Identification of a GPER/GPR30 antagonist with improved estrogen receptor counterselectivity

Article abstract of DOI:10.1016/j.jsbmb.2011.07.002

GPER/GPR30 is a seven-transmembrane G protein-coupled estrogen receptor that regulates many aspects of mammalian biology and physiology. We have previously described both a GPER-selective agonist G-1 and antagonist G15 based on a tetrahydro-3H-cyclopenta[c]quinoline scaffold. The antagonist lacks an ethanone moiety that likely forms important hydrogen bonds involved in receptor activation. Computational docking studies suggested that the lack of the ethanone substituent in G15 could minimize key steric conflicts, present in G-1, that limit binding within the ERα ligand binding pocket. In this report, we identify low-affinity cross-reactivity of the GPER antagonist G15 to the classical estrogen receptor ERα. To generate an antagonist with enhanced selectivity, we therefore synthesized an isosteric G-1 derivative, G36, containing an isopropyl moiety in place of the ethanone moiety. We demonstrate that G36 shows decreased binding and activation of ERα, while maintaining its antagonist profile towards GPER. G36 selectively inhibits estrogen-mediated activation of PI3K by GPER but not ERα. It also inhibits estrogen- and G-1-mediated calcium mobilization as well as ERK1/2 activation, with no effect on EGF-mediated ERK1/2 activation. Similar to G15, G36 inhibits estrogen- and G-1-stimulated proliferation of uterine epithelial cells in vivo. The identification of G36 as a GPER antagonist with improved ER counterselectivity represents a significant step towards the development of new highly selective therapeutics for cancer and other diseases.

Full text of DOI:10.1016/j.jsbmb.2011.07.002

Journal of Steroid Biochemistry & Molecular Biology 127 (2011) 358–366
Contents lists available at ScienceDirect
Journal of Steroid Biochemistry and Molecular Biology
journal homepage: www.elsevier.com/locate/jsbmb
Identification of a GPER/GPR30 antagonist with improved estrogen receptor
counterselectivity
Megan K. Dennisa, Angela S. Fielda, Ritwik Buraib, Chinnasamy Rameshb, Whitney K. Petriea,
Cristian G. Bologac, Tudor I. Opreac,d, Yuri Yamaguchif, Shin-Ichi Hayashig, Larry A. Sklard,e
,
Helen J. Hathawaya,d, Jeffrey B. Arterburnb,d, Eric R. Prossnitza,d,∗
a Department of Cell Biology & Physiology, University of New Mexico Health Sciences Center, Albuquerque, NM 87131, United States
b Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, NM 88003, United States
c Division of Biocomputing, Department of Biochemistry & Molecular Biology, University of New Mexico Health Sciences Center, Albuquerque, NM 87131, United States
d UNM Cancer Center, University of New Mexico Health Sciences Center, Albuquerque, NM 87131, United States
e Department of Pathology, University of New Mexico Health Sciences Center, Albuquerque, NM 87131, United States
f
Research Institute for Clinical Oncology, Saitama Cancer Center, Saitama, Japan
g Department of Molecular and Functional Dynamics, Tohoku University, Sendai, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
GPER/GPR30 is a seven-transmembrane G protein-coupled estrogen receptor that regulates many aspects
of mammalian biology and physiology. We have previously described both a GPER-selective agonist G-1
and antagonist G15 based on a tetrahydro-3H-cyclopenta[c]quinoline scaffold. The antagonist lacks an
ethanone moiety that likely forms important hydrogen bonds involved in receptor activation. Compu-
tational docking studies suggested that the lack of the ethanone substituent in G15 could minimize key
steric conflicts, present in G-1, that limit binding within the ER␣ ligand binding pocket. In this report, we
identify low-affinity cross-reactivity of the GPER antagonist G15 to the classical estrogen receptor ER␣. To
generate an antagonist with enhanced selectivity, we therefore synthesized an isosteric G-1 derivative,
G36, containing an isopropyl moiety in place of the ethanone moiety. We demonstrate that G36 shows
decreased binding and activation of ER␣, while maintaining its antagonist profile towards GPER. G36
selectively inhibits estrogen-mediated activation of PI3K by GPER but not ER␣. It also inhibits estrogen-
and G-1-mediated calcium mobilization as well as ERK1/2 activation, with no effect on EGF-mediated
ERK1/2 activation. Similar to G15, G36 inhibits estrogen- and G-1-stimulated proliferation of uterine
epithelial cells in vivo. The identification of G36 as a GPER antagonist with improved ER counterselectiv-
ity represents a significant step towards the development of new highly selective therapeutics for cancer
and other diseases.
Received 31 May 2011
Received in revised form 1 July 2011
Accepted 3 July 2011
Keywords:
Estrogen
GPR30
GPER
Antagonist
© 2011 Elsevier Ltd. All rights reserved.
1. Introduction
in the case of ER␣ and ER␤ [1], or indirectly, as in the case of GPER
[4,5]. ER␣, ER␤ and GPER can also mediate rapid signaling events
Estrogens mediate a range of physiological processes, including
roles in reproduction, the immune, nervous and cardiovascu-
lar systems [1]. Additionally, estrogen signaling plays a role in
breast, ovarian and other types of cancer [2]. Estrogen signaling
is mediated through at least three receptors, including the soluble
nuclear receptors ER␣ and ER␤, and GPER, a seven transmembrane-
spanning G protein-coupled receptor (GPCR) [3]. All three of these
receptors can mediate gene transcription events, either directly, as
through activation of MAPK, PI3K, Src kinase and related pathways
[2,3]. In vivo, these receptors vary in their tissue distribution and
the estrogen responsiveness of a given tissue is determined both
by receptor expression, co-regulator expression and by signaling
interplay between receptors in response to estrogen [1,3].
Estrogen receptors overlap in some physiological functions as
well as in their ligand specificity. The triphenylethylene derivative
tamoxifen is representative of the selective estrogen receptor mod-
ulator (SERM) anti-estrogen class of therapeutics that inhibit the
binding of the endogenous ligand, 17␤-estradiol (E2, 1), to ER␣ and
ER␤, while fulvestrant (faslodex, ICI182,780) represents the “pure”
antiestrogen class of therapeutics that initiates ER␣ and ER␤ recep-
tor down-regulation. Paradoxically, both of these compounds act
as GPER agonists [6,7], and the resulting agonist/antagonist prop-
∗
Corresponding author at: Department of Cell Biology & Physiology, University
of New Mexico Health Sciences Center, 915 Camino de Salud NE, Albuquerque, NM
87131, United States. Tel.: +1 505 272 5647; fax: +1 505 272 1421.
E-mail address: eprossnitz@salud.unm.edu (E.R. Prossnitz).
0960-0760/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jsbmb.2011.07.002

M.K. Dennis et al. / Journal of Steroid Biochemistry & Molecular Biology 127 (2011) 358–366
359
erties of these compounds vary by receptor status and/or tissue [8].
2.2. Chemical synthesis and characterization of G36
(4-(6-bromo-benzo[1,3]dioxol-5-yl)-8-isopropyl-3a,4,5,9b-
tetrahydro-3H-cyclopenta[c]quinoline)
Thus, compounds with selectivity towards a single receptor are of
great use for probing receptor function in complex systems where
multiple estrogen receptors are expressed. We have previously
developed a GPER-selective agonist, G-1 2 [9], and a GPER-selective
antagonist, G15 3 [10], which have been used to elucidate GPER
function in a variety of systems, both in vitro as well as in vivo.
For example, G-1 has been used to define a role for GPER in pro-
liferation [4,11], protein kinase C activation [12], PI3K activation
[9] and calcium mobilization in a range of cells [9,13,14]. Addition-
ally, G-1 has been used in in vivo systems to investigate the roles of
GPER in uterine epithelial proliferation [10], thymic atrophy [15],
immune regulation in experimental autoimmune encephalomyeli-
tis (a murine model of multiple sclerosis) [15,16] and vascular
regulation [17]. The antagonist G15 has now been used by sev-
eral groups to establish the role of GPER in estrogen-mediated
events, including vasodilation [18], zebrafish oocyte maturation
[19], neuroprotection [20] and inhibition of chondrogenesis [21].
These results indicate that selective ligands for GPER have a wide
range of functional applications and can contribute to our under-
standing of the contributions of GPER signaling in complex systems
expressing ER␣ and/or ER␤ in addition to GPER.
Particularly in systems expressing multiple estrogen receptors,
the utility of these compounds derives from and is limited by their
selectivity for GPER vs. ER␣ and ER␤. These small molecules are
based on a common tetrahydro-3H-cyclopenta[c]quinoline scaf-
fold, with the key difference between agonist G-1 and antagonist
G15 being the presence of an ethanone moiety on G-1 [9,10]. Since
G15 lacks this bulky substituent group and, as we demonstrate here,
exhibits increased ER␣/␤ activity compared to G-1, we synthesized
G36, a G-1 analog that contains an isopropyl moiety substituted for
the ethanone in G-1. We postulated that the increased bulk present
in G-1 and G36 would increase steric clashes within the binding
pocket of ER␣ and ER␤ compared to G15, thus limiting the binding
and downstream signaling activity observed when high doses of
G15 are used.
A catalytic amount of Sc(OTf)3 (0.049 g, 0.1 mmol, 10 mol%)
in anhydrous acetonitrile (1 mL) was added to a mixture of 6-
bromopiperonal (0.229 g, 1.00 mmol), 4-isopropylaniline (0.135 g,
1.0 mmol) and cyclopentadiene (0.33 g, 5.0 mmol) in acetonitrile
(3 mL). The reaction was stirred at ambient temperature (∼23 ◦C)
for 2 h with monitoring of product formation by thin layer chro-
matography using 20% ethyl acetate/hexanes eluent (Rf 0.5). The
product crystallized from acetonitrile, and was filtered and washed
with additional acetonitrile (5 mL) to provide the pure endo isomer
as colorless solid (390 mg, 94%). Mp 135–136 ◦C; FT-IR (KBr): 3435,
2955, 1613, 1504, 1040 cm−1 1H NMR (300 MHz, CDCl3) ␦ 7.18
;
(s, 1H), 7.02 (s, 1H), 6.92 (d, J = 2.0 Hz, 1H), 6.87 (d, J = 8.0, 2.0 Hz,
1H), 6.57 (d J = 8.0 Hz, 1H), 5.99 (d, J = 1.3 Hz, 1H), 5.97 (d, J = 1.3 Hz,
1H), 5.89–5.85 (m, 1H), 5.68–5.64 (m, 1H), 4.88 (d, J = 3.5 Hz, 1H),
4.10 (d, J = 8.2 Hz, 1H), 3.45 (bs, 1H), 3.22–3.11 (m, 1H), 2.84–2.73
(m, 1H), 2.64–2.55 (m, 1H), 1.83–1.75 (m, 1H), 1.22 (d, J = 1.5, 3H),
1.20 (d, J = 1.5 Hz, 3H); 13C NMR (75 MHz, CDCl3) ␦ 147.4, 147.2,
143.1, 139.9, 134.7, 134.0, 130.2, 126.8, 125.9, 124.2, 116.0, 113.0,
112.8, 108.1, 101.7, 56.8, 46.2, 42.2, 33.3, 31.3, 24.3, 24.1; HPLC–MS:
Electrospray positive ion (cone voltage 62 V, capillary voltage 3 kV).
G36 (1 mg/mL CH3CN, 20 ␮L) was injected onto a Symmetry® C18
(5 ␮m, 3.0 mm × 150 mm, Waters) column and eluted with 60–90%
acetonitrile (gradient 1.5% min−1) in water showed a single peak at
22.42 min (Fig. 4). UV–vis ꢀmax 294 nm (Fig. 5). ESI-MS m/z (ES+)
calcd for C22H22BrNO2 (M+H)+ 412.08; found 412.11; HRMS: calcd
[M+H]+ for C22H22BrNO2 412.0905, found 412.0912.
2.3. Reagents
17␤-estradiol was from Sigma (St. Louis, MO). G-1 and G15
were synthesized as previously described [9,10,23]. pERK anti-
bodies were from Cell Signaling (Danvers, MA). Goat–anti-rabbit
HRP was from GE-Amersham (Piscataway, NJ). DMEM and RPMI
1640 were obtained from Fisher Scientific (Pittsburgh, PA). 32%
paraformaldehyde (PFA) was obtained from Electron Microscopy
Sciences (Hatfield, PA).
In this report, we describe the identification and characteriza-
tion of G36 4, a GPER-selective antagonist with enhanced selectivity
for GPER and decreased activity towards ER␣ and ER␤ compared
to the previously described antagonist, G15. We show that high
doses of G15 induce low levels of transcription via ER␣ and that
G36 minimizes this off-target effect. Additionally, G36 maintains
equal efficacy as an antagonist of GPER compared to G15 in a range
of functional assays, both in vitro and in vivo.
2.4. Cell culture and transfection
COS7 cells were maintained in DMEM containing 10% fetal
bovine serum, 2 mM l-glutamine, 100 units/mL penicillin and
100 ␮g/mL streptomycin. SKBr3 cells were maintained in RPMI con-
taining 10% fetal bovine serum, 2 mM l-glutamine, 100 units/mL
penicillin and 100 ␮g/mL streptomycin. Cells were grown as a
monolayer at 37 ◦C, in a humidified atmosphere of 5% CO2 and 95%
air. For microscopy experiments, cells were seeded onto 12 mm
glass coverslips and allowed to adhere for at least 12 h prior to
transfection. Twenty-four hours prior to PI3K or pERK experiments,
medium was replaced with serum-free, phenol red free RPMI 1640.
For experiments requiring transient transfection of PH-mRFP, ER␣-
GFP, ER␤-GFP or GPER-GFP, Lipofectamine2000 (Invitrogen) was
used according to manufacturers instructions. PH-mRFP was trans-
fected at 1/4 the recommended amount to achieve appropriate
expression levels.
2. Materials and methods
2.1. Molecular docking
The crystal structure of the human estrogen receptor (ER)
ligand-binding domain in complex with 17␤-estradiol from the
RCSB Protein Data Bank (PDB ID: 1ERE) was used [22]. The receptor
was prepared for docking using the standard protocol implemented
in fred receptor, a wizard like graphical utility that prepares an
active site for docking with FRED (version 2.2.5, OpenEye Sci-
entific Software, Inc., Santa Fe, NM, USA, www.eyesopen.com,
2010). Three-dimensional conformations for the chosen ligands
(E2, G-1, G15, and G36) were generated using OMEGA (ver-
sion 2.3.2, OpenEye Scientific Software, Inc., Santa Fe, NM, USA,
www.eyesopen.com, 2010) and the ligands were docked in the
binding site of the receptor with FRED, using the default dock-
ing parameters. The Chemgauss3 scoring function, which uses
Gaussian smoothed potentials to measure the complementarity
of ligand poses within the active site, was used to evaluate and
compare how well the different ligands fit in the ligand binding
site.
2.5. ERE activation assays
ERE activity was determined using MCF-7 cells stably trans-
fected with an ERE-GFP reporter construct [24]. Briefly, cells were
deprived of estrogen for 4 days (with one intermediate medium
change) in phenol red free DMEM/F12 supplemented with 10%
charcoal-stripped FBS. In 24 well plates, ∼80,000 cells were seeded,

360
M.K. Dennis et al. / Journal of Steroid Biochemistry & Molecular Biology 127 (2011) 358–366
and 24 h later treated with the indicated compounds (dissolved
in DMSO, final DMSO 0.1% for each compound added) for 24 h in
triplicate, trypsinized, washed and analyzed for green fluorescence
by flow cytometry. Mean fluorescence intensities were determined
and normalized to E2 values following subtraction of vehicle con-
trol values.
was quantitated by immunofluorescence using anti-Ki-67 anti-
body (LabVision) followed by goat anti-mouse IgG conjugated
to Alexa488 (Invitrogen). Nuclei were counterstained with 4ꢀ,6-
diamidino-2-phenylindole (DAPI). At least 4 animals per treatment
were analyzed, and the Ki-67 immunodetection was repeated three
times per mouse.
2.6. Ligand binding assays
3. Results and discussion
Binding assays for ER␣ and ER␤ were performed as previously
described [7]. Briefly, COS7 cells were transiently transfected with
either ER␣-GFP or ER␤-GFP. Following serum starvation for 24 h,
cells (∼5 × 104) were incubated with G36 for 20 min in a final vol-
ume of 10 ␮L prior to addition of 10 ␮L 20 nM E2-Alexa633 in
saponin-based permeabilization buffer. Following 10 min at RT,
cells were washed once with 200 ␮L PBS/2%BSA, resuspended in
20 ␮L and 2 ␮L samples were analyzed on a DAKO Cyan flow
cytometer using HyperCytTM as described [25].
3.1. G-1/G15 ERE-mediated activity and design & synthesis of G36
In our original identification of the GPER antagonist G15 [10],
we speculated that the ethanone moiety of G-1 might be criti-
cal for GPER activation through the formation of hydrogen bonds
with the receptor and that loss of this bonding would result in a
compound that could still bind but not activate GPER, thus acting
as a competitive antagonist. In our initial characterization of G15,
we detected weak binding of G15 to ER␣ and ER␤ at high con-
centrations (≥10 ␮M) of G15; however, there were no effects of
G15 on estrogen receptor-mediated signaling observed in multiple
functional assays [10]. These results supported the use of G15 as
a selective GPER antagonist for in vitro and in vivo investigations.
In order to assess the structural basis of the observed low-affinity
binding, we performed docking studies with the 3-dimensional X-
ray crystal structure of the estrogen-bound ligand binding domain
of ER␣ (PDB ID 1ERE) [22]. E2, as expected, displayed a high binding
score (−99.6) for the receptor site, the docking pose being almost
2.7. Intracellular calcium mobilization
SKBr3 cells (1 × 107 cells/mL) were incubated in Hanks balanced
salt solution (HBSS, Gibco) containing 3 ␮M indo1-AM (Invitrogen)
and 0.05% pluronic acid for 1 h at RT. Cells were then washed twice
with HBSS, incubated at RT for 20 min, washed again with HBSS,
resuspended in HBSS at a density of 108 cells/mL and kept on ice
until assay, performed at a density of 2 × 106 cells/mL. Ca++ mobi-
lization was determined ratiometrically using ꢀex 340 nm and ꢀem
400/490 nm at 37 ◦C in a spectrofluorometer (QM-2000-2, Photon
Technology International) equipped with a magnetic stirrer.
˚
identical to the ligand in the X-ray structure (RMSD = 0.53 A). Dock-
ing of G-1 resulted in a significantly worse score (−52.8), mostly due
to a steric clash with the Arg394 residue in the binding site (Fig. 1),
which could explain the high level of GPER selectivity shown by
this ligand. However, the docking score of G15 (−76.2) indicates
that this ligand has a lower steric clash than G-1 within the binding
site and might therefore exhibit binding to the receptor, but with a
much lower affinity than E2.
To test whether G15 exhibits any activity towards ER␣, we
utilized a highly-sensitive estrogen response element (ERE) tran-
scriptional assay, in which estrogenic activity through either ER␣
(or potentially ER␤) results in activation of a stably transfected ERE-
GFP reporter construct in MCF-7 cells yielding a readout of GFP
2.8. PI3K activation
The PIP3 binding domain of Akt fused to mRFP1 (PH-mRFP1) was
used to localize cellular PIP3. COS7 cells (cotransfected with GPER-
GFP or ER␣-GFP and PH-mRFP1) cells were plated on coverslips and
serum starved for 24 h followed by stimulation with ligands as indi-
cated. The cells were fixed with 2% PFA in PBS, washed, mounted
in Vectashield and analyzed by confocal microscopy using a Zeiss
LSM510 confocal fluorescent microscope. Cells were preincubated
with 1 ␮M G36 as indicated for 20 min prior to stimulation and G36
was present during stimulation.
2.9. Western blotting
Cells were serum starved in phenol red-free RPMI 1640 for 24 h
prior to treatment for pERK Western blots. For activation of pERK,
cells were treated as indicated, washed once with ice-cold PBS
and lysed using NP-40 buffer. Twenty ␮g protein was loaded per
lane. Cells were preincubated with G15 or G36 as indicated prior to
stimulation and inhibitors were present during stimulation. Band
quantitation was carried out using ImageJ [26].
2.10. Mouse uterine estrogenicity assay
C57Bl6 female mice (Harlan) were ovariectomized at 10 weeks
of age. E2 and G36 were dissolved in absolute ethanol at 1 mg/mL
and further diluted in ethanol. For treatment, 10 ␮L of the appro-
priate dilution (of single or of combined compounds) was added
to 90 ␮L aqueous vehicle (0.9% NaCl with 0.1% albumin and 0.1%
Tween-20). Ethanol alone (10 ␮L) was added to 90 ␮L aqueous
vehicle as control (sham). At 12 days post-ovariectomy, mice were
injected subcutaneously with compound. Eighteen hr after injec-
tion, mice were sacrificed and uteri were dissected, fixed in 4%
paraformaldehyde, and embedded in paraffin. Five-micron sec-
tions were placed on slides, and proliferation in uterine epithelia
Fig. 1. Docking poses of selected ligands in the ER binding site: E2 green, G-1
magenta, G15 yellow. Arg394, which displays steric clashes with G-1 is depicted
in orange in the lower left corner. The Pro399-Leu402 loop was hidden in order to
better present the ligands. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of the article.)

M.K. Dennis et al. / Journal of Steroid Biochemistry & Molecular Biology 127 (2011) 358–366
361
with the difference being the presence of an ethanone moiety on
G-1 which is lacking in G15 (Fig. 3A). Based on the docking study,
we postulated that the presence of this side chain on G-1 resulted
in additional steric conflicts and reduced binding ability of G-1 for
ER␣. Since we speculated that the inability of G15 to form hydrogen
bonds through the ethanone group is responsible for its antago-
nist activity, we further hypothesized that replacing the reactive
ethanone group with a hydrophobic isopropyl group might gen-
erate a GPER antagonist with increased selectivity for GPER over
ER␣ and ER␤ compared to G15. Thus, we identified the isopropyl
group as a potential isosteric replacement of the ethanone group
that would lack both H-bond donor and acceptor capacity. The
target compound G36 was synthesized using a three-component
Povarov cyclization of 4-isopropylaniline, 6-bromopiperonal and
cyclopentadiene catalyzed by Sc(OTf)3 in acetonitrile (Fig. 3B).
The product crystallized from acetonitrile to afford G36 in excel-
lent yield as a racemic mixture of the analytically pure syn/endo
diastereomer.
Fig. 2. High concentrations of G15 mediate weak activation of ERE. MCF-7 cells
stably transfected with an ERE-GFP reporter were treated for 24 h with the indi-
cated concentration of E2, G-1 or G15. Whereas E2 shows half maximal activation at
approximately 100 pM, G-1 shows no activation up to 10 ␮M. G15 exhibits limited
activation (∼15–20%) at concentrations ranging from 1 to 10 ␮M.
expression (Fig. 2) [24]. In this assay, high doses (≥1 ␮M) of G15
(concentrations 105–106 greater than that required for E2) resulted
in weak activation of the ERE reporter (≤20% of E2 response), con-
sistent with the initial results that G15 binds weakly to ER␣ and
ER␤ at high concentrations. To establish that the expression of GFP
is mediated by ER␣ (or potentially ER␤), we coincubated E2 and G15
with ICI182,780 (a full ER antagonist that results in ER downregu-
lation), which yielded complete ablation of GFP induction by both
E2 and G15 (data not shown). Together, these results demonstrate
that at high concentrations, G15 is capable of mediating limited
ER-dependent transcriptional activity.
3.2. ER docking, binding and activation of G36
Docking analysis with G36 yielded a score (−52.3) very similar
to that of G-1 with a similar steric clash of the isopropyl group with
Arg 394 (Supplemental Fig. 1). To examine the binding selectivity
of G36 towards ER␣ and ER␤, we compared the binding of G36 to
that of G15 and E2 using a competitive fluorescent estrogen ligand
binding assay. Results from dose-response curves confirm that G15
displays weak but significant binding to ER␣ and ER␤ only at the
highest concentration, 10 ␮M, with G36 lacking detectable binding
activity to both ER␣ and ER␤ over the entire range of concentrations
tested (Fig. 4). It should be noted that the binding of G15 to ER␣ and
ER␤ is nevertheless weak, between 104 and 105 fold less than that
of E2. These results suggested that G36 may exhibit reduced acti-
With these results in mind, we revisited the structural similar-
ity of the GPER-selective agonist G-1 and G15. Both GPER-selective
compounds are based on the same cyclopenta[c]quinoline scaffold,
Fig. 3. (A) Structures of estrogen (17␤-estradiol) 1, G-1 2, G15 3 and G36 4. (B) Synthetic route for G36.

362
M.K. Dennis et al. / Journal of Steroid Biochemistry & Molecular Biology 127 (2011) 358–366
Fig. 5. G36 exhibits reduced activity towards ERE activation and inhibition com-
pared to G15. (A) Activation of ERE-GFP response in MCF7 cells by increasing
concentrations of G-1, G15 and G36. (B) Inhibition of ERE response induced by 1 nM
E2 as a function of increasing concentrations of G-1, G15 and G36. *p < 0.05 vs. DMSO
alone (A) or 1 nM E2 (B).
Fig. 4. G36 exhibits improved binding selectivity towards ER␣ and ER␤ compared
to G15. Dose response profile of E2, G15 and G36 for competition of E2-Alexa633
binding to ER␣-GFP (A) or ER␤-GFP (B). G36 shows decreased binding to ER␣ and
ER␤ at the highest concentration compared to G15.
via either the MAPK or PI3K pathways in ER␣- or ER␤-expressing
cells [10], suggesting that the partial agonist activity seen here
applies only to transcriptional signaling through the classical ERs
and not to ER-mediated rapid signaling events. This separation of
rapid and genomic signaling events has previously been observed
with androgen receptor ligands, with some promoting genomic sig-
naling and others preferentially activating non-genomic pathways
[27,28]. In contrast to G15, G36 shows much lower activity in the
ERE activation assay, with no significant ERE response elicited by
G36 concentrations up to 1 ␮M and only ∼5% activation at 10 ␮M
compared to ∼25% for G15 at 10 ␮M. G36 also shows no ability
to antagonize E2-mediated ERE response at concentrations up to
10 ␮M. It should be noted that the inhibitory effects of G-1 and G15
on estrogen-induced transcription were absent when E2 was used
at 10 nM (data not shown), suggesting that the inhibitory effects
are only present at low estrogen concentration. Nevertheless, these
results, along with the ER␣ and ER␤ binding data, suggest that G36
has decreased activity via ER␣ and ER␤ compared to the previously
described GPER antagonist, G15.
vation of ERE in comparison to G15. In addition to being a highly
sensitive functional assay, one benefit of the ERE assay is that it can
be used to investigate both agonist and antagonist properties of
compounds. We took advantage of this property to probe more fully
the activities of G-1, G15 and G36 against ER␣/␤ (Fig. 5). We con-
firmed that G-1, when administered to MCF7 cells stably expressing
the ERE-GFP reporter, displayed no agonist activity at doses up to
10 ␮M; however, when G-1 is administered at high doses it does
show weak antagonism of the 1 nM E2-mediated ERE response. This
finding is noteworthy, as it suggests that G-1 should optimally be
used at sub-micromolar concentrations when investigating GPER
function in complex systems where transcriptional effects may be
relevant, thus limiting possibly confounding results from potential
inhibition of estrogen-mediated transcriptional effects via ER␣/␤.
Most published work has employed G-1 at sub-micromolar doses
(typically 10–100 nM, occasionally ∼1 ␮M), so the finding that G-
1 antagonizes ER␣/␤ function at high doses does not detract from
such studies, particularly since most studies are carried out over
short (i.e. non-genomic) time frames and in the absence of added
estrogen.
3.3. G36 selectively inhibits estrogen-mediated PI3K activity
through GPER
We next investigated whether G36 maintains the inhibitory
activity towards GPER displayed by G15 [10]. First, to examine PI3K
activation of endogenous receptors, COS7 cells were transfected
with PH-RFP, a fluorescent reporter of PIP3 localization whose
nuclear translocation is used to assess PI3K activation [7]. Neither
E2 nor G-1 nor G36 at a concentration of 10 ␮M demonstrated any
activation of PI3K, indicating that COS7 cells do not exhibit either
receptor-mediated or non-specific responses to any of these lig-
ands (Fig. 6A). Next, COS7 cells were transiently transfected with
In contrast to G-1, G15 demonstrates limited agonism of the
ERE reporter (Fig. 5) with ∼15% activity at 1 ␮M and ∼25% activity
at 10 ␮M compared to the maximal E2 response (10 nM, see Fig. 2).
Additionally, G15 inhibits the E2-induced ERE response by ∼30%
when administered at 10 ␮M with no significant effect at 1 ␮M.
These results suggest that G15 likely functions as a partial ago-
nist of transcription mediated by ERs. Our prior studies however
demonstrated that G15 does not mediate rapid signaling events

M.K. Dennis et al. / Journal of Steroid Biochemistry & Molecular Biology 127 (2011) 358–366
363
either ER␣-GFP or GPER-GFP and PH-RFP. As expected, due to its
minimal activity in ERE assays, G36 did not inhibit E2-mediated
PI3K activation in cells expressing ER␣-GFP (Fig. 6B). Conversely,
in cells expressing GPER-GFP, G36 effectively inhibited PI3K activa-
tion induced by E2 (Fig. 6C), suggesting that the modifications made
to the initial chemical scaffold had maintained the GPER antagonist
activity while decreasing the ER␣ activity of G15.
assay [10], with G36 showing slightly higher potency for inhibit-
ing E2-mediated (G36 IC50 = 112 nM, G15 IC50 = 190 nM) calcium
mobilization and essentially identical potency for G-1-mediated
(G36 IC50 = 165 nM, G15 IC50 = 185 nM) calcium mobilization. To
assess whether G36 might act non-specifically to inhibit calcium
mobilization, we tested G36 at 10 ␮M for its effect on calcium mobi-
lization by an unrelated GPCR, specifically endogenous purinergic
receptors (Fig. 7C). G36 exhibited no inhibition of calcium mobiliza-
tion by 1 ␮M ATP, indicating high selectivity for GPER. Furthermore,
G36 itself did not alter the basal calcium level in cells (Fig. 7C, first
arrow). Thus, G36 affords enhanced selectivity as an antagonist of
GPER with similar potency in the inhibition of GPER function when
compared to G15.
3.4. G36 inhibits estrogen- and G-1-mediated calcium
mobilization
In order to determine the potency of G36 compared to G15, we
assayed the ability of G36 to inhibit both E2- and G-1-mediated cal-
cium mobilization in SKBr3 cells, which endogenously express only
GPER and neither ER␣ nor ER␤. Both E2 and G-1 mediated similar
levels of calcium mobilization at 200 nM (Fig. 7A). To determine the
potency of G36, cells were stimulated with a constant dose of E2
or G-1 and increasing concentrations of G36 were used to gener-
ate a dose-response curve (Fig. 7B). IC50 values for G36 inhibition
of E2 and G-1 mediated calcium mobilization were similar, with
G36 having an IC50 of 112 nM for E2 and 165 nM for G-1. These
values compare favorably with those found for G15 in the same
3.5. G36 inhibits estrogen- and G-1-mediated but not
EGF-mediated ERK activation
Since activation of the ERK/MAPK signaling pathway was one of
the functions first attributed to GPER [6], we verified the activity
of G36 as an antagonist of GPER function in ERK activation (Fig. 8).
SKBr3 cells, which express only GPER and neither ER␣ nor ER␤,
were stimulated with E2 and the GPER-selective agonist G-1. With
Fig. 6. G36 inhibits E2-induced PI3K activation in cells expressing GPER, but not in cells expressing ER␣. (A) COS7 cells (which lack endogenous ER␣, ER␤ and GPER) transiently
transfected with only PH-RFP show no activation of PI3K, as evidenced by the lack of nuclear translocation of PH-RFP, in response to E2, G-1 or G36 (all at 10 ␮M). (B) COS7
cells transiently transfected with ER␣-GFP and PH-RFP activate PI3K, as evidenced by nuclear translocation of PH-RFP in response to E2 and this response is not inhibited by
the presence of G36. (C) In COS7 cells transiently transfected with GPER-GFP and PH-RFP, G36 inhibits the PI3K activation induced by E2.

364
M.K. Dennis et al. / Journal of Steroid Biochemistry & Molecular Biology 127 (2011) 358–366
Fig. 6. (Continued ).
both stimuli, we detected a 5–6-fold increase in the level of pERK.
Upon stimulation in the presence of either G15 or G36, we observed
a significant decrease in the E2- and G-1-induced ERK activation.
To test for possible non-specific effects of G36 or G15, we examined
EGF-induced ERK activation, which was unchanged by the presence
of either G15 or G36, illustrating the selectivity of each of these
molecules for GPER.
3.6. G36 inhibits estrogen- and G-1-mediated uterine
proliferation
In order to extend our findings in vivo, we evaluated the effect of
G36 in the proliferative response of the uterine epithelium (Fig. 9).
Upon administration of E2, uterine epithelia exhibit increased pro-
liferation, as evidenced by the increase in the percent of cells

M.K. Dennis et al. / Journal of Steroid Biochemistry & Molecular Biology 127 (2011) 358–366
365
Fig. 8. G36 inhibits ERK activation induced by E2 and G-1 via GPER. SKBr3 cells
were stimulated with E2 (10 nM), G-1 (10 nM) or EGF (10 ng/mL) in the presence of
vehicle control (DMSO), 1 ␮M G15 or G36 (pre-incubated for 15 min) and analyzed
for pERK and total ERK levels by Western blot. Quantitation of pERK activation was
from three independent experiments. *p < 0.05 vs. DMSO alone.
Fig. 7. G36 inhibits E2 and G-1 mediated calcium mobilization in SKBr3 cells. (A)
SKBr3 cells, which endogenously express only GPER but neither ER␣ nor ER␤, were
monitored for calcium mobilization induced by either E2 (200 nM) or G-1 (200 nM).
DMSO was added at the first arrow as a control and either E2 or G-1, as indicated at
the second arrow. (B) E2- and G-1-mediated calcium mobilization are inhibited by
increasing concentrations of G36 (pre-incubated for ∼2 min, following the scheme in
A). (C) Purinergic receptor activation (by 1 ␮M ATP, second arrow) was not inhibited
by pretreatment with 10 ␮M G36 (first arrow). Data is normalized to E2 (200 nM) or
G-1 (200 nM) activation of calcium mobilization in the presence of vehicle control
(DMSO) as shown in A.
Fig. 9. G36 inhibits GPER-mediated proliferation in vivo. Epithelial uterine cell pro-
liferation was assessed in the presence of E2, G-1, G36, G-1 + G36 or E2 + G36.
Amounts injected were as follows: E2: 10 ␮g/kg; G-1: 10 ␮g/kg; G36 (alone or in
combination): 50 ␮g/kg. Ki-67 positivity of the uterine epithelium was determined
by immunofluorescence microscopy. *p < 0.05 vs. sham; **p < 0.05 vs. E2.
onist and demonstrate the utility of GPER-selective agonists and
antagonists in a complex model system.
staining positive for the proliferation marker Ki-67. As we have
previously observed [10], G-1 can also induce significant prolifera-
tion of uterine epithelia, albeit to a substantially lesser extent than
E2. This result is most likely due to E2 activating the full cohort
of estrogen receptors in the mouse (ER␣, ER␤ and GPER), whereas
G-1 only activates GPER, which alone cannot fully recapitulate the
response of ER␣ and ER␤. G36 administered alone had no effect
on proliferation, whereas G36 administered with G-1 completely
abrogated the proliferative response to G-1, as expected of a GPER
antagonist. Combination treatment with E2 and G36 resulted in
a significant decrease in proliferation compared to E2 treatment,
consistent with the inhibition previously observed with G15. These
in vivo results confirm our in vitro findings that G36 is a GPER antag-
4. Conclusions
In summary, we have developed a second generation GPER
antagonist G36 with improved performance and selectivity that
exhibits significantly decreased off-target effects on ER␣ and ER␤
compared to our previously described antagonist, G15. It is how-
ever important to note the context in which both G-1 and G15/G36
are likely to be used. G-1 has typically been used to determine
the contribution of GPER activation to a cellular or physiological
response in the absence of estrogen, either by its omission in culture
medium or following ovariectomy in mice. Thus, although G-1 may

366
M.K. Dennis et al. / Journal of Steroid Biochemistry & Molecular Biology 127 (2011) 358–366
exhibit slight inhibitory activity on estrogen-mediated ERE activa-
tion, its complete lack of stimulatory activity represents the more
important consideration. For G-1, it is not clear that the inhibitory
activity is a result of ER binding, as we were previously unable to
detect significant competition even at 10 ␮M [9]. It is therefore
possible that this inhibitory effect is indirect as a result of rapid
signaling initiated by GPER. Interestingly, very high doses of G-1
have recently been shown to inhibit estrogen-mediated uterine
epithelial cell proliferation in vivo through inhibition of stromal
ERK1/2 activation and ER␣ phosphorylation [29]. Alternatively,
G15 demonstrates both binding at high concentrations as well
as stimulatory activity in the absence of estrogen and inhibitory
activity in the presence of estrogen. The observation of binding
suggests that these functional effects may be direct via ER␣ bind-
ing. However, G15 would typically not be used alone (except as
a negative control) and thus the stimulatory activity would likely
not be of concern, unless potentially used in combination with
G-1, where ERE activation could confound results. Furthermore,
any effects would be limited to “long-term” assays where ERE
activation could be involved and not rapid or acute events (e.g.
<1 h).
The improved selectivity of G36 in terms of <∼5% activation
and inhibition of ERE-mediated transcription at 10 ␮M will make it
highly useful for probing the functions of GPER in a wide range
of complex assays and model systems that express one or both
of the classical ERs in addition to GPER. Nevertheless, as with any
pharmacological agent, conclusions regarding the involvement of
the presumed target should be confirmed with the use of siRNA
or knockout mice where possible. Building from the numerous
documented successful applications using G15 to examine the con-
tributions of GPER to estrogen physiology, we believe that the
development of G36 will give researchers a more selective tool to
investigate GPER function.
[6] E.J. Filardo, J.A. Quinn, K.I. Bland, A.R. Frackelton Jr., Estrogen-induced activation
of Erk-1 and Erk-2 requires the G protein-coupled receptor homolog, GPR30,
and occurs via trans-activation of the epidermal growth factor receptor through
release of HB-EGF , Mol. Endocrinol. 14 (10) (2000) 1649–1660.
[7] C.M. Revankar, D.F. Cimino, L.A. Sklar, J.B. Arterburn, E.R. Prossnitz, A transmem-
brane intracellular estrogen receptor mediates rapid cell signaling , Science 307
(5715) (2005) 1625–1630.
[8] E.R. Prossnitz, L.A. Sklar, T.I. Oprea, J.B. Arterburn, GPR30: a novel therapeu-
tic target in estrogen-related disease , Trends Pharmacol. Sci. 29 (3) (2008)
116–123.
[9] C.G. Bologa, C.M. Revankar, S.M. Young, B.S. Edwards, J.B. Arterburn, A.S. Kise-
lyov, M.A. Parker, S.E. Tkachenko, N.P. Savchuck, L.A. Sklar, T.I. Oprea, E.R.
Prossnitz, Virtual and biomolecular screening converge on a selective agonist
for GPR30 , Nat. Chem. Biol. 2 (4) (2006) 207–212.
[10] M.K. Dennis, R. Burai, C. Ramesh, W.K. Petrie, S.N. Alcon, T.K. Nayak, C.G. Bologa,
A. Leitao, E. Brailoiu, E. Deliu, N.J. Dun, L.A. Sklar, H.J. Hathaway, J.B. Arterburn,
T.I. Oprea, E.R. Prossnitz, In vivo effects of a GPR30 antagonist , Nat. Chem. Biol.
5 (6) (2009) 421–427.
[11] J. Teng, Z.Y. Wang, E.R. Prossnitz, D.E. Bjorling, The G protein-coupled recep-
tor GPR30 inhibits human urothelial cell proliferation , Endocrinology 149 (8)
(2008) 4024–4034.
[12] J. Kuhn, O.A. Dina, C. Goswami, V. Suckow, J.D. Levine, T. Hucho, GPR30 estrogen
receptor agonists induce mechanical hyperalgesia in the rat , Eur. J. Neurosci.
27 (7) (2008) 1700–1709.
[13] S.D. Noel, K.L. Keen, D.I. Baumann, E.J. Filardo, E. Terasawa, Involvement of
G-Protein coupled receptor 30 (GPR30) in rapid action of estrogen in primate
LHRH neurons , Mol. Endocrinol. 23 (2009) 349–359.
[14] E. Brailoiu, S.L. Dun, G.C. Brailoiu, K. Mizuo, L.A. Sklar, T.I. Oprea, E.R.
Prossnitz, N.J. Dun, Distribution and characterization of estrogen receptor G
protein-coupled receptor 30 in the rat central nervous system , J. Endocrinol.
193 (2) (2007) 311–321.
[15] C. Wang, B. Dehghani, Y. Li, L.J. Kaler, T. Proctor, A.A. Vandenbark, H.
Offner, Membrane estrogen receptor regulates experimental autoimmune
encephalomyelitis through up-regulation of programmed death 1 , J. Immunol.
182 (5) (2009) 3294–3303.
[16] E. Blasko, C.A. Haskell, S. Leung, G. Gualtieri, M. Halks-Miller, M. Mahmoudi,
M.K. Dennis, E.R. Prossnitz, W.J. Karpus, R. Horuk, Beneficial role of the GPR30
agonist G-1 in an animal model of multiple sclerosis , J. Neuroimmunol. 214
(1–2) (2009) 67–77.
[17] E. Haas, I. Bhattacharya, E. Brailoiu, M. Damjanovic, G.C. Brailoiu, X. Gao, L.
Mueller-Guerre, N.A. Marjon, A. Gut, R. Minotti, M.R. Meyer, K. Amann, E.
Ammann, A. Perez-Dominguez, M. Genoni, D.J. Clegg, N.J. Dun, T.C. Resta, E.R.
Prossnitz, M. Barton, Regulatory role of G protein-coupled estrogen receptor
for vascular function and obesity , Circ. Res. 104 (3) (2009) 288–291.
[18] S.H. Lindsey, K.A. Carver, E.R. Prossnitz, M.C. Chappell, Vasodilation in response
to the GPR30 agonist G-1 is not different from estradiol in the mRen2. Lewis
female rat , J. Cardiovasc. Pharmacol. (2011).
Acknowledgments
[19] C. Peyton, P. Thomas, Involvement of epidermal growth factor receptor sig-
naling in estrogen inhibition of oocyte maturation mediated through the G
protein-coupled estrogen receptor (GPER) in Zebrafish (Danio rerio) , Biol
Reprod 85 (1) (2011) 42–50.
[20] S. Gingerich, G.L. Kim, J.A. Chalmers, M.M. Koletar, X. Wang, Y. Wang, D.D.
Belsham, Estrogen receptor alpha and G-protein coupled receptor 30 mediate
the neuroprotective effects of 17beta-estradiol in novel murine hippocampal
cell models , Neuroscience 170 (1) (2010) 54–66.
[21] Z. Jenei-Lanzl, R.H. Straub, T. Dienstknecht, M. Huber, M. Hager, S. Grassel, R.
Kujat, M.K. Angele, M. Nerlich, P. Angele, Estradiol inhibits chondrogenic differ-
entiation of mesenchymal stem cells via nonclassic signaling , Arthritis Rheum.
62 (4) (2011) 1088–1096.
This work was supported by NIH grants R01 CA127731 (ERP,
JBA, TIO), CA118743 (ERP) and CA116662 (ERP), and MH084690
(LAS); the New Mexico Cowboys for Cancer Research Foundation
(JBA, ERP); Oxnard Foundation (ERP); and the Stranahan Foun-
dation (ERP). Data were generated in the Flow Cytometry and
Fluorescence Microscopy (http://hsc.unm.edu/crtc/microscopy/
Facility.html) Shared Resource Centers supported by the Univer-
sity of New Mexico Health Sciences Center and the University of
New Mexico Cancer Center.
[22] A.M. Brzozowski, A.C. Pike, Z. Dauter, R.E. Hubbard, T. Bonn, O. Engstrom, L.
Ohman, G.L. Greene, J.A. Gustafsson, M. Carlquist, Molecular basis of agonism
and antagonism in the oestrogen receptor , Nature 389 (6652) (1997) 753–758.
[23] R. Burai, C. Ramesh, M. Shorty, R. Curpan, C. Bologa, L.A. Sklar, T. Oprea, E.R.
Prossnitz, J.B. Arterburn, Highly efficient synthesis and characterization of the
GPR30-selective agonist G-1 and related tetrahydroquinoline analogs , Org.
Biomol. Chem. 8 (2010) 2252–2259.
[24] Y. Yamaguchi, H. Takei, K. Suemasu, Y. Kobayashi, M. Kurosumi, N. Harada, S.
Hayashi, Tumor–stromal interaction through the estrogen-signaling pathway
in human breast cancer , Cancer Res. 65 (11) (2005) 4653–4662.
[25] S. Ramirez, C.T. Aiken, B. Andrzejewski, L.A. Sklar, B.S. Edwards,
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jsbmb.2011.07.002.
References
High-throughput flow cytometry: validation in microvolume bioassays
Cytometry A 53 (1) (2003) 55–65.
[26] W. Rasband, ImageJ , NIH, Bethesda, MD, 2007.
,
[1] D.P. Edwards, Regulation of signal transduction pathways by estrogen and pro-
gesterone , Annu. Rev. Physiol. 67 (2005) 335–376.
[2] C.A. Lange, D. Gioeli, S.R. Hammes, P.C. Marker, Integration of rapid signaling
events with steroid hormone receptor action in breast and prostate cancer ,
Annu. Rev. Physiol. 69 (2007) 171–199.
[3] E.R. Prossnitz, J.B. Arterburn, H.O. Smith, T.I. Oprea, L.A. Sklar, H.J. Hathaway,
Estrogen signaling through the transmembrane G protein-coupled receptor
GPR30 , Annu. Rev. Physiol. 70 (2008) 165–190.
[4] L. Albanito, A. Madeo, R. Lappano, A. Vivacqua, V. Rago, A. Carpino, T.I. Oprea,
E.R. Prossnitz, A.M. Musti, S. Ando, M. Maggiolini, G protein-coupled recep-
tor 30 (GPR30) mediates gene expression changes and growth response to
17beta-estradiol and selective GPR30 ligand G-1 in ovarian cancer cells , Cancer
Res. 67 (4) (2007) 1859–1866.
[27] S. Kousteni, T. Bellido, L.I. Plotkin, C.A. O’Brien, D.L. Bodenner, L. Han, K. Han, G.B.
DiGregorio, J.A. Katzenellenbogen, B.S. Katzenellenbogen, P.K. Roberson, R.S.
Weinstein, R.L. Jilka, S.C. Manolagas, Nongenotropic, sex-nonspecific signaling
through the estrogen or androgen receptors: dissociation from transcriptional
activity , Cell 104 (5) (2001) 719–730.
[28] A. Gill, M. Jamnongjit, S.R. Hammes, Androgens promote maturation and
signaling in mouse oocytes independent of transcription: a release of inhi-
bition model for mammalian oocyte meiosis , Mol. Endocrinol. 18 (1) (2004)
97–104.
[29] F. Gao, X. Ma, A.B. Ostmann, S.K. Das, GPR30 activation opposes
estrogen-dependent uterine growth via inhibition of stromal ERK1/2 and
estrogen receptor alpha (ER{alpha}) phosphorylation signals , Endocrinology
152 (4) (2011) 1434–1447.
[5] D.P. Pandey, R. Lappano, L. Albanito, A. Madeo, M. Maggiolini, D. Picard, Estro-
genic GPR30 signalling induces proliferation and migration of breast cancer
cells through CTGF , EMBO J. 28 (5) (2009) 523–532.