Structure-based design of substituted hexafluoroisopropanol- arylsulfonamides as modulators of RORc

Article abstract of DOI:10.1016/j.bmcl.2013.10.054

The structure-activity relationships of T0901317 analogs were explored as RORc inverse agonists using the principles of property- and structure-based drug design. An X-ray co-crystal structure of T0901317 and RORc was obtained and provided molecular insight into why T0901317 functioned as an inverse agonist of RORc; whereas, the same ligand functioned as an agonist of FXR, LXR, and PXR. The structural data was also used to design inhibitors with improved RORc biochemical and cellular activities. The improved inhibitors possessed enhanced selectivity profiles (rationalized using the X-ray crystallographic data) against other nuclear receptors.

Full text of DOI:10.1016/j.bmcl.2013.10.054

Bioorganic & Medicinal Chemistry Letters 23 (2013) 6604–6609
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure-based design of substituted hexafluoroisopropanol-aryl-
sulfonamides as modulators of RORc
a
b
a
a
Benjamin P. Fauber a, , Gladys de Leon Boenig , Brenda Burton , Céline Eidenschenk , Christine Everett ,
Alberto Gobbi a, Sarah G. Hymowitz a, Adam R. Johnson a, Marya Liimatta a, Peter Lockey b,
Maxine Norman b, Wenjun Ouyang a, Olivier René a, Harvey Wong a
⇑
a Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080, USA
b Argenta, Units 7–9 Spire Green Centre, Flex Meadow, Harlow, Essex CM19 5TR, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
The structure–activity relationships of T0901317 analogs were explored as RORc inverse agonists using
the principles of property- and structure-based drug design. An X-ray co-crystal structure of T0901317
and RORc was obtained and provided molecular insight into why T0901317 functioned as an inverse ago-
nist of RORc; whereas, the same ligand functioned as an agonist of FXR, LXR, and PXR. The structural data
was also used to design inhibitors with improved RORc biochemical and cellular activities. The improved
inhibitors possessed enhanced selectivity profiles (rationalized using the X-ray crystallographic data)
against other nuclear receptors.
Received 20 September 2013
Revised 23 October 2013
Accepted 25 October 2013
Available online 4 November 2013
Keywords:
RORc
Ó 2013 Elsevier Ltd. All rights reserved.
ROR
c
T0901317
X-ray structure
IL-17
Inflammation
Autoimmune
TH-17 cells are a subset of pro-inflammatory CD4+ T cells whose
production of IL-17 is dependent upon the nuclear receptor (NR)
The benzene-sulfonamide compound T0901317 (Fig. 1,
compound 1) was initially described by a team at Tularik (now
Amgen) as a potent agonist of the liver X receptor (LXR).12 Since
the original disclosure of 1, other investigators have demonstrated
that 1 is an agonist of multiple additional NRs including pregnane
X receptor (PXR) and farnesoid X receptor (FXR).13,14 Recently, a
team from the Scripps Research Institute in Florida described 1
as an inverse agonist of RORc.15 We were intrigued by the binding
of 1 to RORc, and embarked upon a campaign to explore the SAR of
related analogs as inverse agonists of RORc.
retinoic acid receptor-related orphan receptor gamma (ROR
c or
RORc, also known as NR1F3).1 IL-17 has been implicated in the
pathology of several autoimmune diseases including rheumatoid
arthritis (RA), inflammatory bowel disease, psoriasis, and multiple
sclerosis.2 Reduced IL-17 expression via antibody blockade or
genetic knock-out in mice resulted in a marked decrease in inflam-
mation in several murine models.3,4 Monoclonal neutralizing anti-
bodies against IL-17 are also being evaluated in human clinical
trials and have demonstrated proof-of-concept activity in psoriasis,
RA, and non-infectious uveitis.5
In an effort to aid our analog design, we co-crystallized com-
pound 1 with the human RORc ligand-binding domain (LBD) and
obtained a 2.9 Å resolution X-ray structure (Table S1 Supplemen-
tary data).16 The helices of the RORc LBD co-crystal structure
In addition to modulating the expression of the IL-17 family of
cytokines, RORc signaling promotes other pro-inflammatory cyto-
kines including IL-22 and GM-CSF in TH17 cells.6,7 Thus, inhibition
of RORc may lead to enhanced anti-inflammatory activity over the
IL-17-neutralizing antibodies that are currently under investiga-
tion in clinical trials. These observations, coupled with the viability
of RORc genetic knock-out mice,8,9 make RORc an attractive target
for modulating human autoimmune diseases.10,11
adopted the canonical three-layered
a-helical fold common to
most NRs,17 with 1 buried in the hydrophobic core (Fig. S1a, Sup-
plementary Data). The RORc LBD helices were in a similar arrange-
ment to that of the previously reported co-crystal structure of the
RORc LBD with 25-hydroxycholesterol (Fig. 1, compound 2) [PDB:
3L0L],18 with the exception that there is a lack of density for the
C-terminus of helix 11, all of helix 110 and helix 12 in the co-crystal
structure with 1, suggesting these helices lack regular structure
when RORc interacts with compound 1. The disordered helix 12
should disfavor recruitment of co-activator proteins by prohibiting
⇑
Corresponding author. Tel.: +1 650 467 5773.
E-mail address: Fauber.Benjamin@gene.com (B.P. Fauber).
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.10.054

B. P. Fauber et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6604–6609
6605
OH
F3C
F3C
Phe357
25
H
Phe367
O
O
S
H
OH
N
H
H
HO
F3C
3
1
2
Figure 1. Chemical structures of T0901317 (1) and 25-hydroxycholesterol (2).
the interaction of Glu483 on helix 12 with the conserved LXXLL
‘charge-clamp’ motif in co-activator proteins.19 This structural
observation was concordant with the RORc LBD and steroid recep-
tor co-activator-1 (SRC1) recruitment biochemical assay results in
which 2 behaved as an agonist; whereas, 1 was an inverse
agonist.15,18
Gln265
Leu303
Met344
Val340
Arg346
Within the ligand binding pocket of the RORc co-crystal struc-
ture, the hexafluoro-2-isopropanol moiety of 1 could fit in two
equal but opposite orientations of the 2FoꢀFc electron density
map (Fig. S1b, Supplementary data). In both orientations of the
hexafluoroisopropanol moiety, there were no notable hydrogen
bonds (<3.2 Å) from the hydroxyl group to the protein or crystallo-
graphic water molecules. This observation was in stark contrast to
Figure 2b. Co-crystal structure of compound 1 (yellow) in complex with RORc
(grey). The phenyl-sulfonamide group of compound 1 accessed a parallel-displaced
p–p
stacking interaction with Phe367 (3.3–3.6 Å), and Phe367 engaged neighboring
Phe357 in an edge-to-face stacking interaction (3.9–4.0 Å). Leu303 (grey),
p–p
Val340 (grey), and Met344 (grey) made notable hydrophobic contacts with
compound 1. The ligand binding pocket polar residues Gln265 (grey) and Arg346
(grey) are also depicted.
the co-crystal structures of 1 with the LBDs of human LXRa [PDB:
1UHL]21 and LXRb [PDB: 1UPV],22 where the hexafluoroisopropa-
nol hydroxyl group made a strong hydrogen bond interaction with
the ligand, along with other residue differences in the receptors,
gave rise to a unique binding orientation of 1 in RORc relative to
a histidine residue in LXR
a (His421, 2.4 Å) and LXRb (His435,
2.5 Å), resulting in stabilization of helices 11-12, recruitment of
the co-activator, and receptor agonist activity.23 The lack of a
hydrogen bond interaction between 1 and His458 in RORc may ex-
plain why this ligand behaved as an inverse agonist of RORc.
In addition to the lack of ligand hydrogen bonding in the RORc
co-crystal structure, 1 was also rotated ꢁ30° in the binding pocket
relative to its orientation in the LXR co-structures (Fig. 2a). The res-
idue difference that potentially enforced the ꢁ30° rotation of 1 was
the presence of Leu303 in RORc, which was a smaller alanine res-
that observed in co-structures of 1 with LXRa, LXRb, and PXR
[PDB: 2O9I].24 As a result of the unique binding orientation of 1
in RORc, the phenyl-sulfonamide group of 1 was allowed to access
a
parallel-displaced
(3.3–3.6 Å).25 The
further enhanced by the edge-to-face
Phe367 with neighboring Phe357 (3.9–4.0 Å) (Fig. 2b). The afore-
mentioned receptor-ligand stacking interaction was not
observed in the LXR and PXR co-structures with 1. We predicted
that compounds capable of forming receptor-ligand stacking
p
–
p
stacking interaction with Phe367
stacking interaction of 1 and Phe367 was
stacking interaction of
p–p
p–p
p–p
idue in LXR
a
(Ala261) and LXRb (Ala275). The planar rotation of
p–p
interactions with Phe367 in RORc could exhibit increased RORc
potency and selectivity.26 In order to test this hypothesis, we de-
signed and synthesized aryl-sulfonamide analogs of 1, allowing
exploration of their structure–activity relationships.
We evaluated analogs of 1 in two biochemical assays.27 A radi-
oligand competition binding assay provided information on the
binding affinity of a compound for the RORc LBD by displacement
of a tritiated 25-hydroxycholesterol radioligand from the RORc
ligand binding pocket. A co-activator peptide binding assay moni-
tored the ability of the RORc LBD to bind to a co-activator peptide
derived from SRC1. Disrupting the recruitment of the SRC1 co-
activator peptide is indicative of compounds that display inverse
agonist activity with RORc.
Ala261
Leu303
In addition to optimizing the complementarity of the ligand to
the RORc binding pocket, we paid close attention to the calculated
logP value (clogP <5 was preferred) of potential analogs, as our
starting molecule (1) was fairly lipophilic (clogP = 4.5).28 Our goal
was to explore the SAR of T0901317 analogs against RORc while
maintaining favorable physiochemical properties to favor drug-like
pharmaceutical properties.29,30 Thus, we initiated our exploration
of the phenyl-sulfonamide moiety via a Hansch analysis with a ni-
trile substituent (Table 1, compounds 3–5, clogP = 4.4–4.2, respec-
tively).31 It was clear from the Hansch analysis that the para-nitrile
substituent (5) was preferred and provided a slight improvement
in the ligand-lipophilicity efficiency (LLE)32 value versus 1, while
it maintained similar potency values in the RORc binding and
SRC1 recruitment assays. In an effort to probe the effects of various
para-substituents, we synthesized and tested a set of analogs with
His421
His458
Figure 2a. A 2.9 Å resolution co-crystal structure of compound
complex with RORc (grey) and a 2.9 Å resolution co-crystal structure of compound
1 (green) in complex with LXR (magenta) [PDB: 1UHL]. All atoms of the LBD for
each receptor have been overlaid. His435 (grey) in RORc does not make a notable
hydrogen bond interactions (<3.2 Å) with 1 (yellow); whereas, 1 (green) does make
1 (yellow) in
a
a strong hydrogen bond (2.4 Å) with His421 (magenta) in LXR
are shown as dashed lines (black). The Leu303 (grey) side-chain in RORc is a smaller
Ala261 (magenta) in LXR and the residue difference potentially gives rise to
different rotations of the ligand in the respective receptors.
a. Hydrogen bonds
a

6606
B. P. Fauber et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6604–6609
Table 1
Table 2
Structure–activity relationships of the sulfonamide analogs
Structure–activity relationships of the N-alkyl analogs
OH
F3C
OH
F3C
F3C
F3C
O
O
O O
S
S
N
R
N
R
F3C
RORc
Compd
R-group
RORc
IC50
RORc LLEb
RORc SRC1
EC50 (lM) [%eff.]
Compd
R-group
RORc LLEb
RORc SRC1
EC50
(lM) [%eff.]
a
c
(
l
M)
a
c
IC50
(
l
M)
12
13
14
15
16
17
18
19
20
(CH2)2OH
(CH2)2OMe
(CH2)2SO2Me
(CH2)2NHCOMe
(CH2)2CO2H
4
3
2.3
1.8
—
—
—
2.3
1.4
1.0
—
1 [ꢀ80%]
>10
>10
>10
>10
0.190 [ꢀ97%]
4 [ꢀ70%]
8 [ꢀ64%]
ND
1
3
4
5
6
7
8
9
10
11
Ph
0.060
0.520
0.190
0.050
0.051
0.034
0.748
4
2.7
1.9
2.4
3.1
2.3
2.8
2.7
2.1
1.9
2.1
0.054 [ꢀ99%]
0.920 [ꢀ91%]
0.100 [ꢀ93%]
0.061 [ꢀ96%]
0.020 [ꢀ100%]
0.019 [ꢀ99%]
1 [ꢀ88%]
>10
>10
>10
0.245
0.519
0.467
>10
(2-CN)Ph
(3-CN)Ph
(4-CN)Ph
(4-Me)Ph
(4-F)Ph
3-Pyridyl
5-(2-N-Me)pyrazole
c-Pr
Et
(CH2)2CF3
Bn
H
ND
ND
>2 [ꢀ22%]
5
See the Supplementary data for experimental details associated with each assess-
ment. All assay results are reported as the arithmetic mean of at least two separate
runs. ND = not determined.
Bn
0.520
See the Supplementary data for experimental details associated with each assess-
ment. All assay results are reported as the arithmetic mean of at least two separate
runs. ND = not determined.
a
Inhibition of the RORc LBD and [3H2]-25-hydroxycholesterol interaction.
b
Ligand-lipophilicity efficiency (LLE) was calculated using the RORc biochemical
IC50 value and calculated logP.32
Inhibition of the RORc LBD and [3H2]-25-hydroxycholesterol interaction.
c
a
Inhibition of RORc LBD recruitment of the SRC1 co-activator peptide; negative
percent efficacy denotes inverse agonism relative to the basal activity of apo-RORc
LBD.
b
Ligand-lipophilicity efficiency (LLE) was calculated using the RORc biochemical
IC50 value and calculated logP.32
c
Inhibition of RORc LBD recruitment of the SRC1 co-activator peptide; negative
percent efficacy denotes inverse agonism relative to the basal activity of apo-RORc
LBD.
could not directly interact with the Gln265, it may be able to indi-
rectly interact with the Gln265 or Arg346 via a water-mediated
hydrogen bond network. We made additional attempts to directly
engage those RORc binding pocket polar residues, or indirectly
interact with them via water-mediated hydrogen bonding. Several
analogs containing a range of N-alkyl polar side chains were syn-
thesized, and these analogs were also poorly tolerated in the RORc
biochemical assays (Table 2, compounds 13–16). After exploring
polar substitution, we focused our attention on lipophilic N-alkyl
substituents. The N-ethyl analog was synthesized to explore the
influence of the 2,2,2-trifluoroethyl substituent on potency and
also potentially improve the LLE value (Table 2, compound 17).
As it turned out, the potency values of compound 17 in the binding
assay and the SRC1 recruitment assay were approximately fourfold
less than the parent molecule (1) and 17 provided no improvement
in LLE. The N-3,3,3-trifluoropropyl and N-benzyl analogs were syn-
thesized to explore the role of larger lipophilic groups in the RORc
binding pocket (Table 2, compounds 18 and 19). Both analogs had
similar RORc biochemical potency, but they were inferior to the
potency of 1. Additionally, replacement of the N-alkyl group with
a hydrogen resulted in an inactive molecule (Table 2, compound
20), and thus confirmed the importance of the N-2,2,2-trifluoro-
ethyl group in maintaining RORc binding affinity.
The hexafluoro-2-isopropanol moiety of 1 binds near the inter-
face of helix 11 and helix 3 in the RORc co-crystal structure. As
mentioned above, the alcohol on the hexafluoroisopropanol group
did not make any notable hydrogen bonds to the adjacent protein
residues or to the crystallographic water molecules. To investigate
the importance of the alcohol functionality, we synthesized the
hexafluoroisopropanol O-methyl ether. It lost considerable binding
and functional potency in the RORc biochemical assays (Table 3,
compound 21). Computational docking studies initially suggested
that there could be space to accommodate the hexafluoroisopropa-
nol O-methyl ether, given the distribution of the 2Fo–Fc electron
density map in the hexafluoroisopropanol containing region of
the ligand (1) co-crystal structure with RORc. The decrease in RORc
biochemical activity for compound 21 also suggested that there
may have been an interaction with the alcohol that was not appar-
ent in the 2.9 Å resolution co-crystal structure of 1 and RORc. To
different rp values (Table 1, compounds 6 and 7).33,34 The (para-
fluorophenyl)-sulfonamide moiety (7) provided an appropriate
balance between RORc binding potency, LLE, and potent full in-
verse agonism in the SRC1 peptide binding assay. Exploration of
aromatic heterocyclic-sulfonamide analogs did not provide any
potency improvements (Table 1, compounds 8 and 9). The lack of
potency from the cyclopropyl-sulfonamide analog demonstrated
that the partial sp2 character present in the cyclopropane ring
was not sufficient to engage the aforementioned
p–p stacking
interaction with the Phe367 residue in the RORc ligand binding site
(Table 1, compound 10).35 Extending the aromatic moiety by one
atom to generate a benzylic-sulfonamide analog resulted in a loss
of binding affinity, presumably due to a disruption of the
stacking interaction with the receptor (Table 1, compound 11).
As discussed above, the phenyl-sulfonamide group on 1 ac-
cessed a unique stacking interaction in the RORc co-crystal
p–p
p–p
structure that was not observed in the co–structures of 1 with
LXR and PXR. As a result of the phenyl-sulfonamide moiety adopt-
ing a unique orientation, the N-2,2,2-trifluoroethyl group also ac-
cessed a different vector than those previously observed in LXR
or PXR. The N-trifluoroethyl group resided in a region of the RORc
ligand binding pocket rich in lipophilic residues, and it made sev-
eral lipophilic interactions36 (3.3–3.4 Å) with Leu303, Val340, and
Met344 (Fig. 2b). We were intrigued that in the same region of
the RORc ligand binding pocket, the 3b-OH of 25-hydroxycholes-
terol (2) formed a direct hydrogen bond with Gln265 (3.1 Å) and
interacted with Arg346 via a water-mediated hydrogen bond
(2.8 Å) [PDB: 3L0L]. Thus, we initiated an exploration of various
N-alkyl analogs of 1 to probe the potential engagement of the
Gln265 or Arg346 residues in the RORc ligand binding pocket.
Exploration of N-alkyl analogs of 1 commenced with the syn-
thesis of an ethyl-alcohol analog (Table 2, compound 12). The
hydroxyl group on 12 was poorly tolerated and resulted in a signif-
icant loss in potency. Computational modeling studies with this
analog suggest that the lipophilic region surrounding the N-trifluo-
roethyl group in the co-crystal structure of 1 and RORc may not tol-
erate polar functionality. We hypothesized that if the ethyl-alcohol

B. P. Fauber et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6604–6609
6607
Table 3
cellular selectivity of the potent inverse agonists, these compounds
were also subjected to a panel of LXR, PXR, and FXR HEK293-Gal4
construct cell assays and we monitored the activation/agonism
of their basal transcriptional activity to determine if there were
any improvements in selectivity versus compound 1. Ultimately,
the most potent and selective compounds identified in the ROR
and NR cell assay panels were progressed into a cytokine assay
that monitored their ability to inhibit IL-17 production in an
isolated human peripheral blood mononuclear cell (PBMC) frac-
Structure–activity relationships of the hexafluoro-2-isopropanol group
R
O
O
S
N
F3C
RORc
Compd
R-group
RORc LLEb
RORc SRC1
EC50 (lM) [%eff.]
a
c
IC50
(l
M)
tion. We also used interferon gamma (INFc
) and CellTiter-GloÒ
21
22
23
24
25
C(CF3)2OMe
CH2OH
CO2H
CONH2
SO2NH2
0.900
>10
>10
>10
9
0.6
—
—
—
2.7
9 [ꢀ45%]
>10
>10
>10
>10
(CTG) cell culture assays as positive controls to monitor for activity
against non-TH17 cell cytokines and aberrant cytotoxicity,
respectively.27
A key weakness in using 1 to assess the effect of RORc inhibition
in cellular or in vivo studies is its lack of selectivity for RORc over
other NRs. In particular, while compound 1 was a 463 nM inhibit-
ior in the RORc Gal4 cellular assay, it was only approximately ten-
fold selective against the other ROR family members (Table 5).
See the Supplementary data for experimental details associated with each assess-
ment. All assay results are reported as the arithmetic mean of at least two separate
runs.
a
Inhibition of the RORc LBD and [3H2]-25-hydroxycholesterol interaction.
b
Ligand-lipophilicity efficiency (LLE) was calculated using the RORc biochemical
IC50 value and calculated logP.32
c
Further, compound 1 was a weak FXR agonist (EC50 = 1
lM) but
Inhibition of RORc LBD recruitment of the SRC1 co-activator peptide; negative
was a potent LXR and LXRb agonist, with 373 and 156 nM EC50
values, respectively, and 643- and 284-fold activation of these
a
percent efficacy denotes inverse agonism relative to the basal activity of apo-RORc
LBD.
receptors, respectively.
Several of our compounds showed improvements in these re-
spects as they possess increased potency and selectivity relative
to compound 1. For instance, compound 5 demonstrated increased
inverse agonism of RORc (EC50 = 192 nM) and better selectivity
over the other ROR family members. Compound 5 also elicited a
weaker agonist response in FXR, LXRb, and PXR, with no detectable
probe the importance of various ligand hydrogen bond donors in
that region of the ligand binding pocket, we synthesized analogs
containing a benzylic alcohol, a carboxylic acid, a primary amide,
and a primary sulfonamide (Table 3, compounds 22–25, respec-
tively). All of these analogs had negligible activity in both RORc
biochemical assays. These results suggested that the loss of po-
tency observed with compound 21 was probably due to a steric
requirement of the RORc binding pocket that did not favor the O-
methyl ether moiety.
LXRa agonist activity (Table 5). Compound 17 was a modest inhib-
itor of the SRC1 co-activator assay and this potency, in turn, trans-
lated into modest inhibition of the RORc Gal4 cellular assay
(EC50 = 593 nM). Compound 17 also displayed limited agonist
In an attempt to explore alternative binding modes of the li-
gand, we explored several core change analogs of 1 (Table 4).
Two isomers of the N-ethyl analog were examined as comparisons
to the parent N-ethyl sulfonamide compound 17. Cyclization of the
N-ethyl group onto the core phenyl ring to generate an indoline
N--benzene-sulfonamide analog was less potent than the parent
compound (Table 4, compound 26). Movement of the N-ethyl-N-
phenylsulfonamide group to the meta-position of the core phenyl
ring also resulted in loss of potency in the RORc binding assay
and a several-fold loss of potency in the SRC1 co-activator recruit-
ment assay (Table 4, compound 27). In a comparison of functional
groups, the sulfonamide moiety of 1 was converted into an amide,
resulting in a large potency decrease (Table 4, compound 28). This
large drop in potency was somewhat anticipated given the major
differences in the conformational preferences of sulfonamides
and amides.37 Removal of the N-2,2,2-trifluoroethyl group on 28
resulted in a further loss of potency (Table 4, compound 29), sim-
ilar to that observed in the previous comparison of compounds 1
and 20.
activity in the FXR and LXRa cell assays, and comparable levels
of LXRb and PXR agonist activity to that noted with compound 1.
Compound 7 displayed potent inhibition in the RORc Gal4 cellular
assay (EC50 = 89 nM) and was also >60-fold selective over the other
ROR family members. Compound 7 also provided less potent and
lower-fold agonist activity in the FXR and LXR cellular assays than
that observed with compound 1. Compound 7 retained the same-
fold activation of the PXR receptor as seen with compound 1, but
was fivefold less potent.
The improved selectivity profiles noted with compounds 5 and
7 were potentially due to the successful application of structure-
based design principals. Both of these compounds were designed
to have an improved receptor-ligand
the RORc Phe367 ligand binding pocket residue. As noted earlier,
this stacking interaction was not observed in co-structures
p–p stacking interaction with
p–p
of 1 with LXR and PXR. Thus, providing this interaction between
compounds and RORc was not expected to facilitate interactions
with LXR and PXR and, in turn, increased the selectivity window
for RORc.
During the course of our SAR studies, we synthesized several of
the published ROR modulators as control compounds—including
SR1078.38 We found that SR1078 lacked detectable binding po-
tency in our radiometric assay and lacked activity in our SRC1
co-activator recruitment assay (Table 4, compound 30). A similar
observation was recently described by a team at GlaxoSmithKline
using an analogous co-activator recruitment assay.20 The related
sulfonamide analog, SR1001,39 also displayed limited potency in
our RORc biochemical assays (Table 4, compound 31).
We also profiled the potent inverse agonist compounds identi-
fied with the SRC1 co-activator recruitment assay in a series of
HEK293-Gal4 construct ROR cellular assays. Three known isoforms
of ROR (RORc, RORb, and RORa) were profiled under cellular recep-
tor antagonist conditions by which we monitored the suppression
of their basal transcriptional activity. In order to assess the NR
Compounds 1, 5, and 7 were progressed into human primary
cell cytokine production assays to assess their abilities to inhibit
production of the pro-inflammatory cytokine, IL-17.40 Compound
1 displayed modest inhibition of IL-17 (EC50 = 3
lM), while com-
pound 5 was threefold more potent (EC50 = 1 M). The differences
l
in IL-17 inhibition values between the two compounds exhibited
roughly the same difference in potency observed in the RORc
Gal4 cellular assay for the two compounds, providing confidence
that our early-stage assays were accurately rank-ordering com-
pounds. It was also noteworthy that neither compound showed
any activity in the INFc or CTG assays, demonstrating that the com-
pounds were not indiscriminately suppressing cytokine produc-
tion, nor were they grossly cytotoxic. Compound 7 was the most
potent inhibitor of IL-17 expression (EC50 = 132 nM), with no
detectable inhibition in the INFc and CTG assays (Table 6).

6608
B. P. Fauber et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6604–6609
Table 4
Structure–activity relationships of the core change analogs
OH
F3C
F3C
R
Compd
R-group
RORc
IC50
RORc LLEb
RORc SRC1
EC50 (lM) [%eff.]
a
c
(lM)
O
S
O
26
0.883
2.2
5 [ꢀ76%]
N
N
27
28
29
30
0.329
2.2
2.1
—
0.900 [ꢀ94%]
4 [ꢀ79%]
>10
S
O
O
O
2
N
F3C
O
O
>10
>10
N
H
—
>10
N
H
CF3
O
O
S
S
N
H
31
2
4.0
0.911 [ꢀ88%]
NH
O
N
See the Supplementary data for experimental details associated with each assessment. All assay results are reported as the arithmetic mean of at least two separate runs.
a
Inhibition of the RORc LBD and [3H2]-25-hydroxycholesterol interaction.
b
Ligand-lipophilicity efficiency (LLE) was calculated using the RORc biochemical IC50 value and calculated logP.32
c
Inhibition of RORc LBD recruitment of the SRC1 co-activator peptide; negative percent efficacy denotes inverse agonism relative to the basal activity of apo-RORc LBD.
Table 5
Potency and selectivity profiles in Gal4 human NR transcription reporter assaysa
Compd
RORc
RORb
RORa
FXR
EC50
LXR
EC50
a
LXRb
EC50
PXR
EC50
b
b
b
b
EC50
(
l
M)
EC50
(
lM)
EC50
(
lM)
(
lM) [act.]
(lM) [act.]
(
lM) [act.]
(lM) [act.]
1
5
7
0.463
0.192
0.089
0.593
6
5
6
4
5
7
6
5
1 [77]
3 [46]
1 [64]
9 [22]
0.373 [643]
>10
0.719 [303]
3 [349]
0.156 [284]
0.902 [199]
0.385 [237]
0.246 [74]
0.036 [22]
0.571 [27]
0.194 [30]
0.039 [24]
17
See the Supplementary data for experimental details associated with each assessment. All assay results are reported as the arithmetic mean of at least two separate runs.
a
All assays were conducted in HEK293-Gal4 cellular constructs. All ROR assays monitored the suppression of their respective basal transcriptional activities, an outcome
consistent with inverse agonist activity of ligands with these the receptor. The FXR, LXR, and PXR cellular assays monitored the activation of the basal transcriptional
activities, an outcome consistent with agonist activity of the receptor.
b
The maximum-fold activation [act.] was the fold increase of the compound’s agonist activity relative to the apo-receptor.
Syntheses of the analogs discussed above are described
in Scheme S1 (Supplementary data). The route toward the
aryl-sulfonamide analogs commenced with sulfonylation of
the commercially available 4-substituted aniline scaffolds under
the conditions described by Li et al. (Scheme S1, intermediates
33a–k and compounds 20 and 31).41 The secondary aniline was
then alkylated with either an alkyl-halide or alkyl-triflate, in the
presence of potassium carbonate and refluxing acetonitrile, to pro-
duce various N-alkyl analogs (Scheme S1, compounds 1, 3–13, 17,
19).42 The N-ethyl(methylsulfone) analog 14 required a modified
approach in which compound 20 was treated with 2-chloroethyl-
methylsulfide and potassium carbonate in warm DMF, followed
by oxidation with m-CPBA in DCM. Compound 1 was further
transformed into the O-methyl ether analog under basic alkylation
conditions (Scheme S1, compound 21). The 4-carboxamide and 4-
carboxylic acid core analogs were synthesized from the
corresponding nitrile and ester intermediates, respectively
(Scheme S1, intermediates 34 and 35). Hydrolysis of the nitrile
intermediate, under basic peroxide conditions,43 produced the pri-
mary amide analog, and saponification of the ester intermediate
revealed the carboxylic acid analog (Scheme S1, compounds 23
and 24, respectively).
We also generated analogs using an alternative synthetic se-
quence in which the aniline nitrogen was first alkylated under ba-
sic conditions or reductive amination conditions.44 Subsequent
sulfonylation or acylation of the secondary aniline intermediates

B. P. Fauber et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6604–6609
6609
Table 6
8. Kurebayashi, S.; Ueda, E.; Sakaue, M.; Patel, D. D.; Medvedev, A.; Zhang, F.;
Jetten, A. M. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 10132.
9. Sun, Z.; Unutmaz, D.; Zou, Y.-R.; Sunshine, M. J.; Pierani, A.; Brenner-Morton, S.;
Mebius, R. E.; Littman, D. R. Science 2000, 288, 2369.
10. Huh, J. R.; Littman, D. R. Eur. J. Immunol. 2012, 42, 2232.
11. Burris, T. P.; Busby, S. A.; Griffin, P. R. Chem. Biol. 2012, 19, 51.
12. Schultz, J. R.; Tu, H.; Luk, A.; Repa, J. J.; Medina, J. C.; Li, L.; Schwendner, S.;
Wang, S.; Thoolen, M.; Mangelsdorf, D. J.; Lustig, K. D.; Shan, B. Genes Dev. 2000,
14, 2831.
13. Mitro, N.; Vargas, L.; Romeo, R.; Koder, A.; Saez, E. FEBS Lett. 2007, 581, 1721.
14. Houck, K. A.; Borchert, K. M.; Hepler, C. D.; Thomas, J. S.; Bramlett, K. S.;
Michael, L. F.; Burris, T. P. Mol. Gen. Metab. 2004, 83, 184.
15. Kumar, N.; Solt, L. A.; Conkright, J. J.; Wang, Y.; Istrate, M. A.; Busby, S. A.;
Garcia-Ordonez, R. D.; Burris, T. P.; Griffin, P. R. Mol. Pharmacol. 2010, 77, 228.
16. See the Supplementary data for the experimental details associated with the
co-crystal structure of compound 1 and the RORc LBD. The coordinates and
structure factors for the complex between RORc and compound 1 have been
deposited with the protein data bank and assigned the accession code 4NB6.
17. Wurtz, J. M.; Bourguet, W.; Renaud, J. P.; Vivat, V.; Chambon, P.; Moras, D.;
Gronemeyer, H. Nat. Struct. Biol. 1996, 3, 87.
Potency in human IL-17 and INF
c
production assaysa
Compd
IL-17 PBMC
IL-17 PBMC
%max.
inhibition
IFN
EC50
c
CTG
EC50
EC50
(
lM)
(l
M)
(lM)
1
5
7
3
1
81
74
83
>20
>20
>20
>20
>20
>20
0.132
See the Supplementary data for experimental details associated with each assess-
ment. All assay results are reported as the arithmetic mean of at least two separate
runs.
a
All assays were conducted using peripheral blood mononuclear cells (PBMCs)
isolated from human whole blood. Interferon gamma (INFc
) and CellTiter-GloÒ
(CTG) cell culture assays were used as positive controls to monitor for non-TH17 cell
cytokine activity and adverse off-target effects on cell physiology, respectively.
provided the final compounds (Scheme S1, compounds 15, 18, 25,
and 28). Compound 16 was made using a similar reaction sequence
in which the N-alkyl ester group was carried through the synthesis
and saponified in the final step to reveal the carboxylic acid group.
Compounds 29 and 30 were made via the direct acylation of 2-(4-
aminophenyl)-1,1,1,3,3,3-hexafluoropropan-2-ol (32a). Compound
27 was generated from intermediate 41 under basic alkylation
conditions (Scheme S2, Supplementary data).41 Compound 2224
and compound 2641 were synthesized according to literature
procedures.
In summary, the structure activity relationships of T0901317 (1)
analogs were successfully explored to generate improved RORc in-
verse agonists using structure-based drug design principles. An
X-ray co-crystal structure of 1 and the RORc LBD facilitated the
structure-based drug design efforts. The resulting compounds have
improved RORc biochemical activity, cellular activity, and inhibit IL-
17 production in primary human cells. These inhibitors also possess
enhanced selectivity profiles against other NRs that can be rational-
ized using X-ray crystallographic data. This coordinated effort led to
the generation of RORc inverse agonists with suitable biochemical
and cellular properties, and will allow investigation of the role of
RORc in autoimmune diseases and other biological roles.
18. Jin, L.; Martynowski, D.; Zheng, S.; Wada, T.; Xie, W.; Li, Y. Mol. Endocrinol.
2010, 24, 923.
19. Germain, P.; Staels, B.; Dacquet, C.; Spedding, M.; Laudet, V. Pharmacol. Rev.
2006, 58, 685.
20. Zhang, W.; Zhang, J.; Fang, L.; Zhou, L.; Wang, S.; Xiang, Z.; Li, Y.; Wisely, B.;
Zhang, G.; An, G.; Wang, Y.; Leung, S.; Zhong, Z. Mol. Pharmacol. 2012, 82, 583.
21. Svensson, S.; Östberg, T.; Jacobsson, M.; Norström, C.; Stefansson, K.; Hallén, D.;
Johansson, I. C.; Zachrisson, K.; Ogg, D.; Jendeberg, L. EMBO J. 2003, 22, 4625.
22. Hoerer, S.; Schmid, A.; Heckel, A.; Budzinski, R.-M.; Nar, H. J. Mol. Biol. 2003,
334, 853.
23. The stabilization of His421 in LXR
a and His435 in LXRb are important
electrostatic interactions to hold helix 12 (a.k.a. AF2 helix) into a hydrophobic
groove on the surface of the LBD near helices 3–5 and elicit agonist function of
the respective receptors. For additional discussion of this histidine interaction,
see Refs. 21,22 and Williams, S.; Bledsoe, R. K.; Collins, J. L.; Boggs, S.; Lambert,
M. H.; Miller, A. B.; Moore, J.; McKee, D. D.; Moore, L.; Nichols, J.; Parks, D.;
Watson, M.; Wisely, B.; Willson, T. M. J. Biol. Chem. 2003, 278, 27138.
24. Xue, Y.; Chao, E.; Zuercher, W.; Willson, T. M.; Collins, J. L.; Redinbo, M. R.
Bioorg. Med. Chem. 2007, 15, 2156.
25. Salonen, L. M.; Ellermann, M.; Diederich, F. Angew. Chem., Int. Ed. 2011, 50, 2.
26. Both LXR receptor subtypes presented an aromatic amino acid residue (LXR
Phe326 and LXRb: Phe340) in the same region as the Phe367 found in RORc.
The presence of Val355 in RORc may also allow access to the stacking
interaction of 1 with Phe367; whereas, this region of the ligand binding pocket
was partially occluded in the co-structures of 1 with LXR and LXRb due to the
presence of a larger isoleucine residue common to both LXR receptor subtypes
(LXR : Ile313 and LXRb: Ile327).
a:
p–p
a
a
27. See the Supplementary data for the biochemical and cellular assay details.
28. The partition coefficient (logP) was calculated with internal software using the
VolSurf approach. For additional details, see Cruciani, G.; Crivori, P.; Carrupt, P.-
A.; Testa, B. J. Mol. Struc.-Theochem. 2000, 503, 17.
Acknowledgments
29. Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug Disc. 2007, 6, 881.
30. Waring, M. J. Expert Opin. Drug Discov. 2010, 5, 235.
31. Hansch, C.; Fujita, T. J. Am. Chem. Soc. 1964, 86, 1616.
32. See Ref. 29 and Leach, A. R.; Hann, M. M.; Burrows, J. N.; Griffen, E. J. Mol.
BioSyst. 2006, 2, 429.
We thank Drs. Krista Bowman and Jiansheng Wu and their
respective Genentech research groups for performing all required
protein expression and purification activities. We also thank Dr.
Rene Devos for NR vector expression and Dr. David Chantry for
many helpful discussions regarding RORc. In addition, we thank
Crystallographic Consulting, LLC for diffraction data collection.
Use of the Advanced Photon Source was supported by the US
Department of Energy, Office of Science, Office of Basic Energy Sci-
ences, under Contract No. DE-AC02-06CH11357.
33. Hammett, L. P. J. Am. Chem. Soc. 1937, 59, 96.
34.
A broader range of para-substituents for the aryl-sulfonamide was not
considered due to excessively high clog P values (>5) for the final
compounds combined with limited reagent availability.
35. Anslyn, E. V.; Dougherty, D. A. Advanced Concepts in Electronic Structure Theory.
Modern Physical Organic Chemistry, 1st ed.; University Science Books: USA,
2006. pp 850–853.
36. Bissantz, C.; Kuhn, B.; Stahl, M. J. Med. Chem. 2010, 53, 5061.
37. Brameld, K. A.; Kuhn, B.; Reuter, D. C.; Stahl, M. J. Chem. Inf. Model. 2008, 48, 1.
38. Wang, Y.; Kumar, N.; Nuhant, P.; Cameron, M. D.; Istrate, M. A.; Roush, W. R.;
Griffin, P. R.; Burris, T. P. ACS Chem. Biol. 2010, 5, 1029.
Supplementary data
39. Solt, L. A.; Kumar, N.; Nuhant, P.; Wang, Y.; Lauer, J. L.; Liu, J.; Istrate, M. A.;
Kamenecka, T. M.; Roush, W. R.; Vidovic´, D.; Schürer, S. C.; Xu, J.; Wagoner, G.;
Drew, P. D.; Griffin, P. R.; Burris, T. P. Nature 2011, 472, 491.
40. Wilson, N. J.; Boniface, K.; Chan, J. R.; McKenzie, B. S.; Blumenschein, W. M.;
Mattson, J. D.; Basham, B.; Smith, K.; Chen, T.; Morel, F.; Lecron, J.-C.; Kastelein,
R. A.; Cua, D. J.; McClanahan, T. K.; Bowman, E. P.; De Waal Malefyt, R. Nat.
Immunol. 2007, 8, 950.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2013.10.
054.
References and notes
41. Li, L.; Liu, J.; Zhu, L.; Cutler, S.; Hasegawa, H.; Shan, B.; Medina, J. C. Bioorg. Med.
Chem. Lett. 2006, 16, 1638.
42. Supplemental 19F NMR experiments and/or 1H–13C HMBC NMR experiments
were routinely employed to insure the proper N-alkyl connectivity of the final
compounds as drawn. For additional discussion, see Fauber, B. P.; Johnson, A.
R.; René, O.; Bowerman, S.; Burton, B.; Colebrook, A.; Flynn, A.; Harrold, G.;
Huhn, S.; Jones, G.; Lockey, P.; Norman, M.; Wong, H. J. Labelled Comp.
Radiopharm. in press.
1. Ivanov, I. I.; McKenzie, B. S.; Zhou, L.; Tadokoro, C. E.; Lepelley, A.; Lafaille, J. J.;
Cua, D. J.; Littman, D. R. Cell 2006, 126, 1121.
2. Ouyang, W.; Kolls, J. K.; Zheng, Y. Immunity 2008, 28, 454.
3. Lubberts, E.; Koenders, M. I.; Oppers-Walgreen, B.; van den Bersselaar, L.;
Coenen-de Roo, C. J.; Joosten, L. A. Arthritis Rheum. 2004, 50, 650.
4. Nakae, S.; Nambu, A.; Sudo, K.; Iwakura, Y. J. Immunol. 2003, 171, 6173.
5. Patel, D. D.; Lee, D. M.; Kolbinger, F.; Antoni, C. Ann. Rheum. Dis. 2013, 72, 116.
6. Rutz, S.; Eidenschenk, C.; Ouyang, W. Immunol. Rev. 2013, 252, 116.
7. Codarri, L.; Gyülvészi, G.; Tosevski, V.; Hesske, L.; Fontana, A.; Magnenat, L.;
Suter, T.; Becher, B. Nat. Immunol. 2011, 12, 560.
43. Katritzky, A. R.; Pilarski, B.; Urogdi, L.; Katritzky, A. R.; Pilarski, B.; Urogdi, L.
Synthesis 1989, 12, 949.
44. Brameld, K. A.; Carter, D. S.; Chin, E.; Fidalgo, J.-V.; Li, J.; Schoenfeld, R. C.;
Sjogren, E. B.; Talamas, F. X. US 2010/0021423.