Structure-activity studies of 5-substituted pyridopyrimidines as adenosine kinase inhibitors

Article abstract of DOI:10.1016/S0960-894X(00)00602-8

The synthesis and SAR of a novel series of non-nucleoside pyridopyrimidine inhibitors of the enzyme adenosine kinase (AK) are described. It was found that pyridopyrimidines with a broad range of medium and large non-polar substituents at the 5-position potently inhibited AK activity. A narrower range of analogues was capable of potently inhibiting adenosine phosphorylation in intact cells indicating an enhanced ability of these analogues to penetrate cell membranes. Potent AK inhibitors were found to effectively reduce nociception in animal models of thermal hyperalgesia and persistent pain. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0960-894X(00)00602-8

Bioorganic & Medicinal Chemistry Letters 11 (2001) 83±86
Structure±Activity Studies of 5-Substituted Pyridopyrimidines
as Adenosine Kinase Inhibitors
Marlon Cowart,* Chih-Hung Lee, Gregory A. Gfesser, Erol K. Bayburt,
Shripad S. Bhagwat,y Andrew O. Stewart, Haixia Yu, Kathy L. Kohlhaas,
Steve McGaraughty, Carol T. Wismer, Joseph Mikusa, Chang Zhu,
Karen M. Alexander, Michael F. Jarvis and Elizabeth A. Kowaluk
Neurological and Urological Diseases Research, Abbott Laboratories, 100 Abbott Park Road, Abbott Park, IL 60064, USA
Received 12 September 2000; accepted 24 October 2000
AbstractÐThe synthesis and SAR of a novel series of non-nucleoside pyridopyrimidine inhibitors of the enzyme adenosine kinase
(AK) are described. It was found that pyridopyrimidines with a broad range of medium and large non-polar substituents at the 5-
position potently inhibited AK activity. A narrower range of analogues was capable of potently inhibiting adenosine phosphor-
ylation in intact cells indicating an enhanced ability of these analogues to penetrate cell membranes. Potent AK inhibitors were
found to e€ectively reduce nociception in animal models of thermal hyperalgesia and persistent pain. # 2000 Elsevier Science Ltd.
All rights reserved.
Introduction
Our interest has been to prepare AK inhibitors for use
as a novel class of analgesics, free of the side e€ects of
more traditional analgesics such as opiates.
Endogenously produced adenosine (ADO) serves a
number of roles in the body, but it is especially impor-
tant as an extracellular messenger where it acts at spe-
ci®c receptors on the cell surface to modulate neuronal
activity and in¯ammation.1 In spite of extensive e€ort, the
direct pharmacological modulation of adenosine recep-
tors with agonists has not yielded useful drug candidates
for human use, due to the prevalence of mechanism based
side e€ects (prominently hemodynamic e€ects). A ratio-
nale for a therapeutic approach targeting an indirect
modulation of ADO receptors has been proposed as
providing `site and event selectivity', with an enhanced
therapeutic window.2 Damaged tissues produce elevated
levels of ADO, and inhibition of the metabolism of
endogenously produced ADO may selectively amplify
its local action in comparison with undesired systemic
actions.3 AK plays a major regulatory role in metabo-
lizing and inactivating ADO by rephosphorylation to
AMP. ADO and ADO receptor agonists have shown
analgesic actions in both clinical and preclinical models.3
The substrate ADO 1 is highly polar and rapidly meta-
bolized, but has nonetheless been used as the starting
point for the design of AK inhibitors through rational
drug design (Scheme 1).4 However, we have more
recently described a novel series of pyridopyrimidine
AK inhibitors 2 developed by optimization of the high
throughput screening hit 3 (IC50 400 nM).5 We antici-
pated improved membrane penetration and greater
metabolic stability for compounds based on 2, since these
are more lipophilic and lack the hydroxyl groups seen in
nucleoside-like inhibitors. In this report, we describe large
gains in potency upon optimization of the R5 position in
2, in which the best compounds exhibited up to a 2000-
fold boost in potency, in comparison with the unsub-
stituted analogues.
Chemistry
The synthesis of 5-aryl pyridopyrimidines6 was carried
out as has been described, where aryl aldehydes 4 were
condensed with aryl ketones 5 in the presence of NH4OAc
in ethanol or benzene to produce the amino cyano pyridine
intermediates 6 (R5=Ar, Scheme 2). For compounds 6,
*Corresponding author. Tel.: +1-847-938-8170; fax: +1-847-937-
9195; e-mail: marlon.d.cowart@abbott.com
yPresent address: Celgene Corporation, Signal Research Division,
5555 Oberlin Drive, San Diego, CA 92121, USA.
0960-894X/01/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00602-8

84
M. Cowart et al. / Bioorg. Med. Chem. Lett. 11 (2001) 83±86
only successful7 in eliminating the formation of the side
product 12 for compounds where R5 was alkyl, but was
found to produce cleaner products and higher yields
than the standard methodology for all compounds 2.
An alternative method for the formation of 2 where
R5=alkyl was found to be e€ective where X=CH. In
this case, 10 was converted ®rst to an ylide by reaction
with Ph3P, then treatment with NaOH, followed by
reaction with aliphatic aldehydes 4 to produce the chal-
cone 11. Cyclization with CH2(CN)2 gave 6, which was
converted to the target 2 as described above.
Scheme 1. Structures of substrate adenosine 1, pyridopyrimdine AK
inhibitors 2, and screening lead structure 3.
these were subsequently cyclized by heating in formamide
to give target 2.
However these conditions failed on two counts when
attempting to prepare compounds 2 containing an alkyl
group at R5. Firstly, under the standard literature con-
ditions, heating aliphatic aldehydes bearing an a-proton
with NH4OAc in ethanol or benzene gave only aldol-
derived self condensation products. However, it was
found that aliphatic aldehydes 4 could be precondensed
with malononitrile to produce dicyanoethylenes 7, which
after isolation, were then condensed with ketones 5 and
NH4OAc to give high yields of 6. It was found that 1,2-
dichloroethane was a much superior (and in some cases
essential) solvent to benzene or alcohol for this con-
densation.
Biological Results and Discussion
The importance of the R5 group in modulating cytosolic
AK inhibition can be readily seen in Table 1.8 The trend
was for AK inhibitory potency to rapidly increase as the
size of the substituent at R5 was increased from H
(13=733nM and 27=562nM ) to methyl (14=47 nM),
to cyclopropyl (15=55 nM), to butyl (16=11 nM), to
phenyl (18=7nM), to cyclohexyl (17=8nM, 39=3nM)
and 2-bromobenzyl (43=0.17 nM). Variation of the
substitution of the aryl groups led generally to good
potency, with a trend for the compounds with more
polar groups to have reduced activity (compare 15±22 vs
23±26 and 40 vs 39). This pattern of activity suggests that
the AK possesses a binding pocket that prefers pyr-
idopyrimidines of structure 2 to have medium to large R5
substituents with lipophilic properties. When these condi-
tions are met, compounds were found to consistently
possess potency in the low nanomolar range.
The conversion of 6 (where R5=alkyl) to 2 under the
standard conditions reported in the literature (heating
with formamide or formamidine) was not generally suc-
cessful, owing to the unexpected formation of the novel
tricyclic byproducts 12. A more ecient procedure for
this transformation was a three step process where 6 was
heated with HC(OEt)3 to produce an iminoether 8, which
was then stirred with NH3 to give an amidine 9, which was
then thermally cyclized to produce 2. This process was not
In order for a compound to display in vivo activity in
pain models (Table 2), potent inhibitory activity at
cytosolic AK is necessary. However, a very important
Scheme 2. Reagents: (a) CH2(CN)2, NH4OAc, benzene, 80 ꢀC; (b) HCONH2 170 ꢀC; (c) HC(OEt)3, PTSA, 140 ꢀC; (d) NH3; (e) 140 ꢀC, 1,2-
dichlorobenzene; (f) MgO, CH2(CN)2, CH2Cl2; (g) CH2(CN)2, NH4OAc, CH2ClCH2Cl, 80 ꢀC; (h) (i) Ph3P, Et3N, CH3CN; (ii) NaOH; (iii) R5CHO,
PhCH3; (i) (Z=Cl ! Z=NR002) HNR002, DMSO, 100 ꢀC.

M. Cowart et al. / Bioorg. Med. Chem. Lett. 11 (2001) 83±86
Table 1. In vitro inhibition of AK in cytosolic and intact assaysb
85
Cytosolic AKa inhibition
IC50 (nM)
Intact cell AKa inhibition
IC50 (nM)
Compound
X
Z
R5 substituent
Mean
SEM
Mean
SEM
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
N
N
N
N
N
N
N
N
N
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
H
CH3
Cyclopropyl
Butyl
Cyclohexyl
Ph
Ph(3-Cl)
Ph(3-Br)
Ph(3-I)
Ph(3-CF3)
Ph(3-OCH3)
Ph(3-OH)
Ph(3-CONH2)
Ph(3-CN)
H
Ph(2-Br)
Ph(3-Br)
Ph(4-Br)
Ph(3-Cl)
733
47
55
11
8
7
15
5
17
36
85
44
77
35
562
3
2
30
9
62
13
10
5
38
9
58
30
13
10
1
2
10
2
14
25
78
28
28
26
280
1
0.5
10
2
16
10
1
1
18
2
1
1
6
0.07
5
0.02
0.18
>10,000
>1000
825
550
470
470
78
170
120
225
400
353
850
325
>1000
675
50
150
260
230
58
18
58
30
35
140
350
210
35
100
8
115
7
112
210
350
3
94
30
16
4
160
14
90
20
18
430
96
Pyridin-3-yl
Thiophen-3-yl
Furan-2-yl
315
550
488
130
378
115
80
Ph(2,3-dichloro)
Ph(2,5-dichloro)
Pentan-3-yl
Cyclopentyl
Cyclohexyl
Tetrahydro pyran-4-yl
Cyclohexyl cis (3,5-dimethyl)
C(CH3)2CH=CH
CH2Ph(2-Br)
CH(CH3)Ph(2-Br) (racemic)
N
N
N
N
N
N
N
N
2
3
57
17
0.23
21
0.17
0.25
800
102
304
33
N
30
aInhibition of AK and ADO phosphorylation assays were carried out as described by Jarvis et al.7
bValues represent means Æ SEM determined from multiple (ꢁ3) determinations.
additional property should also include the ability to
cross cell membranes, since AK is an intracellular target.
>100 related analogues) only 20 displayed any in vivo
activity in the rat thermal hyperalgesia assay, a model of
in¯ammatory pain.9 However, 20 was not active in pain
models following oral administration. In contrast,
among 27±44 several compounds were active in both the
thermal hyperalgesia model and in the formalin test10
(Table 2). Compounds with more potent AK cellular
inhibition activity tended to exhibit better analgesic
activity. Some analogues (especially 29) also retained
high activity following oral administration. Another
reason for the generally enhanced in vivo and whole cell
AK activity seen in the 7-pyridyl class of analogues may
be in part due to their observed higher solubility than
the 7-phenyl analogues, thereby allowing them to reach
higher concentrations more quickly in cells. These
potent analogues have been used as a starting point for
the design of further modi®ed analogues optimized for
potent oral analgesic activity, and for selectivity versus
the side e€ects that some analogues showed in reducing
spontaneous locomotor activity.11
In both these respects, compounds with a 7-(morpholi-
no)pyridyl group at R7 (X=N, 27±44) generally pos-
sessed improved properties, as well as ecacy in animal
pain models,8 when compared with compounds with
a 7-phenyl group at R7 (X=CH, 13±26). Analogues
(13±26) with a 7-phenyl group at R7 (X=CH) were
often highly insoluble in aqueous media, whereas 27±44
were more soluble, and especially so when formulated
as hydrochloride salts. As re¯ected in the AK intact cell
potencies, the latter class of compounds with R7=
7-(morpholino)pyridyl group also contained the ana-
logues which were most potent in the ability to pene-
trate intact cells and inhibit the AK enzyme (29, 43, and
44). As was expected, since activity in the intact cell AK
assay was anticipated to better predict in vivo activity,
members of this class were also more generally eca-
cious in animal pain models. Among 13±26 (and among

86
M. Cowart et al. / Bioorg. Med. Chem. Lett. 11 (2001) 83±86
Table 2. Activity of selected compounds in in vivo (rat) pain models
References and Notes
Compound
Rat formalin
nociceptiona
Rat carrageenan
hyperalgesiaa
1. (a) Jacobson, K. A.; van Galen, P. J. M.; Williams, M. J. J.
Med. Chem. 1992, 35, 407. (b) Williams, M.; Kowaluk, E. A.;
Arneric, S. P. J. Med. Chem. 1999, 42, 1481. (c) Kowaluk, E.
A.; Bhagwat, S. S.; Jarvis, M. F. Curr. Pharm. Des. 1998, 4, 403.
2. (a) Marangos, P. Med. Hypotheses 1990, 32, 45. (b) Mul-
lane K.; Young M. Drug Dev. Res. 1993 23, 336.
ED50b (mmol/kg)
ED50 (mmol/kg)
b
20
28
29
30
31
32
35
37
38
39
40
41
43
44
0.7
10
25 (60)
30
30
30
0.6 (5)
n.t.
20
>3 0
5
30
10d
10 (5)
2 (>30)
3 0
3
10
3. Kowaluk, E. A.; Jarvis, M. F. Exp. Opin. Invest. Drug
2000, 9, 551.
10
4. (a) Erion, M. D.; Ugarkar, B. G.; DaRe, J.; Castellino, A.
J.; Fujitaki, J. M.; Dixon, R.; Appleman, J. R.; Wiesner, J. B.
Nucleosides Nucleotides 1997, 16, 1013. (b) Cowart, M.; Ben-
nett, M. J.; Kerwin, J. F., Jr. J. Org. Chem. 1999, 64, 2240.
5. Lee, C.-H.; Jiang, M.; Cowart, M.; Gfesser, G.; Perner, R.;
Deecher, D.; Hale, C.; Kim, K. H.; Gu, U. G.; Stewart, A.;
Bhagwat, S. S.; Williams, M.; Jarvis, M. J.; Kowaluk, E. A.
J. Med. Chem., submitted for publication.
n.t.c
n.t.
n.t.
n.t.
20
30
>3 0
30
10 (100)
6. Saito, J. S.; Smith, J. J.; Roe, R. P. J. Am. Chem. Soc. 1968,
90, 8234.
aValues represent compounds administered by ip injection, except for
values in parentheses, which represent compounds administered
orally.
7. With the standard formamide conditions, separation of
products 2 from tricycles 12 was tedious, requiring very careful
chromatography, then fractional crystallization. Of six com-
pounds, average puri®ed yields of 2 were 16%, 12 were 40%.
With the three step amidine cyclization procedure, however,
puri®ed yields of 2 averaged 50%, with no formation of 12.
8. (a) Jarvis, M. F.; Yu, H.; Kohlhaas, K.; Alexander, K.;
Lee, C.-H.; Jaing, M.; Bhagwat, S. S.; Williams, M.; Kowaluk
E. A. J. Pharmacol. Exp. Ther. 2000, 295, in press. (b) Lynch,
J. J., III; Alexander, K. M.; Jarvis, M. F.; Kowaluk, E. A.
Neurosci. Lett. 1998, 252, 207.
bED50s reported for compounds for which a complete dose±response
relationship was obtained with at least six animals at each dose,
p<0.05.
cNot tested.
dCompound administered by subcutaneous injection.
In summary, the study of the SAR of a series of 5-sub-
stituted pyridopyrimidines has led to the discovery of a
family of highly potent AK inhibitors with in vivo
activity.
9. Kowaluk, E. A.; Mikusa, J.; Wismer, C.; Zhu, C. Z.;
Schweitzer, E.; Lynch, J.; Lee, C.-H.; Jiang, M.; Bhagwat, S.
S.; Gomtsyan, A.; McKie, J.; Cox, B. F.; Reinhart, G.; Wil-
liams, M.; Jarvis, M. F. J. Pharmacol. Exp. Ther. 2000, 295, in
press.
10. Poon, A.; Sawynok, J. Eur J. Phamacol. 1995, 286, 177.
11. Matulenko, M. A.; Lee, C.-H.; Frey, R. R.; Jiang, M.;
Cowart, M. D.; Perner, R. J.; Koenig, J. R.; Gfesser, G. A.;
Pratt, J.; Mao, Y.; McKie, J. A.; Gomtsyan, A.; Zheng, G. Z.;
Stewart, A. O.; Bhagwat, S. S.; Yu, H.; Kohlhaas, K. L.;
Alexander, K. M.; McGaraughty, S.; Wismer, C. T.; Mikusa,
J.; van Biesen, T.; Jarvis, M. F.; Kowaluk, E. A. Bioorg. Med.
Chem, submitted for publication.
Acknowledgements
We gratefully acknowledge Dr. Tim van Biesen for
assistance with the in vitro assay, and the repreparation
of some compounds by J. Robert Koenig, and Ernie
Paight.