Pyridopyrimidine analogues as novel adenosine kinase inhibitors

Article abstract of DOI:10.1016/S0960-894X(01)00375-4

A novel series of pyridopyrimidine analogues 9 was identified as potent adenosine kinase inhibitors based on the SAR and computational studies. Substitution of the C7 position of the pyridopyrimidino core with C2′ substituted pyridino moiety increased the in vivo potency and enhanced oral bioavailability of these adenosine kinase inhibitors.

Full text of DOI:10.1016/S0960-894X(01)00375-4

Bioorganic & Medicinal Chemistry Letters 11 (2001) 2071–2074
Pyridopyrimidine Analogues as Novel Adenosine Kinase Inhibitors
Guo Zhu Zheng,a,* Chih-Hung Lee,a John K. Pratt,a Richard J. Perner,a Mei Qun Jiang,a
Arthur Gomtsyan,a MarkA. Matulenko, a Yui Mao,a John R. Koenig,a Ki H. Kim,b
Steve Muchmore,b Haixia Yu,a Kathy Kohlhaas,a Karen M. Alexander,a
Steve McGaraughty,a Katharine L. Chu,a Carol T. Wismer,a Joseph Mikusa,a
Michael F. Jarvis,a Kennan Marsh,c Elizabeth A. Kowaluk,a
Shripad S. Bhagwata,y and Andrew O. Stewarta
aAbbott Laboratories, Dept. 4PM, Bldg. AP9A LL, 100 Abbott Park Road, Abbott Park, IL 60064-6115, USA
bAbbott Laboratories, Dept. 4EK, Bldg. AP9, 100 Abbott Park Road, Abbott Park, IL 60064, USA
cAbbott Laboratories, Dept. 46Y, Bldg. AP10, 100 Abbott Park Road, Abbott Park, IL 60064, USA
Received 15 February 2001; accepted 11 May 2001
Abstract—A novel series of pyridopyrimidine analogues 9 was identified as potent adenosine kinase inhibitors based on the SAR
and computational studies. Substitution of the C7 position of the pyridopyrimidino core with C20 substituted pyridino moiety
increased the in vivo potency and enhanced oral bioavailability of these adenosine kinase inhibitors. # 2001 Elsevier Science Ltd.
All rights reserved.
Adenosine (ADO) is a key homeostatic inhibitory neu-
romodulator that contributes to endogenous antinoci-
ceptive and antiinflammatory responses following tissue
injury. Cellular excitability or tissue trauma increases
the local tissue levels of extracellular ADO.1 Extra-
cellular ADO acts on specific cell-surface P1 purinergic
receptors (A1, A2A, A2B, and A3) proximal to its site of
release to exert its homeostatic effects. Due to its extre-
mely short half-life (measured in seconds) in extra-
cellular fluids, the endogenous actions of ADO are
localized to the tissue and cellular site where it is
released. By blocking the intracellular enzymatic phos-
phorylation of ADO to adenosine monophosphate by
adenosine kinase (AK), selective adenosine kinase inhi-
bitors increase the extracellular concentration of endo-
genously released adenosine,2 and therefore enhance its
neuroprotective, antinociceptive and antiinflammatory
action.3ꢀ5
originally discovered by HTS (high-throughput screen-
ing). Of many potent analogues of this type, compound
9e (ABT-702) was an optimized analogue with good in
vitro and in vivo potencies (in vitro: AK (enzyme):
IC50=1.3 (ꢁ0.58) nM; AK (intact cell): IC50=43.3
(ꢁ5.7) nM; in vivo: inflammatory hyperalgesia (po):
ED50=0.7 mmol/kg).6 However, its relatively short half-
life and moderate oral bioavailability (t1=2=0.9 h;
F=22.6%) needed to be improved. Our goal was to
identify analogues with improved half-life and oral
bioavailability in our pursuit of a drug development
candidate. At the later stages of our lead optimization
period, our efforts were mainly focused on several dif-
ferent pyrido-pyrimidine analogues with substituents at
C5 and C7 positions. There were several subtypes of C7
substituents.7 In this letter, we describe our approaches
to one of the subtypes, 7-(50-pyridinyl) analogues, which
lead to a potent analogue 9h (in vitro: AK (enzyme):
IC50=8.0 (ꢁ2.0) nM; AK (intact cell): IC50=70.9
(ꢁ19) nM; in vivo: inflammatory hyperalgesia (po):
ED50=1 mmol/kg), with better half-life and higher oral
bioavailability (t1=2=3.6 h; F=46.2%) compared to our
lead compound 9e6 (Fig. 1).
Pyridopyrimidine analogues are a novel class of adeno-
sine kinase inhibitors.6 This class of analogues was
*Corresponding author. Tel.: +1-847-935-7513; fax: +1-847-938-
0072; e-mail: guozhu.zheng@abbott.com
yPresent address: Celgene Corporation, Signal Research Division,
5555 Oberlin Drive, San Diego, CA 92121, USA.
Our earlier observations on this subtype of derivatives
indicated that these analogues have very similar adenosine
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00375-4

2072
G. Z. Zheng et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2071–2074
kinase inhibitory potency, regardless of the kind of
substituents at C20 of the pyridinyl ring moiety of the
C7 pyridopyrimidino core (Fig. 1). The implication of
this observation is that the C20 part of the molecule does
not bind to a specific site of the enzyme. In an effort to
find out the docking mode of this subtype of analogues
within adenosine kinase, we have studied a known
crystal structure of adenosine kinase8 and did molecular
modeling with our analogues. These studies suggested
that the terminal of the C7 substituent is likely projected
outside the binding pocket of the enzyme, which may
explain why the terminals of the C7 substituent did not
significantly alter AK inhibitory activity of these analo-
gues (Fig. 2). Furthermore, it has been noticed that the
majority of the metabolic reactions occur at C20 sub-
stituent. These studies and observations suggest that if
we find a suitable group at C20 to avoid rapid metabolism
occurring at that site, we could improve the PK profiles
of these analogues without adversely affecting their AKi
(adenosine kinase inhibitor) potency. Therefore, a wide
range of substituents at the C20 of the C7 pyridino moiety
could be screened to improve half-life and oral bioavail-
ability of analogue 9e in light our rational (Fig. 2).
Table 1 is a summary of typical examples of nitrogen
linked tertiary amino analogues. SAR results indicated
that analogues with cyclic tertiary amines exhibited
better oral activities than those with acyclic tertiary
amines. Compound 9a is a potent AKi in vitro (AK
(enzyme): IC50=4.1 (ꢁ0.020) nM; AK (intact cell):
IC50=42.3 (ꢁ5.6) nM). This compound also potently
reduced carrageenan-induced thermal hyperalgesia in
the rat (ED50=1.0 mmol/kg, ip). Its in vivo potency,
however, was diminished when the compound was
administrated orally (inflammatory hyperalgesia (po):
ED50>10 mmol/kg). The lack of oral activity of this
analogue (9a) was easily explained by its PK profile,
which revealed several unidentified metabolites and
poor plasma concentration of the parent compound
following oral administration in the rat.10 Cyclic tertiary
amine substituents on the C20 position also yielded
potent AK inhibitors. Five-membered ring pyrrolidino
analogues 9b and 9c had in vivo activity when adminis-
tered ip. However, these compounds had very low oral
bioavailability. Better oral activity and bioavailability
were achieved by expanding the five-membered ring to a
six-membered ring system at the C20 position of these
analogues (entries 5–8, Table 1). The 4-morpholinyl
analogue 9e (ABT-702) was one of the best analogues
and showed improved, broad spectrum oral efficacy.6
Although its plasma half-life still needed to be improved
[t1/2: 0.9 h (iv)], its oral bioavailability reached a rea-
sonable level (23% in rat and 59% in dog). Moving the
oxygen heteroatom of the morpholinyl substituent out-
side the piperidino ring system gave 4-hydroxy-piper-
idinyl derivative 9f, which had improved plasma half-
life of 2.1 h, oral bioavailability of 18.6%, and good in
vivo activity (inflammatory hyperalgesia in rat (po):
ED50=3 mmol/kg). Incorporating two oxygen hetero-
atoms in the spiral ketal derivative 9g further enhanced
the oral bioavailability (F=44.7%), and retained good
half-life (t1=2=2.4 h) following iv administration.
Finally, modification of ketal functionality to a 4-
OTHP oxime derivative 9h, improved plasma half-life to
t1=2=3.6 h, retained good oral bioavailability
(F=46.2%, po) in the rat and improved PK profiles in
the dog. Following oral administration, 9h exhibited an
ED50 of 1 mmol/kg in reducing carrageenan-induced
thermal hyperalgesia in the rat (Table 1).
After comparing analogues with different X linkages at
C20, the nitrogen atom was found to be the best among
carbon, oxygen, and sulfur9 in terms of in vitro and in
vivo potencies (Fig. 1). Less metabolic problems were
observed when X is a nitrogen atom. Analogues with
tertiary amines at C20 demonstrated good correlation
between intact cell inhibition values and in vivo efficacy.
The synthesis of the nitrogen-linked analogues was
simple and straightforward, which allowed access to a
variety of analogues in order to compare their physical
chemical properties, such as solubility, PK (pharmaco-
kinetic) profiles, as well as in vivo activities.
Figure 1.
Chemistry
The C20-derivatives are easily accessible and prepared as
shown in Schemes 1, 2, and 3. In the presence of a cat-
alytic amount of glycine, condensation of malono-nitrile
with 3-bromobenzaldehyde gave a quantitative yield of
2-(30-bromophenyl)-1,1-dicyanyl-ethene (3) (Scheme 1).
Figure 2. Molecular modeling shown here demonstrates that the oxy-
gen of the morpholino moiety of 9e (red) is outside the binding pocket
of the adenosine kinase.
Scheme 1.

G. Z. Zheng et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2071–2074
Table 1. PK profiles of selected pyridopyrimidine analogues
2073
Pharmacokinetics
In vitro (IC50, nM)
AK
In vivo (ED50, mmol/kg)
Rat
Dog
Inflammatory hyperalgesia
t1/2
F
t1/2
F
Entry C20-substituent
(enzyme)
(intact cell)
(ip)
(po)
(iv) (po)
(oral, %)
(5.0 mmol/kg)
(iv) (po)
(oral, %)
(5.0 mmol/kg)
1
2
3
4
5
6
7
8
4.1 (ꢁ0.020)
4.12 (ꢁ1.5)
3.4 (ꢁ2.1)
1.3 (ꢁ1.5)
1.3 (ꢁ0.58)
1.8 (ꢁ0.76)
1.4 (ꢁ1.0)
8.0 (ꢁ2.0)
42.3 (ꢁ5.6)
37.7 (ꢁ9.6)
45.0 (ꢁ32)
42.5 (ꢁ26)
43.3 (ꢁ9.6)
33.0 (ꢁ0.71)
28.3 (ꢁ20)
70.9 (ꢁ19)
1
>10
nd
nd
nd
0.7
3
nf
nf
nf
0.0
nd
nd
nd
nd
6
0.28
0.0
4.5
4
0.41 0.64
nd
nd
8
0.25
0.9
2.1
2.4
3.6
nf
nd
nf
0.0
nd
nd
0.6
5
22.6
18.6
44.7
46.2
1.2 1.3
61.7
nd
nd
3
1
3.6
4.7
4.1 3.4
5.2 3.2
38.5
>100
2
1
nd, not determined; nf, unable to estimate plasma eliminatiom half-life.
Scheme 2. (a) (COCl)2, DMF (cat), CH2Cl2; (b) (1) dimethyl mal-
onate, MgCl2, Et3N; (2) DMSO/H2O.
3-Acetyl-6-chloropyridine (6) was made from the corre-
sponding acid in two steps (Scheme 2).11
Condensation of the ketone 6 with 2-(30-bromophenyl)-
1,1-dicyanylethene (3) in the presence of ammonium
acetate gave the pyridino intermediate 7.12 The latter
was converted to a key intermediate, 8, in the presence
of trisformamide in formamide. Different primary and
secondary amines can be used to generate analogues,
which have different amino moieties at the C70 position.
Alternatively, an animo moiety can be put on to the 3-
acetyl-6-pyridine (6) to give a stable 2-amino-5-acet-
ylpyridine 10, which is then converted to the final pro-
duct the same as the previously described routes. The
latter approach may increase the chemical yield of the
final product, but it is less versatile than the previous
approach (Scheme 3).
In conclusion, based on our SAR and computational
studies, we have identified several pyridopyrimidinyl
Scheme 3. (a), (e) NH4OAc, dichloroethane; (b), (f) trisformamide,
formamide; (c), (d) amine, ethanol.

2074
G. Z. Zheng et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2071–2074
analogues, which are orally effective in a spectrum of
animal pain models with potency and full efficacy com-
parable to, or superior to morphine.
177. (f) Sosnowski, A.; Yaksh, T. L. Anesth. Analg. 1989, 69,
587. (g) Yamamoto, T.; Yaksh, T. L. Pain 1992, 121.
4. (a) Cronstein, B. N.; Naime, D.; Firestein, G. Arth. Rheum.
1995, 38, 1040. (b) Li, E.; Perl, E. J. Neurophysiol. 1994, 4, 1611.
5. Cronstein, B. N.; Naime, D.; Firestein, G. Arth. Rheum.
1995, 38, 1040.
6. Lee, C. H. J. Med. Chem. 2001, 44, 0000.
7. Research other than that described in this paper will be
published in several separate papers.
8. Mathews, I.; Erion, M. D.; Ealick, S. E. Biochemistry 1998,
37, 15607.
9. This portion of the research will be covered by a separate
publication.
10. There are two major metabolites and several minor meta-
bolites from the parent compound 9a following oral adminis-
tration. Due to its relatively complicated PK profiles as
compared with other cyclic amino analogues, this analogue
and its metabolites were not characterized further.
11. Kuo, D. L. Tetrahedron 1992, 48, 9233.
References and Notes
1. (a) Williams, M.; Burnstock, G. In Purinergic Approaches
in Experimental Therapeutics; Jacobson, K., Jarvis, M., Eds.;
Willey-Liss: New York, 1997; pp 3–26. (b) Newby, A. C.
Trends Biochem. Sci. 1984, 2, 42.
2. Davies, L. P.; Jamieson, D. D.; Baird-Lambert, J. A.;
Kazlauskas, R. Biochem. Pharmacol. 1984, 33, 347.
3. (a) Newby, A. C.; Holmquist, A. C.; Illingworth, J.; Pear-
son, J. D. Biochem. J. 1983, 214, 317. (b) Holmgren, M.;
Hedner, J.; Mellstrand, T.; Nordberg, G.; Hedner, T. H.
Naunyn-Schmiedberg’s Arch. Pharmacol. 1986, 334, 290. (c)
Keil, G. J.; DeLander, G. E. Life Sci. 1992, 51, 171. (d) Ahli-
janian, M. K.; Takemori, A. E. Eur. J. Pharmacol. 1985, 112,
171. (e) Poon, A.; Sawynok, J. Eur. J. Pharmacol. 1995, 286,
12. Latif, N.; Assad, F. M.; Girgis, N. S. Indian J. Chem.,
Sect. B 1981, 20, 463.