Synthesis and pharmacological evaluation of aminopyrimidine series of 5-HT1A partial agonists

Article abstract of DOI:10.1016/j.bmcl.2008.12.087

Aminopyrimidine 2 (4-(1-(2-(1H-indol-3-yl)ethyl)piperidin-3-yl)-N-cyclopropylpyrimidin-2-amine) emerged from a high throughput screen as a novel 5-HT_1A agonist. This compound showed moderate potency for 5-HT_1A in binding and functional assays, as well as moderate metabolic stability. Implementation of a strategy for improving metabolic stability by lowering the lipophilicity (c Log D) led to identification of methyl ether 31 (4-(1-(2-(1H-indol-3-yl)ethyl)piperidin-3-yl)-N-(2-methoxyethyl)pyrimidin-2-amine) as a substantially improved compound within the series.

Full text of DOI:10.1016/j.bmcl.2008.12.087

Bioorganic & Medicinal Chemistry Letters 19 (2009) 1159–1163
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and pharmacological evaluation of aminopyrimidine series
of 5-HT1A partial agonists
*
Amy B. Dounay , Nancy S. Barta, Jack A. Bikker, Susan A. Borosky, Brian M. Campbell, Terry Crawford,
Lynne Denny, Lori M. Evans, David L. Gray, Pil Lee, Edward A. Lenoir, Wenjian Xu
Pfizer Global Research and Development, Eastern Point Road, Groton, CT 06340, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Aminopyrimidine
2
(4-(1-(2-(1H-indol-3-yl)ethyl)piperidin-3-yl)-N-cyclopropylpyrimidin-2-amine)
Received 10 November 2008
Revised 18 December 2008
Accepted 19 December 2008
Available online 25 December 2008
emerged from a high throughput screen as a novel 5-HT1A agonist. This compound showed moderate
potency for 5-HT1A in binding and functional assays, as well as moderate metabolic stability. Implemen-
tation of a strategy for improving metabolic stability by lowering the lipophilicity (cLogD) led to
identification of methyl ether 31 (4-(1-(2-(1H-indol-3-yl)ethyl)piperidin-3-yl)-N-(2-methoxyethyl)pyr-
imidin-2-amine) as a substantially improved compound within the series.
Keywords:
5-HT1A partial agonist
Buspirone
Ó 2008 Elsevier Ltd. All rights reserved.
Serotonin (5-hydroxytryptamine, 5-HT) is an important neuro-
transmitter that is known to play a role in regulating numerous
physiological functions, including thermoregulation, vasoconstric-
tion, sexual behavior, appetite, and sleep. Additionally, alterations
in 5-HT-mediated mechanisms are associated with a range of neu-
ropsychiatric conditions, such as anxiety, depression, and bipolar
disorder. Among the various 5-HT receptors, the 5-HT1A subtype
has been most extensively studied.1,2 5-HT1A agonists and partial
agonists have demonstrated clinical effectiveness in the treatment
of anxiety3 and depression.4,5 Recent preclinical and clinical stud-
ies also suggest that 5-HT1A agonists may be useful in treating
dementia,6 Parkinson’s disease,7 pain associated with spinal cord
injuries,8 ischemia,9 and ADHD.10 To date, buspirone (1, Fig. 1) is
the only 5-HT1A agonist marketed in the US. Its market potential
has been limited by side effects, a short half-life (2–3 h) that neces-
sitates three-times a day (tid) dosing,11 and formation of a metab-
this compound focused on improving 5-HT1A binding potency,
functional activity, and metabolic stability.
The synthesis of novel analogs of 2 was accomplished using the
versatile chloropyrimidine intermediate 8 (Scheme 1). The synthe-
sis of 8 began with piperidine 3-carboxylic acid (3), which was pro-
tected as the tert-butyl carbamate 4. Subsequent Weinreb amide
formation and methyl Grignard addition afforded ketone 5, which
was further elaborated to vinylogous amide 6 using Bredereck’s
reagent. Cyclocondensation with guanidine provided 2-aminopy-
rimidine 7, and diazotization in the presence of CuCl2 yielded 2-
chloropyrimidine 8. Although the conversion of 7–8 proceeded in
modest yield, 8 was readily accessible on 100 g scale for use in
medicinal chemistry studies.
Chloropyrimidine 8, which presents potential points of diversi-
fication at both the 2-pyrimidyl position and at the piperidine
nitrogen, served as a useful template for parallel synthesis of a ser-
olite, 2-(piperazin-1-yl)pyrimidine, which is
a
potent a2-
adrenergic receptor antagonist.12 As part of an effort to identify no-
vel 5-HT1A agonists, compound 2 (Fig. 1) emerged as a lead com-
N
pound following
a high throughput screen of the Pfizer
compound collection. Whereas the activity of the 3-(2-(piperidin-
1-yl) ethyl) indole moiety at 5-HT receptors is known,13 compound
2 represents a novel template against this receptor class. This com-
pound showed moderate potency for 5-HT1A in binding and func-
tional assays, as well as moderate stability in an in vitro human
liver microsome (HLM) assay. Lead optimization efforts around
N
H
N
N
N
N
N
O
N
O
N
N
H
1
2
5-HT1A Ki = 75 nM
5-HT1A Ki = 15 nM
* Corresponding author.
E-mail address: Amy.Dounay@pfizer.com (A.B. Dounay).
Figure 1. Buspirone and 5-HT1A agonist 2.
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.12.087

1160
A. B. Dounay et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1159–1163
O
O
O
O
d
b, c
a
OH
OH
N
N
N
NMe2
N
H
Boc
Boc
Boc
6
5
4
3
N
N
e
N
NH2
f
N
Cl
N
N
Boc
Boc
8
7
Scheme 1. Reagents and conditions: (a) Boc2O (1.2 equiv), NaOH (2.1 equiv), dioxane/water, 3 °C—rt (88%); (b) HN(Me)(OMe) (1.2 equiv), EDCI (1.2 equiv), Et3N (3.0 equiv),
HOBt (0.2 equiv), CH2Cl2, 3 °C—rt (92%); (c) MeMgCl (2.0 equiv), MTBE, À10 °C—rt; (d) tert-butoxy-bis(dimethyamino)-methane, reflux; (e) NaOMe (2.0 equiv), MeOH,
guanidine (1.5 equiv), 3 °C—rt (87%, three steps); (f) isoamyl nitrite (1.3 equiv), CuCl2 (1.2 equiv), CH3CN, 65 °C (20%).
ies of 2-alkylaminopyrimidines (11–28). Treatment of 8 at 65 °C
with a variety of primary amines (see Tables 1–3) afforded amino-
pyrimidines 9 (see Scheme 2). Liberation of the piperidine amine
functionality using 4 N HCl followed by N-alkylation of primary al-
kyl bromides14 under microwave heating conditions provided the
bulkier alkyl groups (Table 1). However, these improvements in
5-HT1A binding and functional activities were accompanied by a
significant decrease in HLM stability. Although racemates were
used for the majority of our early SAR work in this series, the effect
of the stereocenter at the 3-piperidyl position was examined using
single enantiomers 15 and 16, which were resolved using chiral
HPLC. Our study of 15 and 16 shows no significant difference in
5-HT1A binding affinities between the two enantiomers, but the
(À)-enantiomer, 15, was shown to be a weaker partial agonist than
the (+)-enantiomer, 16.
array of diverse 2-alkylaminopyrimidines (11–28).15
.
One aspect of the SAR studies on lead compound 2 focused on
the effect of minor modifications of the alkylamino group at the
2-pyrimidyl position. Our preliminary studies showed that higher
5-HT1A binding affinity and intrinsic activity could be achieved
by replacement of the cyclopropyl group with slightly longer or
In optimizing lead compound 2, we also investigated modifica-
tion of the 3-ethylindole portion of the molecule. Recognizing the
potential metabolic soft spots presented by the indole motif, a
blocking approach toward improving HLM stability was under-
taken. Compounds 17–21 were prepared in order to study the
effect of blocking the C2, C5, C6, and C7 positions of the indole
nucleus (Table 2). Although 5-HT1A binding potency improved
somewhat with the addition of fluoro or chloro substituents at
the C5 and C7 positions of the indole (Table 2, compounds 17–
21), the HLM stability was generally reduced in these compounds
compared to lead compound 2 (compounds 17–19; 21). Likewise,
replacement of the indole system with alternate fused bicyclic het-
erocycles or carbocycles, including benzodioxane, benzothiophene,
or indane systems (compounds 22–24) led to a modest improve-
ment in 5-HT1A binding affinity and a significant decrease in
HLM stability. Analysis of 2 using the MetaSite software19 for pre-
diction of metabolic soft spots suggested that the observed cyto-
chrome P450-mediated metabolism in this series has a high
probability of occurring at the position on the ethyl tether adjacent
to the piperidine nitrogen. Compound 25 was prepared to test
whether increased steric bulk at this position would interfere with
metabolic degradation. However, this change had little effect on
either 5-HT1A binding or HLM stability. In vitro metabolite identi-
fication studies in this series showed that one major metabolite re-
sulted from oxidation of the pseudobenzylic position on the
ethylene tether. Ketone 26 was prepared in order to test the effect
of blocking this site. Unfortunately, this modification led to com-
plete loss of 5-HT1A binding potency and also did not improve
HLM stability.
Table 1
Binding, functional activity, and microsomal stability of 2-alkylamino pyrimidines
N
N
R
N
NH
Compound
R
Ent 5-HT1A Ki FLIPR
FLIPR
EC50
(nM)
HLM Clintc
(mL/min/kg)
(nM)a
IA (%)b
1
2
—
—
15
72
38
41
148
35
75.3
138
HN
11
12
13
14
21.3
7.4
32
82
88
—
147
140
—
94
134
135
138
HN
HN
HN
Due to the failure of our blocking strategy to produce metabol-
ically stable analogs, compounds 27 and 28 were prepared to test
the hypothesis that polar, monocyclic heterocycles might serve
as suitable substitutes for the indole. These compounds did indeed
show significant improvement in HLM stability; however, 5-HT1A
binding affinity was dramatically reduced. Not surprisingly, in
our studies of compounds 2–28 (Tables 1 and 2), a strong correla-
tion between cLogD20 and HLM stability was observed. Whereas
modification of the size of the alkyl groups on the alkyl amino sub-
stituent was well tolerated by 5-HT1A (Table 1), replacement of the
5.2
—
—
HN
15
16
n-PrNH
n-PrNH
—
+
11.2
16.3
50
90
302
78
81
178
a
Serotonin1A scintillation proximity assay (SPA) binding assay; IC50 values are
the mean of at least two experiments carried out in duplicate.16
b
FLIPR assay for agonism at serontonin1A receptors; EC50 and IA values are the
mean of at least two experiments carried out in duplicate.17
Human Liver Microsome Incubation assay.18
c

A. B. Dounay et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1159–1163
1161
Table 2
Binding, functional activity, and microsomal stability of modified heterocyclic analogs
N
N
H
N
N
R
cLogDc
HLM Clintd (mL/min/kg)
b
Compound
R
5HT1A Kia (nM)
FLIPR IA (%)
FLIPR EC50 (nM)
2
75.3
38
138
169
1.7
35
N
H
F
17
18
23.6
10.4
28
2.1
2.8
40
49
N
H
Cl
–
–
–
N
H
Cl
19
20
26.6
35.1
–
3.1
2.1
135
N
H
29
216
28
N
H
F
21
22
64.5
37.8
–
–
–
–
2.0
2.5
65
60
N
F
H
O
O
23
24
25
38.5
42.5
40.2
–
–
4.2
3.1
1.6
143
106
60
S
10
–
290
–
N
H
O
26
3420
–
–
3.1
36
N
H
(continued on next page)

1162
A. B. Dounay et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1159–1163
Table 2 (continued)
cLogDc
HLM Clintd (mL/min/kg)
b
Compound
R
5HT1A Kia (nM)
FLIPR IA (%)
FLIPR EC50 (nM)
27
780
–
–
0.8
15
N
S
28
4450
–
–
0.3
<7.6
N
N
H
a
IC50 values are the mean of at least two experiments carried out in duplicate.
EC50 and IA values are the mean of at least two experiments carried out in duplicate.
Calculated logD value (pH 7.4).
b
c
d
Human Liver Microsome Incubation assay.16
Table 3
LogD modification via alkyl amino side chain modification
N
N
R
N
NH
c
Compound
R
5HT1A Kia,b (nM)
FLIPR IA (%)
38
FLIPR EC50 (nM)
cLogD
HLM Clintd (mL/min/kg)
35
2
75.3
138
51.4
65
1.7
HN
HN
HN
29
30
175
78.3
66.3
91
91
94
0.8
1.4
1.2
12.3
34.4
9.6
OH
OMe
OMe
31
40.7
HN
a
Calculated LogD value (pH 7.4).
b
c
IC50 values are the mean of at least two experiments carried out in duplicate.
EC50 and IA values are the mean of at least two experiments carried out in duplicate.
Human Liver Microsome Incubation assay.16
d
improved HLM stability (29); however, 5-HT1A binding affinity
was reduced somewhat by this modification. Capping of the alkyl
amino group with a methyl ether proved to be a better solution
to the problem. Although no improvement in metabolic stability
was observed with compound 30, 5-HT1A binding was not
adversely affected by this modification, and functional activity
was significantly improved over that of the simple alkyl amino
derivatives. Ultimately, the best balance of potency and stability
was achieved with methyl ether 31. As predicted, the reduction
of cLogD in compound 31 relative to lead compound 2 afforded a
significant improvement in HLM stability. Furthermore, this mod-
ification did not affect 5-HT1A binding potency and led to improved
5-HT1A functional activity, and the compound showed good in vivo
brain exposure in rat.21
N
N
NHR1
a
b
8
N
Boc
9
N
N
N
NHR1
N
NHR1
c
.2HCl
N
N
H
R2
10
11-29
Scheme 2. Reagents and conditions: (a) R1NH2, 65 °C; (b) 4 N HCl in dioxane, rt; (c)
R2Br, CH3CN/H2O, Et3N or NaOH, 150 °C (microwave), 45 min.
Identification of aminopyrimidine 2 as a lead compound repre-
sented an important first step toward discovery of a novel 5-HT1A
agonist with a superior pharmacokinetic profile to buspirone. As
part of our lead optimization efforts, several strategies toward
improving the pharmacological and pharmacokinetic profile for
aminopyrimidine 2 were investigated. The initial approach of
blocking potential metabolic soft spots was ineffective. One plausi-
ble explanation for this finding is that as single positions were
blocked on the relatively lipophilic analogs, a compensatory
increase in the rate of P450-mediated formation of alternate
indole core with smaller, less lipophilic heterocycles was not well
tolerated (Table 2, compounds 27–28). These findings suggested
that modulating cLogD by modification of the alkyl amino group
might be a fruitful approach toward improving HLM stability while
maintaining 5-HT1A binding affinity and functional activity.
To test the effect of adding polar groups to the alkyl amino
terminus, compounds 29–31 were prepared (Table 3). As expected,
reduction of LogD by the addition of a terminal hydroxyl group

A. B. Dounay et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1159–1163
1163
metabolites resulted in no net improvement in clearance.22 A more
effective strategy utilized lowering the lipophilicity (cLogD) by
addition of a methyl ether at the alkyl amino terminus of the lead
molecule. Using this approach, methyl ether 31 emerged as a
substantially improved compound within the series. Additional
studies required for further advancement of 31 will include study
of the single enantiomers of the compound in both in vitro and
in vivo pharmacological and pharmacokinetic studies.
12. (a) Bianchi, G.; Caccia, S.; Della Vedova, F.; Garattini, S. Eur. J. Pharmacol.
1988, 151, 365; (b) Barbhaiya, R. H.; Shukla, U. A.; Pfeffer, M.; Pittman, K. A.;
Shrotriya, R.; Laroudie, C.; Gammans, R. E. Eur. J. Clin. Pharmacol. 1994, 46,
41.
13. McKenna, D. J.; Repke, D. B.; Lo, L.; Peroutka, S. J. Neuropharmacology 1990, 29,
193.
14. (a) Alkyl bromides were prepared from the corresponding alcohols using
CBr4 (1.5 equiv) and PPh3 (1.5 equiv) in CH2 Cl2, rt, 16 h; Alcohols that
were not commercially available were prepared according to the following
method (b) Campos, K. R.; Woo, J. C. S.; Lee, S.; Tillyer, R. D. Org. Lett.
2004, 6, 79.
15. Compounds 12 and 25 were prepared as diasteromeric mixtures from racemic
starting materials.
16. Assayed according to published protocol: Graham, J. M.; Coughenour, L. L.;
Barr, B. M.; Rock, D. L.; Nikam, S. S. Bioorg. Med. Chem. Lett. 2008, 18, 489.
17. Standard FLIPR techniques and CHO cells transfected with human 5-HT1A
Acknowledgments
We are grateful to D. McKenzie, W. Klopf, and W. Lam for gen-
erating metabolite identification and plasma/brain exposure data.
receptor and
(ꢀ25,000 cells/well). Growth media was removed and cells were loaded at
37 °C with 25 l/well DMEM containing 4.6 Fluo-4 dye and 2.5 mM
probenecid. After 1 h, cells were washed using assay buffer containing
Ga16 were used. Cells were plated on 384-well plates
l
lM
References
a
final ionic composition (mM) of: 1.26 CaCl2, 0.81 Mg2SO4, 5.55 glucose, 10.0
HEPES, 5.37 KCl, 0.44 KH2PO4, 0.34 Na2HPO4, 136.89 NaCl, 2.5 probenecid.
1. Caliendo, G.; Santagada, V.; Perissutti, E.; Fiorino, F. Curr. Med. Chem. 2005, 12,
1721.
2. Childers, W. E.; Robichaud, A. J. Annu. Rep. Med. Chem. 2005, 40, 17.
3. Goldberg, H. L.; Finnerty, R. J. Am. J. Psychiatry 1979, 136, 1184.
4. Feiger, A. D.; Heiser, J. F.; Shrivastava, R. K.; Weiss, K. J.; Smith, W. T.; Sitsen, J.
M.; Gibertini, M. J. Clin. Psychiatry 2003, 64, 243.
5. Stahl, S. M.; Kaiser, L.; Roeschen, J.; Keppel Hesselink, J. M.; Orazem, J. Int. J.
Neuropsychopharmacol. 1998, 1, 11.
6. Sato, S.; Mizukami, K.; Asada, T. Int. J. Neuropsychopharmacol. 2007, 10, 281.
7. Bezard, E.;Gerlach, I.;Moratalla, R.;Gross, C. E.;Jork, R. Neurobiol. Dis. 2006, 23, 77.
8. Wu, W.-P.; Hao, J.-X.; Xu, X.-J.; Wiesenfeld-Hallin, Z.; Koek, W.; Colpaert, F. C.
Eur. J. Pharmacol. 2003, 478, 131.
Plates were equilibrated for 30 min post-loading in
incubator. A 5 min incubation period at rt was initiated within FLIPR with a
20
a 37 °C, CO2-gassed
l
l addition of (3Â) drug-containing buffer or 20 ll addition of a (2Â)
buffer solution containing serotonin. Fluorescence was monitored at 5 s
intervals for 315 s. Data was analyzed as peak height or maximum
fluorescent counts above basal.
18. Assay method adapted from published protocols: (a) Riley, R. J.; McGinnity, D.
F.; Austin, R. P. Drug Metab. Dispos. 2005, 33, 1304; (b) Obach, R. S. Drug Metab.
Dispos. 1999, 27, 1350.
19. Cruciani, G.; Carosati, E.; De Boeck, B.; Ethirajulu, K.; Mackie, C.; Howe, T.;
Vianello, R. J. Med. Chem. 2005, 48, 6970.
9. Mauler, F.; Horvath, E. J. Cereb. Blood Flow Metab. 2005, 25, 451.
10. Tanaka, K.; Shintani, N.; Hashimoto, H.; Kawagishi, N.; Ago, Y.; Matsuda, T.;
Hashimoto, R.; Kunugi, H.; Yamamoto, A.; Kawaguchi, C.; Shimada, T.; Baba, A.
J. Neurosci. 2006, 26, 5091.
20. Calculated LogD values (cLogD) using ACD/Labs 9.03.
21. Compound 31 dosed in rat (n = 4) 10 mg/kg (SC); mean concentrations (ng/mL)
1 h post dose; plasma: 798.3 224.3; brain: 1013.7 80.4.
22. Fisher, M. B.; Henne, K. R.; Boer, J. Curr. Opin. Drug Discovery Dev. 2006, 9,
101.
11. Gammans, R. E.; Mayol, R. F.; LaBudde, J. A. Am. J. Med. 1986, 80, 41.