Synthesis and activity of novel bile-acid conjugated glucocorticoid receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2006.08.133

A series of potent steroidal glucocorticoid receptor antagonists has been discovered. After conjugation to cholic acid, the compounds retained an affinity for GR in vitro and had modest in vivo efficacy.

Full text of DOI:10.1016/j.bmcl.2006.08.133

Bioorganic & Medicinal Chemistry Letters 16 (2006) 6086–6090
Synthesis and activity of novel bile-acid conjugated
glucocorticoid receptor antagonists
Steven J. Richards,a,* Thomas W. von Geldern,a Peer Jacobson,a Denise Wilcox,a
a
b
b
b
b
¨
¨
Phong Nguyen, Lars Ohman, Marie Osterlund, Birgitta Gelius, Marlena Grynfarb,
Annika Goos-Nilsson,b Jiahong Wang,a Steven Funga and Masha Kalmanovicha
aMetabolic Disease Research, Abbott Laboratories, 100 Abbott Park Road, Abbott Park, IL 60064-6098, USA
bKaro Bio AB, Novum, SE-141 57 Huddinge, Sweden
Received 18 July 2006; revised 22 August 2006; accepted 28 August 2006
Available online 20 September 2006
Abstract—A series of potent steroidal glucocorticoid receptor antagonists has been discovered. After conjugation to cholic acid, the
compounds retained an affinity for GR in vitro and had modest in vivo efficacy.
Ó 2006 Elsevier Ltd. All rights reserved.
The glucocorticoid receptor (GR) is ubiquitously ex-
pressed in the body and regulates a myriad of functions.
In the liver, the endogenous GR ligand cortisol leads to
increased hepatic glucose production via the upregula-
tion of key gluconeogenic enzymes.1 The antiglucocorti-
coid RU-486 (1, Fig. 1), or mifepristone, has been
shown to acutely reduce hepatic glucose output suggest-
ing that GR blockade might be a useful approach for
treating the hyperglycemia associated with type 2 diabe-
tes.2 However, RU-486 also has deleterious extrahepatic
effects that limit its utility as a chronic diabetes therapy.3
Recent data in rodents further imply that selective
hepatic GR blockade would be beneficial to diabetic pa-
tients.4 In this study, the authors note that hepatic GR
knockout mice with diabetes had lower hepatic glucose
output when fasted. This is correlated with dramatically
reduced phosphoenolpyruvate carboxykinase (PEPCK)
mRNA induction. The therapeutic utility of steroidal
GR antagonists could potentially be expanded if they
could be targeted to the liver.
to concentrate the drug in the liver.5 The linchpin of this
strategy is the 11b aryl substituent, shown by the GR
crystal structure with RU-486 (1) bound to prevent helix
12 of the GR from adopting its active conformation.6
GR forms complexes with agonists that allow helix 12
to close over the ligand binding site.6 Because the 11b
aryl substituent is solvent exposed, attachment of cholic
acid or other bulky substituents is tolerated. A-348441
(2), a cholic acid conjugate of RU-486, was discovered
by successfully applying this strategy.7,8
As a follow-on strategy, we sought to redesign the ste-
roidal GR antagonist (3) to improve potency and then
prepare the bile-acid conjugates 4 in hopes that im-
proved in vitro potency of the parent GR antagonist
would result in improved potency in vivo.
An 11b aryloxy group might provide an alternate means
of attachment for the bile-acid. The corresponding non-
conjugated ethers were prepared by an Sn20 cuprate
addition to known epoxide 5, followed by acidic hydro-
lysis and ketal deprotection (Scheme 1).9
Our strategy for accomplishing this involves the cou-
pling of a potent GR antagonist with cholic acid in
hopes of engaging the bile-acid transporter machinery
Compounds were first evaluated in a GR binding assay
(Table 1). A radiolabeled form of GR agonist dexameth-
asone was used as the reference ligand. The functional
assay, abbreviated GRAF (‘GR linked to Alkaline Fos-
fatase’), was a cell-based format using CHO cells engi-
neered to express GR.7 They have been further
modified such that GR activation stimulates the
Keywords: Glucocorticoids; Cortisol; Prednisolone; Antagonist; Glu-
cocorticoid receptor; RU-486; Steroid; Bile-acid; Liver-targeting; Ch-
olic acid.
*
Corresponding author. Tel.: +1 8479359575; fax: +1 8479381674;
e-mail: steven.richards@abbott.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.08.133

S. J. Richards et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6086–6090
6087
6i–l, was tolerated only if the substituents were small. A
simple meta-substituent in compounds 6h and 6c led to a
boost in potency, while meta-di-substituted analogs were
less potent than their mono-substituted comparators
(6m and 6d). A significant drop in potency was observed
if both meta-substituents were larger than methyl (6n).
Compounds 6k and 6o indicated that the position of
difluoro substituents on the aromatic ring could modu-
late potency.
Compound 6h stood out as an exceptionally potent
antagonist, likely the most potent steroidal GR antago-
nist reported to date.11 It was 10 times more potent than
RU-486 (1) in the GR binding assay and 5 times more
potent in the GRAF functional assay. The meta-dimeth-
yl analog 6m was also shown to be about twice as potent
in vitro as RU-486 (1). The biphenyl ether 6e was found
to be equipotent. These three compounds were selected
for conjugation to cholic acid and evaluation in vivo.
The cholic acid coupling partner was prepared as de-
scribed in Scheme 2. The synthesis of alcohol 7 was com-
pleted as described in the literature.12 The ethylene
glycol linker was then activated as the mesylate to pro-
vide 8. The GR antagonist portion of the molecule
was again prepared from epoxide 5. Phenols 9a–c were
alkylated with allyl bromide. The protected aryl bro-
mides (10a–c) were then subjected to the cuprate cou-
pling to epoxide 5. After the addition, acidic
hydrolysis and ketal deprotection proceeded smoothly.
The allyl protecting group was removed using phenyl si-
lane under Pd0 catalysis. The two halves of the conju-
gate, 8 and 11a–c, were coupled over the course of
24 h with slight warming in the presence of base. At-
tempts to increase the reaction rate by raising the tem-
perature led to lower product recovery, presumably
due to decomposition of the starting materials. Saponi-
fication of the methyl ester furnished conjugates 4a–c.
Figure 1. Representative steroidal GR antagonists.
Bile-acid conjugates 4a–c were found to be potent GR
antagonists (Table 2). The bile-acid did cause a decrease
in binding affinity and functional potency relative to the
parent compounds (6h, 6e, and 6m). Nevertheless, com-
pounds 4a–c were taken into a prednisolone challenge
experiment using male CD rats.7 In this experiment, effi-
cacy of the GR antagonist was related to how much it
blunted the effects of GR agonist prednisolone on two
hepatic parameters, tyrosine aminotransferase (TAT)
upregulation and glycogen deposition. To assay for liv-
er-targeting, the attenuation of prednisolone induced
lymphopenia, a systemic parameter, was also moni-
tored. The ideal liver-selective GR antagonist would
have effects on TAT and glycogen and no effect on
lymphopenia.
Scheme 1. Reagents and conditions: (a) ArBr, Mg, I2, THF, reflux
15 min, then 0 °C, CuI, 5; (b) 2 M HCl, THF, rt, 5 min (20–50% 2
steps).
secretion of alkaline phosphatase. Effective antagonists
in this assay decreased dexamethasone stimulated pro-
duction of alkaline phosphatase.
The SAR for this series of analogs showed that a variety
of 11b substituents were tolerated by the receptor. para-
OMe substitution led to known compound 6a, which
was roughly equipotent to RU-486 (1) in both binding
and functional assays.10 Similarly, the literature com-
pound 6b, with a para-SMe substituent, also had compa-
rable potency.10 para-O-Alkyl derivatives of various
sizes were tolerated (6c–e and known compound 6f).10
para-Phenols tended to be less active than their O-alkyl-
ated homologs (compare the significant drop in potency
of 6g relative to 6h). ortho-Substitution, as evidenced by
When administered as a single 100 mg/kg oral dose to
rats followed by the prednisolone challenge, RU-486
(1) effectively nullified the effect of prednisolone on
TAT and liver glycogen. It was also systemically avail-
able as was demonstrated by reversal of prednisolone in-
duced lymphopenia (Table 2). A-348441 (2) was also
efficacious against the hepatic parameters but had no ef-
fect on lymphopenia. GR antagonist 6h effectively

6088
S. J. Richards et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6086–6090
Table 1. SAR for 11b-aryl substituents
a
GR IC50 (nM)
b
GRAF IC50 (nM)
Compound
R1
R2
R3
R4
R5
1
H
H
NMe2
OMe
SMe
H
H
H
H
H
H
H
H
H
H
OMe
F
1.1
0.9
5.0
7.6
6a
6b
6c
6d
6e
6f
H
H
H
H
H
H
0.6
0.4
1.8
4.7
H
Cl
Cl
H
O-allyl
O-allyl
4-OMePh
OPh
H
H
Me
H
1.5
0.7
20.9
4.2
H
H
H
H
1.5
4.5
18.8
28.6
1.1
6g
6h
6i
H
Me
Me
H
OH
H
H
OMe
OMe
OMe
OMe
OMe
OMe
OMe
OMe
H
0.1
OMe
OMe
F
H
481.5
6.1
0.7
nd
6j
H
H
43.0
10.1
32.0
2.8
6k
6l
H
H
F
Me
H
H
H
H
H
H
F
9.0
0.4
6m
6n
6o
Me
OMe
F
Me
OMe
H
H
60.7
3.2
nd
15.0
H
Human GR binding assay. GRAF cellular functional assay.
a Values are means of two experiments.
b nd, Not determined.
Scheme 2. Reagents and conditions: (a) MsCl, Et3N, THF, rt, o.n., (37%); (b) allyl bromide, Cs2CO3, rt, 20 h (45%); (c) 10a–c, Mg, I2, THF, reflux
15 min, then 0 °C, CuI; (d) 2 M HCl, THF, rt, 5 min (21% 2 steps); (e) Pd(Ph3)4, PhSiH3, CH2Cl2, rt, 1 h (55%); (f) 8, n-Bu4NI, Cs2CO3, THF, 55 °C,
24 h (93%); (g) 1 M LiOH, THF.
blocked hepatic TAT and lymphopenia in the periphery,
similar to RU-486 (1). Conjugate 4a was only marginal-
ly active in vivo. It had no effect on glycogen and mini-
mal, equal effects on TAT and prednisolone induced
lymphopenia. Compound 4b showed minimal effects
on glycogen deposition. Conjugate 4c was active in this

S. J. Richards et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6086–6090
6089
Table 2. In vitro and in vivo data for bile-acid conjugates 4a–c and comparison to antagonists A-348441 (2), 6h, and RU-486 (1)
GR IC50 (nM) GRAF IC50 (nM) TAT (%) Gly.c (%) Lymph. (%) Plasma Conc.b,c (nmol/mL) Liver Conc.b,c (nmol/g)
a
Compound
1
1.1
1.2
0.1
3.1
4.9
2.5
5
10
101
72
97
31
0
77
86
nd
0
104
9
3.9
4.5
4.8
4.5
nd
26.5
26.5
18.1
22.2
nd
2
6h
4a
4b
4c
1.1
30.7
200
23.5
67
20
0
12
51
44
40
nd
nd
a Values are means of two experiments.
b Data collected 7 h after 100 mg/kg oral dose. There were 5 animals per group.
c nd, Not determined.
gates 4a–c in vivo. Further studies in this area would
also focus on a re-examination of the conjugation of po-
tent GR antagonists such as 6h.
Table 3. Rat ADME properties for selected compounds
Compound
Microsomal stabilitya
(% remaining)
Plasma bindingb
(% bound)
1
70
98
95.6
95.9
nd
2
Acknowledgments
6h
4a
6e
4c
87
95
98.3
nd
We thank Jim Schmidt, Sue Swanson, and Maury Em-
ery for performing the metabolism and plasma protein
binding assays. We also acknowledge JT Link and Andy
Souers for their critical review of this manuscript.
96
100
nd
a Twenty minutes incubation period.
b nd, Not determined.
References and notes
model. However, 4c was significantly less active than
RU-486 (1) and, like 1, had peripheral as well as hepatic
effects.
1. (a) Andrews, R. C.; Walker, B. R. Clin. Sci. 1999, 96, 513;
(b) Kurukulasuriya, R.; Link, J.; Madar, D.; Pei, Z.;
Richards, S.; Rohde, J.; Souers, A.; Szczepankiewicz, B.
Curr. Med. Chem. 2003, 10, 123.
`
2. Garrel, D. R.; Moussali, R.; Oliveira, A. D.; Lesiege, D.;
`
In order to better understand the disparity between
potency in vitro and lack of in vivo efficacy of the O-
linked conjugates, we first measured the concentration
of 4a in rat liver 7 h after administration of the GR
antagonist (Table 2). The concentration of 4a in the liver
compared favorably to effective liver concentrations of
RU-486 (2) and A-348441 (1).
Lariviere, F. J. Clin. Endocrinol. Metab. 1995, 80, 379.
3. Bamberger, C.; Chrousos, G. Ann. NY Acad. Sci. 1995,
761, 296.
4. Opherk, C.; Tronche, F.; Kellendonk, C.; Kohlmuller, D.;
¨
¨
Schulze, A.; Schmid, W.; Schutz, G. Mol. Endocrinol.
2004, 18, 1346.
5. (a) Apelqvist, T.; Wu, J.; Koehler, K International Patent
WO 0058337, 2000; (b) Wess, G.; Kramer, W.; Schulbert,
G.; Enhsen, A.; Baringhaus, K.; Glombik, H.; Mullner, S.;
¨
Conjugates 4a and 4c were also examined for stability in
rat liver microsomes (Table 3). Both compounds were
inert to oxidative metabolism. This was also true for
A-348441 (1). Apparently conjugation to cholic acid
had the effect of protecting these steroidal GR antago-
nists from oxidative clearance as the parent compounds
6e and 6h as well as RU-486 (1) were rapidly degraded.
Bock, K.; Kleine, H.; John, M.; Neckermann, G.; Hoff-
mann, A. Tetrahedron Lett. 1993, 34, 819; (c) Kurukula-
suriya, R.; Link, J.; Madar, D.; Pei, Z.; Richards, S.;
Rohde, J.; Souers, A.; Szczepankiewicz, B. Curr. Med.
Chem. 2003, 10, 99.
˚
6. Kauppi, B.; Jakob, C.; Fa¨rnegardh; Yang, J.; Ahola, H.;
Alarcon, M.; Calles, K.; Engstro¨m; Harlan, J.; Munch-
¨
more, S.; Ramqvist, A.; Thorell, S.; Ohman, L.; Greer, J.;
Gustafsson, J.; Carlstedt-Duke, J.; Carlquist, M. J. Biol.
Chem. 2003, 278, 22748.
After discounting low liver concentration and metabolic
instability as reasons for the modest efficacy of the con-
jugates, we measured protein binding of conjugate 4a in
rat plasma and found it to be >98% bound.
7. von Geldern, T.; Tu, N.; Kym, P.; Link, J.; Jae, H.; Lai,
C.; Apelqvist, T.; Rhonnstad, P.; Hagberg, L.; Koehler,
K.; Grynfarb, M.; Goos-Nilsson, A.; Sandberg, J.;
¨
Osterlund, M.; Barkem, T.; Ho¨glund, M.; Wang, J.;
Fung, S.; Wilcox, D.; Nguyen, P.; Jakob, C.; Hutchins, C.;
¨
Fa¨rnegardh, M.; Kauppi, B.; Ohman, L.; Jacobson, P.
In summary, a series of 11b aryloxy modified steroid
GR antagonists was discovered to have comparable or
superior potency, in vitro, to RU-486 (1). However, de-
spite achieving a comparable liver concentration of 4a,
the conjugate was not as effective as RU-486 (1) or its
conjugate A-348441 (2) in vivo. Differences in protein
binding do not fully explain the observed differences in
in vivo activity. Clearly other mechanisms worked in
concert to limit the efficacy of 4a. Additional metabo-
lism experiments could shed light on the fate of conju-
˚
J. Med. Chem. 2004, 47, 4213.
¨
¨
8. Jacobson, P.; von Geldern, T.; Ohman, L.; Osterland, M.;
Wang, J.; Zinker, B.; Wilcox, D.; Nguyen, P.; Mika, A.;
Fung, S.; Fey, T.; Goos-Nilsson, A.; Grynfarb, M.;
Barkhem, T.; Marsh, K.; Beno, D.; Nga-Nguyen, B.;
Kym, P.; Link, J.; Tu, N.; Edgerton, D.; Cherrington, A.;
Efendic, S.; Lane, B.; Opgenorth, T. J. Pharmacol. Exp.
Ther. 2005, 314, 191.

6090
S. J. Richards et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6086–6090
9. Belanger, A.; Philibert, D.; Teutsch, G. Steroids 1981, 37,
361.
11. In vitro potency data for many steroidal GR antagonists
have not been reported as IC50s but as relative binding
affinities to GR agonist ligands. See Teutsch, G.; Gaillard-
Moguilewsky, ; Lemoine, G.; Nique, F.; Philibert Bio-
chem. Soc. Trans. 1991, 19, 901.
12. Wess, G.; Kramer, W.; Bartmann, W.; Glombik, E.;
Mullner, S.; Bock, K.; Dries, A.; Kleine, H.; Schmitt, W.
Tetrahedron Lett. 1992, 33, 195.
10. (a) Compounds 6a, 6b, and 6f have been reported in the
following references: (a) Philibert, Daniel; Teutsch, J. G.;
Costerousse, G.; Daraedt, R. German Patent DE 3307143,
1983; (b) Philibert, D.; Hardy, M.; Gaillard-Moguilewsky,
M.; Nique, F.; Tournemine, C.; Nedelec, L. J. Steroid
Biochem. 1989, 34, 413.