Tricyclic oxazolo[2,3-f]purinediones: Potency as adenosine receptor ligands and anticonvulsants

Article abstract of DOI:10.1016/j.bmc.2004.06.043

A series of tricyclic mono- or disubstituted oxazolo[2,3-f]purinediones was synthesized and evaluated for their adenosine receptor affinity and anticonvulsant activity. Mainly selective adenosine A_2A receptor antagonists were identified. A 7-be

Full text of DOI:10.1016/j.bmc.2004.06.043

Bioorganic & Medicinal Chemistry 12 (2004) 4895–4908
Tricyclic oxazolo[2,3-f]purinediones: potency as adenosine
receptor ligands and anticonvulsants
a
Anna Drabczynska, Christa E. Muller, Britta Schumacher, Sonja Hinz,
b
b
b
´
Janina Karolak-Wojciechowska, Barbara Michalak, Elzbieta Pe˛kala
¨
c
a
a
_
a,
*
´
and Katarzyna Kiec-Kononowicz
aJagiellonian University Medical College, Faculty of Pharmacy, Department of Chemical Technology of Drugs, Medyczna 9,
´
Pl 30-688 Krakow, Poland
bPharmaceutical Institute Poppelsdorf, University of Bonn, Kreuzbergweg 26, D-53115 Bonn, Germany
c
_
Institute of General and Ecological Chemistry, Technical University of Łodz, Zwirki 36, Pl 90-924 Łodz, Poland
´ ´
´ ´
Received 28 November 2003; accepted 29 June 2004
Available online 4 August 2004
Abstract—Synthesis and physicochemical properties of 7-mono- and 6,7-disubstituted dihydrooxazolo-[3,2-f]purinediones are
described. Oxazolo[2,3-f]purinediones were synthesized by cyclization of 8-bromotheophylline with oxiranes. The obtained com-
pounds (1–22) were evaluated for their affinity at adenosine A1 and A2A receptors. They showed mainly adenosine A2A receptor
affinity at low micromolar concentrations and A2A selectivity, for example, compound 9 with an octyl substituent at the oxazole
ring displayed adenosine A2A receptor affinity (Ki =0.998lM) and at least 25-fold A2A versus A1 selectivity. This compound was
less selective (5-fold) towards human recombinant A2B and A3 adenosine receptors. In this group of compounds active adenosine
A1 receptor antagonists were also identified. Oxazolopurinediones were evaluated in vivo as anticonvulsants in MES and ScMet
tests and examined for neurotoxicity in mice (ip). Compounds with long alkyl chains showed anticonvulsant activity in both tests
(in 100 and 300mg/kg doses), accompanied by significant neurotoxicity. The anticonvulsant activity in rats (po) was higher and
without signs of neurotoxicity. SAR and QSAR studies stressed the importance of lipophilic 7-substituents for both types of phar-
macological activity. The volume of the substituent is, however, limited at the A2A AR, an n-octyl group being optimal.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1.Introduction
subtypes in the brain. The need for the search of sub-
type-selective AR ligands has been recognized. Cur-
rently, selective A1 AR antagonists are developed as
potassium-saving diuretics with kidney protective prop-
erties for the treatment of acute renal failure. Adenosine
A2A receptors offer an attractive target for CNS dis-
eases, for example, adenosine A2A receptor antagonists
have therapeutic potential in neurodegenerative disor-
ders: ParkinsonÕs and HuntingtonÕs disease, as drugs
controlling motor functions and exhibiting neuroprotec-
tive properties.4,6–9 The involvement of ARs interaction
in seizure disorders was discussed.8–11 Stimulation of
adenosine A1 and A2A receptors was described as play-
ing role in the suppression of seizures (pentylenetetra-
zole induced convulsions and audiogenic seizures in
DBA/2 mice), while the ARs antagonists increased the
incidence of both clonic and tonic seizures.4,10,11
The adenosine receptor (AR) family consists of four
subtypes designated A1, A2A, A2B and A3.1 The re-
sponses of these four ARs are mediated by receptor-cou-
pled G proteins, which may activate several different
effector systems including adenylate cyclase, potassium
and calcium channels, phospholipase A2 or C, and
guanylate cyclase.
On the basis of the widespread effects attributed to the
accumulation of endogenously released adenosine it
has been considered that regulation of ARs has substan-
tial therapeutic potential.2–5 Specificity in its central ac-
tions appears to depend on the distribution of AR
Keywords: Adenosine A1; A2A receptor antagonists; 1,3-Oxazolo[2,3-f]-
purinediones; Anticonvulsant activity; Tricyclic xanthine derivatives.
*
In our research we have been dealing with tricyclic xan-
thine derivatives12,13 conceptually designed as potential
Corresponding author. Tel./fax: +48-12-657-04-88; e-mail: mfkonono@
cyf-kr.edu.pl
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.06.043

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4896
mainly of those that were mono- or disubstituted at
O
CH3
the oxazole ring. Syntheses, structure elucidation by
means of X-ray analysis, in vitro evaluation of the affin-
ity to ARs and in vivo examination of anticonvulsant
properties are described.
R1
O
N
N
N
N
R2
R
8-styrylxanthine derivative
ring formation
2.Chemistry
The synthesis of 7- or 6,7-disubstituted tetra-
hydro[1,3]oxazolo[2,3-f]purinediones was accomplished
as shown in Figure 2. 8-Bromotheophylline15 reacted
with oxiranes in the presence of a catalytic amount of
pyridine in anhydrous propanol, butanol or acetonitrile
(Fig. 2). Intermediate hydroxyalkyl derivatives were not
isolated and subsequently cyclization took place. Struc-
tures, physicochemical properties of the synthesized
compounds and reaction conditions are summarized in
Tables 1 and 2. Compounds 2, 3, 4 and 5 were described
previously16–18 with incomplete spectroscopic data and
without pharmacological investigations. Compound 1
had been obtained by de Martiis et al.16 in two steps.
In our approach it was synthesized as described above.
The structures of the compounds were confirmed by
crystallographic studies of two selected compounds
(see below) and by UV spectral analyses, showing a
slight bathochromic shift of kmax (276–280nm) typical19
for non-8-amino xanthines. In the IR spectra they pre-
sented typical absorption bands20 for cyclic ethers and
O
O
(CH2)n
(CH2)n
X2
R1
1
R
N
N
N
N
N
N
R3
X1
N
R3
O
N
R2
O
N
R2
X1
=
CHR
X2 = O
X2 CHR
X1 = O
=
sterically constrained analogs
Oxygen derivatives
Nitrogen derivatives
Figure 1. Oxygen- or nitrogen-containing annelated theophylline
analogs.
A1 or A2A adenosine receptor antagonists. Compounds
with oxygen or nitrogen containing fused rings were ob-
tained as constrained analogs of 8-styrylxanthines,
which are potent and selective A2A AR antagonists
(Fig. 1).14
1
carbonyl groups of xanthines and in the H NMR spec-
tra, the expected chemical shifts were observed. Spectro-
scopic data are presented in Table 3. The existence of
isomeric structures 1a–22a must be taken into consider-
ation, especially in the case of the disubstituted deriva-
tive 22. The structure of 22 was assigned on the basis
Such tricyclic structures exhibited A1 or A2A adenosine
receptor antagonist properties, in some cases selectivity
to one of the above mentioned subtypes of receptors.
The investigated compounds have shown anticonvulsant
activity in the in vivo tests, however the mechanism of
this action was not clear.
1
of its H NMR spectral data. Crucial for the confirma-
tion of the structure of 22 was the chemical shift of 7-H
(absent in isomer 22a), which is observed for all the
structures in the range of 5.13–5.75ppm and similarly
at 5.29ppm for the structure 22. Compounds 2–11,
13–22 possessing at C-7 and 12 at C-6 and C-7 centres
of asymmetry were obtained as racemates.
As a continuation of our studies, the systematic exami-
nation of oxazolo[2,3-f]purinediones was performed
1
O
O
R
2
H3C
H3C
O
R
1
2
R
N
N
R
pyridine
solvent
NH
N
+
OH
O
N
N
O
Br
Br
N
N
H3C
H3C
O
1
O
2
R
R
H3C
H3C
O
4
6
2
1
3N
N
5
N
R
R
N
or
7
O
2
O
O
1N
N
N
N
H3C
H3C
1 - 22
Figure 2. Synthesis of 7-mono or 6,7-disubstituted dihydrooxazolo[2,3-f]purinediones.
1a - 22a

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4897
Table 1. Structures and elemental analysis of dihydro-1,3-oxazolo[2,3-f]purinediones
No. R1
R2
Mol. form., m. wt.
Calculated (%)
Found (%)
H N
C
H
N
C
1
2
H
H
H
CH3
C9H10N4O3, 222.20
C10H12N4O3, 236.21
48.81
Ref. 17
4.83 25.10 48.88 4.93 25.02
3
H
H
H
H
H
H
H
H
H
C10H11N4O3Cl, 270.67 Ref. 17
Cl
4
C11H14N4O3, 250.26
C10H12N4O4, 252.23
C12H16N4O3, 264.29
C13H18N4O3, 278.31
C15H22N4O3, 306.36
C17H26N4O3, 334.41
C13H16N4O3, 276.29
C15H20N4O3, 304.34
C13H16N4O3
Ref. 18
Ref. 16
54.54
56.11
58.80
61.07
56.51
59.21
56.51
5
OH
6
6.11 21.20 54.54 5.92 21.06
6.53 20.13 56.15 6.65 20.15
7.24 18.29 58.52 6.96 18.05
7.85 16.75 61.42 7.82 16.70
5.84 20.28 56.42 5.96 20.30
6.63 18.41 59.16 6.89 18.11
5.84 20.28 56.58 5.93 20.41
7
8
9
10
11
12
13
H
C16H16N4O3, 312.32
61.54
5.17 17.94 61.25 4.92 17.90
14
15
H
H
C13H18N4O4, 294.31
C14H20N4O4, 308.33
53.05
52.83
6.12 19.03 53.06 6.19 19.01
6.55 18.17 53.20 6.73 17.81
O
O
O
16
17
H
H
C14H16N4O5, 320.30
C16H16N4O4, 328.32
52.50
58.54
5.04 17.49 52.58 5.18 17.70
4.92 17.06 58.49 5.03 16.85
O
O
18
19
H
H
C17H18N4O4, 336.36
C17H18N4O5, 358.35
60.72
56.98
5.40 16.66 60.57 5.27 16.77
5.07 15.63 56.62 4.94 16.00
O
O
O
CH3
20
H
C25H34N4O4, 454.55
66.08
7.55 12.32 65.95 7.41 11.94
O
O
21
22
H
C15H16N4O5, 332.31
C12H16N4O4, 280.28
54.22
51.42
5.14 16.82 53.91 4.99 16.90
5.76 19.99 51.49 5.92 19.61
O
CH3
CH3, CH3
O
3.X-ray structure analysis of compounds 12 and 13
The tetracyclic molecule 12 possesses a semi-circular
shape. Mainly because of this shape, an atypical struc-
tural disorder has been observed in the crystal. During
structure refinement it was established that 30% of the
molecules rotate about the O2ꢀ ꢀ ꢀC8 line and have been
located as it is shown in Figure 3.
As it is clear from Figure 3, both structures of 12 and 13
contain the tricyclic oxazolo[2,3-f]purinedione moiety.
Although nonplanarity of the whole skeleton has been
observed, the purinedione rings are coplanar with a
dihedral angle between both rings equaling only about
2ꢁ. In contrast two carbon atoms of the oxazole ring
(C5a, C9a in 12 and C6, C7 in 13) are in sp3 hybridiza-
tion and such rings in both molecules are in screw con-
formation (Table 4). But the bending of the oxazole ring
in both molecules is different. Thus in 12, due to the fact
that the atoms C5a and C9a are part of a cyclohexane
chair, the asymmetry parameter DC2 in the oxazole ring
is much higher than for 13 (see Table 4). In structure 13,
the phenyl ring of the benzyl substituent is inclined to
the tricyclic oxazolo[2,3-f]purinedione skeleton at 42.8ꢁ.
4.Pharmacology
All compounds were tested in vitro in radioligand bind-
ing assays for affinity to A1 and A2A adenosine receptors
in rat cortical membrane and rat striatial membrane
preparations, respectively. The A1 selective agonist
[3H]2-chloro-N6-cyclopentyladenosine ([3H]CCPA) and
the A2A-selective antagonist [3H]1-propargyl-3-(3-
hydroxypropyl)-7-methyl-8-(m-methoxystyryl)xanthine

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4898
Table 2. Physical properties and reaction conditions of dihydro-1,3-oxazolo[2,3-f]purinediones
No.
Mp ꢁC
Yield (%)
50
Reaction medium
Reaction time (h)
10
Crystal. solvent
Methoxyethanol
TLC Rf
1
265–268
Butanol
Ref. 17
Ref. 17
0.22
0.31
0.35
0.47
0.10
0.47
0.65
0.65
0.59
0.50
0.67
0.65
0.50
0.39
0.58
0.45
0.41
0.45
0.36
0.55
0.46
0.60
2
3
4
Ref. 18
Ref. 16
5
6
187–188
185–186
175–176
155–156
176–178
152–153
245–248
190–191
173–175
141–142
188–190
181–182
172–174
173–175
141–142
157–160
156–158
91
69
96
96
82
97
85
75
83
92
78
78
56
74
80
74
45
Butanol
Butanol
Butanol
Butanol
Butanol
Butanol
Butanol
Butanol
Butanol
Propanol
Butanol
Propanol
Butanol
Butanol
Butanol
Butanol
Acetonitrile
4
4
3
4
2
2
10
3
2
5
3
2
4
3
3
3
2
Ethanol
Ethanol
7
8
Ethanol
Ethanol
Ethanol
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Ethanol+H2O
Ethanol
Ethanol
Ethanol
50% Ethanol
Propanol
Ethanol
Ethanol
Methoxyethanol
Ethanol
Ethanol
Ethanol
([3H]MSX-2) were used as radioligands.21,22 The results
are presented in Table 5.
5.Biological results and conclusions
5.1. In vitro tests
Six selected compounds 9, 12, 13, 15, 17 and 20 were fur-
ther tested for their affinity to human recombinant A2B
and A3 receptors recombinantly expressed in Chinese
hamster ovary (CHO) cells. 4-(2-[7-Amino-2-(2-furyl)-
[1,2,4]-triazolo[2,3-a]triazin-5-[1,3,5](4-(amino)-ethyl)phe-
nol ([3H]ZM241385) was used as a radioligand in
A2B binding studies32 and [3H]2-phenyl-8-ethyl-4-methyl-
(8R)-4,5,7,8-tetrahydro-1H-imidazo[2,1-i]purine-5-one
([3H]PSB-11) was used as A3-selective receptor lig-
and23,33 (Table 6).
The results of the binding assays towards adenosine A1
and A2A receptors (Table 5) showed that oxazolopurine
derivatives had moderate affinity towards these two re-
ceptor subtypes mostly with selectivity for A2A receptors.
The most active substances had long alkyl, alkenyl (7, 8,
9, 10), benzyl (13) or isosteric isobutoxymethyl (15) and
phenoxymethyl (17) substituents in position 7 of the
oxazole ring. Steric properties of the substituent are cru-
cial for the affinity to the adenosine A2A receptors: the
replacement of the isobutoxymethyl group in active com-
pound 15 with an isopropoxymethyl substituent in 14
caused a significant decrease in affinity. Shortening of
the alkyl chain and its substitution with chlorine or hyd-
roxyl (2–5), or introduction of substituted arylethers (18–
21) decreased or even abolished affinity. Also annelation
of a fourth ring (12) or additional substitution in position
6 of the oxazole ring (compound 22) was disadvanta-
geous. On the other hand, it is worth noticing that in
one case (benzyl derivative 13) A1 selectivity (4-fold
A1-selective over A2A) was observed. The best A2A recep-
tor ligand of the present series was 9 with a long lipophi-
lic octyl substituent in position 7 of the oxazole ring (Ki
A2A =0.988lM at least, 25-fold A2A-selective versus A1).
This compound also showed some selectivity towards
human recombinant A2B and A3 adenosine receptors.
However, among the six selected compounds addition-
ally tested at A2B and A3 ARs, representing different
structures (9, 12, 13, 15, 17, 20), compound 9 was the best
ligand for A2B and A3 ARs as well. The other investi-
gated compounds caused only low inhibition of
[3H]ZM241385 binding to the A2B receptor, without
showing any adenosine A3 affinity (Table 6). The most
potent selective ligands from the previously examined
class of nitrogen analogs (Fig. 1) of oxazolopurinediones
All compounds were also tested in vivo as anticonvul-
sants, following the Anticonvulsant Drug Development
(ADD) Program at the National Institute of Neurolog-
ical and Communicative Disorders and Stroke (NIN-
CDS) in Bethesda, MD, using their protocols.24–26
Compounds were injected intraperitoneally into the
mice and evaluated in the preliminary anticonvulsant
screening with at least three dose levels (30, 100 and
300mg/kg). The profile of anticonvulsant activity was
established by maximal electroshock (MES) pattern test
and subcutaneous pentylenetetrazole (ScMet) seizure
test. The MES assays have predictive value for agents
as potential therapeutics in the management of grand
mal epilepsy, whereas the ScMet test is for those likely
to be effective against petit mal.24 Minimal motor
impairment was measured by the rotorod toxicity test.
The results are given in Table 7. Some compounds (6,
7, 10, 11, 12, 13, 14) were administered orally to rats
and examined in the MES, ScMet screen and rotorod
test (Table 8). For one compound (4) a quantitative test
was performed (ED50 and TD50 determination). Results
of this test compared with the literature data available
for valproate27 are reported in Table 9.

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4899
Table 3. Spectroscopic data of compounds 1–22
Compd
IR m (cmꢁ1
)
UV kmax, loge
1H NMR d (ppm)
1
2954-Alkyl, 1709-CO (pos. 2), 1658-CO (pos. 4),
1209-COC, 742-CH2
280
5.02
3.17 (s, 3H, N3CH3), 3.31 (s, 3H, N1CH3), 4.30 (t,
J=7.98Hz, 2H, N5CH2), 5.13 (t, J=8.02Hz, 2H,
CH2)
Ref. 18
2
3
4
5
2955-Alkyl, 1710-CO (pos. 2), 1661-CO (pos. 4),
1141-COC, 744-CH2
280
5.09
278
4.92
280
4.99
280
4.98
2955-Alkyl, 1711-CO (pos. 2), 1657-CO (pos. 4),
1208-COC, 744-CH2
Ref. 18
Ref. 18
2952, 2884-alkyl, 1713-CO (pos. 2), 1656-CO
(pos. 4), 1052-COC, 742-CH2
3407-OH, 2956-alkyl, 1710-CO (pos. 2), 1655-CO
(pos. 4), 1081-COC, 746-CH2
3.16 (s, 3H, N3CH3), 3.30 (s, 3H, N1CH3), 3.62–
3.84 (2m, 2H, N5CH2), 4.07–4.39 (2m, 2H,
CH2OH), 5.30 (t, J=5.5Hz, 1H, OH), 5.46–5.61
(m, 1H, CH)
6
7
2959, 2876-alkyl, 1715-CO (pos. 2), 1655-CO
(pos. 4), 1126-COC, 743-CH2
280
5.06
1.01 (t, J=7.43Hz, 3H, (CH2)2CH3), 1.45–1.52
(m, 2H, CH2–CH2CH3), 1.7–2.02 (2m, 2H,
CH2CH2CH3), 3.37 (s, 3H, N3CH3), 3.50 (s, 3H,
N1CH3), 3.99 (q, J=7.7Hz, 1H, N5CH2), 4.48 (q,
J=8.25Hz, 1H, N5CH2), 5.34–5.44 (m, 1H, CH)
0.91 (t, J=7.15Hz, 3H, (CH2)3CH3), 1.35–1.53
(m, 4H, CH2–(CH2)2CH3), 1.81–2.02 (2m, 2H,
CH2(CH2)2CH3), 3.34 (s, 3H, N3CH3), 3.47 (s,
3H, N1CH3), 3.95–4.01 (q, J=7.7Hz, 1H,
N5CH2), 4.43–4.49 (q, J=7.7Hz, 1H, N5CH2),
5.32–5.41 (m, 1H, CH)
2958, 2932-alkyl, 1715-CO (pos. 2), 1655-CO
(pos. 4), 1126-COC, 743-CH2
280
4.97
8
9
2956, 2929, 2858-alkyl, 1715-CO (pos. 2),
1655-CO (pos. 4), 1086-COC, 744-CH2
280
5.06
0.88 (t, J=6.74Hz, 3H, (CH2)5CH3), 1.29–1.54
(m, 8H, CH2–(CH2)4CH3), 1.81–2.01 (2m, 2H,
CH2(CH2)4CH3), 3.37 (s, 3H, N3CH3), 3.50 (s,
3H, N1CH3), 3.95–4.02 (dd, J=7.70Hz, 1H,
N5CH2), 4.45–4.50 (dd, J=7.57Hz, 1H, N5CH2),
5.38–5.42 (m, 1H, CH)
2956, 2926, 2856-alkyl, 1715-CO (pos. 2),
1654-CO (pos. 4), 1051-COC, 743-CH2
280
5.06
0.87 (t, J=6.74Hz, 3H, (CH2)7CH3), 1.26–1.54
(m, 12H, CH2(CH2)6CH3), 1.94–2.01 (m, 1H,
CH2(CH2)6CH3), 1.94–2.06 (m, 1H,
CH2(CH2)6CH3), 3.37 (s, 3H, N3CH3), 3.51 (s,
3H, N1CH3), 3.97–4.02 (q, J=7.70Hz, 1H,
N5CH2), 4.45–4.51 (q, J=8.11Hz, 1H, N5CH2),
5.33–5.41 (m, 1H, CH)
10
11
3075-Alkenyl, 2954-alkyl, 1714-CO (pos. 2),
1655-CO (pos. 4), 1210-COC, 744-CH2
280
5.10
1.92–2.34 (3·m, 4H, (CH2)2), 3.38 (s, 3H,
N3CH3), 3.51 (s, 3H, N1CH3), 4.01 (q, J=7.56Hz,
1H, N5CH2), 4.49 (t, J=8.94Hz, 1H, N5CH2),
5.06–5.14 (m, 2H, CH2@CH), 5.35–5.45 (m, 1H,
CH2@CH), 5.75–5.88 (m, 1H, CH)
1.44–2.12 (m, 8H, (CH2)4), 3.38 (s, 3H, N3CH3),
3.51 (s, 3H, N1CH3), 4.00 (q, J=7.70Hz, 1H,
N5CH2), 4.45–4.50 (m, 1H, N5CH2), 4.95–5.04 (m,
2H, CH2@CH), 5.33–5.43 (m, 1H, CH2@CH),
5.61–5.84 (m, 1H, CH)
3077-Alkenyl, 2991, 2859-alkyl, 1713-CO (pos. 2),
1659-CO (pos. 4), 1097-COC, 744-CH2
280
5.01
12
13
2955–2873-Alkyl, 1706-CO (pos. 2), 1658-CO
(pos. 4), 1199-COC, 744-CH2
280
4.93
1.38–2.00 (m, 6H, cyclohexyl), 2.39–2.45 (m, 1H,
cyclohexyl), 2.87–2.92 (m, 1H, cyclohexyl), 3.38 (s,
3H, N3CH3), 3.52 (s, 3H, N1CH3), 3.65–3.74 (dt,
1H, N5CH), 4.47–4.56 (dt, 1H, OCH)–
3.14–3.21 (m, 1H, CH2), 3.31 (d, J=6.33Hz, 1H,
CH2), 3.36 (s, 3H, N3CH3), 3.51 (s, 3H, N1CH3),
4.10 (q, J=7.70Hz, 1H, N5CH2), 4.39 (q,
J=8.25Hz, 1H, N5CH2), 5.54–5.65 (m, 1H, CH),
7.37 (m, 5H, phenyl)
3025-Phenyl, 2948-alkyl, 1703-CO (pos. 2),
1653-CO (pos. 4), 1059-COC, 746-CH2
280
4.97
14
15
2971, 2934, 2867-alkyl, 1714-CO (pos. 2),
1655-CO (pos. 4), 1149-COC, 744-CH2
280
4.76
1.12–1.15 (m, 6H, 2CH3), 3.38 (s, 3H, N3CH3),
3.51 (s, 3H, N1CH3), 3.61–3.73 (m, 1H,
CH(CH3)2), 3.75–3.83 (m, 2H, N5CH2), 4.30–4.45
(m, 2H, CH2O), 5.42–5.50 (m, 1H, CH)
0.84 (q, J=3.03Hz, 6H, 2CH3), 1.80–1.87 (m, 1H,
CH(CH3)2), 3.28 (d, J=6.60Hz, 2H, CH2CH),
3.38 (s, 3H, N3CH3), 3.51 (s, 3H, N1CH3), 3.78
(m, 2H, N5CH2), 4.31–4.44 (m, 2H, CH2O),
5.43–5.51 (m, 1H, CH)
2957, 2874-alkyl, 1714-CO (pos. 2), 1672-CO
(pos. 4), 1120-COC, 742-CH2
280
4.97
(continued on next page)

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4900
Table 3 (continued)
Compd
IR m (cmꢁ1
)
UV kmax, loge
1H NMR d (ppm)
16
2955-Alkyl, 1719-CO (pos. 2), 1651-CO (pos. 4),
1294-ester, 1163-COC, 743-CH2
280
5.01
1.88 (s, 3H, CH3), 3.38 (s, 3H, N3CH3), 3.51 (s,
3H, N1CH3), 4.24–4.30 (m, 1H, N5CH2), 4.44–4.6
(m, 3H, N5CH2 +CH2O), 5.61–5.67 (m, 2H,
@CH2), 6.04 (s, 1H, CH)
17
18
3025-Phenyl, 2948-alkyl, 1703-CO (pos. 2),
1653-CO (pos. 4), 1059-COC, 746-CH2
276
5.03
3.39 (s, 3H, N3CH3), 3.52 (s, 3H, N1CH3), 4.30
(dd, J=3.99Hz, 1H, N5CH2), 4.40 (dd,
J=3.99Hz, 1H, N5CH2), 4.47–4.60 (m, 2H,
CH2O), 5.66–5.74 (m, 1H, CH), 6.98 (m, 2H,
2060phenyl), 6.95 (m, 1H, 40phenyl), 7.26–7.32 (m,
2H, 3050phenyl)
3064-Phenyl, 2952-alkyl, 1704-CO (pos. 2),
1655-CO (pos. 4), 1240-O phenyl, 1127-COC,
750-CH2
276
4.93
2.04 (s, 3H, CH3), 3.40 (s, 3H, N3CH3), 3.53 (s,
3H, N1CH3), 4.29 (dd, J=3.58Hz, 1H, N5CH2),
4.40 (dd, J=3.58Hz, 1H, N5CH2), 4.49–4.62 (m,
2H, CH2O), 5.68–5.75 (m, 1H, CH), 6.80 (d,
J=7.98Hz, 1H, 60phenyl), 6.91 (t, J=7.4Hz,
30phenyl) 7.11–7.29 (m, 2H, 4050phenyl)
3.40 (s, 3H, N3CH3), 3.53 (s, 3H, N1CH3), 3.76 (s,
3H, OCH3), 4.25 (dd, J=3.85Hz, 1H, N5CH2),
4.34 (dd, J=4.13Hz, 1H, N5CH2), 4.47–4.59 (m,
2H, CH2O), 5.64–5.70 (m, 1H, CH), 6.83 (s, 4H,
phenyl)
19
20
21
22
3064-Phenyl, 2951, 2888-alkyl, 1704-CO (pos. 2),
1654-CO (pos. 4), 1240-O phenyl, 744-CH2
280
4.93
2959-Phenyl, 2929–2872-alkyl, 1709-CO (pos. 2),
1658-CO (pos. 4), 1243-O phenyl, 1123-COC,
743-CH2
280
5.04
0.77–1.63 (3·m, 19H, (CH2)8CH3), 3.40 (s, 3H,
N3CH3), 3.53 (s, 3H, N1CH3), 4.28 (q, J=3.71Hz,
1H, N5CH2), 4.38 (m, 1H, N5CH2), 4.47–4.59 (m,
2H, CH2O), 5.64–5.72 (m, 1H, CH), 6.79–6.82 (m,
2H, 2060phenyl), 7.15–7.27 (m, 2H, 3050phenyl)
3.37 (s, 3H, N3CH3), 3.50 (s, 3H, N1CH3), 3.81
(dd, J=4.12Hz, 1H, N5CH2), 3.86 (dd,
J=4.04Hz, 1H, N5CH2), 4.26–4.58 (m, 4H, CH2–
O–CH2), 5.41–5.49 (m, 1H, CH), 6.33 (m, 2H,
3040furan), 7.40 (s, 1H, 20furan)
3121-Phenyl, 2952-alkyl, 1712-CO (pos. 2),
1655-CO (pos. 4), 1110-COC, 879-phenyl,
744-CH2
280
5.03
2955, 2851-alkyl, 1703-CO (pos. 2), 1656-CO (pos.
4), 1099-COC, 749-CH2
278
5.02
1.58 (d, J=9.08Hz, 6H, 2CH3), 3.39 (s, 3H,
N3CH3), 3.52 (s, 3H, N1CH3), 3.66 (s, 3H, OCH3),
5.29 (s, 1H, CH)
were investigated for their antagonistic properties in ra-
dioligand binding sodium shift assays and could be char-
acterized as antagonists for A2AARs.12 Taking into
account the structural similarity of currently investigated
group of compounds, we may presume that oxazolo[2,3-
f]purinediones have an A2AAR antagonistic profile of
activity as well.
The most active compounds in mice tests were also
examined for their activity in rats showing after oral
administration even much better seizure protection
(Table 8). In the MES test after administration of
30mg/kg dose po anticonvulsant activity without neuro-
toxic symptoms was shown in all the tested compounds
except compound 14. Compound 12 active in mice only
in the ScMet test showed, after oral administration in
rats, activity in the MES test. Compound 6 active in
both tests in mice, showed only protection in ScMet test
in rats at a maximal dose of 50mg/kg. Compound 4 was
advanced to phase II evaluation for quantification of
activities (ED50 and TD50) against ScMet induced sei-
zures. These pharmacological parameters were com-
pared with data for valproate (Table 9). The time to
peak effect (TPE) was comparable but the protective in-
dex (PI) was two times lower than that for valproate.
5.2. In vivo tests
The character of the substituent at the oxazole ring
seems to have a significant influence on the profile and
strength of the anticonvulsant activity (Table 7). Com-
pounds with alkyl and alkenyl substituents (from 2–6
C atoms) and a benzyl moiety (comp. 4, 6, 7, 8, 10,
11, 13) showed protective activity in both tests in short
time (0.5h). For compounds 7 and 10 the protective
activity appeared within 0.25h after injection and lasted
for 4h. Some compounds (3, 12, 14, 15) showed protec-
tion only in chemical seizures. In the neurotoxicity test
all of the substances showed toxicity at the active dose.
Such signs as myoclonic jerks (13), loss of righting reflex
(6, 10, 11), tonic extension (11), continuous seizure
activity (13) and even death (15, 22) appeared. Introduc-
ing a phenoxyalkyl moiety in position 7 (17–21) or a sec-
ond substituent in position 6 of the oxazole ring (22)
completely abolished anticonvulsant activity.
The anticonvulsant activity of the examined compounds
was analyzed for correlation with ARs affinity. Different
effects were observed. Compound 9 the most active and
selective A2A AR antagonist was deprived as well of
anticonvulsant as neurotoxic activity. Structure 4 not
active as A1 or A2A AR antagonist has shown short time
lasting anticonvulsant activity in both used test (being
classified to I ASP class). Contrary to that results com-
pound 13 showing in some degree selective A1 AR affin-

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4901
clusion it should be noticed that the apparent correla-
tion of proconvulsant or anticonvulsant activity (in
MES or ScMet induced seizures) with AR affinity was
not observed.
5.3. SAR and QSAR studies
In preliminary structure and activity studies, the percent
inhibition of oxazolopurinediones (%) of radioligand
binding to both studied adenosine receptors (A1, A2A
)
has been examined as a function of three lipophilicity
descriptors logP, logD and clogP (logP and logD were
calculated34 clogP was obtained experimentally35). All
indispensable numerical data are collected in Table 5
while results in graphical form are presented in Figure
4. For all three descriptors second-degree polynominal
relations have been obtained but rather unsatisfying.
However after omitting only one point (compound 20)
linear relations have been obtained with significantly
higher correlation coefficient. The respective relations
have suggested that the compounds with maximal inhi-
bition should possess clogP, logP, and logD about 3.5.
As a basis for subsequent QSAR analysis, the set of
eight oxazolopurinediones with known Ki versus
[3H]MSX (Table 5) has been taken into account. The
data finally selected as the descriptors are collected in
Table 10. In the first stage, the relationship of Ki =
f(clogP) has been established keeping, as before in the
preliminary study, a second-degree polynominal charac-
ter (r2 =0.52) with minimal Ki at clogP about 3 (Fig.
5a). It agrees well with the relationship from Figure 4
for compounds exhibiting the highest affinity to the
receptor.
Figure 3. ORTEP drawings of 12 and 13. For 12 molecule with
sof=0.3 has been designated by dotted lines.
Due to the lack of suitable descriptors, molecular mode-
ling36 and geometry optimization40 for all eight oxazo-
lopurines were performed. In order to broaden our
knowledge on this subject, the volumes of the substitu-
ents at the main tricyclic oxazolo[2,3-f]purinedione skel-
eton (DV) were calculated and selected as descriptors
(Table 10). From the linear relationship (Fig. 5b) it
ity in dose of 100mg/kg protected against MES and
ScMet induced seizures, but in the same dose presented
toxicity. Similarly 7 and 10 showing although week
A1AR affinity were active in both tests at doses
100mg/kg accompanied with neurotoxicity. As the con-
Table 4. Selected geometry details for 12 (for molecule with sof=0.7 only) and 13
12
13
(a) Torsion angles (ꢁ)
N5–C5a–C9a–O10
C5a–C9a–O10–C10a
C9a–O10–C10a–N5
O10–C10a–N5–C5a
C10a–N5–C5a–C9a
ꢁ17.6
22.7
4.3
N5–C6–C7–O8
C6–C7–O8–C8a
C7–O8–C8a–N5
O8–C8a–N5–C6
C8a–N5–C6–C7
3.7
ꢁ1.9
4.0
11.9
ꢁ18.8
ꢁ1.8
ꢁ4.6
˚
(b) Deviation of the selected atoms from purine ring plane (A)
C5a
C9a
O10
0.327 (5)
ꢁ0.030 (5)
C6
C7
O8
ꢁ0.153 (5)
ꢁ0.156 (5)
ꢁ0.099 (4)
0.021 (4)
˚
(c) Random deviation purine ring atoms (A)
0.033
0.171
(d) Asymmetry parameter (twofold-axis) (ꢁ)
DC2 (C2a. . . O10–C10a)
12.8ꢁ
DC2 (N5. . . O8–C7)
2.0ꢁ

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4902
Table 5. Affinities of oxazolo[2,3-f]purinediones at A1 and A2A adenosine receptors and lipophilicity parameters
No.
R
A1 versus
Ki SEM versus
[3H]CCPA
A2A versus
Ki SEM versus
[3H]MSX-2
LogP
clogP
ꢁ0.50
LogD
(pH7.4)
[3H]CCPAa
% inhib. SEM
concn 25lM
[3H]MSX-2b
% inhib. SEM
concn 25lM
n=3 (lM)
n=3 (lM)
1
2
H
CH3
8
11
2
1
>25
>25
15
2
0
>25
>25
ꢁ0.95
ꢁ0.42
ꢁ0.95
ꢁ0.42
3
7
4
>25
>25
0
2
3
2
4
6
4
9
>25
ꢁ0.36
0.09
ꢁ1.84
0.60
1.11
2.13
3.15
0.76
1.78
ꢁ0.67
0.34
ꢁ0.20
0.09
ꢁ1.68
0.60
1.11
2.13
3.66
0.76
1.78
4
11 11
20
27
50
57
87
90
48
>25
5
6
21
29
3
6
8
>25
>25
ꢁ0.80
0.34
6
>25
26.0 0.80
16.1 2.87
3.70 0.85
0.998 0.05
22.7 5.8
7
>25
0.80
8
40 10
P 25
>25
2.13
9
33
41
7
9
3.60
10
P 25
0.43
11
12
30
20
3
2
>25
>25
47
42
6
7
P 25
P 25
13
69
3
2.07 0.15
80
6
8.39 1.76
1.26
0.51
1.26
14
15
6
2
2
>25
>25
8
4
7
>25
ꢁ0.21
ꢁ0.08
ꢁ0.05
18
55
2.67 0.54
0.06
0.66
0.22
16
17
9
3
6
>25
>25
9
1
>25
ꢁ0.04
ꢁ0.32
0.11
0.78
35
69 2 (n=4)
12.9 3.9
0.63
0.70
18
19
20
37 12
>25
>25
>25
48
3
P 25
>25
>25
1.07
0.58
6.19
1.01
0.96
5.55
1.22
0.74
5.32
9
8
8
33 11
16
37
4
21
22
18
15
6
7
>25
>25
24
32
1
3
>25
>25
ꢁ0.04
0.34
ꢁ0.04
0.34
a At rat brain cortical membranes.
b At rat brain striatal membranes.
Table 6. Binding of oxazolo[2,3-f]purinediones to human recombinant A2B and A3 adenosine receptors
Compd
A2B human versus
[3H]ZM241385
Ki SEM (lM)
A3 human versus
[3H]PSB-11
Ki SEM (lM)
% inhibition SEM
concn 10lM (n=2)
% inhibition SEM
concn 10lM (n=2)
9
56
24
17
29
39
31
1
4
5
9
0
1
5.2 2.7
>10
>10
39
0
2
0
1
0
4
0
12.3 0.6a
>10
>10
12
13
15
17
20
16
0
>10
>10
>10
>10
22
2
>10
>10
a Estimated value by extrapolation; full curve could not be determined due to limited solubility of the compound.

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4903
Table 7. Anticonvulsant activity and neurotoxicity of oxazolo[2,3-f]purinediones
Compda
MESb
0.5h
ScMetb
0.5h
Toxicityb
0.5h
ASP classc
0.25h
1h
4h
0.25h
1h
4h
0.25h
4h
1
––
––
––
––
––
––
––
––
––
––
––
300
300
––
––
––
––
––
300
––
300e
300e
––
300f
100e
300e
––
300d
––
3
3
2
1
3
2
1
2
3
1
1
2
2
3
––
300
100
––g
300
100
300
––
––
––
––
––
4
300
––
5
––
––
––
––
6
300
300
300
––
7
100
100
300
––
100f
100e
––
––
8
9
––
––
10
11
12
13
300
300
––
100
300
300
100i
300
300
––
100
300f
––
300f
––
100
––
100
100
300e
300e
300e
300
––
––
––
––
––
––
––
––
––
––
––
––
––
300
––
––
––
––
––
––
––
––
––
1
2
2
3
3
3
3
3
3
3
14
15
16
17
18
19
20
21
22
––
––
––
––
––
––
––
––
––
300j
300e
––
––
––
––
––
––
––
––
––
––
––
––
––
––
––
h
300f,d
300d
300
300
Test results in mice after intraperitoneal injection.
a Suspension in 0.5% methylcellulose.
b Doses of 30, 100, 300mg/kg were administered. The figures in the table indicate the minimum dose whereby activity was demonstrated in half or
more of the mice. The animals were examined 0.5 and 4h after injections were made. For compounds 7, 10, 11 biological response was after 0.25h.
The dash (––) indicates an absence of activity at the maximum dose administered (300mg/kg).
c Classification is as follows: (1) Anticonvulsant activity was observed (even in less number than 50% of animals) at 100mg/kg or less, (2) anti-
convulsant activity at doses greater than 100mg/kg, (3) compound inactive at 300mg/kg.
d Death.
e Unable to grasp rotorod.
f
Loss of righting reflex.
g Continuous seizure activity at dose 30mg/kg.
h Death following tonic extension at dose 100mg/kg.
i
Myoclonic jerks.
Death following continuous seizure at dose 30mg/kg.
j
Table 8. Anticonvulsant activity and neurotoxicity of selected compounds after oral administration (30mg/kg or 50mg/kg) to rats
Compda
MES
1h
ScMet
1h
Toxicity
0.25h
30
0.5h
30
2h
4h
30
0.25h
0.5h
4h
0.25h
0.5h
1h
4h
6
50
––
––
––
––
––
––
––b
––
––
––
––
––
––
7
30
30
30
10
11
12
13
14
30
30
30
30
30
a Form––suspension in 0.5% methylcellulose.
b The dash (––) indicates an absence of activity and toxicity at the maximum dose administered (50mg/kg).
has been deduced that the substituent volume DV should
tivity for A2A versus A1 receptors. These compounds
were weak and neurotoxic anticonvulsants in mice tests,
after intraperitoneal administration, showing much bet-
ter activity in rats and lacking the neurotoxicity after oral
administration. Although no apparent correlation be-
tween anticonvulsant activity and adenosine recep-
tor affinity (the best A2A ligand 9 was inactive as
3
be limited to 400 cubic (A ).
˚
6.Conclusions
Tricyclic oxazolo[2,3-f]purinediones are adenosine recep-
tor ligands with moderate affinity mostly exhibiting selec-

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4904
Table 9. Quantitative anticonvulsant activity and neurotoxicity of 4 and valproate in mice ip
TPEa (h)
TD50
ED50 ScMetc
PI (TD50/ED50)
b
d
Compd
4
0.25/0.5
151.83
132.96
(108.02–180.07)
[6.82]
483
(110.22–151.66)
[10.53]
209
1.14
2.3
Valproatee
0.25/0.5
a Time to peak effect. The first value is for rotorod test (dose 200mg/kg), the second is for the ScMet test (175mg/kg). In the neurotoxicity assay, 4
was tested at 0.25–0.5h, in ScMet at 0.25–1h.
b Dose (mg/kg) eliciting evidence of minimal neurological toxicity in 50% of animals; 95% confidence interval is shown in parentheses; the slope of the
regression line is shown in brackets.
c Dose (mg/kg) eliciting the ScMet protection in 50% animals.
d Protective index–neurotoxic dose/median effective dose.
e Data from Ref. 27.
Table 10. Similarity descriptors used in QSAR analysis
100
[%] = 20.01 logP + 29.59
% inhibition
Ki versus [3H]MSX-2
clogP
Volume of the
3
substituent (A )
r
2 = 0.62
Compd
˚
[%] = 21.80 clogP +31.08
80
60
40
20
0
r
2 = 0.65
6
26.00
16.10
3.70
0.34
0.80
2.13
3.60
0.43
0.51
0.66
0.70
159.51
212.53
321.57
430.99
197.25
267.81
352.60
309.75
7
[%] = 19.24 logD + 28.28
8
r
2 = 0.61
9
1.00
10
13
15
17
22.70
8.39
2.67
logP
clogP
logD
12.90
A1 vs [3H]CCPA % inhibition
SEM conc. 25 µM
logP, clogP, logD
recorded as KBr discs on an FT Jasco IR 410 spectro-
meter. H NMR spectra were performed with a Var-
-4
-2
0
2
4
6
8
1
100
% inhibition
[%] = 19.52 logP + 28.83
ian-Mercury 300MHz spectrometer in DMSO-d6 (1
and 5) or CDCl3 (the remaining compounds given) in
r units, using TMS as an internal standard.
r
2 = 0.55
[%] = 21.04 clogP+29.41
r 2 = 0.54
80
TLC data were obtained with Merck silica gel 60F254 al-
uminum sheets using benzene/acetone (7:3) as develop-
ing system. The plates were visualized with UV light.
UV spectra were recorded on a Jasco UV–vis V-530
apparatus in a concentration of 10ꢁ5 mol/L in methanol.
Elemental analyses were performed on an Elemental
Vario-EL III apparatus.
[%] = 18.64 logD + 27.62
60
40
20
0
r
2 = 0.54
logP
clogP
logD
A2A vs [3H]MSX % inhibition
SEM conc. 25 µM
logP, clogP, logD
7.2. General procedure for the synthesis of 7-monosub-
stituted or 6,7-disubstituted 1,3-dimethyl-6,7-dihydro-1,3-
oxazolo[2,3-f]purine-2,4(1H,3H)-dione
-4
-2
0
2
4
6
8
Figure 4. Relations between % inhibition of radioligand binding and
lipophilicity descriptors.
A mixture of 8-bromotheophylline (1, 2.58g, 10mmol)
and 20mmol of the corresponding oxirane was heated
under reflux in anhydrous propanol, butanol or aceto-
nitrile (see Tables 1, 2) in the presence of a catalytic
amount of pyridine (0.3mL) for 2–10h. After dissolu-
tion of the starting substance (10min to 2h) the progress
of the reaction was monitored by means of TLC show-
ing three spots (starting material and probably ring-
open and cyclic product). After further heating the spot
of the open-chain product disappeared (2–10h). Some
main products showed only a little amount of starting
material. After cooling the precipitate was collected by
filtration and mixed with 10% NaOH to remove eventu-
ally the rest of unreacted 8-bromotheophylline and sub-
sequently washed with water. A second crop (of worse
quality) could be obtained by evaporating the solvent,
mixing the residue with 10% NaOH and washing with
anticonvulsant) was observed, it seems that a lipophilic
group at the oxazole ring is necessary for both activities
and the volume of the substituent being of the necessary
optimal size is important for high adenosine A2A receptor
affinity.
7.Experimental
7.1. Chemistry
Melting points were determined on a MEL-TEMP II
apparatus and are uncorrected. IR spectra were

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4905
Ki = 3.34 clogP2- 18.28 clogP + 24.46
r2 = 0.52
Ki
Ki = - 0.096 ∆V+ 38.764
r2 = 0.85
Ki
25
20
15
10
5
25
20
15
10
5
0
0
100
∆V
clogP
0
1
2
(a)
3
300
(b)
400
200
Figure 5. (a) Correlation between Ki-value at A2A adenosine receptors and clogP. (b) Correlation between Ki-value at A2A adenosine receptors and
substituent volume.
water. Combined precipitates were purified by
crystallization.
and binding at the test concentration, curves were deter-
mined at the A1 and A2A ARs using 6–7 different
concentrations of test compounds spanning 3 orders of
magnitude. At least three separate experiments were per-
formed each in triplicate. Data were analyzed using the
PRISM program version 3.0 (GraphPad, San Diego,
CA, USA).
7.3. Analytical chromatographic measurements35
Analytical TLC was performed on 20·20cm pre-coated
RP-TLC Al sheets of RP-18, F254S (Merck); samples of
2.5lL of the solutes (0.1mg/mL in methanol) were spot-
ted with a Hamilton microliter syringe. The chromato-
grams were developed over a distance of 17.5cm. The
chambers were saturated with organic solvent vapor
for 20min. In the performed studies the water/methanol
mixtures were used as mobile phases. The concentration
of the organic modifier in the mobile phase ranged from
80–98%. All TLC measurements were performed at
21ꢁC. Spots were visualized under UV light at 254nm.
8.2. Anticonvulsant screening
The anticonvulsant evaluation was carried out using re-
ported procedures.25,26 Male albino mice (CF-1 strain,
18–25g), and male albino rats (Sprague-Dawley 100–
150g) were used as experimental animals. Groups of
1–5 mice were used in MES and ScMet tests, groups
of 2–8 animals in rotorod tests. For the evaluation of
activity after oral administration groups of 4 rats were
used. The tested compounds were suspended in a 0.5%
methyl–cellulose–water mixture.
8.Pharmacology
8.1. Adenosine receptor binding assays
In the preliminary screening each compound was admin-
istered as an ip injection at three dose levels (30, 100 and
300mg/kg) with anticonvulsant activity and neurotoxic-
ity assessed at 30min and 4h intervals after administra-
tion. For some compounds also intervals of 15min and
1h were used.
Radioligand binding assays were performed as previ-
ously described using rat brain cortical membrane prep-
arations for A1 AR assays, and rat brain striatal
membrane preparations for A2A AR assays.28–30 Frozen
rat brains (unstripped) were obtained from Pel-Freezꢂ,
Rogers, Arkansas, USA. For assays at human A3
ARs, CHO cell membranes containing the human A3
AR were used as described.28,31 [3H]2-chloro-N6-cyclo-
pentyladenosine ([3H]CCPA) was used as the A1 radio-
ligand, [3H]3-(3-hydroxypropyl)-7-methyl-8-(m-methox-
ystyryl)-1-propargylxanthine ([3H]MSX-2) as the A2A
ligand,21 [3H]4-(2-[7-amino-2-(2-furyl)-[1,2,4]-triazolo-
[2,3-a][1,3,5]-triazin-5-(4-(amino)-ethyl)phenol ([3H]ZM-
241385) as the A2B receptor ligand,32 and [3H]2-phen-
yl-8-ethyl-4-methyl-(8R)-4,5,7,8-tetrahydro-1H-imidazo-
[2,1-i]purine-5-one ([3H]PSB-11) as the A3 AR radiolig-
and.23,33 Initially, a single high concentration of com-
pound (25lM at A1 and A2A, 10lM at A2B and A3
Anticonvulsant efficacy was measured by maximal elec-
troshock (MES) and subcutaneous pentylenetetrazole
(ScMet), neurological deficit was investigated in the
rotorod tests; the data are presented in Table 7. Some
selected derivatives were examined for oral activity in
the rat MES, ScMet and neurotoxicity screen at 30
and 50mg/kg doses. The results are summarized in
Table 8.
The pharmacological parameters estimated in the prelim-
inary screening were quantified for compound 4 (Table 9).
Anticonvulsant activity was expressed in terms of the
median effective dose (ED50) and neurotoxicity was ex-
pressed as the median toxic dose (TD50). For determina-
tion of the ED50 and TD50, groups of 8 mice were given
receptors) was tested in three (A1, A2A) or two (A2B
A3) independent experiments. For potent compounds,
which showed greater than 50% inhibition of radiolig-
,

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4906
Table 11. Crystal data and structure refinement details for 12 and 13
12
13
Empirical formula
Formula weight
Temperature
C13H16N4O3
276.30
293 (2)K293 (2)K
C16H16N4O3
312.33
˚
0.71073A
˚
0.71073A
Wavelength
Crystal system
Space group
Unit cell dimensions
Monoclinic
P21/n
˚
a=10.238 (2)A
Monoclinic
P21/c
a=9.604 (2)A
˚
˚
˚
b=9.604 (2)A
b=9.344 (2)A
˚
c=13.547 (3)A
˚
c=14.500 (3)A
b=94.88 (3)ꢁ
b=102.45 (3)ꢁ
1498.9 (5)
3
˚
Volume
1291.3 (5)A
Z, Calculated density
Absorption coefficient
F (000)
4, 1.421Mg/m3
0.104mmꢁ1
584
4, 1.384Mg/m3
0.099mmꢁ1
656
Crystal size
0.2·0.2·0.4mm
2.96ꢁ–25.0ꢁ
ꢁ12:h:12, ꢁ11:k:11, ꢁ16:l:15
KM-4 with CCD detector
6499/2208 [R(int)=0.0994]
Full-matrix-block least-squares on F2
2208/263
0.25·0.3·0.5mm
2.98ꢁ–23.0ꢁ
ꢁ10:h:10, ꢁ12:k:12, ꢁ15:l:15
KM-4 with CCD detector
12953/2079 [R(int)=0.1024]
Full-matrix-block least-squares on F2
2079/210
Theta range for data
Index ranges
Diffractometer
Refl. collected/unique
Refinement method
Data/parameters
Goodness-of-fit on F2
Final R [I > 2r (I)]
Largest diff. peak and hole
1.005
R1=0.0686, wR2=0.ꢁ13529
1.099
R1=0.0688, wR2=0.ꢁ13701
˚
˚
0.198 and ꢁ0.175eA
0.285 and ꢁ0.184eA
4
2
3
˚
Table 12. Atomic coordinates (·10 ) and equivalent isotropic displacement parameters (A ·10 ) for (12) and (13)
12
Molecule with sof=0.7
2363 (3)
Molecule with sof=0.3
N1
3964 (2)
3902 (3)
2871 (3)
1867 (2)
2946 (2)
1690 (3)
3982 (2)
3977 (2)
5065 (3)
6408 (2)
7230 (3)
7400 (4)
8419 (4)
9732 (3)
9610 (3)
8541 (4)
8249 (2)
6998 (3)
6226 (2)
5148 (3)
4768 (2)
3968 (3)
4949 (3)
4379 (2)
5682 (2)
5786 (2)
6343 (3)
7029 (2)
6089 (2)
6558 (2)
7480 (3)
8066 (3)
8855 (3)
8476 (3)
7864 (3)
7070 (4)
6339 (2)
6034 (3)
5315 (2)
5377 (3)
57 (1)
69 (1)
57 (1)
68 (1)
61 (1)
81 (1)
48 (1)
60 (1)
56 (1)
55 (1)
76 (1)
93 (2)
15 (2)
192 (2)
83 (2)
01 (2)
171 (1)
69 (1)
59 (1)
56 (1)
N10
C1a0
C20
4009 (6)
4123 (9)
2780 (8)
1753 (6)
2880 (5)
1790 (5)
5129 (6)
6316 (5)
5291 (5)
6101 (5)
7388 (5)
7357 (6)
8464 (6)
9679 (7)
9658 (7)
8520 (5)
6216 (4)
5501 (7)
4360 (6)
4264 (6)
2420 (6)
3482 (9)
1608 (8)
1836 (7)
605 (6)
4808 (5)
3990 (8)
4875 (6)
4408 (5)
5760 (4)
6106 (5)
5343 (5)
5065 (4)
6257 (4)
6803 (5)
7311 (4)
7936 (5)
8821 (5)
8602 (6)
7818 (6)
7075 (5)
8070 (4)
7298 (6)
6957 (6)
6310 (5)
51 (2)
88 (3)
51 (2)
83 (2)
39 (2)
29 (2)
28 (2)
67 (2)
14 (1)
50 (2)
17 (1)
38 (2)
29 (2)
52 (2)
48 (2)
26 (2)
49 (1)
54 (2)
70 (2)
39 (2)
C1a
C2
3388 (3)
1640 (3)
1800 (2)
692 (3)
O2
N3
O20
N30
C3a0
C40
O40
C4a0
N50
C3a
C4
ꢁ35 (4)
296 (3)
ꢁ271 (6)
1911 (6)
2856 (6)
1133 (5)
652 (6)
O4
ꢁ532 (2)
1250 (3)
1164 (2)
853 (4)
C4a
N5
C5a
C6
C5a0
C60
1142 (5)
ꢁ580 (7)
ꢁ172 (7)
165 (8)
ꢁ238 (3)
ꢁ324 (5)
401 (4)
C7
C8
C70
C80
C9
1580 (4)
1513 (4)
441 (2)
C90
1605 (8)
1174 (6)
ꢁ852 (5)
ꢁ312 (8)
ꢁ659 (7)
597 (7)
C9a
O10
C10a
N11
C11a
C9a0
O100
C10a0
N110
C11a0
2157 (3)
2828 (3)
2284 (3)
13
N1
C1a
C2
5486 (3)
6895 (4)
4956 (4)
5671 (3)
3619 (3)
3073 (5)
2714 (4)
1529 (3)
3292 (4)
2783 (3)
1480 (4)
1991 (4)
1860 (2)
2346 (4)
1877 (3)
2303 (3)
1402 (3)
1436 (5)
914 (3)
501 (3)
973 (3)
635 (3)
226 (4)
224 (4)
ꢁ22 (2)
361 (3)
64 (1)
90 (1)
68 (1)
94 (1)
68 (1)
107 (2)
70 (1)
98 (1)
58 (1)
63 (1)
86 (1)
77 (1)
C7a
08
1997 (5)
3468 (3)
3802 (4)
4994 (3)
4621 (4)
2524 (4)
3875 (5)
4359 (5)
3490 (5)
2130 (5)
1642 (4)
ꢁ1033 (3)
708 (2)
921 (3)
2738 (3)
2529 (2)
1702 (3)
1514 (2)
551 (2)
81 (1)
81 (1)
65 (1)
66 (1)
61 (1)
68 (1)
88 (1)
97 (2)
85 (1)
83 (1)
76 (1)
ꢁ992 (3)
ꢁ1518 (2)
ꢁ1323 (2)
ꢁ2357 (3)
ꢁ789 (3)
ꢁ1180 (2)
174 (2)
C8a
N9
O2
N3
C3a
C4
1394 (3)
1414 (3)
ꢁ1079 (3)
ꢁ1444 (4)
ꢁ1496 (4)
ꢁ1192 (4)
ꢁ809 (3)
ꢁ759 (3)
C9a
Cll
3792 (3)
4198 (3)
5160 (3)
5746 (3)
5368 (3)
4399 (3)
C12
C13
C14
C15
C16
O4
C5a
N5
C6
968 (2)
1220 (3)
2319 (3)
C7
Ueq is defined as one third of the trace of the orthogonalized Uij tensor.

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4907
a range of ip doses of the test drug until at least three
points were established in the range of 10–90% seizure
protection or minimal observed neurotoxicity. From
the plot of this data, the respective ED50, TD50 values,
95% confidence intervals, slope of the regression line,
and the standard error of the slope were calculated by
means of a computer program written at NINCDS,
NIH.
and Stroke at the National Institutes of Health, Beth-
esda, MD USA.
This work was supported by Polish State Committee for
Scientific Research (Grant No. 4 P05F 014 18) and by
the Deutsche Forschungsgemeinschaft (GRK677).
C.E.M. is grateful for support by the Fonds der Chem-
ischen Industrie and the Bundesministerium fur Bildung
¨
und Forschung (BMBF), Germany.
8.3. X-ray structure analysis of 12 and 13
References and notes
The crystals of 12 and 13 were obtained by slow evapo-
ration of ethanol solutions. All measurements of the
crystals were performed on a Kuma4CCD j-axis diffrac-
tometer with graphite-monochromated MoKa radiation
at room temperature. The data were corrected for Lo-
rentz and polarization effects. No absorption correction
was applied. All crystallographic data and experimental
details are presented in Table 11. The structures were
solved by direct methods37 and refined using SHEL-
XL.39 The full-matrix least-squares refinement was
based on F2. The anisotropic temperature factors for
all non-H-atoms in structure 13 were applied. In the
crystal of 12 structural disorder was found and for all
non-H atoms in the molecule with structure occupation
factor (sof) equaling 0.7 anisotropic temperature factors
were used while for atoms from the molecule with sof of
0.3 only isotropic ones. The positions of all H-atoms
were found from the electron density Dq map and re-
fined in a raiding model with the isotropic displacement
parameters of 1.5 times the respective Ueq values for the
parent atoms. Atomic scattering factors were those as in
SHELXL.38 Atomic coordinates are gathered in Table
12. Crystallographic data (excluding structural factors)
for structures 12 and 13 reported in this paper have been
deposited with the Cambridge Crystallographic Data
Center. Copies of the data can be obtained free of
charge on application to CCDC, 12 Union Road, Cam-
bridge CB2 1EW, UK(Fax: +441223 336033; E-mail:
deposit@ccdc.cam.ac.uk).
1. Fredholm, B. B.; Ijzerman, A. P.; Jacobson, K. A.; Klotz,
K.-N.; Linden, J. Pharmacol. Rev. 2001, 53, 527.
2. Cunha, R. A. Neurochem. Int. 2001, 38, 107.
3. Hess, S. Expert Opin. Ther. Pat. 2001, 11, 1533.
4. Impagnatiello, F.; Bastia, E.; Ongini, E.; Monopoli, A.
Emerg. Ther. Targets 2000, 4, 635.
5. Ribeiro, J. A.; Cunha-Reis, D.; Lopes, L. V.; Coelho, J.
´
E.; Costenla, A. R.; Correia-de-Sa, P.; Cunha, R. A.; de
Mendonc¸a, A.; Sebastia˜o, A. M. Drug Dev. Res. 2001, 52,
337.
6. Muller, C. E. Drug Fut. 2000, 25, 1043.
¨
7. Ongini, E.; Schubert, P. Drug Dev. Res. 1998, 45, 387.
8. Moreau, J.-L.; Huber, G. Brain Res. Rev. 1999, 31, 65.
9. Ribeiro, J. A.; Sebastia˜o, A. M.; de Mendonc¸a, A. Prog.
Neurobiol. 2003, 68, 377.
10. De Sarro, G.; De Sarro, A.; Di Paola, E. D.; Bertorelli, R.
Eur. J. Pharmacol. 1999, 371, 137.
11. Dunwiddie, T. V. Adv. Neurol. 1999, 79, 1001.
´
12. Kiec-Kononowicz, K.; Drabczynska, A.; Pe˛kala, E.;
Michalak, B.; Muller, C. E.; Schumacher, B.; Karolak-
¨
´
Wojciechowska, J.; Duddeck, H.; Rockitt, S.; Wartchow,
R. Pure Appl. Chem. 2001, 73, 1411.
´
13. Drabczynska, A.; Schumacher, B.; Mu¨ller, C. E.; Karolak-
Wojciechowska, J.; Michalak, B.; Pe˛kala, E.; Kiec-Kon-
´
onowicz, K. Eur. J. Med. Chem. 2003, 38, 397.
14. Sauer, R.; Maurinsh, J.; Reith, U.; Fulle, F.; Klotz, K.-N.;
¨
Muller, C. E. J. Med. Chem. 2000, 43, 440.
¨
15. Eckstein, M.; Gorczyca, M.; Zejc, A. Acta Pharm. Yug.
1972, 11, 133.
´
16. De Martiis, F.; Botre, C.; Tottoli, F. Ann. Inst. Super
Sanita 1965, 1, 1.
17. Eckstein, M. Dissert. Pharm. 1962, 14, 425.
8.4. Computational procedures
´
´
18. Rockitt, S.; Duddeck, H.; Drabczynska, A.; Kiec-Kon-
onowicz, K. Eur. J. Org. Chem. 2000, 3489.
19. Bergmann, F.; Dikstein, S. J. Am. Chem. Soc. 1955, 77,
691.
The starting models of the molecules were based on
crystallographic data for 12 and 13 using the PCMOD.6
program.36 The geometries of the molecules were optim-
ized with MOPAC 6.0 using AM1 Hamiltonians in an
aqueous environment (dielectric constants equals
78.4).40
´
ˇ
20. Rybar, A.; Antos, K .Collect. Czech. Chem. Commun.
1970, 35, 1415.
21. Muller, C. E.; Maurinsh, J.; Sauer, R. Eur. J. Pharm. Sci.
¨
2000, 10, 259.
22. Klotz, K.-N.; Lohse, M. J.; Schwabe, U.; Cristalli, G.;
Vittori, S.; Grifantini, M. N–S Arch. Pharmacol. 1989,
340, 679.
23. Ji, X.-D.; Jacobson, K. A. Drug. Des. Discov. 1999, 16, 89.
24. Porter, R. J.; Cereghino, J. J.; Gladding, G. D.; Kupfer-
berg, H. J.; Scoville, B.; White, B. G. Clev. Clin. 1984, 51,
293.
The values of logP and logD (at pH=7.4) for the com-
pounds investigated were calculated by means of PAL-
LAS (version 1.2) program.34 Multiple Linear
Regression equations were computed by means of the
QSAR-PC:PAR program written by R. C. Coburn.37
25. Krall, R. L.; Penry, J. K.; White, B. G.; Kupferberg, H. J.;
Swinyard, E. A. Epilepsia 1978, 19, 409.
26. Stables, J. P.; Kupferberg, K. J. Molecular and Cellular
Targets for Antiepileptic Drugs; John Libbey:
London, 1998, pp 191–198.
27. Mulzac, D.; Scott, K. R. Epilepsia 1993, 34, 1141.
28. Muller, C. E.; Thorand, M.; Qurishi, R.; Dieckmann, M.;
¨
Acknowledgements
The authors are grateful to Professor James Stables for
providing the biological data through the Antiepileptic
Drug Development (ADD) Program of the National
Institute of Neurological and Communicative Disorders
Jacobson, K. A.; Padgett, W. L.; Daly, J. W. J. Med.
Chem. 2002, 45, 3440.

´
A. Drabczynska et al. / Bioorg. Med. Chem. 12 (2004) 4895–4908
4908
29. Hess, S.; Muller, C. E.; Frobenius, W.; Reith, U.; Klotz,
¨
K.-N.; Eger, K. J. Med. Chem. 2000, 43, 4636.
´
30. Hayallah, A. M.; Sandoval-Ramırez, J.; Reith, U.; Scho-
bert, U.; Preiss, B.; Schumacher, B.; Daly, J. W.; Muller,
¨
Wojciechowska, J. Abstracts of Papers, The Third Multi-
disciplinary Conference on Drug Research, Piła, Poland,
2002; P103.
36. PCMODEL 6 Molecular Modelling Software for the IBM
PC, Serena Software, Bloomington, 1996.
C. M. J. Med. Chem. 2002, 45, 1500.
31. Klotz, K.-N.; Hessling, J.; Hegler, J.; Owman, C.; Kull,
B.; Fredholm, B. B.; Lohse, M. J. N–S Arch. Pharmacol.
1998, 357, 1.
32. Ji, X.-D.; Jacobson, K. A. Drug Des. Discov. 1999, 16, 217.
33. Muller, C. E.; Dieckmann, M.; Thorand, M.; Ozola, V.
¨
Bioorg. Med. Chem. Lett. 2002, 12, 501.
34. PALLAS for Windows 1.2, CompuDrug Chemistry Ltd,
1995.
37. QSAR-PC program, Coburn, R. A., BIOSOFT, Cam-
bridge, UK.
38. Sheldrick, G. M. SHELTXTL PCMT; Siemens Analytical
X-Ray Instruments Inc.: Madison, Wisconsin, USA, 1990.
39. Sheldrick, G. M. SHELXL-93. A FORTRAN-77 Program
for the Refinement of Crystal Structures from Diffraction
Data; University of Go¨ttingen: Germany, 1993.
40. MOPAC program version 6, QCPE No. 455, Department
of Chemistry, Indiana University, Bloomington, IN
47405, 1990.
´
35. Drabczynska, A.; Kiec-Kononowicz, K.; Pe˛kala, E.;
Michalak, B.; Muller, C. E.; Schumacher, B.; Karolak-
´
¨