2,6-Diaryl-4-phenacylaminopyrimidines as potent and selective adenosine A2A antagonists with reduced hERG liability

Article abstract of DOI:10.1016/j.bmcl.2008.01.036

In this report, the design and synthesis of a series of pyrimidine based adenosine A_2A antagonists are described. The strategy and outcome of expanding SAR exploration to attenuate hERG and improve selectivity over A₁ are discussed.

Full text of DOI:10.1016/j.bmcl.2008.01.036

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 1269–1273
2,6-Diaryl-4-phenacylaminopyrimidines as potent and
selective adenosine A2A antagonists with reduced hERG liability
Manisha Moorjani,a,* Xiaohu Zhang,a Yongsheng Chen,a Emily Lin,a Jaimie K. Rueter,a
Raymond S. Gross,a Marion C. Lanier,a John E. Tellew,a John P. Williams,a
Sandra M. Lechner,b Siobhan Malany,c Mark Santos,c Paddi Ekhlassi,d
e
e
e
e
e
´
´
´
Julio C. Castro-Palomino, Marıa I. Crespo, Maria Prat, Silvia Gual, Jose-Luis Dıaz,
John Saundersa and Deborah H. Sleea
aDepartment of Medicinal Chemistry, Neurocrine Biosciences, 12790 El Camino Real, San Diego, CA 92130, USA
bDepartment of Neuroscience, Neurocrine Biosciences, 12790 El Camino Real, San Diego, CA 92130, USA
cDepartments of Pharmacology and Lead Discovery, Neurocrine Biosciences, 12790 El Camino Real, San Diego, CA 92130, USA
dDepartment of Research Analytical, Neurocrine Biosciences, 12790 El Camino Real, San Diego, CA 92130, USA
`
´
Almirall Research Center, Almirall, Ctra. Laurea Miro, 408-410, E-08980 St. Feliu de Llobregat, Barcelona, Spain
e
Received 28 November 2007; revised 8 January 2008; accepted 10 January 2008
Available online 13 January 2008
Abstract—In this report, the design and synthesis of a series of pyrimidine based adenosine A2A antagonists are described. The strat-
egy and outcome of expanding SAR exploration to attenuate hERG and improve selectivity over A1 are discussed. Compound 33
exhibited excellent potency, selectivity over A1, and reduced hERG liability.
Ó 2008 Elsevier Ltd. All rights reserved.
Adenosine receptors belong to the superfamily of G-
protein coupled receptors and are divided into four sub-
types: A1, A2A, A2B, and A3.1 These four receptors are
linked to the secondary messenger adenylyl cylase. Both
subtypes A2A and A2B stimulate adenylyl cyclase,
whereas A1 and A3 inhibit this enzyme.2 A2A receptors
are highly distributed in the central nervous system
and are found in abundance in the basal ganglia, a re-
gion of the brain associated with motor function.3 Not
surprisingly, antagonists of the A2A receptor have dem-
onstrated efficacy in models of Parkinson’s disease (PD)
in addition to exhibiting neuroprotective properties.
Parkinson’s disease is a debilitating motor disorder aris-
ing from the progressive degeneration of dopaminergic
neurons in the nigrostriatal pathway.4 Current dopa-
mine replacement therapies for PD lack neuroprotective
benefits and suffer from poor long-term control and
undesirable side effects, namely dyskinesia (involuntary
movements). In recent clinical trials, the most advanced
A2A antagonist, KW-6002 (istradefylline) from Kyowa
Hakko Kogyo, showed efficacy in alleviating symptoms
of the disease.5 In addition, Schering–Plough has pro-
gressed SCH 420814 into Phase II clinical trials and Ver-
nalis/Biogen Idec are in Phase II clinical trials with
V2006 (structure not disclosed) (Fig. 1).6
Previously, we have reported on the discovery of a new
class of A2A antagonists, based on a pyrimidine core
(Fig. 2).7 Exploration around the pyrimidine core was
carried out, optimizing the R2 and R3 positions with het-
erocycles and the R1 position with amines. Although po-
tent and selective compounds were discovered,
improvements in several areas were desired. The incor-
poration of basic groups within the R1 substituent gave
compounds with desirable physical properties. However,
on further characterization some of the most promising
compounds exhibited inhibition of the hERG channel
(Fig. 3).7,8 It has been cited that blockade of the hERG
K+ channel can lead to prolongation of the heart rate-
corrected QT interval, which in turn elevates the risk
for cardiac arrhythmia and can lead to torsades de points
(sudden death).9 In addition to attenuating hERG
Keywords: A2A antagonists; Adenosine receptor; Parkinson’s disease;
Phenacylaminopyrimidines; hERG.
*
Corresponding author. Tel.: +1 858 617 7600; fax: +1 858 617
7619; e-mail: mmoorjani@neurocrine.com
URL: http://www.neurocrine.com.
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.01.036

1270
M. Moorjani et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1269–1273
O
N
NH2
N
O
O
N
N
O
N
N
N
N
O
O
O
N
N
N
N
KW 6002
SCH 420814
Figure 1. Examples of A2A antagonists in clinical development.
Compounds 14–33 were prepared according to the gen-
eral synthesis outlined in Scheme 1. Starting with either
5-methyl-2-furonitrile (compounds 14–25) or 2-cyano-
pyridine (compounds 27 and 29–33), the corresponding
carboxyamidine 5 was synthesized by treatment with so-
dium methoxide at room temperature followed by reac-
tion with ammonium chloride. The resulting
carboxyamidine was then treated with ethyl cyanoace-
tate in the presence of sodium methoxide to yield 6-
hydroxypyrimidin-4-amine intermediate 6, which was
then treated with phosphorus oxychloride at reflux in
the presence of N,N-diisopropyl ethylamine to give
intermediate 7. The chloro-moiety of intermediate 7
was then displaced with hydrazine and cyclized with
pentane-2-4-dione to give 9 with dimethylpyrazole at
R3. Alternatively, displacement of the chloro on 7 with
pyrazole in the presence of cesium carbonate yields 9
with pyrazole at R3. In the case of compounds 26 and
28 which have a thiazole at position R2, a slightly mod-
ified synthesis was followed. Starting with 2-acetylthiaz-
ole, 3-oxo-3-thiazol-2-yl-propionic acid ethyl ester 11
was generated by treatment with diethyl carbonate and
sodium hydride. Intermediate 11 was then reacted with
carboxyamidines of formula 5, in the presence of potas-
sium tert-butoxide to yield pyrimidin-4-ol intermediate
12. The pyrimidin-4-ol could then be treated with phos-
phorus oxychloride to give 13, followed by reaction with
ammonium hydroxide to give 9. Final compounds were
made by reacting the appropriate phenyl acetic acid with
H
R3
N
R1
O
N
N
R2
1
Figure 2. Pyrimidine core.
N
N
N
N
H
N
H
N
N
N
N
N
N
O
O
N
N
O
O
2
3
Ki (nM) hA2A 12
hA1/hA2A 71
hERG IC50 (nM) 950
Ki (nM) hA2A
hA1/hA2A
hERG IC50 (nM) 653
2
99
Figure 3. Potent A2A antagonists that suffer from hERG inhibition.
inhibition, we wanted to improve selectivity over the A1
receptor. The A1 receptor is present in cardiac muscle
and as a result, may represent a cardio-toxicity risk.1
To address these issues, we broadened the scope of
our right-hand side exploration to incorporate previ-
ously unexplored substituted phenyl groups.
Cl
NH2
NH2NH
NH2
HO
NH2
NH2
(d)
R2
N
N
N
N
(a)
(c)
R2
N
(b)
N
N
N
N
NH
R2
R2
8
R2
6
4
5
7
(e)
(f)
(k)
N
H
N
R3
R3
NH2
R1
O
N
R2
R2
9
10
(j)
N
R2
R3
=
=
O
N
R3
Cl
R3
OH
N
N
O
(g)
(h)
(i)
S
S
N
N
N
S
O
O
R2
13
N
O
R2
12
N
N
N
N
11
Scheme 1. Regents and conditions: (a) NaOMe, NH4Cl, MeOH, rt, 6–12 h, 87–96%; (b) ethyl cyanoacetate, NaOMe, EtOH, 70 °C, 6–12 h, 60–80%;
(c) POCl3, DIPEA, 90 °C, 3–12 h, 50–80%; (d) N2H4, EtOH, 90 °C, 4 h, 80%; (e) pentane-2,4-dione, 0–90 °C, 2 h, 72%; (f) substituted phenylacetic
t
acid, oxalyl chloride, DMF, pyridine, rt, 4–12 h, 20–80%; (g) diethyl carbonate, NaH, 80 °C, 12 h, 50–70%; (h) carboxyamidine, KOtBu, BuOH,
135 °C, 12 h, 50%; (i) POCl3, DIPEA, 90 °C, 3–12 h, 50–80%; (j) NH4OH, MeOH, 80 °C, 12 h, 80–95%; (k) pyrazole, cesium carbonate, 150 °C, 6 h,
80%.

M. Moorjani et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1269–1273
1271
oxalyl chloride in the presence of DMF to generate the
acid chloride. The resulting acid chlorides were then re-
acted with intermediate 9 in the presence of pyridine to
yield compounds of formula 10.
analog 14, IC50 of 300 nM). Furthermore, incorporation
of the pyridine ring enabled us to make hydrochloride
salts of the final compounds which in turn could im-
Table 1. Binding affinities of 14–24 toward the human A2A receptor
and selectivity over the human A1 receptor
Replacement of the right-hand side amine with a substi-
tuted phenyl group increased the lipophilicity. To coun-
ter this increase, substitution was limited to more polar
groups. The 4-methoxy-substituted phenyl (14) was po-
tent (Ki of 4 nM) although selectivity of ꢀ60-fold with
respect to A1 was less than desired (Table 1). Incorpora-
tion of a second methoxy substituent as in the 3,4-dime-
thoxyphenyl (15), and 3,5-dimethoxyphenyl (16)
derivatives improved selectivity while maintaining po-
tency (A1/A2A of 180 and 487, respectively). When the
3,5-dimethoxyphenyl was replaced by small electron-
withdrawing groups such as 3,5-difluorophenyl (17), a
loss in selectivity was observed. A similar trend was
noted upon incorporation of 3,4-difluorophenyl (18).
Further attempts to include more polar substituents on
the phenyl ring led to the incorporation of a para sulfon-
amide group (21). Although once again potency was
maintained with a Ki of 5 nM, selectivity was lower,
being only 90-fold over A1. In general, electronics and
polarity of the phenyl substituent(s) did not greatly af-
fect A2A binding, however, A1 selectivity was more sen-
sitive to these changes. Compounds 22–24 showed
promising potency and selectivity; however, all three
N
N
H
N
R1
N
N
O
O
a
Compound
R1
Ki (nM) hA2A
hA1/hA2A
61
14
4
O
O
O
O
15
16
2
2
180
487
O
F
17
18
5
5
126
152
exhibited inhibition of CYP3A4 or 2D6 (IC50
<
5 lM).11 The most promising compound from this set,
compound 16, showed very good binding activity with
a Ki of 2 nM and selectivity of 487-fold. Profiling of
compound 16 in secondary assays showed that it was
not a potent inhibitor of the major CYP enzymes
(CYP3A4 and 2D6 IC50 > 5 lM). In addition, in a
patch clamp assay, compound 16 did not show signif-
icant inhibition of the hERG channel (IC50 > 3 lM).
By replacing the basic amine side chain with a substi-
tuted phenyl group, we successfully attenuated the
hERG liability. However, this compound exhibited
poor solubility of <0.1 mg/mL at pH 2.12 In an effort
to increase solubility, an SAR exploration was under-
taken to vary the heterocycles at the 2 and 6 positions
on the pyrimidine core.
F
F
F
F
19
20
21
3
3
5
129
411
90
O
O
H
N
O
S
O
The effects of changing the substitution at the R3 and R2
positions were investigated, with particular focus on
reducing the logP contribution of these heterocycles
(Table 2). Replacing dimethyl pyrazole at R3 (14) with
pyrazole (25) decreased the calculated logP by one unit
(from clogP 2.7 to 1.7). Unfortunately, a 5-fold decrease
in selectivity and 2-fold decrease in binding was ob-
served. When R2 was pyridine or methyl furan, incorpo-
ration of a thiazole group at R3 again led to decreased
selectivity as compared to dimethyl pyrazole at R3 (27
vs 28, 14 vs 26). Maintaining the dimethyl pyrazole at
R3 and replacing the methyl furan by a 2-pyridyl group
27 increased potency (Ki of 1 nM) and kept selectivity
(A1/A2A 66-fold) compared to compound 14 (Ki of
4 nM, A1/A2A 61-fold). In addition to maintaining the
binding profile and lowering the clogP, the 2-pyridyl
compound 27 exhibited lower CYP3A4 inhibition with
an IC50 of 1.8 lM (as compared to the methyl furan
22
23
2
162
80
N
N
0.2
24
0.8
179
N
O
a Displacement of specific [3H]-ZM241385 binding at human A2A
expressed in HEK293 cells. Displacement of specific [3H]-DPCPX
binding at human A1 receptors expressed in HEK293 cells. Data are
expressed as geometric means of at least three runs with a standard
deviation less than or equal to 20%.10

1272
M. Moorjani et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1269–1273
Table 2. Binding affinities of 25–28 towards the human A2A receptor, selectivity over the human A1 receptor and calculated logP values
H
R3
N
N
N
O
O
R2
R3
R2
Ki (nM) hA2A
hA1/hA2A
12
clogPb
a
Compound
N
N
25
10
1.7
O
S
26
27
6
1
1
12
66
18
3.2
2.3
2.9
O
N
N
N
N
N
S
28
N
a See footnotes of Table 1.
b Calculated logP values using ACD/Labs logP database.13
Table 3. Binding affinities of 29–33 towards the human A2A receptor,
selectivity over the human A1 receptor and CYP3A4 inhibition
prove solubility. As 2-pyridyl seemed like a promising
alternative to methyl furan at the R2 position, this series
was further explored.
N
N
H
N
R1
Initial results showed a similar SAR trend with the di-
methyl pyrazole/pyridine core (Table 3) as compared
to the dimethyl pyrazole/methyl furan core (Table 1).
The 3,5-dimethoxy phenyl analog 30 was potent (Ki of
1 nM) and exhibited increased selectivity over the mono-
methoxy phenyl 29 (Table 3). However, compound 30
exhibited inhibition of CYP3A4 (IC50 of 1.3 lM). As
encouraging potency and selectivity was seen with elec-
tron-rich phenyl rings, variations of methoxy and
methyl substituents were made. Both 3-methoxy-4-eth-
oxy phenyl 31 and 3,5-dimethylphenyl 32 demonstrated
good potency (Ki of 1 and 3 nM, respectively) and selec-
tivity over 100-fold against A1. The para-sulfone substi-
tuent (33), showed good potency (Ki of 1 nM) and
selectivity of 148-fold. Further in vitro profiling showed
that compound 33 had good functional activity (hcAMP
IC50 of 100 nM), good metabolic stability with an intrin-
sic clearance of 7 mL/min/kg (Human Liver Micro-
somes) (Table 4) and was not a potent CYP3A4 or
2D6 inhibitor (IC50 > 5 lM).14 Head-to-head compari-
son with compound 16 illustrates that compound 33
exhibits increased metabolically stability and has a sig-
nificantly reduced clogP (from 2.5 to 0.7). In addition,
the HCl salt of compound 33 demonstrated good solu-
bility of 0.5 mg/mL at pH 2; whereas, the free base of
compound 16 had a solubility of <0.1 mg/mL at pH 2.
Finally, compound 33 was tested in a hERG patch
clamp assay and showed significantly less activity
(IC50 > 4 lM) than our initial starting points, com-
pounds 2 and 3 (IC50 < 1 lM), not to mention increased
A1 selectivity to almost 150-fold.
N
N
N
O
Compound R1
Ki (nM) hA1/hA2A CYP3A4 inh.
a
b
hA2A
IC50 (lM)
29
1
66
1.8
1.3
7
O
O
30
31
1
1
477
257
O
O
O
32
33
3
1
186
148
>10
>10
O
S
O
a See footnotes of Table 1.
b See note 11.

M. Moorjani et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1269–1273
1273
Table 4. Binding affinities towards the human A2A receptor, selectivity
over the human A1 receptor, intrinsic clearance, and calculated logP
values
Malany, S.; Santos, M.; Gross. R.S.; Williams, J. P.;
Castro-Palomino, J. C.; Crespo, M. I.; Prat, M.; Gual, S.;
´
Dıaz, J. L.; Jalali, K.; Sai, Y.; Zuo, Z.; Yang, C.; Wen, J.;
Compound hA2A Ki hA1/hA2A CLint (HLM)c clogPb
a
O’Brien, Z.; Petroski, R.; Saunders, J. J. Med. Chem., in
press.
(nM)
(mL/min/kg)
8. The hERG potassium current was recorded from a hERG/
HEK cell line using established patch-clamp methods. The
effects of test compounds on the hERG current were
determined at the end of a 5-min application. Test
compounds were tested at six concentrations (0.1 nM,
1 nM, 10 nM, 100 nM, 1 lM and 10 lM). Cisapride
(30 nM) was used as a positive control.
9. Matasi, J. J.; Caldwell, J. P.; Zhang, H.; Fawzi, A.;
Higgins, G. A.; Cohen-Williams, M. E.; Varty, G. B.;
Tulshian, D. B. Bioorg. Med. Chem. Lett. 2005, 15, 3675.
10. On each assay plate, a standard antagonist of comparable
affinity to those being tested was included as a control for
plate-to-plate variability. Overall Ki values were highly
reproducible, the standard deviations were less than or
equal to 20%. All compounds reported were assayed in 3–
6 independent experiments.
16
31
2
1
3
1
487
257
185
148
78
47
113
7
2.5
3.4
2.7
0.7
32
3316
a See footnotes of Table 1.
b See footnotes of Table 2.
c See note 15.
In summary, we have expanded our scope of SAR
around the pyrimidine core to incorporate non-basic
amine side chains, namely substituted phenyls, in an at-
tempt to attenuate hERG liability and improve selectiv-
ity over A1. By balancing lipophilicity with potency and
selectivity, we developed several promising adenosine
A2A antagonists. A number of compounds exhibited
excellent potency and selectivity over A1 to >100-fold.
Furthermore, while maintaining excellent in vitro pro-
files we produced A2A antagonists with good physio-
chemical properties and attenuated the hERG liability.
Further optimization and evaluation of this series will
be reported in due course.
11. Inhibition assays were carried out using microsomes
isolated from transfected cells expressing only CYP3A4,
and in the presence of the fluorescent substrate BFC.
Ketoconazole was used as
a positive control. The
CYP2D6 assay was carried out in the presence of
the fluorescent substrate, AMMC. Quinidine was
used as a positive control. All compounds described
with an IC50 < 30 lM were assayed in two or three
experiments.
12. To determine solubility of compounds, approximately
1 mg of sample was weighed into a 15-mL Falcon
centrifuge tube and the weight recorded to 0.001 mg.
Assay medium, (200 lL, Buffer Solution pH 2.00) was
added and the sample sonicated for 10 min, then shaken
overnight. The sample was then centrifuged and the
supernatant was analyzed by HPLC to determine the
concentration of sample in solution. The concentration in
solution was then calculated based on a standard curve
generated from known dilutions of authentic sample.
13. The calculated logP values stated were obtained using
ACD/Labs LogP database, version 9.02 (2005), Advanced
Chemistry Development Inc., Toronto, Ontario, Canada
(http://www.acdlabs.com).
14. General experimental details for the human cAMP func-
tional assay may be found in the following reference:
Selkirk, J. V.; Nottebaum, L. M.; Ford, I. C.; Santos, M.;
Malany, S.; Foster, A. C.; Lechner, S. M. J. Biomol.
Screen. 2006, 11, 351.
15. General experimental details for this assay may be found
in the following reference: Guo, Z.; Zhu, Y. F.; Gross, T.
D.; Tucci, F. C.; Gao, Y.; Moorjani, M.; Connors, P. J.;
Rowbottom, M. W.; Chen, Y.; Struthers, R. S.; Xie, Q.;
Saunders, J.; Reinhart, G.; Chen, T. K.; Bonneville, A. L.
K.; Chen, C. J. Med. Chem. 2004, 47, 1259.
Acknowledgments
We are indebted to Shawn Ayube, Chris DeVore, and
John Harman for their analytical support.
References and notes
1. Jacobson, K. A.; Zhan-Guo, G. Nat. Rev. Drug Discov.
2006, 5, 247.
2. Van Muijlwijk-Koezen, J. E.; Timmerman, H.; Vollinga,
R. C.; Von Drabbe Kunzel, J. F.; De Groote, M.; Visser,
¨
S.; IJzerman, A. P. J. Med. Chem. 2001, 44, 749.
3. Vu, C. B.; Peng, B.; Kumaravel, G.; Smits, G.; Jin, X.;
Phadke, D.; Engber, T.; Huang, C.; Reilly, J.; Tam, S.;
Grant, D.; Hetu, G.; Chen, L.; Zhang, J.; Petter, R. C. J.
Med. Chem. 2004, 47, 4291.
4. Peng, H.; Kumaravel, G.; Yao, G.; Sha, L.; Wang, J.; Van
Vlijmen, H.; Bohnert, T.; Huang, C.; Vu, C. B.; Ensinger,
C. L.; Chang, H.; Engber, T. M.; Whalley, E. T.; Petter, R.
C. J. Med. Chem. 2004, 47, 6218.
5. Hockemeyer, J.; Burbiel, J. C.; Muller, C. E. J. Org. Chem.
2004, 69, 3308.
6. (a) Neustadt, B. R.; Hao, J.; Lindo, N.; Greenlee, W. J.;
Stamford, A. W.; Tulshian, D.; Ongini, E.; Hunter, J.;
Monopoli, A.; Bertorelli, R.; Foster, C.; Arik, L.; Lach-
owicz, J.; Ng, K.; Feng, K.-I. Bioorg. Med. Chem. Lett.
2007, 17, 1376; (b) Lightowler, S. Presented at the
International Research Conference Targeting Adenosine
A2A Receptors in Parkinson’s Disease and other CNS
Disorders, Boston, MA, May 2006.
16. Compound 33 was prepared in one step from 6-(3,5-
Dimethyl-pyrazole-1-yl)-2-pyridin-2-yl-pyrimidin-4-ylamine
(intermediate 9), using the procedure outlined in Scheme 1
to yield the final compound as an HCL salt (65%): 1H
NMR (300 MHz, DMSO-d6): d 8.77–8.78 (m, 1H), 8.52 (s,
1H), 8.37–8.39 (m, 1H), 8.04–8.07 (m, 1H), 7.89–7.92 (m,
2H), 7.57–7.65 (m, 3H), 6.23 (s, 1H), 3.99 (s, 2H), 3.20 (s,
3H), 2.80 (s, 3H), 2.22 (s, 3H). LCMS: tR = 24.075 (100%);
MS: m/z 463 [M+H]+, expected 463 [M+H]+.
7. Slee, D. H.; Moorjani, M.; Zhang, X.; Lin, E.; Lanier, M.
C.; Chen, Y.; Rueter, J. K.; Lechner, S. M.; Markison, S.;