The identification a novel, selective, non-steroidal, functional glucocorticoid receptor antagonist

Article abstract of DOI:10.1016/j.bmcl.2010.01.133

The identification of novel, potent, non-steroidal/small molecule functional GR antagonist GSK1564023A selective over PR is described. Associated structure-activity relationships and the process of optimisation of an initial HTS hit are also described.

Full text of DOI:10.1016/j.bmcl.2010.01.133

Bioorganic & Medicinal Chemistry Letters 20 (2010) 2340–2343
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The identification a novel, selective, non-steroidal, functional glucocorticoid
receptor antagonist
Joseph Rimland c, Angela Dunne b, Suchete S. Hunjan b, Rosemary Sasse b, Iain Uings e, Dino Montanari c,
Matilde Caivano b, Poonam Shah b, David Standing b, David Gray b, David Brown b, William Cairns b,
Ryan Trump d, Paul W. Smith a, Nicolas Bertheleme a, Pier D’Alessandro a, Sheraz Gul b, Mythily Vimal b,
David N. Smith b, Stephen P. Watson a,
*
a GlaxoSmithKline, Neurosciences Centre of Excellence for Drug Discover, New Frontiers Science Park, Harlow, UK
b Molecular Discovery Research, New Frontiers Science Park, Harlow, UK
c Medicines Research Centre, Verona, Italy
d Research Triangle Park, NC, USA
e Immuno Inflammation Centre of Excellence for Drug Discovery, Medicines Research Centre Stevenage, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
The identification of novel, potent, non-steroidal/small molecule functional GR antagonist GSK1564023A
selective over PR is described. Associated structure–activity relationships and the process of optimisation
of an initial HTS hit are also described.
Received 16 December 2009
Revised 26 January 2010
Accepted 27 January 2010
Available online 4 February 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
GR antagonist
Glucocorticoid receptor
Glucocorticoid receptor antagonist
The glucocorticoid receptor (GR) is a member of the nuclear
receptor (NR) superfamily of ligand activated transcription factors.
Activation of GR leads to either positive regulation of gene expres-
sion (via the trans-activation pathway) or negative regulation (via
the trans-repression pathway).1 GR agonists have found great util-
ity as therapeutic agents, primarily as anti-inflammatory agents
(e.g., dexamethasone).1,2 In contrast, the therapeutic potential of
antagonism of the GR receptor remains largely unexploited despite
a strong therapeutic rationale in a wide range of disease states,
including for example, diabetes, Cushing’s syndrome and psychotic
depression.3
The progesterone receptor (PR) antagonist Mifepristone (RU-
486) also displays potent GR antagonism and has seen some clinical
investigation/utility for its GR mediated effects, but its use remains
compromised by the abortifacient properties associated with its
PR pharmacology.3 In common with RU-486 steroidal GR antago-
nists typically exhibit poor selectivity over other NR family mem-
bers, and historically selective and non-steroidal GR antagonists
have proved elusive. Recently however reports of non-steroidal
selective GR antagonist have begun to emerge, for example,
Argenta/Corcept report4,5 a series pyrimidinediones exemplified
by compound (1) and a series of aryl pyrazoles exemplified by
compound (2) (Fig. 1). However non-steroids reported to date
are typically highly lipophilic and often display disparities be-
tween binding affinity and functional activity. Against this back-
ground we report herein the identification of GSK1564023A, a
potent and novel, small molecule functional GR antagonist, selec-
tive over PR.
O
Cl
N
H
O
N
N
N
N
GSK1564023A
GR MMTV pIC50
7.0
< 4.9
5.5
GR NFkB
PR
pIC50
pIC50
Mwt 410 clogP 1.9 PSA 86
A high throughput screen (HTS) was performed on the GSK
compound library in MMTV/trans-activation functional antagonist
format.6 GSK325971A emerged from this HTS as an exemplar of a
family of novel GR antagonists. This chloro-pyrazolopyrimidine
template is an intermediate in the synthesis of PDE4 inhibitors,7,8
* Corresponding author. Tel.: +44 1279 875004.
E-mail address: steve.p.watson@gsk.com (S.P. Watson).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.01.133

J. Rimland et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2340–2343
2341
Cl
Me2N
OH
O
N
NH
HN
O
O
1
Mifepristone - RU-486
O
O
Cl
O
SO2
N
N
H
N
N
N
N
N
GSK325971A
GR MMTV pIC50
Figure 1. Structures of glucocorticoid receptor antagonists.
2
6.6
F
where at high temperatures the chlorine atom can be displaced by
nucleophiles. Before further progression of GSK325971A and re-
lated compounds we thus took appropriate steps to rigorously con-
firm that the Cl atom was sufficiently stable for assessment in
in vitro assays, and that displacement of the chlorine played no
role in the GR Mechanism of Action (MoA).
Analogues of GSK325971A were typically prepared according to
the well precedented8 chemistry of Scheme 1. Condensation of
ketonitriles with an appropriate hydrazine affords the correspond-
ing 1-substituted-5-aminopyrazole (3) Reaction of the 5-amino-
pyrazole with diethyl-ethoxymethylenemalonate, and subsequent
ring closure and treatment with thionyl chloride afforded chloro-
pyrazolopyrimidine (4) which were saponified and converted to
desired amides (5) via standard coupling procedures.
The data of Table 1 illustrate aspects of the MMTV antagonist
SAR around the GSK325971A template. Appropriate substitution
of the benzyl group facilitates an increase in potency (e.g., 4-Cl,3-
(trifluoromethyl)benzyl analogues 6 and 7), and removal of the
two methyl groups reduces potency (15). Replacement of the ben-
zyl group with simple alkyl groups leads to inactivity (19 and 20)
as does replacement of the benzyl phenyl group with some hetero-
cycles (17 and 18). Simple replacement of the chlorine atom also
leads to inactivity (21 and 22). Interestingly methyl substitution
of the pyrazole 3-position has no significant effect on potency as
illustrated by pairs of compounds 6 and 7, and 8 and 10.
MMTV6 concentration response curves (CRC) of cortisol were
generated in presence of increasing concentrations of GSK325971A
(fig. 2). At lower concentrations of GSK325971A a parallel right-
ward shift of the CRC was observed, consistent with competitive
antagonism of the cortisol response with a pKB of 6.9. At higher
concentrations some suppression of the maximum response was
observed, indicative of a non-competitive component to the inter-
action, a slow dissociation rate or indeed potential cytotoxicity.
The experiments were repeated with a 2 h pre-incubation of com-
pound before cortisol addition, with identical results, indicative of
no mechanistic covalent interaction with the potentially labile
chlorine atom9. Moreover simple analytical experiments indicated
GSK325971A to be unchanged in DMSO/water for 48h at ambient
temperature10
.
In analogues bearing the selectivity conferring isoxazolylbenzyl
amide substituent (see below), MMTV potency proved surprisingly
insensitive to changes to the pyrazole ring nitrogen substituent
(Table 2) although all potencies were slightly reduced relative to
N-methyl parent GSK1564023A itself.
In terms of its modest molecular footprint (Mw = 343,
cLog P = 3.3, PSA = 60) GSK325971A represented an attractive
startpoint for medicinal chemistry exploration. Once the MoA
had been confirmed, and risk of Cl displacement discharged, an
SAR study/optimisation programme based around the template
was undertaken.
Table 3 reports the NR selectivity of GSK325971A and other key
analogues. Whilst GSK325971A itself shows significant activity in
the NF
j
B assay.11 reflecting agonism of the GR trans-repression
pathway, GSK1564023A and compound 12 do not. No compounds
demonstrate activity at the mineralocorticoid receptor13 (MR), and
moreover, in contrast to steroidal antagonist RU-486 (see above)
GSK1564023A demonstrated a 30-fold selectivity over progester-
one receptor (PR) antagonism, and 10 and 12 no significant PR
activity.12
Given the origin of GSK325971A as an intermediate in the syn-
thesis of PDE4 inhibitors, PDE4 activity13 was also investigated.
Whilst GSK325971A itself did show modest PDE4 inhibition this
activity was absent in other analogues such as GSK1564023A.
GSK1564023A displayed the most potent and selective pharma-
cology of the compounds in the series and was selected for more
extensive profiling.
300
14.2µM GSK325971A
7.1µM GSK325971A
3.6µM GSK325971A
200
1.8µM GSK325971A
0.9µM GSK325971A
0.45µM GSK325971A
100
0
0.23µM GSK325971A
0µM GSK325971A
-10
-8
-6
-4
Whilst GSK1564023A showed no activation of the trans-
repression pathway it demonstrated antagonism of an EC80 con-
log [Cortisol] M
centration of cortisol in the NF
GSK1564023A is thus a non-dissociated antagonist, blocking both
j
B assay11 with a pIC50 of 7.2.
Figure 2. MMTV concentration–response curves of cortisol6 raised to increasing
concentrations of GSK325971A.

2342
J. Rimland et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2340–2343
R3
R3\
O
R3
N
O
OEt
O
(ii)
H2N
N
(i)
OEt
CN
NH
N
N
N
NH2
R1
H
R1
R1
3
R3
N
Cl
O
R3
N
Cl
N
O
R4
N
OEt
H
(iii)
(iv)
N
N
N
R1
R1
5
4
Scheme 1. Representative synthesis of GSK325971A analogues. Reagents and conditions: (i) EtOH, Reflux 60–100%; (ii) (a) diethyl-ethoxymethylene malonate, 160 °C, 5 h,
40–60%; (b) SOCl2; (iii) KOH; (iv) R4NH2/DIC/Et3N.
Table 1
Structure–activity relationships around GSK325971A template
O
Y
Z
X
N
N
H
N
N
W
Compound
Z
X
W
Y
MMTV6 pIC50
6
7
4-Cl,3-(Trifluoromethyl)benzyl
4-Cl,3-(Trifluoromethyl)benzyl
4-(4-(3,5-Dimethyl oxazolyl)benzyl
3,5-Di-methylbenzyl
2-(3-Methoxyphenyl)benzyl
3,5-Di-methylbenzyl
2,4-Di-methylbenzyl
4-Cl-benzyl
Indan-2-yl
3-Me-Benzyl
2-Methylbenzyl
Benzyl
2-(3-Methylphenyl) pyrrolidine
Thiazol-5-yl methyl
Pyridyl-2-methyl
n-Propyl
Methyl
H
Et
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
7.2
7.2
7.0
6.8
6.7
6.7
6.7
6.6
6.6
6.4
6.4
6.0
<5.2
<5.2
<5.2
<5.2
<5.2
<5.2
<5.2
Me
Me
H
Me
Me
H
H
Et
H
H
H
Me
H
Me
H
H
Me
Me
Et
Me
Me
Et
Et
Me
Et
GSK1564023A
8
9
10
GSK325971A
11
12
13
14
15
16
17
18
19
20
21
22
Et
Me
Me
Et
Me
Et
Et
Et
Me
Cl
Me
OMe
4-(4-(3,5-Di-methyl oxazolyl)benzyl
4-MeO-Benzyl
Me
Me
the trans-activation and trans-repression GR pathways with com-
parable potency.
Table 2
Effects of modifications to pyrazole N-substituent
GSK1564023A demonstrates very good metabolic stability in
in vitro assays (data not shown) but afforded no systemic exposure
in rat following intravenous administration. Further investigation
revealed that GSK1564023A is unstable in rat blood, and that the
displaceable chlorine atom (see above) was the likely origin of this
instability. As the compound has good aqueous stability (see
above) it was inferred that instability was due to interaction with
biological nucleophiles, but this was not investigated further.
Whilst blood instability of GSK1564023A precluded its further
progression the compound and chemotype remain noteworthy in
displaying a remarkable, if not unprecedented, pharmacological
profile for a small molecule non-steroid. Moreover GSK1564023A
and related compounds represent very useful tools for understand-
ing pharmacophoric interactions with the GR protein and in the
further design of novel GR ligands. Whilst arylpyrazoloazadecalin
O
Cl
X
N
H
O
N
N
N
N
W
Compound
W
X
MMTV6 pIC50
23
24
25
26
27
28
29
30
i-Pr
n-Pr
Me
Me
Me
H
Me
H
6.6
6.6
6.6
6.5
6.5
6.3
6.1
<5.2
Pyridinyl-2-methyl
Benzyl
Phenyl
Cyclopentyl
2-Methylphenyl
4-Methyphenyl
Me
Me

J. Rimland et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2340–2343
2343
Table 3
Selectivity of GSK325971 and analogues
O
Cl
Z
X
N
N
H
N
N
W
Compound
Z
X
W
MMTV6
pIC50
MMTV6
pEC50
NF
j
B11 Inhibition
MR13
pEC50
MR13
pIC50
PR12
pEC50
PR12 pIC50
(% max)
PDE4B14
pIC50
pIC50 (% max)
GSK1564023A
9
12
4-(4-(3,5-Dimethyl oxazolyl)benzyl
2-(3-Methoxyphenyl)benzyl
Indan-2-yl
Me
Me
Et
Me
Me
H
7.0
6.7
6.6
6.7
<5.2
<5.2
<5.2
<5.2
<5.2
5.4 (62%)
<5.2
<4.7
<5
<5
<4.7
<5
<5
<5
<4.7
<5
5.5 (102%)
<5.3
<5
<5.6
4.7
4.9
5.4
GSK325971A
2,4-Di-methylbenzyl
Et
H
5.8 (102%)
<5
<4.7
<5
5.9 (99%)
based GR antagonists2 such as compound 2 show some common
structural features with the GSK1564023A template (e.g., fused
pyrazole), structural data (e.g., protein crystallography) would
be required to rigorously investigate the significance of any
commonality.
In summary, in GSK1564023A we have identified a novel small
molecule functional GR antagonist. Whilst the compound class
may not be progressable due to blood instability the structural fea-
tures and remarkable pharmacology of this molecule/series pro-
vide valuable information to aid the quest for GR antagonists of
clinical utility.
7. Corresponding PDE4 inhibitors derived from displacement of the chlorine with
amines showed no GR antagonism.
8. See for example: preparation of pyrazolo[3,4-b]pyridines as PDE4 inhibitors;
PCT Int. Appl. WO 2005090354A1.
9. We concur with the comment of a reviewer that this experiment does not
constitute definitive proof of a non-covalent MoA. However the MMTV reporter
gene assay measures events which occur in the first 3–4 h after cortisol
addition, thus 2 h pre-incubation is proportionately long enough for any
covalent/pre-organisational mechanistic aspect to manifest itself. Moreover
the high selectivity of compounds of this series over other NRs, and more
importantly the SAR which demonstrate GR activity is ablated by simple
structural changes (e.g., compounds 19 and 22), are also strongly suggestive of
specific MoA rather than simple Cl-displacement. It is planned that
stabilisation of the chlorine substituent will form the subject of a subsequent
publication.
10. Samples showed no significant decomposition whilst held in DMSO solution in
the GSK compound bank over several years.
References and notes
11. NF
line stably transfected with a plasmid containing an ELAM promotor sequence
containing NF B response elements, stimulated with TNF and measuring
jB inhibition measured in a 384 format in human A549 lung epithelial cell
1. Schacke, H.; Rehwinkel, H. Curr. Opin. Invest. Drugs 2004, 5, 524.
2. Takahashi, H.; Razavi, H.; Thomson, D. Curr. Top. Med. Chem. 2008, 8, 521.
3. For a recent detailed review of GR antagonists see: Clark, R. D. Curr. Top. Med.
Chem. 2008, 8, 813.
4. 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.
5. Clark, R. D.; Ray, N. C.; Williams, K.; Blaney, P.; Ward, S.; Crackett, P. H.; Hurley,
C.; Dyke, H. J.; Clark, D. E.; Lockey, P.; Devos, R.; Wong, M.; Porres, S. S.; Bright,
C. P.; Jenkins, R. E.; Belanoff, J. 1H. Bioorg. Med. Chem. Lett. 2008, 18, 1312.
6. Antagonism determined in human A549 lung epithelial fibroblasts expressing
Renilla luciferase reporter gene construct under control of mouse mammary
tumour virus (MMTV). MMTV sequences included glucocorticoid responsive
elements (GRE) which confer glucocorticoid receptor (GR) dependence on
luciferase gene expression. pIC50 values calculated from 11 point
concentration–response curves (CRC) in 384 well format raised against EC80
of cortisol. pEC50 values recorded in presence of compound alone; All values
typically from n = 4 experiments or greater.
j
a
SPAP fluorescence. pIC50 values were calculated from 11 point CRC’s, all values
typically from n = 4 experiments or greater.
12. Determined in 384 well format in African Green Monkey CV-1 cells stably
expressing human progesterone nuclear receptor
B isoform (PR-B). pIC50
values are calculated from 11 point concentration–response curves (CRC’s)
raised against EC80 of progesterone, and pEC50 values of compound alone, all
values typically from n = 4 experiments or greater.
13. Determined in 384 well format in African Green Monkey CV-1 cells stably
expressing human aldosterone-mineralocorticoid nuclear receptor (MR). pIC50
values are calculated from 11 point concentration–response curves (CRC’s)
raised against EC80 of aldosterone, and pEC50 values of compound alone, all
values typically from n = 4 experiments or greater.
14. The inhibitory effect of compounds determined for human phosphodiesterase
4B (PDE4B) enzyme via ATP-luciferase luminescence readout. pIC50 values are
calculated from 11 point concentration–response curves. All values typically
from n = 4 experiments or greater.