Synthesis and biological evaluation of pteridine and pyrazolopyrimidine based adenosine kinase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2004.06.029

Three new approaches have been tested to modify existing pyridopyrimidine and alkynylpyrimidine classes of nonnucleoside adenosine kinase inhibitors 2 and 3. 4-Amino-substituted pteridines 8a-e were generally less active than corresponding 5- and 6-substituted pyridopyrimidines 2. Pyrazolopyrimidine 13c with IC₅₀=7.5nM was superior to its open chain alkynylpyrimidine analog 13g (IC₅₀=22nM) while pyrrolopyrimidines such as 17a were inactive.

Full text of DOI:10.1016/j.bmcl.2004.06.029

Bioorganic & Medicinal Chemistry Letters 14 (2004) 4165–4168
Synthesis and biological evaluation of pteridine and
pyrazolopyrimidine based adenosine kinase inhibitors
Arthur Gomtsyan,* Stanley Didomenico, Chih-Hung Lee, Andrew O. Stewart,
Shripad S. Bhagwat, Elizabeth A. Kowaluk and Michael F. Jarvis
Neuroscience Research, Global Pharmaceutical Research and Development, Abbott Laboratories, 100 Abbott Park Road,
Abbott Park, IL 60064, USA
Received 19 February 2004; accepted 9 June 2004
Available online
Abstract—Three new approaches have been tested to modify existing pyridopyrimidine and alkynylpyrimidine classes of nonnu-
cleoside adenosine kinase inhibitors 2 and 3. 4-Amino-substituted pteridines 8a–e were generally less active than corresponding
5- and 6-substituted pyridopyrimidines 2. Pyrazolopyrimidine 13c with IC50 ¼ 7.5 nM was superior to its open chain alkynylpyr-
imidine analog 13g (IC50 ¼ 22 nM) while pyrrolopyrimidines such as 17a were inactive.
Ó 2004 Elsevier Ltd. All rights reserved.
Adenosine (ADO) is an extracellular signaling agent
within the central and peripheral nervous system.1 It is
released from cells in response to adverse conditions, for
example, tissue trauma, pain, seizures, ischemia. Once
released, ADO induces protective pharmacological
responses upon interaction with specific A1, A2A, A2B,
and A3 G-protein coupled receptors of the P1 family.
However, extremely short half-life of ADO in physio-
logical fluids restricts the protective action of ADO2 to
the tissues and cellular sites where it is released. One
of the approaches of increasing the concentration of
endogenous ADO and potentiation of its action can be
inhibition of ADO metabolizing enzymes. Adenosine
kinase (AK) is a key intracellular enzyme regulating
intra- and extracellular ADO concentrations.3 AK
inhibitors, therefore, may have therapeutic potential as
analgesic and anti-inflammatory agents.
Most of the known early AK inhibitors, for example, 50-
deoxy-5-iodotubercidin4 1 (Fig. 1) and several others
reported recently5 have close structural resemblance to
the natural ligand ADO. AK inhibitors of this structural
class of compounds are commonly referred to as
nucleoside-like AK inhibitors. Our group recently
reported two classes of nonnucleoside AK inhibitors––
pyridopyrimidines 26 and alkynylpyrimidines 37 (Fig. 1).
One of the common motifs that both nucleoside and
nonnucleoside types of AK inhibitors share is the pres-
ence of the lipophilic groups such as I in 1, aryl and alkyl
in 2, and XAr in 3. These groups have been shown to be
key structural elements responsible for filling a putative
hydrophobic pocket of the protein and ensuring effective
binding. In this report, we discuss alternative ap-
proaches to the design of the AK inhibitors that resulted
in structures with different points of attachment of the
NH2 R1
NH2
5 X
NH2
N
I
4
5
6
R2
N
N
Ar
O
H3C
HO
5'-Deoxy-5-Iodotubercidin 1
N
N
N
N
N
N
N
OH
O
N
N
O
3
2
Figure 1. Structures of nucleoside- and nonnucleoside type AK inhibitors.
* Corresponding author. Tel.: +1-847-935-4214; fax: +1-847-937-9195; e-mail: arthur.r.gomtsyan@abbott.com
Present address: Celgene, Signal Research Division, 5555 Oberlin Drive, San Diego, CA 92130, USA.
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.06.029

4166
A. Gomtsyan et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4165–4168
lipophilic group.8 This in turn led to a more diverse set
of AK inhibitors and expanded the scope of SAR in
both classes of nonnucleoside series of AK inhibitors.
In the alkynylpyrimidine series of AK inhibitors, the
reported structures 3 (Fig. 1) contained lipophilic frag-
ment at the 5-position.7 To determine the effect of the
conformational restriction of that lipophilic group in the
known AK inhibitors with an alkynylpyrimidine core,
we designed and synthesized pyrazolopyrimidines 13
(Scheme 2). Deprotonation of 4,6-dichloropyrimidine 9
at the 5-position with LDA followed by the reaction
with different benzaldehydes provided alcohols 10
(yields 45–65%), which were converted to the ketones 11
by CrO3 in acetone with generally >90% yield. Sono-
gashira coupling of 11 with different arylacetylenes in
MeCN–Et3N in the presence of Pd(PPh3)2Cl2 and CuI
gave monoalkynyl pyrimidines 12 with 36–75% yield.
Finally, 12 were reacted with different hydrazines in
EtOH or dichloromethane to give the desired pyrazol-
opyrimidines 13a–f.
Our objective in the pyridopyrimidine series of AK
inhibitors was investigation of SAR at the 4-position.
Modifications at two of the other most relevant sites, the
5- and 6-positions of the pyridopyrimidine core, were
reported5 with the 5-position being the optimal point
of attachment for the lipophilic fragment. In order to
investigate the SAR at the 4-position, we chose to re-
place the pyridopyrimidine core by its surrogate pteri-
dine core. This strategy was based on the earlier ‘SAR
by NMR’ studies9 that had shown that the substitution
at the 5-position with any moiety, including hydrogen,
could potentially cause steric hindrance for any new
group introduced in the vicinity of that site, which in-
6a;10
cludes 4-position. A three-step procedure
of acyl-
ation of dimethyl malonate by commercially available
6-chloro-nicotinoyl chloride 4 (Scheme 1), decarboxyl-
ation and replacement of the chloro group by the mor-
pholino group resulted in acetopyridine 5 in 76% overall
yield. This compound was oxidized with the selenium
dioxide to the ketoaldehyde 6 followed by condensation
with 4,5,6-triaminopyrimidine to provide a regiospecif-
ically11 7-substituted key precursor 7 with 67% yield for
two steps. Target molecules 8a–e with the substitution at
the 4-position of the pteridine core were prepared by
treatment of 7 with a variety of arylalkylamines in
DMSO in the presence of AcOH at 120 °C for 10–12 h.
Yields were generally low and variable (7–54%). Longer
reaction times resulted in even lower yields.
The structure of the pyrrolopyrimidine 17a (Scheme 3)
can be envisaged by closing a ring between the nitrogen
of 5-amino group and a carbon on the 6-alkynyl group
in the alkynylpyrimidine class of compounds (e.g., 3,
where X ¼ NH, Fig. 1) to form bicyclic core. Iodination
and benzylation of 5-amino-4,6-dichloropyrimidine 14
gave 157 in 75% yield. The latter compound was then
subjected to ammonolysis in EtOH at 110 °C in a sealed
tube for 24 h to obtain 16 in 72% yield. Reaction of
16 with 4-(5-ethynyl-pyridin-2-yl)-morpholine6 under
Sonogashira conditions provided the pyrrolopyrimidine
17a in 67% yield in a two-step in situ process of the
cross-coupling followed by the intramolecular hydro-
amination. This sequential reaction represents the
O
O
O
a
b
H
c
Cl
O
N
N
N
N
Cl
N
O
O
4
5
6
R
n
NH2
HN
N
N
d
N
N
N
a R=H, n=0
b R=Ph, n=2
c R=Ph, n=3
d R=Ph, n=4
N
N
N
N
N
N
N
e R=3-indolyl, n=2
O
O
7
8
Scheme 1. Reagents and conditions: (a) (1) dimethyl malonate, MgCl2, Et3N; (2) DMSO–H2O, 150 °C, 80% for two steps; (3) morpholine, EtOH,
95%; (b) (1) SeO2, 1,4-dioxane–H2O; (c) 4,5,6-triaminopyrimidine, BaCl2, 67% for two steps; (d) arylalkyl amine, DMSO, AcOH, 120 °C.
Cl
Cl
N
OH
Cl
N
O
R2
R2
a
b
c
N
N
N
N
9
Cl
Cl
Cl
10
11
R1
N
Cl
O
N
N
N
R2
a R1=R2=H, R3=NMe2
R2
N
d
b R1=H, R2=Br,R3=NMe2
c R1=Me, R2=Br, R3=NMe2
d R1=H, R2=Br, R3=morpholino
e R1=Me. R2=Br, R3=morpholino
f R1=Bn, R2=Br, R3=morpholino
N
13
12
N
R3
N
R3
Scheme 2. Reagents and conditions: (a) LDA, )78 °C, THF, then benzaldehyde; (b) CrO3, acetone, 2 h, rt; (c) 4-(5-ethynyl-pyridin-2-yl)-morpholine,
Pd(PPh3)2Cl2, CuI, MeCN–Et3N (1:1), 60 °C, 6 h; (d) RNHNH2, CH2Cl2 (when R ¼ Me), EtOH (when R ¼ H, Bn).

A. Gomtsyan et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4165–4168
4167
Cl
NH2
N
I
H
N
H
N
NH2
Cl
ref.3
a
N
N
N
N
I
N
I
16
14
15
NH2
N
N
b
N
N
N
17a
Scheme 3. Reagents and conditions: (a) NH3, EtOH, 24 h, 110 °C, sealed tube, 72%; (b) 4-(5-ethynyl-pyridin-2-yl)-morpholine, Pd(PPh3)2Cl2, CuI,
MeCN–Et3N (1:1), rt, 1.5 h, 67%.
application of the Larock’s indole synthesis procedure12
to the heteroaromatic systems such as the pyrimidine
16.
Table 1. In vitro activity of AK inhibitors in cytosolic assay
Compounds
AK inhibition
IC50 (nM)a
R
The results of in vitro inhibition of the cytosolic AK13
by these novel compounds are summarized in Table 1.
Substitution at the 4-position in the pteridine 8a resulted
in compounds 8b–e that exhibited 1.5–6.5-fold increase
in AK inhibitory potency, in comparison to the unsub-
stituted compound 8a. The length of the alkyl group
linking the lipophilic aryl group with the 4-amino group
of the pteridine core does not seem to significantly alter
the AK inhibitory activity as the largest difference in a
series of phenethyl-, phenylpropyl-, and phenylbutyl-
substituted compounds 8b–d is only about 1.5-fold
(120 nM for 8b vs 72 nM for 8c). Although the in vitro
potency of the 4-substituted pteridines 8b–e is lower
than some of the best compounds in the 5-substituted
pyridopyrimidines of general structure 2,6 the present
data suggested the existence of acceptable tolerance
for the 4-substitution and offered the possibility of
approaching the putative hydrophobic pocket from the
4-position of the pyrimidine ring.
8a R ¼ H, n ¼ 0
8b R ¼ Ph, n ¼ 2
8c R ¼ Ph, n ¼ 3
8d R ¼ Ph, n ¼ 4
460 ꢀ 45
120 ꢀ 33
72 ꢀ 23
82 ꢀ 17
n
HN
N
N
N
N
8e R ¼ 3-indolyl, n ¼ 2 300 ꢀ 65
N
N
O
13a R1 ¼ R2 ¼ H,
R3 ¼ NMe2
100 34
75 15
13b R1 ¼ H, R2 ¼ Br,
R1
N
R3 ¼ NMe2
N
N
R2
13c R1 ¼ Me, R2 ¼ Br,
R3 ¼ NMe2
7.5 1.5
22 15
N
13d R1 ¼ H, R2 ¼ Br,
R3 ¼ morpholino
N
NR3
13e R1 ¼ Me, R2 ¼ Br,
R3 ¼ morpholino
82 13
>1000
13f R1 ¼ Bn, R2 ¼ Br,
R3 ¼ morpholino
Conformationally rigid analogs of the alkynylpyrim-
idine AK inhibitors of the general structure 3 were made
by cyclization of the benzylic carbon with 4-amino
nitrogen of the pyrimidine core. Inspection of the bio-
logical data on the resulting pyrazolopyrimidines 13a–f
revealed that the structural constraints were not detri-
mental and may be beneficial for AK inhibition. Thus,
compounds 13c and 13d have IC50 values of 7.5 and
22 nM that are better or equal to the corresponding
inhibition by open chain analog 13g (IC50 ¼ 22 nM).7
The benzyl group in the pyrazole fragment of 13f greatly
reduced the AK inhibitory potency (IC50 >1000 nM)
in comparison with the unsubstituted analog 13d
(IC50 ¼ 22 nM). This could be due to the competition
between the benzyl group and the 3-bromophenyl group
to occupy the hydrophobic binding site. It is more dif-
ficult to explain the trends in the pairs 13b and 13c
versus 13d and 13e. The structural difference between
these pairs is dimethylamino group versus morpholino
group as a part of 7-substituent. These amino groups
have been shown earlier to be similar in their effects on
AK inhibition.6a However, in the first pair, methylation
of the pyrazole ring increases the AK inhibition (7.5 nM
for 13c vs 75 nM for 13b), while in the second pair, the
same exercise decreases the inhibition (82 nM for 13e vs
22 nM for 13d).14
NH2
N
Br
N
13g
17a
17b
22 8
N
N
O
NH2
N
>1000
N
N
N
O
N
Me
N
NH2
N
N
30
4
N
N
O
a The data represent mean SEM from at least three determinations.
The closest ring-opened analog for the pyrrolopyrim-
idine 17a would be 6-alkynylpyrimidine 17b7 with IC50

4168
A. Gomtsyan et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4165–4168
1999, 64, 2240; (b) Erion, M. D.; Ugarkar, B. G.; DaRe,
J.; Castellino, A. J.; Fujitaki, J. M.; Dixon, R.; Appleman,
J. R.; Wiesner, J. B. Nucleos. Nucleot. 1997, 16, 1013; (c)
Ugarkar, B. G.; Castellino, A. J.; DaRe, J. S.; Ramirez-
Weinhouse, M.; Kopcho, J. J.; Rosengren, S.; Erion, M.
D. J. Med. Chem. 2003, 46, 4750, and references cited
therein.
value of 30 nM. N-Demethylation of this compound
with subsequent ring closing by intramolecular hydro-
amination of the triple bond yields the compound 17a
that is virtually inactive (IC50 >1000 nM). This rein-
forces the importance of a rigid linear trajectory of the
alkynyl substituent in the 6-position of pyrimidines like
3 for the optimum inhibition of the enzyme.15
6. (a) Lee, C.-H.; Jiang, M.; Cowart, M.; Gfesser, G.; Perner,
R.; Kim, K.; Gu, Y. G.; Williams, M.; Jarvis, M.;
Kowaluk, E.; Stewart, A.; Bhagwat, S. J. Med. Chem.
2001, 44, 2133; (b) Cowart, M.; Lee, C.-H.; Gfesser, G. A.;
Bayburt, E. K.; Bhagwat, S. S.; Stewart, A. O.; Yu, H.;
Kohlhaas, K. L.; McGaraughty, S.; Wismer, C. T.;
Mikusa, J.; Zhu, C.; Alexander, K. M.; Jarvis, M. F.;
Kowaluk, E. A. Bioorg. Med. Chem. Lett. 2001, 11, 83–86;
(c) Gfesser, G. A.; Bayburt, E. K.; Cowart, M.; Didome-
nico, S.; Gomtsyan, A.; Lee, C.-H.; Stewart, A. O.; Jarvis,
M. F.; Kowaluk, E. A.; Bhagwat, S. S. Eur. J. Med. Chem.
2003, 38, 245; (d) Perner, R. J.; Gu, Y.-G.; Lee, C.-H.;
Bayburt, E. K.; McKie, J.; Alexander, K. M.; Kohlhaas,
K. L.; Wismer, C. T.; Mikusa, J.; Jarvis, M. F.; Kowaluk,
E. A.; Bhagwat, S. S. J. Med. Chem. 2003, 46, 5249.
7. Gomtsyan, A.; Didomenico, S.; Lee, C.-H.; Matulenko,
M. A.; Kim, K.; Kowaluk, E. A.; Wismer, C. T.; Mikusa,
J.; Yu, H.; Kohlhaas, K.; Jarvis, M. F.; Bhagwat, S. S.
J. Med. Chem. 2002, 45, 3639.
8. For a review on nonnucleoside AK inhibitors, see:
Gomtsyan, A.; Lee, C.-H. Curr. Pharmaceut. Des. 2004,
10, 1093.
9. Hajduk, P. J.; Gomtsyan, A.; Didomenico, S.; Cowart,
M.; Bayburt, E. K.; Solomon, L.; Severin, J.; Smith, R.;
Walter, K.; Holzman, T. F.; Stewart, A.; McGaraughty,
S.; Jarvis, M. F.; Kowaluk, E. A.; Fesik, S. W. J. Med.
Chem. 2000, 43, 4781.
In summary, we have investigated three new approaches
of modifying existing pyridopyrimidine and alkynylpyr-
imidine AK inhibitors. It was shown that substitution at
the 4-position of the pteridine core, as a surrogate for the
pyridopyrimidine, resulted in compounds 8a–e which,
while potent, were generally inferior to the previously
reported best examples of the pyridopyrimidine class of
compounds having substitutions at 5- and 6-positions. In
the case of the alkynylpyrimidine AK inhibitors,
restriction of the flexibility of the lipophilic group at the
5-position by forming fused pyrazole ring between the
nitrogen of the 4-amino group and the benzylic carbon of
5-benzyl group can be beneficial as compounds 13c
(IC50 ¼ 7.5 nM) and 13d (IC50 ¼ 22 nM) displayed in-
creased or equal AK inhibition in comparison with ring-
opened analog 13g (IC50 ¼ 22 nM). In contrast, the pyr-
role ring formation between the nitrogen of the 5-sub-
stituent and the alkynyl carbon at the 6-position yielded
compound 17a that was significantly less active (IC50
>1000 nM) than corresponding ring-opened molecule
17b (IC50 ¼ 30 nM). This can potentially be explained by
the loss of specific binding of the morpholinopyridyl
substituent due to an incorrect steric direction.
10. Kuo, D. L. Tetrahedron 1992, 48, 9233.
11. Temple, C.; Laseter, A. G.; Rose, J. D.; Montgomery, J.
A. J. Heterocycl. Chem. 1970, 7, 1195.
12. (a) Larock, R. C.; Yum, E. K. J. Am. Chem. Soc. 1991,
113, 6689; (b) Larock, R. C.; Yum, E. K.; Refvik, M. D.
J. Org. Chem. 1998, 63, 7652.
Acknowledgements
We thank Dr. Tom Pagano for excellent NMR support
and Kathy Kohlhaas for assistance with the in vitro
data.
13. One of the reviewers offered to explain that discrepancy by
a limited overall size of the binding pocket, that is, the
morpholino and methyl groups make 13e larger and
consequently less fit to the binding pocket than morpho-
lino and hydrogen in 13d. Although this could be a
reasonable argument, there are many pyridopyrimidine
AK inhibitors with the 7-substituent much larger than
morpholino group, but still showing single or low double
digit potency, see: Zheng, G. Z.; Mao, Y.; Lee, Ch.-H.;
Pratt, J. R.; Koenig, R. J.; Perner, R.; Cowart, M.;
Gfesser, G.; McGaraughty, S.; Chu, K.; Zhu, C.; Yu, H.;
Kohlhaas, K.; Alexander, K.; Wismer, C.; Mikusa, J.;
Jarvis, M. F.; Kowaluk, E. A.; Stewart, A. O. Bioorg.
Med. Chem. Lett. 2003, 13, 3041.
References and notes
1. Ralevic, V.; Burnstock, G. Pharmacol. Rev. 1998, 50, 413;
Williams, M.; Jarvis, M. Biochem. Pharmacol. 2000, 59,
1173.
2. Arch, J. R. S.; Newsholme, E. A. Essays Biochem. 1978,
14, 82.
3. Moser, G. H.; Schrader, J.; Duessen, A. Am. J. Physiol.
1989, 25, C799.
4. Davies, L. P.; Jamieson, D. D.; Baird-Lambert, J. A.;
Kazlauskas, R. Biochem. Pharmacol. 1984, 33, 347.
5. For more nucleoside-type AK inhibitors, see: (a) Cowart,
M. C.; Bennett, M. J.; Kerwin, J. F., Jr. J. Org. Chem.
14. For the description of AK inhibition assay, see: Jarvis, M.
F.; Yu, H.; Kohlhaas, K.; Alexander, K.; Lee, C.-H.;
Jiang, M.; Bhagwat, S. S.; Williams, M.; Kowaluk, E. A.
J. Pharmacol. Exp. Ther. 2000, 295, 1165.
15. Unpublished data.