Serendipity in drug-discovery: A new series of 2-(benzyloxy)benzamides as TRPM8 antagonists

Article abstract of DOI:10.1016/j.bmcl.2013.09.016

A new series of 2-(benzyloxy)benzamides are presented that are potent functional antagonists of TRPM8 and possess improved LipE and LE compared to the original lead. They were discovered through a series of compound libraries and we present a powerful visualization method for the chemical space explored with each library. Remarkably this new series originated from the highest risk design strategy where compounds were synthesised with the least degree of similarity to the lead structure.

Full text of DOI:10.1016/j.bmcl.2013.09.016

Bioorganic & Medicinal Chemistry Letters xxx (2013) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Serendipity in drug-discovery: A new series
of 2-(benzyloxy)benzamides as TRPM8 antagonists
a
b,
Alan Brown a, , David Ellis , David A. Favor , Tony Kirkup c, , Wolfgang Klute d, , Malcolm MacKenny e,
,
⇑
Gordon McMurray f, Adam Stennett g,
a Worldwide Medicinal Chemistry, Pfizer Neusentis, Granta Park, Great Abington, Cambridge CB21 6GS, UK
b Shanghai ChemPartner, Shanghai City, China
c St Anslem’s RC Secondary School, Canterbury, Kent, UK
d Astra Zeneca, Vastra Gotaland County, Sweden
e Peakdale Molecular, Sandwich, Kent, UK
f Pfizer Neusentis, Granta Park, Great Abington, Cambridge CB21 6GS, UK
g The Kings School, Canterbury, Kent, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
A new series of 2-(benzyloxy)benzamides are presented that are potent functional antagonists of TRPM8
and possess improved LipE and LE compared to the original lead. They were discovered through a series
of compound libraries and we present a powerful visualization method for the chemical space explored
with each library. Remarkably this new series originated from the highest risk design strategy where
compounds were synthesised with the least degree of similarity to the lead structure.
Ó 2013 Elsevier Ltd. All rights reserved.
Received 31 July 2013
Revised 2 September 2013
Accepted 5 September 2013
Available online xxxx
Keywords:
TRPM8
Library design
Multi-dimensional scaling
Lead hopping
1,3-Diamino cyclohexane
TRP (transient receptor potential) channels belong to a large
family of nonselective cation channels that function in a variety
of processes, including temperature sensation.1 Members of the
TRP family have been shown to be multifunctional receptors that
can be activated by physical (heat, cold, voltage, and stress), chem-
ical (osmolality, pH) and specific ligand challenges. Our interest in
TRPM82 stems from its presence in bladder urothelium and
sensory nerves and the potential that modulation of it will have
a beneficial effect in the treatment of urgency and frequency asso-
ciated with overactive bladder.3
Although TRPM8 has no known endogenous ligand, there are a
number of reported small molecules that modulate the activity of
this ion channel. These can either activate (agonise) or deactivate
(antagonize) the ion channel with respect to an external stimuli,
such as cold. A naturally occurring compound that activates the
TRPM8 ion channel is menthol.4 A number of other agonists have
also been described (Fig. 1). These include the menthol derived
WS-12 and icilin.1b,5
Figure 1. Agonists of TRPM8.
Upon initiating our work in this area Bayer had reported on the
discovery of potent small molecular antagonists of TRPM8
(Fig. 2).6 Subsequently, Janssen, Amgen and others7 also disclosed
additional classes of antagonists. Researchers at GlaxoSmithKline
recently reported the in vivo activity in rats of a 2-(benzyloxy)benz-
amide based TRPM8 antagonist (AMTB, 3, described in a Bayer pat-
ent) for overactive bladder and painful bladder syndrome.8 It is this
series, and more specifically analogue 4, that served as the starting
point for the work described in this communication.
At the outset of this work our key goals were to improve the
overall ligand efficiency (LE)9 and lipophilic efficiency (LipE)10 of
compound 4. We were also keen to replace the electron rich thio-
phene and the diamine moiety in 4, since the potential metabolic
instability of the thiophene group and the possibility for internal
⇑
Corresponding author.
E-mail addresses: Alan.D.Brown@pfizer.com (A. Brown), dellis44781@gmail.com
(D. Ellis).
Current address.
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.09.016
Please cite this article in press as: Brown, A.; et al. Bioorg. Med. Chem. Lett. (2013), http://dx.doi.org/10.1016/j.bmcl.2013.09.016

2
A. Brown et al. / Bioorg. Med. Chem. Lett. xxx (2013) xxx–xxx
Scheme 1. Reagents and conditions: (a) MeOH, 30 °C, 3 h; NaBH4, 30 °C, 16 h, then
CH2Cl2, NEt3, 30 °C, 16 h; (b) NaHCO3, MeOH, 50 °C; NaBH4, rt 16 h, then HATU,
DIEA DMA, rt, 16 h; (c) HATU, NEt3, DMF, 60 °C, 16 h or HOBt, EDC, NEt3, DMF, 30 °C,
16 h; (d) TFA, CH2Cl2, 30 °C, 1 h (if applicable).
Figure 2. Reported TRPM8 antagonist (Pfizer generated data).
rearrangement of the ethylene diamine were seen as additional lia-
bilities in this lead series.11
polar and basic residues, derived from primary and secondary
amines. Since compounds from library 3 showed much lower simi-
larity to 4 than the previous strategies, we decided to maximize our
chances of success by increasing the size of the library to 436 mem-
bers. Library compounds were selected from a very large set of com-
mercial and proprietary amines, applying simple cut-offs of clogP
<5 and MW <500, and favoring the presence of a basic side chain.
The chemistry to synthesize these libraries was straightforward
and employed well established amide bond formations, reductive
aminations and deprotections of tert-butoxycarbonyl (Boc) groups
(Scheme 1).
Initially we aimed to fully develop the SAR around the basic side
chain and the thiophene residue of compound 4. We hoped to
make rapid progress towards compounds with improved proper-
ties within a close similarity space around 4; however owing to
the complexity of the issues identified, we also decided to engage
in a more risk-taking lead-hopping exercise with more speculative
structural changes.
Accordingly we synthesised compound libraries evolving from
three design strategies with descending degrees of similarity to-
wards salicylamide 4 (Fig. 3).
In order to visualize a similarity distribution of all library com-
pounds in 2D, we generated a distance matrix for all compounds
based on their structural similarity, followed by a multi-dimen-
sional scaling (MDS) analysis as a dimension reducing technique
(Fig. 4a).12 MDS proved to be an ideal tool to represent the chem-
ical space occupied by libraries following different hypotheses,
with compounds from ‘close-in’ library 1 appearing within the
proximity of the probe and those from ‘lead-hopping’ libraries 2
and 3 occupying distal areas within the MDS space.
Screening results of these libraries are displayed in Figure 4b
using identical MDS coordinates.13 Since the ECFP-4 fingerprints12
used for the similarity calculations have been previously suggested
to correlate with biological activity, one would expect compounds
occurring within close proximity on the MDS plot to have a high
likelihood of similar biological activity.14 Indeed we found that
compounds from library 1, with relatively high similarities
In library 1 the thiophene moiety was varied with more polar
residues derived from aromatic and heteroaromatic aldehydes
(Scheme 1). A total of 130 compounds were synthesized using pre-
dominantly clogP cut-offs (clogP range from 1.5–5.0).
Library 2 replaced the ethylamine unit with a wider selection of
216 cyclic and acyclic residues, using primary amines as a
resourceful source of monomers. The majority of R groups con-
tained an additional basic center; however we also investigated
whether low molecular weight (MW) polar and neutral side chains
would be tolerated.
Unlike the previous examples, library 3 adopted a lead-hopping
approach, which we hoped would lead us into a new, structurally
distinct series. The entire amide N-substituent (thiophene and
ethylamine) was replaced with a wide range of structurally diverse
Figure 4a. 2D MDS derived similarity distribution of libraries according to design
Figure 3. High, medium and low risk library strategies around 4.
strategy.
Please cite this article in press as: Brown, A.; et al. Bioorg. Med. Chem. Lett. (2013), http://dx.doi.org/10.1016/j.bmcl.2013.09.016

A. Brown et al. / Bioorg. Med. Chem. Lett. xxx (2013) xxx–xxx
3
Figure 4b. 2D MDS derived similarity distribution of libraries and TRPM8 activity.
Scheme 2. Reagents and conditions: (a) NiCl2:6H2O, 82%; (b) HCl, 49%; (c) 1 equiv
NaOH, 65%; (d) Boc2O, 65%; (e) NaBH4, AcOH, CHCl3, 33 °C, 64%; (f) o-ben-
zyloxybenzoyl chloride, DIEA, CH2Cl2, 76%; (g) HPLC separation; (h) TFA, CH2Cl2,
97%.; (i) H2/Pd/C, dihydrotoluene EtOH, quant.; (j) p-bromobenzyl bromide, K2CO3,
DMF, rt, 70%.
Figure 4c. Percentage distribution of TRPM8 activity by design strategy.
Figure 5. Asymmetric unit of the 16 (ethyl acetate) crystal structure.
towards 4, on average showed the highest likelihood of activity
(also see Fig. 4c, as measured by the number of compounds with
activities <1
l
M). However, the level of potencies never signifi-
To determine the absolute configuration of the active cyclohex-
ane diamine, a p-brombenzylether analog of 11 (from the less
active series) was prepared and crystallized in two steps from 11
(Scheme 2). The absolute configuration of 16 was established as
1R,3S by single crystal X-ray crystallography (Fig. 5; please note
that the numbering priority changes when the primary amine is
protected with a Boc group). By inference the configuration of
the dicyclohexyl amine of the series derived from compound 12
is the opposite to that of 11.
The individual enantiomers (13 and 14) showed a ꢀ20fold dif-
ference in affinity for TRPM8, with the (1R,3S)-3-aminocyclohexyl
enantiomer having the higher affinity. Compound 14 demon-
strated an improved LE (0.38 from 0.31) compared to 4 and a
significantly improved LipE (4.31–6.16). The racemic trans isomer
17 was >25fold less potent than the more potent enantiomer 14.
None of the other close-in modifications of the (1R,3S)-3-amin-
ocyclohexyl moiety afforded compounds with an affinity of
cantly exceeded that of probe compound 4 or the AMTB analogue
5. A sharp drop-off of average activities was observed for com-
pounds from library 2.
Compounds derived from lead hopping library 3 on average
showed very poor activities, as one would expect based on their
low similarity to lead compound 4. However there was one notable
exception: Compound 6 derived from N-ethyl-cis-1,3-diaminocy-
clohexane displayed significantly improved potency (Kb
<100 nM)15 versus 4. Since it was also lacking the undesirable
thiophene and ethylene diamine moieties, further efforts focused
on this new series.
In order to confirm the activity of 6, the individual enantiomers
were synthesised starting from commercially available cis/trans
1,3-diaminocyclohexane 7 (Scheme 2). The cis isomer 8 was
isolated by means of its insoluble nickel complex, formed from
NiCl2:6H2O in water.16
Selective mono-protection of 8 with Boc group was achieved
using a variation of a literature procedure,17 where the di-HCl salt
was treated with 1 equivalent of NaOH to afford the mono HCl salt
in situ. Mono-ethylation of 9 was achieved via a little used method
from the aminoglycoside field.18,19 Acylation of 10 with o-ben-
zyloxybenzoyl chloride afforded the 2-(benzyloxy)benzamide in
racemic form. At this point the individual enantiomers 11 and 12
were separated via chiral HPLC. Removal of the Boc group with
TFA yielded 13 and 14.20
<2
l
M (Fig. 6).
The SAR around the N-ethyl group was investigated next. Ini-
tially a small range of rac-N-alkyl-cis-1,3-diaminocyclohexanes
were synthesised via N-alkylation of the racemic diamine 9. How-
ever due to poor yields obtained, we used an alternative method
where 3121 was reacted in a reductive amination with primary
amines to form the cis-isomer with good selectivity.22 In a typical
example the reaction with cyclopropylamine afforded
mixture of cis/trans isomers; the cis-isomer 32 (Scheme 3) could
a 9:1
Please cite this article in press as: Brown, A.; et al. Bioorg. Med. Chem. Lett. (2013), http://dx.doi.org/10.1016/j.bmcl.2013.09.016

4
A. Brown et al. / Bioorg. Med. Chem. Lett. xxx (2013) xxx–xxx
Figure 7. SAR of 2-(benzyloxy)benzamide moiety.
Table 2
SAR of benzyl substitutions
Compound
R
Chirality
TRPM8:Kb (nM)
TRPM8:IC50 (nM)
Figure 6. SAR of N-ethyl-cis-1,3-diaminocyclohexane based analogs.
14
47
48
50
51
52
53
54
55
56
57
58
59
H
1R,3S
1R,3S
1S,3R
1R,3S
1S,3R
1R,3S
1S,3R
Racemic
1R,3S
56.3
12.7
246
17.8
367
13
76.2
15.1
291
21.1
435
29.9
85.4
96.9
19.3
254
o-Me
o-Me
o-F
o-F
o-Cl
o-Cl
m-Cl
p-Cl
p-Cl
o-CN
m-CN
p-CN
72
96.6
16.1
214
1250
491
4590
1S,3R
Racemic
Racemic
Racemic
1850
493
6780
Scheme 3. Reagents and conditions: (a) 29a,b: alkyliodide, CsOH, DMF, 29c:
cyclobutanone, NaBH(OAc)3, THF, AcOH; (b) o-benzyloxybenzoyl chloride, DIEA,
CH2Cl2, then TFA, CH2Cl2; (c) t-BuOCONH2, Bi(NO3)3, CH2Cl2, 55%; (d) (1) cyclopro-
pylamine, MeOH, rt, (2) LiBH4, À78 °C, (3) oxalic acid, recrystallisation MeOH/EtOAc
1:2, 37%.
SAR of the 2-(benzyloxy)benzamide was explored with a struc-
turally diverse set of carboxylic acids. A representative set of some
of the analogs made are shown in Figure 7. Of all the compounds
made, only the isomeric analog 43 had potency of <2 lM.
Substitution of the benzyl group was the only modification
investigated that either retained or improved the potency of the
series. Selected compounds are shown in Table 2, with o-methyl,
o-fluoro, o-chloro and p-chloro substituents on the more potent
core affording compounds (47, 50, 52, 55) with 3–4fold improved
potencies over 14. Again, the opposite enantiomers displayed sig-
nificantly weaker activities. Cyano substitutions resulted in a sig-
nificant loss of potency.
Table 1
SAR of N-ethyl substitution of cis-1,3-diamino-cyclohexane based analogs
As detailed in summary Table 3, the most potent compounds
displayed significantly improved LipE over 4 and very low intrinsic
clearance in a human liver microsome assay (HLM Clint app). Com-
Compound
R
TRPM8:Kb (nM)
TRPM8:IC50 (nM)
33
34
35
36
37
38
39
Me
Pr
c-Butyl
c-Propyl
iso-Propyl
n-Bu
3670
1430
794
2250
5130
1560
665
3150
1230
1360
2010
8780
1400
593
Table 3
Properties of the most potent compounds
CH2-c-Pr
Compound
4
14
50
55
TRPM8 IC50 (nM)
TRPM8 Kb (nM)
LogD(7.4)
LipE
LE
1240
443
1.60
4.31
0.32
149
76.2
56.0
0.60
6.65
0.38
<8
21.1
17.8
0.70
6.98
0.40
<8
19.3
16.1
1.30
6.41
0.40
<10
subsequently be obtained in 98% purity after re-crystallisation of
its oxalate salt.
The SAR with respect to the N-ethyl group again proved to be
tight with the best (39) being ꢀ10fold less potent than 14
(Table 1).
HLM CLint, app (lL/min/mg)
LipE and LE were calculated based on IC50 values.
Please cite this article in press as: Brown, A.; et al. Bioorg. Med. Chem. Lett. (2013), http://dx.doi.org/10.1016/j.bmcl.2013.09.016

A. Brown et al. / Bioorg. Med. Chem. Lett. xxx (2013) xxx–xxx
5
Petrocellis, L.; Morera, L.; Moriello, A. S.; Orlando, P.; Morera, E.; Nalli, M.; Di
Marzo, V. Bioorg. Med. Chem. Lett. 2010, 20, 2729.
8. Lashinger, E. S. R.; Steiginga, M. S.; Hieble, J. P.; Leon, L. A.; Gardner, S. D.;
Nagilla, R.; Davenport, E. A.; Hoffman, B. E.; Laping, N. J.; Su, X. Am. J. Physiol.
2008, 295, F803 (3, Pt. 2).
9. (a) Kuntz, I. D.; Chen, K.; Sharp, K. A.; Kollman, P. A. Proc. Natl. Acad. Sci. U.S.A.
1999, 96, 9997; (b) Hopkins, A. L.; Groom, C. R.; Alex, A. Drug Discovery Today
2004, 9, 430.
10. (a) Ryckmans, T.; Edwards, M. P.; Horne, V. A.; Correia, A. M.; Owen, D. R.;
Thompson, L. R.; Tran, I.; Tutt, M. F.; Young, T. Bioorg. Med. Chem. Lett. 2009, 19,
4406; (b) Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug Disc. 2007, 6, 881.
11. The conversion of BMAC (2) to i was observed in DMSO and aqueous buffer at a
range of pH’s with a half life of <2 h
pound 14 was also screened in a CEREP bioprint screen (against a
wide range [>80] of receptors, enzymes and ion channels) and only
rat calcium channel L-site diltiazem had inhibition of >60% (61%) at
10
lM.
In conclusion, we have demonstrated how a lead-hopping de-
sign strategy allowed us to replace an unattractive side-chain in
4 with a novel 1,3-diamino cyclohexane residue. Compounds from
this series are potent functional antagonists of TRPM8 and have
improved LiPE and intrinsic clearance compared to the original
lead. Clearly the success of this strategy was highly dependent
on employing a quality monomer selection, allowing us to cover
a wide area of chemical space in our lead-hopping library.
H
MeO
O
NH2
MeO
O
N
DMSO
or
S
N
NH
Acknowledgments
O
O
i
aq. buffer
S
We thank Mark Gardner for bringing to our attention 2D MDS
plots as a method for visualizing the chemical space of libraries
with different design hypotheses. We acknowledge Cheryl Doherty
for solving the X-ray structure and determining the absolute con-
figuration of 16 and Laure Hitzel for carrying out the HPLC separa-
tion of 11 and 12.
.
12. Similarities were calculated using ECFP-4 fingerprints, MDS x-coordinates were
calculated in Scitegic Pipeline Pilot™ using a ‘classical’ CMDS approach based
on R statistics components available in Pipeline Pilot™. For more information
see: Rodgers, D.; Hahn, M. J. Chem. Inf. Model. 2010, 50, 742.
13. All TRPM8 data was generated in a HEK cell line transfected with a human
TRPM8 receptor construct. TRPM8 inhibitory activity was measured by
monitoring WS12 induced elevation of [Ca2+] using FLIPRÒ detection. The
geometric mean of the agonist EC50 and slope were used to generate modified
Kb values in addition to IC50 values. Experiments were carried out in the
presence of the activating agonist WS12 at a final concentration of 350 nM. All
References and notes
1. (a) Tsavaler, L.; Shapero, M. H.; Morkowski, S.; Laus, R. Cancer Res. 2001, 61,
3760; (b) McKemy, D. D.; Neuhausser, W. M.; Julius, D. Nature 2002, 416, 52; (c)
Peier, A. M.; Moqrich, A.; Hergarden, A. C.; Reeve, A. J.; Andersson, D. A.; Story,
G. M.; Earley, T. J.; Dragoni, I.; McIntyre, P.; Bevan, S.; Patapoutian, A. Cell 2002,
108, 705; (d) Ferrer-Montiel, A.; Fernández-Carvajal, A.; Planells-Cases, R.;
Fernández-Ballester, G.; González-Ros, J. M.; Messeguer, A.; González-Muñiz, R.
Expert Opin. Ther. Pat. 2012, 22, 999; DeFalco, J.; Duncton, M. A.; Emerling, D.
Curr. Top. Med. Chem. 2011, 11, 2237.
reported TRPM8 data is based on
experiments.
a minimum of three independent
14. Muchmore, S. W.; Debe, D. A.; Metz, J. T.; Brown, S. P.; Martin, Y. C.; Hajduk, P. J.
J. Chem. Inf. Model. 2008, 48, 941.
15. Original sample prepared via
a library was not tested at low enough
2. Voets, T.; Nilius, B. J. Membr. Biol. 2003, 192, 1.
concentrations to determine an accurate Kb. A few other compounds from
library three showed activities of <500 nM, however due to their inferior LE
they were not pursued further.
3. See for example: Leon, L. A.; Gardner, S. D.; Nagilla, R.; Davenport, E. A.;
Hoffman, B. E.; Laping, J. N.; Su, X. Am. J. Physiol. Renal Physiol. 2008, 295, F803.
4. Behrendt, H.-J.; Germann, T.; Gillen, C.; Hatt, H.; Jostock, R. Br. J. Pharmacol.
2004, 141, 737.
5. Beck, B.; Bidaux, G.; Bavencoffe, A.; Lemonnier, L.; Thebault, S.; Shuba, Y.;
Barrit, G.; Skryma, R.; Prevarskaya, N. Cell Calcium 2007, 41, 285.
6. (a) Lampe, T.; Alonso-Alija, C.; Stelte-Ludwig, B.; Sandner, P.; Bauser, M.; Beck,
H.; Lustig, K.; Rosentreter, U.; Stahl, E.; Takagi, H. WO2006040136.; (b) Lampe,
T.; Alonso-Alija, C.; Beck, H.; Rosentreter, U.; Sandner, P.; Stahl, E.; Stelte-
Ludwig, B. WO2007017093.; (c) Lampe, T.; Alonso-Alija, C.; Beck, H.;
Rosentreter, U.; Sandner, P.; Stahl, E.; Stelte-Ludwig, B. WO2007017094.
7. (a) Parks, D. J.; Parsons, W. H.; Colburn, R. W.; Meegalla, S. K.; Ballentine, S. K.;
Illig, C. R.; Qin, N.; Liu, Y.; Hutchinson, T. L.; Lubin, M. L.; Stone, D. J., Jr.; Baker, J.
F.; Schneider, C. R.; Ma, J.; Damiano, B. P.; Flores, C. M.; Player, M. R. J. Med.
Chem. 2011, 54, 233; (b) Norman, M. H.; Bo, Y. Y.; Gore, V. Keshav; Horne, D.;
Kaller, M.; Ma, V. V.; Monenschein, H.; Nguyen, T.; Nishimura, N.; Tamayo, N.
WO2009073203.; (c) DeFalco, J.; Steiger, D.; Dourado, M.; Emerling, D.;
Duncton, M. A. J. Bioorg. Med. Chem. Lett. 2010, 20, 7076; (d) Ortar, G.; De
16. Munk, V. P.; Cham, S. T.; Fenton, R. R.; Hocking, R. K.; Hambley, T. W. Aus. J.
Chem. 2002, 55, 523.
17. Lee, D. W.; Ha, H.-J.; Lee, W. K. Synth. Comm. 2007, 37, 737.
18. (a) Salvatore, R. N.; Nagle, A. S.; Jung, K. W. J. Org. Chem. 2002, 67, 674; (b)
Sajiki, H.; Ikawa, T.; Hirota, K. Org. Lett. 2004, 6, 4977.
19. Nam, G.; Kim, S. H.; Kim, J.-H.; Shin, J.-H.; Jang, E.-S. Org. Process Res. Dev. 2002,
6, 78.
20. Separation of 11 and 12 was achieved on Chiralpak IC column (21.2 mm id)
using Heptane/IPA (70:30) as eluent. Spectroscopic data of 14: 1H NMR
(400 MHz, D6-DMSO): d = 8.02 (br, 2H); 7.92 (br, 1H); 7.26–7.40 (m, 6H); 7.13
(m, 2H); 6.96 (m, 1H); 5.16 (m, 2H); 4.11 (m, 1H); 3.68 (m, 1H); 3.04–3.30 (m,
2H); 2.03–1.68 (m, 8H); 0.87 (m, 3H); ES+ve MS m/z: 353 (MH+).
21. Glatthar, R.; Troxler, T. J. WO2006114260.
22. Cabral, S.; Hulina, B.; Kawai, M. Tetrahedron Lett. 2007, 48, 7134.
Please cite this article in press as: Brown, A.; et al. Bioorg. Med. Chem. Lett. (2013), http://dx.doi.org/10.1016/j.bmcl.2013.09.016