Arylpropanolamines: Selective β3 agonists arising from strategies to mitigate phase I metabolic transformations

Article abstract of DOI:10.1016/j.bmcl.2007.05.030

Utilization of N-substituted-4-hydroxy-3-methylsulfonanilidoethanolamines 1 as selective β₃ agonists is complicated by their propensity to undergo metabolic oxidative N-dealkylation, generating 0.01-2% of a very potent α₁ adrenergic

Full text of DOI:10.1016/j.bmcl.2007.05.030

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4290–4296
Arylpropanolamines: Selective b3 agonists arising from strategies
to mitigate phase I metabolic transformations
W. N. Washburn,* T. W. Harper, G. Wu, J. D. Godfrey, P. McCann, R. Girotra, C. Shao,
H. Zhang, A. Gavai, A. Mikkilineni, T. Dejneka, S. Ahmed, Y. Caringal, J. Hangeland,
M. Zhang, P. T. W. Cheng, A. D. Russell, S. Skwish, D. A. Slusarchyk, G. T. Allen,
B. H. Frohlich, B. E. Abboa-Offei, M. Cap, T. L. Waldron, R. J. George, B. Tesfamariam,
K. E. Dickinson, A. A. Seymour and P. M. Sher
Bristol-Myers Squibb Pharmaceutical Research Institute, PO Box 4000, Princeton, NJ 08543, USA
Received 9 April 2007; revised 7 May 2007; accepted 9 May 2007
Available online 16 May 2007
Abstract—Utilization of N-substituted-4-hydroxy-3-methylsulfonanilidoethanolamines 1 as selective b3 agonists is complicated by
their propensity to undergo metabolic oxidative N-dealkylation, generating 0.01–2% of a very potent a1 adrenergic agonist 2. A
summary of the SAR for this hepatic microsomal conversion precedes presentation of strategies to maintain the advantages of chem-
otype 1 while mitigating the consequences of N-dealkylation. This effort led to the identification of 4-hydroxy-3-methylsulfonani-
lidopropanolamines 15 for which the SAR for the unique stereochemical requirements for binding to the b adrenergic receptors
culminated in the identification of the potent, selective b3 agonist 15f.
Ó 2007 Elsevier Ltd. All rights reserved.
In prior papers we described the SAR of 4-hydroxy-3-
methylsulfonoanilidoethanolamines which yielded
nitrogen bonds.6 However, this transformation was a
particularly serious problem for chemotype 1 since the
N-substituent determined the adrenergic agonist profile.
Although b3 adrenergic agonists containing chemotype
1 have been reported by others, no mention of metabo-
lite 2 exists.7 Due to our concerns regarding release of 2,
the safety of compounds 1 was confirmed by iv admin-
istration of 1 at 10 mg/kg to mice, which are extremely
sensitive to 2,8 prior to administration of 1 to African
green monkeys (AGM). Our efforts to surmount the for-
midable challenge posed by 0.01 to ꢀ2% in vivo meta-
bolic conversions of 1–2 are described herein.
1
three selective b3 adrenergic agonists as clinical candi-
dates for treatment of diabetes and obesity.1–3 Given
the uncertainty as to whether adequate depots of brown
adipose tissue were present in adult humans to elicit a
thermogenic response,4 we believed that full b3 agonists
represented the best chance for achieving Proof of Prin-
ciple. The discovery that, out of 50 different catechol
amine surrogates evaluated, only chemotype 1 were full
b3 agonists ensured our commitment to compounds pos-
sessing the 4-hydroxy-3-methyl-sulfonanilido moiety
found in 1 despite the propensity for liberation of low
levels of a potent a1 adrenergic agonist 2.5 This paper,
after delineating the SAR for metabolic conversion of
1–2, focuses on our efforts to maintain the advantages
of series 1 while mitigating the consequences of phase
I metabolic N-dealkylation.
OH
OH
H
H
H
N
NH2
N
N
MeSO2
MeSO2
R
HO
HO
1
2
Incubation of representative examples of 1 with human,
AGM, and rat hepatic microsomes suggested that the
conversions of 1–2 were 3- and 10-fold greater, respec-
tively, for AGM and rat than for human.9 The 10,000-fold
difference in amounts of 2 released by AGM hepatic
microsomal oxidation illustrates the highly structure
dependent nature of this metabolic pathway (Table 1).
Cytochrome P-450 mediated N-dealkylation of 1 was
not unexpected, given the facile cleavage of carbon–
Keywords: b3 Agonists; Metabolic oxidative N-dealkylation; Arylpro-
panolamine; Drug metabolism.
*
Corresponding author. E-mail: William.Washburn@bms.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.05.030

W. N. Washburn et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4290–4296
4291
Moreover, the SAR suggested that, depending on ‘R’,
N-dealkylation could occur via one of three different
pathways (Scheme 1). Mechanistic studies suggested
that N-dealkylation of amines such as 1a–1c for which
Ca is not quaternary proceeds via oxidative conversion
to a carbinolamine with subsequent collapse. As out-
lined in path ‘a’, the carbinolamine arises via electron
transfer to generate an amine radical cation and sub-
sequent loss of a proton followed by oxygen rebound
capture of the benzylic radical to generate a carbinol-
amine intermediate.6a If Ca is quaternary and an ade-
quately stabilized aryl radical cation can be generated,
we envisage a second route ‘b’ liberating 2 could be-
come operative. This pathway entails electron transfer
to generate an aryl radical cation that is subsequently
hydroxylated to generate an ortho substituted phe-
nol.6b Conversion of the latter to an ortho quinine
methide would simultaneously release 2. Formation
of 2 by this pathway would predict a marked depen-
dence on the a-aryl ring substituent(s), since the latter
would modulate the oxidation potential. Detection of
10,000-fold more
2
generated from microsomal
incubation of 1e than from 1d is consistent with this
pathway, as is the 25-fold enhanced formation of 2
from 1f versus 1b. The finding that phenol 1g sponta-
neously released 2 presumably via the non-enzyme
mediated quinone methide route ‘c’ was particularly
noteworthy since the amine substituent ‘R’ of 1
routinely contained a para phenolic ether attached to
Ca to enhance b3 selectivity. This latter pathway,
necessitating P450 mediated oxidative O-dealkylation,
would account for the inability to totally suppress
release of 2 from compounds such as 1i and 1j and
possibly 1h for which the first two pathways would
not be expected to be operative. Efforts to further
modify ‘R’ to preclude any formation of 2 were not
compatible with maintenance of high b3 activity and
selectivity.10
Table 1. % Conversion of 1a–1j to 2 after 3-h incubation at 37 °C with
AGM hepatic microsomes9
OH
H
MeSO2NH
HO
N
R
Since structural modification of the N-substituent of 1
failed to suppress in vivo N-dealkylation to greater
than 99.95% for b3 agonists of interest, we sought
to maintain the hydroxysulfonoanilido/b3 receptor
interactions conducive to a full agonist response yet
modify the structure of 1 to insure that the biological
activity of the primary amine metabolic counterpart of
2 would not be of concern. Specifically, our approach
was to explore both substitution of the hydroxyanilide
moiety of 1 and elongation of the ethanolamine
spacer to ascertain whether full b3 agonist activity
would be maintained while ablating the a adrenergic
activity of the primary amine metabolite counterpart
of 2.
Compound1–3
R
% Conversion
in vitro to 2
1a
1b
1c
1d
1e
1f
(S)-CH(Ph-p-OMe)CONPh
–CH(Ph-p-OCHF2)2
0.05
0.2
1.1
0.001
28
(R)-CH(Ph-p-OCHF2)CH2Ph
–CMe2Ph
–CMe2-Ph-4-OMe
–CH(Ph-p-OMe)2
(R)-CH(Ph-p-OH)CH2(Ph-p-F) 2.2
5
1g
1h
1i
–CMe(Ph-p-OCHF2)2
(S)-CMe(Ph-p-OMe)CONPh
(S)-CMe(Ph-p-OMe)
0.1
0.57
0.49
1j
CONPh-4-PO(OEt)2
OH
Ar
H
N
OH
H
N
OH
H
N
H
R'''
OH
R'''
.
O
.
+
Ar
R'''
R'''
Ar
-2
Path a
R'
R'
R'
R'
R'' = H
OH
H
R''
R'''
N
Ar
OH
OH
H
H
R''
R'''
R''
R''
R'''
N
N
Ar
R'''
O
Ar
-2
Path b
.
+
OH
1
R'
R'
R'
R'
Path c
R' = R''''O
OH
R''
H
R''
R'''
R'''
N
-2
Ar
O
HO
Scheme 1. Oxidative pathways leading to release of 2.

4292
W. N. Washburn et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4290–4296
amine moiety was prompted by the report of Beeley
et al. of aryloxypropanolamines exhibiting b3 agonist
activity.17 Our finding that 6, the 4-hydroxy-3-meth-
ylsulfonoamidophenoxypropanolamine counterpart of
2, exhibited little a adrenergic activity was sufficiently
promising to warrant the synthesis of analogs 7a–7c.
The key synthetic step was a Baeyer–Villiger oxidation
of a SEM protected 4-hydroxy-3-nitrobenzaldehyde to
generate the prerequisite phenol 8 necessary for subse-
quent transformation to the enantiomerically pure chi-
ral epoxide 9 (Scheme 3). Heating of 9 with an
appropriate amine generated 7a–7c. Unlike the aryl
ethanolamine counterparts 1c and 1k, N-substitution
with (R)-1,2-diarylethyl moieties generated high affinity
albeit partial b3 agonists such as 7b and 7c (Tables 2
and 3). However, once the susceptibility of 7a–7c to
air oxidation became apparent, SAR delineation ceased
for this series.18
F
OH
F
OH
HO
HO
NH2
MeSO2NH
HO
NH2
3
4
Prior investigators had reported that the ortho fluorine
of 3 virtually eliminated the a adrenergic activity of nor-
adrenaline and adrenaline while maintaining b1 and b2
activity.11 The abolition of a activity was attributed to
the conformational constraints imposed on the ethanol-
amine side chain due to hydrogen bonding of the ben-
zylic hydroxyl to the aryl fluoride.12 In our in vitro
assays, the b3 agonist profile of 3 was similar to that
of noradrenaline: b3 affinity of 3 (8100 nM) was twice
that of noradrenaline (18,000 nM); moreover, the b3
intrinsic activities (IA) were 75% and 91%, respectively,
for 3 and noradrenaline. Both compounds exhibited
similar binding selectivities of 0.05 and 0.3 versus b1
and b2, respectively, where selectivity is defined as the
ratio of b1 Ki or b2 Ki to b3 Ki.
The oxidatively more robust arylbutanolamine counter-
parts 10a–10c were briefly examined. Synthesis of
arylbutanolamines 10a–10c entailed conversion of 4-
hydroxybenzaldehyde to 11.19 Sequential treatment of
11 with allyl magnesium Grignard, mesylation, and
LiAlH4 reduction generated 4-(4-benzyloxy-3-meth-
ylsulfonamido-phenyl)-1-butene 12 which after epoxida-
tion with MCPBA, condensation with the appropriate
amine, and debenzylation yielded compounds 10a–10c.
However, their markedly inferior b3 affinity and intrinsic
activity relative to chemotype 1 discouraged sustained
efforts.
Accordingly, 4, the o-fluoro counterpart of 2, was pre-
pared (Scheme 2). As expected, 4 was devoid of a ago-
nist activity; however, 4 was a b3 antagonist exhibiting
10-fold diminished b3 affinity relative to 2, although
the b1 and b2 affinities of 2 and 4 were comparable. In
hopes of enhancing b3 agonist activity, a series of N-
substituted analogs of 4 were prepared. The 10-fold or
greater diminution in b3 affinity of these fluorinated
compounds 5a–5d when compared to a corresponding
set of des-fluoro b3 agonists 1a, 1c, 1e, and 1f was disap-
pointing (Table 2). In addition, these fluorinated com-
pounds tended to be partial b3 agonists.
H
N
H
N
CHO
MeSO2
MeSO2
O
O
12
11
Bn
Bn
An alternative approach entailing elongation of the
spacer between the catechol surrogate and the ethanol-
The finding that insertion of a two-atom spacer be-
tween the ethanolamine moiety and the catechol sur-
rogate abolished a agonist activity of 6 prompted
homologation of 2 to generate the corresponding
3-arylpropan-2-ol-1-amine 13. Diazotization of chiral
3-nitrotyrosine in aq H2SO4 proceeded with retention
to generate chiral a-hydroxy acid 14. Conversion of
14 to the primary amine 13 and selected analogs 15
by solution chemistry entailed condensation with an
appropriate amine, followed by borane reduction
(Scheme 4). However, most N-substituted analogs
15 were prepared combinatorially via solid state
synthesis.20
F
O
F
O
MeSO2NH
BnO
Br
CH3
c, d, e
a, b
HO
H
Ph
Ph
F
OTBS
g,h
N
O
MeSO2NH
BnO
I
B
Me
f
Encouraged by the finding that 13 was devoid of a
adrenergic activity, the SAR of this chemotype was fur-
ther elucidated. All three b receptors exhibited an atyp-
ical stereochemical preference for ligands containing
homobenzylic alcohols with S chirality, which is oppo-
site to the projection of the hydroxyl preferred for ary-
lethanolamines or aryloxypropanol-amines.21 In
particular, the affinity of the b3 receptor was 10-fold
greater for the S configured homobenzylic alcohols
15b and 15d than the corresponding R isomers 15a
and 15c (Table 4).
F
OH
H
N
R
RNH2
MeSO2NH
HO
j, k
4 R = H
5 R = H
i
Scheme 2. Reagents and conditions: (a) HNO3, AcOH, 50 °C, 25%; (b)
NaH, BnBr, DMF, 90%; (c) SnCl2, EtOH, 89%; (d) MsCl, Pyr, 94%;
(e) Br2, dioxane, 60 °C, 86%; (f) BH3ÆMe2S, THF, 96% (98.2% ee); (g)
NaI, acetone, 65 °C, 96%; (h) TBSCl, DMAP, ImH, DMF, 100%; (i)
EtN(i-Pr)2, heat, 120 °C, 50–75%; (j) KF, THF; (k) H2, MeOH, 85% (2
steps).

W. N. Washburn et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4290–4296
4293
Table 2. b3 Affinity and intrinsic activity for 2-fluoro4-hydroxy-3-methyl-sulfonoanilidoethanolamines 4 and 5 relative to 1 and 213
X
OH
H
N
MeSO2NH
R
HO
R
X = H16
X = F16
Compound
b3 Ki (nM)
b3 IA
Compound
b3 Ki (nM)
120,000
b3 IA
H
2
17,000
77
21
93
4
0
(S)-CH(Ph-p-OMe)CONPh
(R)-CH(Ph-4-OCHF2)-CH2Ph
–CMe2Ph-4-OMe
1a
1c
1e
1f
1k
119
107
86
5a
5b
5c
5d
58
77
88
24
270
200
7
81
–CH(Ph-4-OMe)2
(R)-CHPh-3,4-(OMe)2-CH2Ph
100
110
1700
11
O
15f was not only the most potent b3 agonist of these four
diastereomers but also was the only full b3 agonist.23
Both b1 and b2 affinities were greatest for 15g, the S,S
diastereomer. In addition, b1 IA was somewhat reduced
for 15f versus for 15g; whereas, b2 IA remained an
invariant 100% for both of these diastereomers.24 These
fortuitous divergences in the stereochemical preferences
for the three b receptors offered additional means to en-
hance b3 in vivo selectivity. These findings also under-
score the divergence in IA sensitivity to the ligand/
receptor interactions, b1 being the most sensitive, b2
the least.
NO2
a, b,c
NO2
OH
NO2
SEMO
CO2H
H
+
HO
SEMO
8
O
MeSO2NH
SEMO
O
1)
2)
d
e,f
O
S
O
NO2
9
O
O
OH
RNH2
h
H
N
MeSO2NH
HO
O
R
g
6 R = H
7 R = H
Proper orientation of this ligand in the b3 receptor ap-
peared dependent on the a-aryl ring being 3,4-disubsti-
tuted with oxygen substituents. The diminished affinity
and b3 IA of 15b or 15d relative to 15f underscore the
importance of both methoxyls on the a-aryl ring to elicit
a strong b3 agonist response; whereas, only the 4-meth-
oxyl is essential for the corresponding arylethanolamine
series 1.1 The effect was not steric in origin, since selec-
tivity was ablated upon replacement of both methoxyls
with chlorine (15i). Reduction in the acceptor hydrogen
bonding potential of the phenolic oxygen appendages
correlated with loss of selectivity; compare 15f or 15j–
15k.
Scheme 3. Reagents and conditions: (a) SEMCl, EtN(i-Pr)2, CH2Cl2,
82%; (b) MCPBA, CH2Cl2; (c) NaOH/H2O/MeOH, 19% (2 steps); (d)
NaH, DMF, 94%; (e) H2, EtOAc, PtO2; (f) Ms2O, CH2Cl2, 95% (2
steps); (g) excess RNH2, 100 °C, 65%; (h) TFA, EtOH, 40–70%.
A more complete SAR pattern emerged following char-
acterization of the four diastereomers 15e–15h. b3 affin-
ity was enhanced by >5-fold if the chirality of the
homobenzylic alcohol was S (compare Ki b3 of 15f
and 15g to that of 15e or 15h). The S,R diastereomer
Table 3. b3 AR agonist activity and selectivity versus b1 and b2 for 4-hydroxy-3-methylsulfonoamidophenoxypropanolamines 6 and 7a–7c, and
4-hydroxy-3-methylsulfonoamidophenoxybutanolamines 10a–10c13
OH
H
CH3SO2NH
HO
N
X
R
R16
Compound
b3 Ki (nM)
b3 IA (% act)
Selectivitya versus
b1
b2
X = O
6
7a
H
–CH2Ph-4-OMe
15,000
47
46
48
67
0.08
3
0.04
360
5
0.4
7
(R)-CH(Ph-4-OCHF2)-CH2Ph
(R)-CHPh-3,4-(OMe)2-CH2Ph
7b
7c
66
120
3
15
X=CH2
10a
10b
10c
–CH2Ph-4-OMe
(R)-CH(Ph-4-OCHF2)-CH2Ph
(R)-CHPh-3,4-(OMe)2-CH2Ph
3700
880
42
40
60
1
38
1
7
160
112
27
a Selectivity is defined as the ratio of b1 Ki or b2 Ki to b3 Ki.

4294
W. N. Washburn et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4290–4296
two aryl rings in a symmetrical amine substituent was
4
NO2
HO
CO2H
a, b, c
CH3SO2NH
HO
CO2H
OH
not encouraging; both the 1,1-diarylmethyl substituent
15q and the 1,3-diaryl-2-propyl moiety 15r produced a
50-fold decrease in b3 affinity and loss of selectivity ver-
sus b1 and b2 receptors.
NH2
14
CH3SO2NH
HO
R
N
d, e
However, the b adrenergic profile of 16a, arising from
homologation of 15f, encouraged further SAR elucida-
tion (Table 6). The high b3 affinity of 16a was depen-
dent on both methoxyls just as was the case for 15f.
The interaction between the para methoxyl and the
b3 receptor appears the more important; compare
the 5-fold reduced b3 selectivity and affinity of 16b
to the 40-fold reduction in b3 affinity of 16c. Replace-
ment of the methoxyls with difluoromethoxy (16d)
reduced b3 affinity and increased b1 and b2 affinity.
Substitution of the c-aryl ring was not promising: an
ortho (16e) or meta (16f) chlorine enhanced b1 and
b2 affinity while decreasing that of b3, whereas a para
chlorine (16g) produced little effect on b3 selectivity or
affinity.
H
OH
13 R = H
15 R = H
Scheme 4. Reagents and conditions: (a) NaNO2, aq H2SO4, 0 °C 71%;
(b) H2, 10% Pd/C, MeOH, 95%; (c) MesCl, py 63% (2 steps); (d)
RNH2, EDC, HOAt, DMF, 20 °C; (e) BH3, THF).
Further modifications of the N-substituent are shown in
Table 5. Chlorination of the b-aryl ring to generate 15l–
15n was not beneficial, since b3 IA decreased compared
to 15f. Deletion of either the b-aryl ring (15o) or a-aryl
ring (15p) severely reduced b3 affinity and essentially
eliminated any selectivity for b3. Incorporation of the
Table 4. SAR for agonist activity for b3 AR compared to b1 and b2 AR for 4-hydroxy-3-methylsulfonoamidophenylpropanolamines 15a–15k13
CH3SO2NH
HO
R
X
N
H
*
OH
#
Compound16
R
X
Chirality
b3 Kia (% b3 IA)
b1 Kia (% b1 IA)24
b2 Kia
#
*
15a
15b
15c
15d
15e
15f
15g
15h
15i
OMe
OMe
H
H
R
S
R
S
R
S
S
R
S
S
S
R
R
R
R
R
R
S
1100 (27)
110 (44)
460 (30)
53 (65)
25 (43)
6 (86)
>10,000 (0)
3600
2700
350
3100
510
1500
66
H
OMe
OMe
OMe
OMe
OMe
OMe
Cl
>10,000 (0)
6000
H
OMe
OMe
OMe
OMe
Cl
>10,000 (0)
2600 (60)
560 (75)
47,000 (48)
1500
9 (62)
24
S
150 (70)
133 (70)
8 (70)
820
38
R
R
R
15j
15k
OH
OCHF2
OMe
OCHF2
3200
690 (0)
1070
102
19 (60)
a Expressed in nM.
Table 5. b3 AR agonist activity and selectivity versus b1 and b2 for 4-hydroxy-3-methylsulfonoamidophenylpropanolamines 15l–15r13
CH3SO2NH
R
*
N
H
HO
OH
Compound16
R
b3 Kia (% IA)
Selectivityb
*
b1
b2
15l
(R)-CH(Ph-3,4-(OMe)2)CH2Ph-2-Cl
(R)-CH(Ph-3,4-(OMe)2)CH2Ph-3-Cl
(R)-CH(Ph-3,4-(OMe)2)CH2Ph-4-Cl
CH2Ph-4-OMe
S
7 (65)
6 (68)
58
52
24
2
13
13
6
15m
15n
15o
15p
15q
15r
S
S
24 (68)
1100(50)
7300(35)
340 (60)
170 (75)
Rac
Rac
S
0.3
0.1
0.6
0.2
CH2CH2Ph-3,4-(OMe)2
CH(Ph-4-OCHF2)2
CH(CH2Ph-3,4-(OMe)2)2
0.6
3
S
3
a Expressed in nM.
b Selectivity is defined as the ratio of b1 Ki or b2 Ki to b3 Ki.

W. N. Washburn et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4290–4296
4295
Table 6. SAR for agonist activity for b3 AR compared to b1 and b2 AR for 4-hydroxy-3-methylsulfonoamidophenylpropanolamines 16a–16g13
X
R
CH3SO2NH
N
Z
H
HO
OH
Compound16
R
X
Z
b3 Kia (% b3 IA)
b1 Kia (b3 Sel)b
b2 Kia (b3 Sel)b
16a
16b
16c
16d
16e
16f
16g
OMe
H
OMe
OMe
H
H
9 (84)
1000 (111)
730 (24)
570 (1)
390 (43)
280 (9)
200 (0.5)
71 (1)
H
31 (93)
390 (44)
59 (86)
7 (73)
OMe
OCHF2
OMe
OMe
OMe
H
OCHF2
OMe
OMe
OMe
H
490 (8)
2-Cl
3-Cl
4-Cl
270 (39)
420 (14)
1050 (88)
57 (8)
26 (1)
120 (10)
29 (57)
12 (89)
a Expressed in nM.
b Selectivity is defined as the ratio of b1 Ki or b2 Ki to b3 Ki.
Table 7. In vivo response of African Green Monkeys
Table 8. Dose–response of African Green Monkeys (iv administration)
Compound
AGM response to 0.1 mg/kg iv dose
Compound
Lipolysis
ED50 (mg/kg)
b1 margin
before
onset of
b2 margin
before
D HR
D K+
D FFA
decrease in
serum K+a
+26*
+4
+0.08
1.05*
+0.39
1.06*
0.92*
0.95*
0.87*
tachycardiaa
1c
5da
15f
À0.49
À0.08
À0.55*
À0.25*
À0.32*
15f
0.009
0.08
0.03
0.1
>3 < 11
>6 < 12
>3 < 5
>25
>55
>60
>25
>25
+24
6
+29*
32*
1b (194449)
1c (196085)
1j (201620)
15m
16a
16g
aThe margin of separation was the ratio of the dose that produced the
onset (statistically significant) of a b1 or b2 event to the ED50 for
lipolysis.
a Administered iv at 0.5 mg/kg.
* P < 0.05.
The in vivo b3 activity of the more promising com-
pounds was ascertained following iv administration to
ketamine sedated African green monkeys (AGM) by
monitoring the b3 induced increase in free fatty acid lev-
els (FFA). A dose–response was determined for those
compounds which produced a statistically significant
elevation in free fatty acid (D FFA) unaccompanied by
b2 mediated decrease in serum K+ concentrations26 (D
K+) or b1 or b2 induced tachycardia (HR). The fluori-
nated hydroxysulfonanilides 5 produced weak responses
consistent with partial b3 agonist activity (Table 7). For
example, 0.5 mg/kg of 5d induced an increase in FFA
levels less than that induced by 1c (BMS-196085) at
0.1 mg/kg. The lack of separation between the b3 and
b2 mediated responses for compounds 15m, 16g, and
particularly 16a was disappointing.
Acknowledgments
We acknowledge the contributions of members of the
Bristol-Myers Squibb analytical chemistry department,
D. Kirk for a sample of 3, and D. Strosberg for transfec-
ted CHO cells.
References and notes
1. Washburn, W. N.; Sher, P. M.; Poss, K.; Girotra, R. N.;
McCann, P. J.; Gavai, A. V.; Mikkilineni, A. B.; Mathur,
A.; Cheng, P.; Dejneka, T. C.; Sun, C. Q.; Wang, T. C.;
Harper, T. W.; Russell, A. D.; Slusarchyk, D. A.; Skwish,
S.; Allen, G. T.; Hillyer, D. E.; Frohlich, B. H.; Abboa-
Offei, B. E.; Cap, M.; Waldron, T. L.; George, R. J.;
Tesfamariam, B.; Ciosek, C. P.; Young, D. A.; Dickinson,
K. E.; Seymour, A. A.; Arbeeny, C. M.; Gregg, R. E.
Bioorg. Med. Chem. Lett. 2001, 11, 3035.
2. Gavai, A. V.; Sher, P. M.; Mikkilineni, A. B.; Poss, K. M.;
McCann, P. J.; Girotra, R. N.; Fisher, L. G.; Wu, G.;
Bednarz, M. S.; Mathur, A.; Wang, T. C.; Sun, C. Q.;
Slusarchyk, D. A.; Young, D. A.; Dickinson, K. E.;
Seymour, A. A.; Arbeeny, C. M.; Washburn, W. N.
Bioorg. Med. Chem. Lett. 2001, 11, 3041.
Only 15f offered any promise as a selective thermogenic
agent; however the margins between the ED50 dose for
lipolysis and the doses inducing the onset of significant
b1 or b2 mediated events were not superior to that
previously determined for the three clinical candidates
(Table 8).
Efforts to continue to progress either 15f or alternatively
to identify a superior analog ceased once the inability of
BMS-196085 to induce a persistent thermogenic re-
sponse in the clinic became apparent.2
3. Washburn, W. N.; Sun, C. Q.; Bisacchi, G.; Wu, G.;
Cheng, P.; Sher, P. M.; Ryono, D.; Gavai, A. V.; Poss, K.;
Girotra, R. N.; McCann, P. J.; Mikkilineni, A. B.;

4296
W. N. Washburn et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4290–4296
Dejneka, T. C.; Wang, T. C.; Merchant, Z.; Morella, M.;
13. Receptor binding assays were run in triplicate with
membranes prepared from CHO cells expressing the
three cloned human b receptors in the presence of 125I-
iodocyanopindolol.14 b3 AR agonist activity was
assessed by measurement of cAMP accumulation levels
in CHO membranes expressing human b3 AR.15 IA is
given as a percentage of the maximal stimulation with
isoproterenol.
14. All compounds were evaluated using the truncated human
b3 adrenergic receptor reported by Emorine, L. T.;
Marullo, S.; Briend-Sutren, M. M.; Patey, G.; Tate, K.;
Delavier-Klutchko, C.; Strosberg, A. D. Science 1989,
245, 1118.
15. Sher, P. M.; Fisher, L. G.; Skwish, S.; Michel, I. M.;
Seiler, S. M.; Washburn, W. N.; Dickinson, K. E. J. Med.
Chem. Res. 1997, 7, 109.
16. Spectral data were fully consistent with structures.
17. Beeley, L. J.; Berge, J. M.; Chapman, H.; Dean, D. K.;
Kelly, J.; Lowden, K.; Kotecha, N. R.; Morgan, H. K. A.;
Rami, H. K.; Thompson, M.; Vong, A. K. K.; Ward, R.
W. Bioorg. Med. Chem. Lett. 1997, 7, 219.
Arbeeny, C. M.; Harper, T. W.; Slusarchyk, D. A.;
Skwish, S.; Russell, A. D.; Allen, G. T.; Tesfamariam, B.;
Frohlich, B. H.; Abboa-Offei, B. E.; Cap, M.; Waldron, T.
L.; George, R. J.; Young, D.; Dickinson, K. E.; Seymour,
A. A. Bioorg. Med. Chem. Lett. 2004, 14, 3525.
4. Himms-Hagen, J.; Ricquier, D. In Handbook of Obesity;
Bray, G. A., Bouchard, C., James, W. P. T., Eds.; Marcel
Dekher: New York, 1998; p 415.
5. Larsen, A. A.; Gould, W. A.; Roth, H. R.; Comer, W. T.;
Uloth, R. H. J. Med. Chem. 1967, 10, 462.
6. (a) Guengerich, F. P.; Yun, C. H.; Macdonald, T. L.
J. Biol. Chem 1996, 271, 27321; (b) Guengerich, F. P.
Chem. Res. Toxicol. 2001, 14, 611.
7. Hu, B.; Ellingboe, J.; Han, S.; Largis, E.; Mulvey, R.;
Oliphant, A.; Sum, F.-W.; Tillet, J. J. Med. Chem. 2001, 44,
1456.
8. LD100 for mice was 57 lg/kg of 2 administered iv; iv doses
of 0.5–5.7 lg/kg produced progressively more pronounced
piloerection; 0.05 lg/kg was a no effect dose. This dose
dependent response suggested that piloerection following
iv administration of 10 mg/kg of 1 during the murine
safety screen was indicative of 0.01–0.1% of the dose being
converted to 2. A few compounds, such as 1f, were lethal
at 10 mg/kg but not at 1 mg/kg implying that metabolic
conversion to 2 was ꢀ0.1–1%.
18. In 95% EtOH at 20 °C, t1/2 was ꢀ40 min to presumably
generate quinones.
19. Larson, A. A.; Gould, W. A. US Patent 1,171,477.
20. Wu, G.; Cai, Z. W.; Bednarz, M. S.; Kocy, O.; Gavai, A.
V.; Godfrey, J. D.; Washburn, W. N.; Sher, P. M.
J. Comb. Chem. 2005, 7, 99.
21. Stereochemical preferences had not previously been
addressed in the few reports concerning arylpropanol-
amine as adrenergic ligands.22
9. Typically, a 100 lM solution of compounds of structure 1
was incubated for 3 h in pH 7.4 phosphate buffer (0.1 M)
containing African green monkey hepatic microsomes
(3 mg/mL protein) and 3 mM NADPH whereupon the
protein was precipitated by addition of 2 volumes MeCN
prior to determination of the conversion to 2 by LC/MS.
Since control studies entailing incubation of compounds
of structure 1 as described above in the absence of
NADPH generated minuscule amounts of 2, solvolytic
release of 2 due to chemical instability of 1 after
protonation was of minimal importance.
22. (a) Fuhrer, W.; Ostermayer, F.; Zimmermann, M.; Meier,
M.; Muller, H. J. Med. Chem. 1984, 27, 831; (b) Lo¨vgren,
¨
K.; Hedberg, H.; Nilsson, J. L. G. J. Med. Chem. 1981, 24,
ˇ
´
ˇ
451; (c) Alderova, E.; Trcka, V.; Smejkal, V.; Protiva, M.
Collect. Czech. Chem. Commun. 1969, 34, 479; (d) WO
2000/012462; (e) WO 1999/051564.
23. Absolute stereochemistry of 15f assigned by X-ray
crystallography.
10. Subsequently during clinical studies with BMS-196085
(1c), BMS-194449 (1b), and BMS-201620 (1j) dose
dependent incidences of sweaty palms, flushing, and
piloerection, all classic manifestations of an a adrenergic
agonist, as well as detection of trace levels of 2 validated
these concerns regarding metabolic liberation of 2.
11. (a) Kirk, K. L.; Cantacuzene, D.; Creveling, C. R. Biomed.
Aspects Fluorine Chem. 1982, 75; (b) Brasili, L.; Cantala-
messa, F.; Picchio, M. T.; Quaglia, W. Eur. J. Pharm.
1987, 144, 141.
24. b1 IA was determined by measuring the acceleration in
contraction of spontaneously beating guinea pig atria
relative to that induced by isoproterenol.25 b2 IA was
determined by measuring relaxation of guinea pig trachea
rings submaximally contracted with carbacol relative to
that induced by isoproterenol.
25. Tesfamariam, B.; Allen, G. T. Br. J. Pharmacol. 1994, 112,
55.
26. Lipworth, B. J.; McFarlane, L. C.; Coutie, W. J.;
McDevitt, D. Eur. J. Clin. Pharmacol. 1989, 37,
297.
12. Hieble, J. P.; Hehr, A.; Li, Y.; Ruffolo, R. R. Proc.
Western Pharmacol. Soc. 1998, 41, 225.