Discovery of a New Class of Liver Receptor Homolog-1 (LRH-1) Antagonists: Virtual Screening, Synthesis and Biological Evaluation

Full text of DOI:10.1002/cmdc.201200307

MED
DOI: 10.1002/cmdc.201200307
Discovery of a New Class of Liver Receptor Homolog-1 (LRH-1) Antagonists:
Virtual Screening, Synthesis and Biological Evaluation
Jullien Rey,[a] Haipeng Hu,[b] Fiona Kyle,[c] Chun-Fui Lai,[c] Laki Buluwela,[c] R. Charles Coombes,[c]
Eric A. Ortlund,[d] Simak Ali,*[c] James P. Snyder,*[b] and Anthony G. M. Barrett*[a]
After lung cancer, breast cancer in women is the most
common cancer in the world with more than a million new
cases diagnosed every year.[1] Estrogens, acting through the es-
trogen receptor, are important drivers of breast cancer growth.
Much progress has been made in the fight against breast
cancer with the development of antiestrogen agents, such as
tamoxifen (1) and raloxifene (2), capable of competing with
the agonistic action of estradiol (3) at the estrogen receptor a
(ERa).[2] Estrogen binding to the ligand binding domain (LBD)
of ERa generates a receptor conformation that reveals a co-ac-
tivator recruitment groove involving a-helices 3, 4, 5 and 12,
allowing co-activator recruitment to agonist-bound ER and
consequent stimulation of gene expression. Through helix 12
(H12) displacement, antiestrogen agents can prevent the for-
mation of the ERa LBD co-activator binding surface, thereby
preventing co-activator recruitment and subsequent stimula-
tion of gene expression.[3] However, resistance to these thera-
pies is common, and so the identification of new molecular
targets for breast cancer treatment is an important goal.
Co-activator recruitment to the LBD of NRs is mediated by
a-helical motifs in co-activator proteins conforming to the con-
sensus sequence LXXLL, where L is leucine and X is any
amino acid.[8,9] As is the case for ERa, agonist binding to other
NRs reveals a co-activator recruitment groove that can accom-
modate the a-helical LXXLL region.[10] This well-studied molec-
ular switch, with conformational changes in the LBD allowing
co-activator recruitment and antagonist binding blocking co-
activator recruitment, provides a powerful screening strategy
for identifying NR agonists and antagonists. Using a high-
throughput screening (HTS) strategy based on interaction be-
tween the LRH-1 LBD and a LXXLL motif in the NR co-activa-
tor, transcriptional intermediary factor 2 (TIF2) in a fluorescence
resonance energy transfer (FRET)-based assay, Whitby et al. dis-
covered a series of substituted cis-bicyclo[3.3.0]oct-2-enes
acting as LRH-1 agonists.[11] However, to the best of our knowl-
edge, no antagonist has yet been reported in the literature,
making the development of such molecules an attractive and
challenging task.
Liver receptor homolog-1 (LRH-1) is a member of the steroi-
dogenic factor subfamily of nuclear receptors (NR) and plays
a prominent role in both adult and developmental biology.[4] It
has also been implicated in the control of aromatase expres-
sion in breast tumor-associated stroma, stimulating local estro-
gen biosynthesis.[5] Recently, LRH-1 was identified as a key reg-
ulator of ER expression in breast cancer cells[6] and was shown
to be associated with invasive breast cancer and estrogen-
dependent cell proliferation.[7] Together, these findings suggest
that the development of a small molecule capable of antago-
nizing LRH-1 could provide a powerful new strategy for inhibit-
ing estrogen signaling for the treatment of breast cancer.
[a] Dr. J. Rey, Prof. A. G. M. Barrett
Department of Chemistry, Imperial College London
London, SW7 2AZ (England)
While LRH-1 is known to bind phospholipids, recent structur-
al studies of LRH-1 suggest that these phospholipids are ex-
changeable with exogenous compounds and that a pool of
apoLRH-1 exists in cells.[12,13] Given the availability of large
chemical libraries, we employed virtual HTS and molecular
modeling strategy to identify potential LRH-1 antagonists.
Such compounds were assumed to bind within the LRH-1 LBD,
displacing the mobile a-helix H12 by means of a-helix–small
molecule steric clashes, and thereby, generate an inactive con-
formation. Using raloxifene (2) as a search template, a two-
dimensional virtual HTS of the ChemNavigator library,[14] which
contains 50 million-plus compounds, delivered 974 hits. Subse-
quent docking-based three-dimensional screening of the latter
dataset with the Glide/molecular mechanics generalized Born
E-mail: agm.barrett@imperial.ac.uk
[b] Dr. H. Hu, Prof. J. P. Snyder
Department of Chemistry and Emory Institute for Drug Development
Emory University
1515 Dickey Drive, Atlanta, GA 30322 (USA)
E-mail: jsnyder@emory.edu
[c] Dr. F. Kyle, C.-F. Lai, Dr. L. Buluwela, Prof. R. C. Coombes, Prof. S. Ali
Department of Surgery and Cancer, Imperial College London
London, W12 0NN (England)
E-mail: simak.ali@imperial.ac.uk
[d] Prof. E. A. Ortlund
Department of Biochemistry, Emory University, School of Medicine
1510 Clifton Rd NE, Atlanta, GA 30322 (USA)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201200307.
ChemMedChem 0000, 00, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

MED
surface area (MMGBSA) algorithms[15,16] and the hLRH-1 LBD
crystal structure housing phosphatidyl glycerol (PG) (PDB:
1YOK;[17] PG removed for docking) led to identification of the
benzothiophene analogue 4. Given the flexibility of nuclear re-
ceptors,[18,19] it would have been ideal to work with LRH-
1 bound by an antagonist. However, in the absence of such
a structure, 1YOK was selected since PG is a large, weak ago-
nist whose binding pocket is expanded by comparison with
a smaller potent agonist; the pocket created by PG can be
considered as more suitable for antagonist binding (see Fig-
ure S1 in the Supporting Information).
hand, extends deeply into the space occupied by the side
chains of residues Asn530 and Leu532 of H12, causing very se-
rious steric clashes (Figure 2). This modeled juxtaposition was
precisely that sought to displace H12.
Figure 2. Interpenetration of the ethoxypiperidine moiety of 4 (space-filled
structure) into helix 12 (H12). Structure 4 is docked into LRH-1 (PDB:
1YOK[17]) without H12 and overlayed with the same receptor retaining H12
in place (red). Leu532 (red sticks) pierces the C-7 CH2O-phenyl segment of
4, while Asn530 passes through the center of the distal piperidine ring.
Docked into LRH-1, structure 4 adopted the desired antago-
nistic pose in which the C-7 side chain promotes virtual H12
displacement (Figure 1a). The benzothiophene core of 4 coin-
cides with the binding pose of the GSK8470 agonist in the X-
ray-determined complex (PDB: 3PLZ;[11b] Figure 1b). The antag-
onist is anchored in the pocket primarily by hydrophobic inter-
actions between the core and the C-3 and C-7 side chains. The
C-3 side chain is surrounded by H6 and H7, and the hairpin
linking H5 and H6. The C-7 piperidine side chain, on the other
In order to validate the docking predictions, the synthesis
and biological evaluation of benzothiophene 4 were pursued.
With no precedence in the literature for the selective introduc-
tion of benzoyl moieties at the C-7 position of a benzothio-
phene ring, a novel and efficient synthetic approach needed to
be developed. Benzothiophene 9 was obtained in two steps
by nucleophilic substitution of bromoketone 6 with thiophenol
5 and subsequent cyclization in neat polyphosphoric acid at
908C (Scheme 1).[20] Under these conditions, a mixture of iso-
mers was formed as the cyclization can occur either at the
ortho- or para-position of the methoxy group (positions 2 and
6 from 7). Pleasingly, the compounds could be separated by
selective crystallization from refluxing acetone, giving the less
soluble and major benzothiophene isomer 9. Mechanistically,
the formation of benzothiophene 9 goes via the formation of
C-3-substituted intermediate 8 through acid-catalyzed hydroal-
kylation by the protonated ketone 7 followed by aromatiza-
tion. Protonation of intermediate 8 followed by a 3,2-shift of
the methoxyphenol ring and aromatization led to desired
product 9.[21]
A double Friedel–Crafts acylation was examined in an at-
tempt to functionalize both the C-3 and C-7 positions of the
benzothiophene ring (Scheme 2). Aluminum-chloride-mediated
double acylation using an excess of 4-fluorobenzoyl chloride
was accompanied by partial selective demethylation and gave
benzophenone 10. Selective mono-deprotection of related aryl
methyl ethers is known to proceed via the formation of a six-
membered ring chelate between the aluminum, the ortho-
ketone carbonyl, and the phenolic oxygen atom leading to, in
this case, phenol 10.[22] This was further demethylated using
boron tribromide to produce diphenol 11. Finally, the piperi-
dine side chains were introduced by nucleophilic aromatic sub-
stitutions generating benzothiophene 4.[23]
Figure 1. Small molecules in the ligand binding domain (LBD) of LRH-1.
a) Molecular modeling (docking pose) of benzothiophene antagonist 4 in
LRH-1. Helix 12 (H12) in the phosphatidyl glycerol-bound protein is high-
lighted in red (PDB: 1YOK[17]). Also shown is the hairpin loop (HL) linking H5
and H6. The residues Asn 530 and Leu 532 of H12 are highlighted as red
sticks. b) Whitby agonist GSK8470/LRH-1 X-ray structure (PDB: 3PLZ[11b]).
&2
&
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 6
ÝÝ
These are not the final page numbers!

MED
PGC1a peptide. Gratifyingly, compound 4 exhibits
LRH-1 antagonism in the low micromolar range
(IC50 =3.1 mm; Table 1), which is ideal for potency im-
provement by molecular manipulation. For example,
the LRH-1 agonist GSK8470 has reported EC50 values
of 0.43–0.63 mm, while chemical modification of the
structure achieved derivatives with potencies as low
as 10–100 nm.[11]
An initial attempt to optimize 4 was motivated by
a desire to decrease the molecular weight (4, MW=
704.9 amu) and improve the analogue physical prop-
erties within the context of the present synthetic
scheme. The concept eliminates the C-7 ethoxypiperi-
dine in 4, but compensates its predicted steric effect
by replacing the C-7 aroyl meta-hydrogen atoms with
methyl groups. Figure 2 illus-
Scheme 1. Synthesis of benzothiophene intermediate 9. Reagents and conditions: a) KOH,
EtOH, H2O, RT, 18 h, 81%; b) polyphosphoric acid, 908C, 1 h, 63%.
trates an apparent severe steric
encounter between the H12
Leu532 residue and the meta-
centers of the aromatic ring.
Docking structures 19 and 20
carrying one and two meta-
methyl groups, respectively, into
the LRH-1 LBD pocket (H12 re-
moved) revealed predicted bind-
ing poses similar to 4. A pro-
spective clash of the alkylated
aromatic rings with Leu532, con-
sistent with antagonism, is evi-
dent (see Figure S2 in the Sup-
porting Information). According-
ly, methylated analogues 19 and
20 (MW=591.7 and 605.7 amu,
respectively) were chosen for
synthesis and bioassay.
Both substituents at C-3 and
Scheme 2. Synthesis of benzothiophene 4. Reagents and conditions: a) 4-fluorobenzoyl chloride, AlCl3, CH2Cl2, RT,
18 h, 70%; b) BBr3, CH2Cl2, 08C, 3 h, 79%; c) 1-(2-hydroxyethyl)piperidine, NaH, DMF, 508C, 5 h, 52%.
C-7 positions were introduced
To assay for inhibition of co-activator binding, benzothio-
Table 1. Biological and computational evaluation of benzothiophene an-
phene 4 was tested in an in vitro assay using the bead-based
proximity AlphaScreen method that allows sensitive measure-
ment of protein–protein interactions by transfer of energy
from the excited donor bead to an acceptor bead in close
proximity,[24] allowing screening for compounds that promote
or inhibit protein–protein interactions. To determine inhibition
of LRH-1/co-activator interaction, an assay was developed to
measure co-activator recruitment by the LBD of LRH-1, using
a biotinylated peptide derived from human PGC1a, with the
sequence 138-AEEPSLLKKLLLAPANT, and his-tagged, purified
recombinant human LRH-1 LBD (amino acids 291–541).
alogues conceived by virtual database screening and molecular model-
ing.
Entry
Compd
IC50[a] [mm]
DG [Kcalmolꢀ1
]
1
2
3
4
19
20
3.1 (2.6–3.7)
5.8 (4.3–7.7)
8.8 (5.1–15.0)
ꢀ136.7
ꢀ132.2
ꢀ130.9
[a] Data are the mean values of three experiments. The corresponding
95% confidence range is given in parentheses. [b] The MMGBSA algo-
rithm was used to predict the binding affinities of these compounds for
LRH-1.
Streptavidin-coated donor beads were incubated with the
biotinylated co-activator peptide, together with his-tagged
LRH-1 and nickel-nitrilotriacetic acid (Ni-NTA) acceptor beads.
Since recombinant LRH-1 LBD takes up the activated confor-
mation due to bound phospholipids,[17,25] potential antagonists
can readily be identified by a decrease in interaction with the
by sequential Friedel–Crafts acylation reactions. Reaction of
benzothiophene 9 with one equivalent of 4-fluorobenzoyl
chloride and aluminum chloride gave ketone 12 in a selective
manner without formation of bis-acylated intermediate 10
ChemMedChem 0000, 00, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
&
3
&
ÞÞ
These are not the final page numbers!

MED
space, while simultaneously
maintaining suitable drug-like
physical properties.
A
mammalian two-hybrid
assay was used to confirm the in
vitro findings. For this, COS-
1 cells were co-transfected with
GAL4(DBD)-PGC1a DNA binding
domain and VP16-LRH-1, togeth-
er with a GAL4-responsive luci-
ferase reporter gene. As expect-
ed, GSK8470 stimulated the in-
teraction of LRH-1 with PGC1a
(Figure 3). Raloxifene (2) weakly
inhibited the interaction at con-
centrations up to 10 mm. Inhibi-
tion by compounds 4, 19 and 20
was again observed, with the
degree of inhibition at 10 mm
being in the order 4>19>20,
as observed in the AlphaScreen
assay. Investigation of the action
Scheme 3. Synthesis of benzothiophenes 17 and 18 by sequential Friedel–Crafts acylations. Reagents and condi-
tions: a) 4-fluorobenzoyl chloride, AlCl3, CH2Cl2, 18 h, 69%; b) AlCl3, CH2Cl2, RT, 18 h; c) BBr3, CH2Cl2, 08C, 3 h, for
17: 42%, for 18: 46% (two steps).
(Scheme 3).[23] Subsequent acylation of benzothiophene 12
with 3-methylbenzoyl chloride (13) gave 15, which was directly
demethylated using boron tribromide to yield phenol 17.
Finally, nucleophilic aromatic substitution of fluoride 17 by
1-(2-hydroxyethyl)piperidine gave the desired benzothiophene
19 in moderate yield (Scheme 4). A similar strategy was used
for the synthesis of benzothiophene 20 starting from inter-
mediate 12 and 3,5-dimethylbenzoyl chloride (14) (Schemes 3
and 4).
of these compounds towards ERa showed that they do indeed
inhibit ERa in reporter gene assays, albeit considerably less po-
tently than raloxifene (Figure S3 in the Supporting Informa-
tion). GSK8470 had little or no effect on ERa activity.
In summary, three novel benzothiophene derivatives identi-
fied as modest LRH-1 antagonists using Glide/MMGBSA dock-
ing methodology were synthesized and biologically assayed.
The identification of these analogues, derived by weak similari-
ty comparisons to the selective estrogen receptor modulator
(SERM) raloxifene (2), arose from high-throughput virtual
screening performed on a library of millions of compounds.
LRH-1 docking analysis suggests that the C-7 moiety of the
benzothiophene scaffold is directed toward H12 and can cause
a significant steric clash between the ligand terminal atoms
and the H12 a-helix. Efforts to further chemically modify the
C-7 side chain to develop novel benzothiophenes with im-
proved potency are underway.
Benzothiophenes 19 and 20 were then tested in the in vitro
AlphaScreen assay to show a two to threefold decrease in an-
tagonistic activity relative to 4 (Table 1). This result was exam-
ined by evaluating the MMGBSA free energies of binding[16] for
the three compounds. The order of binding is estimated to be
4>19>20 precisely in parallel with the in vitro results
(Table 1). Consistently, the geometries of the docked structures
suggest productive ligand contacts, and the degree of pertur-
bation of Leu532 on H12 by the ligands falls in the same order
(see Figure S2 in the Supporting Information). It would appear
that to achieve more effective antagonists, the longer C-7 side
chain will need to be retained so as to penetrate well into H12
Experimental Section
Chemistry
All chemicals were used as re-
ceived or purified using standard
procedures. Solvents were dried by
standard techniques and distilled
under N2 before use. All experi-
ments were carried out in oven-
dried glassware under an inert at-
mosphere of N2 or Ar. Analytical
thin layer chromatography (TLC)
was performed using pre-coated
aluminum- or glass-backed plates
(Merck silica gel 60
F254), and
plates were visualized by ultravio-
let light and/or treatment with
Scheme 4. Synthesis of benzothiophenes 19 and 20. Reagents and conditions: a) 1-(2-hydroxyethyl)piperidine,
NaH, DMF, 508C, 5 h, for 19: 37%, for 20: 35%.
&4
&
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 6
ÝÝ
These are not the final page numbers!

MED
2.40–2.37 (m, 4H), 1.52–1.44 (m, 8H), 1.40–1.34 ppm (m, 4H);
13C NMR (100 MHz, [D6]DMSO): d=193.3, 192.3, 162.8, 162.5, 157.8,
141.9, 138.4, 132.7, 131.8, 131.7, 129.9, 129.7, 129.5, 129.3, 125.8,
123.4, 118.4, 116.1, 115.6, 114.4, 114.2, 65.9 (2C), 57.1, 57.0, 54.3,
54.2, 25.5, 25.4, 23.8 (2C) ppm (one quaternary carbon obscured);
IR (neat): n˜ =1597, 1356, 1236, 1155, 900, 830 cmꢀ1; HRMS-ESI: m/z
[M+H]+ calcd for C42H45N2O6S: 705.2998, found: 705.3010.
7-[(3-Methylphenyl)carbonyl]-2-(4-hydroxyphenyl)-3-({4-[2-(pi-
peridin-1-yl) ethoxy]phenyl}carbonyl)-1-benzothiophen-6-ol (19):
1-(2-Hydroxyethyl)piperidine (33 mL, 0.25 mmol) was added with
stirring to NaH (25 mg, 0.62 mmol, 60% in mineral oil) in DMF
(1.4 mL) at RT. After 5 min, benzothiophene 17 (60 mg, 0.12 mmol)
in DMF (0.20 mL) was added in one portion, and the mixture was
stirred at 508C for 5 h. The reaction was quenched with H2O, and
the aqueous phase was extracted with EtOAc (ꢁ3). The combined
organic extracts were washed with brine, dried (MgSO4), filtered
and concentrated in vacuo. Purification by flash chromatography
(CH2Cl2/MeOH: 99:1!19:1+0.25% NH3·H2O) gave amine 19 as an
amorphous yellow solid (27 mg, 37%): 1H NMR (400 MHz,
[D6]DMSO): d=10.4 (br s, 1H), 9.74 (br s, 1H), 7.68 (d, J=9.3 Hz,
2H), 7.59 (s, 1H), 7.54 (d, J=7.8 Hz, 1H), 7.51 (d, J=8.8 Hz, 1H),
7.45 (d, J=7.3 Hz, 1H), 7.40 (dd, J=7.8, 7.3 Hz, 1H), 7.17 (d, J=
8.8 Hz, 2H), 7.05 (d, J=8.8 Hz, 1H), 6.93 (d, J=9.3 Hz, 2H), 6.66 (d,
J=8.8 Hz, 2H), 4.08 (t, J=5.4 Hz, 2H), 2.62 (t, J=5.4 Hz, 2H), 2.39–
2.36 (m, 7H), 1.48–1.45 (m, 4H), 1.37–1.35 ppm (m, 2H); 13C NMR
(125 MHz, [D6]DMSO): d=195.2, 192.4, 162.8, 157.9, 155.0, 142.3,
138.8, 138.2, 137.7, 133.4, 132.9, 131.9, 129.8, 129.6, 129.3, 129.1,
128.3, 126.7, 126.3, 123.5, 117.9, 116.3, 115.7, 114.5, 65.9, 57.1, 54.3,
25.5, 23.8, 20.8 ppm; IR (neat): n˜ =1653, 1594, 1503, 1429, 1387,
Figure 3. Mammalian two-hybrid analysis of compound activities in COS-
1 cells: 0.01 mm: &; 0.1 mm: &; 1.0 mm: &; 10 mm: &. COS-1 cells were
transfected with GAL4(DBD)-PGC1a L2 (encoding the second LXXLL motif
in PGC1a), VP16-LRH-1(LBD), together with a GAL4-responsive Firefly lucifer-
ase gene and pRL-CMV, which encodes Renilla luciferase and acts as a control
for transfection efficiency. Firefly activities are shown relative to Renilla luci-
ferase activities, the activity for vehicle (DMSO) being taken as 1 and all
other activities calculated relative to this. The mean activities (Log10) of
three replicates are shown; error bars represent the standard error of the
mean (SEM).
KMnO4 or vanillin stains followed by heating as deemed appropri-
ate. Flash column chromatography was carried out using silica
(Merck 9385 Kieselgel 60; 230–400 mesh) under a positive pressure
of N2. 1H and 13C NMR spectra were recorded at 400 or 500 MHz
and 100 or 125 MHz, respectively. Chemical shifts (d) are quoted in
parts per million (ppm) and are referenced to a residual solvent
peak. HSQC was used to confirm peak integration in 13C NMR spec-
tra. Mass spectra were recorded using a Micromass Platform II and
Micromass AutoSpec-Q spectrometer. The purity of the final com-
pounds tested in vitro was assessed by LC/MS and HRMS. Infrared
(IR) spectra were recorded using the attenuated total reflectance
(ATR) technique, monitoring from 4000–700 cmꢀ1. Melting points
(mp) were determined using a hot-stage microscope and are un-
corrected.
1287, 1248, 1164, 1030, 939, 841, 761 cmꢀ1
; HRMS-ESI: m/z
[M+H]+ calcd for C36H34NO5S: 592.2158, found: 592.2156.
7-[(3,5-Dimethylphenyl)carbonyl]-2-(4-hydroxyphenyl)-3-({4-[2-
(piperidin-1-yl) ethoxy]phenyl}carbonyl)-1-benzothiophen-6-ol
(20): 1-(2-Hydroxyethyl)piperidine (32 mL, 0.24 mmol) was added
with stirring to NaH (25 mg, 0.61 mmol, 60% in mineral oil) in DMF
(1.4 mL) at RT. After 5 min, benzothiophene 18 (60 mg, 0.12 mmol)
in DMF (0.20 mL) was added in one portion, and the mixture was
stirred at 508C for 5 h. The reaction was quenched with H2O, and
the aqueous phase was extracted with EtOAc (ꢁ3). The combined
organic extracts were washed with brine, dried (MgSO4), filtered
and concentrated in vacuo. Purification by flash chromatography
(CH2Cl2/MeOH: 99:1!19:1+0.25% NH3·H2O) gave amine 20 as an
amorphous yellow solid (25 mg, 35%): 1H NMR (400 MHz,
[D6]DMSO): d=10.4 (br s, 1H), 9.75 (br s, 1H), 7.68 (d, J=8.8 Hz,
2H), 7.50 (d, J=8.8 Hz, 1H), 7.37 (s, 2H), 7.27 (s, 1H), 7.17 (d, J=
8.3 Hz, 2H), 7.04 (d, J=8.8 Hz, 1H), 6.92 (d, J=8.8 Hz, 2H), 6.66 (d,
J=8.3 Hz, 2H), 4.08 (app-br t, 2H), 2.62 (app-br t, 2H), 2.39 (app-br
s, 4H), 2.32 (s, 6H), 1.46 (app-br s, 4H), 1.36 ppm (app-br s, 2H);
13C NMR (125 MHz, [D6]DMSO): d=195.4, 192.4, 162.8, 157.9, 154.8,
142.2, 138.7, 138.3, 137.5, 134.2, 132.8, 131.9, 129.8, 129.6, 129.3,
126.5 (3C), 123.5, 118.1, 116.3, 115.7, 114.5, 65.9, 57.1, 54.3, 25.5,
23.8, 20.7 ppm; IR (neat): n˜ =1655, 1595, 1504, 1429, 1384, 1361,
1244, 1163, 1140, 1035, 939, 842, 763 cmꢀ1; HRMS-ESI: m/z [M+H]+
calcd for C37H36NO5S: 606.2314, found: 606.2287.
Experimental protocols and characterization data for all intermedi-
ates are given in the Supporting Information.
2-(4-Hydroxyphenyl)-3,7-bis-({4-[2-(piperidin-1-yl)ethoxy]phenyl}-
carbonyl)-1-benzothiophen-6-ol (4): 1-(2-Hydroxyethyl)piperidine
(66 mL, 0.49 mmol) was added with stirring to NaH (34 mg,
0.86 mmol, 60% in mineral oil) in DMF (1.4 mL) at RT. After 5 min,
benzothiophene 11 (60 mg, 0.12 mmol) in DMF (0.20 mL) was
added in one portion, and the mixture was stirred at 508C for 5 h.
The reaction was quenched with H2O, and the aqueous phase was
extracted with EtOAc (ꢁ3). The combined organic extracts were
washed with brine, dried (MgSO4), filtered and concentrated in
vacuo. Purification by flash chromatography (CH2Cl2/MeOH: 49:1!
24:1+0.25% NH3·H2O) gave diamine 4 as a yellow solid (45 mg,
Molecular modeling
1
52%): mp: 129–1328C (MeOH); H NMR (400 MHz, [D6]DMSO): d=
10.3 (br s, 1H), 9.74 (br s, 1H), 7.76 (d, J=8.8 Hz, 2H), 7.68 (d, J=
8.8 Hz, 2H), 7.46 (d, J=8.8 Hz, 1H), 7.15 (d, J=8.8 Hz, 2H), 7.05 (d,
J=8.8 Hz, 2H), 6.92 (d, J=8.8 Hz, 2H), 7.04 (d, J=8.8 Hz, 1H), 6.65
(d, J=8.8 Hz, 2H), 4.16 (t, J=5.9 Hz, 2H), 4.08 (t, J=5.9 Hz, 2H),
2.67 (t, J=5.9 Hz, 2H), 2.61 (t, J=5.9 Hz, 2H), 2.44–2.41 (m, 4H),
For the preliminary virtual high-throughput screen, the ERa antag-
onist raloxifene was employed as a search template for mining of
the ChemNavigator database. With compound similarity thresholds
set at 55%, 974 structures were returned. The corresponding
three-dimensional structures were generated by using LigPrep 2.3,
ChemMedChem 0000, 00, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
&
5
&
ÞÞ
These are not the final page numbers!

MED
[3] A. M. Brzozowski, A. C. W. Pike, Z. Dauter, R. E. Hubbard, T. Bonn, O. Eng-
strçm, L. ꢂhman, G. L. Greene, J.-ꢃ. Gustafsson, M. Carlquist, Nature
1997, 389, 753.
[4] E. Fayard, J. Auwerx, K. Schoonjans, Trends Cell Biol. 2004, 14, 250.
[5] C. D. Clyne, C. J. Speed, J. Zhou, E. R. Simpson, J. Biol. Chem. 2002, 277,
20591.
[6] P. T. R. Thiruchelvam, C.-F. Lai, H. Hua, R. S. Thomas, A. Hurtado, W.
Hudson, A. R. Bayly, F. J. Kyle, M. Periyasamy, A. Photiou, A. C. Spivey,
E. A. Ortlund, R. J. Whitby, J. S. Carroll, C. R. Coombes, L. Buluwela, S. Ali,
Breast Cancer Res. Treat. 2011, 127, 385.
[7] A. L. Chand, K. A. Herridge, E. W. Thompson, C. D. Clyne, Endocr.-Relat.
Cancer 2010, 17, 965.
and conformational searches were performed with MacroModel
9.7.[26] The ten conformers of each structure with the lowest calcu-
lated energy were selected as starting structures for docking. The
LRH-1 LBD crystal structure (PDB: 1YOK[17]) and the Glide docking
procedure (Glide 5.5[15]) in Maestro 9.0 were used to perform the
docking studies.[26] In order to mimic the inactive state of LRH-1,
residues on H12 and the adjacent loop between H11 and H12 (resi-
dues 522–538) in the LRH-1 LBD were removed from the protein
structure during the docking investigations. In the final step,
a Prime 2.1[26] MMGBSA rescoring[16] was performed to predict the
binding affinities of the test compounds for LRH-1.
[8] D. M. Heery, E. Kalkhoven, S. Hoare, M. G. Parker, Nature 1997, 387, 733.
[9] J. Torchia, D. W. Rose, J. Inostroza, Y. Kamei, S. Westin, C. K. Glass, M. G.
Rosenfeld, Nature 1997, 387, 677.
Biology
[10] K. W. Nettles, G. L. Greene, Annuv. Rev. Physiol. 2005, 67, 309.
[11] a) R. J. Whitby, S. Dixon, P. R. Maloney, P. Delerive, B. J. Goodwin, D. J.
Parks, T. M. Willson, J. Med. Chem. 2006, 49, 6652; b) R. J. Whitby, J. Stec,
R. D. Blind, S. Dixon, L. M. Leesnitzer, L. A. Orband-Miller, S. P. Williams,
T. M. Willson, R. Xu, W. J. Zuercher, F. Cai, H. A. Ingraham, J. Med. Chem.
2011, 54, 2266.
[12] P. M. Musille, M. C. Pathak, J. L. Lauer, W. H. Hudson, P. R. Griffin, E. A. Or-
tlund, Nat. Struct. Mol. Biol. 2012, 19, 532.
[13] J. M. Lee, Y. K. Lee, J. L. Mamrosh, S. A. Busby, P. R. Griffin, M. C. Pathak,
E. A. Ortlund, D. D. Moore, Nature 2011, 474, 506.
AlphaScreen (amplified luminescent proximity homogenous assay)
technology is a bead-based proximity assay consisting of a nickel-
chelate-coated acceptor bead bound to a His-tagged LRH-1 LBD
and a streptavidin-coated donor bead bound to a biotinylated co-
activator peptide (PGC1a). Following excitation with a high-intensi-
ty laser at 680 nm, a singlet ambient oxygen molecule forms on
the surface of the donor bead with the ability to diffuse up to
200 nm. If an acceptor bead is within this proximity, the oxygen
singlet reacts with a thioxene derivative on the acceptor bead,
generating chemiluminescence and further activating a cascade of
fluorophorescence emitting light at 520–620 nm, which can be
used as a measurement of proximity/binding.
[14] ChemNavigator, 10919 Technology Place, Suite B, San Diego, CA 92127,
USA; http://www.chemnavigator.com/.
[15] Glide (version 5.5), Schrçdinger, LLC, 120 W. 45th Street, New York, NY
10036-4041, USA. For product details, see: http://www.schrodinger.-
com/products/14/5/ (last accessed: August 22, 2012). For further read-
ing, see: R. A. Friesner, J. L. Banks, R. B. Murphy, T. A. Halgren, J. J. Klicic,
D. T. Mainz, M. P. Repasky, E. H. Knoll, D. E. Shaw, M. Shelley, J. K. Perry,
L. C. Sander, P. S. Shenkin, J. Med. Chem. 2004, 47, 1739.
[16] A. P. Graves, D. M. Shivakumar, S. E. Boyce, M. P. Jacobson, D. A. Case,
B. K. Shoichet, J. Mol. Biol. 2008, 377, 914.
[17] I. N. Krylova, E. P. Sablin, J. Moore, R. X. Xu, G. M. Waitt, J. A. MacKay, D.
Juzumiene, J. M. Bynum, K. Madauss, V. Montana, L. Lebedeva, M.
Suzawa, J. D. Williams, S. P. Williams, R. K. Guy, J. W. Thornton, R. J. Flet-
terick, T. M. Willson, H. A. Ingraham, Cell 2005, 120, 343.
[18] V. J. Hilser, E. E. Thompson, J. Biol. Chem. 2011, 286, 39675.
[19] M. Nocker, P. Cozzini, Curr. Topics Med. Chem. 2011, 11, 133.
[20] C. D. Jones, M. G. Jevnikar, A. J. Pike, M. K. Peters, L. J. Black, A. R.
Thompson, J. F. Falcone, J. A. Clemens, J. Med. Chem. 1984, 27, 1057.
[21] H. Galons, J.-F. Girardeau, C. C. Farnoux, M. Miocque, J. Heterocycl.
Chem. 1981, 18, 561.
[22] F. M. Dean, J. Goodchild, L. E. Houghton, J. A. Martin, R. B. Morton, B.
Parton, A. W. Price, N. Somvichien, Tetrahedron Lett. 1966, 7, 4153.
[23] C. R. Schmid, J. P. Sluka, K. M. Duke, Tetrahedron Lett. 1999, 40, 675.
[24] J. Peppard, F. Glickman, Y. He, S.-I. Hu, J. Doughty, R. Goldberg, J.
Biomol. Screening 2003, 8, 149.
This system was optimized using LRH-1 with both PGC1a and
SRC2 (EC50 =35 nm and 369 nm, respectively, against 50 nm LRH-1)
and robust assay “z” scores (z >0.6). The system was validated
using the known LRH-1 agonist GSK8470 with an EC50 value of
220 nm.[11a] The AlphaScreen system has been utilized previously to
look at LRH-1 LBD co-factor recruitment.[27] Using the interaction
between a set concentration of LRH-1 and PGC1a as a baseline,
a twelve-point titration range of the experimental compounds was
added to assess for inhibition of this interaction. The results were
analyzed using Prism software, correcting for the background
signal, and expressed as a percentage relative to a vehicle-treated
control (DMSO). IC50 values with 95% confidence intervals were
then calculated. Further assay details have been included in the
Supporting Information.
Acknowledgements
The authors are grateful to Cancer Research UK (CRUK) for finan-
cial support of this research. J.P.S. and H.H. thank Prof. Dennis
Liotta (Emory University) for support of the modeling work. The
authors also thank Prof. Donald McDonnell (Duke University) for
his gift of VP16-LRH-1 and Gal4-PGC1a constructs.
[25] E. A. Ortlund, Y. Lee, I. H. Solomon, J. M. Hager, R. Safi, Y. Choi, Z. Guan,
A. Tripathy, C. R. H. Raetz, D. P. McDonnell, D. D. Moore, M. R. Redinbo,
Nat. Struct. Mol. Biol. 2005, 12, 357.
[26] MacroModel (version 9.7), LigPrep (version 2.3), Maestro (version 9.0),
Prime (version 2.1), all from Schrçdinger, LLC, 120 W. 45th Street, New
York, NY 10036-4041, USA. For product details, see: http://www.schro-
dinger.com/productsguide/.
[27] Y. Li, M. Choi, K. Suino, A. Kovach, J. Daugherty, S. A. Kliewer, H. E. Xu,
Proc. Natl. Acad. Sci. USA 2005, 102, 9505.
Keywords: antagonists
· breast cancer · high-throughput
virtual screening · ligand binding domains · nuclear receptor
Received: June 20, 2012
[1] F. Kamangar, G. M. Dores, W. F. Anderson, J. Clin. Oncol. 2006, 24, 2137.
[2] S. Ali, L. Buluwela, C. R. Coombes, Annu. Rev. Med. 2011, 62, 217.
Published online on && &&, 0000
&6
&
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 6
ÝÝ
These are not the final page numbers!

COMMUNICATIONS
Targeting LRH-1: Virtual screening and
molecular modeling were used to iden-
tify novel antagonists of liver receptor
homolog-1 (LRH-1), an emerging thera-
peutic target for breast cancer. Hit com-
pounds were synthesized and biologi-
cally assayed, and the preliminary re-
sults suggest that raloxifene-based ana-
logues, substituted at the position C-7
of the benzothiophene ring, might gen-
erate an inactive protein conformation
through binding and thus antagonize
this nuclear receptor.
J. Rey, H. Hu, F. Kyle, C.-F. Lai, L. Buluwela,
R. C. Coombes, E. A. Ortlund, S. Ali,*
J. P. Snyder,* A. G. M. Barrett*
&& – &&
Discovery of a New Class of Liver
Receptor Homolog-1 (LRH-1)
Antagonists: Virtual Screening,
Synthesis and Biological Evaluation
ChemMedChem 0000, 00, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
&
7
&
ÞÞ
These are not the final page numbers!