Design, optimization, and in vivo evaluation of a series of pyridine derivatives with dual NK1 antagonism and SERT inhibition for the treatment of depression

Article abstract of DOI:10.1016/j.bmcl.2012.11.094

A series of substituted pyridines, ether linked to a phenylpiperidine core were optimized for dual NK₁/SERT affinity. Optimization based on NK₁/SERT binding affinities, and minimization of off-target ion channel activity lead to the

Full text of DOI:10.1016/j.bmcl.2012.11.094

Bioorganic & Medicinal Chemistry Letters 23 (2013) 407–411
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
jour’nal homepage: www.elsevier.com/locate/bmcl
Design, optimization, and in vivo evaluation of a series of pyridine
derivatives with dual NK1 antagonism and SERT inhibition for the treatment of
depression
Kevin W. Gillman a, , Michael F. Parker a, Mark Silva a, Andrew P. Degnan a, George O. Tora a,
⇑
Nicholas J. Lodge b, Yu-Wen Li b, Snjezana Lelas b, Matthew Taber b, Rudolf G. Krause c, Robert L. Bertekap c,
Amy E. Newton b, Rick L. Pieschl b, Kelly D. Lengyel b, Kim A. Johnson d, Sarah J. Taylor d,
Joanne J. Bronson a, John E. Macor a
a Neuroscience Discovery Chemistry, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, United States
b Neuroscience Discovery Biology, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, United States
c Lead Evaluation and Mechanistic Biochemistry, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, United States
d Pharmaceutical Candidate Optimization-Metabolism and Pharmacokinetics, Bristol-Myers Squibb Research and Development, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of substituted pyridines, ether linked to a phenylpiperidine core were optimized for dual NK1/
SERT affinity. Optimization based on NK1/SERT binding affinities, and minimization of off-target ion chan-
nel activity lead to the discovery of compound 44. In vivo evaluation of 44 in the gerbil forced swim test
(a depression model), and ex-vivo NK1/SERT receptor occupancy data support the potential of a dual
acting compound for the treatment of depression.
Received 26 October 2012
Revised 16 November 2012
Accepted 20 November 2012
Available online 30 November 2012
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
Neurokinin 1 antagonists
Serotonin transporter inhibitors
Receptor occupancy
Depression
The therapeutic potential of tachykinin receptor antagonists for
the treatment of depression has been a subject of discussion for
many years.1 Although Phase II clinical trials with the NK1 receptor
antagonist Aprepitant2 showed evidence of antidepressant efficacy,
Phase III trials with Aprepitant as well as a follow-on NK1 receptor
antagonist Casopitant did not support these initial positive find-
ings.3 While these results have negatively impacted the develop-
ment of NK1 antagonists as single agents for the treatment of
depression, there has been interest in the combination of NK1
receptor antagonism with Serotonin Transporter (SERT) inhibitor
activity as a potential means of potentiating the antidepressant
effect of NK1 antagonism.4
off-target ion channel activity (Fig. 1). Initial SAR focus revolved
around introducing a substituted pyridine on the ether side-chain
as a means of increasing polarity in this series (Scheme 1).
Comparison of 1 with its direct pyridine analogue 6 showed
that incorporation of a pyridine ring was tolerated, although po-
tency was diminished three–four-fold for both NK1 antagonism
and SERT inhibition compared to 1 (Table 1). Modification around
the left-hand phenyl ring significantly impacted both NK1 and SERT
binding affinities. In general, decreases in hERG flux activity paral-
leled the decrease in NK1/SERT activity in this series. It should be
noted that to date a 4-fluorophenyl substituent on the piperidine
We set out to develop a series of molecules with dual acting NK1
antagonism and SERT inhibition, with high affinity for both targets.
Early efforts focused on identifying chemotypes that had good
physicochemical properties and potential for brain uptake. A (4-
phenylpiperidine-4-yl)-methyl ether chemotype, exemplified by
1, was chosen as our starting point.5 We designed a series of ana-
logs directed toward improving NK1/SERT affinity,6 and decreasing
2
R
CF
X
CF
3
3
O
O
SAR Modifications
X
X
N
H
N
1
Cl
R
H, R
1
hNK IC = 3.6 nM
X = C or N
1
50
hSERT IC = 1.7 nM
50
Ca Flux IC = 5.2 µM
50
hERG Flux IC = 7.2 µM
50
⇑
Corresponding author. Tel.: +1 203 677 6785; fax: +1 203 677 7884.
E-mail address: kevin.gillman@bms.com (K.W. Gillman).
Figure 1. SAR strategy around the 4-methyl-4-phenylpiperidine core.
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.11.094

408
K. W. Gillman et al. / Bioorg. Med. Chem. Lett. 23 (2013) 407–411
Table 2
Br
In vitro profiling data for ‘right-hand’ substituted pyridine analogues 12–16 produced
via Scheme 2
b
a
N
N
Cl
CF3
Cl
CF3
3
2
CF3
CF3
1) c or d
2) e,f
O
O
N
N
Compd.
R
R00
hNK1
IC50
hSERTa
IC50
Ca flux
hERG flux
a
N
N
R
Cl
Boc
IC50
(lM) IC50 (lM)
(nm)
(nM)
5
4
Scheme 1. Reagents and conditions: (a) NBS/CCl4; (b) KOtBu/THF, tert-butyl 4-
(hydroxymethyl)-4-phenylpiperidine-1-carboxylate; (c) Pd(Ph3P)4/MeCN/KOH/
1200C, R-B(OH)2; (d) Pd(OAc)2, NaOtBu, rac-BINAP, toluene, 1200C, R = 10 or 20
amine; (e) TFA, CH2Cl2; (f) formaldehyde, NaCNBH4, MeCN.
12
13
14
15
H
600
480
10
21
15
13
CH3
3.1
30
Table 1
In vitro profiling data for the first generation pyridine analogues 6–11 produced via
Scheme 1
H
34
74
66
6.6
6.3
14
7
7
7
33
25
12
H
16
CH3
a
Compd.
R
X
hNK1
hSERTa
Ca flux
hERG flux
IC50 (nM)
IC50 (nM)
IC50
(lM)
IC50 (lM)
a
Values are the mean of two experiments.
6
7
8
9
10
11
H
H
H
4-Cl
4-F
4-F
2-F
11
12
430
29
45
52
7.4
6.3
23
9.0
4.3
88
2
7
11
7
5
28
25
23
20
6
CH3
CH3
H
CH3
CH3
findings the pyridine nitrogen position had minimal effects on
SERT potency.
6.3
43
Additional modification of the pyridine ring was explored fixing
the trifluoromethyl group in the 4-position while altering the 2-po-
sition with a variety of acyclic and cyclic amines. This exercise lead
to analogues with significant improvements in NK1 and SERT affin-
ities, as well as in some examples, decreasing off-target ion chan-
nel activity in the flux assays (Table 3). Both acyclic and cyclic
amines possessed good NK1/SERT binding affinities. There was a
clear preference for tertiary amines having better dual activity than
secondary amines, although branching alpha to the nitrogen im-
proved potency as in secondary amine 20. Extended branching of
the alkyl groups decreased both NK1 and SERT affinities as in
example 22. Addition of a second basic amine as in the 4-methyl
piperazine 28 also significantly decreased both NK1 and SERT
affinities.
a
Values are the mean of two experiments.
was the best tolerated modification on the 4-phenylpiperidine
ring. Alkylation of the piperidine nitrogen with a methyl group im-
proved SERT binding affinity slightly, while having little impact on
NK1 affinity. Additional alkyl groups were explored on the piperi-
dine nitrogen, but a methyl substituent was determined to be
optimal.7
Altering the position of the pyridine nitrogen impacted NK1
affinity significantly, demonstrating a clear preference for 2- and
4-substituted pyridines. Comparison of 6 and 15 (and 7/16) shows
a further preference for having the trifluoromethyl group in the 4-
position relative to the pyridine nitrogen (Table 2). Removal of sub-
stitution on the 4-position significantly decreased NK1 binding
affinity as demonstrated by example 13. In contrast to the above
In vitro profiling of one of the better analogues in this series
compound 25, demonstrated that introduction of an amino func-
tionality has a positive impact on protein free fraction (human,
Cl
Cl
a,b,c
d
e,f
Br
Scheme 1
Br
F3C
N
O
31
CO2CH3
F3C
N
Br
F3C
N
CO2CH3
F3C
N
30
33
32
CO2H
Br
Br
Br
g
Scheme 1
h,f
F3C
N
F3C
N
F3C
N
34
36
35
Scheme 2. Reagents and conditions: (a) BuLi, THF, CO2; (b) mCPBA, CH2Cl2; (c) MeOH, HCl; (d) POCl3, 1000C; (e) NaBH4, MeOH; (f) CBr4, Ph3, CH2Cl2.

K. W. Gillman et al. / Bioorg. Med. Chem. Lett. 23 (2013) 407–411
409
Table 3
In vitro profiling data for 2-aminopyridine analogues 17–29 produced via Scheme 1
hSERTa IC50 (nM)
Ca flux IC50
(
lM)
hERG flux IC50 (lM)
a
Compd.
X
R
R0
hNK1 IC50 (nM)
17
18
19
20
21
22
4-F
4-F
4-F
4-F
4-F
4-F
H
–N(CH3)2
–N(CH2CH3)2
–NHCH2CH3
–NHCH(CH2CH3)2
–N(CH3)CH2CH3
–N(CH2CH(CH3)2)2
20
7.3
7.3
4.8
15
3.9
140
6.0
4.7
56
6.5
5.3
10
8.4
14
80
7.1
6.1
9.5
CH3
CH3
CH3
CH3
CH3
1.6
760
3.0
1.0
180
23
24
25
26
H
H
2.7
1.2
1.9
6.3
58
3.5
2.9
5.8
5.5
2.3
7.4
12
H
CH3
CH3
CH3
6.6
6.4
4.0
4-F
4-F
8.7
27
28
4-F
H
CH3
H
29
3.3
94
8.6
9.5
8.3
16
790
29
4-F
CH3
13
2.4
3.4
7.7
a
Values are the mean of two experiments.
6.9% free fraction, gerbil, 6.6% free fraction) and a modest decrease
on hERG channel activity. Unfortunately, the microsomal stability
data (human/rat/mouse: 100/43/51% remaining after 10 min
As in initial studies the 4-fluoro-substituent on the 4-phenyl
piperidine ring afforded the best analogs (44 vs 54–56). Interest-
ingly, replacing the 4-trifluoromethyl group with a cyclopropane
as in 52, afforded an equipotent analog to 44. However, compound
44 had significantly better in vitro microsomal stability than 52.8
In addition, compound 44 had acceptable plasma protein free
fraction across species (human% free = 2.0%, gerbil% free = 2.6%,
rat% free = 3.9%, mouse% free = 2.9%). In contrast to cyclopropane
52, significantly decreased NK1 affinity was seen in 53 with the
introduction of a nitrile in place of the trifluoromethyl group in
the 4-position (44 vs 53). Based on these data, we chose 44 as a
superior compound for in vivo evaluation.
Compound 44 was dosed po (0.25% methylcellulose suspension)
120 min prior to testing in our gerbil Forced Swim Test assay
(gFST).9 Test animals were compared against vehicle, with
10 mg/kg po fluoxetine included as a positive control. Total plasma
and brain exposures were determined, as well as NK1/SERT ex vivo
receptor occupancy. In the gFST studies, compound 44 significantly
decreased immobility (the measure of anti-depressant activity) at
1, 3, and 10 mg/kg po compared to vehicle, and demonstrated effi-
cacy comparable to fluoxetine at the 3 mg/kg dose po (Fig. 2).
Exposure data from these studies indicated that compound 44
was efficacious at modest plasma levels. For example at 1 mg/kg,
the total plasma concentration of 44 was 9.2 nM 8.1 nM. At
3 mg/kg and 10 mg/kg, plasma levels were 37 nM 19 nM, and
140 nM 40 nM respectively at t = 120 min. More importantly,
compound 44 had total brain exposures that were on average
10-fold higher than total plasma levels over the dose range.
Ex vivo receptor occupancy was determined for both NK1 and
SERT target engagement. For NK1 ex vivo studies we used 125I-sub-
@0.5 lM conc.) was modest for 25 and for the majority of
compounds in this series. The major pathway for metabolism in
the 2-aminopyridine series was oxidative dealkylation of the
amine R groups.
As a means to eliminate this metabolic ‘soft spot’ we explored
additional substituents at the 2-position, focusing on R groups that
potentially were resistant to oxidation. Utilizing standard substitu-
tion chemistry on a 2-chloropyridine scaffold (Scheme 1) we ex-
plored a variety of substituents (Table 4). Reduction of the parent
chloro compound 10 with hydrogen and Pd/C afforded analog 37
which suffered a significant loss in NK1 receptor potency. Displace-
ment of the chlorine with sodium cyanide or methoxide afforded
compounds 38 and 39. Compound 39 possessed good NK1/SERT
affinities, yet as in the amine series, this analogue had poor meta-
bolic stability. Analogues 40–51 were synthesized via standard Su-
zuki chemistry from their corresponding boronic acids followed by
either Pd/C reduction in the cases of the saturated analogues, or
bycyclopropanation of the alkenylboronic esters using Diazometh-
ane/Pd(OAc)2, to afford the cyclopropyl analogues. From the
in vitro screening data in Table 4, the 2-cyclopropyl analogue 44
possessed excellent dual NK1/SERT affinity with
a two-fold
improvement in hERG flux activity over the lead 2-chloro-4-tri-
fluoromethyl-phenyl analogue 1.
Lastly, we turned our efforts to improving on compound 44 by
fixing the 2-cyclopropyl position on the pyridine, modifying substi-
tution at the 4-position of the pyridine, and the substituents on the
4-phenylpiperidine ring (Table 5).

410
K. W. Gillman et al. / Bioorg. Med. Chem. Lett. 23 (2013) 407–411
Table 4
In vitro profiling data for substituted pyridine analogues 17–28 produced via substitution of a 2-chloropyridine precursor
hSERTa IC50 (nM)
Ca flux IC50
(lM)
hERG flux IC50 (lM)
a
Compd
X
R
R’
hNK1 IC50 (nM)
37
38
39
40
41
42
43
4-F
H
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
CN
–OCH3
–CH(CH3)2
–CH2CH(CH3)2
CH@CHCH3
CH@C(CH3)2
700
380
22
12
16
2.2
4.3
2.8
12
8.9
10
11
16
7.3
-
4.4
5.3
11
29
33
9.9
26
12
13
18
4-F
4-F
4-F
4-F
4-F
29
39
12
44
45
46
47
48
49
50
4-F
H
CH3
CH3
H
1.9
2.0
2.0
9.1
8.7
5.3
5.3
2.1
6.8
13
28
16
5.6
56
5.0
2.8
1.8
6.5
6.8
3.5
7.1
13
9.0
5.5
34
9
H
4-F
4-F
4-F
H
CH3
CH3
CH3
CH3
4.3
12
51
4-F
CH3
6.1
6.8
8.2
12
a
Values are the mean of two experiments.
Table 5
In vitro profiling data for 4-substituted-2-cyclopropyl pyridine analogues 52–56
250
200
150
100
50
*
*
hNK1a
hSERTa
Ca flux
hERG
*
Compd.
X
R
*
IC50 (nM)
IC50 (nM)
IC50
(l
M)
IC50
(lM)
52
4-F
1.8
1.5
7.4
17
53
54
55
56
4-F
–CN
–CF3
–CF3
–CF3
290
11
6.4
3.7
1.6
32
9.1
8.3
8.9
3.4
4.5
3.6
17
0
4-Cl
3-Cl
3-F
3.1
9.2
9.0
Veh
Flx@10
1
3
10
Figure 2. Gerbil forced swim test. Dose (mg/kg, po, t = 120 min). Fluoxetine at
10 mg/kg po as a positive control (n = 4). ANOVA p <0.001 ⁄p <0.01 Dunnett’s t-test.
a
Values are the mean of two experiments.
stance P.10 SERT occupancy studies were performed with 3H-
Citalopram.11 For all doses, we observed significant target engage-
ment at both the NK1 and SERT binding sites in the gerbil brain.
Receptor occupancy increased proportionally with dose (Fig. 3).
Observations from the occupancy studies demonstrated that it
is not necessary to saturate either target in order to see efficacy
in the gFST. In fact, at the minimum efficacious dose of 1 mg/kg,
only modest target engagement at both targets was required in
order to obtain a significant behavioral effect in this assay.12 These
data support the synergistic relationship between modulation of
the NK1 and SERT pathways with respect to behavioral outcomes
in models for depression.
In conclusion, we optimized a series of substituted pyridines
with an ether-linkage to a 4-methyl-4-phenylpiperidine core.
SAR studies focused on optimization of brain penetration, meta-
bolic stability, and decreasing ion channel activity, while maintain-
ing dual target affinities to afford compound 44. Compound 44
(BMS-817693) demonstrated efficacy in our gerbil forced swim
test at reasonable doses. Based on in vitro ion channel activity
data, and the modest exposures necessary for efficacy, we have

K. W. Gillman et al. / Bioorg. Med. Chem. Lett. 23 (2013) 407–411
411
Ladduwahetty, T.; Merchant, K. J.; Metzger, J. M.; MacIntyre, D. E.; Sadowski,
S.; Sohal, B.; Owens, A. P. J. Med. Chem. 1998, 41, 4623; (f) Lowe, J. A.; Droxda, S.
E.; McLean, S.; Bryce, D. K.; Crawford, R. T.; Snider, R. M.; Longo, K. P.; Nagahisa,
A.; Tsuchiya, M. J. Med. Chem. 1994, 37, 2831.
2. Kramer, M. S.; Winokur, A.; Kelsey, J.; Preskorn, S. H.; Rothschild, A. J.; Snavely,
D.; Ghosh, K.; Ball, W. A.; Reines, S. A.; Munjack, D.; Apter, J. A.; Cunningham, L.;
Kling, M.; Bari, M.; Getson, A.; Lee, Y. Neuropsychopharmacology 2004, 29, 385.
3. (a) Keller, M.; Montgomery, S.; Ball, W.; Morrison, M.; Snavely, D.; Liu, G.;
Hargreaves, R.; Hietala, J.; Lines, C.; Beebe, K.; Reines, S. Biol. Psychiatry 2006,
59, 216; (b) Ratti, E.; Bellew, K.; Bettica, P.; Bryson, H.; Zamuner, S.; Archer, G.;
Squassante, L.; Bye, A.; Trist, D.; Krishnan, R.; Fernandes, S. J. Clin.
Psychopharmacol. 2011, 31, 727.
4. (a) Millan, M. J.; Dekeyne, A.; Gobert, A.; Mannoury La Cour, C.; Brocco, M.;
Rivet, J.; DiCara, B.; Lejeune, F.; Cremers, T. I.; Flik, G.; de Jong, T. R.; Olivier, B.;
De Nanteuil, G. Eur. Neuropsychopharmacol. 2010, 20, 599; (b) Guiard, B. P.;
Lanfumey, L.; Gardier, A. M. Curr. Drug Targets 2006, 7, 187; (c) Huang, Y.;
Williams, W. A. Exp. Opin. Ther. Patents 2007, 17, 889; (d) Guiard, B. P.; Froger,
N.; Hamon, M.; Gardier, A. M.; Lanfumey, L. J. Neurochem. 2005, 95, 1713; (e)
Chenu, F.; Guiard, B. P.; Bourin, M.; Gardier, A. M. Behavioural Brain Res. 2006,
172(2), 256; (f) Gobert, A.; Brocco, M.; Dekeyne, A.; Di Cara, B.; Bouchez, G.;
Lejune, F.; Gannon, R. L.; Millan, M. J. Neuropsychopharmacol. 2009, 34, 1039; (g)
Ryckmans, T.; Berton, O.; Grimee, R.; Kogej, T.; Lamberty, Y.; Pasau, P.; Talaga,
P.; Genicot, C. Bioorg. Med. Chem. Lett. 2002, 12, 3195; (h) Ryckmans, T.;
Balancon, L.; Berton, O.; Genicot, C.; Lamberty, Y.; Lallemand, B.; Pasau, P.;
Pirlot, N.; Quere, L.; Talaga, P. Bioorg. Med. Chem. Lett. 2002, 12, 261.
5. Wu, Y.J. et al. Manuscript in preparation.
6. (a) IC50 values were determined by filtration binding assays for both NK1 and
SERT. The NK1 assay utilized 125I-Substance P, and the SERT assay used 125I-RTI-
55 as the competitive inhibitors. Membranes were derived from stably
expressing cultured cells; (b) for synthetic details on the preparation for
compounds disclosed in this publication see: Parker, M. F.; Bronson, J. J.; Silva,
M. V.; Gillman, K. W. U.S. Patent Application Publication 2009096941.
7. Data not shown. We explored a variety of alkyl substituents on the piperidine
nitrogen and all of these significantly decreased SERT binding affinity
compared to the methyl congener.
Figure 3. NK1/SERT ex vivo receptor occupancy in the gerbil forced swim test
(n = 4).
minimized the potential for significant off target cardiovascular
toxicities at pharmaceutically relevant doses. Additional studies
on the progress of compound 44 will be disclosed in due course.
Acknowledgments
The authors thank Drs. Kimberley A. Lentz and Joseph J. Raybon
for helpful discussions on pharmacokinetic data supporting
compound optimization.
8. In vitro microsomal stability for 44 (human/rat/mouse: 99/43/87% remaining
after 10 min @ 0.5
lM conc.); compound 52 (human/rat/mouse: 50/77/7%
remaining after 10 min @ 0.5
9. Wallace-Boone, T. L.; Newton, A. E.; Wright, R. N.; Lodge, N. J.; McElroy, J. F.
Neuropsychopharmacology 2008, 33, 1919.
10. Duffy, R. A.; Varty, G. B.; Morgan, C. A.; Lachowics, J. E. J. Pharmacol. Exp. Ther.
2002, 301, 536.
11. Lengyel, K.; Pieschl, R.; Strong, T.; Molski, T.; Mattson, G.; Lodge, L. J.; Li, Y.
Neuropharmacology 2008, 55, 63.
l
M conc.).
References and notes
1. (a) Mathew, S. J.; Manji, H. K.; Charney, D. S. Neuropsychopharmacology 2008,
33, 2080; (b) Mclean, S. Curr. Pharm. Des. 2005, 11, 1529; (c) Genicot, C.;
Christophe, B.; Collart, P.; Gillard, M.; Goosens, L.; Henichart, J. P.; Lassoie, M.
A.; Moureau, F.; Neuwels, M.; Nicolas, J. M.; Pasau, P.; Quere, L.; Ryckmans, T.;
Stiernet, F.; Taverne, T.; Van Keulen, B. J. Bioorg. Med. Chem. Lett. 2003, 13, 437;
(d) Santarelli, L.; Gobbi, G.; Blier, P.; Hen, R. J. Clin. Psychiatry 2002, 63(11), 11;
(e) Stevenson, G. I.; Huscroft, I.; MacLeod, A. M.; Swain, C. J.; Cascieri, M. A.;
Chicchi, G. C.; Graham, M. I.; Harrison, T.; Kelleher, F. J.; Kurtz, M.;
12. Additional data relating to NK1/SERT% occupancy and how these levels relate to
efficacy in the gerbil forced swim test will be disclosed in the near future.