Discovery of acylurea isosteres of 2-acylaminothiadiazole in the azaxanthene series of glucocorticoid receptor agonists

Article abstract of DOI:10.1016/j.bmcl.2014.06.010

Acylureas and acyclic imides are found to be excellent isosteres for 2-acylamino-1,3,4-thiadiazole in the azaxanthene-based series of glucocorticoid receptor (GR) agonists. The results reported herein show that primary acylureas maintain high affinity and selectivity for GR while providing improved CYP450 inhibition and pharmacokinetic profile over 2-acylamino-1,3,4-thiadiazoles. General methods for synthesis of a variety of acylureas and acyclic imides from a carboxylic acid were utilized and are described.

Full text of DOI:10.1016/j.bmcl.2014.06.010

Bioorganic & Medicinal Chemistry Letters 24 (2014) 3268–3273
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of acylurea isosteres of 2-acylaminothiadiazole
in the azaxanthene series of glucocorticoid receptor agonists
Hua Gong, Michael Yang, Zili Xiao, Arthur M. Doweyko, Mark Cunningham, Jinhong Wang, Sium Habte,
Deborah Holloway, Christine Burke, David Shuster, Ling Gao, Julie Carman, John E. Somerville,
⇑
Steven G. Nadler, Luisa Salter-Cid, Joel C. Barrish, David S. Weinstein
Department of Discovery Chemistry, Immunoscience, Bristol-Myers Squibb Company, Research and Development, Princeton, NJ 08543-4000, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Acylureas and acyclic imides are found to be excellent isosteres for 2-acylamino-1,3,4-thiadiazole in the
azaxanthene-based series of glucocorticoid receptor (GR) agonists. The results reported herein show that
primary acylureas maintain high affinity and selectivity for GR while providing improved CYP450 inhibi-
tion and pharmacokinetic profile over 2-acylamino-1,3,4-thiadiazoles. General methods for synthesis of a
variety of acylureas and acyclic imides from a carboxylic acid were utilized and are described.
Ó 2014 Elsevier Ltd. All rights reserved.
Received 23 April 2014
Revised 3 June 2014
Accepted 5 June 2014
Available online 13 June 2014
Keywords:
Acylurea
Imide
Isostere
Glucocorticoid receptor
Non-steroidal glucocorticoid receptor
agonists
Agonists of the glucocorticoid receptor (GR), such as dexameth-
asone (dex) 1 and prednisolone (pred) 2, have found broad utility
in autoimmune and inflammatory diseases for over sixty years.1
However, the unparalleled efficacy of these glucocorticoids (GCs),
all of which are steroids, is countered by a host of side effects such
as glucose intolerance (diabetes), muscle wasting, skin thinning,
and osteoporosis.2 In addition, cross-reactivity of steroids with
other members of nuclear hormone receptors (NHRs), such as pro-
gesterone receptor (PR), androgen receptor (AR), and mineralocor-
ticoid receptor (MR), can provoke off-target pharmacology.3
Glucocorticoids provide anti-inflammatory efficacy through the
transcriptional repression (transrepression, TR) of pro-inflamma-
GR-dependent TA contribute to the anti-inflammatory and immu-
nosuppressive properties of GCs, complete dissociation may be
expected to provide GR modulators of sub-optimal efficacy in
treating autoimmune disease.5
In our effort to identify selective and dissociated non-steroidal
GR agonists, we started from dihydro-ethanoanthracene carbox-
amides such as compound 3, which displaying a promising dissoci-
ated profile in discrimination between TR and TA activities.6a We
demonstrated in a subsequent report that several conformational
constraints of this system were unnecessary for potent receptor
binding, as uncyclized diphenylpropionamides (e.g., 4) and xanth-
enes (e.g., 5) maintained high GR binding affinity and agonist activ-
tory transcription factors such as AP-1 and NF
j
B, while side effects
ity.6b Recently, we reported that introduction of
a pyridine
are induced through the transactivation (TA) of genes bearing GC
response elements in their promoter regions. Ligands which mod-
ulate GR function in a pathway-selective manner (‘dissociated ago-
nists’), maintaining TR while minimizing TA, would be expected to
maintain the clinical efficacy of traditional GR agonists, while pro-
viding a relatively improved side effect profile.4 Growing evidence
suggests that ligands which completely dissociate these pathways
may not be expected to meet this idealized clinical profile.
For example, as various gene products which are subject to
nitrogen to the xanthene core of 5 and arylation adjacent to the
pyridine nitrogen to 2-aryl-5H-chromeno[2,3-b]pyridine (azax-
anthene) could not only address metabolic liabilities associated
with phenols 4 and 5, but also provided selective GR ligands which
display a broad range of pharmacologic profiles.7 Among them, two
close benzamide analogs (6 and 7, BMS-776532 and BMS-791826,
respectively) have been found to maintain distinct levels of partial
agonist efficacy, displaying anti-inflammatory activity comparable
to that of prednisolone (2), a known full GR agonist.7 Both com-
pounds were reported to display non-linear pharmacokinetics,
with systemic exposures increasing after oral dosing in a greater
than dose proportional manner.7 We subsequently determined
⇑
Corresponding author. Tel.: +1 609 252 3060; fax: +1 609 252 6804.
E-mail address: David.Weinstein@bms.com (D.S. Weinstein).
http://dx.doi.org/10.1016/j.bmcl.2014.06.010
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

H. Gong et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3268–3273
3269
that this behavior can be attributed to potent CYP450 inhibition,
with concentrations of 6 and 7 exceeding Km for the isoform
responsible for the metabolism of these compounds (saturable
clearance) at relatively low doses (unpublished results). Therefore,
work was initiated to identify a developable GR compound which
would avoid the liabilities associated with both 6 and 7 while
maintaining their promising pharmacological profiles (see Fig. 1).
We suspected that the metabolic liability of 6 and 7 may reside
within the 2-acylamino-1,3,4-thiadiazole moiety. The X-ray co-
crystal structure of GR ligand-binding domain (LBD) bound inda-
zole 2-acyl-1,3,4-aminothiadiazole ligand was useful in our efforts
to identify surrogates for this functionality.8a It revealed that
Asn564 and Gln642 play a role in binding through a three-centered
hydrogen bond network to the 2-acyl-1,3,4-aminothiadiazole moi-
ety as shown in the left panel of Figure 2. Specifically, Asn564
engages in two H-bonds, one to the amide NH and another to
one nitrogen of the thiadiazole, while Gln642 forms an additional
H-bond with the amide carbonyl oxygen. These interactions effec-
tively mimic the engagement of both residues within the steroidal
C-11 b hydroxyl and D ring C-17 substituents of 1 and 2.9 We envi-
sioned that both acyclic imide and acylurea structures could mimic
the three-centered hydrogen bond network of thiazole and thiadi-
azole amide ligands. Docking the acyclic acylurea into the pub-
lished crystal structure fit very well in the GR LBD pocket as
shown in the right panel of Figure 2. Herein we report our results
on synthesis and biological characterization of the acyclic imide
and acylurea GR ligands.
The synthesis of a series of acyclic imides is outlined in
Scheme 1. The preparation of 5H-chromeno[2,3-b]pyridine carbox-
ylic acid 8a has been described before.7 Suzuki–Miyaura coupling
of 8a,b with aryl boronate provided 2-aryl-5H-chromeno[2,3-
b]pyridine carboxylic acid 9. Following the method of Andrus
et al.,10 activation of the acid 9 with carbodiimide followed by
treatment with pentafluorophenol gives an activated pentafluor-
ophenyl ester 10. Condensation of ester 10 with an amide anion,
which is separately generated from treatment of an alkyl or cyclo-
alkyl amide with an appropriately strong base such as sodium
hexamethyl-disilazide gives acyclic imides 12 in 65% yield. Alter-
natively, the acyclic imides can also be synthesized in the following
two step sequence. The carboxylic acid 8 was first converted into a
primary amide 11, which was treated with sodium hexamethyldi-
silazide to generate an amide anion, and then condensation with
an acyl chloride gives imides 13ꢀ18 in 66–70% yield (Scheme 1).
The synthesis of N-methyl acylurea is outlined in Scheme 2. The
acid 8 was converted into a primary amide 19, using HATU as an
ammonium chloride as an ammonium source. Condensation of
19 with methyl carbamoyl chloride in the presence of sodium hex-
amethyldisilazide or sodium hydride gives N-methyl acylurea
compound 20 in 80% yield.
The structure of the N-methyl acylurea 20 was confirmed by an
X-ray crystal structure (Fig. 3). In the similar manner, the synthesis
of N,N-dimethyl acylurea 21 can be achieved as shown in Scheme 3.
Arylation of 20 provided 22.
Another approach for preparation of secondary acylureas is
shown in Scheme 4. Condensation of 11 with isocyanate in an inert
solvent such as toluene, at 90°C gives compounds 23–25 in 65–82%
yields.
One method for synthesis of primary acylureas involved
cleavage of the para-methoxy benzyl (PMB) protected secondary
acylurea 25 with neat trifluoroacetic acid (TFA), or 2,3-dichloro-
5,6-dicyano-1,4-benzoquinone (DDQ) to provide the primary
acylurea 26 in an excellent yield (Scheme 5).
The primary acylurea can also be prepared according to the
method of Xiao et al.11 Therefore, conversion of carboxylic acid 8
to an acylcyanamide 28 was effected using BOP as an activation
agent, in the presence of diisopropylethyl amine and cyanamide.
The acylcyanamide 28 proved to be stable in the subsequent
Suzuki coupling reaction. Thus, condensation with aryl boronate,
in the presence of tetrakis catalyst and potassium phosphate triba-
sic solution, gives the 2-aryl-5H-chromeno[2,3-b]pyridine acylcya-
namide 29. The acyl-cyanamide 30 can be hydrolyzed to primary
acylurea 30 under strongly acidic conditions. Thus, treatment with
4 N HCl in water provided the useful acid intermediate 30 in 86%
yield. The primary acylureas bearing different benzamides 31–36
were readily prepared by condensation of 30 with various amines
as shown in Scheme 6.
The in vitro assays used to characterize biological activities of
the GR ligands including nuclear receptor binding, transrepression
in an A549 cell line [AP-1 and E-selectin (NFjB dependent)], and
transactivation in GAL-4 reporter in a HeLa cell line [NP-1 agonist
assay] have been previously reported.7
GR binding and functional activity for acyclic imides and acylu-
reas are summarized in Table 1. The 2-acylamino-1,3,4-thiadiazole
of 6 (GR Ki = 1.9 nM)7 was simply replaced with an acyclic imide to
get compound 13 (GR Ki = 69.4 nM). A significant loss of of GR
binding affinity and functional activities were observed. While it
is not a trend when the left hand amide was modified. The acyclic
imide compounds 14 and 15, both of which are bearing a different
amide, do show good GR binding affinity and functional activities
in both transrepression and transactivation assays (Table 1). The
GR selectivity over PR is also affected by the left hand amide.
Among three methyl imides 13–15, the morpholine amide 15
showed the least PR selectivity. However, the PR selectivity can
be improved by varying the methyl imide with others, such as imi-
des 16–18. It is also noted that SAR of the acyclic imides is very
sensitive. The imides 16–18 showed weaker human whole blood
(hWB) activity, which is consistent with their weak transactivation
activity. The major issue for further development the imide series
is hydrolytic instability of acyclic imides in low pH. Under acidic
conditions, the acyclic imide can be readily hydrolyzed back to a
primary amide.
activation reagent, diisopropylethyl amine as
a base and
OH
O
O
S
S
O
OH
R'
HO
N
H
N
N
N
H
H
H
R
O
HO
3
4
1 R = F, R' = Me, dexamethasone (dex)
2 R = H, R' = H, prednisolone (pred)
The relatively weak GR binding affinity of the N,N-dimethyl
acylurea compound 21 was improved by removing one methyl
group to N-monomethyl acylurea 22. This can be rationalized
based on what is likely a requirement for a cyclic hydrogen bond
between the terminal NH and the carbonyl oxygen adjacent to
the quaternary carbon, maintaining the acylurea in the most pro-
ductive trans conformation. Accordingly, the GR binding affinity
of 22 increased more than 10-fold relative to 21. Like compound
22, other secondary acylureas 23-25 also exhibited potent GR bind-
ing affinity and transrepression activity (AP-1 and E-selectin).
O
S
O
S
N
N
N
N
N
N
H
H
R
N
N
O
HO
O
6 R = Me, BMS-776532
7 R = Et, BMS-791826
5
O
Figure 1. Synthetic glucocorticoid agonists.

3270
H. Gong et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3268–3273
Figure 2. (Left) Azaxanthene 2-acylaminothiadiazole model based on published co-crystal structure of indazole 2-acylaminothiadiazole ligand (Ref. 8a: Sheppeck et al. BMCL,
2013, 23, 5442). (Right) Acyclic imide/acylurea modeled to mimic three-centered hydrogen bond network of thiazole and thiadiazole amide ligands.
COOH
COOH
O
O
O
B
a
b
Cl
N
O
R1
N
N
O
X
F
R2
8a: X = H
8b: X = F
9
R1
O
N
d
R2
2.849Å
O
2.875Å
NH2
R1
N
O
N
F
R2
11
O
O
F
e
R
O
Cl
F
F
F
Figure 3. X-ray structure of N-methyl acylurea 20.
O
O
O
c
O
N
R
H
H2N
R
F
O
O
O
R1
N
O
R1
N
N
O
N
H
N
N
F
F
N
Cl
R2
R2
12,13-18
10
O
O
11
NaHSi(TMS)2 (1.5 eq)
or NaH, THF, 80%;
O
N
O
Scheme 1. Acyclic imides synthesis. Reagents and conditions: (a) Pd(PPh3)4, DMF,
K3PO4 (2 M), 90 °C, 92%; (b) pentafluorophenol, DCC, DMF, room temperature; (c)
NaHMDS, THF, 0 °C, 65% for
N
F
21
O
2 steps; (d) NH4Cl, HATU, DIEA, DMF, room
temperature, 90%; (e) NaHMDS, THF, 0 °C, 66–70%.
Scheme 3. N,N-Dimethyl acylurea synthesis.
O
O
R
N
H
N
H
R
N
C
O
O
11
Toluene, 90oC
65% ~ 82%
NH2
O
N
O
O
N
F
a
b
8b
23 R = Et
N
H
Cl
O
Cl
O
N
O
24 R = CyBu
25 R = PMB
F
19
Scheme 4. Secondary acylurea synthesis.
O
O
O
O
O
B
N
H
N
H
N
H
N
H
c
However, neither N,N-dimethylacylurea 21 nor secondary acylure-
as 22-25 had good human whole blood activity (Table 1).
O
N
O
Cl
N
O
Gratifyingly, primary acylurea 26, showed very good GR binding
affinity (GR Ki = 2.1 nM) and potent functional activity in both AP-1
(EC50 = 2.1 nM) and E-selectin (EC50 = 1.2 nM) assays. More impor-
tantly, the human whole blood activity of 26 (hWB EC50 = 20 nM)
increased dramatically to become as efficacious as dex (100% of
dex) but with weaker NP-1 transactivation activity (74% of dex).
N
F
F
N
O
20
22
O
O
Scheme 2. N-Methyl acylurea synthesis. Reagents and conditions: (a) NH4Cl, HATU,
DIEA, DMF, room temperature, 5 h, 85%; (b) NaHSi(TMS)2 (1.5 equiv) or NaH, THF,
80%; (c) Pd(PPh3)4, DMF, K3PO4 (2M), 90 °C, 62%.

H. Gong et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3268–3273
3271
O
O
O
O
26 was found to be somewhat higher than that of 27. Most impor-
tantly, primary acylurea 26 showed improved CYP450 inhibition,
particularly against isoform 3A4, and metabolic stability. The half
life of 26 upon incubation with human liver microsomes was found
to be about four times longer than that of 27 (Table 2). Clearly, the
issue of the relatively low PR selectivity relies on the morpholine
amides of compounds 26 and 27. Since the leading compound 7,
PMB
N
H
N
H
N
H
NH2
TFA (neat)
or DDQ
O
N
O
O
N
O
95%
N
F
N
F
26
25
O
O
Scheme 5. Primary acylurea synthesis, a general method.
which bearing
a N-methylethyl amide, possesses more than
1000-fold selectivity of PR/GR.
With the goals of both improving selectivity over PR and weak-
ening transactivation activity, structure activity relationships
around the pendant aryl ring were undertaken (Table 3). We first
explored different benzamides. Not surprisingly (cf., 14 vs 15),
the results showed that pyrrolidine amides 32 and 33 displayed
improved PR selectivity over corresponding morpholine amides
26 and 31. We also found that PR affinity decreased for des-fluoro
compounds, such as 31 and 33. By removing the fluorine from the
C-8 of the azaxanthene ring, the PR potency decreased about 4-
fold. The 3S-fluoro-pyrollidine acylurea compound 34 showed no
PR binding affinity (PR Ki >33330 nM).
At this point, we turned our attention to explore SAR to attenu-
ate transactivation activity of the primary acylureas, since all of
these primary acylureas displayed higher level of NP-1 transactiva-
tion activity compared against corresponding 2-acylamino-1,3,4-
thiadiazoles. We previously reported that a desirable activity
profile of a ‘dissociated’ GR agonist is one with good partial agonist
activity in assays for transrepression, while showing some level of
transactivation to achieve anti-inflammatory efficacy paralleling
that of steroidal full GR agonists. NP-1 agonism efficacy in the
range of 50–65% relative to dex (100%) provided useful efficacy
both in whole blood and animal inflammatory models. We found
that if the pendant phenyl was replaced with a 5-pyridinyl group,
the primary acylureas 35 and 36 did show favorably attenuated
NP-1 transactivation activity (53% and 55% of dex for 35 and 36,
O
O
N
N
N
O
O
O
N
H
B
H
a
b
8b
H2
N
N
N
O
Cl
N
O
HO
F
F
29
28
OH
O
O
O
O
O
N
NH2
N
H
NH2
H
c
d
N
O
N
O
1R2RN
F
F
HO
30
31-36
O
O
Scheme 6. Primary acylurea synthesis, an improved method. Reagents and
conditions: (a) BOP, DIEA, DMF, 80 °C, 12 h, 80%; (b) Pd(PPh3)4, DMF, K3PO4 (2 M),
90 °C, 4 h, 65%; (c) 4 N HCl, dioxane, 60 °C, 2 h, 86%; (d) amine, BOP, DIEA, DMF,
room temperature, 2 h, 80–90%.
Table 2 summarizes the comparison of primary acylurea 26
with the corresponding 2-acylaminothiadiazole 27. Overall, both
primary acylurea 26 and azole azide compound 27 not only have
similar GR binding affinity and other NHRs selectivity, but also
share similar functional activities in AP-1 repression assay and
hWB LPS/THF activity. The NP-1 transactivation activity of acylurea
Table 1
Acyclic imide and acylurea replacement for azole amidea
O
O
N
H
R
R1
N
O
N
F
R2
O
Compd
Structure
NR1R2
GR bindinga
PR bindinga
AP-1 repressionb
E-selectin repressionb
NP-1 agonismc
hWB LPS/TNF
EC50 (nM)
%dexd
6.9
830
R
Ki (nM)
Ki (nM)
EC50 (nM)
%dexd
EC50 (nM)
%dexd
EC50 (nM)
%dexd
Dex
6
7
1.2
1.9
1.6
778
1529
1829
2.5
33.2
17.7
100
70
79
1.1
33.8
6.89
100
63
77
4.5
242
100
32
100
83
98.5
57
379
91
13
14
15
Me
Me
Me
NMe2
Pyrrolidine
Morpholine
69.4
2.5
5.9
12380
12380
268
212.6
5.1
5
110
100
91
109.1
3.6
3.2
72
89
90
1046
27
14
55
67
76
1892
218
88
103
96
99
16
17
18
iPr
CyPr
CyBu
Morpholine
Morpholine
Morpholine
12.9
8.5
8.7
5043
2040
5337
23.7
18.4
17.5
61
76
73
74.8
17.6
32.8
61
62
56
1043
177
586
14
67
6
389
418
2378
81
68
72
21
22
23
24
25
26
NMe2
NHMe
NHEt
NHcyBu
NHPMB
NH2
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
Morpholine
110.5
6.9
5.7
8.7
7.4
10630
1168
2909
5337
623
907
72
77
80
73
46
89
203.2
16.1
36.4
32.8
148.8
1.2
56
66
61
56
52
90
8912
302
241
676
10530
5.5
55
67
86
87
—
—
1035
2143
2378
—
—
89
88
72
—
10.7
34.2
17.5
243.6
2.1
2.1
52.8
74
20.3
100
a
b
c
Values are means of two or more experiments performed in triplicate.
Activation protein (AP-1) and E-selectin assays were performed in an A549 lung epithelial line.
GR transactivation NP-1 assay (run in agonist mode) was performed in the HeLa cell line.
Efficacy represented as percentage of the maximal response of dexamethasone (100%).
d

3272
H. Gong et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3268–3273
Table 2
Comparison of acylurea 26 and 2-acylaminothiadiazole 27a
O
O
O
S
R
26: R =
N
H
NH2
∗
∗
O
N
O
N
N
27: R =
N
H
N
F
O
Compd
26
27
GR binding Ki, nM
PR Binding Ki, nM
AR Binding Ki, nM
2.1
52.8
>50000
>75
>5
1.6
44.9
>50000
>75
>5
ER
a Binding Ki, lM
MR agonistb
AP-1 repression EC50, nM (%dex)
NP-1 agonism EC50, nM (%dex)
hWB LPS/THF EC50, nM (%dex)
2.1 (89)
5.5 (74)
20.3 (100)
2.1 (93)
10.8 (58)
21.5 (108)
Figure 4. Compound 38 in rat LPS-induced TNFa.
HLM t1/2, min
41
9.1
CYP IC50,
l
M (3A4/2C8/2C9/2C19)
>40/8.9/14/11
0.3/3/7/1.0/1.3
a
determined to be 191 min, 14 min, and 98 min, respectively.
Accordingly, moderate clearance (20.8 mL/min/kg) was deter-
mined for 38 after a single 1 mg/kg intravenous dose to rats. A bio-
availability of 25.7% was determined after a single 10 mg/kg oral
dose of a solution of 38 to rats. The pharmacokinetics of 38 was
also evaluated in dogs. Clearance with IV dose was low (9.0 mL/
min/kg) and consistent with high microsomal stability. High expo-
sures and 82% oral bioavailability was recorded after a single 5 mg/kg
oral dose to dogs. In order to evaluate a pharmacodynamic effect of
Values are means at least two experiments.
In A549 cell line, EC50 (lM)/(%maximal efficacy) determination (aldosterone as
b
a positive control).
respectively). Potent and efficacious AP-1 and hWB activities were
maintained. The attached amide group of the pendent pyridine
could also be further modified to other groups such as tert-carbinol
(37) and isopropoxyl (38). For example, compound 38 did show
potent GR binding affinity and functional activity (GR Ki = 0.7 nM,
AP-1 EC50 = 6.7 nM), selectivity over PR and AR (ꢀ450 fold and
>7000 fold, respectively), and no potency below 10,000 nM in a
in vitro safety assessment panel comprised of GPCR, neurotrans-
mitter transporter, ion channel and enzyme off-targets (data not
shown).
38 in vivo, generation of TNF-
a in LPS-challenged rats was deter-
mined. Thus, Lewis male rats received either a single 10 mg/kg oral
dose of 38 (Fig. 4), prednisolone (30 mg/kg), or vehicle. Ninety min-
utes later, the LPS was administered IP, and serum TNF-
a was
determined after another 90 min. 38 was found to provide nearly
equivalent efficacy to prednisolone (69% and 75% inhibition
relative to vehicle, respectively). At 3 h post dose, the serum con-
In light of its promising in vitro pharmacologic profile, acylurea
38 was selected for further evaluation. Half lives of 38 upon incu-
bation with human, rat, and dog liver microsomes (1 mg/mL) were
centration of 38 was determined to be 0.5
lM, equivalent to
Table 3
Primary acylurea structure–activity relationshipa
NH2
O
O
O
NH2
O
O
NH2
O
N
H
N
H
N
H
R1
N
N
O
N
O
N
O
X
F
F
R2
Y
N
O
N
Compd 37
Compd 26, 31-36
Compd 38
NP-1 agonismc hWB LPS/TNFc
O
OH
Compd
Structure
NR1R2
GR
binding
PR
binding
AP-1 repressionb
E-selectin
Metabolic stability
repressionb
X
Y
Ki (nM)
Ki (nM)
EC50
(nM)
%Dexd EC50
%Dexd EC50
(nM)
%Dexd EC50
(nM)
%Dex % remaining (h/r/
m)
(nM)
26
31
32
33
34
F
H
F
H
F
CH Morpholine
CH Morpholine
CH Pyrrolidine
CH Pyrrolidine
CH 3S-F-
2.1
1.5
0.7
1.9
1.2
52.8
282.2
604.1
2322
33330
2.1
6.2
1.6
8.7
0.8
89
111
90
95
104
1.2
4.8
1.4
5.0
0.6
90
90
88
94
90
5.5
11
5.3
49.3
1.3
74
79
98
87
94
20.3
3054
—
207
—
100
98
—
101
—
83/90/80
95/86/75
89/71/73
78/49/76
56/67/77
pyrrolidine
35
36
37
38
F
F
N
N
Pyrrolidine
Morpholine
2.0
7.3
4.0
0.7
1362
22.7
106.2
312.5
16.5
20.3
21.9
6.7
90
96
97
82
12.2
16.3
14.7
3.8
79
77
85
79
112.6
142.9
107.4
59.2
53
55
49
54
211
277
199
355
99
93
106
101
73/86/86
—/—/—
73/83/76
97/58/86
a
b
c
Values are means of two or more experiments performed in triplicate.
Activation protein (AP-1) and E-selectin assays were performed in an A549 lung epithelial line.
GR transactivation NP-1 assay (run in agonist mode) was performed in the HeLa cell line.
d
Efficacy represented as percentage of the maximal response of dexamethasone (100%). %Dex is not reported where <5%.

H. Gong et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3268–3273
3273
4. For recent reviews on developments in the research of dissociated
glucocorticoid agonists, see: (a) Hudson, A. R.; Roach, S. L.; Higuchi, R. I. Curr.
Top. Med. Chem. 2008, 8, 750; (b) Schacke, H.; Berger, M.; Hansson, T. G.;
McKerrecher, D.; Rehwinkel, H. Expert Opin. Ther. Patents 2008, 18, 339; (c)
Regan, J.; Razavi, H.; Thomson, D. Chapter 9: in Ann. Rep. Med. Chem. 2008, 43,
141; (d) Norman, P. Expert Opin. Ther. Pat. 2011, 21, 1137; (e) Norman, P. Expert
Opin. Ther. Pat. 2013, 23, 1239.
5. (a) Nicolaides, N. C.; Galata, Z.; Kino, T.; Chrousos, G. P.; Charmandari, E.
Steroids 2010, 75, 1; (b) Haller, J.; Mikics, E.; Makara, G. B. Frontiers
Neuroendocrinol. 2008, 29, 273; (c) Newton, R.; Leigh, R. Pharmacol. Ther.
2010, 125, 283; (d) Newton, R.; Holden, N. S. Mol. Pharmacol. 2007, 72, 799; (e)
Beck, I. M. E.; Berghe, W. V.; Vermeulen, L.; Yamamoto, K. R.; Haegeman, G.; De
Bosscher, K. Endocr. Rev. 2009, 30, 830.
6. (a) Yang, B. V.; Vacarro, W.; Doweyko, A. M.; Doweyko, L. M.; Huynh, T.;
Tortolani, D.; Nadler, S.; Holloway, D.; Habte, S.; Weinstein, D. S.; Barrish, J. C.
Bioorg. Med. Chem. Lett. 2009, 19, 2139; (b) Yang, B. V.; Weinstein, D. S.;
Doweyko, L. M.; Gong, H.; Vaccaro, W.; Huynh, T.; Xiao, H.-Y.; Doweyko, A. M.;
Mckay, L.; Holloway, D. A.; Somerville, J. E.; Habte, S.; Cunningham, M.;
McMahon, M.; Townsend, R.; Shuster, D.; Dodd, J.; Nadler, S. G.; Barrish, J. C. J.
Med. Chem. 2010, 53, 8241.
7. Weinstein, D. S.; Gong, H.; Doweyko, A. M.; Cunningham, M.; Habte, S.; Wang, J.
H.; Holloway, D. A.; Burke, C.; 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,
7318.
8. (a) Sheppeck, J. E., II; Gilmore, J. L.; Xiao, H. Y.; Dhar, T. G. M.; Nirschl, D.;
Doweyko, A. M.; Sack, J. S.; Corbert, M. J.; Malley, M. F.; Gougoutas, J. Z.; Mckay,
L.; Cunningham, M. D.; Habte, S. F.; Dodd, J. H.; Nadler, S. G.; Somerville, J. E.;
Barrish, J. C. Bioorg. Med. Chem. Lett. 2013, 23, 5442; (b) Gilmore, J. L.; Sheppeck,
J. E., II; Wang, J.; Murali Dhar, T. G.; Cavallaro, C.; Doweyko, A. M.; Mckay, L.;
Cunningham, M. D.; Habte, S. F.; Dodd, J. H.; Nadler, S. G.; Somerville, J. E.;
Barrish, J. C. Bioorg. Org. Chem. Lett 2013, 2013(23), 5448; (c) 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.; Murali Dhar, T. G. J. Med. Chem. 2010,
53, 1270; (d) Xiao, H. Y.; Wu, D. R.; Sheppeck, J. E., II; Habte, S. F.; Cunningham,
M. D.; Somerville, J. E.; Dodd, J. H.; Barrish, J. C.; Nadler, S. G.; Murali Dhar, T. G.
Bioorg. Med. Chem. Lett. 2013, 23, 5571.
eight-fold the IC50 for inhibition of LPS-induced TNF
blood.
a in rat whole
In conclusion, acyclic imides and acylureas have been found to
serve as excellent isosteres for 2-acylamino-1,3,4-thiadiazole in
the azaxanthene series of GR agonists. A structure-based design
approach was used in their identification, rationalizing that a net-
work of three key hydrogen bonds to the receptor would be main-
tained, and SAR favoring primary and secondary (over tertiary)
acylureas is consistent with the proposed binding mode. Primary
acylureas were found to be particularly potent and efficacious in
cellular assays of transrepression and transactivation. Microsomal
stability and CYP inhibition profiles of primary acylureas were
found to be generally improved relative to 2-acylaminothiadiaz-
oles. SAR to weaken transactivation while maintaining anti-
inflammatory activity (cytokine inhibition) in human whole blood
was identified, leading to the identification of 38, a compound with
promising pharmacokinetic properties and which proved effica-
cious in a pharmacodynamic assay in vivo. Ultimately, the utility
of partial agonists of GR such as 38 in treating human disease while
improving on the safety profile of traditional glucocorticoids
remains to be determined.
Acknowledgments
The authors gratefully acknowledge the following individuals
for their support to the project: Mary Ellen Cvijic, Ding Ren Shen
and Melissa Yarde, Dauh-Rurng Wu, Leslie Leith, and Peng Li.
9. (a) Suino-Powell, K.; Xu, Y.; Zhang, C.; Tao, Y.-G.; Tolbert, W. D.; Simons, S. S.;
Xu, H. E. Mol. Cell. Biol. 2008, 28, 1915; (b) Bledsoe, R. K.; Montana, V. G.;
Stanley, T. B.; Thomas, 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.
References and notes
1. Glucocorticoids; Goulding, N. J., Flower, R. J., Eds.; Birkhäuser: Basel,
Switzerland, 2001.
10. Andrus, M. B.; Li, W.; Keyes, R. F. Tetrahedron Lett. 1998, 39, 5465.
11. Xiao, Z. L.; Yang, M. G.; Tebben, A. J.; Galella, T. A.; Weinstein, D. S. Tetrahedron
Lett. 2010, 51, 5843.
2. Buttgereit, F.; Burmester, G.-R.; Lipworth, B. J. Lancet 2005, 365, 801.
3. McMaster, A.; Ray, D. W. Nat. Clin. Pract. Endocrinol. Metab. 2008, 4, 91.