2,9-Disubstituted-N6-(arylcarbamoyl)-8-azaadenines as new selective A3 adenosine receptor antagonists: Synthesis, biochemical and molecular modelling studies

Article abstract of DOI:10.1016/j.bmc.2005.04.063

A number of N⁶-(N-arylcarbamoyl)-2-substituted-9-benzyl-8- azaadenines, obtained by a modification of the synthetic scheme used to prepare selective A₁ ligands, by only three or two steps, are described. At first we prepared a series

Full text of DOI:10.1016/j.bmc.2005.04.063

Bioorganic & Medicinal Chemistry 13 (2005) 4679–4693
2,9-Disubstituted-N6-(arylcarbamoyl)-8-azaadenines as
new selective A3 adenosine receptor antagonists:
Synthesis, biochemical and molecular modelling studies
Giuliana Biagi,a Anna Maria Bianucci,a Alessio Coi,c Barbara Costa,b Laura Fabbrini,b
Irene Giorgi,a,* Oreste Livi,a Iolanda Micco,a Federica Pacchini,a Edoardo Santini,a
Michele Leonardi,a Fatena Ahmad Nofal,a Oreste LeRoy Salernid and Valerio Scartonia
a
`
Dipartimento di Scienze Farmaceutiche, Universita di Pisa, via Bonanno 6, 56126 Pisa, Italy
b
`
Dipartimento di Psichiatria, Neurobiologia, Farmacologia e Biotecnologie, Universita di Pisa, via Bonanno 6, 56126 Pisa, Italy
c
`
Dipartimento di Chimica, Universita degli Studi di Sassari, Via Vienna 2, 07100 Sassari, Italy
dButler University College of Pharmacy, 4600 Sunset, Indianapolis, IN 46028, USA
Received 20 December 2004; revised 22 April 2005; accepted 22 April 2005
Available online 23 May 2005
Abstract—A number of N6-(N-arylcarbamoyl)-2-substituted-9-benzyl-8-azaadenines, obtained by a modification of the synthetic
scheme used to prepare selective A1 ligands, by only three or two steps, are described. At first we prepared a series of 2-phenyl-
9-benzyl-8-azaadenines having as N6 substituent a variously substituted N-phenylcarbamoyl group. Some of these derivatives dem-
onstrated good affinity towards the A3 subtype but low selectivity. Compounds having p-CF3, p-F and p-OCH3, as substituents on
the phenylcarbamoyl group were selected as lead compounds for the second part of this study. Without modifying the N6 substi-
tuent, which would assure A3 affinity, we varied the 9 and 2 positions on these molecules to enhance selectivity. Some compounds
having a p-methyl group on the 2-phenyl substituent showed a very good affinity and selectivity for the A3 subtype, revealing the first
class of A3 adenosine receptor selective antagonists with a bicyclic structure strictly correlated to the adenine nucleus. The molecular
modelling work, carried out using the DOCK program, supplied two models which may be useful for a better understanding of the
binding modes. Both models highlighted the preferred interacting tautomeric forms of the antagonists for human A1 and A3
receptors.
Ó 2005 Elsevier Ltd. All rights reserved.
1. Introduction
nearly all A3 agonists are adenosine derivatives, such
as N6-substituted-phenyl-carbamoyl-adenosine-50-uro-
namides, prepared in 1996 by Baraldi et al.9, or the
new compound CP-608039.10
Adenosine receptors mediate many physiological func-
tions by interaction with adenosine. Four subtypes
(A1, A2A, A2B and A3) are recognised; human A3 aden-
osine receptors were the last to have been cloned1,2 and
intensively studied: in humans they are found not only
in the lung and liver, but also in the CNS, testes, heart
and the immune system.3 Activation of A3 adenosine
receptors mediates both inflammatory and pro-inflam-
matory responses4 and plays an important role in brain
ischemia,5,6 immunosuppression,7 and bronchospasm in
several animal models.8 As for A1 and A2A subtypes,
Selective A3 adenosine receptor antagonists have been
indicated as potential drugs for the treatment of asthma,
and despite anti-inflammatory responses to A3-receptor
activation, for the treatment of inflammatory condi-
tions.11 Other possible roles for these agents include
the therapy of rhinitis, an antiallergenic role and the
treatment of glaucoma and strokes.3
The research of A3 selective antagonists started from
xanthine derivatives, but these efforts largely failed;12,13
Keywords: A3 adenosine receptor antagonists; N6-(N-arylcarbamoyl)-
2-substituted-9-benzyl-8-azaadenines; A1/A3 selectivity; Molecular
modelling; Ligand–receptor interactions.
recently
a
series of pyrido[2,1-f]purine-2,4-dione
derivatives has been described as fused xanthine
derivatives with high affinity for A3 receptors.14 Other
classes of compounds have been studied, such as
*
Corresponding author. Tel.: +39 0502219549; fax: +39
0502219605; e-mail: igiorgi@farm.unipi.it
0968-0896/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.04.063

4680
G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
triazolonaphthyridines, pyridines, pyrans, flavonoids,
triazoloquinazolines, isoquinolines, triazolopyrimi-
dines15 and imidazopurines,16 and some compounds
with high affinity and selectivity have been discovered.
in the phenyl substituent increased both affinity and
selectivity.3 The carbonyl group in these structures
may function as an important H-bond acceptor.24
Previously, we have synthesised a number of N6-substi-
tuted-2-phenyl-9-benzyl-8-azaadenines as very potent
A1 ligands25–27 in which the N6-substituent was an alkyl,
cycloalkyl or aralkyl group. Our review of the literature
(see above) suggested that replacement of these N6-sub-
stituents with a phenylcarbamoyl moiety could lead to
a new class of A3 receptor antagonists. A modification
of the simple synthetic scheme used previously in our lab-
oratories to prepare selective A1 ligands25 would lead to
N6-ureidosubstituted-2-phenyl-9-benzyl-8-azaadenines
(see Scheme 1), whose adenine-like structure would en-
sure an adenosine receptor antagonist functional profile
and whose phenylcarbamoyl group would ensure a bio-
logical activity directed towards the A3 subtype.
It is known that adenine derivatives, analogous to the
physiological agonist but lacking the ribose moiety,
are antagonists of A1 adenosine receptors.17 On the
contrary, until now few A3 antagonists have a bicyclic
structure correlated to the adenine nucleus. For exam-
ple, a 7-deazaadenine18 and an adenosine derivative,
8-phenylethyl-9-ethyladenosine,19 resulted in good but
not very selective A3 antagonists.
Considering the chemical classes of A3 agonists and
antagonists one can find some different structures hav-
ing a N-phenylcarbamoyl function (Fig. 1). This func-
tional group appears to be an important recognition
element in the interaction with A3 receptors. In some
cases, insertion of that function on A1 or A2A ligands
has resulted in the discovery of very potent and selective
A3 compounds.9,20–22 For example, a series of adeno-
sine-50-uronamide derivatives bearing N6-phenylurea
groups (1, Fig. 1) are selective A3 agonists;9 or a series
of 5-amino-pyrazolo[4,3-e]-1,2,4-triazole[1,5-c]pyridines
containing a 2-furyl substituent at the 2 position, a car-
bamoyl residue at the exocyclic amine and a small group
for R (2, Fig. 1) proved to be potent A3 receptor antag-
onists.21 Small groups in the p-position of the phenyl
group interacted with the A3 receptors including F, Cl,
OCH3 and CH3. Only hydrogen is tolerated in the m-po-
sition, and the o-position could be substituted with a
chloro group.3
The first lead compounds obtained (compounds 40, 43
and 46, Table 1) proved to have good A3 activity but lit-
tle A1/A3 selectivity. Then they were submitted to a
number of structural changes to increase activity and
selectivity on A3 adenosine receptors.
In our laboratory, theoretical models for the human A1
adenosine receptors had been previously built, improved
and validated through interaction studies with
ligands.27,28 Residues that appear to be relevant for
ligand–receptor interaction and chiral selectivity in
subtype A1 had been highlighted. Here, we focus our
interest on molecular features required for antagonists
to show A1/A3 selectivity, in order to develop a predic-
tive model for designing new selective A3 antagonists.
A number of isoquinoline (3, Fig. 1) and quinazoline
derivatives (4, Fig. 1) were developed as A3 receptor
antagonists.23 In these compounds, the ureido-linker
can be replaced by an amide linker. A methoxy group
1.1. Chemistry
In the scheme are depicted the simple routes to synthe-
sise compounds 35–90. Compounds 35–80, 85 and 86
were obtained by a three-step procedure and com-
pounds 81–84, 87–90 by a method which needs only
two steps. 2,9-Disubstituted-8-azaahypoxanthines 5–17
(see Fig. 2), obtained by following procedures de-
scribed,29 were transformed by the reaction with thionyl
chloride in the corresponding chlorides, which were not
characterised and reacted with NH3 to give the corre-
sponding 8-azaadenines 18–29 and 32 (see Fig. 2). Ade-
nines 30, 31, 33 and 34 (see Fig. 2) were obtained by a
Ôone-potÕ reaction30 from the suitable azides, malononit-
rile and suitable nitriles. All 8-azaadenines synthesised
(18–34) reacted with the proper isocyanates to give
compounds 35–90. Some of the intermediate com-
pounds (5–7, 11–15, 18–20 and 25–29) have been de-
scribed in the previous papers (see Fig. 2 for references).
O
O
N
H
N
HN
N
H
N
HN
N
O
N
N
N
O
N
O
N
N
R
H
H
N
HN
R
H
H
OH
HO
2
1
O
O
N
HN
N
HN
H
H
1.2. Biochemistry
N
N
N
N
The new compounds 35–90 were tested in radioligand
binding assays to determine their affinities towards re-
combinant human A1, A2A and A3 adenosine receptors.
In particular, affinities for human A1, A2A and A3 aden-
osine receptors expressed in CHO cells were determined
in competition assays using, respectively, [3H]DPCPX,
N
3
4
Figure 1. Phenylureido derivatives with high activity towards A3
adenosine receptors.

G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
4681
Scheme 1. Reagents and conditions: (a) EtONa,110 °C; (b) RCOOEt, EtONa, 110 °C; (c) (1) POCl3, N,N-diethylaniline, 90 °C; (2) NH3, EtOH,
110 °C; (d) R2C6H4NCO, CH3CN, reflux, 60 min; (e) RCN, EtONa, 110 °C.
[3H]NECA and [125I]AB-MECA as radioligands and
have been reported in Table 1.
model for the human A1 and a new model for the human
A3 receptor were built, by exploiting their homology
with bovine rhodopsin, for which an accurate model
had been recently released in PDB (PDB entry
1L9H).33 This work started by aligning the sequences
of human A1 and A3 receptors and bovine rhodopsin,
using the CLUSTAL W program.34 The alignment is
shown in Figure 4. The biopolymer module of InsightII
was used in order to create the model for both human
A1 and A3 receptors. The correctness of fold was evalu-
ated using the Homology module of InsightII. The
structure was quite good even if not optimised, although
a series of steric clashes needed to be resolved with a fur-
ther and accurate geometry optimisation.
With the aim to evaluate the pharmacological profile of
the new synthesised compounds the agonist-induced stim-
ulationof binding of [35S]GTP-c-Stoactivated G proteins
was employed. This assay has been used as a functional
one for a variety of receptors, including adenosine recep-
tors.31,32 In particular, the effects of the most potent and
A3-selective compounds (81, 82, 83 and 84) on agonist-in-
duced stimulation of [35S]GTP-c-S binding from mem-
branes of CHO cells expressing the human A3 adenosine
receptors were studied. Each of the new synthesised com-
pound caused a concentration-dependent loss of binding
of [35S]GTP-c-S in the presence of a constant high concen-
tration of NECA (10 lM) (Fig. 3). IC50 values ranged
from 83 10 nM for the new synthesised compound
84 to 398 40 nM for the new synthesised compound
82 (Table 2). In each case the IC50 value was between
4.3- and 14-fold the Ki value obtained in a radioligand
binding assay at human A3 adenosine receptors.
In order to build models for the complexes of interest we
needed to accurately identify the binding site locations.
With regard to the A1 receptor subtype, we exploited
previous studies already carried out in our laboratory,
while for the A3 receptor subtype, the identification of
the binding site was based upon the work of Baraldi
et al.35 Then we proceeded to pre-position ligand 82,
in the binding site of both A1 and A3 receptors. The
above compound was chosen for its high affinity and
selectivity towards the A3 receptor.
In conclusion, by the functional assay it was found that
these compounds behave as antagonists for A3 adeno-
sine receptors.
The AMBER36 program was used to obtain a low-en-
ergy structure for both human A1 and A3 receptor in
complex with ligand 82. With regard to the solvation,
the explicit solvent model TIP3P water was used, resem-
bling a cap of water around the protein filled to 20 A
around the ligand. As for the minimisation parameters,
the SHAKE procedure was employed to constrain all
2. Molecular modelling
This work is a development of the previous modelling
project started with the aim of analysing the interactions
between the human A1 receptor with some antago-
nists.27,28 Based on this fact, an improved theoretical
˚

4682
G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
Table 1. Binding activity at human A1, A2A and A3 adenosine receptors
R1
R2
Kia (nM) or I (%)
A1/A3
e
Compd
R
b
c
d
hA1
hA2A
hA3
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
C6H5
C6H5
CH2C6H5
CH2C6H5
2-Cl
3-Cl
1120 100
489 50
304 31
197 20
190 18
168 15
46%
672 58
5
333 36
300 40
507 52
3.4
1.6
0.6
3.9
7.0
3.1
2.6
4.1
5.1
1.2
2.0
4.9
0.9
—
C6H5
C6H5
CH2C6H5
CH2C6H5
4-Cl
15%
42%
1
4
2-CF3
3-CF3
4-CF3
2-F
50
27
54
36
27
18
5
3
5
4
3
2
C6H5
C6H5
CH2C6H5
CH2C6H5
2260 203
25%
28%
18%
3
3
2
C6H5
C6H5
CH2C6H5
CH2C6H5
92
110 10
91
9
3-F
4-F
C6H5
C6H5
CH2C6H5
CH2C6H5
9
409 51
2-OCH3
3-OCH3
4-OCH3
4-I
1180 130
500 48
190 21
47%
43%
5
4
1000 90
250 26
C6H5
C6H5
CH2C6H5
CH2C6H5
CH2C6H5
656 59
39
5
C6H5
C6H5
1600 150
28%
35%
5%
3
4
1
2
1
1
3
2
1800 210
0%
0%
(CH2)6CH3
(CH2)6CH3
(CH2)6CH3
(CH2)2C6H5
(CH2)2C6H5
(CH2)2C6H5
o-CH2C6H4Cl
o-CH2C6H4Cl
o-CH2C6H4Cl
o-CH2C6H4F
o-CH2C6H4F
o-CH2C6H4F
p-CH2C6H4CH3
p-CH2C6H4CH3
p-CH2C6H4CH3
CH2C6H5
4-OCH3
4-CF3
4-F
30%
40%
6%
4
4
1
4
1
C6H5
C6H5
—
—
20%
3%
22%
0%
3
C6H5
C6H5
4-OCH3
4-CF3
4-F
45%
7%
—
—
3%
0%
C6H5
C6H5
0%
24%
16%
0%
28%
35%
25%
3
4
3
—
—
4-OCH3
4-CF3
4-F
42%
34%
3
4
C6H5
C6H5
—
1900 180
400 36
10%
0%
2
5
346 31
700 59
5.5
—
—
C6H5
C6H5
4-OCH3
4-CF3
4-F
56% 6
2200 200
0%
32% 3
295 30
C6H5
C6H5
48%
0%
7.5
—
—
4-OCH3
4-CF3
4-F
30%
18%
3
2
22%
29%
2
3
C6H5
C6H5
0%
800 74
16%
45%
5%
1
5
1
490 50
0%
0%
1.6
—
—
CH2C6H5
CH2C6H5
CH2C6H5
(CH2)3CH3
(CH2)3CH3
(CH2)3CH3
(CH2)2CH3
(CH2)2CH3
(CH2)2CH3
CH3
4-OCH3
4-CF3
4-F
5%
15%
7%
1
2
1
2
CH2C6H5
CH2C6H5
0%
6%
22%
0%
2
—
—
CH2C6H5
CH2C6H5
4-OCH3
4-CF3
4-F
14%
0%
1
3
1
1
3
3
2
2
30%
11%
10%
25%
30%
17%
14%
0%
—
—
CH2C6H5
CH2C6H5
7%
1
5
2
2
1
3
3
1
4
3
9
7
23%
15%
10%
25%
26%
3%
2
1
1
3
3
1
1
4-OCH3
4-CF3
4-F
48%
15%
16%
10%
25%
32%
10%
32%
27%
100
—
—
CH2C6H5
CH2C6H5
—
—
CH2C6H5
CH2C6H5
4-OCH3
4-CF3
4-F
CH3
CH3
—
—
CH2C6H5
CH2C6H5
2600 240
5%
4% 0.3
11%
H
4-OCH3
4-CF3
4-F
1
3% 0.3
5
—
—
H
CH2C6H5
CH2C6H5
30%
H
13%
1
1
6%
93
24
28
33
35
28
6
1
—
p-C6H4CH3
p-C6H4CH3
p-C6H4CH3
p-C6H4CH3
p-C6H4CH3
p-C6H4CH3
p-C6H4CH3
p-C6H4CH3
p-C6H4CH3
m-C6H4CH3
m-C6H4CH3
m-C6H4CH3
CH2C6H5
CH2C6H5
CH2C6H5
4-OCH3
4-CF3
F
13%
0%
8
1.1
3.3
24.2
c.s.
c.s.
146
71.7
17.6
31.8
9.8
4.2
19.2
80
3
678 57
0%
29%
0%
3
3
3
o-CH2C6H4Cl
o-CH2C6H4Cl
o-CH2C6H4F
o-CH2C6H4F
p-CH2C6H4CH3
p-CH2C6H4CH3
o-CH2C6H4Cl
o-CH2C6H4Cl
o-CH2C6H4F
4-CF3
4-F
4
12%
1
29%
0%
3
4-CF3
4-F
4100 320
430 39
4
01
8050 740
0%
4-CF3
4F
5000 510
700 63
284 29
22
7030 670
0%
2
4-CF3
4-F
4-CF3
4270 380
1480 150
4750 380
437 40
355 35
248 25
25%
0%
2

G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
4683
Table 1 (continued)
e
Compd
R
R1
R2
Kia (nM) or I (%)
A1/A3
b
c
d
hA1
hA2A
hA3
90
Theophillinef
DPCPXf
m-C6H4CH3
o-CH2C6H4F
4-F
1600 160
6200 530
3.2 0.2
20%
2
45
5
35.9
21,000 1800g
337 28g
86,000 1800
1300 125
a The Ki values are the mean SEM of four separate assays, each performed in triplicate.
b Inhibition of specific [3H]DPCPX binding in human A1 recombinant adenosine receptor or inhibition percentage (I %) at 10 lM concentration.
c Inhibition of specific [3H]NECA binding in human A2A recombinant adenosine receptor or inhibition percentage (I %) at 10 lM concentration.
d Inhibition of specific [125I]AB-MECA binding in human A3 recombinant adenosine receptor or inhibition percentage (I %) at 10 lM concentration.
e A1/A3 selectivity ratio values, c.s.: completely selective.
f Ref. 43.
g Value obtained using bovine striatal membranes.
100
75
50
25
0
-9
-8
-7
-6
-5
log [compound] (M)
Figure 3. Concentration–response curves for inhibition of binding of
35S]GTPcS by new compounds in the presence of a single concentra-
[
tion of the agonist NECA. Membranes of CHO cells expressing human
A3 adenosine receptors were incubated in the presence of 10 lM
NECA and increasing concentrations of the new compounds 81 (j),
82 (.) and 84 (d) as described in Experimental procedures. Each data
point represents the mean SEM of at least three independent
experiments.
Table 2. Ability of new ligands to inhibit [35S]GTP-c-binding in
membranes of cells expressing the cloned human A3 adenosine
receptor compared with receptor binding affinities
Figure 2. Chemical structures of compounds 5–34. See Refs. 25,26,29,
44–46.
Compound A3 receptor binding affinity
Ki (nM)a
[
35S]GTP-c-S binding
IC50 (nM)b
bonds involving at least one hydrogen atom. A cut-off of
˚
12 A for the non-bonded interactions was used. The
81
82
83
84
33
35
6
3
4
1
3
141 15
398 40
83 10
receptors were then minimised with harmonic position
constraints for all heavy atoms. The constraining force
constants applied were 5000, 1000, 100 and 10 kcal/
28
148 16
2
a Ki values determined in radioligand binding assays in membranes
from transfected CHO cells ( SEM).
b IC50 for inhibition by antagonists in the presence of 10 lM NECA in
membranes of CHO cells expressing the human A3 adenosine
receptors ( SEM).
˚
mol/A . Subsequently, a cycle of minimisation was done
to relax all the atoms without constraints. The maxi-
mum number of minimisation steps was set to 10,000
and the convergence criterion for energy gradient was
2
˚
set to 0.01 kcal/mol/A . All molecular dynamics (MD)
simulations were performed with AMBER. The MD
simulation involved: (1) a 20 ps equilibration step, (2)
a 50 ps molecular dynamics simulation for the actual
data collection. The time step of the simulations was
followed by an additional cycle of minimisation with
no constraints (relaxation). After the models for human
A1 and A3 receptors were obtained, ligand 82 was then
docked into the binding site by using the DOCK pro-
gram37 exploiting the flexible docking option. The dock-
ing step was followed by a further sequence of energy
minimisation, using the same gradient threshold as in
the previous one, thus leading to the optimised three-di-
˚
1.5 fs with a cut-off of 12 A for the non-bonded interac-
tions. In the data collection stage, the snapshot was
recorded in the trajectory file every 200 fs. The 10
best complexes for both the receptors were considered
and further minimised through decreasing constraints,

4684
G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
Figure 4. Multiple alignment of human A1, human A3 receptors and bovine rhodopsin.
mensional model for both A1 and A3 human receptors.
The computational analysis was carried out on com-
pounds 41, 43, 82, 84, 86 and 90, selected as significant
molecules. The selection was made to achieve a good
sampling of both selective and non-selective molecules,
thus, ensuring the discovery of selectivity requirements
(see Technical Details for the description of the minimi-
sation and docking protocols). The critical amino acids
surrounding the full antagonist 82 (Leu91, His95,
Phe182, Ile186, Val235, Leu236, Asn250, Asn274 and
Asn278) are shown in Figure 5. They appear to be the
most significant moieties of the binding site for the mol-
ecules analysed, according to details highlighted below.
The purine ring is located in a hydrophobic pocket de-
fined by Leu91 and His95. The carbamoyl moiety is ori-
ented towards Asn274 and Asn278, both located on
TM7. Interestingly, these two asparagine residues are
conserved among a large number of G protein-coupled
receptors. The interactions with this site orient the car-
bamoyl phenyl ring in the middle of TM6 and TM7.
Figure 5. Binding site for 82 in human A3 receptor.
The substituent at position 2 of the adenine nucleus,
mostly a phenyl or a p-methyl-phenyl group, is
placed in a hydrophobic pocket, delimited by apolar
amino acids, such as Ile186, Leu190 (TM5) and
Leu236 (TM6).
Additional aromatic–aromatic stabilisation of the
9-benzyl group, interacting with Phe182, may be respon-
sible for the increase in binding affinity. Site-directed

G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
4685
mutagenesis showed the inability of the N250A mutant
of A3 to bind a radio-labelled antagonist, leading to a
proposed direct interaction of this residue with
ligands.38 This could suggests a better interaction for
ligands carrying an H-bond acceptor in the o-position
of the substituent at position 9.
independent of their position in the phenyl ring. The
A1/A3receptor selectivity for these compounds ranged
from 3 to 7. Replacement of the p-CF3 or p-F with a
OCH3 substituent did not markedly change the activity,
but the OCH3 group in the o- and m-positions proved
detrimental to A3 and A1 receptor affinity. The fact that
the F atom can be an electron-donating group, whereas
the CF3 cannot, suggests that electron-withdrawing abil-
ities of both substituents is an important consideration
in their interactions with macromolecular species. Thus,
the electron-withdrawing properties of the p-OCH3 may
contribute to affinity for the A3 receptors. Compounds
40, 43 and 46 having p-CF3, p-F and p-OCH3, respec-
tively, as substituents on the phenylcarbamoyl group
were selected as lead compounds for the next phase of
the study. Maintaining the N6 substituent unchanged,
which would assure A3 affinity, we sought to vary the
9- and 2-positions on these molecules to lower affinity
towards the A1 subtype and enhance selectivity.
The p-position of the phenyl ring belonging to the N6-sub-
stituent (aromatic moiety attached to the carbamoyl
substituent), turns out to face a mostly hydrophobic area,
due to the presence of Val235 (TM6) and Asn278 (TM7).
We exploited the DOCK program with the aim of iden-
tifying the most favourable binding conformations and
orientations for the selected ligands on both models of
A1 and A3 receptors. Moreover, the DOCK program en-
ables a rough evaluation of interaction energies. These
values do not represent rigorous thermodynamic quanti-
ties, since changes in entropy and solvation effects are
only approximately taken into account. Nevertheless,
they can be reasonably used in comparisons with
relative interaction energies obtained from binding assay
experiments, in order to validate three-dimensional (3D)
theoretical models, that subsequently may be exploited
with predictive purposes. Taking into account the above
considerations, we then compared the DOCK scores
with biological data. An energetically favourable model
correlating well with the experimental results was
selected for both receptor subtypes.
3.1.2. The 9-position. Replacement of the 9-benzyl
group with alkyl or aralkyl substituents (compounds
48–53) resulted in a dramatic fall in affinity for the
three adenosine subtypes. These results reveal the 9-
benzyl group as an important structural criterion for
binding to A1 and A3 receptors in these types of aden-
osine ligands. A substituted benzyl group, however,
those with o-Cl, o-F or p-CH3, confers very little affin-
ity for A2A receptors, low affinity for A1 receptors, and
an affinity for A3 receptors in the submicromolar range
(compounds 56, 59 and 62). Based on the conforma-
tional considerations for the fit of isoquinoline and qui-
nazoline derivatives at the A3 receptors,23 one may
deduce that the benzyl substituent is oriented away
from the pyrimidine portion of the azapurine ring
and can be in the plane or rotated at 90° to the plane
of the azapurine ring.
3. Results and discussion
3.1. Structure–activity relationships
Modification on the 8-azaadenine nucleus made in the
series of compounds synthesised in this paper is at posi-
tions 2, 6 and 9. On the basis of the biological results
(Table 1) of affinity on A1, A2A and A3 some interesting
considerations about structure–activity relationships are
possible.
3.1.3. The 2-position. At position 2, replacement of the
phenyl with a variety of substituents such as H, benzyl,
or alkyl groups (butyl, propyl, methyl) caused nearly
complete loss of activity. Biological assays revealed a
low percentage of displacement at 10 lM (<32%) to-
wards A1, A2A and A3 adenosine receptors. However,
insertion of a methyl group on the m- or p-position of
the phenyl ring resulted in compounds with good affinity
for A3 receptors and showed a very good A1/A3 and
A2A/A3 selectivity (compounds 80–84, 86 and 90). In
fact, compounds 81 and 82, which resulted in complete
selectivity showing no affinity on A1 and A2A receptors
and good affinity for A3 (Ki 32.5 and 35 nM, respec-
tively), bear a p-CH3 group at the phenyl moiety on
C(2), an o-chlorobenzyl at N(9) and a p-CF3 or p-F on
the phenyl moiety at the C(6)-position, respectively.
Compound 84, the most potent of the A3-receptor
blockers, also displays a p-methylphenyl group at the
C(2), an o-fluorobenzyl at N(9) and a p-F at phenylcar-
bamoyl moiety. Also compounds 86 and 90 can be con-
sidered very interesting as A3 selective ligands having
KiA1 700 nM and 1600 nM and KiA3 22.4 and
44.6 nM, respectively. All these compounds have low
affinity for A2A receptors showing percentage inhibition
at 10 lM concentrations between 59 and 0.
3.1.1. The 6-position. The series of compounds 35–53
were synthesised to confirm the hypothesis that carbam-
oylation of the 6-amino group of 2-phenyl-6-amino-9-
benzyl-8-azaadenines would confer A3 ligand activity
on the molecules. Various substituents (Cl, CF3, F,
OCH3 and I) were introduced to the o-, m- and p- posi-
tions of the phenylcarbamoyl function to ascertain the
influence of the type and position of the substituent on
affinity towards the adenosine receptors. The biological
data were encouraging as they revealed activity in the
nanomolar range. For example, compounds 38, 39, 40,
41, 42, 43 and 46 showed Ki values ranging from
18 nM (compound 43) to 54 nM (compound 40) (see Ta-
ble 1). These compounds showed very low affinity to-
wards A2A receptors, but retained good affinity for A1
receptor subtypes with Ki values between 91 and
1600 nM. It is evident from Table 1 that the chlorophe-
nyl groups do not contribute significantly to the affinity
for the adenosine receptors, nor does the p-iodophenyl
moiety. The potent electron-withdrawing groups CF3
and the F groups produce nearly the same response,

4686
G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
3.2. Molecular modelling
3.2.2. Quality fitting and prediction reliability. With re-
gard to the predictive power of the theoretical models
developed here, it has to be pointed out that the crite-
rion followed in assessing their validity consisted of
splitting the receptor antagonists into two classes on
the basis of their affinity value. For example, in the case
of the human A1 receptor, the trend of calculated inter-
action energies was found to be in good accordance with
data coming from binding experiments for all the high
affinity antagonists (Ki up to 90 nM). For the other com-
pounds showing lower affinity (higher than 400 nM) also
the calculated interaction energies were observed to have
quite unfavourable values. An analogous observation
can be made with regard to the A3 antagonists. So both
models appear capable of discriminating between high
and low affinity molecules.
From molecular modelling studies the following findings
were obtained.
3.2.1. Analysis of tautomerism. For each one of the six
antagonists of both human A1 and A3 receptors, sub-
jected to the docking procedure by means of the DOCK
program, three runs were performed, corresponding to
one of the three tautomeric forms, here referred as to
ureic, iminol-1 and iminol-2 (Fig. 6, b–d). The DOCK
results suggest the ureic form as the preferred one for
compounds 41, 84 and 86 in A1 and compounds 43,
86 and 82 in A3. The iminol-1 tautomeric form appears
to be preferred for compounds 90 and 82 in A1. Finally,
the iminol-2 tautomeric form was suggested to be the
preferred one for compound 43 in A1 and compounds
84, 41 and 90 in A3. In most of the cases, DOCK sug-
gested the ureic form as the best for the interaction in
both receptors. On the other hand, the iminol-2 form
turns out to be the best for the interactions of 43 and
84, the most active antagonist of A1 and A3, respec-
tively. Finally, the iminol-1 tautomeric form seems to
show the lowest population in interaction with the bind-
ing site.
4. Conclusions
After the first findings regarding compounds 35–46,
which showed in some cases remarkable A3 affinity,
our efforts were addressed to improving the A1/A3 selec-
tivity. This goal was reached maintaining the N6 substi-
tuent unchanged to ensure A3 affinity and varying the
structure of groups at N(9) and C(2) positions. In sev-
eral cases, especially for compounds 80–86, 89 and 90,
we managed to lower affinity towards the A1 subtype
and to enhance A3/A1 selectivity, obtaining the first
examples of A3 potent and selective antagonists having
an adenine-like nucleus. Further, by molecular model-
ling methods we could obtain some information regard-
ing the preferred tautomeric structure of the ureic group
in interaction with binding sites, and good correlation
between selectivity ratio values and differences in calcu-
lated interaction energies with A3 and A1 subtypes.
As an example, Figure 7a shows the different orientation
of 86 in the A1 receptor. The ureic form seems to take a
more relaxed (quite planar) arrangement in the binding
site; on the other hand the iminol-2 tautomeric form is
more distorted along the carbamoyl moiety, leading to
a rotation of the aromatic ring. This fact can justify low-
er interaction energy predicted by DOCK for the ureic
form. Figures 7b and c show a stereo-view for two differ-
ent orientations of 86 in the A3 receptor. Also in this
case, the ureic form seems to keep a quite planar confor-
mation into the binding site, whereas the iminol-2 tauto-
meric form results to be slightly distorted. It may
account for the less favourable interaction energy pre-
dicted by DOCK with regard to the ureic form, even
though the scores for the two inhibitors differ very
slightly, as in the case of the A1 receptor.
The molecular modelling work, carried out by the
DOCK program, supplied two receptor models, which
appear to be extremely useful for understanding the
binding modes, and suggesting the most probable tau-
tomeric forms of antagonists for human A1 and A3
receptors. In particular, the ureic form is thought to
be the most relevant for the interaction in both the
receptors, whereas the iminol-2 form seems to be
the best for the binding of 43 and 84 in A1 and A3
receptors, respectively. The good correlation between
the experimental and calculated values for the A1
receptor and the good estimation of the A1/A3 selec-
tivity, can identify DOCK as a suitable tool for the
design of new selective antagonists towards human
A3 receptor.
5. Experimental
5.1. Chemistry
Melting points were determined on a Kofler hot-stage
apparatus and are uncorrected. IR spectra in Nujol
mulls were recorded on a Mattson Genesis series FTIR
spectrometer. 1H NMR spectra were recorded on a Bru-
ker AC 200 spectrometer in d units from TMS as an
Figure 6. (a) Main template shared by all the analysed antagonists of
A1 and A3 adenosine receptors; (b) ureic form; (c) iminolic-1 form; (d)
iminolic-2 form.

G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
4687
Figure 7. (a) Different conformations for ureic and iminol-2 forms of compound 86 in human A1 receptor, respectively coloured in magenta and light
blue. (b) Best conformer of the ureic form of compound 86 in human A3 receptor. (c) Best conformer of the iminol-2 form of compound 86 in human
A3 receptor.
internal standard; the compounds were dissolved in the
solvent indicated in Table 4. TLC was performed on
precoated silica gel F254 plates (Merck). Microanalyses
(C H N) were carried out on a Carlo Erba elemental
analyser (Model 1106) and were within 0.4% of the
theoretical values.
which was filtered and crystallised or flash-chromato-
graphed (see Tables 3 and 4).
5.3. General procedure to prepare 6-aminosubstituted-2,9-
disubstituted-8-azaadenines 18–29, 32
A mixture of a suitable 2,9-disubstituted-8-azahypoxan-
thine (2 mmol), N,N-diethylaniline (2 mmol) and POCl3
(1.5 mL, 16.3 mmol) was heated at 90 °C for 6 h. After
cooling, the residue was diluted with diethyl ether and
then filtered. The solid obtained (6-chloro-2,9-disubsti-
tuted-8-azaadenine) was not characterised but reacted
as soon as possible with absolute ethanol saturated with
NH3 (15 mL) in a well-stoppered steel flask. After 10 h
at 110 °C the mixture was concentrated at reduced pres-
sure, and the solid precipitated filtered, washed with
water and crystallised or flash-chromatographed (see
Tables 3 and 4).
5.2. General procedure to prepare 2,9-disubstituted-8-
azahypoxanthines 5–17
To a stirred solution of sodium ethoxide obtained from
sodium (0.46 g, 20 mmol) and absolute ethanol (20 mL),
cyanoacetamide (0.84 g, 0.01 mol) and suitable azide
(0.01 mol) were added and the mixture was refluxed
for 1 h. Then the appropriate ethyl ester (0.05 mol)
was added and heating was continued for 8 h. Evapora-
tion of the solvent at reduced pressure, dilution with
water and acidification with 10% acetic acid gave a solid
Table 3. Chemical and physical properties of compounds 8–10, 16, 17, 20–24 and 30–33
Compound
Yield (%)
Crystall. Solvent
Mp (°C)
Formula
17H12ClN5O
C17H12FN5O
18H15N5O
Analysis
8
26.3
29
MeOH/Isopropanol
MeOH/Isopropanol
MeOH/Isopropanol
283–285
280
C
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
9
10
16
17
20
21
22
23
24
30
31
32
33
34
50
40
262–264
270
C
MeOH
a
C18H15N5O
C19H17N5O
C18H16N6
55
64
240–242
165
198
b
46
50
CH3CN
C17H13ClN6
Isopropanol
Isopropanol
Isopropanol
THF/n-hexane
Isopropanol
Isopropanol
180–182
193–195
202
C
17H13FN6
C18H16N6
18H16N6
C18H15ClN6
18H15FN6
C19H18N6
18H15ClN6
C18H15FN6
71
47.5
26
C
223–225
205
212
20
50
C
37
35
THF/n-hexane
c
215–216
195
C
a Purification by flash-chromatography (eluent CHCl3/MeOH 100:4).
b Purification by flash-chromatography (eluent CHCl3/MeOH 100:2).
c Purification by flash-chromatography (eluent CHCl3).

4688
G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
Table 4. 1H NMR of the new synthesised compounds
Aliphatic H
2.27 (s, 3H)
Aromatic H
Benzylic H
Exch. H
8a
8.15 (m, 2H), 7.56 (m, 4H), 7.36 (m, 3H)
8.17 (m, 2H), 7.60–7.17 (m, 7H)
5.88 (s, 2H)
5.83 (s, 2H)
5.75 (s, 2H)
12.86 (s, 1H)
12.85 (s, 1H)
12.84 (s, 1H)
9a
10a
8.17 (m, 2H), 7.60 (m, 3H), 7.30 (d, J = 8.0 Hz,
2H), 7.17 (d, J = 8.0 Hz, 2H)
16b
17a
2.48 (s, 3H)
8.10 (d, J = 8.0 Hz, 2H), 7.53–7.28 (m, 7H)
8.13 (d, J = 8.4 Hz, 2H), 7.33 (d, J = 8.4 Hz, 2H),
7.28 (d, J = 8.4 Hz, 2H ), 7.15 (d, J = 8.4 Hz, 2H)
8.35 (m, 2H), 7.50 (m, 3H), 7.15 (m, 5H)
5.80 (s, 2H)
5.69 (s, 2H)
11.19 (s, 1H)
12.82 (s, 1H)
2.38 (s, 3H), 2.26 (s,
3H)
4.84 (t, J = 7.2 Hz, 2H)
20a
3.32 (t, J = 7.0 Hz, 2H)
8.38 (br s, 1H)
8.07 (br s, 1H)
6.17 (br s, 2H)
6.22 (s, 2H)
21b
22b
23b
8.46 (m, 2H), 7.47 (m, 4H), 7.25 (m, 3H)
8.47 (m, 2H), 7.50–7.07 (m, 7H),
5.98 (s, 2H)
5.89 (s, 2H)
5.78 (s, 2H)
2.33 (s, 3H)
8.48 (m, 2H), 7.50 (m, 3H), 7.43 (d, J = 8.0 Hz,
2H), 7.16 (d, J = 8.0 Hz, 2H)
7.47–7.31 (m, 10H)
5.89 (s, 2H)
24b
30b
31a
5.73 (s, 2H), 4.17 (s, 2H)
5.96 (s, 2H)
5.84 (s, 2H)
6.06 (s, 2H)
6.14 (s, 2H)
2.43 (s, 3H)
2.37 (s, 3H)
8.35 (d, J = 8.4 Hz, 2H), 7.46–7.20 (m, 6H)
8.29 (d, J = 8.2 Hz, 2H), 7.45–7.20 (m, 6H)
8.45 (br s, 1H)
8.13 (br s, 1H)
8.43 (br s, 1H)
32a
2.37 (s, 3H), 2.24 (s,
3H)
8.31 (d, J = 8.0 Hz, 2H), 7.30 (d, J = 8.0 Hz, 2H),
7.29 (d, J = 7.8 Hz, 2H) 7.15 (d, J = 7.8 Hz, 2H)
5.74 (s, 2H)
8.10 (br s, 1H)
3.45 (s, 2H)
33a
34a
2.40 (s, 3H)
2.40 (s, 3H)
7.96 (m, 2H), 7.56–7.36 (m, 6H)
8.20 (m, 2H), 7.46–7.17 (m, 6H)
5.89 (s, 2H)
5.85 (s, 2H)
8.48 (br s, 1H)
8.20 (br s, 1H)
10.99 (1H)
35a
7.14–7.60 (m, 11H), 8.14 (d, 1H), 8.39 (m,
2H + 1H exch.)
5.95 (s, 2H)
36a
37b
38b
7.18–7.82 (m, 12H), 8.41 (m, 2H + 1H exch.)
7.25–7.65 (m, 12H), 8.40 (m, 2H + 1H exch.)
7.37–7.88 (m, 12H), 8.33 (m, 2H)
5.95 (s, 2H)
5.90 (s, 2H)
5.91 (s, 2H)
11.34 (1H)
11.87 (1H)
8.52 (1H)
11.30 (1H)
8.48 (1H)
12.14 (1H)
8.48 (1H)
12.08 (1H)
11.54 (1H)
11.93 (1H)
39b
40b
7.36–8.05 (m, 12H), 8.43 (m, 2H)
7.34–7.83 (m, 12H) , 8.42 (m, 2H)
5.91 (s, 2H)
5.91 (s, 2H)
41b
42b
7.15–7.60 (m, 12H), 8.45–8.50 (m, 2H + 1H exch.)
6.90 (m, 1H), 7.35–7.64 (m, 11H), 8.42 (m,
2H + 1H exch.)
5.90 (s, 2H)
5.90 (s, 2H)
43b
44b
7.07–7.67 (m, 12H), 8.42 (m, 2H + 1H exch.)
6.92–7.59 (m, 11H), 8.25 (m, 1H), 8.55 (m, 2H)
5.90 (s, 2H)
5.89 (s, 2H)
11.80 (1H)
3.85 (s, 3H)
3.86 (s, 3H)
3.85 (s, 3H)
8.38 (s, 1H)
11.44 (s, 1H)
11.81 (s, 1H)
45b
6.75 (m, 1H), 7.21–7.62 (m, 11H), 8.44 (m,
2H + 1H exch.)
5.90 (s, 2H)
46b
47b
48b
6.95–7.62 (m, 12H), 8.42 (m, 2H + 1H exch.)
7.37–7.74 (m, 12H), 8.37 (m, 2H + 1H exch.)
8.41 (m, 2H), 7.60 (m, 5H), 6.96 (d, 2H)
5.91 (s, 2H)
5.90 (s, 2H)
12.08 (s, 1H)
11.87 (s, 1H)
11.69 (s, 1H)
8.27 (s, 1H)
4.73 (t, J = 7.0 Hz,
3H), 3.85 (s, 3H), 2.10
(m, 2H), 1.41–1.27 (m,
8H) 0.89 (t, J = 6.4 Hz,
3H)
49a
50b
4.71 (t, J = 7.0 Hz,
2H), 2.02 (m, 2H), 1.28
(m, 8H),0.82 (t,
8.41 (m, 2H), 7.81–7.61 (m, 7H + 1H exch.)
11.46 (s, 1H)
11.83 (s, 1H)
J = 6.4 Hz, 3H)
4.74 (t, J = 7.0 Hz,
2H), 2.12 (m, 2H),
1.41–1.27 (m, 8H), 0.89
(t, J = 6.4 Hz, 3H)
8.43–8.35 (m, 2H), 7.69–7.57 (m, 5H), 7.13 (d, 2H)
8.34 (s, 1H)
11.23 (s, 1H)
51a
3.74 (s 3H), 4.96 (t,
J = 6.6 Hz, 2H)
8.31 (m, 2H), 7.62 (m, 2H), 7.51 (d, J = 9.0 Hz,
2H), 720–7.11 (m, 5H), 6.98 (d, J = 9.0 Hz, 2H)
3.32 (t, J = 6.6 Hz, 2H)
11.12 (s, 1H)
12.08 (s, 1H)
8.56 (s, 1H)
52b
53a
5.00 (t, J = 7.2 Hz, 2H)
4.97 (t, J = 6.6 Hz, 2H)
8.35 (m, 2H), 7.82 (d, J = 9.0 Hz, 2H), 7.69–7.60
(m, 4H), 7.29–7.20 (m, 6H)
8.34 (m, 2H), 7.63 (m, 5H), 7.30–7.13 (m, 7H)
3.44 (t, J = 7.2 Hz, 2H)
3.37 (t, J = 6.6 Hz, 2H)
11.27 (s, 1H)
11.10 (br s, 1H)
(continued on next page)

G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
4689
Table 4 (continued)
Aliphatic H
Aromatic H
Benzylic H
6.01 (s, 2H)
Exch. H
54a
3.75 (s, 3H)
8.35 (m, 2H), 7.61–7.38 (m, 9H), 6.98 (d,
J = 8.8 Hz, 2H)
11.20 (s, 2H)
55a
56a
57a
8.39 (m, 2H), 7.81–7.38 (m, 11H)
8.38 (m, 2H), 7.67–7.22 (m, 11H)
8.39 (m, 2H), 7.63–6.20 (m, 9H), 6.98 (d,
J = 9.2 Hz, 2H)
6.03 (s, 2H)
6.03 (s, 2H)
5.98 (s, 2H)
11.38 (s, 2H)
11.24 (s, 2H)
11.20 (s, 2H)
3.75 (s, 3H)
58b
59a
60b
61b
8.40 (m, 2H), 7.84–7.15 (m, 11H)
5.98 (s, 2H)
5.99 (s, 2H)
5.86 (s, 2H)
12.08 (s, 1H)
8.55 (s, 1H)
11.26 (s, 1H)
11.24 (s, 1H)
11.65 (s, 1H)
8.40 (m, 2H), 7.66–7.19 (m, 11H)
3.85 (s, 3H) 2.34 (s,
3H)
2.34 (s, 3H)
8.41 (m, 2H + 1H exch.), 7.61–6.93 (m, 11H)
8.40 (m, 2H), 7.82 (d, J = 8.4 Hz, 2H), 7.61 (m,
5H), 7.44 (d, J = 8.2 Hz, 2H), 7.18 (d, J = 8.2 Hz,
2H)
5.86 (s, 2H)
2.34 (s, 3H)
12.09 (s, 1H)
8.63 (s, 1H)
62a
63b
2.26 (s, 3H)
3.83 (s, 3H)
8.41 (m, 2H), 7.62 (m, 5H), 7.36–7.16 (m, 6H)
7.50–7.27 (m, 10H), 7.19 (d, J = 8.8 Hz, 2H), 6.86
(d, J = 8.8 Hz, 2H)
5.89 (s, 2H)
5.82 (s, 2H)
4.36 (s, 2H)
5.84 (s, 2H)
4.41 (s, 2H)
5.83 (s, 2H)
4.38 (s, 2H)
5.84 (s, 2H)
11.26 (s, 2H)
11.14 (s, 1H)
8.28 (s, 1H)
11.56 (s, 1H)
8.35 (s, 1H)
11.26 (s, 1H)
8.20 (s, 1H)
11.73 (s, 1H)
11.17 (s, 1H)
64b
65b
66a
7.55–7.27 (m, 14H)
7.49–7.14 (m, 14H)
3.74 (s, 3H), 3.00 (t,
J = 6.8 Hz, 2H), 1.84
(m, 2H), 1.38 (m, 2H),
0.91 (t, J = 7.2 Hz, 3H)
3.01 (t, J = 6.8 Hz,
2H), 1.83 (m, 2H), 1.41
(m, 2H), 0,91 (t,
7.50 (d, J = 9.0 Hz, 2H), 7.35 (m, 5H), 6.96 (d,
J = 9.0 Hz, 2H)
67a
68a
69a
7.77 (m, 4H), 7.34 (m, 5H)
5.84 (s, 2H)
5.84 (s, 2H)
5.85 (s, 2H)
11.85 (s, 1H)
11.26 (s, 1H)
J = 7.2 Hz, 3H)
3.00 (t, J = 6.8 Hz,
2H), 1.81 (m, 2H), 1.36
(m, 2H), 0.89 (t,
7.59 (m, 2H), 7.34–7.18 (m, 7H)
11.85 (s, 1H)
11.15 (s, 1H)
J = 7.2 Hz, 3H)
3.74 (s, 3H), 2.97 (t,
J = 7.2 Hz, 2H), 1.87
(m, 2H), 0.97 (t,
7.50 (d, J = 9.0 Hz, 2H), 7.34 (m, 5H), 6.95 (d,
J = 9.0 Hz, 2H),
11.72 (s, 1H)
11.15 (s, 1H)
J = 7.2 Hz, 3H)
70b
71a
3.07 (t, J = 7.2 Hz,
2H), 2.00 (m, 2H), 1.11
(t, J = 7.2 Hz, 3H)
2.99 (t, J = 7.2 Hz,
2H), 1.88 (m, 2H), 0.97
(t, J = 7.2 Hz, 3H)
3.74 (s, 3H), 3.73 (s,
3H)
7.75–7.31 (m, 9H)
5.84 (s, 2H)
5.85 (s, 2H)
12.14 (s, 1H)
8.40 (s, 1H)
7.61 (m, 2H), 7.37–7.19 (m, 7H)
11.85 (s, 1H)
11.26 (s, 1H)
72a
73b
7.52 (d, J = 8.0 Hz, 2H), 7.34 (m, 5H), 6.94 (d,
J = 8.0 Hz, 2H)
5.83 (s, 2H)
5.81 (s, 2H)
11.62 (s, 2H)
2.86 (s, 3H)
12.12 (s, 1H)
8.35 (s, 1H)
8.35 (s, 2H)
11.44 (s, 1H)
11.19 (s, 1H)
11.86 (s, 1H)
11.42 (s, 1H)
11.61 (s, 1H)
11.30 (s, 1H)
11.65 (s, 1H)
74a
75a
3.34 (s, 3H)
3.74 (s, 3H)
8.74 (m, 2H), 8.22 (m, 2H), 7.34 (m, 5H)
8.84 (s, 1H), 7.53 (d, J = 8.8 Hz, 2H), 7.34 (m, 5H)
6.94 (d, J = 8.8 Hz, 2H)
5.92 (s, 2H)
5.88 (s, 2H)
76a
77a
78b
8.88 (s, 1H), 7.85 (d, J = 8.8 Hz, 2H), 7.72 (d,
J = 8.8 Hz, 2H), 7.35 (m, 5H)
8.86 (s, 1H), 7.67 (m, 2H), 7.36–7.22 (m, 7H)
5.90 (s, 2H)
5.90 (s, 2H)
5.88 (s, 2H)
3.85 (s, 3H)
8.30 (d, J = 8.2 Hz, 2H + 1H exch.), 7.61–7.35 (m,
9H), 6.96 (d, J = 9.0 Hz, 2H)
2.50 (s, 3H)
2.52 (s, 3H)
79b
8.30 (d, J = 8.2 Hz, 2H + 1H exch.), 7.81 (d,
J = 8.8 Hz, 2H), 7.67 (d, J = 8.8 Hz, 2H), 7.56–
7.35 (m, 7H)
5.89 (s, 2H)
12.10 (s, 1H)
11.82 (s, 1H)
80b
81b
2.51 (s, 3H)
2.51 (s, 3H)
8.29 (d, J = 8.2 Hz, 2H + 1H exch.), 7.64–7.07 (m,
11H)
5.89 (s, 1H)
6.04 (s, 2H)
8.28 (d, J = 8.8 Hz, 2H), 7.82 (d, J = 8.4 Hz, 2H),
7.68 (d, J = 8.4 Hz, 2H) 7.42–7.30 (m, 6H)
12.11 (s, 1H)
8.54 (s, 1H)
(continued on next page)

4690
G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
Table 4 (continued)
Aliphatic H
Aromatic H
Benzylic H
6.01 (s, 2H)
Exch. H
82a
83a
2.41 (s, 3H)
2.51 (s, 3H)
8.27 (d, J = 8.2 Hz, 2H), 7.67–7.22 (m, 10H)
11.28 (s, 1H)
11.25 (s, 1H)
12.11 (s, 1H)
8.57 (s, 1H)
11.30 (s, 2H)
11.47 (s, 1H)
11.32 (s, 1H)
8.29 (d, J = 8.4 Hz, 2H), 7.82 (d, J = 8.6 Hz, 2H),
7.67 (d, J = 8.6 Hz, 2H), 7.57–7.15 (m, 6H)
8.28 (d, J = 8.4 Hz, 2H), 7.78–7.06 (m, 10H)
8.32 (d, J = 8.0 Hz, 2H), 7.80 (m, 4H), 7.44 (d,
J = 8.0 Hz, 2H), 7.33 (d, J = 7.8 Hz, 2H) 7.17 (d,
J = 7.8 Hz, 2H)
5.97 (s, 2H)
84a
85a
2.40 (s, 3H)
2.43 (s, 3H)
2.26 (s, 3H)
5.90 (s, 2H)
5.88 (s, 2H)
86a
2.42 (s, 3H)
2.26 (s, 3H)
2.42 (s, 3H)
2.42 (s, 3H)
8.30 (d, J = 8.0 Hz, 2H), 7.63 (m, 2H), 7.43 (d,
J = 8.0 Hz, 2H) 7.31–7.16 (m, 6H)
8.23 (s, 2H), 7.95–7.38 (m, 10H)
5.88 (s, 2H)
11.31 (s, 1H)
11.22 (s, 1H)
12.68 (s, 2H)
11.45 (s, 1H)
11.25 (s, 1H)
12.15 (s, 1H)
8.54 (s, 1H)
11.90 (s, 1H)
8.48 (s, 1H)
87a
88a
5.98 (s, 2H)
6.02 (s, 2H)
8.17 (m, 2H), 7.64–7.27 (m, 10H)
89b
90b
2.54 (s, 3H)
2.52 (s, 3H)
8.21 (m, 2H), 7.84 (d, J = 8.6 Hz, 2H) 7.66 (d,
J = 8.6 Hz, 2H) 7.51–7.16 (m, 6H)
8.20 (m, 2H), 7.66–7.07 (m, 10H)
5.98 (s, 2H)
5.98 (s, 2H)
a Solvent = DMSO.
b Solvent = CDCl3.
5.4. General procedure to prepare 2,9-disubstituted-8-
azaadenines 30, 31, 33 and 34
Bio-Wittaker (Walkersville, MD, USA). (ꢀ)-N6-(2-phe-
nylisopropyl) adenosine (R-PIA), 50-N-ethylcarbamoyl-
adenosine (NECA) and other agents were purchased
from Sigma–Aldrich, srl.
To a stirred solution of sodium ethoxide obtained from
sodium (0.46 g, 20 mmol) and absolute ethanol (20 mL),
malononitrile (0.66 g, 0.01 mol) was added and the mix-
ture was heated at reflux for 15 min. Then a solution of a
suitable azide (0.01 mol) and a suitable nitrile (0.1 mol)
in absolute ethanol (2 mL) was added and the mixture
was refluxed for 5 h. Evaporation of the solvent at re-
duced pressure, dilution with water and acidification
with 10% acetic acid gave a solid, which was repeatedly
washed with hot diethyl ether and then with water. The
product obtained was purified by crystallisation or flash
chromatography (see Tables 3 and 4).
Affinity of new synthesised compounds towards human
A1 adenosine receptors (ARs) was evaluated by compe-
tition experiments as described by Baraldi et al.,39 using
[3H]DPCPX (111Ci/mmol) as a radioligand.
CHO cells, stably transfected with human A1ARs, were
grown and maintained in DulbeccoÕs modified EagleÕs
medium with F12 nutrient mixture without nucleosides,
containing 10% fetal calf serum, penicillin (100 U/mL),
streptomycin (100 lg/mL), L-glutamine (2 mM), and
geneticin (0.2 mg/mL) at 37 °C in 5% CO2/95% air.
The cells were washed with phosphate-buffered saline
solution and were scraped from flasks in ice-cold hypo-
tonic buffer (5 mM Tris–HCl, 2 mM EDTA, pH 7.4).
The cell suspension was homogenised with Polytron,
and the homogenate was centrifuged at 48,000g for
30 min. The membrane pellet was resuspended in T1
buffer (50 mM Tris–HCl, 2 mM MgCl2, 2 U/mL ADA,
pH 7.4). Aliquots of membranes (50 lg of proteins) ob-
tained from A1 CHO cells were incubated at 25 °C for
180 min in 500 lL T1 buffer, containing 3nM
[3H]DPCPX and different concentrations of new synthe-
sised compounds. Nonspecific binding was determined
in the presence of 50 lM R-PIA. The dissociation con-
stant (Kd) of [3H]DPCPX in CHO cell membranes was
3 nM.
5.5. General procedure to prepare 2,9-disubstituted-6-
[(aryl)amino]carbonyl]amino-8-azapurines 35–90
A solution of 1 (100 mg, 0.34 mmol) and of a suitable
phenylisocyanate (0.68 mmol) in acetonitrile (5 mL)
was refluxed for 1 h. Before cooling, the hot reaction
mixture was filtered to collect the solid precipitated,
which resulted in the pure reaction product. Some of
these products were not recrystallised (see Tables 4
and 5).
6. Biochemistry
6.1. Materials and methods
N6-[4-Amino-3-[125I]iodobenzyl) 50-N-methycarbamoyl-
adenosine ([125I]ABMECA), [3H]-8-cyclopenty-1,3-
dipropylxanthine ([3H]DCPCX) and [3H]-50-N-ethyl-
carboxamidoadenosine ([3H]NECA) were obtained
from NEN Life Science Products, Inc. (Koln, Ger-
many). Adenosine deaminase (ADA) was obtained from
Boeringer–Mannhein (Mannhein, Germany). Cell cul-
ture media and fetal calf serum were obtained from
Affinity towards human A2AARs was evaluated by com-
petition experiments as described by Klotz et al.,40 using
[3H]NECA as a radioligand. CHO cells, stably transfec-
ted with human A2AARs, were grown adherently and
the membrane fraction was prepared as described above.
Aliquots of cell membranes (80 lg of proteins) were
incubated at 25 °C for 90 min in 500 lL T2 buffer
(50 mM Tris–HCl, 2 mM MgCl2, 2 U/mL ADA,

G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
Table 5. Chemical and physical properties of the compounds 35–90
4691
Compd
Yield (%)
Crystall. Solventa
Mp (°C)
Formula
24H18ClN7O
Analysis
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
96
80
90
95
92
86
90
91
90
94
92
85
92
75
58
58
84
53
99
55
60
70
55
60
92
40
56
89
95
56
55
96
90
85
85
80
80
50
65
62
82
80
72
61
70
69
80
53
50
42
58
80
70
40
30
50
Isopropanol
Isopropanol
Isopropanol
Isopropanol
Isopropanol
Isopropanol
Isopropanol
Isopropanol
Isopropanol
Isopropanol
Isopropanol
Isopropanol
Isopropanol
Isopropanol
183
193
C
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
(C,H,N)
C24H18ClN7O
C24H18ClN7O
C25H18F3N7O
224
169
177–180
235–237
206
C
C
25H18F3N7O
25H18F3N7O
C24H18FN7O
24H18FN7O
C24H18FN7O
25H21N7O2
C25H21N7O2
25H21N7O2
210
220
C
176
190–191
187
C
C
242–245
182
200
C24H18IN7O
C22H29N7O2
C25H26F3N7O
188
202
C24H26FN7O
C26H23N7O2
C26H20F3N7O
212
210–211
225
268
C
25H20FN7O
C25H20ClN7O2
25H17ClF3N7O
C24H17ClFN7O
25H20FN7O2
THF/n-hexane
THF
THF/n-hexane
C
210
220
C
THF
THF/n-hexane
240
215
C25H17F4N7O
C24H17F2N7O
C26H23N7O2
230
255
C26H20F3N7O
THF/n-hexane
EtOH
216
174
C25H20FN7O
C26H23N7O2
C26H20F3N7O
183–185
183
180
C25H20FN7O
C23H25N7O2
C23H22F3N7O
C22H22FN7O
185
180–181
190
198
C22H23N7O2
C22H20F3N7O
190–193
260
191
C
C
21H20FN7O
20H19N7O2
C20H16F3N7O
C19H16FN7O
C19H17N7O2
THF
THF/Isopropanol
Isopropanol
232–233
218
225
C19H14F3N7O
224
235
C18H14FN7O
C26H23N7O2
C26H20F3N7O
THF/n-hexane
THF/n-hexane
THF/n-hexane
THF/isopropanol
THF/isopropanol
232
222
C
C
25H20FN7O
26H19ClF4N7O
244
220
C25H19ClFN7O
C26H19F4N7O
C25H19F2N7O
250
246–248
253
250
C27H22F3N7O
C26H22FN7O
C26H19ClF3N7O
CHCl3/MeOH
243–244
262 dec.
222
C
25H19ClFN7O
C26H19F4N7O
C25H19F2N7O
CHCl3
245
a Where crystallisation solvent is not indicated, mp was determined on the pure crystalline product directly precipitated from reaction mixture and
not recrystallised.
pH 7.4) in the presence of 30 nM [3H]NECA and differ-
ent concentrations of new synthesised compounds. Non-
specific binding was determined in the presence of
100 lM NECA. The dissociation (Kd) of [3H]NECA in
A2A CHO cell membranes was 30 nM.
Affinity towards human A3ARs was evaluated by compe-
tition experiments as described by Colotta et al.,41 using
[
fected with human A3ARs, were grown adherently and
the membrane fraction was prepared as described above.
125I]ABMECA as a radioligand. CHO cells, stably trans-

4692
G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
Aliquots of cell membranes (40 lg of proteins) were incu-
bated at 25 °C for 90 min in 100 lL T3 buffer (50 mM
Tris–HCl, 10 mM MgCl2, 1 mM EDTA, 2 U/mL ADA,
pH 7.4) in the presence of 1.4 nM [125I]ABMECA and dif-
ferent concentrations of new synthetised compounds.
Nonspecific binding was determined in the presence of
50 lM R-PIA. The dissociation (Kd) of [125I]ABMECA
in A3 CHO cell membranes was 1.4 nM.
were followed by at least 20,000 steps of conjugated gra-
˚
dient, with a threshold of 0.01 kcal/A. The subsequent
molecular docking procedure was performed by using
the DOCK program (version 4.0.1).37 The search for
optimal orientations and conformations of the ligands
in the binding site was carried out by the use of the flex-
ible docking option in the program. All torsion angles of
the ligands were so allowed to change and many com-
plex configurations were collected. The corresponding
approximate interaction energies were also collected
and compared in order to identify optimal orientations
and conformations of the ligands. Further, energy min-
imisations (same protocol as above) were performed on
the best configurations of both models for complexes
involving A1 and A3 receptors, so that the whole struc-
ture could relax to the best near energy minima. It made
the analysis of interactions, detailed at molecular level,
more reliable.
In all competition experiments, the compounds were dis-
solved in DMSO and added to the assay mixture
(DMSO concentration maximum 2%). The experiments
(n = 3), carried out in triplicate, were analysed by an
iterative curve-fitting procedure (GraphPad, Prism pro-
gram, San Diego, CA), which provided IC50, and SEM
values for compounds tested. IC50 values were converted
to Ki values using the Cheng and Prusoff equation.42
Protein levels were evaluated through the method of
Lowry, using bovine serum albumin as standard.
The binding of [35S]GTPcS was carried out as described
by Jacobson et al.32 using CHO cells expressing human
A3 receptors. Briefly, membranes were suspended in a
buffer containing 50 mM Tris, 3 U/mL adenosine deam-
inase, 100 mM NaCl, and 10 mM MgCl2, pH 7.4 at a
protein concentration of 5–10 lg per tube. The mem-
brane suspension was preincubated with 0.5 lM GDP,
10 lM NECA and an antagonist in a final volume of
450 lL buffer at 30 °C for 20 min and transferred to
ice for 20 min. [35S]GTPcS was added to a final concen-
tration of 0.1 nM in a total volume of 500 lL and the
mixture was determined in the presence of 10 lM
GTPcS. Incubation of the reaction mixture was termi-
nated by filtration over GF/C glass fibres using a Bran-
del cell harvester and washed with the same buffer.
Acknowledgements
We thank Dr. Karl-Norbert. Klotz (Institut fur Pharma-
¨
¨
kologie und Toxikologie, Universita¨t Wurzburg, Wurz-
¨
burg, Germany) for providing cloned human A1, A2A
and A3 receptors expressed in CHO cells. This research
`
was supported by the Ministero Istruzione Universita
Ricerca Scientifica (MIUR).
References and notes
1. Zhou, Q. Y.; Li, C. Y.; Olah, M. E.; Johnson, R. A.; Stiles,
G. L.; Civelli, O. Proc. Natl. Acad. Sci. U.S.A. 1992, 89,
7432.
2. Salvatore, C. A.; Jacobson, M. A.; Taylor, H. E.; Linden,
J.; Johnson, R. G. Proc. Natl. Acad. Sci. U.S.A. 1993, 15,
10365.
6.2. Molecular modelling (technical details)
3. Muller, C. E. Curr. Top. Med. Chem. 2003, 3, 445.
4. Fishman, P.; Bar-Yehuda, S. Curr. Top. Med. Chem. 2003,
3, 463.
5. von Lubitz, D. K. J. E.; Carter, M. F.; Deutsch, S. I.; Lin,
R. C. S.; Mastropaolo, J.; Meshulam, Y.; Jacobson, K. A.
Eur. J. Pharmacol. 1995, 275, 23.
6. von Lubitz, D. K. J. E.; Lin, R. C. S.; Popik, P.; Carter,
M. F.; Jacobson, K. A. Eur. J. Pharmacol. 1994, 263, 59.
7. Mackenzie, W. M.; Hoskin, D. W.; Blay, J. Cancer Res.
1994, 54, 3521.
¨
All calculations were carried out on a Silicon Graphics
Octane R12000 workstation, equipped with the Accelrys
package containing a graphic tool (Insight) [Insight,
Accelrys Inc., San Diego CA] and a computation tool
(Discover) [Discover, Accelrys Inc., San Diego CA],
which is capable of performing molecular mechanics
(MM) calculations and molecular dynamics (MD) simu-
lations. Insight was exploited for building the three-di-
mensional models for human A1 and A3 receptors and
some antagonists, while energy minimisation was per-
formed through Discover.
8. Meade, C. J.; Mierau, J.; Leon, I.; Ensinger, H. A. J.
Pharmacol. Exp. Ther. 1996, 279, 1148.
9. Baraldi, P. G.; Cacciari, B.; Spalluto, G.; Ji, X-d.; Olah,
M.; Stiles, G.; Dionisotti, S.; Zocchi, C.; Ongini, E.;
Jacobson, K. A. J. Med. Chem. 1996, 39, 802.
10. DeNinno, M. P.; Masamune, H.; Chenard, L. K.; DiRico,
K. J.; Eller, C.; Etienne, J. B.; Tickner, J. E.; Kennedy, S.
P.; Knight, D. R.; Kong, J.; Oleynek, J. J.; Tracey, W. R.;
Hill, R. J. J. Med. Chem. 2003, 46, 353.
The two initial models for complexes involving ligand 82
and both receptors were energy minimised by using the
cff91 force field of Discover according to the protocol
˚
described below. A sphere of water molecules of 18 A
radius, centred on the ligand, was added to each model
of the complexes. As a consequence of explicitly repre-
senting the solvent, a value of 1 was used for the dielec-
tric constant in the MM calculations aimed at energy
minimisation. The dielectric constant was also set to as
distance dependent. First of all 100 steps of iterations
based on the steepest descents algorithm were performed
until the threshold on the energy derivative with respect
11. Ramkumar, V.; Stiles, G. L.; Beaven, M. A.; Ali, H. J.
Biol. Chem. 1993, 268, 16887.
12. Kim, H. O.; Ji, X-d.; Melman, N.; Olah, M.; Stiles, G.;
Jacobson, K. A. J. Med. Chem. 1994, 37, 4020.
13. Kim, H. O.; Ji, X-d.; Melman, N.; Olah, M.; Stiles, G.;
Jacobson, K. A. J. Med. Chem. 1994, 37, 3373.
14. Priego, E-M.; von Frijtag Drappe Kuenzel, F.; IJzerman,
` `
A. P.; Camarasa, M. J.; Perez-Perez, M-J. J. Med. Chem.
2002, 45, 3337.
˚
to the atomic coordinates was less than 1 kcal/A. They

G. Biagi et al. / Bioorg. Med. Chem. 13 (2005) 4679–4693
4693
15. van Muijlwijk-Koezen, J. E.; Timmerman, H.; IJzerman,
A. P. Prog. Med. Chem. 2001, 38, 61.
16. Saki, M.; Tsumuki, H.; Nonaka, H.; Shimada, J.; Ichim-
ura, M. Eur. J. Pharmacol. 2002, 444, 133.
17. Muller, C. Exp. Opin. Ther. Patents 1997, 7, 419.
¨
18. Hess, S.; Muller, C. E.; Frobenius, W.; Reith, U.; Klotz,
¨
32. Jacobson, K. A.; Park, K. S.; Jiang, J. L.; Kim, Y. C.;
Olah, M. E.; Stiles, G. L.; Ji, X. D. Neuropharmacology
1997, 36, 1157.
33. Okada, T.; Fujiyoshi, Y.; Silow, M.; Navarro, J.; Landau,
E. M.; Shichida, Y. Proc. Natl. Acad. Sci. U.S.A. 2002, 99,
5982.
K.-N.; Eger, K. J. Med. Chem. 2000, 43, 4636.
19. Volpini, R.; Costanzi, S.; Lambertucci, C.; Vittori, S.;
Klotz, K.-N.; Lorenzen, A.; Cristalli, G. Bioorg. Med.
Chem. Lett. 2001, 11, 1931.
20. Baraldi, P. G.; Cacciari, B.; Pineda de las Infantas, M. J.;
Romagnoli, R.; Spalluto, G.; Volpini, R.; Costanzi, S.;
Vittori, S.; Cristalli, G.; Melamn, N.; Park, K-S.; Ji, X-d.;
Jacobson, K. A. J. Med. Chem. 1998, 41, 3174.
21. Baraldi, P. G.; Cacciari, B.; Romagnoli, R.; Spalluto, G.;
Moro, S.; Klotz, K.-N.; Leung, E.; Varani, K.; Grassi, S.;
Merighi, S.; Borea, P. A. J. Med. Chem. 2000, 43, 4768.
22. Baraldi, P. G.; Tabrizi, M. A.; Bovero, A.; Avitabile, B.;
Preti, D.; Fruttaro, F.; Romagnoli, R.; Varani, K.; Borea,
P. A. Eur. J. Med. Chem. 2003, 38, 367.
34. (a) Pearson, W. R.; Lipman, D. J PNAS 1988, 85,
2444–2448; (b) Pearson, W. R. Methods Enzymol. 1990,
183, 63.
35. Baraldi, P. G.; Cacciari, B.; Moro, S.; Spalluto, G.;
Pastorin, G.; Da Ros, T.; Klotz, K.-N.; Varani, K.; Gessi,
S.; Borea, P. A. J. Med. Chem. 2002, 45, 770.
36. Pearlman, D. A.; Case, D. A.; Cladwell, J. W.; Ross, W.
S., III; Cheatman, T. E., III; DeBolt, S.; Ferguson, D.;
Seibel, G.; Kollman, P. A. Comput. Phys. Commun. 1995,
91, 1.
37. Ewing, T. J. A.; Kuntz, I. D. J. Comput. Chem. 1997, 18,
1175.
38. Gao, Z. G.; Chen, A.; Barak, D.; Kim, S. K.; Muller, C.
E.; Jacobson, K. A. J. Biol. Chem. 2002, 277, 19056.
39. Baraldi, P. G.; Romagnoli, R.; Pavani, M. G.; Nunez
Mdel, C.; Tabrizi, M. A.; Shryock, J. C.; Leung, E.;
Moorman, A. R.; Uluoglu, C.; Iannotta, V.; Merighi, S.;
Borea, P. A. J. Med. Chem. 2003, 46, 794.
23. van Muijlwijk-Koezen, J. E.; Timmerman, H.; van der
Goot, H.; Menge, W. M. P. B.; Frijtag von Drabbe
Kunzel, J.; de Groote, M.; IJzerman, A. P. J. Med. Chem.
¨
2000, 43, 2227.
24. Colotta, V.; Catarsi, D.; Varano, F.; Cecchi, L.; Filacchi-
oni, G.; Martini, C.; Trincavelli, L.; Lucacchini, A. J.
Med. Chem. 2000, 43, 1158.
40. Klotz, K. N.; Hessling, J.; Hegler, J.; Owman, C.; Kull, B.;
Fredholm, B. B.; Lohse, M. J. Naunyn Schmiedebergs Arch
Pharmacol. 1998, 357, 1.
25. Biagi, G.; Giorgi, I.; Livi, O.; Scartoni, V.; Lucacchini, A.;
Martini, C.; Tacchi, P. V. Farmaco 1994, 49, 187.
26. Biagi, G.; Giorgi, I.; Livi, O.; Scartoni, V.; Lucacchini, A.
Farmaco 1996, 51, 395.
27. Bianucci, A. M.; Biagi, G.; Coi, A.; Giorgi, I.; Livi, O.;
Pacchini, F.; Scartoni, V.; Lucacchini, A.; Costa, B. Drug
Dev. Res. 2001, 54, 52.
28. Bianucci, A. M.; Bigi, M. U.; Biagi, G.; Giorgi, I.; Livi, O.;
Scartoni, V. A. Drug Des. Discovery 1998, 15, 149.
29. Barili, P. L.; Biagi, G.; Livi, O.; Scartoni, V. J. Heterocycl.
Chem. 1985, 22, 1607.
41. Colotta, V.; Catarzi, D.; Varano, F.; Filacchioni, G.;
Martini, C.; Trincavelli, L.; Lucacchini, A. Arch. Pharm.
(Weinheim) 2004, 337, 35.
42. Cheng, Y.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22,
3099.
43. Colotta, V.; Catarzi, D.; Varano, F.; Calabri, F. R.;
Lenzi, O.; Filacchioni, G.; Martini, C.; Trincavelli, L.;
DeFlorian, F.; Moro, S. J. Med. Chem. 2004, 47,
3580.
44. Biagi, G.; Giorgi, I.; Livi, O.; Scartoni, V.; Lucacchini, A.
XI. Farmaco 1992, 47, 1457.
30. Biagi, G.; Franchi, M.; Giorgi, I.; Livi, O.; Scartoni, V. J.
Heterocycl. Chem. 1989, 26, 39.
31. Lorenzen, A.; Guerra, L.; Vogt, H.; Schwabe, U. Mol.
Pharmacol. 1996, 49, 915.
45. Biagi, G.; Giorgi, I.; Livi, O.; Scartoni, V.; Lucacchini, A.;
Martini, C.; Tacchi, P. IV. Farmaco 1994, 49, 183.
46. Barili, P. G.; Biagi, G.; Mucci, L.; Livi, O.; Scartoni, V. J.
Heterocycl. Chem. 1987, 24, 997.