Identification and SAR around N-{2-[4-(2,3-dihydro-benzo[1,4]dioxin-2-ylmethyl)-[1,4]diazepan-1-yl]-ethyl}-2-phenoxy-nicotinamide, a selective α2C adrenergic receptor antagonist

Article abstract of DOI:10.1016/j.bmcl.2008.08.055

The discovery of the CNS-penetrant and selective α_2C adrenergic receptor antagonist N-{2-[4-(2,3-dihydro-benzo[1,4]dioxin-2-ylmethyl)-[1,4]diazepan-1-yl]-ethyl}-2-phenoxy-nicotinamide, 13 is described. Structure-activity studies demonstrate the structural requirements for binding affinity, functional activity, and selectivity over other α₂-AR subtypes.

Full text of DOI:10.1016/j.bmcl.2008.08.055

Bioorganic & Medicinal Chemistry Letters 18 (2008) 5689–5693
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Identification and SAR around N-{2-[4-(2,3-dihydro-benzo[1,4]dioxin-
2-ylmethyl)-[1,4]diazepan-1-yl]-ethyl}-2-phenoxy-nicotinamide,
a selective a2C adrenergic receptor antagonist
Snahel D. Patel a, Wendy M. Habeski a, Hyunsuk Min a, Jiansu Zhang a, Robin Roof b,à, Bradley Snyder b,à
,
Gary Bora b,à, Brian Campbell b,à, Cheryl Li b,à, Debra Hidayetoglu b, Douglas S. Johnson b,à
,
Archana Chaudhry b, , Maura E. Charlton a, Natasha M. Kablaoui a,
*
a Pfizer Global Research and Development, Cambridge Laboratories, Cambridge, MA 02139, USA
b Pfizer Global Research and Development, Ann Arbor Laboratories, Ann Arbor, MI 48109, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The discovery of the CNS-penetrant and selective a2C adrenergic receptor antagonist N-{2-[4-(2,3-dihy-
dro-benzo[1,4]dioxin-2-ylmethyl)-[1,4]diazepan-1-yl]-ethyl}-2-phenoxy-nicotinamide, 13 is described.
Structure–activity studies demonstrate the structural requirements for binding affinity, functional activ-
Received 19 June 2008
Revised 11 August 2008
Accepted 15 August 2008
Available online 22 August 2008
ity, and selectivity over other a2-AR subtypes.
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords:
a2C
Adrenergic receptors
Antagonist
The
to the Adrenergic Receptor (AR) subfamily, which contains nine AR
subtypes grouped into a1 2, and b subfamilies. The 2-ARs are in-
a
2C-AR is a G-protein Coupled Receptor (GPCR) that belongs
other known CNS neurotransmitter targets. Through optimization
of HTS hit 1, we discovered compound 13 which meets all of our
criteria.
,
a
a
volved in presynaptic control of neurotransmitter release and
modulate norepinephrine (NE) and serotonin (5HT) neurotrans-
mission.1–3 There is considerable evidence for 5HT and NE dysfunc-
tion in depression, as many marketed antidepressants are known
A high-throughput screen of our internal compound file utiliz-
ing an in vitro competitive radioligand binding assay for
a2C-AR
was run, using cell membranes prepared from CHO cells over-
expressing recombinant human
in a similar assay format against
pounds with selectivity for a
a
a
2C-AR. Hits were then screened
2A-AR and 2B-AR. Potent com-
2C-AR were then evaluated in whole
cell functional assays utilizing cell lines overexpressing the 2-AR
to elevate these neurotransmitters.4 There are three
a
2-AR sub-
2A-AR is expressed in brain
2B-AR is localized in the periphery, while
2C-ARs are highly localized in specific brain regions associated
with mood disorders. Data from 2C-AR genetically modified
2C-AR selec-
a
types:
a
2A-AR,
a
2B-AR, and
a2C-AR;
a
and the periphery,
a
a
a
isoforms to measure the inhibition of AR (NE) induced calcium
mobilization using a FLIPR (fluorescent imaging plate reader).
Compound 1, was identified as a promising lead, with good po-
a
(
a
2C-ARÀ/À) mice, though inconclusive, suggest that
a
tive antagonists could have therapeutic utility in treating depres-
tency and 15-fold selectivity over
over 2B-AR in the binding assays, but marginal selectivity in the
functional assays, being equipotent between 2B-AR and 2C-AR,
2A-AR (Table 1).
a2A-AR and 5-fold selectivity
sion without inducing some of the side effects generally linked
a
with non-subfamily selective
have been reports of selective
in vivo animal antidepressant models.6–10 We therefore set out to
identify novel 2C-AR selective antagonists that are CNS-penetrant,
selective for 2C-AR over 2A-AR, and 2B-AR, and selective over
a
2-AR inhibitors.5 Furthermore, there
a
a
a2C-AR antagonists that are active in
and only 7Â selective over
a
In order to probe the chemical space around compound 1, we
designed a virtual library of 1134 analogs as depicted in Figure 1.
The 1,4-benzodioxan structural feature of compound 1 has been
a
a
a
a
previously incorporated into competitive
a
-AR antagonists,11 and
this moiety was kept constant in the library design. The central
diamine moiety was varied with the first diversity element com-
prised of 9 cyclic or bicyclic examples with varying lengths and
presentation angles, and a second diversity element comprised of
126 alky, heteroalkyl, aryl, and heteroaryl containing building
* Corresponding author. Tel.: +1 617 551 3473; fax: +1 860 686 8090.
E-mail address: natasha.m.kablaoui@pfizer.com (N.M. Kablaoui).
Present address: Asterand, Inc., 440 Burroughs, Detroit, MI 48202, USA.
Present address: Pfizer Global Research and Development, Groton Laboratories,
Eastern Point Road, Groton, CT 06340, USA.
à
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.08.055

5690
S. D. Patel et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5689–5693
Table 1
Compound structure, binding affinities, and functional data illustrating the development of compound 13 and SAR for the series
Compound
Structure
Binding Ki (nM)
Binding sel
Functional Ki (nM)
Functional sel
a2C
a2B
a2A
a2B/
a2C a2A
/
a2C a2C
a2B
a2A
a2B
/
a2C a2A/a2C
MeO
O
N
N
1
33
194
506
220
6
15
23
17
167
1
7
N
O
O
O
O
O
2
3
4
5
5
49
10
44
25
132
NT
474
NT
5
19
N
N
N
H
O
O
N
O
O
3
8
27
3
9
NT
N
N
N
H
O
O
O
O
O
CN
N
N
0.8
0.5
36
28
208
138
48
58
275
286
1.9
27.5
103
501
366
14
262
N
N
N
N
H
Me
Me
Me
O
O
HN
O
30.5
3
12
N
H
N
N
O
O
O
O
O
Me
N
N
N
N
N
N
6
7
8
0.8
10.5
10.9
17.9
81
13
18
1
103
195
28
4.8
3.8
NT
13.3
14.1
NT
73.6
48.4
NT
3
4
15
13
N
H
O
O
N
O
Me
0.6
118
443
N
H
O
O
15.7
N
H
N
N
O
O
O
N
9
2.5
1.9
69.5
15.3
66.1
44.8
28
26
6.8
1.7
132
275
19
41
N
O
O
10
8
4
23
5.2
15.4
3
9
9
8
O
O
N
N
N
N
N
O
N
N
O
11
12
19.6
72.8
146.1
>500
7
21.1
NT
180
NT
167
NT
N
>500
>500
N
O

S. D. Patel et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5689–5693
5691
Table 1 (continued)
Compound
Structure
Binding Ki (nM)
Binding sel
Functional Ki (nM)
Functional sel
a2C
a2B
a2A
a2B/
a2C
a2A
/
a2C
a2C
a2B
a2A
a2B
/
a2C
a2A/a2C
O
O
O
O
N
13
0.9
14.3
94.4
16
106
1.9
31.3
>442
16
>232
N
N
N
H
sized was a ‘cherry-picked’ selection14 based on desired calculated
properties, rather than a traditional combinatorial cross. Com-
pounds were synthesized on a Tecan liquid handler following the
procedure outlined in Scheme 1.
R1
N
N
n
R2
alkyl
aryl
O
O
n
N
n
R1
N
R2
aryl
The compounds were tested in the
binding assays, and the results are depicted in a three dimensional
2-AR activity matrix in Figure 2. Selectivities of the compounds
for 2C-AR over 2A-AR are highlighted by color (green is optimal,
>30-fold), and the selectivities of the compounds for 2C-AR over
2B-AR are highlighted by shape (spheres are optimal, >5-fold).
From this first round ‘cherry-picked’ library only compound 2, with
44-fold selectivity over 2A-AR and 10-fold selectivity over 2B-AR
a2C-AR, a2B-AR, and a2A-AR
het
n
R1
N
N
R2
a
SO2R
n
a
a
n
a
1
X
9
X
126
a
Figure 1. Library design for ‘cherry-picked’ synthesis.
a
a
(Table 1), emerges as a viable lead. The binding potency and selec-
tivity of 2 are mirrored in the functional subtype data. The remain-
der of the library contains compounds with potent binding
blocks. We picked approximately 500 compounds for synthesis
based on the adherence of their calculated properties to the ‘rule
of 5’,12 the constraints of the polar surface area that would hypo-
thetically allow for the compounds to cross the blood–brain barrier
(BBB),13 and structural diversity. The compound library synthe-
affinities to
example is compound 3, the piperazine analog of 2, with 9-fold
selectivity over 2A-AR and 3-fold selectivity over 2B-AR; this
a2C-AR, but with poor selectivity profiles. One such
a
a
highlights the importance of the central homopiperazine to the
selectivity of compound 2 (Table 1).
We then made a series of analogs probing the SAR around the
amide portion of compound 2. Scheme 2 outlines the general pro-
cedure used to vary the amide; commercially unavailable carbox-
ylic acids were synthesized according to the literature
precedent,15 but adapted to the microwave which reduces side
products and increases the yield.
The binding potencies and selectivity profiles for the analogs in
this targeted set are depicted in Figure 3. The visualization shows
that compound 2 represents a vein of subtype selective chemical
space rather than an isolated selective compound. Optimally selec-
O
O
O
O
a
Br
+
N-Boc
N
N-Boc
H-N
O
O
b,c
N
N
R
Scheme 1. General synthetic route for library synthesis (compounds 2 and 3, Table
1). Reagents and conditions: (a) DIPEA, TEA, DMF, 70 °C, 18 h; (b) HCl in dioxane,
methanol, rt, 4 h; (c) RCH2Br, i-Pr2EtN, DMF, 70 °C, 18 h, 5–90% overall.
tive compounds (>50-fold selective for
a2C-AR over a2A-AR and
>10-fold selective for 2C-AR over 2B-AR) are again represented
a
a
by green spheres. Examination of the plot reveals many green
spheres, indicating improved selectivity distribution of this tar-
geted set of compounds.
α
2C-AR
Table 1 shows a sampling of compounds (4–8) that illustrate
the SAR of the series. Compound 4 is very potent and selective in
both binding and functional assays (>250-fold selectivities over
a2A-AR for both binding and functional activities). This illustrates
that meta-cyano substitution of the phenyl ring of the nicotinic
acid is beneficial for both potency and selectivity in binding and
functional determinations. However, we found that all of the com-
pounds with alkylamino or alkoxy substituents instead of the
phenoxy have excellent binding potencies and selectivities, but
poor functional selectivities. For example, the tert-butyl amine
analog (5) has high levels of potency and selectivity in the binding
Compound 3
α
2B-AR
α
2A-AR
assays (58-fold and 286-fold over
tively), but low levels of potency and selectivity in the functional
setting (3-fold and 12-fold over 2B-AR and 2A-AR, respectively).
a2B-AR and a2A-AR, respec-
Compound 2
a
a
Compounds 6 and 7 illustrate this effect with an O-n-butyl moiety
either directly replacing the phenoxy in compound 2 or with the
alkoxy and pyridyl nitrogen transposed. Complete removal of the
phenoxy moiety results in low levels of binding selectivity (8, 1-
Figure 2. Representation of the ‘cherry-picked’ library in 3D ‘selectivity’ space, with
axes depicting binding affinity in each of the 2-AR subtypes. Best compounds are
green spheres near the bottom of the visualization.
a
fold and 28-fold over a2B-AR and a2A-AR, respectively).

5692
S. D. Patel et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5689–5693
O
O
O
O
O
b, c, b
a
N
Br
N
NBoc
+
NBoc
N
HN
N
O
O
O
d
N
.
NH2 HCl
N
R
O
H
Scheme 2. General synthetic route to access nicotinic amide variants of 2, compounds 4–8. Reagents and conditions: (a) DIPEA, TEA, DMF, 70 °C, 18 h, 94%; (b) HCl in dioxane,
methanol, rt, 4 h, 95%; (c) tert-butyl 2-bromoethylcarbamate, DIPEA, DCE, 60 °C, 18 h, 90%; (d) RCOOH, HBTU, TEA, DMF, rt, 5–30%.
Compounds 9 and 10 are examples of modifications to the
spacer between the homopiperazine and the nicotinic acid, and
were synthesized according to Scheme 3. In general, we found that
the less constrained linkers offer better selectivity, which may be
true only because the constrained linkers tested are not in the opti-
mal conformation. For example, replacement of the amide with an
ether linkage results in good functional activity and selectivity
and 12 contain the regioisomers of chroman as replacements for
the benzodioxan. Although they are analogs of the less optimal
benzyl piperazine, they illustrate that the aminoethanol moiety
in the benzodioxan is crucial for a2-AR activity. The chroman ana-
log of compound 2 could be an interesting analog to test. Com-
pounds 11 and 12 were synthesized analogously to Scheme 1;
the bromomethyl chroman starting material for analog 12 is syn-
thesized according to the literature precedent,16 and the regio-
isomer used in compound 11 is commercially available.
Considering that compounds 1 through 12 are chiral, we chose
to resolve compound 2 into its enantiomers in order to study the
effect of stereochemistry on potency and selectivity. We found that
compound 13, the S-enantiomer of compound 2 (94% ee), has
excellent potency and selectivity in both the binding and func-
tional assays, with functional activity of 2 nM and functional
(7 nM, 19-fold and 41-fold over a2B-AR and a2A-AR, respectively);
however, the more conformationally constrained oxadiazole
spacer does not show good binding or functional selectivity,
although it is very potent in both settings (2 nM). Compounds 11
α
2C-AR
selectivites of >230-fold over
a2A-AR and 16-fold over a2B-AR.
The R-enantiomer of compound 2 (60% ee, not shown) is less po-
tent than the S-enantiomer (9.2 nM, binding assay), and is less
selective, at 12-fold and 53-fold over
a2B-AR and a2A-AR, respec-
tively, in the binding assay (functional data not available). The
enantiomerically enriched benzodioxans were synthesized accord-
ing to the literature procedures.17
Although compounds 13 and 4 essentially have equivalent po-
tency and selectivity profiles, compound 13 was chosen for further
testing. Resolution of compound 4 could provide an extremely po-
tent and selective compound, but we chose to move forward with
13 due to its lower molecular weight and PSA, which are predictive
of better brain penetration. Binding affinity of 13 for the rat ortho-
α
2B-AR
α
2A-AR
log of
petitive radioligand binding assay analogous to that applied for the
human ortholog. Rat 2A-AR and 2B-AR binding affinities were
177 nM and 7.6 nM, respectively, providing a sufficient window
of >100-fold selectivity for 2C-AR over 2A-AR; the 4-fold selectiv-
a2C-AR was determined to be 1.7 nM, using an in vitro com-
a
a
Figure 3. Representation of the second round targeted library in 3D ‘selectivity’
space, with axes depicting binding affinity in each of the 2-AR subtypes. Best
compounds are green spheres near the bottom of the visualization.
a
a
a
ity over
a2B-AR in rat is recognized as low, but is acceptable for
O
O
O
N
a, b
N
N
NH
OH
O
PhO
N
N
O
c
N
d
Cl
OPh
Br
N
A
PhO
O
O
O
N
N
O
OPh
N
N
N
N
N
N
O
O
10
9
Scheme 3. Synthetic routes to compounds 9 and 10. Reagents and conditions: (a) ethylchloroacetate, TEA, DMF, 72%; (b) LAH, THF, 56%; (c) NaH, THF, 10% (3-(bromomethyl)-
2-phenoxypyridine from NBS bromination of 2-phenoxy-3-pyridinemethanol, 47%); (d) A, DIPEA, DMF, 180 °C, 14%; note compound A was made analogously to the literature
precedent.18

S. D. Patel et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5689–5693
5693
Table 2
In conclusion, through high-throughput screening, ‘sparse ma-
trix’ library synthesis and traditional medicinal chemistry, we have
identified in compound 13 a useful chemical tool for exploring the
Compound 13 activities in a panel of human CNS targets
a
a
2C-AR
1A-AR
1 nM
59 nM
D4
10% at 1
4% at 10
1600 nM
27% at 10
1790 nM
3480 nM
1480 nM
l
l
M
M
M1
M2
M3
NET
DAT
SERT
H1
role of
a2C-AR in rat CNS models. Compound 13 has excellent bind-
5HT1a
5HT1b
5HT2a
5HT2b
5HT7
D2
141 nM
44% at 10
2400 nM
690 nM
530 nM
820 nM
75 nM
ing affinity and functional activity for
a
2C-AR, outstanding selectiv-
lM
l
M
ity among CNS-related targets, excellent binding affinity and
selectivity in rats, and adequate exposure in the rat CSF and brain.
33% at 1
1900 nM
l
M
References and notes
D3
H2
1. Bylund, D. B.; Eikenberg, D. C.; Hieble, J. P. Pharmacological Reviews 1994, 46,
121.
2. Kamibayashi, T.; Maze, M. Anesthesiology 2000, 93, 1345.
3. Ruffolo, R. J.; Nichols, A. J., et al Annual Reviews in Pharmacology and Toxicology
1993, 32.
Table 3
Drug exposure of 13 in plasma, brain and cerebrospinal fluid (CSF) of Male Wistar–
Kyoto rats following subcutaneous administration
4. Kent, J. M. Lancet 1997, 355, 911.
5. Scheinin, M.; Sallinen, J.; Haapalinna, A. Life Sciences 2001, 68, 2277.
6. Hoglund, L.; Koivisto, A.-P.; Tauber, A.; Kallatsa, O.; Sallinen, J.; Silver, S.;
Hoffren, A.-M.; Iles, M.; Wurster, S.; Oy Juvantia Pharma Ltd: Finland, 2004, p
WO 2004/067513 A1.
7. Sallinen, J.; Hoglund, I.; Engstrom, M.; Lehtimaki, J.; Virtanen, R.; Sirvio, J.;
Wurster, S.; Savola, J.-M.; Haapalinna, A. British Journal of Pharmacology 2007,
150, 391.
Dose (mg/kg)
Plasma (ng/ml)
Brain (ng eq/q)
CSF (ng/ml)
CSF/a2C Ki
3
10
30
744
2100
6190
173
630
2230
7.39
26.9
113
9
33
137
8. Sallinen, J.; Sirvio, J.; Orion Corporation: Finland, 2003, p WO 03/082825 A1.
9. Wiker, C.; Linner, L.; Svensson, T. Poster presented at the European College of
Neuropsychopharmacology, 16th Annual Meeting and Congress, Prague, 2003.
10. Wurster, S.; Engstrom, M.; Svola, J.-M.; Hoglund, I.; Sallinen, J.; Haapalinna, A.;
Tauber, A.; Hoffren, A.-M.; Salo, H.; Orion Corporation Oy Juvantia Pharma Ltd:
Finland, 2001, p WO 01/64645 A2.
11. Nelson, W. L.; Wennerstrom, J. E. Journal of the Chemical Society, Chemical
Communications 1976, 22, 921.
12. Lipinski, C. A.; Lombardo, F.; Dominy, B.; Feeney, P. Advanced Drug Delivery
Reviews 1997, 23, 3.
13. Kelder, J.; Grootenhuis, P. D. J.; Bayada, D. M.; Delbressine, L. P. C.; Ploemen, J. P.
Pharmaceutical Research 1999, 16, 1514.
14. Schneider, G. Current Medicinal Chemistry 2002, 9, 2095.
15. Chiu, B. K.; Lam, M. H.; Lee, D. Y.; Wong, W. Y. Journal of Organometallic
Chemistry 2004, 689, 2888.
16. Matesanz, E.; Alcazar, J.; Andres, J. I.; Bartolome, J. M.; Bruyn, M. D.; Fernandez,
J.; Emelen, K. V. Tetrahedron Letters 2003, 44, 2275.
17. Birch, A. M.; Bradley, P. A.; Gill, J. C.; Kerrigan, F.; Needham, P. L. Journal of
Medicinal Chemistry 1999, 42, 3342.
looking at CNS effects in rats, since a2B-AR is only expressed in the
periphery. We also tested compound 13 in an assay panel of other
CNS targets (Table 2) and found that it is >50-fold selective over
a
1A-AR and D3, and >100-fold selective over all other targets
tested, again supporting that the compound can be used to identify
CNS effects of 2C-AR modulation. There is some concern about the
a
relatively high molecular weight and polar surface area of com-
pound 13 (488.6 and 76.2, respectively), which could limit the
CNS exposure of the compound. However, subcutaneous drug dos-
ing of compound 13 in male Wistar–Kyoto rats showed adequate
drug exposure in the cerebrospinal fluid (CSF). Table 3 shows the
exposure of the compound in the plasma, brain, and CSF; the ratio
of compound exposure in the CSF versus the a2C-AR Ki’s indicates
that adequate exposure should be achieved for observation of
CNS activity.
18. Zablocki, J.; Kalla, R.; Perry, T.; Palle, V.; Varkhedkar, V.; Xiao, D.; Pisopio, A.;
Maa, T.; Gimbel, A.; Hao, J.; Chu, N.; Leung, K.; Zeng, D. Bioorganic & Medicinal
Chemistry Letters 2005, 15, 609.