Optimization of arylindenopyrimidines as potent adenosine A2A/A1 antagonists

Article abstract of DOI:10.1016/j.bmcl.2010.03.024

Two reactive metabolites were identified in vivo for the dual A_2A/A₁ receptor antagonist 1. Two strategies were implemented to successfully mitigate the metabolic liabilities associated with 1. Optimization of the arylindenopyrimidines led to a number of amide, ether, and amino analogs having comparable in vitro and in vivo activity.

Full text of DOI:10.1016/j.bmcl.2010.03.024

Bioorganic & Medicinal Chemistry Letters 20 (2010) 2868–2871
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Optimization of arylindenopyrimidines as potent adenosine A2A/A1 antagonists
*
Brian C. Shook , Stefanie Rassnick, Devraj Chakravarty, Nathaniel Wallace, Mark Ault, Jeffrey Crooke,
J. Kent Barbay, Aihua Wang, Kristi Leonard, Mark T. Powell, Vernon Alford, Daniel Hall, Kenneth C. Rupert,
Geoffrey R. Heintzelman, Kristen Hansen, James L. Bullington, Robert H. Scannevin, Karen Carroll,
Lisa Lampron, Lori Westover, Ronald Russell, Shawn Branum, Kenneth Wells, Sandra Damon, Scott Youells,
Derek Beauchamp, Xun Li, Kenneth Rhodes, Paul F. Jackson
Johnson & Johnson Pharmaceutical Research and Development, L.L.C., Welsh and McKean Roads, PO Box 776, Spring House, PA 19477, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Two reactive metabolites were identified in vivo for the dual A2A/A1 receptor antagonist 1. Two strategies
were implemented to successfully mitigate the metabolic liabilities associated with 1. Optimization of
the arylindenopyrimidines led to a number of amide, ether, and amino analogs having comparable
in vitro and in vivo activity.
Received 20 January 2010
Revised 2 March 2010
Accepted 4 March 2010
Available online 7 March 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
A2A Antagonist
A1 Antagonist
Adenosine antagonists
Parkinson’s disease
Mouse catalepsy
Dopamine
Adenosine receptors are members of the G-protein-coupled
receptor super family.1 The adenosine receptors have been divided
into four different subtypes A1, A2A, A2B, and A3.1 The A2A receptor,
which is highly expressed in the striatum,2 is a much sought after
target in the pharmaceutical industrybecause of its potentialto treat
Parkinson’s disease (PD) and other neurodegenerative disorders.3
We have previously reported a series of methylene amine
substituted arylindenopyrimidines as dual adenosine A2A/A1
antagonists.4 Parkinson’s disease severely affects motor function
but the disease is also associated with anxiety, depression, and
cognitive impairment.5 Therefore the idea of a dual antagonist
could offer improved benefit over the traditional selective A2A
O
O
major
N
N
N
N
N
N
1
2
NH2
NH2
minor
O
N
N
O
NH2
3
Scheme 1. In vivo metabolism of compound 1.
* Corresponding author. Tel.: +1 215 628 7047; fax: +1 215 540 4612.
E-mail address: bshook@its.jnj.com (B.C. Shook).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.03.024

B. C. Shook et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2868–2871
2869
O
O
Ar
Ar
NMO, MeCN
4A ms; (85-90%)
N
N
N
N
Br
O
4
5
N(Boc)2
N(Boc)2
KMnO4,
acetone
(80-92%)
O
O
1. TFA, CH2Cl2
(91-98%)
Ar
N
Ar
R'RN
HO
N
2. R'RNH, HATU,
i-Pr2NEt, THF
(57-88%)
N
N
O
O
7-18
6
NH2
N(Boc)2
Scheme 2. Synthesis of amides 7–18.
antagonist because of its inherent A1 activity which is known to
show positive effects in animal models of learning and
memory.5c,d
We have reported that 1 was a potent A2A/A1 antagonist that
had very potent activity in haloperidol induced catalepsy in mice
(ED50 = 0.2 mg/kg) as well as other animal models of PD.4 Further
characterization of this compound however, showed the formation
of two reactive intermediates, 2 and 3, that were formed in vivo
(Scheme 1). The major metabolite was the iminium ion 2 which
was formed via oxidation alpha to the nitrogen inside the pyrroli-
dine ring followed by elimination. Members of our DMPK group
were able to identify this iminium ion and successfully trap it with
cyanide. The minor metabolite 3 was formed through an analogous
oxidation at the benzylic position followed by elimination of pyr-
rolidine. These metabolites did not prove to be liabilities until
some adverse events were detected in the 28 day non-human pri-
mate studies.
At this point the focus was to address this issue and eliminate
the potential for these metabolites to be formed. The strategy con-
sisted of two approaches to manipulate the current scaffold while
maintaining good functional in vitro and in vivo activity.6 One ap-
proach was to oxidize the benzylic position of the methylene
amines, to form the corresponding amides, which would eliminate
the benzylic oxidation and also decrease the electron rich nature of
the pyrrolidine making it less susceptible to ring oxidation. The
second approach was to convert the benzylic carbon to heteroat-
oms like oxygen and nitrogen, to eliminate benzylic oxidation,
while moving the solubilizing group out further in hope to elimi-
nate the iminium ion formation.
The first approach took advantage of a common intermediate 4
used to synthesize 1 and a number of its analogs (Scheme 2).4 The
benzyl bromide 4 was oxidized to the corresponding aldehyde 5
using N-methylmorpholine-N-oxide (NMO).7 Further oxidation of
the aldehyde with KMnO4 gave the carboxylic acid 6. Deprotection
of the amino group using TFA followed by amide coupling using
HATU afforded the desired amide analogs 7–18.
A wide variety of amides were prepared and most substitu-
tion was tolerated as seen in Table 1. Both acyclic and cyclic
amides gave desired functional activity for both A2A and A1.
For the acyclic analogs 7–12 both secondary and tertiary amides
maintained good in vitro acitivity. The majority of cyclic analogs
prepared were substituted piperazines as these kept a tertiary
amine present for solubility. The cyclic analogs also maintained
good in vitro activity, and a number of these compounds showed
very good activity in our primary animal model of mouse cata-
lepsy,8 comparable to that of compound 1. One difference, how-
ever, is that the amides typically had lower brain concentrations
in rat compared to those having methylene amine substituents.
Table 1
A2A/A1 antagonists having an amide linker. In vitro activity for A2A and A1 functional
assays, and ED50’s for mouse catalepsy
O
Ar
R'RN
N
N
O
NH2
Compound NRR0
Ar
A2A Ki A1 Ki
Mouse catalepsy
ED50 (mg/kg), po
(nM)
12.4
7.8
(nM)
N
N
7
Ph
Ph
208
<1.0
<1.0
Me2N
8
9
94.7
Et2N
Et2N
N
H
N
Ph
Ph
13.5
28.4
45.2
215
ND
ND
H
N
10
11
O
N
Ph
Ph
18.6
16.8
258
119
<1.0
0.4
N
N
H
N
12
13
14
15
16
17
18
Ph
32.5
8.2
12.3
58.4
101
67.4
191
84.1
ND
HN
N
N
N
N
N
N
N
Ph
0.4
4-F-Ph
Ph
11.1
14.9
32.1
20.7
<1.0
<1.0
ND
N
N
For example, the brain Cmax for 1 was 3.6
lM while the brain
4-F-Ph
Ph
N
Cmax for 14 was 0.6 M. The decreased brain exposure, however,
l
did not dramatically decrease its in vivo potency. Most impor-
tantly, the amides did not have the metabolic liability associated
with 1. As we previously reported for the methylene amines,
ND
N
ND = not determined.

2870
B. C. Shook et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2868–2871
ArCHO, NaOH
EtOH; (80-92%)
Ar
MeO
MeO
O
O
19
20
1. K2CO3, DMF,
guanidine-HCl
2. Air
(37-70% for 2 steps)
O
O
1. LiCl, NMP,
180 oC; (41-69%)
Ar
Ar
R'RN
O
N
MeO
N
2. t-BuOK, DMF,
N
N
22-33
Cl
21
R'RN
NH2
NH2
o
70 C; (50-75%)
Scheme 3. Synthesis of ethers 22–33.
only phenyl, 4-fluoro and 3-cyano substituents were tolerated on
the aromatic ring of the amide analogs.
with compounds having a 4-fluoro substituent (compounds 23, 25,
27, 30, and 32). Several of the compounds showed excellent activ-
ity in vivo having ED50’s <1.0 mg/kg. Interestingly, 32 was signifi-
cantly less potent in vitro for A2A than the other analogs tested
in vivo, yet it still remains quite good at reversing mouse catalepsy
The second strategy was implemented next by introducing a
heteroatom to replace the methylene in 1. The commercially avail-
able 6-methoxyindanone 19 was condensed with a variety of aryl
aldehydes under basic conditions to afford the benzylidenes 20
(Scheme 3).4,9 Similar to our previous Letter, 20 was reacted with
guanidine to form an intermediate dihydroaminopyrimidine that,
upon aromatization to the aminopyrimidine, was oxidized up to
the corresponding ketone 21.10 Removal of the methyl ether was
achieved using LiCl11 in NMP at 180 °C to give the intermediate
phenols, not shown. The use of BBr3 was ineffective in removing
the methyl ether. Lastly, the phenol was alkylated with a variety
of aminoethyl chlorides using t-BuOK12 in DMF to give the desired
ethers 22–33.
Table 3
A2A/A1 antagonists having an amino linker. In vitro activity for A2A and A1 functional
assays, and ED50’s for mouse catalepsy
O
R'RN
N
N
NH2
Compound NRR’
A2A Ki
(nM)
A1 Ki
(nM)
Mouse catalepsy
ED50 (mg/kg), po
Like the amides, most of the ether linkages were tolerated (Ta-
ble 2), but there is a consistent ꢀ3- to 10-fold decrease in potency
N
37
4.7
0.9
57.3
12.7
ND
ND
N
H
N
Table 2
38
A2A/A1 antagonists having an ether linker. In vitro activity for A2A and A1 functional
assays, and ED50’s for mouse catalepsy
N
H
N
O
N
Ar
39
0.9
4.4
3.1
15.4
32.7
46.7
ND
R'RN
O
N
N
H
N
NH2
40
<1.0
N
N
O
Compound NRR’
Ar
A2A Ki
A1 Ki
(nM)
Mouse catalepsy ED50
(mg/kg), po
(nM)
H
N
41
Not active at 3 mg/kg
Me2N
22
23
24
25
26
27
Ph
17.3
37.3
7.9
33.3
37.4
103
414
187
468
<3.0
ND
O2S
Me2N
Et2N
4-F-Ph
Ph
H
N
<1.0
ND
42
43
12.5
4.9
142
Not active at 3 mg/kg
ND
Et2N
N
N
4-F-Ph
Ph
50.6
13.7
76.6
i-Pr2N
i-Pr2N
<1.0
ND
N
16.4
Ph
28
29
30
31
32
33
Ph
35.3
6.5
194
48.2
203
159
600
334
ND
N
44
45
46
2.9
2.7
32.4
35.6
256
ND
O
N
N
N
N
N
N
N
Ph
<1.0
O
O
N
N
N
N
N
<1.0
<1.0
4-F-Ph
Ph
18.1
19.3
117
<1.0
11.4
<1.0
47
23.2
10.0
253
137
ND
ND
Ph
0.8
48
N
Ph
35.1
Not active at 10 mg/kg
N
ND = not determined.
ND = not determined.

B. C. Shook et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2868–2871
2871
PhCHO, NaOH
EtOH; (94%)
F
F
O
O
34
35
1. K2CO3, DMF,
guanidine-HCl
2. Air
(51% for 2 steps)
O
O
R'RNH, NMP, µW
150 oC; (42-86%)
R'RN
N
F
N
N
N
37-48
36
NH2
NH2
Scheme 4. Synthesis of amines 37–48.
having an ED50 = 0.8 mg/kg. This may be due to increased brain
exposure of the ether analogs. Unlike the amides, the ethers
showed much improved brain exposure comparable to that of 1.
References and notes
1. (a) Fredholm, B. B.; Abbraccio, M. P.; Burnstock, G.; Daly, J. W.; Harden, T. K.;
Jacobson, K. A.; Leff, P.; Williams, M. Pharmacol. Rev. 1994, 46, 143; (b) Fredholm,
B. B.; Cunha, R. A.; Svenningsson, P. Curr. Top. Med. Chem. 2002, 3, 1349.
2. (a) Augood, S. J.; Emson, P. C. Mol. Brain Res. 1994, 22, 204; (b) Dunwiddie, T. V.;
Masino, S. A. Annu. Rev. Neurosci. 2001, 24, 31.
3. (a) Klaasse, K. C.; Ijzerman, A. P.; Grip, W. J.; Beukers, M. Purinergic Signal. 2008,
4, 21; (b) Simola, N.; Morelli, M.; Pinna, A. Curr. Pharm. Des. 2008, 14, 1475; (c)
Palmer, T. M.; Stiles, G. L. Neuropharmacology 1995, 34, 683.
4. (a) Shook, B. C.; Rassnick, S.; Hall, D.; Rupert, K. C.; Heintzelman, G. R.; Hansen,
K.; Chakravarty, D.; Bullington, J. L.; Scannevin, R. H.; Magliaro, B.; Westover, L.;
Carroll, K.; Lampron, L.; Russell, R.; Branum, S.; Wells, K.; Damon, S.; Youells, S.;
Li, X.; Osbourne, M.; Demarest, K.; Tang, Y.; Rhodes, K.; Jackson, P. F. Bioorg.
Med. Chem. Lett. 2010, 20, 2864; (b) Heintzelman, G. R.; Bullington, J. L.; Rupert,
K. C. WO Patent 2005042500.
5. (a) Muller, C. E.; Scior, T. Pharm. Acta Helv. 1993, 68, 77; (b) DeNinno, M. P.
Annu. Rep. Med. Chem. 1998, 33, 111; (c) Pires, V. A.; Pamplona, F. A.; Pandolfon,
P.; Fernandes, D.; Prediger, R. D. S.; Takahashi, R. N. Behav. Pharmacol. 2009, 20,
134; (d) Davidson, D. A.; Weidley, E. Life Science 1976, 18, 1279–1284.
6. (a) Shook, B. C.; Jackson, P. F. WO Patent 2009055548.; (b) Shook, B. C.; Jackson,
P. F. WO Patent 2009055308.
7. Griffith, W. P.; Jolliffe, J. M.; Ley, S. V.; Springhorn, K. F.; Tiffin, P. D. Synth.
Commun. 1992, 22, 1967.
8. (a) Zarrindast, M. R.; Modabber, M.; Sabetkasai, M. Psychopharamcology 1993,
113, 257; (b) Schwarzschild, M. A.; Agnati, L.; Fuxe, K.; Chen, J. F.; Morelli, M.
Trends Neurosci. 2006, 29, 647.
9. (a) Kuck, D.; Paisdor, B.; Gruetzmacher, H. F. Chem. Ber. 1987, 120, 589; (b)
Bullington, J. L.; Cameron, J. C.; Davis, J. E.; Dodd, J. H.; Harris, C. A.; Henry, J. R.;
Lee Pellegrino-Gensey, J.; Rupert, K. C.; Siekierka, J. J. Bioorg. Med. Chem. Lett.
1998, 8, 2489.
10. Walters, M. E.; Kolonko, R. P.; Wehmeyer, R. M. U.S. Patent 1996-5545760.
11. Bernard, A. M.; Ghiani, M. R.; Piras, P. P.; Rivoldini, A. Synthesis 1989, 287.
12. Vicker, N.; Burgess, L.; Chuckowree, I. S.; Dodd, R.; Folkes, A. J.; Hardick, D. J.;
Hancox, T. C.; Miller, W.; Milton, J.; Sohal, S.; Wang, S.; Wren, S. P.; Charlton, P.
A.; Dangerfield, W.; Liddle, C.; Mistry, P.; Stewart, A. J.; Denny, W. A. J. Med.
Chem. 2002, 45, 721.
Compound 29 had a brain Cmax of 4.1 lM and an extremely potent
ED50 <0.1 mg/kg in the mouse catalepsy model. The ether substi-
tuted analogs showed superior metabolic stability compared to 1.
Finally, the ether linker was replaced with an amino linker
using similar chemistry. The commercially available 6-fluoroinda-
none 34 was condensed with benzaldehyde under basic conditions
to afford the benzylidene 35 (Scheme 4). The aminopyrimidine 36
was formed analogously to 21 using guanidine and air. The fluoride
was then displaced, in the microwave, using a number of different
primary and secondary amines to give the target compounds 37–
48. The attempted the displacement on the 4-fluoro analogs, re-
sulted in competing reactions giving mixtures of products, so the
focus was on the phenyl substituent.
Similarly to the ethers and amides, the amino analogs were also
well tolerated (Table 3). In general, most of the compounds showed
superior in vitro activity for A2A than the amides and ethers, and
most showed increased A1 activity as well. Although a limited
number of compounds were tested in vivo, 40, 45, and 46, did show
comparable activity to the previously described analogs. No brain
concentrations were determined for this set of analogs. Again, it
is worth noting that the compounds in Table 3 did not show the
metabolic liabilities present in 1.
In summary, it was shown that amide, ether and amino substi-
tuted arylindenopyrimidines have potent functional in vitro activ-
ity for both A2A and A1. Several analogs of each set also
demonstrate very potent activity in the mouse catalepsy model.