8-Substituted-9-deazaxanthines as adenosine receptor ligands: Design, synthesis and structure-affinity relationships at A2B

Article abstract of DOI:10.1016/j.ejmech.2004.07.008

A number of 8-substituted-9-deazaxanthine derivatives (1,3-dialkyl-6- substituted-1H-pyrrolo[3,2-d]pyrimidine-2,4(3H,5H)-diones) were prepared and tested for their antagonistic activity at the recombinant human adenosine receptors, in particular at the A<

Full text of DOI:10.1016/j.ejmech.2004.07.008

European Journal of Medicinal Chemistry 39 (2004) 879–887
www.elsevier.com/locate/ejmech
Short communication
8-Substituted-9-deazaxanthines as adenosine receptor ligands:
design, synthesis and structure-affinity relationships at A2B
Angelo Carotti a,*, Angela Stefanachi a, Enrique Raviña b, Eddy Sotelo b, Maria Isabel Loza c,
Maria Isabel Cadavid c, Nuria B. Centeno d, Orazio Nicolotti a
a Dipartimento Farmacochimico,Università di Bari, via Orabona 4, 70125 Bari, Italy
b Departamento de Quimica Organica, Universidade de Santiago de Compostela, 15782 Santiago de Compostela, Spain
c Departamento de Farmacologia, Universidade de Santiago de Compostela, 15782 Santiago de Compostela, Spain
d Grup de Recerca en Informàtica Biomèdica (GRIB), IMIM-UPF, C/Dr. Aiguader 80, 08003 Barcelona, Spain
Received 12 March 2004; received in revised form 24 June 2004; accepted 7 July 2004
Abstract
A number of 8-substituted-9-deazaxanthine derivatives (1,3-dialkyl-6-substituted-1H-pyrrolo[3,2-d]pyrimidine-2,4(3H,5H)-diones) were
prepared and tested for their antagonistic activity at the recombinant human adenosine receptors, in particular at the A2B and A2A receptor
subtypes. Compounds endowed with micromolar to nanomolar binding affinities, but with poor A2B/A2A selectivity, were obtained.
Preliminary quantitative structure–affinity relationships suggested that the binding potency at the A2B receptor is mainly modulated by the
electronic and lipophilic properties of the ligands.
© 2004 Elsevier SAS. All rights reserved.
Keywords: 1,3-Dialkyl-6-substituted-1H-pyrrolo[3,2-d]pyrimidine-2,4(3H, 5H)-diones; Adenosine receptor antagonists; A2B and A2A binding affinities;
A2A/A2B selectivity; Quantitative structure–affinity relationships
1. Introduction
patic glucose balance [21], cell growth [22], gene expression
[23], mast cell degranulation [24] and intestinal water secre-
tion [25]. More recently, a growing interest for the discovery
of selective A2B receptor antagonists stemmed from their
potential use as antidiabetic [26] and antiasthmatic agents
[27,28].
Adenosine is a nucleoside that modulates many important
physiological functions [1,2] by interacting at four different
G-protein coupled receptor subtypes [3] named A1, A2A, A2B
and A3.[4] A1 and A2A adenosine receptors (ARs) are stimu-
lated by low adenosine concentrations, whereas higher ade-
nosine levels are required for the activation of A2B and A3
ARs [5].A huge number ofAR agonists and antagonists have
been synthesized so far [6–9] and potent and subtype selec-
tive ligands, with high potential in many therapeutic fields,
have been discovered. Selective ligands for A1, A2A and,
more recently, A3 ARs have been made available for biologi-
cal and pharmacological studies [10–14]. As far as selective
A2B receptor ligands are concerned, only few compounds
with a relatively low selectivity have been described [15–19].
Selective A2B receptor antagonists are therefore actively pur-
sued since they play a pivotal role in the control of a variety
of physiological functions including vascular tone [20], he-
Xanthine derivatives constitute, along with purine nucleo-
side analogues and condensed tricyclic nitrogen heterocyclic
derivatives [6,7,9,13], one of the most exploited class of AR
ligands [17–19,29–31]. In spite of that, xanthines and closely
related analogues, appear still appealing since the antiasth-
matic activity shown by some xanthine drugs, (i.e. theophyl-
line and enprofylline), has been associated to a slight, but
significant, antagonism at the A2B receptor subtypes [32].
As anticipated, many xanthine derivatives have been
already investigated as AR antagonists whereas
9-deazaxanthines (9-dAXs) have been only rarely studied
[31,33], especially for their antagonist activity at A2B recep-
tors [17]. Therefore, some years ago we begun a systematic
study addressing the synthetic accessibility to fully functio-
nalized 9-dAXs and a deep exploration of their biological
and pharmacological properties. Herein we report the design,
* Corresponding author. Tel.: +39-080-5442782; fax: +39-080-5442230.
E-mail address: carotti@farmchim.uniba.it (A. Carotti).
0223-5234/$ - see front matter © 2004 Elsevier SAS. All rights reserved.
doi:10.1016/j.ejmech.2004.07.008

880
A. Carotti et al. / European Journal of Medicinal Chemistry 39 (2004) 879–887
Table 1
Chemical structures and affinity binding data of 9-dAX derivatives 1b–19b
Cp
1b
2b
3b
4b
5b
6b
7b
R
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
n.C3H7
n.C3H7
n.C3H7
n.C3H7
n.C3H7
n.C3H7
R1
C6H5
R2
H
H
H
H
H
H
H
H
H
H
CH3
CH3
CH3
H
H
OH
H
pKi a hA2B
6.74
7.16
7.24
20%
0%
pKi a hA2A
6.32 b
6.73
pKi hA1
pKi hA3
4-OH-C6H4
4-NH2-C6H4
4-NHAc-C6H4
2-F-C6H4
2,6-F2-C6H3
2-Thienyl
3-Thienyl
5-Br-2-thienyl
2-Furyl
C6H5
2-Thienyl
2-Furyl
C6H5
4-NH2-C6H4
C6H5
6.66
21%
13%
0%
6.09
6.58
9%
7.00
7.14
7.41
6.73
1% c
3%
12%
6.64
7.78
6.17
7.59
6.79
6.02
8b
9b
6.98
6.42
10b
11b
12b
13b
14b
15b
16b
17b
18b
19b
10% c
16%
18%
6.19 d
6.95
7.47 d
7.60
7.02
30%
32%
6.80
5.87
4-OCH3-C6H4
4-OC3H7-C6H4
4-OCH2C6H5-C6H4
H
H
5.60
a Binding affinity is expressed as pKi or percent of displacement at 0.1 µmol (1 µmol for compound 16b at hA3). Ki and percent of displacement had a
SEM <10%. For compound 16b, the EC50 at hA2B is reported.
b Lit. pKi (rat brain, rA2), 6.29 [33].
c Lit. pKi value (rat brain) 4.40 and 4.95 at rA2 and rA1, respectively, [33].
d Lit. pKi values (rat brain) 6.35 and 7.88 at rA2 and rA1, respectively, [33].
synthesis and AR binding affinities, mainly at the A2B and
2A subtypes, of a series of 1,3-dialkyl-(7,)8-(di)substituted-
2. Chemistry
A
Compounds 1b, 2b, 4b–10b, 14b, 18b and 19b were
prepared starting from the corresponding 1,3-dialkyl-6-
methyl-5-nitrouracyls according the synthetic pathway outli-
ned in Scheme 1 [33].
The amino derivatives 3b and 15b were obtained trough
the basic hydrolysis of the corresponding 1,3-dimethyl [4b]
and 1,3-dipropyl [35] acetamido derivatives, respectively,
whereas the N7-methyl derivatives 11b–13b were synthesi-
9-deazaxanthines, whose chemical structures are reported in
Table 1. Our main goals were the discovery of new molecular
entities endowed with high affinity and, possibly, good selec-
tivity towards the A2B receptor subtype and the development
of preliminary structure–affinity relationships, at A2B recep-
tors, and structure–selectivity (A2B/A2A) relationships able
to suggest further structural modifications of the 9-dAX
skeleton to improve both potency and selectivity.
Scheme 1. Synthetic pathway to 9-dAX derivatives.

A. Carotti et al. / European Journal of Medicinal Chemistry 39 (2004) 879–887
881
zed by alkylation of the corresponding N7H derivatives 1b,
7b and 10b with methyl iodide and K2CO3 in DMF. Finally,
the p-OCH3 phenyl derivative 17b the N7–OH derivative 16b
were prepared from the corresponding 5-nitro-6-styryl-
uracyl precursor through an efficient reductive cyclization
reaction, recently reported by some of us [35].
water partition coefficient can be assessed by ACD [39] and
c-LOGP [40] software, respectively, for 9-dAXs compared
with the corresponding xanthine analogues.
The above observations were in full agreement with the
quite diverse biological profiles, structure–activity and struc-
ture–selectivity relationships observed for the two classes of
ligands [41].
As far as the AR affinity of 9-dAXs is concerned, an
examination of the data in Table 1 revealed that relatively
potent, but poorly selective, A2B receptor ligands were obtai-
ned.
3. Pharmacology
Compounds were tested for their ability to displace [3H]-
DPCPX, [3H]-ZM241385, [3H]-DPCPX and [3H]-NECA
from cloned human A1, A2A, A2B and A3 ARs, respectively.
Assays were carried out by coincubation of compounds, in at
least six different concentrations, with the appropriate ra-
dioactive ligand. Adenylyl cyclase functional assay was per-
formed to determine the intrinsic activity of compound 16b
at A2B receptors.
Interestingly, affinities atA2B receptor subtype were found
to be generally higher than those observed at A2A receptor,
with the most selective compounds 18b, 7b and 15b having a
Ki (A2A)/Ki (A2B) ratio of 8.3, 8.1 and 6.8, respectively.
Isosteric substitution of the 8-phenyl ring of the 1,3-
dimethyl-9-dAXs generally improved the activity; the most
active isoster, i.e. the 3-thienyl derivative 8b, had a
pKi = 7.14. The introduction of a 5-bromo substituent in the
thienyl ring of compound 7b, yielded derivative 9b that
displayed a higher affinity at both A2A and A2B receptors.
Notably, the less lipophilic 2-furyl derivative 10b, showed at
the A2B receptor subtype an affinity lower than the thienyl
isosters 7b and 8b. Substitution at the para position of the
8-phenyl ring with hydrophilic and/or electron-donor groups
afforded a significant increase of activity (see cps 2b, 3b and
17b) whereas a hydrophilic substituent with a low electronic
effect induced a drastic decrease of affinity (i.e. cp 4b). An
even more dramatic drop of affinity was observed in the
ortho-substituted fluoro derivatives 5b and 6b. Most likely, a
strong intramolecular HB between the H at N7 and the ortho-
fluoro substituent of the phenyl at C8 might preclude, or
strongly limit, the formation of a favourable HB (donor)
interaction of the N7H with an acceptor group on the recep-
torial counter part. This hypothesis is supported by the lack
of activity observed in N7-methylated derivatives 11b–13b.
Interestingly, the N7–OH derivative 16b, which may be able
4. Results and discussion
Preliminarily, the 9-dAX ring system was compared to
that of xanthines, to detect some physicochemical differen-
ces, especially in terms of electronic characteristics, ability to
make hydrogen bonds (HBs), water solubility and partition
in apolar/polar immiscible solvents. A straightforward com-
parison of the molecular electrostatic potentials (MEPs)
[36,37], performed by means of the MIPSIM software [38],
of two model compounds, that are 8-methyltheophylline and
its corresponding 9-deaza analogue, depicted in Fig. 1, indi-
cates the lack of the deepest MEP minimum of
8-methyltheophylline in the corresponding 9-deaza analo-
gue, a fact that may strongly influence the ability of making
HB and/or participating to other electrostatic (polar) interac-
tions.
As for the physicochemical properties relevant to the
ADME profile, both a higher water solubility and octanol/
Fig. 1
. MEP[36] minimum regions of 8-methyltheophylline (yellow contours, left-hand side) and of its 9-deaza analogue (red contours, right-hand side). The
values of the local deepest minima are indicated in kcal/mol.

882
A. Carotti et al. / European Journal of Medicinal Chemistry 39 (2004) 879–887
to engage an HB (donor) interaction, recovered some affinity
at A2B and, at least in part, at A2A AR subtypes.
pointing findings can be to some extent related to the pre-
sence of some peculiar structural patterns that the model was
still unable to quantify. And in fact, a close look at the
structural features of the molecules in the dataset easily led to
the identification of two particular groups of compounds: the
first is constituted by two ortho-fluoro substituted derivatives
(5b and 6b) at the 8-phenyl ring and the second by three
N7–CH3 derivatives (11b–13b). As expected, the removal of
these two groups of molecules consistently increased the
model fitting capability (r2 = 0.653). It should be added that
such a model came from only 13 compounds whose activity
was spread in a very narrow range (6.20–7.78). The need to
recover most of the structural and biological information
contained in our dataset prompted us to make use of nominal
variables to explain the impact on QSAR of the ortho-fluoro
and of N7–CH3 substituted derivatives. Our efforts was awar-
ded by the following equation:
It is worth noting that the N7–OH derivative 16b is a
potent and highly selective ligand at the A1 receptor. Finally,
a further structural variation was explored by replacing the
methyl groups at positions 1,3 of 1b and 3b with n-propyl
groups. While for the former, as already reported [33], a little
change of activity was observed (i.e. cp 14b), a strong in-
crease of activity was measured for its para-NH2-substituted
congener 15b. Within the examined series of compounds, the
anilino derivative 15b proved to be the most active ligand at
the A2B receptor subtype.
To gain, at a quantitative level, further insights on the most
important physicochemical properties modulating the bin-
ding affinity at A2B receptor, biological data in Table 1 were
subjected to a QSAR study by means of a multivariate linear
regression analysis [42] (Hansch-type approach) [43] with
cross validation. Unfortunately, our dataset was too small for
a secure application of 3D QSAR methods.
p Ki
=
5.310 0.472 - 0.230 0.091
r
+
1.200
͑
͒
͑
͒
0.310
cLOGP–0.181
0.043
cLOGP2–1.937
͑
͒
͑
͒
Classical substituent (molecular) descriptors assessing
electronic (r, EHOMO, ELUMO) steric (MR, van der Waals
volumes, Verloop (L, B1, B5)) and lipophilic parameters
(LOGP, p were taken from standard compilations [44] or
calculated by means of known software such as the c-LOGP
[40]. The ones used for the development of regression equa-
tion shown below are listed in Table 2.
Compound 16b was not considered in the QSAR analysis
as it presented unique structural elements compared to the
congeneric series of compounds listed in Table 1. Moreover,
for this compound only the intrinsic activity was measured.
Preliminary one- and two-term models derived from r and
c-LOGP, the latter tested both as a linear and as a quadratic
term, resulted in very poor statistics (r2 < 0.30). These disap-
0.276 I1–2.097
0.211 I2
͒
͑
͒
͑
n = 18, r2 = 0.933, s = 0.308, q2 = 0.851
where n, r2, s and q2 are statistical parameters indicating the
numbers of data points, the squared correlation coefficient,
the standard deviation and the squared cross-validation cor-
relation coefficient (based on the leave-one-out procedure
[45]), respectively. Figures in parentheses represent the stan-
dard errors of the regression coefficients.
It must be underlined that for ligands with low binding
affinity at A2B receptor (up to 12% displacement at 0.1 µM),
that are compounds 5b, 6b, 11b, 12b and 13b, a censored
(truncated) pKi value, arbitrarily set to 5.00, was used for the
Table 2
Affinity binding data and descriptors used for the derivation of QSAR equation
c
c
Cp
pKi (obs)
6.74
7.16
7.24
6.20
5.00
5.00
7.00
7.14
7.41
6.73
5.00d
5.00d
5.00d
6.64
7.78
7.59
6.79
6.02
pKi (calc) a
7.16
7.06
6.92
6.64
5.07
4.93
6.98
7.05
6.90
6.81
5.16
4.98
4.87
7.10
7.80
7.29
6.68
6.05
r b
c-LOGP
2.35
1.69
1.13
1.37
2.51
2.66
2.22
2.01
2.81
1.74
2.64
2.50
2.03
4.47
3.24
4.40
5.46
6.17
I1
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
I2
0
0
0
0
0
0
0
0
0
0
1
1
1
0
0
0
0
0
1b
2b
3b
4b
5b
6b
7b
8b
–0.01
–0.10
–0.21
–0.01
0.08
0.16
0.05
–0.02
0.16
9b
10b
11b
12b
13b
14b
15b
17b
18b
19b
0.02
–0.01
0.05
0.02
–0.01
–0.21
–0.08
–0.08
–0.08
a Values calculated from our QSAR equation.
b Hammett sigma constants refer to the R1 substituent.
c Indicator variables (see text for definition).
d Censored (truncated) values.

A. Carotti et al. / European Journal of Medicinal Chemistry 39 (2004) 879–887
883
derivation of QSAR. Moreover, in the development of the
5. Conclusion
QSAR models, the r values reported in Table 2 were multi-
plied by 10 to obtain a new set of values, more or less
equiscalar with those of c-LOGP, I1 and I2.
The results reported in the present paper demonstrated
that 8-aryl(heteroaryl)substituted-9-deazaxanthines might
be considered a promising class of compounds to develop
potent and, hopefully, selective ligands at the ARs, in parti-
cular at theA2B receptor subtype. The most salient features of
the SAFIR and, to a lesser extent, SSR (structure–selectivity
relationship) coming from the present study have been used
as a starting point for the design of new selectiveA2B receptor
ligands. The outcome of these ongoing studies will be repor-
ted in due course in forthcoming papers.
In the above equation, r refers to the Hammett constant of
the R1 substituents whereas I1 and I2 are indicator variables
that take into account the presence/absence of ortho-fluoro
substituents at the 8-phenyl ring and of N7–CH3 groups,
respectively. Briefly, I1 is set to 1 for compounds 5b and 6b
and to 0 for all the remaining ones while I2 is set to 1 for
compounds 11b–13b with all the others being set to 0. As
repeatedly observed in QSAR modelling, [43] a linear rela-
tionship with lipophilicity can hold only up to a certain
maximum LOGP (or p) value (formally termed LOGP0 or
p0) beyond which a decrease of activity can be observed, and
this is the case also for our equation: the compound with the
highest activity (15b, pKi = 7.78) showed a c-LOGP = 3.24,
a value close to the theoretical LOGP0 (nearly 3.30). Moreo-
ver, the high activity of 15b is thought to derive also by its
strong electron-donor properties. A drastic lowering of acti-
vity was observed for ligands with c-LOGP >>3.24 such as
compounds 14b, 18b and 19b. A parabolic profile (data not
shown) resulted by plotting c-LOGP vs. pKi being the three
compounds just mentioned located on the bottom descending
part of the curve.
6. Experimental protocols
6.1. Chemistry
General Information. All reactions were carried out under
an argon atmosphere. Reagent grade solvents were dried
according to standard techniques. Magnesium sulphate was
used as a drying agent for water containing organic phases.
Thin-layer chromatography (TLC) was performed by using
aluminium sheets precoated with silica gel 60 F254 (0.2 mm)
E. Merck. Chromatographic spots were visualised by UV
light or Hannessian reagent. Purification of crude com-
pounds and separation of reaction mixtures were carried out
by column chromatography on silica gel 60 (0.040–
0.063 mm, E. Merck) using appropriate eluents. Melting
points (uncorrected) were determined in a Gallenkamp mel-
ting point apparatus. Chemical shifts (d) are reported in parts
per million (ppm) relative to the solvent central peak . Cou-
A further way to prove the goodness of the applicability
domain of the above equation is the analysis of the leverages
[46] which represent the distance of a given compound from
the centre of the experiments. The ability of making a predic-
tion is smoothly reduced when the leverage value moves
away from such a centre. It is generally accepted that a
warning leverage results when the value 3p/n is overcome. In
this ratio, p represents the number of parameters of a model
and n the number of objects, respectively. Very satisfying all
the leverages values calculated for our equation were lower
than 3p/n.
1
pling constant (J values) are given in Hertz (Hz). HNMR
spectra were recorded at 300 MHz on a Bruker AMX-300
spectrometer. Low resolution mass spectra were determined
in a quadrupolar Hewlett-Packard 5988 mass spectrometer
by electronic impact (MS-EI). The molecular weights and
the fragmentation patterns from the MS spectra (not repor-
ted) were in full agreement with the proposed chemical
structures. The purity of intermediates 1a–10a, 14a, 18a was
checked by HPLC (Waters 1525 Binary HPLC pump, Phe-
nomenex C18). Elemental analyses were performed on a
Carlo Erba C, H, N analyser. All final products had satisfac-
tory C, H, N analyses (within 0.40 of theoretical values).
The syntheses of 8-aryl-9-deazaxanthines 1b [33], 14b [33]
16b [35] and 17b [35] have been already described.
In order to confirm that the model shown in the equation
did not represent a spurious regression, a randomisation test
was conducted. In doing that, the dependent variables of
Table 2 were scrambled 30 times and a linear regression
analysis was performed for each resulting data sequence. The
best model, among the 30 ones obtained, showed statistics
(r2 = 0.667 and q2 = 0.324) much poorer than those observed
in the unscrambled equation. This result made us more confi-
dent on the statistical reliability of our QSAR model.
Within the limit of the low number of data points and the
narrow physicochemical space explored, our QSAR equation
suggested that the binding affinity at A2B receptors increases
as the electron-donor character of substituents at position
8 and the overall molecular lipophilicity increase, but only up
to the optimal c-LOGP values which is close to 3.30. The
high and negative regression coefficients associated with the
I1 and I2 indicator variables, quantified well the dramatic
drop of affinity caused by both the ortho-phenyl substitution
and the N7-methylation.
6.1.1. General procedure for the condensation between
aromatic aldehydes and 1,3-dialkyl-6-methyl-5-nitrouracils
A solution of the 6-methyl-5-nitro-pyrimidine-2,4-dione
(30 mmol), the appropriate aldehyde (30 mmol) and piperi-
dine (45 mmol) in dry dioxane was refluxed for 5–8 h. The
solution was concentrated under vacuum and the residue was
treated with ethanol under stirring until a precipitate was
formed. The solid was collected by filtration and dried under
vacuum to yield the expected 6-alkenyl-substituted uracyl
derivative.

884
A. Carotti et al. / European Journal of Medicinal Chemistry 39 (2004) 879–887
6.1.1.1. 6-[(E)-2-(4-hydroxyphenyl)-vinyl]-1,3-dimethyl-5-
nitro-pyrimidine-2,4(1H,3H)-dione [2a]. Yield 35%
1HNMR (acetone d6): 7.50 (d, J = 14.9 Hz, 1H), 6.90 (m,
5H), 3.50 (s, 3H), 3.45 (s, 3H).
triethyl phosphite (5 ml) was refluxed for 7 h. The solid
formed upon cooling was collected, washed with ethyl ether
and crystallized from a water/MeOH mixture to yield the
9-deazaxanthine derivatives listed below:
6.1.1.2. N-{4[(E)-2-(1,3-dimethyl-5-nitro-2,6-dioxo-1,2,3,6-
tetrahydropyrimidin-4-yl)-vinyl]-phenyl}-acetamide [4a].
Yield 38% 1HNMR (DMSO d6): 7.58 (m, 4H), 6.83 (m, 2H),
3.28 (s, 3H), 3.17 (s, 3H), 2.17 (s, 3H).
6.1.2.1. 6-(4-hydroxyphenyl)-1,3-dimethyl-1H-pyrrolo-[3,2-
d]-pyrimidine-2,4(3H,5H)-dione [2b]. Yield: 15%;
m.p.>250 °C 1HNMR (DMSO d6): 12.1 (s, 1H), 9.70 (s, 1H),
7.72 (d, 2H, J = 8.6 Hz), 6.79 (d, 2H, J = 8.6 Hz), 6.53 (s, 1H),
3.40 (s, 3H), 3.24 (s, 3H).
6.1.1.3.
6-[(E)-2-(2-fluorophenyl)-vinyl]-1,3-dimethyl-5-
nitro-pyrimidine-2,4(1H,3H)-dione [5a]. Yield 42%
1HNMR (CDCl3): 7.31 (d, J = 14.9 Hz, 1H), 7.40 (m, 4H),
6.95 (d, J = 14.9 Hz, 1H), 3.44 (s, 3H), 3.36 (s, 3H).
6.1.2.2. N-[4-(1,3-dimethyl-2,4-dioxo-2,3,4,5-tetrahydro-
1H-pyrrolo[3,2-d]-pyrmidin-6-yl)-phenyl]-acetamide [4b].
Yield: 25%; m.p.>250 °C HNMR (DMSO d6): 12.27 (s,
1
6.1.1.4. 6-[(E)-2-(2,6-difluorophenyl)-vinyl]-1,3-dimethyl-
5-nitro-pyrimidine-2,4(1H,3H)-dione [6a]. Yield 22%
1HNMR (CDCl3): 7.29 (m, 1H), 7.18 (d, 1H, J = 18.0 Hz),
7.05 (d, 1H, J = 18.0 Hz), 6.79 (m, 2H), 3.51 (s, 3H), 3.43 (s,
3H).
1H), 10.05 (s, 1H), 7.83 (d, 2H, J = 8.7 Hz), 7.62 (d, 2H,
J = 8.7 Hz), 6.63 (s, 1H), 3.40 (s, 3H), 3.24 (s, 3H), 2.43 (s,
3H).
6.1.2.3. 6-(2-Fluorophenyl)-1,3-dimethyl-1H-pyrrolo-[3,2-
d]-pyrimidine-2,4(3H,5H)-dione [5b]. Yield 37% m.p.>
6.1.1.5. 1,3-dimethyl-5-nitro-6-[(E)-2-(thien-2-yl)-vinyl]-
pyrimidine-2,4(1H,3H)-dione [7a]. Yield 53% 1HNMR
(CDCl3): 7.49 (m, 1H), 7.30 (m, 1H), 7.26 (d, 1H,
J = 16.4 Hz), 7.11 (m, 1H), 6.47 (d, 1H, J = 16.4 Hz), 3.56 (s,
3H) 3.47 (s, 3H).
1
250 °C HNMR (DMSO d6): 12.42 (s, 1H), 7.94 (m, 1H),
7.34 (m, 3H), 6.56 (s, 1H), 3.42 (s, 3H), 3.26 (s, 3H).
6.1.2.4. 6-(2,6-Difluorophenyl)-1,3-dimethyl-1H-pyrrolo-
[3,2-d]-pyrimidine-2,4(3H,5H)-dione [6b]. Yield 12%
1
m.p.>250 °C HNMR (DMSO d6): 12.40 (s, 1H), 7.51 (m,
6.1.1.6.
1,3-dimethyl-5nitro-6-[(E)-2-(thien-3-yl)-vinyl]-
pyrimidine-2,4(1H,3H)-dione [8a]. Yield 35% 1HNMR
(CDCl3): 7.92 (m, 1H), 7.67 (m, 2H), 7.10 (m, 2H), 3.42 (s,
3H), 3.36 (s, 3H).
1H), 7.27 (m, 2H), 6.44 (s, 1H), 3.40 (s, 3H), 3.25 (s, 3H).
6.1.2.5.
1,3-Dimethyl-6-(thien-2-yl)-1H-pyrrolo-[3,2-d]-
pyrimidine-2,4(3H,5H)-dione [7b]. Yield 22% m.p.>250 °C
1HNMR (DMSO d6) 12.50 (s, 1H), 7.68 (dd, 1H, J = 3.7,
1.0 Hz), 7.56 (dd, 1H, J = 5.0, 1.0 Hz), 7.12 (m, 1H), 6.45 (s,
1H), 3.39 (s, 3H), 3.24 (s, 3H).
6.1.1.7. 1,3-dimethyl-5-nitro-6-[(E)-2-(5-bromo-thien-2-yl)-
vinyl]-pyrimidine-2,4(1H,3H)-dione [9a]. Yield 47%
1HNMR (CDCl3): 7.11 (d, 1H, J = 16.5 Hz), 7.00 (m, 2H),
6.30 (d, 1H, J = 16.5 Hz), 3.44 (s, 3H) 3.34 (s, 3H).
6.1.2.6.
1,3-Dimethyl-6-(thien-3-yl)-1H-pyrrolo-[3,2-d]-
6.1.1.8. 6-[(E)-2-(fur-2-yl)-vinyl]-1,3-dimethyl-5-nitro-pyri-
midine-2,4(1H,3H)-dione [10a]. Yield 39% 1HNMR
(CDCl3): 7.68 (m, 1H), 7.62 (m, 4H), 3.56 (s, 3H) 3.47 (s,
3H).
pyrimidine-2,4(3H,5H)-dione [8b]. Yield 20% m.p.>250 °C
1HNMR (DMSO d6): 12.33 (s, 1H), 8.06 (dd, 1H, J = 2.4,
1.6 Hz), 7.65 (m, 2H), 6.60 (s, 1H), 3.40 (s, 3H), 3.25 (s, 3H).
6.1.1.9. 6-[(E)-2(4-propoxyphenyl)-vinyl]-1,3-dipropyl-5-
nitro-pyrimidine-2,4(1H,3H)-dione [18a]. Yield 41%
1HNMR (CDCl3): 7.38 (d, 2H, J = 8.8 Hz), 6.99 (d, 1H,
J = 16.5 Hz), 6.91 (d, 2H, J = 8.8 Hz), 6.45 (d, 1H,
J = 16.5 Hz), 3.95 (m, 6H), 1.77 (m, 6H) 1.00 (m, 9H).
6.1.2.7. 1,3-Dimethyl-6-(5-bromo-thien-2-yl)-1H-pyrrolo-
[3,2-d]-pyrimidine-2,4(3H,5H)-dione [9b]. Yield 17%
m.p.>250 °C HNMR (DMSO d6): 12.57 (s, 1H), 7.43 (d,
1H, J = 3.9 Hz), 7.20 (d, 1H, J = 3.9 Hz), 6.49 (s, 1H), 3.38 (s,
3H), 3.24 (s, 3H).
1
6.1.1.10. 6-[(E)-2(4-benzyloxyphenyl)-vinyl]-1,3-dipropyl-
5-nitro-pyrimidine-2,4(1H,3H)-dione [19a]. Yield 41%
1HNMR (CDCl3): 7.34 (m, 7H), 6.99 (m, 3H), 6.45 (d, 1H,
J = 16.5 Hz), 5.10 (s, 2H) 3.90 (m, 4H), 1.71 (m, 4H) 0.96 (m,
6H).
6.1.2.8. 6-(Fur-2-yl)-1,3-dimethyl-1H-pyrrolo-[3,2-d]-pyri-
midine-2,4(3H,5H)-dione [10b]. Yield 14% m.p.>250 °C
1HNMR (DMSO d6):12.50 (s, 1H), 7.76 (d, 1H, J = 1.3), 7.06
(d, 1H, J = 3.2 Hz), 6.61 (dd, 1H J = 3.2, 1.3 Hz), 6.46 (s, 1H),
3.40 (s, 3H), 3.24 (s, 3H).
6.1.2. General procedure for the reductive cyclization
reaction of 1,3-dialkyl-5-nitro-6-alkenyl-substituted ura-
cyls to 1-H-pyrrolo-[3,2-d]-pyrimidin-2,4-dione derivati-
ves
A solution of the appropriate 1,3-dialkyl-5-nitro-6-
alkenyl-substituted uracyl intermediate (3.00 mmol) in
6.1.2.9. 6-(4-Propoxyphenyl)-1,3-dipropyl-1H-pyrrolo-[3,2-
d]-pyrimidine-2,4(3H,5H)-diones [18b]. Yield 33%.
m.p. >179–180 °C. 1HNMR (DMSO d6): 12.18 (s, 1H), 7.83
(d, 2H, J = 8.7 Hz), 6.96 (d, 2H, J = 8.7 Hz), 6.62 (s, 1H), 3.91
(m, 6H), 1.62 (m, 6H), 0.81 (m, 9H).

A. Carotti et al. / European Journal of Medicinal Chemistry 39 (2004) 879–887
885
6.1.2.10. 6-(4-Benzyloxyphenyl)-1,3-dipropyl-1H-pyrrolo-
[3,2-d]-pyrimidine-2,4(3H,5H)-diones [19b]. Yield 37%.
m.p. >198–199 °C. 1HNMR (DMSO d6): 12.18 (s, 1H), 7.84
(d, 2H, J = 8.8 Hz), 7.38 (m, 5H) 7.06 (d, 2H, J = 8.8 Hz),
6.63 (s, 1H), 5.13 (s, 2H) 3.84 (m, 4H), 1.61 (m, 4H), 0.87(m,
6H).
6.1.4.2. 1,3,5-Trimethyl-6-(thien-2-yl)-1H-pyrrolo-[3,2-d]-
pyrimidine-2,4(3H,5H)-dione[12b]. Yield 95%. m.p. 187–
188 °C. 1HNMR (CDCl3): 7.42 (dd, 1H, J = 4.9, 1.0 Hz), 7.22
(m, 1H), 7.13 (m, 1H), 6.01 (s, 1H), 4.08 (s, 3H), 3.45 (s, 3H)
and 3.41 (s, 3H).
6.1.4.3.
1,3,5-Trimetyl-6-(fur-2-yl)-1H-pyrrolo-[3,2-d]-
6.1.3. Synthesis of 6-(4-aminophenyl)-1,3-dialkyl-1H-pyr-
rolo-[3,2-d]-pyrimidine-2,4(3H,5H)-diones [3b] and[15b]
N-[4-(1,3-dimethyl-2,4-dioxo-2,3,4,5-tetrahydro-1H-pyr-
rolo[3,2-d]-pyrimidin-6-yl)-phenyl]-acetamide [4b] or N-[4-
(1,3-dipropyl-2,4-dioxo-2,3,4,5-tetrahydro-1H-pyrrolo[3,2-
d]-pirimidin-6-yl)-phenyl]-acetamide [35] (0.160 mmol)
was dissolved in a mixture of 10% aqueous solution of NaOH
(1 ml) and EtOH (1 ml), followed by 2 h refluxing. The solid
precipitate was filtered, washed with EtOH and crystallized
from a water/MeOH mixture.
pyrimidine-2,4(3H,5H)-dione[13b]. Yield 95%. m.p. 187–
188 °C. 1HNMR (CDCl3): 7.53 (d, J = 1.5 Hz, 1H), 6.63 (d,
J = 3.4 Hz, 1H), 6.50 (m, 1H), 6.17 (s, 1H), 4.17 (s, 3H), 3.46
(s, 3H) and 3.40 (s, 3H).
6.2. Pharmacology
6.2.1. Radioligand binding assays
Radioligand binding competition assays were performed
in vitro using A1, A2A, A2B and A3 human receptors ex-
pressed in transfected CHO (A1), HeLa (A2A and A3) and
HEK-293 (A2B) cells. The experimental conditions used are
given in details in Table 3.
6.1.3.1. 6-(4-Aminophenyl)-1,3-dimethyl-1H-pyrrolo-[3,2-
d]-pyrimidine-2,4(3H,5H)-dione
[3b].
Yield
45%.
m.p. >250 °C. 1HNMR (DMSO d6): 11.96 (s, 1H), 7.59 (d,
2H, J = 8.6 Hz), 6.59 (d, 2H, J = 8.6 Hz), 6.43 (s, 1H), 5.45
(br, 2H), 3.40 (s, 3H) and 3.25 (s, 3H).
In each instance aliquots of membranes (15 µg for A1,
10 µg forA2A, 18 µg forA2B and 90 µg forA3) in bufferA (see
Table 3) were incubated for the specified period of time at
25 °C with the radioligand (2–35 nM) and six different
concentrations (ranging from 1 nM to 5 µM) of the test
molecule, or standard, in a final volume of 200 µl. The
binding reaction was stopped by rapid filtration in a multis-
creen manifold system (Millipore Iberica, Madrid, Spain).
Unbound radioligand was removed by washing 4× with
250 µl ice-cold buffer B for A1 and A2A receptors and 6×
250 µl ice-cold buffer B for A2B and A3 receptors. Non-
specific binding was determined using a 50–400 µM NECA
soln for A2A and A2B receptors and 10–100 µM R-PIA soln
for A1 and A3. Radioactivity retained on filters was determi-
ned by liquid scintillation counting using Universol (ICN
Biochemicals, Inc.). The binding affinities were determined
using [3H]-DPCPX as the radioligand for A1 and A2B, [3H]-
ZM241385 for A2A and [3H]-NECA for A3. The inhibition
constant (Ki) was calculated from IC50 by the Cheng and
Prusoff [47] equation Ki = IC50/(1+(C/KD)), where IC50 is
the concentration of compound that displaces the binding of
radioligand by 50%, C is the free concentration of radioli-
gand and KD is its apparent dissociation constant.
6.1.3.2. 6-(4-Aminophenyl)-1,3-dipropyl-1H-pyrrolo-[3,2-
d]-pyrimidine-2,4(3H,5H)-dione[15b].
Yield
35%.
m.p. >250 °C. 1HNMR (DMSO d6): 11.89 (s, 1H), 7.57 (d,
2H, J = 8.3 Hz), 6.56 (d, 2H, J = 8.3 Hz), 6.43 (s, 1H), 5.38
(br, 2H), 3.81 (m, 4H), 1.60 (m, 4H), 0.87 (m, 6H).
6.1.4. Synthesis of 6-(aryl or heteroaryl)-1,3,5-trimethyl-
1H-pyrrolo-[3,2-d]-pyrimidine-2,4(3H,5H)-diones [11b],
[12b] and[13b]
One mmol of the appropriate 9-deazaxanthine was dissol-
ved in DMF (3 ml). K2CO3 (2.76 g, 20 mmol) and MeI
(0.62 ml, 10 mmol) were added, followed by heating at 50 °C
for 4 h. Upon the addition of water (10 ml) the desired
compound, precipitated from the solution, was filtered and
recrystallized from EtOH.
6.1.4.1.
1,3,5-Trimethyl-6-phenyl-1H-pyrrolo-[3,2-d]-
pyrimidine-2,4(3H,5H)-dione[11b]. Yield 93%. m.p. 190–
191 °C. 1HNMR (CDCl3): 7.44 (m, 5H), 5.95 (s, 1H), 3.95 (s,
3H), 3.47 (s, 3H) and 3.41 (s, 3H).
Table 3
Experimental conditions used for radioligand binding assays on recombinant human A1, A2A, A2B, and A3 receptors
hA1
hA2A
hA2B
hA3
Buffer A
Buffer B
20 mM Hepes, 100 mM NaCl,
10 mM MgCl2, 2 units/ml ade-
nosine deaminase (pH 7.4)
50 mM Tris–HCl, 1 mM EDTA, 50 mM Tris–HCl, 1 mM EDTA, 50 mM Tris–HCl, 1 mM EDTA,
10 mM MgCl2, 2 units/ml ade-
nosine deaminase (pH 7.4)
10 mM MgCl2, 0.1 mM benza- 5 mM MgCl2, 2 units/ml adeno-
midine, 2 units/ml adenosine
deaminase (pH 6.5)
sine deaminase (pH 7.4)
20 mM Hepes, 100 mM NaCl,
10 mM MgCl2, (pH 7.4)
GF/C
[3H]-DPCPX 2 nM
10 µM (R)-PIA
50 mM Tris–HCl, 1 mM EDTA, 50 mM Tris–HCl (pH 6.5)
10 mM MgCl2 (pH 7.4)
50 mM Tris–HCl (pH 7.4)
Plate
GF/C
GF/B
GF/B
Radioligand
Non-specific
binding
[3H]-ZM241385 3 nM
50 µM NECA
[3H]-DPCPX 35 nM
400 µM NECA
[3H]-NECA 30 nM
100 µM (R)-PIA
Incubation
25 °C/60 min
25 °C/30 min
25 °C/30 min
25 °C/180 min

886
A. Carotti et al. / European Journal of Medicinal Chemistry 39 (2004) 879–887
[12] C.E. Müller, Curr. Top. Med. Chem. 3 (2003) 445–462.
6.2.2. cAMP assay
[13] P.G. Baraldi, M.A. Tabrizi, A. Bovero, B. Avitabile, D. Preti, F. Frut-
tarolo, R. Romagnoli, K. Varani, P.A. Borea, Eur. J. Med. Chem. 38
(2003) 367–382.
The assay was performed on human A2B receptors trans-
fected in CHO cells by the method described by Salomon
[48]. Briefly cells were seeded in 12-well culture plates and
incubated at 37 °C in an atmosphere of 5% CO2 in Dul-
becco’s Modified Eagle’s Medium Nutrient Mixture F-12
(DMEM F-12), containing 10% foetal calf serum (FCS) and
1% L-glutamine. This medium was replaced 24 h before the
assays by a medium containing dialyzed FCS. Prior to the
assay, [3H]-adenine was added to the medium (3 µCi/ml) and
cells were incubated for 2 h in 5% CO2 atmosphere at 37 °C.
Cells were washed 3× with 1 ml of a medium, constituted by
DMEM-F12 and 25 mM HEPES (pH 7.4) and pre-incubated
with assay medium containing 30 µM rolipram at 37 °C for
15 min. Compound 16b was incubated for 15 min at 37 °C.
Reaction was stopped by adding ice-cold 300 mM perchloric
acid containing [14C]-cAMP and cells were maintained at
4 °C for 30 min. The [3H]-cAMP elicited in each well was
isolated by chromatographic methods and [14C]-cAMP
content allowed to calculate the yield of this isolation. The
potency of the compound was expressed as EC50, that is the
concentration that elicited 50% of maximal response.
[14] M.P. De Ninno, H. Masamune, L.K. Chenard, K.J. Di Rico, C. Eller,
J.B. Etienne, J.E. Tickner, S.P. Kennedy, D.R. Knight, J. Kong,
J.J. Oleynek, R. Tracey, R.J. Hill, J. Med. Chem. 46 (2003) 353–355.
[15] R. Volpini, S. Costanzi, S. Vittori, G. Cristalli, K.-N. Klotz, Curr. Top.
Med. Chem. 3 (2003) 427–433.
[16] Y.C. Kim, X. Jin, N. Melman, J. Linden, K.A. Jacobson, J. Med.
Chem. 43 (2000) 1165–1172.
[17] A.M. Hayallah, J. Sandoval-Ramírez, U. Reith, U. Schobert, B. Pre-
iss, B. Schumacher, J.W. Daly, C.E. Müller, J. Med. Chem. 45 (2002)
1500–1510.
[18] S.-A. Kim, M.A. Marshall, N. Melman, H.S. Kim, C.E. Müller,
J. Linden, K.A. Jacobson, J. Med. Chem. 45 (2002) 2131–2138.
[19] P.G. Baraldi, M.A. Tabrizi, D. Preti,A. Bovero, R. Romagnoli, F. Frut-
tarolo, N.A. Zaid, A.R. Moorman, K. Varani, S. Gessi, S. Merighi,
P.A. Borea, J. Med. Chem. 47 (2004) 1434–1447.
[20] A. Rubino, V. Ralevic, G. Burnstock, Br. J. Pharmacol. 115 (1995)
648–652.
[21] H. Harada, O. Osamu Asano, Y. Hoshino, S. Yoshikawa, M. Mat-
sukura, Y. Kabasawa, J. Nijima, Y. Kotake, N. Watanabe, T. Kawata,
T. Inoue, T. Horizoe, N.Yasuda, H. Minami, K. Nagata, M. Murakami,
J. Nagaoka, S. Kobayashi, I. Tanaka, S. Abe, J. Med. Chem. 44 (2001)
170–179.
[22] M.B. Grant, R.W. Tarnuzzer, S. Caballero, M.J. Ozeck, M.I. Davis,
P.E. Spoerri, I. Feoktistov, I. Biaggioni, J.C. Shryock, L. Berladinelli,
Circ. Res. 85 (1999) 699–706.
Acknowledgements
We gratefully thankAlmirall Prodesfarma SA (Barcelona-
Spain) for promoting and funding the present project, Dr.
Bernat Vidal (Almirall Prodesfarma) for the preparation of
compound 15b, CESCA-CEPBA for a mobility grant to AC
and Prof. Ferran Sanz (GRIB, UPF, Barcelona) for his hel-
pful comments and suggestions in the preparation of the
manuscript.
[23] C.E. Müller, Drugs Future 25 (2000) 1043–1052.
[24] J.A. Auchampach, J. Jin, T.C. Wan, G.H. Caughey, J. Linden, Mol.
Pharmacol. 52 (1998) 846–860.
[25] G.R. Strohmeier, S.M. Reppert, W.I. Lencer, J.L. Madara, J. Biol.
Chem. 270 (1995) 2387–2394.
[26] I. Feostikov, I. Baggioni, Pharmacol. Rev. 49 (1997) 381–402.
[27] K.A. Jacobson, A.P. Ijzerman, J. Linden, Drug Dev. Res. 47 (1999)
45–53.
[28] J.R. Fozard, Curr. Opin. Pharmacol. 3 (2003) 264–269.
References
[29] A. Maconi, G. Pastorin, T. Da Ros, G. Spalluto, Z.G. Gao, K.A. Jacob-
son, P.G. Baraldi, K. Varani, S. Moro, P.A. Borea, J. Med. Chem. 45
(2002) 3579–3582.
[1] V. Ralvic, G. Burnstock, Pharmacol. Rev. 50 (1998) 413–492.
[2] S. Hess, Expert Opin. Ther. Pat. 11 (2000) 1–29.
[30] E. Ongini, A. Monopoli, B. Cacciari, P.G. Baraldi, Farmaco 56 (2001)
87–90.
[3] B.B. Fredhom, G. Arslan, L. Haldner, G. Kull, G. Schulte, W. Wasser-
man, Nauyn-Schmied. Arch. Pharmacol. 362 (2000) 364–374.
[31] C.E. Müller, Farmaco 56 (2001) 77–80.
[32] A.S. Robeva, R. Woodward, X. Jin, Z. Gao, S. Bhattacharya,
H.E. Taylor, D.L. Rosin, J. Linden, Drug. Dev. Res. 39 (1996) 243–
252.
[4] B.B. Fredhom, A.P. Ijzerman, K.A. Jacobson, K.-N. Klotz, J. Linden,
Pharm. Rev. 53 (2001) 527–552.
[5] C.E. Müller, B. Stein, Curr. Pharm. Des. 2 (1996) 501–530.
[33] B. Grahner, S. Winiwarter, W. Lanzner, C.E. Müller, J. Med.Chem. 37
(1994) 1526–1534.
[6] P.G. Baraldi, P.A. Borea, Trends Pharmacol. Sci. 21 (2000) 456–459.
[7] P.G. Baraldi, B. Cacciari, R. Romagnoli, S. Meringhi, K. Verani,
P.A. Borea, G. Spalluto, Med. Res. Rev. 20 (2000) 103–128.
[34] The amino derivatives 3b and 15b were obtained through the basic
hydrolysis of the corresponding 1,3-dimethyl [4b] and 1,3-dipropyl
(see Ref. [35]) acetamido derivatives, respectively.
[8] E.M. van der Wenden, A.P. Ijzerman, W.A. Soudijn, J. Med. Chem. 35
(1992) 629–635.
[35] A. Stefanachi, F. Leonetti, A. Cappa, A. Carotti, Tetrahedron Lett. 44
(2003) 2121–2123.
[9] P.G. Baraldi, B. Cacciari, S. Moro, G. Spalluto, G. Pastorin, T. Da Ros,
K.-N. Klotz, K. Varani, S. Gessi, P.A. Borea, J. Med. Chem. 45 (2002)
770–780.
[36] Quantum mechanical MEP computation was carried by means of the
MIPSIM software [38]. The required ab initio wave functions were
obtained with the GAMESS program: M.W. Schmidt, K.K. Bald-
ridge, J.A. Boatz, S.T. Elbert, M.S. Gordon, J.H. Jensen, S. Koseki, N.
Matsunaga, K.A. Nguyen, S.J. Su, T.L. Windus, M. Dupuis, J.A.
Montgomery, J. Comput. Chem. 14 (1993) 1347–1363.
[10] W. Soudijn, I. Van Wijngaarden, A.P. Ijzerman, Curr. Top. Med.
Chem. 3 (2003) 355–367.
[11] G. Cristalli, C. Lambertucci, S. Taffi, S. Vittori, R. Volpini, Curr. Top.
Med. Chem. 3 (2003) 387–401.

A. Carotti et al. / European Journal of Medicinal Chemistry 39 (2004) 879–887
887
[37] M. Orozco, F.J. Luque, Molecular Electrostatic Potentials: Concepts
and Applications. Theoretical and Computational Chemistry, J.S.
Murray, K. Sen (Eds.), Elsevier, Amsterdam vol. 3, 1996, pp. 181–
218.
[43] H. Kubinyi, QSAR: Hansch Analysis and Related Approaches; R.
Mannhold, P. Krogsgaard-Larsen, H. Timmerman (Eds.),VCH, Wein-
heim, New York 1993.
[44] C. Hansch, A. Leo, D. Hoekman, Exploring QSAR Hydrophobic,
Electronic, and Steric Constants, vol. 2, ACS Professional Reference
Book, American Chemical Society, Washington, DC, 1995.
[38] M. de Cáceres, J. Villà, J.J. Lozano, F. Sanz, Bioinformatics 16 (2000)
568–569.
[39] ACD, version 7.0, Toronto, Canada.
[45] S. Wold, L. Eriksson, Statistical validation of QSAR results. Chemo-
metrics Methods in Molecular Design, H. Van de Waterbeemd (Ed.),
VCH, Weinheim, 1996, pp. 309–318.
[40] c-LOGP, version 4.5, Biobyte, Claremont, CA, USA.
[41] Biological data of xanthine derivatives are not shown in this paper;
they came from a collaborative study with Prof. Franco Fernandez of
the University of Santiago de Compostela (Spain) and are unpub-
lished yet.
[46] R.G. Brenton, Chemometrics, Data Analysis for Laboratory and
Chemical Plant, Wiley, England, 2003 15–117.
[47] Y.C. Cheng, W.H. Prusoff, Biochem. Pharmacol. 22 (1973) 3099–
3108.
[42] The multiple linear regression analysis has been carried out by means
of the QSAR module of the molecular modeling software package of
Sybyl (version 6.8, Tripos Ass., St. Louis, Missouri).
[48] Y. Salomon, Adv Cyclic Nucleotide Res. 10 (1979) 35–55.