Discovery of potent and selective nonsteroidal indazolyl amide glucocorticoid receptor agonists

Article abstract of DOI:10.1016/j.bmcl.2013.06.089

Modification of a phenolic lead structure based on lessons learned from increasing the potency of steroidal glucocorticoid agonists lead to the discovery of exceptionally potent, nonsteroidal, indazole GR agonists. SAR was developed to achieve good selectivity against other nuclear hormone receptors with the ultimate goal of achieving a dissociated GR agonist as measured by human in vitro assays. The specific interactions by which this class of compounds inhibits GR was elucidated by solving an X-ray co-crystal structure.

Full text of DOI:10.1016/j.bmcl.2013.06.089

Bioorganic & Medicinal Chemistry Letters 23 (2013) 5442–5447
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of potent and selective nonsteroidal indazolyl amide
glucocorticoid receptor agonists
a
a
a
a
James E. Sheppeck II b, , John L. Gilmore , Hai-Yun Xiao , T. G. Murali Dhar , David Nirschl ,
Arthur M. Doweyko a, Jack S. Sack a, Martin J. Corbett a, Mary F. Malley a, Jack Z. Gougoutas a,
Lorraine Mckay a, Mark D. Cunningham a, Sium F. Habte a, John H. Dodd a, Steven G. Nadler a,
John E. Somerville a, Joel C. Barrish a
⇑
a Research and Development, Bristol-Myers Squibb Company, Princeton, NJ 08543-4000, United States
b Ironwood Pharmaceuticals, 301 Binney St., Cambridge, MA 02142, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 9 July 2013
Modification of a phenolic lead structure based on lessons learned from increasing the potency of steroi-
dal glucocorticoid agonists lead to the discovery of exceptionally potent, nonsteroidal, indazole GR ago-
nists. SAR was developed to achieve good selectivity against other nuclear hormone receptors with the
ultimate goal of achieving a dissociated GR agonist as measured by human in vitro assays. The specific
interactions by which this class of compounds inhibits GR was elucidated by solving an X-ray co-crystal
structure.
Keywords:
GR agonist
Glucocortocoid receptor
Dissociated GR inhibitors
Ó 2013 Published by Elsevier Ltd.
For over 60 years, glucocorticoid receptor (GR) agonists such as
hydrocortisone, prednisone, and dexamethasone have been a
mainstay in the treatment of inflammatory and allergic diseases.
However, their prolonged systemic use has been limited by numer-
ous accompanying side effects such as diabetes, osteoporosis, skin
thinning, muscle atrophy, and others.1 While inhaled glucocorti-
coids such as fluticasone have avoided these side effects by directly
targeting the lung for the treatment of asthma coupled with a very
short systemic half-life, there remains a significant clinical need for
a GR agonist that could be administered systemically that pos-
sesses a safer side effect profile.2 A number of studies have demon-
strated that glucocorticoid-mediated agonism of GR leads to
nuclear translocation of the receptor that, in concert with addi-
pro-inflammatory genes could in principle be ‘dissociated’ from
transactivation of genes ultimately associated with the adverse ef-
fects of glucocorticoid treatment.5–8
In the course of our research towards nonsteroidal GR agonists
that might exhibit a dissociated profile, we attempted to improve
OAc
OH
O
OH
O
HO
HO
OH
R'
H
H
N
H
H
R
H
N
O
tional cofactors, inhibits the transcription factors NF-jB and acti-
vator protein-1 (AP-1) resulting is the suppression of pro-
inflammatory genes under their control (transrepression).3,4 Alter-
natively, the agonist-activated GR complex can homodimerize and
bind to glucocorticoid response elements (GREs) to initiate the
transcription of genes directly that are primarily involved in
metabolism (transactivation).5 The report of a GR dimerization-
deficient transgenic mouse that remained sensitive to the anti-
inflammatory effects of dexamethasone (croton oil-induced ear
edema was reduced compared to wild type) but lacked the typical
dexamethasone-induced side effects provided pharmacological
justification for the hypothesis that transrepression of
1
2
3
R = F, R' = Me, dexamethasone
R = H, R' = H, prednisolone
fluorocortivazol10
"super" GR agonist
F
O
S
O
S
N
N
N
N
N
H
N
H
N
N
HO
BMS / Yang Lead9
weak GR partial agonist
5
4
a non-steroidal
GR full agonist
F
⇑
Corresponding author.
E-mail address: jim.sheppeck@bms.com (J.E. Sheppeck).
Figure 1. Genesis of the indazolyl class nonsteroidal GR agonist.
0960-894X/$ - see front matter Ó 2013 Published by Elsevier Ltd.
http://dx.doi.org/10.1016/j.bmcl.2013.06.089

J. E. Sheppeck II et al. / Bioorg. Med. Chem. Lett. 23 (2013) 5442–5447
5443
upon the modest GR activity of a previously discovered phenolic
series 49 by structurally combining it with fluorocortivazol 3, one
of the most potent glucocorticoids ever reported. (see Fig. 1)10
We hypothesized that the A-ring of prednisolone was being mim-
icked by the phenol of 4 which was corroborated by modeling of 4
in the dexamethasone/GR ligand binding domain co-crystal struc-
ture (vide infra). Since the glucocorticoid receptor has evolved to
bind the steroid ring system (even nonnatural glucocorticoids such
as 3), the fusion of 3 and 4 to create indazole 5 did not appear to be
an unreasonable proposition. Furthermore, if only some of the po-
tency of fluorocortivazol was imparted to lead series 4, it would
improve its chances to progress farther in lead optimization. Use
of 4-fluorophenyl-substituted indazoles and pyrazoles to enhance
the potency of GR agonists was initially reported by the Scanlan
group in 2005—a strategy employed by a number of groups
since.8c,d If successful, the indazole chemotype would dramatically
simplify the structural requirements of a GR agonist from a steroi-
dal 4-ring architecture containing 8 contiguous chiral centers to
only aromatic rings and a single chiral center.
The indazoles in this and the accompanying Letter were all syn-
thesized from the common synthetic intermediate 5-hydroxyme-
thylindazole (Scheme 1) prepared using the method of Sun
et al.11 Oxidation of the hydroxymethyl moiety using Dess–Martin
reagent followed by the addition of a Grignard reagent or alkyl/aryl
lithium reagent provided an appropriately substituted secondary
alcohol. The alcohol reacted smoothly with silyl ketene acetals in
the presence of TiCl4 using a modification of the Mukaiyama aldol
reaction,12 to provide the indazole esters which were hydrolyzed
to give the corresponding acids. Substitution at the N1 position
of the indazole was accomplished using a Buchwald arylation pro-
cedure13 or by alkylation using sodium hydride and alkyl iodides to
give the appropriately substituted indazoles. Finally, coupling with
a variety of amines using standard coupling conditions afforded
the desired amides.
invariably had greater activity in the transactiviation assay. To this
end we strove to discover compounds that had good partial agonist
activity (EC50 <50 nM) whose agonist efficiencies were in the range
of 65–90% relative to dexamethasone (100%).
Table 1 summarizes the GR binding and functional activity for a
series of indazole-based GR agonists possessing a diverse set of
indazole N-linked substituents using dexamethasone and com-
pound 4 as positive controls. Our initial SAR effort was directed
at establishing whether the phenol of compound 4 could indeed
be replaced by an indazole. The enantiomeric16 indazoles 5a and
5b both exhibited good binding to the GR (as did essentially all
of the analogs in Table 1) but even similar activity than the phenol
4 in the AP-1 and E-selectin assays with the (3S) stereoisomer 5a
being the more potent of the two. Nevertheless, the efficacy of 5a
relative to dexamethasone showed it to be a partial agonist just
like the phenol 4. The indazole nitrogen was substituted with an
phenyl group to give 5b which showed a dramatic increase in
AP-1 activity. The 4-fluorophenyl analog 5c which is the same
moeity as in fluorocortivasol also showed good activity as a race-
mic mixture. This compound was separated into its enantiomers
(3S)-5c and (3R)-5c. The (3S) stereoisomer showed exceptional
activity both in the AP-1 and E-selectin assays (2.4 and 4 nM,
respectively) and was 40ꢀ more potent than the (3R) enantiomer.
Though close in activity to dexamethasone, compound (3S)-5c
showed strong NP-1 agonism as well, implying no dissociation be-
tween transrepression and transactivation like dexamethasone it-
self. Changing from a thiadiazole amide to a thiazole amide (5d)
gave compounds with similar GR binding as well as functional
activities including the undesired potent activity in the NP-1 assay.
The thiazole amide derivatives also showed much greater potency
in the (3S) stereoisomer. Next, a series of changes to the phenyl
ring were studied (5e–5l). Moving from a 4-F to a 4-Cl gave a com-
pound with similar activities in the binding and functional assays,
however changing to a 4-Br group led to a loss of binding and func-
tional activity. The activity profile of the three pyridyl isomers are
shown in compounds 5g–5i. The 3-pyridyl compound (5h) was the
most potent of these isomers and was further analyzed as the
homochiral analogs.16 Compound (3S)-5h shows strong GR binding
and functional activities but again exhibited significant activity in
the transactivation assay though it was less active in this assay
than dex. Compounds 5m–5t show a series of changes in which
the N-linked indazole substituent was changed from aryl to alkyl.
Most of the N-alkyl analogs gave significantly less active com-
pounds in the functional assays with the exception of the isopentyl
(5o) and cycloheptyl (5t) substitutions. Unfortunately, the latter
two more active compounds also showed increased potency in
the NP-1 transactivation assay.
Structure–activity relationships for the indazole GR agonist ser-
ies are detailed in Tables 1–4 and selectivity against related nucle-
ar hormone receptors is found in Table 5. The in vitro assays used
to assess the biological activity of the GR ligands prepared in
Scheme 1 consisted of: a GR binding assay, two assays to measure
transrepression in an A549 cell line (an AP-1 and an E-selectin (NF-
jB-dependent) repression assay), and a previously reported trans-
activation assay (a NP-1 agonist assay that contains a GR chimera
with a GAL4 reporter in a HeLa cell line).14,15 Based on literature
precedence, the desired activity profile of a ‘dissociated’ GR agonist
is a compound which is a partial agonist of the in vitro functional
assays for transrepression while showing little to no activity in the
transactivation assay.8,9 Full agonists in the transrepression assays
OH
OH
d, e
a,b,c
OH
f, g
Y
N
N
N
H
H2N
N
H
O
Y
O
Y
O
R
R
R
Z
OH
N
OH
H
h or i
j
N
Y
N
N
N
N
N
H
X
X
Scheme 1. Reagents and conditions: (a) Ac2O (3 equiv), KOAc (2 equiv), CHCl3 reflux, 2 h; (b) isoamyl nitrite (2.2 equiv), 18-crown-6 (0.05 equiv), reflux 24 h; (c)1 M NaOH
(1 equiv), MeOH, 1 h; (d) Dess–Martin reagent (1.2 equiv), DCM, 2 h; (e) Y-MgBr or Y-Li (1.1 equiv), THF, 0 °C or ꢁ78 °C, 3 h; (f) (1-methoxy-2-methylprop-1-
enyloxy)trimethylsilane [R = CH3] or (E)-(1-methoxyprop-1-enyloxy)trimethylsilane [R = H] (1.2 equiv), TiCl4, (1.1 equiv), DCM, 0 °C, 12 h; (g) 1 M NaOH, DMSO, MeOH,
100 °C, 12 h; (h) CuI (0.05 equiv), K3PO4 (1.8 equiv), t-cyclohexanediamine (0.5 equiv), X-I (1 equiv), dioxane, 110 °C, 24 h; or (i) NaH (2 equiv), X-I (1.5 equiv), THF, 3 h; (j) Z-
NH2 (1 equiv), EDC (1.3 equiv), HOBt (1.3 equiv), TEA (2 equiv), DCM, 12 h or cyanuric fluoride (1.1 equiv), pyridine (1.1 equiv), DCM, 2 h, then Z-NH2 (1 equiv), DMAP (cat),
THF, 80 °C, 12 h.

5444
J. E. Sheppeck II et al. / Bioorg. Med. Chem. Lett. 23 (2013) 5442–5447
Table 1
In vitro GR binding and functional activity for indazole N-linked substituents
O
S
J
N
N
H
N
N
X
Compd
X
J
GR bindinga Ki (nM)
AP-1 repressionb
E-selectin repressionb
NP-1 agonismc
%Dexd
EC50 (nM)
%Dexd
EC50 (nM)
% dexd
a
a
a
EC50 (nM)
Dex
4
(3S)-5a
(3R)-5a
5b
1.2
1.1
2.7
13
5.0
18
5.1
6.7
15
6.0
7.9
18
2.5
21
24
617
20
23
2.4
92
4.1
1.4
400
23
956
127
19
349
106
1200
56
271
301
114
45
550
1020
1932
>10,000
29
100
71
79
35
73
81
93
50
79
92
56
96
x
62
74
58
56
x
76
69
82
76
96
32
45
62
x
1.1
21
15
100
71
77
x
4.5
526
208
>10,000
ND
ND
5.8
ND
ND
7.6
435
256
ND
ND
132
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
132
ND
ND
ND
ND
100
28
41
H
H
N
N
N
N
N
N
C
>5000
ND
ND
4
ND
ND
0.5
58
x
Phenyl
ND
ND
86
ND
ND
97
81
84
ND
ND
77
35
56
ND
62
x
44
77
87
27
ND
ND
ND
86
ND
ND
106
ND
ND
94
59
55
ND
ND
57
5c
4-F-phenyl
4-F-phenyl
4-F-phenyl
4-F-phenyl
4-F-phenyl
4-F-phenyl
4-Cl-phenyl
4-Br-phenyl
2-Pyridyl
3-Pyridyl
3-Pyridyl
4-Pyridyl
3-MeOPh
4-MeOPh
4-HOPh
(3S)-5c
(3R)-5c
5d
(3S)-5d
(3S)-5d
5e
5f
5g
(3S)-5h
(3R)-5h
5i
5j
5k
5l
5m
5n
5o
5p
5q
C
C
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
10
29
ND
ND
19
185
78
ND
143
>5000
195
1823
49
1047
ND
ND
ND
21
8.1
0.9
4.0
1.9
36
8.8
7.2
10
19
5.7
47
17
14
x
x
x
x
x
x
66
ND
ND
ND
ND
60
Isopropyl
Isobutyl
Isopentyl
2-Hydroxyethyl
3-Hydroxypropyl
Benzyl
5r
5s
5t
PhCH2CH2-
cycloheptyl
46
12
85
188
Dex is dexamethasone, x means <30%, and ND means not determined.
a
Values are means of two or more experiments performed in triplicate. All data is on racemates unless specified otherwise.
Activator protein-1 (AP-1) and E-selectin assays were performed in an A549 lung epithelial cell line.
GR transactivation NP-1 assay (run in agonist mode) was performed in the HeLa cell line.
b
c
d
efficacy is represented as a percentage of the maximal response of dexamethasone (100%).
Table 2 summarizes the GR binding and AP-1 activity for a ser-
4-methyl derivative resulted in a compound with comparable
AP-1 activity and agonist efficiency. Larger groups on the thiazole
and thiadiazole led to compounds with decreased activity (data
not shown).
ies of indazole-based GR agonists possessing a diverse set of inda-
zole side chains (Y) using dexamethasone and compounds 5c and
5d as positive controls. The compound lacking any substituent at
the Y position (6a) had much lower GR binding and was devoid
of AP-1 activity. Substitution at this position with a variety of alkyl
groups (6a–6h) led to increased activity in both the GR binding and
AP-1 assay underscoring the requirement for a hydrophobic but
not necessarily an aromatic group at this position of the ligand.
Of the alkyl substituents tested, the most potent compounds were
the propyl and isobutyl substitutions (6c and 6e) which have a Ki’s
of 1.7 and 7.5 nM in the GR binding assay and an EC50 of 18 and
29 nM respectively in the AP-1 assay. Replacing the phenyl group
with benzyl or pyridyl led to compounds with significantly less po-
tent activity in the AP-1 assay.
Table 3 summarizes the GR binding and AP-1 activity for a ser-
ies of indazole-based GR agonists possessing a diverse set of amide
groups (Z) again using dexamethasone and compounds 5c and 5d
as positive controls. In general, the amide groups presented in this
table exhibited good GR binding. Changing the amide to a 3-thienyl
led to some loss of AP-1 activity and the tetrazolyl led to complete
loss of AP-1 activity. The 2-triazolyl compound had comparable
activity to the 2-thiadiazolyl. Methylation on the various positions
of either the thiazole or thiadiazole was then explored. Both
5-methyl derivatives led to some loss of AP-1 activity while the
Table 4 shows the activity profiles of the compounds in which
the gem dimethyl substitution at the C2-position is replaced with
a single methyl group. Each of the four stereoisomers of the mono-
methyl product was isolated using chiral HPLC and tested dis-
cretely.16 One of the isomers, 8a, gave modest GR binding
activity and poor activity in the AP-1 assay. The other three iso-
mers were all very potent in both the AP-1 and NP-1 assays with
compound 8c showing exceptional activity. This compound was
in fact more potent than dexamethasone with a GR binding Ki of
0.48 nM and EC50 of 0.30 nM in the AP-1 assay. Yet again, potent
AP-1 agonism appeared to be accompanied by potent NP-1 agon-
ism with an EC50 of 0.10 nM in the NP-1 assay.
To monitor selectivity against other nuclear hormone receptors,
a subset of compounds was screened in binding or activation as-
says for the androgen receptor (AR), the progesterone receptor
(PR), and the mineralocorticoid receptor (MR).17 Compounds
(3S)-5c, (3R)-5c, 6c and 8c all exhibited over 1000 fold selectivities
between GR and AR. Compound 8b showed the lowest selectivity
between GR and AR at around 400 fold. A modest selectivity was
observed between GR and PR with a 10–50 fold separation of bind-
ing affinities for all the compounds except 8b which has a 350 fold

J. E. Sheppeck II et al. / Bioorg. Med. Chem. Lett. 23 (2013) 5442–5447
5445
Table 4
Table 2
Conformational effects of 2-methyl groups
Structure–activity relationship of indazole side chain
O
S
O
S
N
J
2
3
N
N
N
H
N
H
Y
N
N
N
N
F
F
GR bindinga Ki (nM)
AP-1 repressionb
Compd
GR bindinga Ki (nM)
AP-1 repressionb
NP-1 agonismc
Compd
Y
J
EC50 (nM) %Dexd EC50 (nM) %Dexd
a
a
%Dexc
a
EC50 (nM)
Dex
(3S)-5c
(3R)-5c 6.7
8a
8b
8c
8d
1.2
5.1
2.5
2.4
92
453
13
100
93
50
4.5
5.8
—
>100
27
100
106
—
—
76
Dex
5c
5d
6a
6b
6c
6d
6e
6f
1.2
18
15
2.5
23
4.1
4322
87
18
81
29
705
82
100
81
79
36
62
60
29
60
37
47
55
60
Phenyl
Phenyl
H
Ethyl
N
C
N
C
N
N
N
N
C
44
42
84
0.59
0.48
2.4
103
106
89
9.8
1.7
2.4
7.5
1.1
21
0.30
20
0.10
20
122
100
Propyl
Isopropyl
Isobutyl
Benzyl
Cyclopentyl
Cyclohexyl
4-Pyridyl
a
Values are means of two or more experiments performed in triplicate. 8b and 8c
are diastereomeric at C2 and assumed to be (3S).
Activator protein-1 (AP-1) and E-selectin assays were performed in an A549
lung epithelial cell line.
6g
6h
6i
b
C
C
18
12
66
389
c
GR transactivation NP-1 assay (run in agonist mode) was performed in the HeLa
a
cell line.
Values are means of two or more experiments performed in triplicate. All data is
on racemates.
d
efficacy is represented as a percentage of the maximal response of dex (100%).
b
Activator protein-1 (AP-1) was performed in an A549 lung epithelial cell line.
Efficacy is represented as a percentage of the maximal response of dex (100%).
c
ocortivazol with the GR LBD from Suino-Powell et al20 for compar-
ison. In this structure, only Gln570 is engaged by the pyrazole of
cortivasol and the large phenyl group forces an expansion of the
binding pocket to accommodate it. Interestingly, the unsubstituted
indazole of (3S)-5a is rotated 180 degrees relative to the orienta-
tion of cortivazol and hydrogen bonds to Gln570 as a hydrogen
bond donor and an acceptor to Arg611. This indazole orientation
may be an artifact, however, since Met604 has been mutated to a
Leu to achieve diffraction-quality crystals and does contact the li-
gand. The 2-amidothiadiazole forms a bidentate hydrogen bond
to Asn564 on helix 3 but also accepts a hydrogen bond from
Gln642 on helix 11—the same residues engaged by cortivazol but
selectivity over PR. Finally, neither compound (3S)-5c nor (3R)-5c
showed agonist activity in a cellular MR activation assay.
During the course of our lead optimization effort on the inda-
zole series, we were able to successfully solve the crystal structure
of compound (3S)-5a co-crystallized with the GR ligand binding
domain (GR LBD).18 In Figure 2 above, the left panel shows the
2.8 Å resolution co-crystal structure of compound (3S)-5a with
the GR ligand binding domain. The indazole forms two hydrogen
bonds to Arg611 and Gln570 which have been observed crystallo-
graphically to engage the enone of dexamethasone.19 The far right
panel shows a 2.5 Å resolution co-crystal structure of fluor-
Table 3
Structure–activity relationship of the amide group
O
Z
N
H
N
N
F
Compd
Z
GR bindinga Ki (nM)
AP-1 repressionb
a
EC50 (nM)
%Dexc
Dex
5c
5d
7a
7b
7c
7d
7e
7f
1.2
18
15
11
11
17
11
14
21
2.5
23
4.1
73
17
>10,000
49
29
141
100
81
79
68
73
—
48
74
56
2-Thiadiazolyl
2-Thiazolyl
3-Thienyl
2-Triazolyl
2-Tetrazolyl
2-(5-Me-thiazolyl)
2-(4-Me-thiazolyl)
2-(5-Me-1,3,4-thiadiazolyl)
a
b
c
Values are means of two or more experiments performed in triplicate. All data for on racemates.
Activator protein-1 (AP-1) was performed in an A549 lung epithelial cell line.
Efficacy is represented as a percentage of the maximal response of dex (100%).

5446
J. E. Sheppeck II et al. / Bioorg. Med. Chem. Lett. 23 (2013) 5442–5447
Table 5
In vitro NHR selectivity of selected examples
Compd
GR bindinga Ki (nM)
AR binding Ki (nM)
PR binding Ki (nM)
MR activation
(3S)-5c
(3R)-5c
6c
8b
8c
6.7
5.1
1.7
0.59
0.48
>25000
>25000
2513
243
>4,166
107
56
109
214
22
>5000
>5000
—
—
—
a
Values are means of two or more experiments performed in triplicate.
Figure 2. (Left) 2.8 Å resolution co-crystal structure of compound (3S)-5a with the GR ligand binding domain. The M604L point mutation shown in purple was necessary to
achieve diffraction-quality crystals. (Center) molecular model of compound (3S)-5c docked into the GR LBD-dex co-crystal structure (1m2z) after removal of dex. (Right) 2.5 Å
resolution co-crystal structure of deacylcortivazol in the GR LBD.20
in a fundamentally different way. Asn564 has been identified as a
critical amino acid that acts as a conformational trigger by closing
the distance between helix 11 and 3 which results in a protein con-
formational change of helix 12 (not shown), the position of which
determines cofactor recruitment to the GR LBD-ligand complex
and receptor agonism or antagonism.21 The decrease in the func-
tional activity of 5c and 5d relative to 7a (replacement of the 2-
thiazolyl or 2-thiadiazolyl rings with an isosteric 2-thienyl group)
is readily explained in light of the crystal structure since the latter
can only hydrogen bond to Asn564 via a monodentate interaction.
The center panel contains a docked model of compound (3S)-5c
in the GR LBS crystal structure wherein an interaction of the fluo-
rophenyl-substituted indazole with Gln570 is seen to mimic that of
cortivazol which cannot rotate the indazole to the conformation
observed in the crystal structure of (3S)-5a (left panel).22 The
amidothiadiazole of (3S)-5c was found to hydrogen bond to
Asn564 and Gln642 as observed for (3S)-5a. The improved potency
of (3S)-5c over (3S)-5a observed by adding the 4-fluorophenyl
group to the indazole is thought to be due to the forced accommo-
dation of the 4-fluorophenyl group between helix 1 and 3—an anal-
ogous structural change and potency enhancement observed in
changing the enone of dexamethasone to deacylcortivazol.
In summary, we have described a novel series of indazoyl
amides which are potent and selective ligands of the glucocorticoid
receptor. These compounds have much simpler structures than
glucocorticoids while maintaining in vitro potencies comparable
and in some cases better than dexamethasone and fluorocortivasol.
Crystallographic evidence established that these indazoyl amides
bind the GR ligand binding domain using the same residues as en-
gaged by glucocorticoids but in a unique manner. While com-
pounds in this series were extremely effective at inhibiting
transrepression as measured in an AP-1 and ELAM assay, they also
show concomitant potency increases in the NP-1 transactivation
assay. An accompanying Letter describes further work on this scaf-
fold in search of ‘dissociated’ glucocorticoid inhibitors.
References and notes
1. Hirschmann, R. Angew. Chem. 1991, 103, 1305. and lit. cited therein; Hatz, H. J.
Glucocorticoide, 2nd ed.; Wissenschaftliche Verlagsgesellschaft mbH: Stuttgart,
2005.
2. Buttgereit, F.; Burmeister, G.-R.; Lipworth, B. J. Lancet 2005, 365, 801.
3. (a) Hayashi, R.; Wada, H.; Ito, K.; Adcock, I. M. Eur. J. Pharmacol. 2004, 500, 51;
(b) Smolen, J. S.; Steiner, G. Nat. Rev. Drug Disc. 2003, 2, 473; (c) Caldenhoven,
E.; Liden, J.; Wissink, S.; Van de Stolpe, A.; Raaijmakers, J.; Koenderman, L.;
Okret, S.; Gustafsson, J. A.; Van der Saag, P. T. Mol. Endocrinol. 1995, 4, 401.
4. (a) Stocklin, E.; Wissler, M.; Gouileux, F.; Groner, B. Nature 1996, 383, 726; For a
review on transrepression, see (b) Laudet, V.; Gronemeyer, H. The Nuclear
Receptor Factsbook; Academic Press, 2002; p 42.
5. Newtown, R.; Holden, N. S. Mol. Pharmacol. 2007, 72, 799.
6. (a) Reichardt, H. M.; Kaestner, K. H.; Tuckerman, J.; Kretz, O.; Wessely, O.; Bock,
R.; Gass, P.; Schmid, W.; Herrlich, P.; Angel, P.; Schutz, G. Cell 1998, 93, 531;
Reichardt, H. M.; Tuckerman, J. P.; Gottlicher, M.; Vujic, M.; Weih, F.; Angel, P.;
Herrlich, P.; Schutz, G. EMBO J. 2001, 20, 7168.
7. Recent reviews: (a) Schacke, H.; Berger, M.; Hansson, T. G.; McKerrecher, D.;
Rehwinkel, H. Expert Opin. Ther. Pat. 2008, 18, 339; (b) Hudson, A. R.; Roach, S.
L.; Higuchi, R. I. Curr. Top. Med. Chem. 2008, 8, 750; (c) Takahashi, H.; Razavi, H.;
Thomson, D. Curr. Top. Med. Chem. 2008, 8, 521; (d) Mohler, M. L.; He, Y.; Wu,
Z.; Hong, S.-S.; Miller, D. D. Expert Opin. Ther. Pat. 2007, 17, 37.
8. Some examples from the literature: (a) Vayssiere, B. M.; Dupont, S.; Choquart,
A.; Petit, F.; Garcia, T.; Marchandeau, C.; Gronemeyer, H.; Resche-Rigon, M. Mol.
Endocrinol. 1997, 11, 1245; (b) Schacke, H.; Schottelius, A.; Docke, W.-D.;
Strehlke, P.; Jaroch, S.; Schmees, N.; Rehwinkel, H.; Hennekes, H.; Asadullah, K.
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 22; (c) Shah, N.; Scanlan, T. S. Bioorg. Med.
Chem. Lett. 2004, 14, 5199; (d) Thompson, C. F.; Quraishi, N.; Ali, A.; Tata, J. R.;
Hammond, M. L.; Balkovec, J. M.; Einstein, M.; Ge, L.; Harris, G.; Kelly, T. M.;
Mazur, P.; Pandit, S.; Santoro, J.; Sitlani, A.; Wang, C.; Williamson, J.; Miller, D.
K.; Yamin, T. D.; Thompson, C. M.; O’Neill, E. A.; Zaller, D.; Forrest, M. J.;
Carballo-Jane, E.; Luell, S. Bioorg. Med. Chem. Lett. 2005, 15, 2163; (e) Barker, M.;
Clackers, M.; Copley, R.; Demaine, D. A.; Humphreys, D.; Inglis, G. G. A.;
Johnston, M. J.; Jones, H. T.; Haase, M. V.; House, D.; Loiseau, R.; Nisbet, L.;
Pacquet, F.; Skone, P. A.; Shanahan, S. E.; Tape, D.; Vinader, V. M.; Washington,
M.; Uings, I.; Upton, R.; McLay, I. M.; Macdonald, S. J. F. J. Med. Chem. 2006, 49,
4216; (f) Kuzmich, D.; Kirrane, T.; Proudfoot, J.; Bekkali, Y.; Zindell, R.; Beck, L.;
Nelson, R.; Shih, C.-K.; Kukulka, A. J.; Paw, Z.; Reilly, P.; Deleon, R.; Cardozo, M.;
Nabozny, G.; Thomson, D. Bioorg. Med. Chem. Lett. 2007, 17, 5025; (g) Ray, N. C.;
Clark, R. D.; Clark, D. E.; Williams, K.; Hickin, H. G.; Crackett, P. H.; Dyke, H. J.;
Lockey, P. M.; Wong, M.; Devos, R.; White, A.; Belanoff, J. K. Bioorg. Med. Chem.
Lett. 2007, 17, 4901; (h) Robinson, R. P.; Buckbinder, L.; Haugeto, A. I.; McNiff, P.
A.; Millham, M. L.; Reese, M. R.; Schaefer, J. F.; Abramov, Y. A.; Bordner, J.;
Chantigny, Y. A.; Kleinman, E. F.; Laird, E. R.; Morgan, B. P.; Murray, J. C.; Salter,

J. E. Sheppeck II et al. / Bioorg. Med. Chem. Lett. 23 (2013) 5442–5447
5447
E. D.; Wessel, M. D.; Yocum, S. A. J. Med. Chem. 2009, 52, 1731; (i) Yang, B. V.;
Vacarro, W.; Doweyko, A.; Doweyko, L.; Huynh, T.; Tortolani, D.; Nadler, S.;
McKay, L.; Somerville, J.; Holloway, D. A.; Habte, S.; Weinstein, D. S.; Barrish, J.
Bioorg. Med. Chem. Lett. 2009, 19, 2139; (j) Xiao, H. Y.; Wu, D. R.; Malley, M. F.;
Gougoutas, J. Z.; Habte, S. F.; Cunningham, M. D.; Somerville, J. E.; Dodd, J. H.;
Barrish, J. C.; Nadler, S. G.; Dhar, T. G. M. J. Med. Chem. 2010, 53, 1270; (k)
Weinstein, D. S.; Gong, H.; Doweyko, A. M.; Cunningham, M.; Habte, S.; Wang, J.
H.; Holloway, D. A.; Burke, C.; Gao, L.; Guarino, V.; Carman, J.; Somerville, J. E.;
Shuster, D.; Salter-Cid, L.; Dodd, J. H.; Nadler, S. G.; Barrish, J. C. J. Med. Chem.
2011, 54(20), 7318.
unambiguously determined by X-ray crystallography of 3-(1H-indazol-5-yl)-
2,2-dimethyl-3-phenylpropanoic acid (deposited into the Cambridge
Crystallography Center, CCDC 922816) and it (and in parallel, its enantiomer)
was used to make the (3S) and (3R) enantiomers of 5a, 5c, 5d, and 5h. (b)
Respective enantiomers of compounds 8a, 8b, 8c, and 8d were prepared by
separating the precursor methyl-3-phenylpropanoic acid on
a preparative
Chiralcel OJ column using an SFC method with a mobile phase of 27% MeOH/
73% CO2. Absolute stereochemistry was not determined for these compounds.
17. GR, PR, and AR ligand binding assays were conducted in fluorescence
polarization format which measures the competition between
a test
9. Yang, B. V.; Weinstein, D. S.; Doweyko, L.; Gong, H.; Vacarro, W.; Huynh, T.;
Xiao, H. Y.; Doweyko, A.; McKay, L.; Holloway, D. A.; Somerville, J.; Habte, S.;
Cunningham, M. D.; McMahon, M.; Townsend, R.; Shuster, D.; Dodd, J. H.;
Nadler, S.; Barrish, J. J. Med. Chem. 2010, 53, 8241.
compound and a fluorescently labeled ligand for binding to the full length or
ligand binding domain of the nuclear hormone receptor. IC50 values were
determined by fitting the fluorescence polarization signal data using the four
parameter logistic equation. The Ki values were determined by application of
10. Hirschmann, R.; Steinberg, N. G.; Buchschacher, P.; Fried, J. H.; Kent, G. J.;
Tishler, M. J. Am. Chem. Soc. 1963, 85, 120.
11. Sun, J.-H. et al J. Org. Chem. 1997, 62, 5627.
the
Cheng–Prusoff
equation
to
the
IC50
values,
where
Ki ¼ IC50=ð1 þ ligand concentration=KdÞ. Data shown represent the means of
duplicate experiments. The MR agonist assay was determined in an A549 cell
line. EC50 was determined as (mM)/(% maximal efficacy) using aldosterone as a
positive control. Data shown represent the means of duplicate experiments.
18. >60 GR mutants, multiple co-activator peptides, and numerous compounds
were screened. Manuscript in preparation.
19. Bledsoe, R. K.; Montana, V. G.; Stanley, T. B.; Delves, C. J.; Apolito, C. J.; McKee,
D. D.; Consler, T. G.; Parks, D. J.; Stewart, E. L.; Willson, T. M.; Lambert, M. H.;
Moore, J. T.; Pearce, K. H.; Xu, H. E. Cell 2002, 110, 93.
12. Mayr, H.; Bug, T.; Gotta, M. F.; Hering, N.; Irrgang, B.; Janker, B.; Kempf, B.; Loos,
R.; Ofial, A. R.; Remennikov, G.; Schimmel, H. J. Am. Chem. Soc. 2001, 123, 9500.
13. (a) Antilla, J. C.; Baskin, J. M.; Barder, T. E.; Buchwald, S. L. J. Org. Chem. 2004, 69,
5578; (b) Klapars, A.; Antilla, J. C.; Huang, X.; Buchwald, S. L. J. Am. Chem. Soc.
2001, 123, 7727.
14. Webster, N. J. G.; Green, S.; Jin, J. R.; Chambon, P. Cell 1988, 54, 199.
15. Cellular assays: AP-1 activity is measured using an AP-1 response element
(containing 5 copies) cloned into a luciferase reporter vector. This reporter is
stably transfected into the human A549 lung epithelial cell line. AP-1 activity is
induced by PMA (15 ng/ml), and inhibition of the induction by compounds is
20. Suino-Powell, K.; Xu, Y.; Zhang, C.; Tao, Y.-G.; Tolbert, W. D.; Simons, S. S., Jr.;
Xu, H. E. Mol. Cell. Biol. 1915, 2008, 6.
21. (a) Kauppi, B.; Jakob, C.; Farnegardh, M.; Yang, J.; Ahola, H.; Alarcon, M.; Calles,
K.; Engstrom, O.; Harlan, J.; Mcuhmore, S.; Ramqvist, A.-K.; Thorell, S.; Ohman,
L.; Greer, J.; Gustafsson, J.-A.; Carlstedt-Duke, J.; Carlquist, M. J. Biol. Chem.
2003, 278, 22748–22754; (b) Schoch, G. A.; D’Arcy, B.; Stihle, M.; Burger, D.;
Bar, D.; Benz, J.; Thoma, R.; Ruf, A. J. Mol. Biol. 2010, 395, 568.
quantitated by measuring decreased luciferase activity. NF
using a truncated, NF B-dependent, E-selectin promoter (300 bp) cloned into a
luciferase reporter vector. This reporter is stably transfected into the human
A549 lung epithelial cell line. NF B activity is induced using IL-1b (0.5 ng/ml)
jB is measured
j
j
and inhibition of the induction by compounds is quantitated by measuring
decreased luciferase activity. NP-1 GRE activation activity is measured using a
GR ligand binding domain (GR-LBD) chimera cloned into a GAL4 luciferase
reporter system. This reporter system is stably transfected into a HeLa cell line
(NP-1). Response to ligand induced binding is quantitated by measuring
luciferase activity. Direct activation of the GR-LBD by compounds (agonist) can
be measured as increased luciferase activity.
22. Illustrated in Figure 2, panel 2 are the major H-bonding contacts surrounding
ligand (3S)-5c. These were arrived at by substituting the dexamethasone
molecule in the published Bledsoe et al. X-ray structure (1M2Z.PDB),
positioning the compound in the conformation that was observed
crystallographically for compound (3S)-5a in panel 1, and allowing contact
residues to relax in place. Crystallographic waters at the Arg611/Gln570 end
were retained. Gln642 was repositioned to allow an H-bond to form with the
pendant thiadiazole. The molecular mechanics minimization was conducted
using the Amber force field as implemented in Flo (Thistlesoft, CT) software.
Accommodation of the 4-fluorophenyl group as shown required some manual
repositioning of contacting residues.
16. Respective enantiomers of compounds 5a, 5c, 5d, and 5h were prepared by
separating the precursor 3-(1H-indazol-5-yl)-2,2-dimethyl-3-phenylpropanoic
acid on a preparative Chiralcel OJ column using an SFC method with a mobile
phase of 20% EtOH/MeOH (1:1)/80% CO2 and 0.1% TFA. Retention times of the
enantiomers were 12.3 and 17.1 min. Absolute stereochemistry was