2,4-Thiazolidinediones as potent and selective human β3 agonists

Article abstract of DOI:10.1016/S0960-894X(01)00063-4

Methylsulfonamide substituted 2,4-thiazolidinedione 22c is a potent (EC₅₀ = 0.01 μM, IA = 1.19) and selective (more than 110-fold over β₁ and β₂ agonist activity) β₃ agonist. This compound has also been proven to be active and selective in an in vivo mode.

Full text of DOI:10.1016/S0960-894X(01)00063-4

Bioorganic & Medicinal ChemistryLetters 11 (2001) 757–760
2,4-Thiazolidinediones as Potent and Selective Human ꢀ3 Agonists
Baihua Hu,a,* John Ellingboe,a Iwan Gunawan,a Stella Han,b Elwood Largis,b Zenan Li,a
Michael Malamas,a Ruth Mulvey,b Alexander Oliphant,a Fuk-Wah Sum,a Jeff Tillettb
and Victoria Wonga
aChemical Sciences, Wyeth-Ayerst Research, Pearl River, NY 10965, USA
bCardiovascular/Metabolic Diseases Research, Wyeth-Ayerst Research, Pearl River, NY 10965, USA
Received 4 December 2000; accepted 12 January2001
Abstract—Methylsulfonamide substituted 2,4-thiazolidinedione 22c is a potent (EC50=0.01 mM, IA=1.19) and selective (more than
110-fold over b1 and b2 agonist activity) b3 agonist. This compound has also been proven to be active and selective in an in vivo
mode. # 2001 Elsevier Science Ltd. All rights reserved.
The b3-adrenergic receptor (b3-AR) has been shown to
mediate various pharmacological and physiological
effects such as lipolysis in white adipocyte tissue (WAT),
thermogenesis in brown adipocyte tissue (BAT), and
The current b-amino alcohols for b3-AR agonist LHS
moieties belong to two general structural classes:1
1
relaxation of urinarybladder detrusor tissue. Conse-
quently, several pharmaceutical firms, including our-
selves, are engaged in developing potent and selective
b3-AR agonists for the treatment of obesity, type II
diabetes, and frequent urination. CL316243,2 dis-
covered in our laboratories, has been shown to be an
effective anti-obesityand anti-diabetic agent in rodents.
However, human clinical trials with CL316243 have
been disappointing due to structural differences between
the human and the rodent b3-AR. To discover potent
human b3-AR agonists we have established Chinese
hamster ovary(CHO) cell line expressing the cloned
human b3-AR receptor that would more accurately
predict the effects in humans. Using this human b3-AR
assaythe 4-piperidino-benzoic acid derivative 1 was
aryloxypropanolamines and arylethanolamines. The
aryloxypropanolamines that are not known in the lit-
erature were generallyprepared from epoxides 3 by
regioselective ring opening with dibenzylamine followed
bydebenzylation (Scheme 1). The required epoxides for
21a–c, e–f (the substituents a–f for compounds 43 and
21 are defined in Table 1) were prepared according to
found to be
a modestlypotent human agonist
(EC50=0.22 mM, IA=1.2).2c We decided to investigate
whether incorporation of a carboxylic acid mimetic 2,4-
thiazolidinedione would have a positive effect on b3-AR
agonist 1. The synthesis and b3-AR agonist activityof
2,4-thiazolidinedione 2 with several b-amino alcohols
on the left-hand side (LHS) will be described in this
communication.
*Corresponding author. Fax: +1-845-732-5561; e-mail: hub@
war.wyeth.com
Scheme 1. (a) Bn2NH, MeOH; (b) HCO2NH4, Pd/C.
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00063-4

758
B. Huet al. / Bioorg. Med. Chem. Lett. 11 (2001) 757–760
Table 1. Variation of LHS of 2,4-thiazolidinediones
Compound
Ar
b3-ARa EC50 mM (IA)
b2-ARa EC50 mM (IA)
b1-ARa EC50 mM (IA)
CL316243
21a
21b
21c
21d
21e
21f
22a
22b
1.15 (0.63)
0.04 (0.71)
0.012 (0.89)
0.003 (1.1)
0.38 (0.84)
0.013 (1.2)
0.024 (1.1)
0.17 (0.56)
0.9 (0.9)
262 (0.43)
(0.03)
(0.01)
(0)
111 (0.21)
2.02 (0.34)
2.86 (0.59)
1.17 (0.99)
NT
1.53 (1.17)
0.021 (0.41)
NT
Phenyl
4-OH-Phenyl
4-OH-3-MeSO2NH-Phenyl
4-OH-3-PhSO2NH-Phenyl
5-Carbostyril
4-(2-Benzimidazolone)
3-Cl-Phenyl
3-Pyridyl
NTb
(0.05)
(0.03)
NT
NT
NT
22c
22d
22e
22f
4-OH-3-MeSO2NH-Phenyl
4-Cl-3-MeSO2NH-Phenyl
3-MeSO2NH-Phenyl
4-OH-3-PhSO2NH-Phenyl
0.01 (1.19)
0.66 (0.59)
>10
1.19 (0.67)
(0.02)
NT
2.72 (0.82)
(0.03)
NT
0.006 (0.93)
0.62 (0.5)
0.43 (0.46)
ab-ARs agonistic activities were assessed bymeasurement of cAMP accumulation levels in CHO cells expressing human b-ARs; the intrinsic
activities (IA) were given as a fraction of the maximal stimulation with isoproterenol.
bNT=not tested.
literature procedures.4 The corresponding 3-benzene-
sulfonamide epoxide 3d for 21d was synthesized as
shown in Scheme 2. Benzylation of phenol 55 with ben-
zyl bromide followed by reduction of the nitro group
and sulfonylation with benzenesulfonyl chloride gave 6.
The benzene-sulfonamide 6 was then benzylated again
with benzyl chloride, and the benzoic ester was hydro-
lyzed. Alkylation of the resulting phenol with (2S)-gly-
cidyl 3-nitro-benzene sulfonate gave the requisite 3d
(Scheme 2).
prepared from 108 bya similar route as shown in
Scheme 4. In this case, the azide instead of amine was
used as a LHS moiety.
As indicated in Scheme 5, reduction of bromoketone 13
with Corey’s CBS-borane followed by cyclization with
NaOH gave the epoxide which underwent ring opening
upon treatment with ammonia. The major product 14
was readilyseparated from the minor amount of the
9
kineticallyless favored isomer
byrecrsytallization.
After protection of the amine moiety, reduction of the
nitro group, sulfonylation of the resulting aniline fol-
lowed bydeprotection of the carbamate group provided
15. The 3-pyridyl analogue 18 was prepared analogously
to 14 from the bromoketone 1610 as shown in Scheme 6.
Several routes were used to prepare the arylethanol-
amines. The bromide 8b was derived from aniline 76 by
sulfonamide formation, bromination and enantio-
selective reduction using Corey’s (R)-2-meth-CBS-oxaz-
aborolidine as the catalyst. Nucleophilic substitution of
bromide 8 with sodium azide in DMSO7 followed by
debenzylation and azide reduction furnished the desired
arylethanolamine 9 (Scheme 3). The azide 12 was
Thiazolidinediones 21a–f or 22a–f (the substituents a–f
are defined in Table 1) were convenientlyprepared as
outlined in Scheme 7. A Knoevenagel condensation
between aldehyde 1911 and 2,4-thiazolidinedione fol-
lowed by catalytic hydrogenation and acidic hydrolysis
Scheme 2. (a) BnBr, K2CO3, DMF, 90%; (b) H2, PtO2, MeOH, 79%;
(c) PhSO2Cl, Et3N, EtOAc, 61%; (d) BnCl, NaH, DMF, 70%; (e)
NaOH, 82%; (f) (2S)-glycidyl 3-nitrobenzene sulfonate, K2CO3,
DMF, 76%.
Scheme 4. (a) PhSO2Cl, Py, CH2Cl2, 57%; (b) CuBr2, CHCl3, 96%;
(c) (R)-2-methyl-CBS-oxazaborolidine, BH3, THF; (d) NaN3, DMSO;
64% over steps c and d.
Scheme 3. (a) PhSO2Cl, Py, CH2Cl2, 47%; (b) CuBr2, CHCl3, 53%;
(c) (R)-2-methyl-CBS-oxazaborolidine, BH3, THF, 61%; (d) NaN3,
DMSO; (e) H2, Pd/C, MeOH, 71% for 9a (steps d and e), 69% for 9b
(steps d and e).
Scheme 5. (a) (R)-2-Methyl-CBS-oxazaborolidine, BH3, THF, 90%;
(b) NaOH, THF, 82%; (c) NH3, MeOH, 70%; (d) (Boc)2O, DMF,
78%; (e) Fe, NH4Cl, MeOH/H2O, 83%; (f) MsCl, Py, CH2Cl2, 49%;
(g) TFA, CH2Cl2, 64%.

B. Huet al. / Bioorg. Med. Chem. Lett. 11 (2001) 757–760
759
provided piperidones 20. The desired final products
(21a–f or 22a–f)12 were prepared byutilizing reductive
amination of piperidones 20 with the appropriate aryl-
oxypropanolamines, arylethanolamines or arylethanol-
azides.
also examined, and the 5-carbostyril (8-hydroxy-3,4-
dihydro-1H-quinolin-2-one) analogue 21e is potent
(EC50=0.013 mM, IA=1.2) and 117-fold selective over
the b1 agonist activitywhile the benzimidazolone
analogue 21f is also potent (EC50= 0.024 mM, IA=1.1)
but not selective over the b1 receptor.
We examined a variety of aryloxypropanolamines sub-
stituted 2,4-thiazolidinediones for their abilityto stimu-
late an increase in cAMP in CHO cells expressing the
cloned human b3-AR receptor.13 Phenoxy 21a is a
potent (EC50=0.04 mM) agonist at the b3 receptor,
however, it is a partial agonist (defined as IA=0.2–0.90)
with 71% activation relative to (À)-isoproterenol and is
only50-fold selective over the b1 agonist activity. The 4-
hydroxylphenoxy 21b is also a partial agonist, but it is
more selective (238-fold) against the b1 receptor. While
the 3-methylsulfonamide phenoxy 21c is a potent
(EC50=0.003 mM) full agonist (defined as IA >0.90)
with IA value of 1.1, the bulkybenzenesulfonamide
phenoxy 21d is 127-fold less potent. These results
suggest that in the aryloxypropanolamine series the
size of sulfonamide substituent (methylsulfonamide
of 21c vs benzenesulfonamide of 21d) is important for
b3 agonist activitywhile the hydrogen bonding capacity
(21a–c, 4-OH vs 4-H or 3-MeSO2NH-4-OH) appears to
be less important. Two heterocyclic derivatives were
Next we examined various arylethanolamine substituted
2,4-thiazolidinediones. Introduction of a 3-chlorophenyl
or 3-pyridyl (both are widely used as b3 agonist LHS
moieties1) group on 22 resulted in low potencycom-
pounds (22a and 22b). To our delight, the 3-methylsul-
fonamide-4-hydroxy analogue 22c is a potent agonist at
the b3 receptor (EC50=0.01 mM, IA=1.19) with more
than 110-fold selectivityagainst both b1- and b2-ARs.
Further modifications byreplacement of 4-hydroxywith
either 4-chloro or 4-hydrogen gave less potent com-
pounds (22d and 22e). In contrast to 21d, replacement
of the 3-methylsulfonamide (22c) with 3-benzenesulfo-
namide
(22f)
retained
b3
agonist
activity
(EC50=0.006 mM, IA=0.93) while selectivityover the
b1-AR is somewhat (ꢀ4-fold) decreased. These results
suggest that, unlike the aryloxypropanolamine series, in
the arylethanolamine series the hydrogen bonding
capacity(4-OH vs 4-chloro or 4-H in 22c, 22d, and 22e)
is important for b3 agonist activitywhile the size of
the sulfonamide substituent (methylsulfonamide of 22c
vs benzenesulfonamide of 22f) appears to be of little
importance.
Selected compounds with a good agonist activityprofile
were examined in b1 and b2 binding assays, and aryloxy-
propanolamine derivatives 21e and 21f were found to
bind tightlyto b1 and b2 receptors with a binding con-
stant Ki (determined byinhibition of 125I-iodocyano-
pindolol)14 in the range of 0.004–0.055 mM. These
results suggest that the aryloxypropanolamines do in
fact exhibit strong antagonist activityat the b1 and b2
receptors and therefore maycause unwanted side
effects.1 In contrast, the arylethanolamine 22c has a Ki
value of 3.28 mM for b2 and 1.57 mM for b1, confirming
its selectivityover the b1- and b2-ARs.
Scheme 6. (a) (R)-2-Methyl-CBS-oxazaborolidine, BH3, THF, 34%;
(b) NaOH, THF, 72%; (c) NH3, MeOH, 72%.
The abilityof b3 agonist 22c to treat or inhibit disorders
related to obesityor type II diabetes was confirmed in
an in vivo procedure15 which compared thermogenesis
in human b3-AR transgenic mice (Tg mice) with b3-AR
knock out mice (KO mice). Administered 10 mg/kg (ip)
to Tg mice and KO mice, 22c is active (30Æ4% ther-
mogensis) in Tg mice and inactive (À2Æ4% thermo-
gensis) in KO mice.
In conclusion, LHS modifications on 2,4-thiazolidine-
diones have led to the identification of a potent and
selective agonist 22c with an EC50 of 0.01 mM, >110-
fold selectivityover b1- and b2-AR agonist activityand
active and selective in the in vivo procedure. It was
subsequentlyfound that slight modification on the
2,4-thiazolidinedione moiety(such as alkylation on the
N-H, replacement of 2,4-thiazolidinedione with 1,2,4-
oxadiazolidine-3,5-dione) led to more potent and selec-
tive compounds. The details of these modifications will
be presented separately.
Scheme 7. (a) 2,4-Thiazolidinedione, piperidine, EtOH, 99%; (b) H2,
Pd/C, MeOH; (c) HCl, 54% over steps b and c; (d) aryloxy-
propanolamines, arylethanolamines or arylethanolazides, NaBH
(OAc)3, DMF.

760
B. Huet al. / Bioorg. Med. Chem. Lett. 11 (2001) 757–760
5. Svensson, M.; Helgee, B.; Bertil, K.; Andersson, G. J.
Mater. Chem. 1998, 8, 353.
Acknowledgements
6. For the synthesis of 7 and 8a, see: Washburn, W. N.;
Girotra, R. N.; Sher, P. M.; Mikkilineni, A. B.; Poss, K. M.;
Mathur, A.; Gavai, A.; Bisacchi, G. S. EP 659737, 1995;
Chem. Abstr. 1995, 124, 8408
The authors acknowledge the members of the Wyeth-
Ayerst Discovery Analytical Chemistry group for ana-
lytical and spectral determinations.
7. Alvarez, S. G.; Alvarez, M. T. Synthesis 1997, 413.
8. Oelschlaeger, H.; Schreiber, O. Liebigs Ann. Chem. 1961,
641, 81.
References and Notes
9. Guy, A.; Ferroud, D. C.; Garreay, R.; Godefroy-Fal-
guieres, A. Synthesis 1992, 821.
10. Sarodnick, G.; Kempter, G. Pharmazier 1985, 40, 384.
11. Taylor, E. C.; Skotnicki, J. S. Synthesis 1981, 606.
12. All final compounds and novel aryloxypropanolamines
1. (a) For recent reviews on b3-agonists see: Katsilambros, N.
Eur. J. Intern. Med. 2000, 11, 65. (b) Weyer, C.; Gautier, J. F.;
Danforth, E., Jr. Diabetes Metab. 1999, 25, 11. (c) Kordik,
C. P.; Reitz, A. B. J. Med. Chem. 1999, 42, 181. (d) Weber,
A. E. Annu. Rep. Med. Chem. 1998, 33, 193. (e) Dow, R. L.
Exp. Opin. Invest. Drugs 1997, 6, 1811.
2. (a) Sum, F. W.; Gilbert, A.; Venkatesan, A. M.; Lim, K.;
Wong, V.; O’Dell, M.; Francisco, G.; Chen, Z.; Grosu, G.;
Baker, J.; Ellingboe, J.; Malamas, M.; Gunawan, I.; Primeau,
J.; Largis, E.; Steiner, K. Bioorg. Med. Chem. Lett. 1999, 9,
1921. (b) Bloom, J. D.; Dutia, M. D.; Johnson, B. D.; Wissner,
A.; Burns, M. G.; Largis, E. E.; Dolan, J. A.; Claus, T. H. J.
Med. Chem. 1992, 35, 3081. (c) Hu, B. Unpublished results.
3. (a) For the synthesis of 4b and 4f see: Siegel, M. G.; Shu-
ker, A. J.; Droste, C. A.; Hahn, P. J.; Jesudason, C. D.;
Mcdonald, J. H., III; Matthews, D. P.; Rito, C. J.; Thorpe, A.
J. Mol. Div. 1998, 3, 113. (b) Bell, M. G.; Crowell, T. A.;
Matthews, D. P.; McDonald, J. H., III; Nee, D. A.; Shuker,
A. J.; Winter, M. A. U.S. Patent 708,621, 1998; Chem. Abstr.
1998, 129, 148910.
1
and arylethanolamines were characterized by H NMR, mass
spectrometry, and high-resolution mass spectrometry. All
of the compounds that are being biologicallyevaluated are
mixtures of diastereomers (the 2,4-thiazolidinedione adds a
chiral center to the structure). For experimental details, see:
Hu, B.; Sum, F.-W.; Malamas, M. PCT application, sub-
mitted, July, 2000.
13. The human b3 receptor was cloned as described in Van,
S. A.; Nahmias, C. K. S.; Briend-Sutren, M. M.; Strosberg,
A. D.; Emorine, L. J. Eur. J. Biochem. 1993, 213, 1117. The
human b2 and b1 receptors were cloned as described in Tate,
K. M.; Briend-Sutren, M. M.; Emorine, L. J.; Delavier-
Klutchko, C.; Marullo, S.; Strosberg, A. D. Eur. J. Biochem.
1993, 196, 357.
14. For radioligand binding procedure see: Zheng, W.; Niku-
lin, V. I.; Konkar, A. A.; Vansal, S. S.; Shams, G.; Feller,
D. R.; Miller, D. D. J. Med. Chem. 1999, 42, 2287.
15. For measuring thermogenesis in human b3-AR transgenic
mice and b3-AR knock out mice see: Susulic, V. S.; Frederich,
R. C.; Lawitts, J.; Tozzo, E.; Khahn, B. B.; Harper, M.-E.;
Himms-Hagen, J.; Flier, J. S.; Lowell, B. B. J. Biol. Chem.
1995, 270, 29483.
4. 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 and references
therein.