Discovery of indazole ethers as novel, potent, non-steroidal glucocorticoid receptor modulators

Article abstract of DOI:10.1016/j.bmcl.2016.10.052

A structure-based design approach led to the identification of a novel class of indazole ether based, non-steroidal glucocorticoid receptor (GR) modulators. Several examples were identified that displayed cell potency in the picomolar range, inhibiting LPS-induced TNF-α release by primary peripheral blood mononuclear cells (PBMCs). Additionally, an improved steroid hormone receptor binding selectivity profile, compared to classical steroidal GR agonists, was demonstrated. The indazole ether core tolerated a broad range of substituents allowing for modulation of the physiochemical parameters. A small sub-set of indazole ethers, with pharmacokinetic properties suitable for oral administration, was investigated in a rat antigen-induced joint inflammation model and demonstrated excellent anti-inflammatory efficacy.

Full text of DOI:10.1016/j.bmcl.2016.10.052

Accepted Manuscript
Discovery of indazole ethers as novel, potent, non-steroidal glucocorticoid re-
ceptor modulators
Martin Hemmerling, Karl Edman, Matti Lepistö, Anders Eriksson, Svetlana
Ivanova, Jan Dahmén, Hartmut Rehwinkel, Markus Berger, Ramon Hendrickx,
Matthew Dearman, Tina Jellesmark Jensen, Lisa Wissler, Thomas Hansson
PII:
S0960-894X(16)31090-3
DOI:
http://dx.doi.org/10.1016/j.bmcl.2016.10.052
BMCL 24355
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
12 August 2016
17 October 2016
18 October 2016
Please cite this article as: Hemmerling, M., Edman, K., Lepistö, M., Eriksson, A., Ivanova, S., Dahmén, J.,
Rehwinkel, H., Berger, M., Hendrickx, R., Dearman, M., Jellesmark Jensen, T., Wissler, L., Hansson, T., Discovery
of indazole ethers as novel, potent, non-steroidal glucocorticoid receptor modulators, Bioorganic & Medicinal
Chemistry Letters (2016), doi: http://dx.doi.org/10.1016/j.bmcl.2016.10.052
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Discovery of indazole ethers as novel, potent, non-steroidal
glucocorticoid receptor modulators
Martin Hemmerling a,*, Karl Edman b, Matti Lepistö a, Anders Eriksson a, Svetlana Ivanova c,
Jan Dahmén c, Hartmut Rehwinkel d, Markus Berger d, Ramon Hendrickx a, Matthew
Dearman a, Tina Jellesmark Jensen a, Lisa Wissler b and Thomas Hansson a
a Respiratory, Inflammation & Autoimmunity, Innovative Medicines and Early Development Biotech Unit, AstraZeneca, Pepparedsleden 1,
Mölndal 43183, SE
b Discovery Sciences, Innovative Medicines and Early Development Biotech Unit, AstraZeneca, Pepparedsleden 1, Mölndal 431 83, SE
c AstraZeneca R&D Lund, Scheelevägen 1, Lund 22187, Sweden, SE
d Medicinal Chemistry Berlin, Candidate Generation & External Innovation, Drug Discovery, Bayer Pharma AG, Berlin 13353, DE
* Corresponding author. Tel.: +46 31 7761647; fax: +46 31 7763818
e-mail address: martin.hemmerling@astrazeneca.com
Keywords:
Non-steroidal glucocorticoid
Glucocorticoid receptor modulators
Indazoles
Anti-inflammatory activity
ABSTRACT
A structure-based design approach led to the identification of a novel class of indazole ether based,
non-steroidal glucocorticoid receptor (GR) modulators. Several examples were identified that
displayed cell potency in the picomolar range, inhibiting LPS-induced TNF-α release by primary
peripheral blood mononuclear cells (PBMCs). Additionally, an improved steroid hormone receptor
binding selectivity profile, compared to classical steroidal GR agonists, was demonstrated. The
indazole ether core tolerated a broad range of substituents allowing for modulation of the
physiochemical parameters. A small sub-set of indazole ethers, with pharmacokinetic properties
suitable for oral administration, was investigated in a rat allergen-induced joint inflammation model
and demonstrated excellent anti-inflammatory efficacy.
The glucocorticoid receptor (GR) is a transcription factor belonging to the nuclear hormone receptor
family. Ligand activation triggers both positive and negative regulation of gene transcription and
subsequently pleiotropic effects on metabolism and the immune system. Glucocorticoids are widely
prescribed for the treatment of inflammatory and autoimmune conditions and are highly effective1, but
their therapeutic use can be limited due to numerous adverse effects following chronic administration.2
and 3
In the strive to dissociate anti-inflammatory effects from systemic metabolic side effects, non-
steroidal GR modulators have been described4, 5, 6, 7 and 8 wherein the rigid steroidal core is replaced by

alternative scaffolds. This approach allows for modified ligand-induced receptor conformations
compared to steroidal GR ligands. The distinct modulation of the receptor with selective GR
modulators (SGRMs) is envisioned to result in differential co-regulator recruitment and DNA
recognition, potentially leading to dissociated pharmacology. We recently reported the discovery of a
series of SGRMs with a differentiated co-regulator peptide interaction profile, exemplified by
compound 1 (Fig. 1).9 In parallel, a joint effort with Bayer Pharma AG focused on the exploration of
two series of GR modulators, structurally distinct from 1 and exemplified by compounds 2 and 3.10 and
11
Subsequent combination of these starting points provided novel GR modulators of the generic
structures 7 and 8 with improved potency and a property profile allowing for oral administration.
1
Prednisolone (Pred): X = H, R = H
Dexamethasone (Dex): X = F, R = Me
2
3
Scheme 1: Synthetic GR ligands
The lead compound 1 exhibited efficacy in inflammatory pharmacodynamic (PD) models following
inhaled administration. Unfortunately, the analysis of metabolites in hepatocytes and human liver
microsomes revealed the formation of potentially mutagenic 4-aminoindazoles.9 and 12 The primary
objective for optimizing the lead compound became to modify the structural motif of the linker to
avoid the formation of these undesired metabolites. In the initial exploration, the secondary amine was
replaced by an ether linkage or stabilized by ring-closure to a pyrrolidine. However, both compounds
4 and 5 demonstrated reduced cellular transrepression (TR) activity as assessed in a reporter gene
assay measuring the agonistic effect on GR as an inhibition of AP-1 mediated transcriptional activity
(Table 1).13 These findings were in accordance with the X-ray structure of the GR ligand binding
domain (LBD) in complex with 1, indicating that the secondary aniline forms a hydrogen bond to the
backbone of Leu563 in helix 3.9 As such, it constitutes an integral part of the ligand pharmacophore
highly sensitive to change. Compound 6 was made in an attempt to reduce the potential for steric clash
with the Leu563 carbonyl by shifting the ether oxygen to the indazole 5-position, but again cellular TR
activity was reduced.

Table 1: Non-steroidal GR ligands
4
6
5
Comp
Cell TRa, b
IC50 (nM)
Eff.c (%)
HLMd
CLint
1
2
3
4
5
6
1.47
1.15
0.47
60
11
98
89
75
100
94
71
47
78
<10
16
150
52
>200
a Values as mean of at least two experiments
b Transrepression activity, assessed as inhibition of AP-1 signalling
c Efficacy, maximum response relative to Dexamethasone (100 %)
d Intrinsic clearance in human liver microsomes (µL/min/mg)
As minor structural changes were not successful, alternative scaffolds mimicking the lead
compound 1 were considered. An entirely new approach was identified when the GR-ligands 2 and 3
were evaluated. X-ray structure determination of GR in complex with ligand 214 enabled the docking
of the related and more potent ligand 3 into the GR-LBD with high confidence. Guided by the
superimposition of the binding conformations of ligands 1 and 3 (Fig. 1A), it was suggested to move
the linker attachment from the 4-position, as seen in 1, to the 5-position. Again, the anilinic motif was
exchanged for an ether linkage to mitigate the potential risk of releasing mutagenic metabolites. As the
5-ether linker aligns along a new trajectory, similar to that defined by the linker unit of compound 3, a
phenyl group was introduced as a branch from the ether α-carbon to match the vector of the equivalent
trisubstituted phenyl of 3 for positioning into the lipophilic pocket which is targeted by the 17α-
substituent of corticosteroids (17α-pocket). Finally, as the bulky sulfonamide mesityl group could no
longer be accommodated by the 17α-pocket, a smaller cyclopropyl residue was incorporated to avoid
potential steric clashes. The resulting indazole ether 7a (Fig. 1C), the prototype of the new indazole
ether series, retained the transrepression (TR) agonist potency of the original starting points (0.91 nM,
Table 2). In the following we report the optimization of 7a to compounds with improved potency
which confers some examples with improved metabolic stability allowing for oral dosing in preclinical
in vivo models. (Tables 2 and 3).

Figure 1. (A) Overlay of compound 3 docking (cyan) on x-ray structure of GR-LBD in complex with compound 1
(magenta). (B) Overview of the design. The ligand motifs have been color coded according to ligand binding pocket region;
blue binds near the gate-keeper residues (Gln570 and Arg611), red motif interacts with Asn564, green motif extend into the
17α-pocket. (C) Docking of hybrid ligand 7a on GR-LBD.
All indazole ethers were prepared following the route exemplified in Scheme 2 based on a building
block strategy wherein modified norephedrines are coupled with an iodo-indazole. The N-(4-
fluorophenyl)-5-iodo-indazole 9 is readily accessible from a condensation reaction of ortho-
fluorinated benzaldehyde with phenyl hydrazine under basic conditions at elevated temperatures. The
primary amine of the indazole ether scaffold 10 was formed from an Ullmann coupling reaction of (-)-
norephedrine with the indazole 9, along with N-coupled byproducts limiting the yields for this step.15
The sulfonamides 7 were obtained by condensation of the primary amine 10 with the corresponding
sulfonyl chlorides. Amides 8 were obtained typically by condensation of 10 with acids using standard
amide coupling conditions, e.g. reaction with propylphosphonic anhydride (T3P) and triethylamine.
The trifluoroacetamide 8b was prepared by an alternative protocol using the corresponding anhydride
in presence of triethylamine.

b
a
9
10
c
d
8b
7a
Scheme 2. Representative synthesis of indazole ethers. Reagents and conditions: a) Cs2CO3 (3.0 equiv.) NMP, r.t. to
150°C; b) 1R,2S-(-)-norephedrine, CuI (0.05 equiv.) Cs2CO3 (2.2 equiv.) PrCN, 125°C; c) c-propylsulfonyl chloride
(3.0 equiv.) triethylamine (1.4 equiv.), DCM; d) 2,2,2-trifluoroacetic anhydride (2.0 equiv.), triethylamine (6.0 equiv.),
DCM, 0°C to r.t.
Norephedrine analogues were prepared as exemplified for (1R,2S)-2-amino-1-(6-methoxy-3-
pyridyl)propan-1-ol, the precursor for example 8s (Scheme 3). Applying a variant of a Weinreb-Nahm
α-amino-ketone synthesis, an aryl Grignard reagent was added to the deprotonated Weinreb amide 11
to yield the aryl ketone 12.16 Meerwein-Ponndorf-Verley reduction then proceeded with a high degree
of diastereoselectivity to afford the protected anti-amino alcohol 13 (de > 99%) which was deprotected
to yield the norephedrine analogue 14 (ee > 99%).17
a, b, c
d
e
14
11
12
13
Scheme 3. Synthesis of norephedrine analogues by example. Reagents and conditions: a) i-PrMgCl*LiCl (1.25 equiv.), 0°C
to 10°C; b) 2-methoxy-5-bromopyridine, i-PrMgCl (1.25 equiv.), THF, 0°C – r.t.; c) 20% aq. HCl, 0°C to 10°C; d) Al(OiPr)3
(0.2 equiv.), i-PrOH (11 equiv.), toluene, 50°C; e) HCl (4N in dioxane, 10 equiv.).
The structure activity relationship of the indazole ethers was evaluated by measuring
transrepression activity (AP-1 pathway) in a reporter gene assay13 and inhibition of LPS-induced
TNFα release in PBMCs. In addition to potency, intrinsic clearance in human liver microsomes was
measured to provide an indication of the potential for systemic efficacy (Table 2 and 3). For the
sulfonamide residue R1, methyl (7c) was preferred over longer alkyl chains exemplified by 7b, 7g and
7h. Best potencies were observed for short secondary alkyl groups as demonstrated by the examples

7a, 7e and 7f, ranked in the order c-Pr > i-Pr > 2-Bu. The trifluoromethyl group (R1 in 7d) resulted in
reduced potency compared to secondary alkyl groups or the smaller methyl. Substitution of the phenyl
branch of the central norephedrine moiety was well tolerated and here lipophilic or electron donating
groups (7i, 7k) seemed to be preferred over the polar, electron deficient substituents (7j).
Table 2: Sulfonamide series of indazole ethers
7
Comp
R1
R2
Cell TRa
human PBMCa, b
IC50 (nM)
Eff. (%)d
HLMd
ClogPe
IC50 (nM)
Eff. (%)c
CLint
5.5
7a
7b
7c
7d
7e
7f
c-Pr
Pr
H
0.91
13
97
90
3.1
79
16
76
H
ND
ND
ND
1.1
-
-
6.0
5.0
7.1
5.8
6.3
6.4
7.1
6.1
3.9
5.5
Me
H
7.4
37
83
<10
<10
22
CF3
i-Pr
H
75
-
H
H
2.8
6.6
44
96
86
110
-
2-Bu
iBu
93
5.1
120
>200
180
78
7g
7h
7i
H
81
ND
ND
0.67
0.56
2.3
pentyl
c-Pr
c-Pr
c-Pr
H
100
0.75
1.4
0.62
66
-
4-SMe
4-S(O)2Me
3-OMe
110
110
120
77
83
77
7j
<3
7k
53
a Values as mean of at least two experiments
b LPS-induced TNF-α release in PBMCs
c Efficacy, maximum response relative to Dexamethasone (100 %)
d Intrinsic clearance in human liver microsomes (μL/min/mg)
e Calculated using BioByte, version 4.3
ND: not determined
For the original lead 1, the sulfonamide functionality is essential for functional activity as the N-
S(O)2-C dihedral angle positions the mesityl group into the 17α-pocket. The indazole ether scaffold
however accesses this lipophilic cleft through its central norephedrine motif and therefore the rationale
behind the need for the sulfonamide group was invalidated. Therefore, it was possible to replace the
sulfonamide by a carboxamide group with the aim to align the peripheral amide residues R1 along the
trajectory of the ether linker (Table 3). Comparison of the matched pair 7a (Table 2) and 8a (Table 3)
demonstrated the validity of this approach as these examples showed very similar potency. The tert-
butyl amide was equally well tolerated, and 8h was equipotent to the corresponding cyclopropyl
reference 8a. Also in analogy to the sulfonamides, R1 residues as the small methyl (8c) or the extended
methoxyethyl (8d) resulted in decreased potency. However, the trifluoroacetamide in 8b (0.13 nM)
shifted the TR cell potency in a different direction to that observed for the trifluoromethyl sulfonamide

7d (37 nM). Introduction of 5-membered heterocycles as R1, exemplified by compounds 8f and 8g,
resulted in a similar gain in potency as observed for the trifluoromethyl residue, affording low
picomolar IC50 activity. In contrast, the larger lipophilic phenyl amide (8e) only yielded nanomolar
potency.
Table 3: Indazole ether amides
oxadiazolyl A =
5-Me-1,3,4-oxadiazole-2-yl
oxadiazolyl B =
8
5-Me-1,2,4-oxadiazole-3-yl
Comp
R1
Ar
R2
Cell TRa
IC50 (nM)
human PBMCa, b
IC50 (nM)
Eff. (%)c
HLMd
ClogPe
Eff. (%)c
CLint
8a
8b
8c
8d
8e
8f
c-Pr
CF3
Ph
H
0.62
0.13
3.10
5.09
1.41
0.09
0.09
0.66
0.25
0.08
0.05
0.21
0.03
0.84
0.04
0.04
8.83
0.40
94
92
6.01
78
<10
<10
<10
>200
<10
<10
<3
5.4
5.9
4.8
5.2
6.5
4.4
5.4
6.0
6.0
6.0
7.1
5.7
6.5
3.0
7.1
5.8
3.9
4.7
Ph
H
0.50
ND
83
-
Me
Ph
H
96
C2H4OMe
Ph
Ph
H
H
94
64.1
ND
77
-
Ph
87
oxadiazolyl A
oxadiazolyl B
t-Bu
Ph
H
104
115
108
113
121
114
108
97
0.16
0.27
1.20
0.89
0.43
0.31
1.44
0.10
3.15
0.29
0.15
ND
79
71
83
79
84
84
77
80
77
87
84
-
8g
8h
8i
Ph
H
Ph
Ph
H
62
t-Bu
4-MeO
3-MeO
4-Et
4-Et
4-MeS
4-Me(O)2S
H
>200
130
>200
82
8j
t-Bu
Ph
8k
8l
t-Bu
Ph
CH2OH
CF3
Ph
8m
8n
8o
8p
8r
8s
Ph
110
<10
40
CH2OH
CF3
Ph
111
98
2-naphthyl
3-quinolinyl
3-pyridyl
3-pyridyl
CF3
H
102
91
<10
20
c-Pr
H
c-Pr
4-MeO
102
2.16
73
7.3
a Values as mean of at least two experiments
b LPS-induced TNF-α release in PBMCs
c Efficacy, maximum response relative to Dexamethasone (100 %)
d Intrinsic clearance in human liver microsomes (μL/min/mg)
e Calculated using BioByte, version 4.3
ND: not determined
The significant increase in potency for the most favorable amides was perpetuated when lipophilic
substitution R2 was introduced to the aryl branch. Within the small series of tert-butyl amides 8h to 8k
it can be seen that the R2 substituents methoxy and ethyl have the potential to increase potency by
approximately one order of magnitude. Equally, methylthio substitution was beneficial in combination
with a trifluoroacetamide (8m) which further improved the potency by about 4-fold (33 pM) in
comparison with the unsubstituted reference 8b. Similarly, improved TR agonism was observed when

lipophilic bicyclic aryls were used as aryl branch, exemplified by compounds 8o and 8p. This pair
indicates that N-heterocycles might be tolerated without significant loss of potency. This general
statement is brought into context by the pyridyl pair 8r and 8s which provided diverging results. While
the unsubstituted 3-pyridyl 8r lost potency by about one order of magnitude compared with the phenyl
analogue 8a, para-methoxy substitution as in 8s compensated for the heteroatom insertion and fully
restored picomolar potency. It is noteworthy that in example 8n, the electron deficient methylsulfonyl
substituent added polarity without apparent loss of potency. This, combined with the polar
hydroxymethyl amide group, led to 8n having equivalent potency to the reference example 8a but with
lipophilicity (ClogP) lowered by two orders of magnitude.
In terms of metabolic stability the indazole ethers gave variable results. For the subseries of
cyclopropyl sulfonamides 7a, 7j and 7k it was shown that the modest intrinsic clearance in 7a could
be moved in opposite directions depending on the nature of the phenyl substituent. Microsomal
clearance decreased in line with lipophilicity when a polar methylsulfonyl substituent was introduced
in 7j. However, addition of a metabolically labile methoxy substituent in compound 7k resulted in
accelerated metabolic degradation at conserved lipophilicity; 7a vs 7k (both ClogP 5.5). In summary,
no strong correlation between stability in microsomes and calculated lipophilicity was observed. For
instance, some of the most lipophilic examples, such as the examples 8b or 8p, exhibited good
metabolic stability in vitro. Apparently, the indazole ether core is not prone to rapid metabolic
degradation unless labile groups R1 or R2 are introduced.
For a more detailed understanding of the structure activity relationship for the indazole ether amide
series, the X-ray structure of the GR ligand binding domain (LBD) in complex with the compound 8b
was determined (Fig. 3).18
Figure 2. (A) Overview of the GR LBD in complex with compound 8b (magenta). The dotted lines represent putative
hydrogen bonds. (B) The central phenyl moiety extends directly into the ligand entry pocket where helices 3 and 7 meet.

In agreement with the compound 1 complex structure,9 the fluorophenyl rearranges the gatekeeper
residues (Arg611 and Gln570) to open up the inducible pocket between helices 3 and 5. At the other
end of compound 8b, the trifluoroacetamide forms several hydrogen bonding interactions with the
protein. Specifically, the amide proton and the trifluoromethyl group form a bidentate interaction by
combining hydrogen bonding and electrostatic interactions with Asn564. Similar bivalent interactions
with Asn564 have been observed previously and are associated with increased potency of GR
agonists.19 and 20 Accordingly, the preference for the R1 amide residues present in 8b, 8f and 8g could be
linked to the additional interaction, either electrostatic (CF3) or hydrogen bonding (oxadiazoles), with
the amide residue of Asn564. Gln642 has rearranged into a conformation, which facilitates another
direct hydrogen bond to the amide carbonyl of 8b. As such, the amide group forms a distinct bridge
linking helices 3 and 7 together. At the center of the ligand, the phenyl extends into the 17α pocket
where helices 3 and 7 meet.
The diverging structure activity relationship for the trifluoromethyl residue R1 in 7d and 8b might
be rationalized from docking the sulfonamide 7d into the co-crystal structure of 8b bound to the GR-
LBD. Potential repulsive interactions of sulfonamide groups with increased steric demand, such as
trifluoromethyl or isobutyl, with the norephedrine phenyl moiety or interference with the Gln642
residue would prevent the formation of the hydrogen bonding network as described above for the
preferred binding conformation of compound 8b (Fig. 3).
Figure 3. Trifluoromethyl sulfonamide 7d (cyan C) in the same conformation as 7a in Fig. 1C, superimposed with
trifluoroacetamide 8b (magenta C). Purple dotted lines denote hydrogen bonds, red double arrows denote repulsive
interactions.
The distinct binding mode of the indazole ethers suggests a strong preference of the (1R,2S)
configuration of the central norephedrine moiety. In support of that, the (1S,2R) enantiomer of 8j was
found to be more than 200-fold less active than compound 8j. Therefore, other stereoisomers were no
longer pursued in this study.

Selectivity versus other nuclear hormone receptors, based on relative binding affinities, was
excellent throughout the entire indazole ether series and is shown for representative compounds (Table
4). While no examples of binding to mineralocorticoid (MR), androgen (AR) or estrogen receptors
(ER) could be detected, a modest binding activity versus the progesterone receptor (PR) was observed
in a few cases.
Table 4. Cross-reactivity to steroidal hormone receptors21
Comp
GR binding
PR binding
AR binding
MR binding
ERα binding
IC50 (µM)
ERβ binding
IC50 (µM)
IC50 (µM)
IC50 (µM)
IC50 (µM)
IC50 (µM)
7a
0.0060
0.0212
0.0025
0.0027
0.0029
0.0090
0.0028
0.0061
0.0061
1.16
0.79
>12
0.91
>10
0.63
>20
2.08
0.44
>5
>10
>10
>20
>10
>10
>10
>10
0.018
0.040
>10
>10
>20
>10
>10
>10
>10
>10
ND
>10
>10
>20
>10
>10
>10
>10
>10
ND
7i
>5
7j
>10
>5
8b
8n
8o
>10
>10
>20
>2
8s
Pred
Dex
>20
Values as mean of at least two experiments
ND: not determined
The lowest selectivity (37-fold) was obtained for compound 7i and is derived from the unexpected
low binding affinity for GR (IC50 = 22 nM), in contrast to the potent cellular transrepression activity
(cell TR IC50 = 0.75 nM, Table 2).
The activities obtained in the AP-1 repression reporter gene assay (cell TR) translated well into an
anti-inflammatory effect, assessed as inhibition of TNFα release following LPS-stimulation in both
human PBMCs and, for a smaller set of compounds, in human and rat whole blood (Tables 5 and 7).
Analysis of the underlying data set for the indazole ethers revealed a modest correlation between the
anti-inflammatory activities in PBMCs and whole blood (R2 = 0.44, Fig. 4A). When accounting for
binding to plasma protein in both assays,22, 23 and 24 a highly improved correlation between the
calculated unbound potencies was found (R2 = 0.91, Fig. 4B), suggesting that unspecific binding to
plasma protein (PPB) modulates GR driven activity in vitro.
Table 5: In vitro pharmacology and unspecific binding
Comp
human PBMCa, b
IC50 (nM)
Eff. (%)c
human whole blooda, b
IC50 (nM)
Eff. (%)c
human PPB
(%)-free

7j
8b
0.56
0.50
3.1
8
3
29
14
8
9
1.60
0.01
4.00
0.15
>41
83
78
73
70
90
89
88
89
8n
16
8s
2.2
42
Pred
45
160
a Values as mean of at least two experiments
b LPS-induced TNF-α release
c Efficacy, maximum response relative to Dexamethasone (100%)
Figure 4: In vitro inhibition of LPS-stimulated TNFα expression; correlation of cellular (PBMC) and whole blood
derived activities. A) Uncorrected potencies. B) Corrected for unspecific binding to assay protein content
To examine the in vivo potential of this new compound class, compounds 7j, 8b, 8n and 8s were
selected for oral PK studies. These compounds are stable in human liver microsomes but are
differentiated in their property profile, in particular with regard to solubility and plasma protein
binding due to differences in lipophilicity of three orders of magnitude (Tables 3 and 5). Compound
8b deserves particular attention since it lies outside the typical oral drug space as defined by Lipinski’s
rule of five25 and seems to exhibit an unfavorable oral drug profile due to low solubility and excessive
plasma protein binding. Nonetheless, all studied compounds showed good exposure in the rat and the
amides 8b, 8n and 8s gave similar results for the area under the curve (AUC). Compound 7f displayed
unusual rat pharmacokinetics, characterized by a vastly extended half-life (t1/2 = 82 h), which could
pose additional challenges if further progressed and was therefore deselected.

Table 6: Properties and rat pharmacokinetics
Sola
Hep CLint
CL (i.v.)c
Vss (i.v.)c
[L/kg]
t1/2 (i.v.)c
[h]
AUC (p.o.)c
F (p.o.)c
[%]
b
[µM]
[µL/min/1E6]
[mL/min/kg]
[µM*h]
7j
<0.1
< 0.02
14
<1.0
3.2
0.62
19
4.2
14
81.9
11.4
5.1
41.5
7.1
109
53
69
66
13
8b
8n
<2.7
5.8
16
4.8
15
8.1
8s
3
28
10.4
0.27
5.9
Pred
162
27
54
0.63
0.35
a
Thermodynamic solubility in 0.1M phosphate buffer pH 7.4 at 25°C
b Intrinsic clearance in rat hepatocytes
c Blood PK Han Wistar Rat; oral (p.o.) dose: 10 μmol/kg; intravenous (i.v.) dose: 1 μmol/kg; values as mean of at least
two experiments; CL: clearance; Vss: volume of distribution at steady state; t : terminal half-life; AUC: area under the
½
curve, extrapolated to infinity; F: oral bioavailibility. For experimental details refer to the supporting information.
With in vivo pharmacokinetic profiles suitable for oral administration established, the efficacy of
the selected compounds was investigated in the Streptococcal cell wall (SCW) reactivation model of
arthritis in rat by their ability to reduce antigen challenge-induced ankle swelling (Fig. 5).26
Figure 5. Inhibition of ankle swelling in the rat SCW model: Rats were orally dosed qd with either compound as a
solution/suspension in vehicle (0.5% HPMC/0.1% Tween in water) or with vehicle alone for 8 days. Data shows the
mean SEM of ankle AUC for each animal using all data combined from either 2 (8b), 4 (8s) or a single (8n)
experiment(s) with at least n=10 animals/treatment group. Percent inhibition of ankle AUC in the vehicle treated
animals was calculated and the statistical significance evaluated by ANOVA (Kruskal-Wallis) with post test analysis
(Dunn’s multiple comparison): (*) p<0.05, (**) p<0.01, (***) p<0.001. For experimental details refer to the supporting
information.
All examined indazole ethers achieved a reduction of ankle swelling at lower doses compared to
prednisolone (Pred) (Table 7). The observed effects, reported as ED50 SCW, ranked the compounds
aligned with lipophilicity and in vitro potency. Compound 8b proved most efficacious followed by 8s,
with the more polar indazole ether 8n substantially less potent. The excellent efficacy displayed by
compound 8b is a result of both the excellent in vitro potency and the unexpected high exposure seen
in the PD animals. When reviewing the PK-PD relationship by relating exposure (AUC) at ED50 to
whole rat blood potency, all compounds, including the steroid prednisolone, show similar ratios AUC /
ED50 WB within a 3-5-fold range, indicating a robust PK / PD relationship. The highly diverging
affinity to plasma protein for the compounds studied appears not to be a factor modulating in vivo
efficacy beyond the effects seen in vitro. This finding is in agreement with an earlier reported minor

impact of protein binding on the systemic bioactivity of highly potent corticosteroids27. Compound 8s
(AZD2906) was nominated for clinical development.
Table 7. Pharmacology of indazole ethers in cell, blood and SCW rat model
8b
8n
8s
Pred
Rat PBMC IC50 (nM)a, b
Rat Whole Blood EC50 (nM)a, b
Rat PBB, (%)-free
0.33
3.3
1.1
0.29
7.5
13
18
50
0.011
12 [ 8]
80
4.0
0.63
40 [ 10]
35
31
ED50 SCW (µg/kg) [CI]c
AUC (nM*h)d
800 [ 600]
2550 [ 800]
246
14
590
12
AUC / EC50 Whole Blood (h)
24
5
a Values as mean of at least two experiments
b LPS-induced TNFα release
c Processed from experiments of 10 individual animals (Lewis rat) per dose (CI: 95% confidence interval)
d Exposure in PD animals at ED50 dose, during 24h at final day, quantified as area under the curve
In conclusion, indazole ethers represent a novel, potent, nonsteroidal class of GR modulators.
Exchanging the sulfonamide function by amides enhanced in vitro potency that could be further
increased by lipophilic substitution of the central aryl moiety. The most potent examples demonstrated
efficacy in a cellular transrepression assay at picomolar concentrations. The indazole ether scaffold
resists rapid metabolic clearance and enabled the discovery of GR modulators suitable for oral drug
administration. Selected examples were investigated in the SCW rat PD model where in vitro potency
translated well into in vivo efficacy. We propose that this class of GR modulators shows potential for
treatment of inflammatory diseases. Considering their high intrinsic potency, the indazole ethers might
also show potential for the inhaled treatment of inflammatory diseases. The nonsteroidal indazole
ether scaffold offers multiple opportunities for further structural elaboration and therefore presents an
excellent starting point for investigation and development of selective GR modulators with
differentiated pharmacology.
Acknowledgements
We thank Anders Jerre, Irene Mile and Charlotte Kåse for the in vitro pharmacology assays, Stefan
Eirefelt for the in vivo PK studies, Simon Cruwys for project leadership and in vivo co-coordination
and Mark Furber and Frank Narjes for proof reading this manuscript.
References and notes
1. Coutinho, A.E.; Chapman, K.E. Mol. Cell. Endocrinol. 2011, 335, 2.
2. Moghadam-Kia, S.; Werth, V.P. Int. J. Dermatol. 2010, 49, 239.
3. Dahl R. Respir. Med. 2006, 100, 1307.

4. Jaroch, S.; Berger, M.; Huwe, C.; Krolikiewicz, K.; Rehwinkel, H.; Schäcke, H.; Schmees, N.;
Skuballa, W. Bioorg. Med. Chem. Lett. 2010, 20, 5835.
5. Schäcke, H.; Berger, M.; Hansson, T.; McKerrecher, D.; Rehwinkel, H. Expert Opin. Ther.
Patents, 2008, 18, 339.
6. Reuter, K.C.; Grunwitz, C.R.; Kaminski, B.M.; Steinhilber, D.; Radeke, H.H.; Stein, J. J.
Pharmacol. Exp. Ther. 2012, 341, 68.
7. Biggadike, K.; Bledsoe, R.K.; Coe, D.M.; Cooper, T.W.; House, D.; Iannone, M.A.; Macdonald,
S.J.; Madauss, K.P.; McLay, I.M.; Shipley, T.J.; Taylor, S.J.; Tran, T.B.; Uings, I.J.; Weller, V.;
Williams, S.P. Proc. Nat. Acad. Sci. U. S. A. 2009, 106, 18114.
8. Gilmore, J.L.; Sheppeck II, J.E.; Wang, J.; Dhar, T.G.; Cavallaro, C.; Doweyko, A.M.; McKay,
L.; Cunningham, M.D.; Habte, S.F.; Nadler, S.G.; Dodd, J.H.; Somerville, J.E.; Barrish, J.C.
Bioorg. Med. Chem. Lett. 2013, 23, 5448.
9. Edman, K.; Ahlgren, R.; Bengtsson, M.; Bladh, H.; Backstrom, S.; Dahmen, J.; Henriksson, K.;
Hillertz, P.; Hulikal, V.; Jerre, A.; Kinchin, L.; Kåse, C.; Lepistö, M.; Mile, I.; Nilsson, S.;
Smailagic, A.; Taylor, J.; Tjornebo, A.; Wissler, L.; Hansson, T. Bioorg. Med. Chem. Lett. 2014,
24, 2571.
10. Berger, M.; Rehwinkel, H.; Schäcke, H.; Bäurle, S.; Schmees, N. WO 2007/104582.
11. Berger, M.; Dahmén, J.; Rehwinkel, H.; Jaroch, S.; Schaecke, H. WO 2008/006627.
12. Shamovsky, I.; Ripa, L.; Borjesson, L.; Mee, C.; Norden, B.; Hansen, P.; Hasselgren, C.;
O'Donovan, M.; Sjö, P. J. Am. Chem. Soc. 2011, 133, 16168.
13. Cellular transrepression assay: Human bronchogenic carcinoma cells, ChaGo K1(ATCC: HTB
168), transfected with a 12-O-Tetradecanoylphorbol-13-acetate (TPA) response elements (TRE)-
LacZ construct (5x TRE-LacZ) are incubated for 24h in the presence of test compounds. The
transrepression activity is measured as the inhibition of phorbol myristate acetate (PMA)-
stimulated TRE-LacZ activity, resulting in reduced β-galactosidase activity, monitored in a
fluorometric assay format using 4-methylumbelliferyl-β-D-galatoside (MUG) as substrate.
14. Berger, M.; Edman, K.; Wissler, L.; Neuhaus, R.; Rehwinkel, H.; Schäcke, H.; Jaroch, S. in
manuscript. The coordinates for the complex of GR LBD with compound 2 have been deposited
in the PDB with accession code 5g3j.
15. Job, G.E.; Buchwald, S.L. Org. Lett. 2002, 4, 3703.
16. Conrad, K.; Hsiao, Y.; Miller, R. Tetrahedron Lett. 2005, 46, 8587.
17. Yin, J.; Huffman, M.A.; Conrad, K.M.; Armstrong III, J.D. J. Org. Chem. 2006, 71, 840.
18. The experimental details are described in the attached Supplementary material: Structure
determination. The coordinates for the complex of GR LBD with compound 8b have been
deposited in the PDB with accession code r5g5w.
19. Biggadike, K.; Bledsoe, R.K.; Hassell, A.M.; Kirk, B.E.; McLay, I.M.; Shewchuk, L.M.; Stewart,
E.L. J. Med. Chem. 2008, 51, 3349.
20. Sheppeck II, J.E.; Gilmore, J.L.; Xiao, H.Y.; Dhar, T.G.; Nirschl, D.; Doweyko, A.M.; Sack, J.S.;
Corbett, 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.
21. Binding to hormone receptors: Ligand binding to GR, PR, AR and ER was assessed in
fluorescence polarisation (FP) ligand competition assays, which measures the relative affinity of
test compounds for the full length or ligand binding domain of the nuclear receptor in the
presence of a fluorescently labelled ligand. Ligand binding to MR was assessed in a ligand
competition binding scintillation proximity assay (SPA), which measures binding of test
compounds to the MR ligand binding domain in the presence of 3H-aldosterone. IC50 values were
determined by fitting the data to a 4-parameter logistic equation.
22. Calculation of unbound drug concentration: Unbound blood concentrations were obtained from
measurements in rat plasma with an albumin concentration of 489 µM and total drug
concentrations of 10 µM. The equilibrium constant K was calculated under the assumption of a
1:1 binding to albumin only (eq 1).
ꢀꢁ
K ꢀ
(eq. 1)
ꢀ
ꢂ ꢁ
¤°££
¤°££
PD = protein (albumin) drug complex); Pfree = unbound protein concentration (albumin); Dfree
unbound drug concentration. The free drug concentrations in the PBMC assay could
=

subsequently be calculated using this equilibrium constant, K, and the albumin concentration of
35.6 µM in the diluted fetal calf serum in the assay medium. The detailed description of the
derived equations are included in the supplementary material.
23. Davies, B.; Morris, T. Pharm. Res. 1993, 10, 1093.
24. Price, P.J.; Gregory, E.A. In vitro, 1982, 18, 576.
25. Lipinski, C.A.; Lombardo, F.; Dominy, B.W.; Feeney, P.J. Adv. Drug Delivery Rev. 1997, 46, 3.
26. Schimmer, R.C.; Schrier, D.J.; Flory, C.M.; Dykens, J.; Tung, D.K.; Jacobson, P.B.; Friedl, H.P.;
Conroy, M.C.; Schimmer, B.B.; Ward, P.A. J. Immunol. 1997, 159, 4103.
27. Daley-Yates, P.T. Br. J. Clin. Pharmacol. 2015, 80, 372.

Graphical abstract