4-Aminopiperidine ureas as potent selective agonists of the human β3-Adrenergic receptor

Article abstract of DOI:10.1016/S0960-894X(01)00645-X

The preparation and structure-activity relationships (SARs) of potent agonists of the human β₃-adrenergic receptor (AR) derived from a 4-aminopiperidine scaffold are described. Examples combine human β₃-AR potency with selectivity over human β₁-AR and/or human β₂-AR agonism. Compound 29s was identified as a potent (EC₅₀ = 1 nM) and selective (greater than 400-fold over β₁- with no β₂-AR agonism) full β₃-AR agonist with in vivo activity in a transgenic mouse model of thermogenesis.

Full text of DOI:10.1016/S0960-894X(01)00645-X

Bioorganic & Medicinal Chemistry Letters 11 (2001) 3123–3127
4-Aminopiperidine Ureas as Potent Selective Agonists of the
Human ꢀ3-Adrenergic Receptor
Mark A. Ashwell,a,* William R. Solvibile, Jr.,a Stella Han,b Elwood Largis,b
Ruth Mulveyb and Jeffrey Tilletb
aChemical Sciences, Wyeth-Ayerst Research, PO Box CN-8000, USA
bCardiovascular/Metabolic Diseases Research, Wyeth-Ayerst Research, PO Box CN-8000, USA
Received 2 July 2001; accepted 21 September 2001
Abstract—The preparation and structure–activity relationships (SARs) of potent agonists of the human b3-adrenergic receptor
(AR) derived from a 4-aminopiperidine scaffold are described. Examples combine human b3-AR potency with selectivity over
human b1-AR and/or human b2-AR agonism. Compound 29s was identified as a potent (EC50=1 nM) and selective (greater than
400-fold over b1- with no b2-AR agonism) full b3-AR agonist with in vivo activity in a transgenic mouse model of thermogenesis.
# 2001 Elsevier Science Ltd. All rights reserved.
b-Adrenoceptors (b-ARs) were subclassified as b1-AR
and b2-AR in 1967. Increased heart rate is the primary
consequence of b1-AR stimulation, while bronchodila-
tion and smooth muscle relaxation typically result from
b2-AR stimulation. Adipocyte lipolysis (breakdown of
fat) was initially thought to be solely a b1-AR mediated
process. However, results from the early 1980s indicated
that the receptor mediating lipolysis was atypical in
nature. These atypical receptors were designated b3-
ARs and are found on the cell surface of both white and
brown adipocytes where their stimulation promotes
both lipolysis and energy expenditure.1 More recently
the b3-AR has been shown to play a role in the relaxa-
tion of human urinary bladder detrusor tissue.2 Early
compounds with greater agonist activity for the stimu-
lation of lipolysis (b3-AR) than for the stimulation of
atrial rate (b1-AR) and tracheal relaxation (b2-AR)
showed potential as anti-obesity and anti-diabetic
agents in rodents.3
ovary (CHO) cell line expressing the cloned human b3-
AR receptor5 in addition to the human b1- and b2-ARs.
In vivo activity was assessed using a transgenic mouse
model of thermogenesis.6
This paper describes a systematic approach undertaken
to explore the utility of the 4-piperidine scaffold (Fig. 1,
Structures Ia and Ib) in the design of human b3-AR
agonists.
Figure 1. Urea derivatized 4-aminopiperidine template for human b3-
AR agonism.
Disappointing human clinical trials with one such agent
(CL-316243, BTA-243), optimized for the rodent b3-
AR, have been attributed to partial rather than full
agonism at the human b3-AR and poor oral bioavail-
ability.4 A new phase of b3-AR agonist discovery at
Wyeth-Ayerst Research thus utilized a Chinese hamster
Division of the molecule into two pieces on either side of
the secondary alkyl amine suggested either asymmetric
epoxides such as 4 or asymmetric b-amino alcohols such
as 23 as key structural building blocks. Routes to these
are shown in Schemes 1–3. Mitsunobu reaction of com-
mercial R-(+)-glycidol with phenols 37 and 7 (Scheme 1)
provided epoxides 4 and 8,8 respectively. Other more
elaborate epoxides (e.g., 119 and 1410) (Scheme 2) were
provided by alkylation of the corresponding phenols
*Corresponding author at present address: ArQule, Inc., 19 Presidential
Way, Woburn, MA 01801, USA. Fax:+1-781-994-0670; e-mail:
mashwell@arqule.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00645-X

3124
M. A. Ashwell et al. /Bioorg. Med. Chem. Lett. 11 (2001) 3123–3127
t
Scheme 1. (a) BuPh2SiCl, imidazole, CH2Cl2, 82%; (b) Pd/C, cyclo-
hexene, EtOH, 94%; (c) R-(+)-glycidol, PPh3, DEAD, THF, 56%;
t
(d) BuPh2SiCl, imidazole, CH2Cl2, 98%; (e) MCPBA, CHCl3, 85%;
(f) Raney Ni, H2, THF, 73%; (g) mesyl chloride, THF, iPr2NEt, 66%;
(h) Boc2O, CH2Cl2, 96%; (i) NaOH, MeOH, 79%; (j) R-(+)-glycidol,
DEAD, PPh3, 70%.
Scheme 2. (a) b-Ethoxyacryl chloride, Et3N, benzene, 99%; (b) HCl,
78%; (c) Raney Ni, H2 iPrOH, THF, 47%; (d) 48% HBr, heat, 80%;
(e) acetone, water, K2CO3, PhCH2Br, 27%; (f) K2CO3, 2-butanone,
(2S)-(+)-glycidyl-3-nitrobenzenesulfonate, 82 and 32% for 13 to 14;
(g) NaNO2, H2O, HCl (concd); (h) NaOH, 3-hydroxypyridine, 22%
steps g and h; (i) K2CO3, acetone, (2S)-(+)-glycidyl-3-nitrobenzene-
sulfonate, 28%; (j) Raney Ni, ethanol; (k) phosgene, CH2Cl2, 51% for
steps j and k.
with the commercially available (2S)-(+)-glycidyl-3-
nitrobenzenesulfonate.
In the case of epoxide 17, an alternate synthesis to that
previously disclosed11 was developed starting from
trisubstituted phenol 15. Here the alkylation with (2S)-
(+)-glycidyl-3-nitrobenzenesulfonate was performed
prior to the reduction of nitroaniline 16 and ring closure
with phosgene yielding 17 (Scheme 2).
Alternatively, when phenethylamines were the target ago-
nists epoxides such as 20 were constructed or purchased
[(R)-(+)-3-chlorostyrene oxide for the preparation of
37a]. An alternative strategy required chiral amino
alcohol 23, and this was prepared as shown in Scheme 3.
The preparation of the scaffold piperidine 27 is shown
in Scheme 4. Selective protection of the primary amine
of 24 as the BOC amide was followed by reductive
amination with 4-benzylpiperidinone to yield 26.
Removal of the benzyl protecting group gave the sec-
ondary amine 27.
Scheme 3. (a) PhCH2Br, NaOMe, MeOH, 80%; (b) Br2, CHCl3, 87%;
(c) NaBH4, ethanol, THF, 96%; (d) K2CO3, 2-butanone, 74%; (e)
MeSO2Cl, pyridine, CH2Cl2, 47%; (f) CuBr2, CHCl3, 57%; (g) (R)-2-
methyl-CBS-oxazaborolidine, BH3, THF, 61%; (h) NaN3, DMSO;
(i) H2, Pd/C, MeOH, 71% for steps h and i.
Reaction of 27 with isocyanates or with triphosgene
followed by an amine provided ureas 28 depending on
reagent availability. Following purification and removal
of the BOC group with formic acid the formate salt 28
was used directly or treated with NaOH to liberate the
free base.
eration of the aldehyde 38 and subsequent reductive
amination provided phenethylamines such as 37c.12
Ureas 35 proved more difficult to prepare. The instabil-
ity of the intermediate nitro aldehyde, generated by
oxidation of 30, could be avoided if the Dess–Martin
oxidation was followed immediately by protection as the
dimethyl acetal to give 31 as shown in Scheme 5. Reduc-
tion of the nitro group of 31 provided aniline 32. Ureas
35 were obtained from 32 as described for 27 above.
The in vitro data for a selection of simple alkyl ureas
based on the 4-aminopiperidine scaffold is presented in
Table 1. In general molecules of this type (29a–f) have
comparable human b3-AR agonism relative to the stan-
dard isoproterenol, considerably weaker b2-AR agon-
ism and some selectivity over b1-AR agonism. The
greatest receptor subtype selectivity is found when the
alkyl group is an extended straight chain (cf. 29a and
29b). Importantly the potency of b1-AR agonism
appears to fall offas the chain length increases.
Intermediate ureas 28 were reacted directly with epox-
ides in the presence of a hindered organic base, thus
providing access to either aryloxypropanolamines or
phenethylamines as shown in Scheme 6. This reaction
was not regiospecific and access to the starting epoxides
was not always feasible. An alternative approach is
illustrated in the preparation of 37c. Here, in situ gen-
An extensive selection of aryloxypropanolamines was
prepared holding the alkyl urea portion of 29a constant.
The importance of the 4-hydroxy group of 29a was
demonstrated by its removal to give 29g which is an

M. A. Ashwell et al. /Bioorg. Med. Chem. Lett. 11 (2001) 3123–3127
3125
Scheme 4. (a) BOC2O, CH2Cl2, 92%; (b) 4-benzylpiperidinone,
Na(OAc)3BH, AcOH, Na2SO4, 90%; (c) Pd/C, cyclohexene, ethanol,
96%; (d) isocyanate, THF (75–90%) or amine, triphosgene, CH2Cl2,
(50–80%); (e) formic acid (80–95%).
Scheme 5. (a) Dess–Martin oxidation, CH2Cl2; (b) trimethyl ortho-
formate, PTSA, methanol, 89% for steps a and b; (c) Pd/C, ammo-
nium formate, ethanol, 76%; (d) 4-benzylpiperidinone, Na(OAc)3BH,
AcOH, Na2SO4, 90%; (e) isocyanate, THF (75–90%) or amine, tri-
phosgene, CH2Cl2, (50–80%); (f) formic acid, (80–95%).
Scheme 6. (a) Aryloxy epoxide, ethanol, iPr2NEt, heat (30–75%);
(b) ethanol, Pr2NEt, heat, 16%; (c) H2 Pd/C, ethanol, 5% for steps b
and c (20–37%); (d) NaI, MeSiCl3, CH3CN; (e) MeOH, NaCNBH3,
AcOH, 7% for steps d and e.
i
Table 1. Agonist activity at cloned human b-ARs of alkyl urea substituted piperidines
Compd
Ar
R1
R2
b3 EC50 (nM) IA (%)a
b1 EC50 (nM) IA (%)a
b EC50 (nM) IA (%)a
29a
29b
29c
29d
29e
29f
29g
4-OHPh
4-OHPh
4-OHPh
4-OHPh
4-OHPh
4-OHPh
4-FPh
Octyl
Me
H
iPr
Chexl
Et
Octyl
H
H
H
H
H
8 (>100)
47 (85)
30 (100)
90 (105)
80 (112)
36 (96)
37%@10 mM
610 (53)
350 (106)
511 (77)
943 (54)
85 (68)
164 (74)
ndb
iac
iac
iac
iac
77 (17)
iac
ndb
ethyl
H
29h
29i
29j
Octyl
Octyl
Octyl
H
H
H
290 (68)
ia@0.1mM
66 (67)
ndb
ndb
ndb
ndb
88 (60)
ndb (16)
37a
3-ClPh
Octyl
Octyl
H
H
1020 (77)
ndb
ndb
ndb
ndb
37b (racemic)
42%@10 mM
ab-ARs agonistic activities were assessed by measurement of cAMP accumulation levels in CHO cells expressing human b-ARs; the intrinsic
activities (IA) are given as a percentage of maximal stimulation relative to isoproterenol.
bnd, not determined.
cia, inactive.

3126
M. A. Ashwell et al. /Bioorg. Med. Chem. Lett. 11 (2001) 3123–3127
Table 2. Agonist activity at cloned human b-ARs of urea substituted piperidines
Compd
Ar
R1
R2
H
b3 EC50 (nM) IA (%)a
b1 EC50 (nM) IA (%)a
b2 EC50 (nM) IA (%)
29k
4-OHPh
50 (90)
187 (27)
iac
29l
4-OHPh
4-OHPh
4-OHPh
4-OHPh
4-OHPh
4-OHPh
4-Fbenzyl
2,4-diFbenzyl
2,4-diClbenzyl
2-Fbenzyl
2,6-diFbenzyl
2,5-diFbenzyl
H
H
H
H
H
H
41 (93)
30 (84)
250 (91)
37 (91)
32 (104)
23 (92)
109 (63)
460 (106)
254 (56)
260 (36)
750 (75)
522 (48)
iac
iac
iac
iac
iac
iac
29m
29n
29o
29p
29q
29r
29s
37c
Octyl
H
H
H
200 (99)
1 (100)
5 (100)
ndb (22)
423 (75)
2640 (78)
iac
2,5-diFbenzyl
2,5-diFbenzyl
iac
10 (110)
aSee footnotes for Table 1.
bSee footnotes for Table 1.
cSee footnotes for Table 1.
essentially inactive compound. Similarly more complex
replacements (29h–j) also significantly reduced activity.
Phenethylamines (37a,b) were prepared but offered no
improvement.
NHSO2Me substituent was introduced adjacent to
the phenol. Importantly, switching to the 2,5-difluoro
capped urea and the corresponding phenethylamine
provided the potent b3-AR agonist 37c. However, little
selectivity over b2-AR agonism was observed.
The induction of thermogenesis in human b3-AR trans-
genic mice (Tg mice) by 29a was compared with its
ability to induce thermogenesis in b3-AR knock-out
mice (KO mice). This assay was employed as an in vivo
measure of the potential of these agonists to treat or
inhibit disorders related to obesity or type II diabetes.
Disappointingly 29a had weak in vivo activity in the
human b3-AR transgenic mouse (11%Æ8% at 10 mg/
kg ip). One possible reason for the low in vivo activity
was suggested by Phase I metabolism studies using iso-
lated P450 rat microsomes. It was demonstrated in this
assay that the long alkyl chain was a site of oxidation. A
number of capped alkyl ureas were prepared in order to
overcome this liability. In the case of the cyclopentyl-
propyl urea 29k the b3-AR agonism was reduced com-
pared to 29a and no improvement in in vivo activity was
observed.
Significantly, when the ethanolamine portion of 37c was
replaced with the corresponding aryloxypropanolamine
to give 29s, potent b3-AR agonism (EC50=1 nM) was
combined with high selectivity over both b1-AR and
b2-AR. Compound 29s was also demonstrated to be
a selective human b3-AR agonist in vivo (22%Æ7% at
10 mg/kg ip) and thus shows promise as an agent for
the treatment of obesity, type II diabetes and frequent
urination.
In conclusion 4-aminopiperidine has been demonstrated
to be a viable scaffold for the preparation of human b3-
AR agonists. Optimization of piperidine alkyl ureas
derived from this structural element led to the identifi-
cation of many in vitro selective b3-AR agonists. In vivo
results in a b3 transgenic mouse of thermogenesis and
supporting data from in vitro P450 microsomal studies
led to the identification of aromatic capped ureas. The
selectivity of these aromatic capped ureas was improved
by varying the substituents and substitution pattern
around the phenyl ring. The resulting 29s was found to
be a potent human b3-AR full agonist (EC50=1 nM)
with greater than 400-fold selectivity over b1-AR and no
b2-AR agonism. In addition 29s was active at a dose of
10 mg/kg in the acute in vivo b3 transgenic mouse model
of thermogenesis.
A more successful approach is illustrated with 29l. Here
a robust and selective in vivo response (46%Æ4% at
10 mg/kg ip) was measured. However, there has been a
decrease in both b3-AR agonism and selectivity. Sys-
tematic variations of the substituents on the phenyl ring
led to the identification of 29q (Table 2). Here the 2,5-di
fluoro substituent pattern provided a b3-AR agonist
with over 20-fold selectivity against b1-AR partial
agonism and devoid of b2-AR agonism.
Further Phase I metabolic studies highlighted the 4-
hydroxyphenyloxy portion of the molecule as a poten-
tial liability and a further round of optimization was
undertaken. Firstly, comparing 29a with 29r it is clear
that there is a significant loss of activity when the
References and Notes
1. For a recent review, see: Weyer, C. J.; de Souza, C. Drug
Dev. Res. 2000, 51, 80.

M. A. Ashwell et al. /Bioorg. Med. Chem. Lett. 11 (2001) 3123–3127
3127
2. Takeda, M.; Obara, K.; Mizusawa, T.; Tomita, Y.; Arai,
K.; Tsutsui, T.; Hatano, A.; Takahashi, K.; Nomura, S. J.
Pharm. Exp. Ther. 1999, 288, 1367.
3. Largis, E. E.; Burns, M. G.; Muenkel, H. A.; Dolan, J. A.;
Claus, T. H. Drug Dev. Res. 1994, 32, 69.
8. For a related synthesis, see: Beeley, L. J.; Thompson, M.;
Dean, D. K.; Kotecha, N. R.; Berge, J. M.; Ward, R. W. PCT.
Int. Appl. WO 9604233, 1996. Chem. Abstr. 1996, 125, 58092.
9. For a synthesis of the racemic epoxide see: Tominaga, M.;
Tone, H.; Nakagawa, K.; Takada, K.; Hoshino, Y.; Wata-
nabe, K. Chem. Pharm. Bull. 1981, 29, 2166.
10. Fisher, M. H.; Parmee, E. R.; Mathvink, R. J.; Weber,
A. E.; Ok, H. O. Eur. Pat. Appl. EP 611003, 1994. Chem.
Abstr. 1994, 121, 300591.
11. Elsinga, P. H.; van Waarde, A.; Jaeggi, K. A.; Schreiber, G.;
Heldoorn, M.; Vaalburg, W. J. Med. Chem. 1997, 40, 3829.
12. For full experimental details, see: Ashwell, M. A.;
Solvibile, W. R.; Molinari, A. J.; Quagliato, D. PCT appli-
cation, submitted July 2000.
4. Weyer, C.; Tataranni, P. A.; Snitker, S.; Danforth, E., Jr.;
Ravussin, E. Diabetes 1998, 47, 1555.
5. Lelias, J. M.; Kaghad, M.; Rodriguez, M.; Chalon, P.;
Bonnin, J.; Dupre, I.; Delpech, B.; Bensaid, M.; LeFur, G.;
Ferrara, P.; Caput, D. FEBS Lett. 1993, 324, 127.
6. Susulic, V. S.; Frederich, R. C.; Lawitts, J.; Tozzo, E.;
Kahn, B. B.; Harper, M.-E.; Himms-Hagen, J.; Flier, J. S.;
Lowell, B. B. J. Biol. Chem. 1995, 270, 29483.
7. Stern, A.; Swenton, J. S. J. Org. Chem. 1987, 52, 2763.