Synthesis of 1-(2-chloro-2-phenylethyl)-6-methylthio-1H-pyrazolo[3,4-d] pyrimidines 4-amino substituted and their biological evaluation

Article abstract of DOI:10.1016/j.ejmech.2003.11.007

A new series of 4-amino-6-methylthio-1H-pyrazolo[3,4-d]pyrimidines (2a-m) bearing the 2-chloro-2-phenylethyl chain at the N1 position, has been synthesized. The affinity of these compounds for A₁ adenosine receptor (A₁AR) was measure

Full text of DOI:10.1016/j.ejmech.2003.11.007

European Journal of Medicinal Chemistry 39 (2004) 153–160
www.elsevier.com/locate/ejmech
Original article
Synthesis of
1-(2-chloro-2-phenylethyl)-6-methylthio-1H-pyrazolo[3,4-d]pyrimidines
4-amino substituted and their biological evaluation
Silvia Schenone a,*, Olga Bruno a, Francesco Bondavalli a, Angelo Ranise a, Luisa Mosti a,
Giulia Menozzi a, Paola Fossa a, Fabrizio Manetti b, Lucia Morbidelli c, Letizia Trincavelli d,
Claudia Martini d, Antonio Lucacchini d
a Dipartimento di Scienze Farmaceutiche, Università degli Studi di Genova, viale Benedetto XV, 16132 Genoa, Italy
b Dipartimento Farmaco Chimico Tecnologico, Università degli Studi di Siena, via Aldo Moro, 53100 Siena, Italy
c Dipartimento di Biologia Molecolare, Università degli Studi di Siena, via Aldo Moro, 53100 Siena, Italy
d Dipartimento di Psichiatria, Neurobiologia, Farmacologia e Biotecnologie, Università degli Studi di Pisa, via Bonanno, 56126 Pisa, Italy
Received 12 June 2003; received in revised form 19 November 2003; accepted 19 November 2003
Abstract
A new series of 4-amino-6-methylthio-1H-pyrazolo[3,4-d]pyrimidines (2a–m) bearing the 2-chloro-2-phenylethyl chain at the N1
position, has been synthesized. The affinity of these compounds for A1 adenosine receptor (A1AR) was measured. The compounds showed
poor affinity. A more interesting result was obtained by 2a, 2d, 2g, which demonstrated inhibitory activity on cell proliferation of the A-431
cell line stimulated by epithelial growth factor (EGF) and on EGF receptor tyrosine kinase (EGFR-TK) phosphorylation.
© 2003 Elsevier SAS. All rights reserved.
Keywords: Pyrazolo-pyrimidines; Adenosine; Epidermal growth factor
1. Introduction
rivatives were found to be selective ligands with antagonist
activity for A1 adenosine receptors (A1AR). They may have
therapeutical use as cognitive enhancers, anti-dementia
drugs (e.g. for Alzheimer’s disease and cerebrovascular de-
mentia), psycostimulants, anti-depressant drugs and amelio-
rants of cerebral function [14]. A1AR antagonists have also
demonstrated promising therapeutic potential for renal and
cardiac failure [15,16].
Adenosine is an endogenous neuromodulator, which me-
diates its biological effects by interacting with four receptor
subtypes named A1, A2A, A2B, and A3, identified and cloned
from various mammalian tissue types [1–3].
Despite the large number of potential therapeutic applica-
tions, relatively few compounds have entered into clinical
trials and are used in the clinic [4,5]. This fact is also due to
the poor selectivity of inhibitors for unique adenosine recep-
tor subtype. Consequently, there has been an expanded inter-
est in finding selective molecules that lack undesirable side
effects.
We recently synthesized a series of 4-amino-1-(2-chloro-
2-phenylethyl)-1H-pyrazolo[3,4-b]pyridine-5-carboxylic
acid ethyl ester derivatives 1 [17] (see Fig. 1). The affinity
data showed that these new molecules bind the A1AR in a
completely selective way. Some of the compounds show
interesting affinity and antagonistic activity.
In recent years, a wide variety of nitrogen-containing
heterocycles have been synthesized as possible A1 selective
antagonists [6,7]. In particular, among these, pyrazolo[3,4-
b]pyridine [8,9] and pyrazolo[3,4-d]pyrimidine [10–13] de-
In order to find a more active and A1 selective antagonist,
we recently developed a pseudoreceptor model [18] (con-
structed taking into consideration structure–activity data of
molecules derived both from our own research and also
collected from current literature) capable of predicting theA1
binding activity of compounds not yet synthesized.
* Corresponding author.
E-mail address: schensil@unige.it (S. Schenone).
© 2003 Elsevier SAS. All rights reserved.
doi:10.1016/j.ejmech.2003.11.007

154
S. Schenone et al. / European Journal of Medicinal Chemistry 39 (2004) 153–160
with acetic acid (80% yield). Alkylation of 5 with methyl
iodide in THF at reflux afforded the 6-thiomethyl derivative
6. The latter was in turn treated with the Vilsmeier complex
(POCl3:DMF) to obtain the dihalogenated compound 8 bear-
ing the chlorine atom either at position 4 of the pyrimidine
nucleus and on the N1 side chain. Compound 8 was purified
in good yield by chromatography on a Florisil® column.
It is interesting to point out that treatment of 6 with POCl3
at reflux afforded the undesired product 7, resulting from
dehydrohalogenation of the N1 side chain.
Fig. 1
Finally, regioselective substitution of the C4 chlorine
atom with an excess of various amines afforded the desired
compounds 2a–m in yields ranging between 60% and 80%.
Notably, the chlorine atom on the side chain was not substi-
tuted by the amine in spite of its benzylic position, as shown
by the 1H NMR chemical shifts of the CH2–CH side chain,
which give an ABX complex pattern owing to the non-
equivalence of CH2-protons.
Based on this theoretical model and taking into account
that the SAR data evidenced both the 2-chloro-2-phenylethyl
group in position 1 of the pyrazole ring and the substituted
nitrogen at the 4 position of the pyridine nucleus as important
features for activity of compounds 1 towardA1AR, we envis-
aged the family of 1-aryl-4-amino substituted-6-methylthio-
1H-pyrazolo[3,4-d]pyrimidine derivatives 2, as an attractive
field of investigation. These new compounds share a hetero-
cyclic structure more strictly isosteric to the purinic moiety
of adenosine and appear as interesting compounds to be
synthesized and tested as A1AR inhibitors. In fact, binding
activity of the N4 propyl, butyl, and cyclopropyl derivatives
were predicted by the pseudoreceptor model to be 13, 2, and
160 nM, respectively. Moreover, it is worth pointing out that
6-thio-substituted pyrazolo[3,4-d]pyrimidines that show an
excellent activity against A1AR, are already present in cur-
rent literature [19,20].
3. Pharmacology
All the compounds were tested for their ability to displace
[3H]N6-cyclohexyladenosine ([3H]CHA) from A1AR and
[3H]-2-{[4-(2-carboxyethyl)phenylethyl]amino}-5-(N-
ethylcarbamoyl) adenosine ([3H]CGS21680) fromA2AAR in
bovine cortical and striatal membranes, respectively [16].
The A1 and A2A receptor binding affinities for compounds
2a–m (see Table 1), are expressed as Ki or % of inhibition of
binding (at 10 µM compound concentration).
2. Chemistry
The synthetic route to the target compounds is outlined in
the Scheme 1.
4. Result and discussion
The starting material is ethyl ester of 5-amino-1-(2-
hydroxy-2-phenylethyl)-1H-pyrazole-4-carboxylic acid (3),
obtained following our reported procedure [18]. Reaction of
3 with benzoyl isothiocyanate in THF at reflux for 8 h yielded
the intermediate 4. Cyclization to the pyrazolo[3,4-
d]pyrimidine intermediate 5 was achieved by treating 4 with
1 M NaOH at 100 °C for 10 min, followed by acidification
This new family of 6-methylthio-pyrazolo[3,4-
d]pyrimidines is evidently less active than the pyrazolo[3,4-
d]pyridines synthesized by us and bearing the same 2-chloro-
2-phenylethyl side chain on the N1 of the pyrazole ring. In
fact, the most potent compounds in this series were 2a and 2b
having the cyclopropylamino and n-propylamino group at
the C4, showing A1AR affinity of 0.5 and 0.9 µM, respec-
Scheme 1

S. Schenone et al. / European Journal of Medicinal Chemistry 39 (2004) 153–160
155
Table 1
proach is simplified, since the functional groups of inhibitors
(i.e. erlotinib) interacting with the enzyme and the amino
acid residues of the binding site are well known from X-ray.
Thus, the docking problem is essentially reduced to locate
the possible conformations of the new pyrazolo-pyrimidino
derivatives within the active site of the enzyme.
Binding activity of compounds 2a–m at bovine A1 and A2A adenosine
receptors
Compounds
R
A1AR affinity
A2AAR affinity
(Ki or % inhibition (Ki or %
at 10 µM) a
inhibition at
10 µM) a
60%
51%
34%
25%
20%
37%
19%
6.7%
0%
All the complexes generated by docking simulations were
then submitted to statistical conformational search with
Monte Carlo Multiple Minimum (MCMM) method involv-
ing all the rotatable bonds of the ligand and following a
computational protocol reported in detail in Section 6.
In a first attempt, the crystallographic coordinates of the
4-aminopyrimidine nucleus of erlotinib were used to drive
the superposition of the corresponding ring of 2a into the
experimental EGFR-TK binding site. Such an orientation
allowed us to find a hydrogen bond contact between the N7
atom of 2a and the amide NH group of Met769 (Fig. 2), as
previously described in the EGFR–erlotinib complex [23].
Upon minimization of this complex, a conformational rear-
rangement was found to force both Thr766 and Gln767 away
from their original position and to eliminate steric clashes
involving these amino acids and the thiomethyl group of the
inhibitor. Such a structural reorganization led to complexes
lacking all the hydrogen bonds involving the pyrimidine ring
of 2a, that on the basis of literature reports were considered
as non-interesting complexes and consequently not further
investigated.
2a
2b
2c
2d
2e
2f
2g
2h
2i
NH cyclopropyl
NHC3H7
0.52
0.90
2.08
15
NHC4H9
NHC(CH3)3
N(C2H5)2
11%
37%
5.84
46%
34%
45%
NH(CH2)2OC2H5
NH cyclohexyl
1-Pyrrolidinyl
1-Piperidinyl
4-Morpholinyl
2j
2k
2l
19%
0%
1-Hexahydroazepinyl 28%
NHCH2C6H5
18%
9.44
20%
2.6%
2m
NHCH2CH2C6H5
a The Ki values are mean S.E.M. of three separate assays, each perfor-
med in triplicate.
tively. Increasing the length or bulkiness of the side chain led
to a significant decrease in affinity, the n-butyl (2c) and
tert-butyl (2d) derivatives showing affinity values of 2 and
15 µM, respectively. The remaining compounds of the series
were inactive, showing that a short and not too bulky side
chain (n-propyl and cyclopropyl) on the C4 nitrogen atom is
necessary for these ligands to bind A1AR.
Based on these findings, several conclusions can be put
forward. (i) The pharmacophore model developed by us to
predict A1 activity has to be revised according to the new
results. Work in this direction is already ongoing. (ii) The
presence of the ester moiety at C5 of 1 is necessary for the
A1AR activity, according to the suggestion that this group
could give a profitable internal hydrogen bond with the NH at
C4 [18]. (iii) Compounds lacking the C6 substituent show
even lower A1 affinity, suggesting the preparation of new
derivatives with a longer chain at C6 [21].
On the other hand, a different orientation of 2a (Fig. 3),
corresponding to that proposed by Traxler for 1-aryl-
pyrazolopyrimidines gave more interesting results. In fact,
Recently, Traxler et al. [22], on the basis of a pharmacoph-
ore model for inhibitors competing for the ATP-binding site
of the epidermal growth factor receptor (EGFR), reported the
design and synthesis of a potent and selective EGFR tyrosine
kinase (EGFR-TK) inhibitor belonging to the
4-phenylaminopyrazolo[3,4-d]pyrimidine family. Moreover,
the crystal structure of the kinase domain from EGFR has
been determined at a 2.6 Å resolution, in complex with a
4-anilinoquinazoline inhibitor, erlotinib, currently in Phase
III clinical trials as an anti-cancer agent [23]. Based on both
these literature reports, we have performed molecular dock-
ing simulations using the crystallographic coordinates of
EGFR as the input macromolecular target and the position of
erlotinib into the EGFR-TK binding site to orient our mol-
ecules, with the purpose of ascertaining whether 2a–m could
interact profitably with EGFR.
Fig. 2. Complex between compound 2a and the EGFR-TK binding site, as
obtained by superposition of the pyrazolo-pyrimidine derivative onto the
crystallographic structure of the 4-aminopyridine nucleus of erlotinib. The
N7 endocyclic nitrogen atom is at a hydrogen bond distance from the
backbone amide group of Met769. For sake of clarity, only few amino acids
of the macromolecule were displayed.
It is important to note that in the present study, the com-
putational step corresponding to the molecular docking ap-

156
S. Schenone et al. / European Journal of Medicinal Chemistry 39 (2004) 153–160
the EGFR, AG1478, showed similar inhibitory activity of
cell proliferation (33%). Lower concentration of 2a (10 nM)
produced marginal inhibition of cell proliferation (10%),
while a 1000-fold higher concentration (10 µM) provided the
maximal effect. A similar trend was found for 2d and 2g
when tested at the same concentrations. While a 10 µM
concentration led to the maximal effect for both compounds,
a 50 nM concentration produced a cell proliferation inhibi-
tion of 28% and 19%, respectively. Any significant anti-
proliferative activity was found at 10 nM concentrations.
In cytotoxicity assay, a nanomolar concentration of 2a
was not effective, while a higher concentration (10 µM)
produced significant cell death (36%). Regarding 2d and 2g,
the latter concentration showed a 34% and 41% cell death.
Finally, AG1478 and the three pyrazolo-pyrimidines were
compared for their ability to prevent phosphorylation of
EGFR in cell stimulated with EGF at 10 ng ml–1. At 1 µM,
whileAG1478 produced a 40% inhibition of EGFR phospho-
rylation, 2a, 2d, and 2g inhibited it by 25%, 18%, and 19%,
respectively.
Compound 2a was also investigated for its effects in the in
vitro human EGF tyrosine kinase assay. As a result, an
interesting 48% inhibition of control values obtained in the
presence of 2a was found, demonstrating that such a com-
pound is an inhibitor of EGFR at its kinase site. The enzyme
assay on EGF-TK was performed on A431 cells using
PD153035 as the reference compound, according to the pro-
cedure reported by Carpenter et al. [29].
Fig. 3. Complex between compound 2a and the EGFR-TK binding site,
with the pyrazolo-pyrimidine derivative oriented in agreement with what
proposed by Traxler. A bidentate hydrogen bond acceptor/donor system
anchoring 2a into the ATP binding site, was found, involving Met769 and
Gln767 of the macromolecule and the N5 endocyclic nitrogen and the
exocyclic amino nitrogen, respectively.
the N5 pyrimidine nitrogen atom together with the secondary
amino group at the 4-position constitute the bidentate hydro-
gen bond acceptor–donor system anchoring 2a within the
ATP binding site of the EGFR. The corresponding receptor
counterparts are represented by the NH amide backbone of
Met769, and the carbonyl group of Gln767, respectively.
According to this binding mode, the 2-chloro-2-phenylethyl
side chain lies into a large pocket mainly defined by Phe699,
Val702, Lys721, andAsp831. Moreover, the cyclopropyl ring
is accommodated within a region constituted by three sulfo-
rated residues (namely, Met742, Cys751, and Met769) and
two threonine amino acids (namely, 766 and 830), in addition
to Leu820. Finally, the methylthio substituent at the
6-position is located into a region of the binding site (delim-
ited by Leu694 and Gly772) where no profitable interaction
with the macromolecule was found, similarly to some
6-amino derivatives reported by Traxler et al. [22].
5. Conclusion
In conclusion, although compounds 2 did not show rel-
evant A1AR ligand activity, they may represent reasonable
input structure to design and synthesize compounds with
potential inhibitory activity toward tyrosine kinase receptors.
Further efforts are ongoing in our labs to test the remaining
compounds of the series 2.
Moreover, it is important to point out that, to the best of
our knowledge, 2a is the first example of 6-substituted
pyrazolo[3,4-d]pyrimidine showing anti-proliferative activ-
ity, probably acting on tyrosine kinase catalytic domain with
possible anti-tumor applications.
The biological activity of 2a, 2d and 2g was assessed on
A-431 cell line expressing EGFR. Proliferation and cytotox-
icity 1 of A-431 cells and the autophosphorylation of EGFR
in response to the three pyrazolo-pyrimidines, were evalu-
ated. Compound 2a tested at 50 nM in the presence of EGF
inhibited cell proliferation by 40%. The known inhibitor of
6. Experimental protocols
6.1. Chemistry
Starting materials were purchased from Aldrich-Italia
(Milan).
Melting points (m.p.) were determined with a Büchi B
540 apparatus and are uncorrected. IR spectra were recorded
in KBr with a Perkin–Elmer 398 spectrophotometer. 1H
NMR spectra were recorded in CDCl3 solution on a Varian
Gemini 200 (200 MHz) instrument, chemical shifts (d) are
1 Cell proliferation was evaluated, as total cell number counted micros-
copically, after 24 h incubation with test compound in the presence of the
specific growth factor. Cell cytotoxicity was evaluated by trypan blue exclu-
sion following 1 h incubation with test compound. EC50 was calculated as
the concentration of the compound able to give 50% of dead cells.

S. Schenone et al. / European Journal of Medicinal Chemistry 39 (2004) 153–160
157
reported in ppm relative to TMS as internal standard; J in Hz.
1H patterns are described using the following abbreviations:
s, singlet; t, triplet; q, quartet; sext, sextet; m, multiplet; br,
broad.
All compounds were tested for purity by TLC (Merk,
Silica gel 60 F254, CHCl3 as eluant).
The excess of POCl3 was removed by distillation under
reduced pressure. H2O (20 ml) was carefully added to the
residue and the suspension was extracted with CHCl3 (3 ×
20 ml). The organic solution was washed with H2O (10 ml),
dried (MgSO4), filtered and concentrated under reduced
pressure. The crude brown oil was purified by column chro-
matography (Florisil® 100–200 Mesh), using CHCl3 as elu-
ant to afford the pure product 7 (0.9 g, 30%)) as a white solid,
m.p. 108–109 °C. 1H NMR: d 2.70 (s, 3H, CH3), 7.25–7.58
(m, 6H, 5H,Ar + 1H, CH=), 7.96 (d, J = 14.0, 1H, CH=), 8.14
(s, 1H, H-3). IR (cm–1): 1658 (C=C).
Analyses for C, H, N were within 0.3% of the theoretical
value.
6.1.1. Ethyl 5-{[(benzoylamino)carbonothioyl]amino}-1-
(2-hydroxy-2-phenylethyl)-1H-pyrazole-4-carboxylate (4)
A suspension of 3 (2.7 g, 10 mmol) and benzoyl isothio-
cyanate (1.7 g, 11 mmol) in anhydrous tetrahydrofuran
(20 ml) was refluxed for 12 h. The solvent was evaporated
under reduced pressure; the oil residue crystallized adding
diethyl ether (30 ml) to afford the pure product 4 (4.07 g,
93%) as a white solid, m.p. 171–172 °C. 1H NMR: d 1.29 (t,
J = 7.0, 3H, CH3), 3.97–4.20 (m, 5H, 2CH2 + OH, 1H
disappears with D2O), 4.58–4.68 (m, 1H, CHOH), 7.05–7.98
(m, 10H, Ar), 8.02 (s, 1H, H-3), 8.70 (s, 1H, NH, disappears
with D2O), 12.05 (s, 1H, NH, disappears with D2O). IR
(cm–1): 3221 (NH), 3190–2940 (OH), 1708 and 1671 (2CO).
6.1.5. 4-Chloro-1-(2-chloro-2-phenylethyl)-6-(methylthio)-
1H-pyrazolo[3,4-d]pyrimidine (8)
The Vilsmeier complex, previously prepared from POCl3
(6.13 g, 40 mmol) and anhydrous DMF (2.92 g, 40 mmol)
was added to a suspension of 6 (3.02 g, 10 mmol) in CHCl3
(20 ml).
The mixture was refluxed for 4 h. The solution was
washed with H2O (2 × 20 ml), dried (MgSO4), filtered and
concentrated under reduced pressure. The crude oil was puri-
fied by column chromatography (Silica gel, 100 Mesh), using
a mixture of diethyl ether/petroleum ether (b.p. 40–60 °C)
1/1 as eluant, to afford the pure product 8 (2.2 g, 65%) as a
6.1.2. 1-(2-Hydroxy-2-phenylethyl)-6-thioxo-1,5,6,7-
tetrahydro-4H-pyrazolo[3,4-d]pyrimidin-4-one (5)
1
white solid, m.p. 95–96 °C. H NMR: d 2.62 (s, 3H, CH3),
A solution of 4 (4.38 g, 10 mmol) in 2 M NaOH (40 ml)
was refluxed for 10 min and successively diluted with H2O
(40 ml). The solution was acidified with glacial acetic acid.
Standing in a refrigerator for 12 h a solid crystallized, and
was filtered and recrystallized from absolute ethanol to give 5
4.77–5.05 (m, 2H, CH2N), 5.45–5.56 (m, 1H, CHCl), 7.29–
7.46 (m, 5H, Ar), 8.02 (s, 1H, H-3).
6.1.6. General procedure for
1-(2-chloro-2-phenylethyl)-6-methylthio-1H-
1
(2.3 g, 80%) as a white solid, m.p. 264–265 °C. H NMR
pyrazolo[3,4-d]pyrimidines 4-amino substituted (2a–m)
To a solution of 8 (3.4 g, 10 mmol) in anhydrous toluene
(20 ml), the proper amine (40 mmol) was added, and the
reaction mixture was stirred at room temperature for 24 h.
After it was extracted with H2O, the organic phase was dried
under reduced pressure; the oil residue crystallized by adding
a mixture of diethyl ether/petroleum ether (b.p. 40–60 °C)
1/1 (5 ml), to give products 2a–m.
Compound 2a: White solid, m.p. 132–133 °C, yield 75%.
1H NMR: d 0.69–0.81 (m, 2H, CH2), 0.92–1.08 (m, 2H,
CH2), 2.57 (s, 3H, CH3S), 2.86–2.99 (m, 1H, CH), 4.72–4.98
(m, 2H, CH2N), 5.51–5.62 (m, 1H, CH–Cl), 5.81 (br s, 1H,
NH, disappears with D2O), 7.22–7.58 (m, 5H, Ar), 8.08 (s,
1H, H-3). IR (cm–1): 3410 (NH).
Compound 2b: White solid, m.p. 126–127 °C, yield 76%.
1H NMR: d 1.02 (t, J = 5.0, 3H, CH3), 1.73 (sext, J = 5.0, 2H,
CH2), 2.58 (s, 3H, CH3S), 3.55 (q, J = 5.0, 2H, CH2NH),
4.70–4.97 (m, 2H, CH2N), 5.20 (br s, 1H, NH, disappears
with D2O), 5.49–5.58 (m, 1H, CHCl), 7.25–7.47 (m, 5H,Ar),
7.78 (s, 1H, H-3). IR (cm–1): 3414 (NH).
(d6-DMSO): d 4.15–4.30 and 4.55–4.72 (2m, 2H, CH2N),
4.85–5.00 (m, 1H, CHO), 5.66 (br s, 1H, OH, disappears with
D2O), 7.20–7.51 (m, 5H, Ar), 8.02 (s, 1H, H-3), 12.20 (s, 1H,
NH, disappears with D2O), 13.40 (s, 1H, NH, disappears
with D2O). IR (cm–1): 3362 (NH), 3142–2773 (OH), 1681
(CO).
6.1.3. 1-(2-Hydroxy-2-phenylethyl)-6-(methylthio)-1,5-
dihydro-4H-pyrazolo[3,4-d]pyrimidin-4-one (6)
A solution of 5 (2.88 g, 10 mmol) and methyl iodide
(7.10 g, 50 mmol) in anhydrous tetrahydrofuran (20 ml) was
refluxed for 12 h. The solvent and the excess of methyl iodide
were removed by distillation under reduced pressure; the oil
residue crystallized by adding chloroform (10 ml) and was
purified by recrystallization with absolute ethanol to give 6
(2.17 g, 72%) as a white solid, m.p. 208–209 °C. 1H NMR: d
2.51 (s, 3H, CH3), 4.27–4.50 (m, 2H, CH2N), 5.04–5.18 (m,
1H, CHO), 5.68 (d, 1H, OH, disappears with D2O), 7.20–
7.42 (m, 5H, Ar), 7.97 (s, 1H, H-3), NH not detectable. IR
(cm–1): 3544 (NH), 3450–3350 (OH), 1678 (CO).
Compound 2c: White solid, m.p. 106–107 °C, yield 45%.
1H NMR: d 0.97 (t, J = 7.0, 3H, CH3), 1.33–1.84 (m, 4H,
2CH2), 2.57 (s, 3H, CH3S), 3.59 (q, J = 7.0, 2H, CH2NH),
4.70–4.95 (m, 2H, CH2N), 5.49–5.60 (m, 1H, CH–Cl), 7.23–
7.49 (m, 5H Ar), 7.78 (s, 1H, H-3), NH not detectable. IR
(cm–1): 3413 (NH).
6.1.4. 4-Chloro-6-(methylthio)-1-(2-phenylvinyl)-1H-
pyrazolo[3,4-d]pyrimidine (7)
POCl3 (14 g, 91 mmol) was added to 6 (3.02 g, 10 mmol),
the mixture was refluxed for 12 h and then cooled to room
temperature.

158
S. Schenone et al. / European Journal of Medicinal Chemistry 39 (2004) 153–160
Compound 2d: White solid, m.p. 80–81 °C, yield 50%. 1H
300 server using the software package MacroModel/
BatchMin [17] equipped with the AMBER* united-atom
force field.
The enzyme models used in this study were derived from
the coordinates of the structure labeled 1m17 in the
Brookhaven Protein Data Bank, which represent the com-
plexes between the kinase domain of EGFR and a
4-anilinoquinazoline inhibitor (erlotinib), refined to 2.6 Å
resolution.
To create initial coordinates for the docking studies, all the
water molecules of the crystal structure were removed and
excluded from the calculations. Similarly the 40 amino acids
from the carboxy terminal tail of the EGFR structure were
deleted. Hydrogen atoms were added in their idealized posi-
tions in such a way as to protonate all Lys andArg side chains
and deprotonate all Glu and Asp side chains.
In the present study, the computational step corresponding
to the molecular docking approach is simplified, since the
functional groups of the substrates interacting with the en-
zyme and the amino acid residues of the catalytic machinery
are well known in advance. Thus, the docking problem is
essentially reduced to locate the possible conformations of
the ligand within the active site of the enzyme. The pyrimi-
dine ring in the X-ray crystal structures was used to guide the
building of the EGFR–2a complex providing a good tem-
plate to adjust torsional angles of the inhibitor. In particular,
the 4-aminopyrimidine ring of 2a was superposed onto the
corresponding ring of erlotinib.
In a second attempt, following the Traxler’s model, 2a was
located into the binding site in such a way as to contact (with
its endocyclic nitrogen atom at the 5-position and the second-
ary amino group at the 4-position) by hydrogen bonds the
amide NH group of Gln767 and the carbonyl group of
Met769, respectively.
The complexes generated were then submitted to a statis-
tical conformational search with MCMM method involving
the rotatable bonds of the inhibitor. MCMM causes an input
structure to be modified by random changes in its torsion
angles as given by the TORS command. Values of 0° and
180° are, respectively, the minimum and maximum angular
increments to be added or subtracted from the current dihe-
dral angles. It is important to note that conformational mobil-
ity of the active site residues was not a complicating factor
since it has been shown that these residues undergo only
minor structural changes upon the binding of different inhibi-
tors [23].
The least square superimposition routine was selected
(COMP command) to compare each new complex with all
conformers previously found in order to eliminate duplicate
minima.
Because of the large number of atoms in the model, to
correctly optimize the EGFR–2a complexes, the following
constraints had to be imposed: (a) a subset, centered on the
inhibitor and comprising only the small molecules and a shell
of residues surrounding the binding site of the enzyme within
a radius of 10 Å from the ligand, was created and subjected to
NMR: d 1.57 (s, 9H, t-but), 2.61 (s, 3H, CH3S), 4.68–4.98
(m, 2H, CH2N), 5.08 (br s, 1H, NH disappears with D2O),
5.51–5.62 (m, 1H, CHCl), 7.28–7.50 (m, 5HAr), 7.73 (s, 1H,
H-3).
Compound 2e: White solid, m.p. 91–92 °C, yield 80%. 1H
NMR: d 1.29 (t, J = 9.0, 6H, 2CH3), 2.57 (s, 3H, CH3S), 3.71
(q, J = 9.0, 4H, 2CH2), 4.68–4.98 (m, 2H, CH2N), 5.53–5.63
(m, 1H, CHCl), 7.26–7.50 (m, 5H, Ar), 7.73 (s, 1H, H-3).
Compound 2f: White solid, m.p. 115–116 °C, yield 68%.
1H NMR: d 1.21 (t, J = 7.0, 3H, CH3), 2.56 (s, 3H, CH3S),
3.53 (q, J = 7.0, 2H, OCH2CH3), 3.55–3.85 (m, 4H, 2CH2),
4.66–4.97 (m, 2H, CH2N), 5.45–5.60 (m, 1H, CHCl), 5.72
(br s, 1H, NH, disappears with D2O), 7.27–7.44 (m, 5H, Ar),
7.77 (s, 1H, H-3). IR (cm–1): 3441 (NH).
Compound 2g: White solid, m.p. 138–139 °C, yield 72%.
1H NMR: d 1.14–1.57, 1.70–1.89 and 2.01–2.18 (m, 11H
cyclohexyl), 2.57 (s, 3H, CH3S), 3.98 (br s, 1H, NH, disap-
pears with D2O), 4.69–4.98 (m, 2H, CH2N), 5.50–5.61 (m,
1H, CHCl), 7.23–7.49 (m, 5H, Ar), 7.73 (s, 1H, H-3). IR
(cm–1): 3402 (NH).
Compound 2h: White solid, m.p. 151–152 °C, yield 52%.
1H NMR: d 1.90–2.06 (m, 2H, CH2), 2.07–2.23 (m, 2H,
CH2), 2.58 (s, 3H, CH3S), 3.69–3.86 (m, 4H, 2 CH2Npyrr),
4.70–4.97 (m, 2H, CH2N), 5.54–5.62 (m, 1H, CHCl), 7.24–
7.57 (m, 5H, Ar), 7.79 (s, 1H, H-3).
Compound 2I: White solid, m.p. 94–95 °C, yield 60%. 1H
NMR: d 1.60–1.82 (m, 6H, 3CH2pip), 2.57 (s, 3H, CH3S),
3.82–3.98 (m, 4H, 2CH2Npip), 4.70–4.97 (m, 2H, CH2N),
5.53–5.64 (m, 1H, CHCl), 7.28–7.49 (m, 5H, Ar), 7.81 (s,
1H, H-3).
Compound 2j: White solid, m.p. 116–117 °C, yield 75%.
1H NMR: d 2.57 (s, 3H, CH3S), 3.79–3.98 (m, 8H,
4CH2morph), 4.71–4.98 (m, 2H, CH2N), 5.51–5.61 (m, 1H,
CHCl), 7.26–7.49 (m, 5H, Ar), 7.82 (s, 1H, H-3).
Compound 2k: White solid, m.p. 99–100 °C, yield 60%.
1H NMR: d 1.46–1.69 and 1.75–1.98 (2m, 8H, 4CH2), 2.57
(s, 3H, CH3S), 3.73–3.99 (m, 4H, CH2NCH2), 4.69–4.96 (m,
2H, CH2N), 5.52–5.63 (m, 1H, CHCl), 7.24–7.50 (m, 5H,
Ar), 7.76 (s, 1H, H-3).
Compound 2l: White solid, m.p. 142–143 °C, yield 81%.
1H NMR: d 2.58 (s, 3H, CH3S), 4.70–4.94 (m, 4H, CH2N +
CH2NH), 5.50–5.60 (m, 1H, CH–Cl), 7.24–7.96 (m, 10H,
Ar), 7.72 (s, 1H, H-3), NH not detectable. IR (cm–1): 3442
(NH).
Compound 2m: White solid, m.p. 73–74 °C, yield 76%.
1H NMR: d 2.59 (s, 3H, CH3S), 2.98 (q, J = 6.0, 2H,
CH2C6H5), 3.87 (q, J = 6.0, 2H, CH2NH), 4.70–4.95 (m, 2H,
CH2N), 5.30 (br s, 1H, NH, disappears with D2O), 5.50–5.60
(m, 1H, CHCl), 7.19–7.48 (m, 10H Ar), 7.73 (s, 1H, H-3). IR
(cm–1): 3445 (NH).
6.2. Computational details
All calculations and graphical manipulations were per-
formed on a Silicon Graphics Octane workstation and Origin

S. Schenone et al. / European Journal of Medicinal Chemistry 39 (2004) 153–160
159
energy minimization. The inhibitor and all amino acid side
6.3.2. Inhibitory activity on cell proliferation
chains of the shell were unconstrained during energy mini-
mization to allow for reorientation and thus proper hydrogen-
bonding geometries and vdW contacts; (b) all the atoms
external to the subset remained fixed, even if their non-bond
interactions with all the relaxing atoms have been calculated.
Two intramolecular distances were monitored during the
docking experiments in order to ensure that the relative
orientations of the residues of the binding site remained
compatible with the postulated interaction pattern. These
were the distances between the amide hydrogen atom
Met769 and the N5 atom of the pyrimidine ring of 2a, and
between the carbonyl oxygen atom of Gln767 and the hydro-
gen atom of the secondary amino group at the 4-position of
2a.
Energy minimization of the complexes was performed
using the Polak–Ribière conjugate gradient method until the
derivative convergence was 0.01 kcal Å–1 mol–1. The default
non-bonded cutoff protocol employed by the program was
modified. Thus, a vdW cutoff of 12.0 Å, an electrostatic
cutoff of 20 Å, and a hydrogen bonding cutoff of 2.5 Å were
utilized for all calculations.
6.3.2.1. Cell lines and culture conditions. Cell culture
reagents and materials were from Sigma Chemical Co. FCS
was from Hyclone, Logan, UT, USA. Human recombinant
EGF was from Calbiochem-Novabiochem Int., San Diego,
CA, USA. Diff-Quik was from Mertz-Dade AG, Dudingen,
Switzerland.
The human epidermoid carcinoma A-431 cells were ob-
tained from American Type Culture Collection (Rockville,
MD). A-431 cells were maintained in culture in DMEM
supplemented with 4500 g l–1 glucose and 10% FCS. Cells
were split 1:5 twice a week.
6.3.2.2. Cytotoxicity. The cytotoxic effect of compounds
was evaluated by trypan blue exclusion. Cells were stimu-
lated for 1 h at 37 °C with the increasing concentrations
(10 nM–10 µM) of the test compound in 1% FCS medium.
Cells were then stained with 0.4% trypan blue in phosphate
buffer saline (PBS) for 5 min. The number of dead and living
cells was counted at the microscope in a blind manner. The
percentage of dead cells over the total number of cells was
calculated.
6.3. Biological
6.3.1. A1 and A2A adenosine receptor binding assays
[3H]CHA, [3H]CGS 21680 were obtained from DuPont-
NEN (Boston, MA). DPCPX was purchased from RBI (Na-
tik, MA). Adenosine deaminase was from Sigma Chemical
Co. (St. Louis, MO). All other reagents were from standard
commercial sources and of the highest commercially avail-
able grade.
6.3.2.3. Proliferation studies. Cell proliferation was quanti-
fied by total cell number as reported [27]. The compounds
were tested at concentration ranging from 10 nM to 10 µM in
the absence and in the presence of the growth factors EGF
0.1 pM, for 24 h.
6.3.2.4. Autophosphorylation assay. A-431 cells (3 × 106)
were seeded in a 10 cm plate with 10% serum at 37 °C
overnight and starved for 24 h, after which they were exposed
to 1 µM of 2a, 2d, 2g or AG1478 for 2 h and subsequently
treated with 10 ng ml–1 EGF for 10 min at 37 °C. Proteins
were extracted and 80 µg of lysate for each experimental
point were subjected to 10% SDS-polyacrilamide gel elec-
trophoresis and transferred to a PDVF membrane. Western
blot analysis was performed as described [28]. The anti-
phosphotyrosine antibody 1:2000 (Upstate Biotechnology)
or anti-EGFR (Upstate Biotechnology) for the determination
of corresponding receptor levels were incubated with the
membranes for 18 h. Blots were incubated with HRP-goat
anti-mouse antibody (Upstate Biotechnology) and the bands
visualized with an enhanced chemiluminescence system
(Amersham Pharmacia Biotech). Band intensities were mea-
sured using the ScioImage software package.
Displacement of [3H]CHA (31 Ci mmol–1) from A1AR in
bovine cortical membranes and of [3H]CGS 21680 (42.1 Ci
mmol–1) from A2AAR in bovine striatal membranes was
performed as described [24] Adenosine A1 receptor affinities
with [3H]DPCPX as radioligand were determined according
to Pirovano et al. [25]. Measurements with [3H]DPCPX were
performed in the presence and in the absence of 1 mM GTP.
Incubations were carried out in duplicate for 90 min at
25 °C. Non-specific binding was determined in the presence
of 50 M R-PIA and constituted approximately 30% of the
total binding. Binding reaction was terminated by filtration
through a Whatman GF/C filter, washing three times with
5 ml of ice-cold buffer.
All compounds were routinely dissolved in DMSO and
diluted with assay buffer to the final concentration, where the
amount of DMSO never exceeded 2%.
At least six different concentrations spanning three orders
of magnitude, adjusted appropriately for the IC50 of each
compound, were used. IC50 values, computer-generated us-
ing a non-linear regression formula on a computer program
(GraphPad, San Diego, CA), were converted to Ki values,
knowing the Kd values of radioligands in the different tissues
and using the Cheng and Prusoff equation [26]. The dissocia-
tion constant (Kd) of [3H]CHA and [3H]CGS 21680 were
1.2 and 14, respectively.
6.3.2.5. Human EGF tyrosine kinase assay. The experimen-
tal conditions are based on scintillation counting of the reac-
tion product ([c-33P]poly GAT) obtaine d by combination of
substrate/tracer (poly GAT, 0.48 mg ml–1, and [c-33P]ATP)
during a 60 min incubation at 30 °C. Results are expressed as
percent inhibition of control values obtained in the presence
of 2a.

160
S. Schenone et al. / European Journal of Medicinal Chemistry 39 (2004) 153–160
[15] G. Biagi, I. Giorgi, O. Livi, F. Pacchini, P. Rum, V. Scartoni, B. Costa,
Acknowledgements
M.R. Mazzoni, L. Giusti, Farmaco 57 (2002) 221–233.
[16] F. Da Settimo, G. Primofiore, S. Taliani, A.M. Marini, C. La Motta,
E. Novellino, G. Greco, A. Lavecchia, L. Trincavelli, C. Martini, J.
Med. Chem. 44 (2001) 316–327.
[17] S. Schenone, O. Bruno, P. Fossa, A. Ranise, G. Menozzi, L. Mosti,
F. Bondavalli, C. Martini, L. Trincavelli, Bioorg. Med. Chem. Lett. 11
(2001) 2529–2531.
[18] F. Bondavalli, M. Botta, O. Bruno, A. Ciacci, F. Corelli, P. Fossa,
A. Lucacchini, F. Manetti, C. Martini, G. Menozzi, L. Mosti, A. Ran-
ise, S. Schenone, A. Tafi, L. Trincavelli, J. Med. Chem. 45 (2002)
4875–4887.
[19] M. Chebib, J.R. Quinn, Bioorg. Med. Chem. 5 (1997) 311–322.
[20] S.A. Poulsen, D.J. Young, J.R. Quinn, Bioorg. Med. Chem. Lett. 11
(2001) 191–193.
[21] S. Schenone, O. Bruno, F. Bondavalli, et al., Abstract of Papers, 16th
National Meeting of the Italian Chemical Society, Pharmaceutical
Division, Sorrento, Italy, 18–22 September 2002.
[22] P. Traxler, G. Bold, J. Frei, M. Lang, N. Lyndon, H. Mett, E. Buch-
dunger, T. Meyer, M. Mueller, P. Furet, J. Med. Chem. 40 (1997)
3601–3616.
[23] J. Stamos, M.X. Sliwkowski, C. Eigenbrot, J. Biol. Chem. 277 (2002)
46265–46272.
[24] V. Colotta, D. Catarzi, F. Varano, L. Cecchi, G. Filacchioni, C. Mar-
tini, L. Trincavelli, A. Lucacchini, J. Med. Chem. 43 (2000) 1158–
1164.
Financial support from Italian MIUR (Cofin 2002, prot.
2002038577_002) is gratefully acknowledged. The study
was partially supported by funds from AIRC (Lucia Mor-
bidelli). We are grateful to Sandra Donnini for technical
assistance in the biological evaluation.
References
[1] J.W. Daly, J. Med. Chem. 25 (1982) 197–207.
[2] C.E. Müller, T. Scior, Pharm. Acta Helv. 68 (1993) 77–111.
[3] K.A. Jacobson, P.J.M. Van Galen, M. Williams, J. Med. Chem. 35
(1992) 407–422 (and references cited therein).
[4] A. Erfurth, M. Schmauss, Fortschr. Neurol. Psychiatr. 63 (1995)
93–98.
[5] P.J. Scammells, S.P. Baker, A.R. Beauglehole, Bioorg. Med. Chem. 6
(1998) 1517–1524.
[6] M. Williams, Med. Res. Rev. 9 (1989) 219–243.
[7] S.A. Poulsen, R. Quinn, Bioorg. Med. Chem. 6 (1998) 619–641.
[8] J.W. Daly, K.D. Hutchinson, S.I. Secunda, D. Shi, W.L. Padgett,
M.T. Shamin, Med. Chem. Res. 4 (1994) 293–306.
[9] S. Psychoyos, C.J. Ford, M.A. Phillips, Biochem. Pharmacol. 31
(1982) 1441–1442.
[10] H.W. Hamilton, D.F. Ortwine, D.F. Worth, J.A. Bristol, J. Med. Chem.
30 (1987) 91–96.
[11] L.P. Davies, S.C. Chow, J.H. Skerrit, D.J. Brown, G.A.R. Johnston,
Life Sci. 34 (1984) 2117–2128.
[25] I.M. Pirovano, A.P. Ijzermann, P.J. van Galen, W. Soudijn, Eur. J.
Pharmacol. 172 (1989) 185–193.
[26] Y. Cheng, W.H. Prusoff, Biochem. Pharmacol. 22 (1973) 3099–3108.
[27] M. Ziche, L. Morbidelli, R. Choudhuri, H.T. Zhang, S. Donnini,
H.J. Granger, R.J. Bicknell, Clin. Invest. 99 (1997) 2625–2634.
[28] Q. Qiu, F. Dudouit, S. Matheson, F. Brahimi, R. Banerjee,
P. McNamee, B. Jean-Claude, Cancer Chemother. Pharmacol. 51
(2003) 1–10.
[12] S.A. Poulsen, J.R. Quinn, J. Med. Chem. 39 (1996) 4156–4161.
[13] F.A. Harden, J.R. Quinn, P.J. Scammells, J. Med. Chem. 34 (1991)
2892–2898.
[29] G. Carpenter, L. King Jr, S. Cohen, J. Biol. Chem. 254 (1979) 4884–
4891.
[14] V. Ralevic, G. Burnstock, Pharmacol. Rev. 50 (1998) 413–492.