2804
R. J. Perner et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2803–2807
penetrate the cell membrane. Medicinal chemistry efforts
directed at optimizing the potency and improving
the membrane permeability of these AK inhibitors was
initiated. Herein, we report our investigation on the
development of 6,7-disubstituted 4-aminopyrido[2,3-d]pyri-
midines 2 as inhibitors of adenosine kinase.
lipophilic, groups. For example, the p-bromo analogue
10 had an IC50 as low as 20 nM while the p-methoxy
analogue 12 was 92 nM. Para substitution was preferred
over meta substitution as can be seen by comparing the
bromo derivatives 10 and 11 and methoxy derivatives 12
and 13, while 3,4-disubstitution was tolerated as evi-
denced by analogues 14 and 29,17 even when the 3-sub-
stituent was methoxy as in analogues 15 and 30. To
further the 6-position SAR, the aromatic ring was ex-
tended out from the pyridopyrimidine core by one to
three methylene units. It was found that the benzyl ana-
logue (compound 17, 114 nM) gave similar potencies as
the phenyl series while the phenethyl analogue 18
showed only modest potency (433 nM) and the phenyl-
propyl analogue 19 was greater than 2 lM. Finally, cer-
tain alkyl groups could also be placed at the 6-position
of the pyridopyrimidine core to give potent AK inhibi-
tors. For example, longer chain groups such as pentyl
(compound 20, 30 nM) were well tolerated while shorter
chains and branched alkyl groups such as isobutyl (com-
pound 21, 200 nM) gave analogues with modest
potency.
The synthesis of 6,7-disubstituted 4-aminopyrido[2,3-d]-
pyrimidines was accomplished using the route shown in
Scheme 1. An appropriately substituted acetylene 3 was
treated with catecholborane followed by Suzuki cou-
pling of the resulting boronic acid 5 with iodopyrimidine
6 to give the trans disubstituted alkene 7, typically in 50–
80% yield after purification. The desired products 2 were
then obtained via cyclization by aza-Cope rearrange-
ment of the intermediate formed by reaction of 7 with
an aryl aldehyde at high temperature in 1,2,4-trichloro-
benzene or diphenyl ether, typically in 15–25% yield.
Compound 6 was prepared by iodination of commer-
cially available 4,6-diaminopyrimidine (4),9 while the
acetylenes 3 were either obtained from commercial
sources or prepared according to various literature
procedures.16
A variety of groups were investigated at the 7-position.
In addition to the dimethylamino group, a number of
other substituted phenyl moieties were tried and, in gen-
eral, most gave analogues with AK potencies slightly
worse to much worse than the dimethylamino group,
and usually with much worse solubility. A variety of
other aromatic and heteroaromatic groups, some with
potential for improving solubility, were also studied.
A number of pyridine and pyrimidine analogues such
as 22, 23, and 24 showed modest potency while the
benzofuran analogue 25 and thiazole derivative 26 had
potencies near 1 lM. Of all the groups studied at the
7-position, the most striking was the thiophene moiety.
While the 3-thiophene group gave an analogue (com-
pound 27) that was 1.5 times more potent than the cor-
responding dimethylaminophenyl derivative, it was the
2-thiophene group that consistently yielded compounds
with very good AK inhibitory activity, often in the sin-
gle-digit nanomolar range. For example, thiophene ana-
logues with disubstituted phenyl (29), benzyl (31), and
alkyl (32) groups at the 6-position all had very good
potencies.
As was the case with the pteridine series, incorporation
of a substituent at the 6-position was found to be impor-
tant to the potency of the 6,7-disubstituted pyridopyrim-
idine compounds. Placement of an unsubstituted phenyl
group at C-6 yielded an analogue that was about 2-fold
more potent than in the unsubstituted case (cf. com-
pounds 16 vs 8, Table 1). Substituents on the phenyl ring
were also key to the potency of these analogues. While
not of the same magnitude as in the pteridine series
(IC50 = 25 nM vs 440 nM),9 the p-dimethylaminophenyl
analogue 9 was 4-fold more potent than 8, and addi-
tional increases in potency could be achieved by substi-
tuting the phenyl ring with other, particularly more
In stark contrast to the other two 4-aminopyridopyrimi-
dine series studied (i.e., 5,7-disubstituted and 5,6,7-tri-
substituted), it was found that all of the analogues
from the 6,7-disubstituted series of compounds appar-
ently had poor cell membrane permeability as evidenced
by the lack of activity in the intact cell assay. In fact, out
of more than 110 compounds synthesized the most po-
tent analogue had an IC50 of 225 nM (compound 14).
While they share similar calculated physical chemical
properties (solubility, clog P, etc), it is this characteristic
which makes the 6,7-disubstituted pyridopyrimidine ser-
ies distinctly different from the 5,7-disubstituted and
5,6,7-trisubstituted pyridopyrimidine series. The latter
two series, in particular the 5,7-disubstituted series, were
both able to produce analogues with intact cell potencies
less than 50 nM. Compounds from these two series,
which had single-digit nanomolar potency at the enzyme
Scheme 1. Reagents and conditions: (a) catecholborane, THF, reflux,
100%; (b) I2, K2CO3, DMF/H2O, 45 ꢁC, 65%; (c) (PPh3)4Pd,
(aq)Na2CO3, THF, reflux, 50–80%; (d) Ar7CHO, 1,2,4-trichloroben-
zene or diphenyl ether, 215–260 ꢁC, air, 15–25%.
Perner, Richard J.
