo[3,4-b]pyridine-5-carboxylate 5a. A mixture of ethyl
4-chloro-1-(2-chloro-2-phenylethyl)-6-oxo-6,7-dihydro-1H-
pyrazolo[3,4-b]pyridine-5-carboxylate 4 (0.5 g, 1.3 mmol),
propylamine (0.12 g, 2 mmol), glacial acetic acid (0.12 g,
2 mmol) and dioxane (10 cm3) in a closed vessel was exposed to
microwave irradiation (power 300 W; final temperature 150 ◦C;
pressure 35 psi) for 10 min. After cooling, the mixture was
diluted with chloroform (50 cm3), washed in succession with
water (20 cm3), saturated sodium carbonate solution (10 cm3)
and water (20 cm3), dried by magnesium sulfate and evaporated
in vacuo. The solid residue was recrystallized from 95% ethanol
to yield the title compound 5a as a white solid (0.44 g, 85%); mp
200–202 ◦C; (found: C, 59.33; H, 5.68; N, 13.84%. C20H23ClN4O3
requires C, 59.62; H, 5.75; N, 13.91%); mmax (CHCl3)/cm−1
3403, 1643, 1611 and 1586. dH (200 MHz; CDCl3; Me4Si): 1.15
(3H, t, J 7.6, NHCH2CH2CH3), 1.46 (3H, t, J 7.0, CH2CH3),
1.86 (2 H, sextet, J 7.2, NHCH2CH2CH3), 3.5–3.7 (2 H, m,
NHCH2CH2CH3), 4.48 (2 H, q, J 7.0, CH2CH3), 4.65–4.80
and 4.95–5.10 (2 H, 2 dd, CH2N), 5.50–5.65 (1 H, m, CHCl),
7.25–7.55 (5 H, m, Ph), 7.96 (1 H, s, 3-H), ∼8.7 (1 H, very br s,
NH, ex).
Ethyl 1-(2-chloro-2-phenylethyl)-4-cyclopropylamino-6-oxo-
6,7-dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylate 5f. As
described for 5a, from 4 and cyclopropylamine (0.11 g, 2 mmol).
Recrystallization from 95% ethanol gave 5f as a white solid
(0.39 g, 75%); mp 206–208 ◦C; (found: C, 59.99; H, 5.26;
N, 13.80%. C20H21ClN4O3 requires C, 59.92; H, 5.28; N,
13.98%); mmax (CHCl3)/cm−1 3407, 1644, 1613 and 1585. dH
(200 MHz; CDCl3; Me4Si): 0.60–0.75 and 0.95–1.10 (4 H, 2
m, CH2CH2), 1.25 (3 H, t, J 7.0, CH2CH3), 2.95–3.10 (1 H,
m, CH cyclopropyl), 4.23 (2 H, q, J 7.0, CH2CH3), 4.50–4.70
and 4.95–5.15 (2 H, 2 dd, CH2N), 5.50–5.65 (1 H, m, CHCl),
7.35–7.70 (5 H, m, Ph), 8.32 (1 H, s, 3-H), ∼9.60 (1 H, very br s,
NH, ex) and ∼11.70 (1 H, very br s, NH, ex).
Ethyl 1-(2-chloro-2-phenylethyl)-4-cyclopentylamino-6-oxo-
6,7-dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylate 5g. As
described for 5a, from 4 and cyclopentylamine (0.17 g, 2 mmol).
Recrystallization from 95% ethanol gave 5g as a white solid
(0.40 g, 71%); mp 192–193 ◦C; (found: C, 61.51; H, 5.87; N,
13.02%. C22H25ClN4O3 requires C, 61.61; H, 5.87; N, 13.06%);
mmax (CHCl3)/cm−1 3396, 1644, 1610 and 1584. dH (200 MHz;
CDCl3; Me4Si): 1.46 (3 H, t, J 7.2, CH2CH3), 1.65–1.90 (4 H,
m, CH2CH2), 2.10–2.25 (4 H, m, CH2CHCH2), 4.30–4.50 (1 H,
m, cyclopropyl CH), 4.47 (2 H, q, J 7.0, CH2CH3), 4.65–4.80
and 4.95–5.10 (2 H, 2 dd, CH2N), 5.50–5.60 (1 H, m, CHCl),
Ethyl 4-(benzylamino)-1-(2-chloro-2-phenylethyl)-6-oxo-6,7-
dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylate 5b. As de-
scribed for 5a, from 4 and benzylamine (0.21 g, 2 mmol).
Recrystallization from 95% ethanol gave 5b as a white solid
(0.51 g, 86%); mp 174–176 ◦C; (found: C, 64.07; H, 5.12; N,
12.44%. C24H23ClN4O3 requires C, 63.93; H, 5.14; N, 12.42%);
mmax (CHCl3)/cm−1 3395, 1644, 1611 and 1583. dH (200 MHz;
CDCl3; Me4Si): 1.20 (3 H, t, J 7.0, CH2CH3), 4.35 (2 H, q, J
7.0, CH2CH3), 4.65–4.80 and 4.90–5.10 (2 H, 2 dd, CH2N), 4.84
2 2 6 8
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 2 6 2 – 2 2 7 0

View Article Online
7.25–7.55 (5 H, m, Ph), 7.96 (1 H, s, 3-H), ∼8.75 (1 H, very br s,
Receptor binding assays. A1 and A2A receptor binding. The
affinity of the new synthetised compounds towards A1 and A2A
ARs was evaluated by competition experiments assessing their
ability to displace [3H]DPCPX and [3H]CGS21680 binding from
bovine cortical and striatal membranes, respectively. Binding
assays were carried out as previously described.17–19
NH, ex).
Ethyl 1-(2-chloro-2-phenylethyl)-4-cycloexylamino-6-oxo-6,7-
dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylate 5h. As de-
scribed for 5a, from 4 and cyclohexylamine (0.2 g, 2 mmol).
Recrystallization from 95% ethanol gave 5h as a white solid
(0.46 g, 81%); mp 233–234 ◦C; (found: C, 62.67; H, 6.11; N,
12.58%. C23H27ClN4O3 requires C, 62.37; H, 6.14; N, 12.65%);
mmax (CHCl3)/cm−1 3391, 1634, 1604 and 1584. dH (200 MHz;
CDCl3; Me4Si): 1.47 (3 H, t, J 7.0, CH2CH3), 1.30–1.95 (8 H, m,
4 CH2 cyclohexyl), 2.10–2.25 (2 H, m, CH2 cyclohexyl), 3.80–
3.95 (1 H, m, CH), 4.49 (2 H, q, J 7.0, CH2CH3), 4.65–4.75 and
4.95–5.10 (2 H, 2dd, CH2N), 5.50–5.60 (1 H, m, CHCl), 7.25–
7.55 (5 H, m, Ph), 7.84 (1 H, s, 3-H), ∼8.70 (1 H, very br s, NH,
ex).
A3AR binding. Compound affinity towards A3 ARs were
evaluated in membranes obtained from CHO cells trans-
fected with human A3 AR, kindly supplied by Dr Klotz
(Wurzburg, Germany). Competition experiments were per-
formed using [125I]AB-MECA as radioligand following a de-
scribed procedure20
All compounds were routinely dissolved in DMSO and diluted
with assay buffer to the final concentration, where the amount
of DMSO never exceeded 2%.
At least six different concentrations spanning three orders
of magnitude, adjusted appropriately for the IC50 of each com-
pound, were used. IC50 values, computer-generated using a non-
linear regression formula on a computer program (GraphPad,
San Diego, CA), were converted to Ki values, knowing the Kd
values of radioligands in the different tissues and using the
Cheng and Prusoff equation.21 The dissociation constant (Kd)
of [3H]CHA, [3H]CGS 21680, and [125I]AB-MECA were 1.2, 14,
and 1.02 nM, respectively.
Ethyl 1-(2-chloro-2-phenylethyl)-6-oxo-4-pyrrolidin-1-yl-6,7-
dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylate 5i. As de-
scribed for 5a, from 4 and pyrrolidine (0.14 g, 2 mmol).
Recrystallization from 95% ethanol gave 5i as a white solid
(0.37 g, 68%); mp 247–248 ◦C; (found: C, 61.04; H, 5.78; N,
13.63%. C21H23ClN4O3 requires C, 60.79; H, 5.59; N, 13.50%);
mmax (CHCl3)/cm−1 1708, 1626 and 1574. dH (200 MHz; CDCl3;
Me4Si): 1.27 (3 H, t, J 7.0, CH2CH3), 2.04 (4 H, br s, 2CH2
pyrrolidinyl), 3.70 (4 H, br s, 2CH2N pyrrolidinyl), 4.23 (2 H, q,
J 7.0, CH2CH3), 4.75–5.10 (2 H, 2 dd, CH2N), 5.35–5.50 (1 H,
m, CHCl), 7.20–7.50 (5 H, m, Ph), 7.88 (1 H, s, 3-H), ∼13.60
(1 H, very br s, NH, ex).
Computational methods
Data set. For this investigation 65 A1 AR selective antago-
nists 6–71 chosen from the literature,8,22–30 plus a pyrazole[3,4-
b]pyridine derivative 62 previously published by us9 and two of
the most interesting compounds among the newly synthesized,
namely 5b and 5g, were collected. In Table 2 are reported their
molecular structures and their affinity value towards the A1 AR
from rat cerebral cortex, expressed as Ki (nM). The reduced
availability of affinity data on A1 agonists for the human A1
receptor still hampers the study of agonist–receptor interactions
in the A1AR model, this is why we therefore employed a larger
dataset of affinity values as determined in rat brain, considering
that the sequences of the A1AR in the two species have a
percentage identity of 94.8% and the amino acid differences
are located in the TM4–5 loop and in the carboxy-terminal
cytoplasmic segment.
Ethyl 4-anilino-1-(2-chloro-2-phenylethyl)-6-oxo-6,7-dihydro-
1H-pyrazolo[3,4-b]pyridine-5-carboxylate 5j. As described for
5a, from 4 and aniline (0.19 g, 2 mmol). Recrystallization from
95% ethanol gave 5j a white solid (0.41 g, 72%); mp 195–196 ◦C;
(found: C, 63.22; H, 5.13; N, 12.63%. C23H21ClN4O3 requires C,
63.23; H, 4.84; N, 12.82%); mmax (CHCl3)/cm−1 3381, 1647, 1614
and 1579. dH (200 MHz; CDCl3; Me4Si): 1.25 (3 H, t, J 6.8,
CH2CH3), 4.31 (2 H, near q, J 7.0, CH2CH3), 4.45–4.65 and
4.75–4.95 (2 H, 2 dd, CH2N), 5.25–5.40 (1 H, m, CHCl), 6.21
(1 H, s, 3-H), 7.10–7.50 (10 H, m, 2 Ph), ∼10.30 (1 H, very br s,
NH, ex).
Method B: general procedure for the preparation of pyrazolo-
[3,4-b]pyridones 5c, 5f and 5j. Representative preparation of
ethyl 1-(2-chloro-2-phenylethyl)-6-oxo-4-(2-phenylethyl)amino-
6,7-dihydro-1H-pyrazolo[3,4-b]pyridine-5carboxylate 5c. A so-
lution of ethyl 4-chloro-1-(2-chloro-2-phenylethyl)-6-oxo-6,7-
Calculations were performed transforming the original affin-
ity in pKi and using it as CoMFA dependent variable.
Docking Studies. Molecular structures of ligands 5b and
5g and 6–71 were built and energy minimized within
MacroModel.31 Conformational analysis was carried out using
the AMBER* force field, as included in MacroModel. For
all compounds, the resulting geometries of the lower energy
conformers were re-optimized with semi-empirical quantum
mechanic calculations, using the Hamiltonian AM1 as imple-
mented in Spartan32 and atomic charges were calculated.
dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylate
4 (0.5 g,
1.3 mmol) and 2-phenylethylamine (0.24 g, 2 mmol) in chlo-
roform (10 cm3) in a closed vessel was heated (oil bath) at 60 ◦C,
with stirring, for 2 h. After cooling, the mixture was diluted with
chloroform (40 cm3), washed with water (2 × 20 cm3), dried by
magnesium sulfate and evaporated in vacuo. The solid residue
was recrystallized from 95% ethanol to yield the title compound
5c (0.21 g, 33%).
The theoretical three-dimensional model of the human A1AR
previously published by us11 was used for a two-step docking
protocol. In a first phase, each inhibitor was docked into the
active site by means of the FlexX module, as implemented in
Sybyl version 6.833 with the macromolecule and the ligands being
flexible. Preparation of the protein for FlexX requires definition
of the binding pocket in terms of “interaction points”. In this
work the active site was defined as all atoms within a distance
Ethyl 1-(2-chloro-2-phenylethyl)-4-cyclopropylamino-6-oxo-
6,7-dihydro-1H-pyrazolo[3,4-b]pyridine-5-carboxylate 5f. As
described for 5c, from 4 and cyclopropylamine (0.11 g, 2 mmol).
Recrystallization from 95% ethanol gave 5f (0.31 g, 69%).
Ethyl 4-anilino-1-(2-chloro-2-phenylethyl)-6-oxo-6,7-dihydro-
1H-pyrazolo[3,4-b]pyridine-5-carboxylate 5j. As described for
5a, from 4 and aniline (0.19 g, 2 mmol). Recrystallization from
95% ethanol gave 5j (0.09 g, 16%).
˚
of 10 A from adenosine, the natural substrate, whose binding
mode has been defined in ref. 11. This specific distance was
determined in order to ensure a significant portion of the active
site for the docking experiments. Starting from the best-docked
geometries, as obtained with FlexX, the second step consisted in
a further refinement of the complex performed with QXP.34 Also
the algorithm implemented in the QXP program allows for fully
flexibility of the inhibitors and simultaneous flexibility of the
active site side-chains. Each docking run included 15 000 steps
Pharmacological assays
Biological methods. Materials. [3H]DPCPX, [3H]CGS 21680
and [125I]AB-MECA were obtained from DuPont-NEN (Boston,
MA). Adenosine deaminase was from Sigma Chemical Co. (St.
Louis, MO). All other reagents were from standard commercial
sources and of the highest commercially available grade.
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 2 6 2 – 2 2 7 0
2 2 6 9

View Article Online
of Monte Carlo perturbation, subsequent fast searching, and
final energy minimisation. The results were evaluated in terms
of total estimated binding energy, internal strain energy of the
ligand, Van der Waals and electrostatic interaction energies and
for each ligand, the 20 best docked conformations, according to
the QXP score, were saved and analysed further.
8 H. W. Hamilton, D. F. Ortwine, D. F. Worth and J. A. Bristol, J. Med.
Chem., 1987, 30, 91–96; S. A. Poulsen and J. R. Quinn, J. Med. Chem.,
1996, 39, 4156–4161.
9 F. Bondavalli, M. Botta, O. Bruno, A. Ciacci, F. Corelli, P. Fossa, A.
Lucacchini, F. Manetti, C. Martini, G. Menozzi, L. Mosti, A. Ranise,
S. Schenone, A. Tafi and M. L. Trincavelli, J. Med. Chem., 2002, 45,
4875–4887 and references cited therein.
10 S. Schenone, O. Bruno, F. Bondavalli, A. Ranise, L. Mosti, G.
Menozzi, P. Fossa, F. Manetti, L. Morbidelli, L. Trincavelli, C.
Martini and A. Lucacchini, Eur. J. Med. Chem., 2004, 2, 153–160
and references cited therein.
11 F. Giordanetto, P. Fossa, G. Menozzi, S. Schenone, F. Bondavalli, A.
Ranise and L. Mosti, J. Comput.-Aided Mol. Des., 2003, 17, 39–51.
12 A. Bolognese, G. Correale, M. Manfra, A. Lavecchia, E. Novellino
and V. Barone, Org. Biomol. Chem., 2004, 2, 2809–2813 and literature
cited therein.
13 T. Y. H. Wu, P. G. Schultz and S. Ding, Org. Lett., 2003, 5, 3587–3590.
14 R. A. F. de Ligt, P. A. M. van der Klein, J. K. von Frijtag Drabbe
Ku¨nzel, A. Lorenzen, F. A. El Maate, S. Fujikawa, R. van Westhoven,
T. van den Hoven, J. Brussee and A. P. Ijzerman, Bioorg. Med. Chem.,
2004, 12, 139–149.
15 M. Rarey, B. Kramer, T. Lengauer and G. Klebe, J. Mol. Biol., 1996,
261, 470–489.
16 R. D. Cramer, III, D. E. Patterson and J. D. Bunce, J. Am. Chem.
Soc., 1988, 110, 5959–5967.
17 F. Da Settimo, G. Primofiore, S. Taliani, A. M. Marini, C. La Motta,
E. Novellino, G. Greco, A. Lavecchia, L. Trincavelli and C. Martini,
J. Med. Chem., 2001, 44, 316–327.
18 V. Colotta, D. Catarzi, F. Varano, L. Cecchi, G. Filacchioni, C.
Martini, L. Trincavelli and A. Lucacchini, J. Med. Chem., 2000, 43,
1158–1164.
19 I. M. Pirovano, A. P. IJzerman, P. J. M. van Galen and W. Soudijn,
Eur. J. Pharmacol., 1989, 172, 185–193.
CoMFA analysis. The conformers of 14–16, 35, 46, 56, 64
and 67–71, selected through the docking procedure described
above, were used as templates for all other antagonists in
order to elaborate a useful alignment of compounds 6–71. The
DISCO module implemented in Sybyl was used to optimize this
superimposition. Compound 15 (KW3902), the most active
antagonist among our data set, was selected as template
molecule for DISCO calculations.
The aligned molecules were placed one by one in a 3D cubic
˚
lattice with 2 A grid; a methyl probe with a +1 charge was used
to calculate steric and electrostatic fields, represented by Van
der Waals potential and columbic term, respectively. A 30 Kcal
mol−1 energy cut-off was applied, which means the steric and
electrostatic energies greater than 30 Kcal mol−1 are truncated
to the value and, thus, can avoid infinity of energy values
inside molecule. Regression analyses were performed applying
the partial least squares (PLS) algorithm in Sybyl; the steric field
alone and the combination of the steric and electrostatic fields
were used as structural descriptors to evaluate their correlation
with affinity (pKi) data. Cross-validations in PLS was used in
the meaning of obtaining the optimal number of components
to be used in the subsequent analyses. PLS analysis based on
least squares fit gave a correlation with a cross-validated r2 of
20 V. Colotta, D. Catarzi, F. Varano, L. Cecchi, G. Filacchioni, C.
Martini, L. Trincavelli and A. Lucacchini A, J. Med. Chem., 2000,
43, 3118–3124.
cv
0.540 with a maximum number of components set equal to five.
Final CoMFA model was generated using non-cross-validated
PLS analysis with the optimum number of components to give
21 Y. C. Cheng and W. H. Prusoff, Biochem. Pharmacol., 1973, 22,
3099–3108.
an r2 0.937. To obtain statistical confidence limits, the non-
cv
22 R. H. Erickson, R. N. Hiver, S. W. Feeney, P. R. Blake, W. J.
Rzeszotarsky, R. P. Hicks, D. G. Costello and M. E. Abreu, J. Med.
Chem., 1991, 34, 1431–1435.
23 R. F. Bruns, J. H. Fergus, E. W. Badger, J. A. Bristol, L. A. Santay,
J. D. Hartman, S. J. Hays and C. C. Huang, Naunyn-Schmiedeberg’s
Arch. Pharmacol., 1987, 335, 59–63.
24 H. Nonaka, M. Ichimura, M. Takeda, T. Kanda, J. Shimada, F.
Suzuki and H. Kase, Br. J. Pharmacol., 1996, 117, 1645–1652.
25 C. H. Liu, B. Wang, W. Z. Li, L. H. Yun, Y. Liu, R. B. Su, J. Li and
H. Liu, Bioorg. Med. Chem., 2004, 12, 4701–4707.
cross-validated analysis was repeated with ten bootstrap groups,
which yielded an r2 of 0.960 (optimum number of components
was five), SEP = 0.229, std dev = 0.014, steric contribution =
0.510 and electrostatic contribution = 0.490 (Table 4).
In ref. 35 the CoMFA protocol followed is reported in detail.
All calculations were carried out on SGI O2 workstations and
on a standard personal computer running under Linux.
26 B. K. Trivedi and R. F. Bruns, J. Med. Chem., 1988, 31, 1011–1014.
27 V. Colotta, L. Cecchi, D. Catarzi, G. Filacchioni, C. Martini, P.
Tacchi and A. Lucacchini, Eur. J. Med. Chem., 1995, 30, 133–139.
28 R. A. F. De Ligt, P. A. M. Van der Klein, J. K. von Frijag Drabbe
Kunzel, A. Lorenzen, F. Ait el Mate, S. Fujikawa, R. van Westhoven,
J. Brussee and A. P. Ijzerman, Bioorg. Med. Chem., 2004, 12, 139–149.
29 P. J. M. van Galen, P. Nissen, I. van Wijngaarden and A. P. Ljzerman,
J. Med. Chem., 1991, 34, 1202–1206.
Acknowledgements
Financial support from Italian MIUR (Cofin 2002, prot.
2002038577_002) is gratefully acknowledged.
References
1 T. V. Dunwiddie and S. Masino, Annu. Rev. Neurosci., 2001, 24, 31–
55.
2 N. J. Emptage, C. A. Reid and A. Fine, Neuron, 2001, 29, 197–208.
3 T. Takigawa and C. Alzheimer, J. Physiol., 2002, 539, 67–75.
4 B. B. Fredholm, A. P. Ijzerman, K. A. Jacobson, K.-N Klotz and J.
Linden, Pharmacol. Rev., 2001, 53, 527–552.
5 S. Hess, Expert Opin. Ther. Pat., 2001, 11, 1533–1561.
6 C. E. Muller, Expert Opin. Ther. Pat., 1997, 7, 419–440.
7 J. W. Daly, K. D. Hutchinson, S. I. Secunda, D. Shi, W. L. Padgett
and M. T. Shamin, Med. Chem. Res., 1994, 4, 293–306; S. Psychoyos,
C. J. Ford and M. A. Phillips, Biochem. Pharmacol., 1982, 31, 1441–
1442.
30 T. Terai, Y. Kita, T. Kusumoki, T. Himazaki, T. Ando, H. Horiai, A.
Akamane, Y. Shiokawa and K. Yoshida, Eur. J. Pharmacol., 1995,
279, 217–225.
31 F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp, M.
Lipton, C. Caufield, G. Chang, T. Hendrickson and W. C. Still,
J. Comput. Chem., 1990, 11, 440–467.
32 Spartan ꢀ02, http://www.wavefun.com.
33 Sybyl v. 6.8, TRIPOS Associates, St. Louis, MO.
34 C. McMartin and R. Bohaceck, J. Comput.-Aided Mol. Des., 1997,
11, 333–344.
35 A. Li, S. Moro, N. Forsyth, N. Melman, X. Ji and K. A. Jacobson,
J. Med. Chem., 1999, 42, 706–721.
2 2 7 0
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 2 6 2 – 2 2 7 0