Discovery of a potent and dissociated non-steroidal glucocorticoid receptor agonist containing an alkyl carbinol pharmacophore

Article abstract of DOI:10.1016/j.bmcl.2014.03.005

Synthesis and structure-activity relationship (SAR) of a series of alkyl and cycloalkyl containing non-steroidal dissociated glucocorticoid receptor (GR) agonists is reported. This series of compounds was identified as part of an effort to replace the CF<

Full text of DOI:10.1016/j.bmcl.2014.03.005

Bioorganic & Medicinal Chemistry Letters 24 (2014) 1934–1940
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of a potent and dissociated non-steroidal glucocorticoid
receptor agonist containing an alkyl carbinol pharmacophore
a
b
a
a
b
Hossein Razavi a, , Doris Riether , Christian Harcken , Jörg Bentzien , Roger M. Dinallo , Donald Souza ,
Richard M. Nelson a, Alison Kukulka a, Tazmeen N. Fadra-Khan a, Edward J. Pack Jr. a,
Ljiljana Zuvela-Jelaska b, Josephine Pelletier b, Mark Panzenbeck b, Carol A. Torcellini b, John R. Proudfoot a,
Gerald H. Nabozny b, David S. Thomson a
⇑
a Department of Medicinal Chemistry, Boehringer Ingelheim Pharmaceuticals, Inc., 900 Ridgebury Road, PO Box 368, Ridgefield, CT 06877, USA
b Department of Immunology & Inflammation, Boehringer Ingelheim Pharmaceuticals, Inc., 900 Ridgebury Road, PO Box 368, Ridgefield, CT 06877, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Synthesis and structure–activity relationship (SAR) of a series of alkyl and cycloalkyl containing non-
steroidal dissociated glucocorticoid receptor (GR) agonists is reported. This series of compounds was
identified as part of an effort to replace the CF3 group in a scaffold represented by 1a. The study culminated
in the identification of compound 14, a t-butyl containing derivative, which has shown potent activity for
GR, selectivity against the progesterone receptor (PR) and the mineralocorticoid receptor (MR), in vitro
anti-inflammatory activity in an IL-6 transrepression assay, and dissociation in a MMTV transactivation
counter-screen. In a collagen-induced arthritis mouse model, 14 displayed prednisolone-like efficacy,
and lower impact on body fat and free fatty acids than prednisolone at an equivalent anti-inflammatory
dose.
Received 23 December 2013
Revised 27 February 2014
Accepted 3 March 2014
Available online 12 March 2014
Keywords:
Nuclear hormone receptors
Glucocorticoids
Non-steroidal
Dissociated
Ó 2014 Elsevier Ltd. All rights reserved.
Agonist
For well over a half of century, clinicians have relied on their
staple arsenal of steroidal glucocorticoids (GCs) to combat inflam-
matory, immune and metabolic maladies.1 However, the therapeu-
tic uses of GCs in chronic diseases are associated with adverse
effects such as glucose intolerance, muscle wasting, skin thinning,
weight gain and osteoporosis that detract from their remarkable
immuno-modulatory and anti-inflammatory activities.2,3
The glucocorticoid receptor (GR) is a member of the nuclear
receptor superfamily and is comprised of multiple domains: an N-
terminal transcription activation domain, a DNA binding domain
that contains a dimerization interface, and a C-terminal ligand bind-
ing domain (LBD) that also contains a transcription activation do-
main. In its unbound state, the GR resides in the cytoplasm and is
associated with, and stabilized by chaperones such as heat-shock
proteins 70 and 90.4 According to the proposed classical mechanism
of GC signaling,5 the binding of a GC agonist such as cortisol, the
endogenous ligand in human, to the cytosolic receptor results in a
cascade of events such as conformational changes, dissociation of
the ligand bound receptor from chaperones, and translocation of
the receptor–ligand complex into the nucleus. The ligand-activated
GR functions as a monomer or a homo-dimer. Once inside the
nucleus, the monomer binds to transcription factors such as NF-
jB and AP-1, which prevents them from interacting with the DNA
resulting in gene down-regulation. The glucocorticoid-mediated
down-regulation of pro-inflammatory genes has been called tran-
scriptional repression or ‘trans-repression’ (TR), and is thought to
be the main mechanistic pathway for the anti-inflammatory activi-
ties of GCs.6 Alternatively, the receptor–ligand homo-dimer binds to
Glucocorticoid Response Elements (GREs) in the promoter region of
genes that ultimately results in gene transcription. This glucocorti-
coid-mediated gene up-regulation, referred to as transcriptional
activation or ‘trans-activation’ (TA), has been primarily associated
with the aforementioned undesirable effects.3
The mechanism of GC action is intricate and still poorly under-
stood. For instance, GC analogues with minor structural differences
are reported to regulate differential gene expression in the same cell
line.7 The GR-mediated signaling is further complicated by factors
such as post-transcriptional mechanisms and selective cofactor
recruitment.8 While the beneficial and the deleterious effects of
GCs are generally considered to be due to on-target activities, their
off-target cross-reactivity with other nuclear hormone receptors
(NHR), such as the androgen receptor (AR), the mineralocorticoid
receptor (MR), and the progesterone receptor (PR), also contributes
to the observed side effects.9
⇑
Corresponding author. Tel.: +1 (203) 778 7926; fax: +1 (203) 791 6072.
E-mail address: hossein.razavi@boehringer-ingelheim.com (H. Razavi).
http://dx.doi.org/10.1016/j.bmcl.2014.03.005
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

H. Razavi et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1934–1940
1935
Despite complexities, insights into the GC mechanism have
rekindled hopes of finding an agent capable of eliciting the desired
anti-inflammatory response while minimizing the unwanted side
effects.10 These discoveries have fueled the hypothesis that the
two GC-mediated mechanistic pathways (i.e., TR and TA) could
be segregated, thus separating the anti-inflammatory response
from the unwanted side effects.11 Since marketed GCs such as
prednisolone and dexamethasone (Fig. 1) lack the desired NHR
selectivity and are unable to differentiate between the two afore-
mentioned mechanistic pathways, the search for a GR selective
synthetic agent that is capable of dissociating the anti-inflamma-
tory effects from the undesirable side effects has intensified.12 Pre-
vious communications from our project group have described
selective and dissociated GR agonists from a novel non-steroidal
chemotype (e.g., 1a), which is comprised of three main structural
elements: an A-ring mimetic, a D-ring mimetic, and a four carbon
linker bearing a hydroxyl and a trifluoromethyl group.13 The tri-
fluoromethyl carbinol group embedded in this linker is also fea-
tured in other GR agonists (e.g., 1b, and 1c).14 The initial attempt
to replace the CF3 moiety in 1b and 1c with large non-fluorinated
alkyl groups such as cyclohexylmethyl or benzyl provided moder-
ate to good GR binders; however, these structural modifications re-
sulted in the alteration of the functional activity from agonism to
antagonism, thereby highlighting the critical role of this group as
a function-regulating pharmacophore.14 Herein, we report the con-
tinuation of our SAR efforts directed at the CF3 group in the series
depicted by 1a.
As part of a strategy to prosecute the SAR in this scaffold, we
examined the role of the CF3 moiety on GR activity and strived to
identify a suitable replacement for this group. We envisioned that
an ideal replacement would retain the agonist activity that was
displayed by its CF3 counterpart. Moreover, it would serve as a
function-regulating pharmacophore (vide supra) providing a han-
dle to fine-tune the in vitro functional activity, which could trans-
late into a desirable in vivo profile. Finally, the inclusion of such
structural diversity might lead to an improvement in physico-
chemical properties, and mitigate any potential risks associated
with unforeseen findings from this scaffold. Although the SAR
based on 1b and 1c had indicated that the CF3 group was requisite
for agonist activity, we believed that these compounds contained
sub-optimal A-ring mimetics and linkers. In fact, a recent report
from Bayer Schering AG on this series has highlighted the deficien-
cies of the methylbenzoxazine group and the carboxamide linker
(i.e., the A-ring mimetic and the linker depicted in compound
1c), and has shown that their replacement resulted in improved
GR selectivity, cellular activity and dissociation.15 In this context,
1a appeared to be a better starting point than 1b and 1c. Com-
pound 1a is a potent and selective GR ligand (Table 1). It is also a
potent and efficacious inhibitor of IL-1-stimulated IL-6 production
in human foreskin fibroblasts (HFF) cell lines, and has a dissociated
profile in a TA reporter gene assay measuring induction at
the MMTV promoter (10% maximal induction vs 100% for
dexamethasone).13b
To rapidly access the alkyl derivatives in this series, we devised
a synthesis whereby analogs could be generated from trifluoro-
methyl ketones. Syntheses of various CF3 ketones have already
been reported.13b According to the reaction sequence outlined in
Scheme 1, CF3 ketones were quantitatively converted to the corre-
sponding carboxylic acids using the haloform reaction. The resul-
tant acids were reacted with thionyl chloride in dichloromethane
to afford the acid chlorides, which upon treatment with morpho-
line provided the corresponding amides in excellent yields. Next,
morpholinyl amides in THF were reacted with various alkyl lithium
reagents at À78 °C to provide the desired alkyl ketones in moder-
ate to good yields. Finally, 2-methyl-azaindoles, prepared accord-
ing to a previously reported method,16 were sequentially treated
with n-BuLi and t-BuOK at À78 °C. Reactions of the resultant carba-
nions with alkyl ketones furnished compounds 2–13 in low to
moderate yields (Scheme 1).
Compounds 2–9 were initially evaluated for their affinity
towards human GR, PR, and MR in a competitive binding fluores-
cence polarization assays using
a
tetramethyl rhodamine
(TAMRA)-labeled dexamethasone probe for GR and MR, and a
TAMRA-labeled mifepristone for PR.17 The SAR indicated that ana-
logs containing small alkyl groups (i.e., methyl and ethyl) were less
potent GR binders than the CF3 comparator (Table 1 cf. 2, and 3 vs
1a). These observed differences in potency were consistent with
the reported effects of fluorine on increasing the affinity of a ligand
towards its target and modulating its physicochemical properties.18
Accordingly, we attributed the higher GR affinity of 1a to the im-
proved hydrophobic interactions between the lipophilic CF3 group
and the receptor, as well as the inductive effect of fluorine atoms
that lowered the pKa of the nearby hydroxyl group thereby enhanc-
ing its hydrogen bond donating ability (cf. calculated pKa of OH
group: 1a = 12.1 vs 2 = 14.8).19 The calculated pKa values for the hy-
droxyl groups of the non-halogenated alkyl derivatives 2–9 were
between 14.7 and 14.8. The minor pKa differences among these ana-
logs (e.g., calculated pKa of OH group: 2 = 14.8; 4 = 14.7) suggested
that the acidity of the OH group did not play a significant role in
their observed binding potencies. In this series of compounds, the
improvement in GR potency was correlated with the increase in
the size of the alkyl moiety; however, the binding affinity plateaued
beyond the isopropyl group indicating an upper limit for the rela-
tionship between size and potency (Table 1. cf. 2, 3, 4, 5 and 6).
Compounds 4–6 also demonstrated good selectivity against PR
and MR (i.e. >100 fold). In fact, the GR potency and NHR selectivity
displayed by these compounds are comparable to those of the CF3
analog (1a). Cycloalkyl derivatives 7 and 8 also showed potent GR
activity, and >100 fold selectivity against PR and MR. However,
increasing the ring size from cyclobutyl to cyclopentyl resulted in
higher MR affinity and lower selectivity (Table 1 cf. 8 vs 9).
OH
OH
O
OH
O
OH
11
B
HO
HO
H
C
D
A
F
H
H
O
O
Dexamethasone
Prednisolone
O
F
N
HO CF3
A-Ring Mimetic
N
H
D-Ring Mimetic
1a
OH
HO CF3
HO CF
3
H
N
H
N
N
O
O
O
O
Next, the panel was tested for their agonist activity in the
IL-1-induced IL-6 production TR assay in HFF cell lines. Overall,
all alkyl and cycloalkyl analogs were less potent and efficacious
than 1a. However, among the non-fluorinated alkyl groups, the
O
O
F
1b
1c
Figure 1. Examples of synthetic steroidal and non-steroidal glucocorticoids.

1936
H. Razavi et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1934–1940
Table 1
Receptor binding and functional assay data for dexamethasone, prednisolone, and compounds in the 6-azaindole series (1a and 2–9)
R2
OH
O
N
N
H
F
a
a
b
Compound no.
R2
GR IC50 (nM)
PR IC50 (nM)
MR IC50a(nM)
IL-6 IC50 (nM)
IL-6 efficacyc (%)
Dexamethasone
Prednisolone
1a
2
3
4
5
6
7
—
—
CF3
Me
3
15
4
230
51
9
>2000
>2000
>2000
>2000
>2000
>2000
1800
>2000
>2000
>2000
>2000
33
44
580
0.5
6
10
>2000
220
73
36
>2000
47
100
91
92
35
50
62
87
15
87
34
12
>2000
>2000
1000
1800
1900
>2000
>2000
1100
Et
i-Propyl
t-Butyl
t-Pentyl
c-Propyl
c-Butyl
c-Pentyl
8
13
15
17
26
8
9
260
>2000
a
b
c
IC50 values are from duplicate 11-point dose response curves.
IC50 values and % efficacies are from two experiments conducted on two different occasions.
Reported efficacies are at the top concentration of 2
lM and are compared to 100% for dexamethasone.
R1
F
O
R1
F
O
R1
F
O
a
b,c
CF3
OH
N
O
X
Y
X
Y
R1
F
O
R1
HO R2
N
H
d
R2
N
H
e
2-13
F
Scheme 1. Reagents and conditions: (a) NaOH, EtOH, H2O 80 °C (quant.); (b) SOCl2, CH2Cl2, 0 °C; (c) morpholine, pyridine, CH2Cl2 (quant. over two steps); (d) R2Li, THF,
À78 °C (50–98%); (e) n-BuLi, tert-BuOK, THF, À78 °C (3–72%). R1 and R2 are defined in Tables 1 and 2. 6-Azaindole or 5-azaindole are defined by X = CH and Y = N or X = N and
Y = CH, respectively.
t-butyl and cyclopropyl had the optimal size and conferred the best
combination of potency and efficacy (Table 1 5 and 7). In line with
the observed GR potencies, the IL-6 potencies and efficacies grad-
ually improved as the size of the alkyl group increased (Table 1
cf. 2, 3 and 4 vs 5). In contrast, increasing the size of the t-butyl
group by one carbon led to a dramatic reduction in potency and
efficacy (Table 1 cf. 5 vs 6). The retention of GR binding potency
and the concomitant erosion of the IL-6 potency and efficacy sug-
gested that the functional activity in this series had switched from
agonist to antagonist by moving from the t-butyl to the t-pentyl
group. A more subtle trend was observed in the cycloalkyl deriva-
tives. As the size of the cyclic group increased, the IL-6 potency and
efficacy gradually decreased resulting in a complete loss of func-
tional activity for the cyclopentyl derivative (Table 1 cf. 8, 9 vs 7).
To gain additional insights, compound 7 was docked into the
X-ray co-crystal structure of the GR-LBD bound to dexamethasone
(Fig. 2).20 According to this docking pose, the 6-azaindole moiety
overlays with the steroid’s A-ring. Analogous to the carbonyl oxy-
gen of dexamethasone, the N(6) of the azaindole forms a pair of
hydrogen bonds with Arg611 and Gln570. Additionally, the indole
NH is hydrogen bonded to Leu563, and the central hydroxyl group,
which mimics the C(11)-OH moiety of dexamethasone, is involved
in a key hydrogen bonding interaction with Asn564. This illustra-
tion also suggests that the cyclopropyl group of ligand 7 resides
in a hydrophobic cavity lined by Cys736, Phe749, Leu753, and
Ile757, which are located on helix 12 and its preceding loop
(Fig. 3). As depicted in Figure 3, the cyclopropyl group of com-
pound 7 appears to optimally fill this hydrophobic pocket, which
may help to stabilize helix 12 resulting in agonist activity. This is
in agreement with the observed SAR that indicates lower GR bind-
ing affinity and reduced agonist activity for smaller alkyl groups
such as methyl and ethyl (Table 1 cf. 7 vs 2 and 3). Compounds
containing larger alkyl groups (e.g., t-pentyl and c-pentyl deriva-
tives, 6 and 9, respectively) are still capable of binding to the GR;
however, their alkyl moieties are too large to be accommodated
in this small pocket, and likely force helix 12 out of its position
leading to a conformational change that result in the loss of
functional activity. This hypothesis is consistent with the ligand-
induced conformational changes that are observed in the X-ray
co-crystal structure of RU486 (a GR antagonist) with GR-LBD,21
and is in line with the reported switch of GR functional activity
in the scaffold that is represented by compounds 1b and 1c (vide
supra).14 In fact, compound 9 (eutomer) displayed the ability to
antagonize the dexamethasone inhibition of IL-1-induced IL-6 pro-
duction in HFF cells (IL-6 antagonist IC50 = 970 nM, n = 3).
We have previously reported that compound 1a and its corre-
sponding 5-azaindole analog had comparable GR binding potency,
and displayed a similar TR activity as measured by potency and

H. Razavi et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1934–1940
1937
Figure 2. Compound 7 (yellow) docked in the GR-LBD. This docking pose is generated with GLIDE22 using the dexamethasone (orange) and GR-LBD X-ray co-crystal structure
(PDB entry 1M2Z),20 and rendered with MOE.23 In this illustration, the cyclopropyl group of 7 is shown by yellow spheres, potential hydrogen bonds between 7 and the GR-
LBD are indicated by white dotted lines, and the surface of the binding pocket is colored according to lipophilicity (i.e., lipophilic in green and hydrophilic in magenta).
Compounds 10 and 11 were also evaluated for their ability to
induce TA in a luciferase reporter gene assay, which utilized a
mouse mammary tumor virus (MMTV) promoter in HeLa cells. In
this assay, dexamethasone induced the MMTV-promoted gene
expression, and showed potent and efficacious profiles in this
and the IL-6 assay (Table 2). A ligand was considered to be dissoci-
ated in this in vitro assay if it displayed lower potency and/or effi-
cacy as compared to the IL-6 TR assay. As such, the extent of the
dissociation depended on the size of the window between the
respective read-outs. Based on these criteria, 10 and 11 were both
dissociated in the MMTV assay. Compounds 10 showed negligible
potency and efficacy at the highest tested concentration as com-
pared to its potent and efficacious profile in IL-6 (Table 2).
Although 11 showed measurable potency in the MMTV assay, it
was dissociated based on efficacy (Table 2, MMTV: 20% maximal
efficacy vs IL-6: 92% maximal efficacy).
The 2-methoxy-5-fluorophenyl group featured in 10 and 11 as a
D-ring mimetic has previously been reported to be a metabolic lia-
bility.13b Therefore, we decided to replace it with metabolically sta-
Figure 3. Docking pose of 7 (yellow) in the GR-LBD generated with GLIDE22 using
the dexamethasone/GR-LBD X-ray co-crystal structure (PDB entry 1M2Z),20 and
rendered with MOE.23 In this picture, the cyclopropyl group is shown by yellow
spheres, helix 12 is depicted by a red ribbon, and a potential hydrogen bond is
indicated by a white dotted line. For clarity, some structural features such as a
portion of helix 12 and its preceding loop are omitted. Lipophilic and hydrophilic
surfaces are shown in green and magenta, respectively.
ble groups that would afford comparable in vitro profiles. One such
expedient was the 2-methyl-5-fluorophenyl moiety, which was
previously identified as part of our SAR efforts in the CF3 containing
compounds (e.g., 1a).13b This group addressed the metabolic con-
cerns and maintained a desirable profile. In particular, the t-butyl
and cyclopropyl analogs were selective and potent GR binders
(Table 2, compounds 12, and 13). Compounds 12 and 13 were 3-
4 fold less potent and had comparable efficacies in the IL-6 assay
relative to their 2-methoxy-5-fluorophenyl comparators (i.e., 10
and 11, respectively). Additionally, both compounds were dissoci-
ated in the MMTV counter-screen.
Since compounds in this series contain a tertiary alcohol group,
we were concerned about the solution stability of 12 and 13 at low
pH. Indeed, the cyclopropyl analog 13 was unstable in aqueous
solution at pH 2.0. In contrast, the t-butyl derivative 12 was
remarkably stable at pH 7.4 and pH 2.0, and retained its structural
integrity under these conditions at 37 °C for 24 h.
efficacy in the IL-1 induced IL-6 production assay.13b Therefore, the
5-azaindole derivatives of 5 and 7 were also synthesized and eval-
uated in the aforementioned NHR binding and IL-6 functional as-
says. In general, the 5-azaindole analogs were slightly more
potent GR binders than their corresponding 6-azaindoles, and
maintained >100 fold selectivity over PR and MR (cf. Table 2, com-
pound 10 and 11 vs Table 1, compound 5 and 7, respectively).
Moreover, the 5-azaindole moiety seemed to have imparted higher
IL-6 potency to compounds in this series. For instance, compounds
10 and 11 were 4-6 fold more potent than 5 and 7, respectively.
However, the effect on IL-6 efficacy appeared to be marginal as
the 5- and 6-azaindole derivatives (i.e., t-butyl 10 vs 5, and cyclo-
propyl 11 vs 7) displayed comparable maximal inhibitions.
In order to profile 12 in vivo, the racemate was resolved to af-
ford enantiomers 14 and 15. Both pure enantiomers (>99% ee)
were tested in the NHR binding assays. The active enantiomer 14
was an excellent GR binder with >100 fold selectivity over PR

1938
H. Razavi et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1934–1940
Table 2
Nuclear hormone receptor binding, and cellular data for dexamethasone and compounds in the 5-azaindole series (10–13)
N
R1
R2
OH
N
H
F
a
a
a
b
b
Compound no.
R1
R2
GR IC50 (nM)
PR IC50 (nM)
MR IC50 (nM)
IL-6 IC50 (nM) eff.c (%)
MMTV EC50 (nM) eff.c (%)
Dexamethasone
—
—
3
4
6
5
9
>2000
790
>2000
780
33
640
1800
720
0.5 (100)
9 (88)
8 (92)
35 (80)
24 (89)
17 (100)
>2000 (10)
90 (20)
>2000 (1)
>2000 (11)
10
11
12
13
OMe
OMe
Me
t-Butyl
c-Propyl
t-Butyl
c-Propyl
Me
>2000
1550
a
b
c
IC50 values are from duplicate 11-point dose response curves.
IC50 values and % efficacies are from two experiments conducted on two different occasions.
Reported efficacies are at the top concentration of 2 M and are compared to 100% for dexamethasone.
l
Table 3
Nuclear hormone receptor binding, and in vitro cellular data for compounds 14 and 15
N
N
OH
OH
N
H
N
H
15
14
F
F
a
a
a
b
b
Compound no.
GR IC50 (nM)
PR IC50 (nM)
MR IC50 (nM)
IL-6 IC50 (nM), eff.c (%)
MMTV IC50 (nM), eff.c (%)
14
15
2
1300
290
>2000
420
1300
27 (76)
—
>2000 (1)
—
a
b
c
IC50 values are from triplicate 11-point dose response curves.
IC50 values and % efficacies are from averages of three experiments conducted on three different occasions.
Reported efficacies are at the top concentration of 2 M and are compared to 100% for dexamethasone.
l
Table 4
Pharmacokinetic data in Sprague-Dawley ratsa
Compound no.
CL (iv) [(ml/min)/kg]
48
VSS (iv) [L/kg]
MRT (iv) [hr]
1.38
Cmax (po) [ng/ml]
AUCinf (po) [h ng/ml]
6728
F (po) [%]
14
4.0
708
64
a
Mean of three rats dosed at 2 mg/kg i.v. in a PEG400/water (70:30) vehicle and at 30 mg/kg p.o. in a PEG400/water/Tween (80/18/2) vehicle.
and MR (Table 3). As expected, 14 displayed good potency and effi-
cacy in the IL-6 assay, and was dissociated in the MMTV TA assay.
The absolute stereochemical configuration of the chiral center in
14 was not determined; however, the configuration was assigned
based on the previously established configuration for the active
enantiomer in the CF3 series.24
To assess the pharmacokinetic properties of this compound, 14
was dosed in male Sprague-Dawley rats (Table 4). Intravenous
administration of 14 revealed a moderate to high clearance (i.e.,
CL = 48 mL/min/kg = 68% of hepatic blood flow), which was consis-
tent with the observed in vitro hepatocytes and liver microsome
data (14: rat hepatocytes = 51% Qh; MLM = 59% Qh; HLM = 79%
Qh). Compound 14 also displayed a moderate volume of distribu-
tion (VSS) and mean residence time (MRT). It had good oral exposure
as determined by the AUC and the Cmax, which resulted in good oral
bioavailability. The oral exposure for 14 was in agreement with
good equilibrium solubility (Eq. solubility: pH 7.4 = 15 lg/mL; pH
Graph 1. Anti-inflammatory effects of compound 14 and prednisolone in a mouse
collagen-induced arthritis study at various oral doses in 30% cremophor.
2.2 = 840 lg/mL), as well as high permeability and low efflux ratio
observed in the Caco-2 assay (AB = 13.4 Â 10À6 cm/s, BA =
16.9 Â 10À6 cm/s, ratio = 1.3).
vehicle. Compound 14 demonstrated 82% reduction of TNF-
a at
In order to evaluate the effectiveness of the compound in an
acute inflammatory setting, 14 was orally dosed in a mouse model
1 mg/kg (t-test: p < 0.05 vs vehicle), which was comparable to
the inhibition displayed by Prednisolone at 3 mg/kg (i.e., 87%) in
of LPS-challenged TNF-
a production using 30% cremophor as

H. Razavi et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1934–1940
1939
Table 5
Disease scores of prednisolone and 14 in a mouse CIA study, as well as the side effect profiles of prednisolone and 14 in healthy mice at various oral doses
Compound
no. (mg/kg)
CIA Disease Score AUC SE Body fat [%fat] Mean SE
Triglycerides [mEq/L]
Mean SE (% change vs
vehicle)
Free Fatty Acids [mEq/
L]Mean SE (% change vs
vehicle)
Insulin [ng/mL]Mean SE
(% change vs vehicle)
(% change vs vehicle)
(% change vs vehicle)
Vehiclea
14 (30)
21.5 2.0
10.5 1.7
(À51%)#
3.4 0.7
(À84%)#
12.5 2.3
(À42%)#
4.3 2.0
(À80%)#
18.5 1.8
17.1 0.97
(À7%)
9.3 0.8
13.1 1.4
(+29%)#,*
19.8 1.5
(+113%)#
8.3 0.4
(À11%)
1.16 0.08
1.45 0.13
(+25%)
0.85 0.09
(À27%)#,**
1.35 0.15
(+16%)
0.38 0.08
0.61 0.07
(+60%)#
1.23 0.12
(+224%)#
0.52 0.19
(+37%)
14 (100)
Pred (3)
Pred (30)
17.9 1.0
(À3%)**
21.0 1.6
(+13%)
28.3 1.3
14.4 2.4
(+55%)#
1.48 0.10
1.09 0.20
(+53%)#
(+27%)#
(+187%)#
a
30% cremophor.
p < 0.05 versus vehicle.
#
*
p < 0.05 versus, pred. at 3 mg/kg group.
p < 0.05 versus, pred. at 30 mg/kg group.
**
the same study. Based on these results, 14 was estimated to have
an ED50 of less than 1 mg/kg.
scaffold. The t-butyl containing 5-azaindole derivative (14) demon-
strated good potency for GR and >100 fold selectivity over PR and
MR. This ligand displayed good in vitro anti-inflammatory activity,
To further evaluate this compound in vivo for its anti-inflamma-
tory activity and to determine its side effect profile, 14 was tested
in a therapeutic mouse collagen-induced arthritis (CIA) study. CIA
is a chronic model of inflammatory polyarthritis that shares many
features of human rheumatoid arthritis such as joint inflammation
and bone erosion. In this study, female B10.RIII mice were immu-
nized with porcine type II collagen and complete Freund’s adjuvant
then monitored for signs of arthritis. Animals were enrolled into
the study at the first signs of disease characterized by joint and
paw swelling and treated daily with compound, prednisolone or
vehicle (i.e., 30% cremophor) for a period of five weeks. The animals
were monitored regularly and arthritis severity was assessed
through clinical scoring of all four paws based on a score of 1–4,
with 4 indicating the highest degree of severity. The clinical score
per paw was summed to give a maximal severity score of 16 for
each animal. The results suggested that once daily oral treatment
with compound 14 was effective in completely eradicating the dis-
ease progression at 100 mg/kg resulting in 84% inhibition by AUC
as compared to the vehicle (Mann–Whitney test: p < 0.05). Fur-
thermore, the lower dose of 14 reduced the disease severity by
51% (Mann–Whitney test: p < 0.05 vs vehicle), which suggested
an ED50 of 30 mg/kg (Graph 1 and Table 5). It is interesting to note
that comparable efficacies were observed in the LPS and the CIA
studies (82% in LPS vs 84% in CIA) at different doses. This phenom-
enon is not uncommon and can be attributed in part to the funda-
mental differences between the acute inflammatory and the
chronic disease-relevant animal models.
The impact of the disease on biomarkers in a CIA mouse is com-
plex. To avert such complexities and to gauge the effect of com-
pound 14 on the metabolic side effect markers more accurately,
the CIA protocol was repeated with healthy non-immunized
B10.RIII mice that were of the same age as those enrolled in the
CIA study. At the end of the study, serum was collected and ana-
lyzed for biomarkers such as body fat, triglycerides, free fatty acids,
and insulin (Table 5). In this study, 14 did not increase body fat and
free fatty acids at the 100 mg/kg dose, which was statistically sig-
nificant relative to prednisolone at the equivalent anti-inflamma-
tory dose of 30 mg/kg. For these dose groups, a similar trend was
also observed in body weight changes (data not reported). These
findings were in line with the in vitro dissociated profile of 14 in
the MMTV assay. However, 14 demonstrated a dose-dependent in-
crease in the triglycerides and the serum insulin levels that sug-
gested a lack of dissociation based on these biomarkers.
which translated into excellent potency in the LPS-induced TNF-
a
production mouse model and good efficacy in the CIA mouse study.
Additionally, 14 showed less impact on body fat and free fatty acids
in the CIA study when compared to prednisolone at an equivalent
anti-inflammatory dose. Although 14 has an attractive profile, fur-
ther optimization of drug-like properties (e.g., Cytochrome P450
3A4 inhibition) are required before a compound from this series is
advanced into development. Our optimization of the drug-like prop-
erties in this scaffold will be reported in due course.
References and notes
1. (a) Gupta, R.; Jindal, D. P.; Kumar, G. Bioorg. Med. Chem. 2004, 12, 6331; (b)
Buttgereit, F.; Saag, K. G.; Cutolo, M.; da Silva, J. A. P.; Bijlsma, J. W. J. Scand. J.
Rheumatol. 2005, 34, 14; (c) Parente, L. In Glucocorticoids; Goulding, N. J.,
Flowers, R. J., Eds.; Birkhauser: Boston, 2001; p 35; (d) Kirwan, J. R.; Balint, G.;
Szebenyi, B. Rheumatology 1999, 38, 100.
2. Pisu, M.; James, N.; Sampsel, S.; Saag, K. G. Rheumatology 2005, 44, 781.
3. Schäcke, H.; Döcke, W.-D.; Asadullah, K. Pharmacol. Ther. 2002, 96, 23.
4. Bodine, P. V.; Litwack, G. Receptor 1990, 1, 83.
5. Heck, S.; Kullmann, M.; Gast, A.; Ponta, H.; Rahmsdorf, H. J.; Herrlich, P.; Cato,
A. C. B. EMBO J. 1994, 13, 4087.
6. (a) Buckbinder, L.; Robinson, R. P. Curr. Drug Targets: Inflammation Allergy 2002,
1, 127; (b) Barnes, P. J. Clin. Sci. 1998, 94, 557; (c) Resche-Rigon, M.;
Gronemeyer, H. Curr. Opin. Chem. Biol. 1998, 2, 501; (d) Burnstein, K. L.;
Cidlowski, J. A. Mol. Cell. Endocrinol. 1992, 83, C1.
7. Wang, J.-C.; Shah, N.; Pantoja, C.; Meijsing, S. H.; Ho, J. D.; Scanlan, T. S.;
Yamamoto, K. R. Genes Dev. 2006, 20, 689.
8. (a) Ronacher, K.; Hadley, K.; Avenant, C.; Stubsrud, E.; Simons, S. S., Jr.; Louw,
A.; Hapgood, J. P. Mol. Cell. Endocrinol. 2009, 299, 219; (b) van der Laan, S.;
Meijer, O. C. Eur. J. Pharmacol. 2008, 585, 483; (c) Newton, R.; Holden, N. S. Mol.
Pharmacol. 2007, 72, 799.
9. (a) McMaster, A.; Ray, D. W. Nat. Clin. Pract. Endocrinol. Metab. 2008, 4, 91; (b)
Coghlan, M. J.; Elmore, S. W.; Kym, P. R.; Kort, M. E. In Annual Reports in
Medicinal Chemistry; Doherty, A. M., Ed.; Academic Press: New York, 2002; Vol.
37, p 167.
10. (a) Zhou, J.; Cidlowski, J. A. Steroids 2005, 70, 407; (b) Adcock, I. M. Pulm.
Pharmacol. Ther. 2000, 13, 115; (c) Katzenellenbogen, J. A.; Katzenellenbogen, B.
S. Chem. Biol. 1996, 3, 529.
11. Miner, J. N.; Hong, M. H.; Negro-Vilar, A. Expert Opin. Invest. Drugs 2005, 14,
1527.
12. For reviews see, (a) Razavi, H.; Harcken, C. In Anti-Inflammatory Drug Discovery;
Levin, J., Laufer, S., Eds.; RSC Publishing: London, 2012; p 481; (b) Berlin, M.
Expert Opin. Ther. Pat. 2010, 20, 855; (c) DeBosscher, K. J. Steroid Biochem. Mol.
Biol. 2010, 120, 96; (d) Schäcke, H.; Berger, M.; Hansson, T. G.; McKerrecher, D.;
Rehwinkel, H. Expert Opin. Ther. Pat. 2008, 18, 339; (e) Hudson, A. R.; Roach, S.
L.; Higuchi, R. I. Curr. Top. Med. Chem. 2008, 8, 750; (f) Regan, J.; Razavi, H.;
Thomson, D. In Annual Reports in Medicinal Chemistry; Macor, J. E., Ed.;
Academic Press: New York, 2008; Vol. 43, p 141; (g) Mohler, M. L.; He, Y.; Wu,
Z.; Hong, S.-S.; Miller, D. D. Expert Opin. Ther. Pat. 2007, 17, 37.
13. (a) Betageri, R.; Gilmore, T.; Kuzmich, D.; Kirrane, T. M.; Bentzien, J.;
Wiedenmayer, D.; Bekkali, Y.; Regan, J.; Berry, A.; Latli, B.; Kukulka, A. J.;
Fadra, T. N.; Nelson, R. M.; Goldrick, S.; Zuvela-Jelaska, L.; Souza, D.; Pelletier, J.;
Dinallo, R.; Panzenbeck, M.; Torcellini, C.; Lee, H.; Pack, E.; Harcken, C.;
In summary, we have identified the t-butyl group as a suitable
replacement for the CF3 moiety in the trifluoromethyl-aryl-pentanol

1940
H. Razavi et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1934–1940
Nabozny, G.; Thomson, D. S. Bioorg. Med. Chem. Lett. 2011, 21, 6842; (b) Riether,
18. (a) Meanwell, N. A. J. Med. Chem. 2011, 54, 2529; (b) Böhm, H.-J.; Banner, D.;
Bendels, S.; Kansy, M.; Kuhn, B.; Müller, K.; Obst-Sander, U.; Stahl, M.
ChemBioChem 2004, 5, 637.
19. The pKa values are calculated with ACDlabs v 12.00 (2008) from Advanced
Chemistry Development, Inc. 110 Yonge Street 14th Floor, Toronto, Ontario
M5C 1T4, Canada.
20. Bledsoe, R. B.; 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.
21. Kauppi, B.; Jakob, C.; Farnegardh, M.; Yang, J.; Ahola, H.; Alarcon, M.; Calles, K.;
Engstrom, O.; Harlan, J.; Muchmore, S.; Ramqvist, A.-K.; Thorell, S.; Ohman, L.;
Greer, J.; Gustafsson, J.-A.; Carlstedt-Duke, J.; Carlquist, M. J. Biol. Chem. 2003,
278, 22748.
D.; Harcken, C.; Razavi, H.; Kuzmich, D.; Gilmore, T.; Bentzien, J.; Pack, E. J., Jr.;
Souza, D.; Nelson, R. M.; Kukulka, A.; Fadra, T. N.; Zuvela-Jelaska, L.; Pelletier, J.;
Dinallo, R.; Panzenbeck, M.; Torcellini, C.; Nabozny, G. H.; Thomson, D. S. J. Med.
Chem. 2010, 53, 6681; (c) Regan, J.; Lee, T. W.; Zindell, R. M.; Bekkali, Y.;
Bentzien, J.; Gilmore, T.; Hammach, A.; Kirrane, T. M.; Kukulka, A. J.; Kuzmich,
D.; Nelson, R. M.; Proudfoot, J. R.; Ralph, M.; Pelletier, J.; Souza, D.; Zuvela-
Jelaska, L.; Nabozny, G.; Thomson, D. J. Med. Chem. 2006, 49, 7887.
14. Betageri, R.; Zhang, Y.; Zindell, R. M.; Kuzmich, D.; Kirrane, T. M.; Bentzien, J.;
Cardozo, M.; Capolino, A. J.; Fadra, T. N.; Nelson, R. M.; Paw, Z.; Shih, D.-T.; Shih,
C.-K.; Zuvela-Jelaska, L.; Nabozny, G.; Thomson, D. S. Bioorg. Med. Chem. Lett.
2005, 15, 4761.
15. Jaroch, S.; Berger, M.; Huwe, C.; Krolikiewicz, K.; Rehwinkel, H.; Schäcke, H.;
Schmees, N.; Skuballa, W. Bioorg. Med. Chem. Lett. 2010, 20, 5835.
16. Harcken, C.; Ward, Y.; Thomson, D.; Riether, D. Synlett 2005, 3121.
17. For detail descriptions of the binding (GR, PR, and MR) and the cellular (IL-6
and MMTV) assays, as well as the in vivo mouse models (LPS and collagen-
induced arthritis) see Ref. 13b.
22. Small-Molecule Drug Discovery Suite 2013-2: Glide, version 6.0, Schrödinger,
LLC, New York, NY, 2013.
23. MolecularOperatingEnvironment (MOE), 2012.1001;ChemicalComputingGroup
Inc., 1010 Sherbooke St. West, Suite 910, Montreal, QC H3A 2R7, Canada, 2012.
24. Lee, T. W.; Proudfoot, J. R.; Thomson, D. S. Bioorg. Med. Chem. Lett. 2006, 16, 654.