Modulation of estrogen-related receptors subtype selectivity: Conversion of an ERRβ/γ selective agonist to ERRα/β/γ pan agonists

Article abstract of DOI:10.1016/j.bioorg.2020.104079

Estrogen Related Receptors (ERRs) are key regulators of energy homeostasis and play important role in the etiology of metabolic disorders, skeletal muscle related disorders, and neurodegenerative diseases. Among the three ERR isoforms, ERRα emerged as a potential drug target for metabolic and neurodegenerative diseases. Although ERRβ/γ selective agonist chemical tools have been identified, there are no chemical tools that effectively target ERRα agonism. We successfully engineered high affinity ERRα agonism into a chemical scaffold that displays selective ERRβ/γ agonist activity (GSK4716), providing novel ERRα/β/γ pan agonists that can be used as tools to probe the physiological roles of these nuclear receptors. We identified the structural requirements to enhance selectivity toward ERRα. Molecular modeling shows that our novel modulators have favorable binding modes in the LBP of ERRα and can induce conformational changes where Phe328 that originally occupies the pocket is dislocated to accommodate the ligands in a rather small cavity. The best agonists up-regulated the expression of target genes PGC-1α and PGC-1β, which are necessary to achieve maximal mitochondrial biogenesis. Moreover, they increased the mRNA levels of PDK4, which play an important role in energy homeostasis.

Full text of DOI:10.1016/j.bioorg.2020.104079

Journal Pre-proofs
Modulation of Estrogen-Related Receptors Subtype Selectivity: Conversion
of an ERRβ/γ Selective Agonist to ERRα/β/γ Pan Agonists
Mohamed Shahien, Mohamed Elagawany, Sadichha Sitaula, Shaimaa S.
Goher, Sheryl L. Burris, Ryan Sanders, Amer Avdagic, Cyrielle Billon,
Lamees Hegazy, Thomas P. Burris, Bahaa Elgendy
PII:
S0045-2068(20)31376-6
DOI:
https://doi.org/10.1016/j.bioorg.2020.104079
YBIOO 104079
Reference:
To appear in:
Bioorganic Chemistry
Received Date:
Revised Date:
Accepted Date:
7 May 2020
30 June 2020
4 July 2020
Please cite this article as: M. Shahien, M. Elagawany, S. Sitaula, S.S. Goher, S.L. Burris, R. Sanders, A. Avdagic,
C. Billon, L. Hegazy, T.P. Burris, B. Elgendy, Modulation of Estrogen-Related Receptors Subtype Selectivity:
Conversion of an ERRβ/γ Selective Agonist to ERRα/β/γ Pan Agonists, Bioorganic Chemistry (2020), doi:
https://doi.org/10.1016/j.bioorg.2020.104079
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Elsevier Inc. All rights reserved.

Modulation of Estrogen-Related Receptors Subtype Selectivity: Conversion of an ERRβ/γ
Selective Agonist to ERRα/β/γ Pan Agonists
Mohamed Shahiena, Mohamed Elagawanyb,c,d, Sadichha Sitaulac, Shaimaa S. Gohera, Sheryl L.
Burrisc, Ryan Sandersc, Amer Avdagicc, Cyrielle Billonb,c, Lamees Hegazyb,c, Thomas P.
Burrisb,c, Bahaa Elgendya,b,c,*
a
Chemistry Department, Faculty of Science, Benha University, Benha 13518, Egypt
bDepartment of Pharmaceutical and Administrative Sciences, St. Louis College of Pharmacy, St. Louis, MO
63110, USA
cCenter for Clinical Pharmacology, Washington University School of Medicine and St. Louis College of
Pharmacy, St. Louis, MO 63110, USA
dDepartment of Pharmaceutical Chemistry, Faculty of Pharmacy, Damanhour University, Damanhour,
Egypt
Abstract
Estrogen Related Receptors (ERRs) are key regulators of energy homeostasis and play important
role in the etiology of metabolic disorders, skeletal muscle related disorders, and
neurodegenerative diseases. Among the three ERR isoforms, ERRα emerged as a potential drug
target for metabolic and neurodegenerative diseases. Although ERR/ selective agonist chemical
tools have been identified, there are no chemical tools that effectively target ERR agonism. We
successfully engineering high affinity ERR agonism into a chemical scaffold that displays
selective ERRβ/ agonist activity (GSK4716), providing novel ERR/β/ pan agonists that can be
used as tools to probe the physiological roles of these nuclear receptors. We identified the structural
requirements to enhance selectivity toward ERRα. Molecular modeling shows that our novel
modulators have favorable binding modes in the LBP of ERRα and can induce conformational
changes where Phe328 that originally occupies the pocket is dislocated to accommodate the
ligands in a rather small cavity. The best agonists up-regulated the expression of target genes PGC-
1α and PGC-1β , which are necessary to achieve maximal mitochondrial biogenesis. Moreover,
they increased the mRNA levels of PDK4, which play an important role in energy homeostasis.

Key Words: Estrogen-Related Receptors; Pan Agonists; Molecular Modeling; N-Acyl
Hydrazones
1. Introduction
Estrogen Related Receptors (ERRs) are members of the nuclear hormone receptors (NRs)
superfamily. The ERRs subfamily comprise three members, ERRα, ERRβ and ERRγ; and are
closely related to the estrogen receptors (ERα and ERβ). ERRα and ERRβ were identified by
Giguère and his co-workers in 1988 [1], and ERRγ was identified in 1999 by Chen et al. [2].
The ERRs are orphan receptors since no natural ligands have been identified for any of the three
ERR isoforms. Unlike ERs, ERRs have constitutive activity and can function in absence of ligands.
Although ERRs are structurally related to ERs and share sequence similarity with these receptors,
they do not bind with estrogens, which are the endogenous ligands of the ERs. ERRα is expressed
mainly in tissues of high energy demand such as brown adipose tissue, intestine, and skeletal
muscles and functions as a sensor of energy metabolism. ERRβ is expressed in low levels in adult
skeletal muscle, heart, kidney, eyes, and ears. This isoform is known to play an important role in
the early development and its postnatal expression is highly restricted. ERRγ is expressed mainly
in skeletal muscle, heart, pancreas, kidney, placenta, spinal cord and brain [3-5]. Within the ERR
subfamily, ERRα is the most divergent of the // trio, especially with regard to the ligand
binding domain (LBD). The ligand binding pocket (LBP) of ERRα is only 100 Å3, which makes
it the smallest LBP among NRs. ERRα is constitutively active because the LBP is partially filled
with the amino acid residue Phe328, which allows helix 12 (H12) to adopt an active conformation
that can interact with coactivators. The crystal structure of ERRα LBD in complex with coactivator
PGC-1α shows that estrogen-like ligands cannot fit the small LBP and most likely will disrupt the
active conformation [6]. This is probably why the majority of ERRα identified ligands are inverse

agonists. Co-crystal structure of inverse agonist 1 with ERRα LBD shows that only part of this
ligand fits in the LBD and the rest of the ligand sticks out from the LBP. The LBP undergoes
significant conformational changes to accommodate the ligand and Phe510 (on H12) was
dislocated to avoid steric clash with Phe328. Consequently, H12 was displaced to fill the
coactivator groove of the AF-2 and therefore inactivate the receptor. This molecular mechanism is
unique to ERRα and highlights the possibility of rationally designing drugs to modulate ERRα
despite the small size of the LBP. Even though ERRα LBP is small in size and partially filled, a
well-designed small-molecule ligand can be accommodated if it can promote conformational
changes and cause the cavity to open up.
Several ERRα inverse agonists have been identified. Kaempferol (2), dietary flavonol found in
many natural products, can reduce gene transcription by inhibiting the interaction between ERRα
and DNA response elements [7]. XCT790 (3) is another ERRα inverse agonist that works by
inducing ERRα ubiquitin-dependent proteasomal degradation [8,9]. This compound can inhibit the
proliferation of A549 lung cancer cell [10] and HepG2 hepatocarcinoma and its multi-drug
resistance (MDR) sub-line R-HepG2 by inducing mitochondrial reactive oxygen species [11] and
reducing mitochondrial mass. XCT790 (3) reduces the expression of PGC-1α [12] and PGC-1β
[13], thus suppressing mitochondrial biogenesis and increasing glucose uptake.
Scientists at GlaxoSmithKline suggested that ERRα may be intractable as a drug target after they
failed to identify ERRα agonists using high-throughput screening and structure guided design [14].
However, natural phytoestrogens were identified by virtual screening as potential ERRα agonists.
For example, in addition to functioning as ER and ER agonists, genistein (4a), daidzein (4b),
and biochanin (4c) are three isoflavones, which behaved as agonists for all ERRs and were shown
to activate ERRα in mammalian cell transfection and mammalian two-hybrid experiments [15].

N
N
O
OH
O
O
F3C
OH
O
O
S
N
OH
N
H
H
N
CF3
N
HO
XCT790
3
1
2
CF3
O
R2
R1
O
O
B
N
C
N
N
A
H
N
HO
HO
O
GSK4716
6
. Genistein: R1 = OH, R2 = OH
4a
5
. Daidzein:
. Biochanin A: R = OH, R = OCH3
R1 = H, R2 = OH
4b
1
2
4c
Figure 1. Examples of ERRα modulators.
Ding et al. identified a series of pyrido[1,2-a]pyrimidin-4-ones as ERRα agonists by screening
their carbonyl focused library [16]. These compounds enhanced the uptake of glucose and fatty
acid in C2C12 muscle cells through the elevation of mRNA and protein levels of ERRα in
downstream targets. Compound 5 has an agonistic effect on ERRα and was shown to increase the
transcription of ERRα in a dose dependent manner. Moreover, compound 5 activated the receptor
in 293FT cells pretreated with kaempferol (2), which is known inverse agonist of ERRα.
Interestingly, 5 increased the mRNA levels of key target genes involved in oxidative metabolism
such as MCAD, PDK4, and ATP5b. This compound was selective for ERRα over ERα, ERβ, and
ERRβ. However, 5 moderately agonized ERRγ [16].
As part of our ongoing research program to develop chemical tools for various nuclear receptors
(e.g. LXR [17-20], ROR [21], REV-ERB [22]), we became interested in developing ERRα/γ dual
agonists with the aim of eventually developing small molecule agonists specific to ERRα to
evaluate the function of this receptor in animal models as well as determine the potential utility of
ERRα ligands in human diseases. The lack of chemical probes to this drug target has hindered our

ability to study the relationship between ERRα activity and diseases states and to launch the
clinical translatability of this target. GSK4716 (6) (Fig. 1) was reported as an agonist of ERRβ and
ERRγ with no observed activity on ERRα [23]. In our hands, GSK4716 (6) showed very weak
activity against ERRα, which prompted us to revisit this lead compound to identify ERRα/γ dual
agonists. Herein we describe our probe of the structure activity relationships (SARs) of GSK4716
(6) through the chemical modification of rings A and B while maintaining the acyl hydrazone
moiety to identify the structural requirements to promote selectivity toward the ERRα isoform.
2. Results and discussion
2.1. Chemistry
The synthesis of compounds 10-46 is illustrated in Scheme 1. The acyl hydrazides (8) were
prepared from their corresponding methyl esters (7) through reaction with hydrazine in methanol
under reflux. All acyl hydrazides (8) were obtained in quantitative yields and their identity were
confirmed by comparing their melting points and mass to literature. The acyl hydrazones (10-46)
were synthesized by condensation of acyl hydrazides (8) with aldehydes (9) in ethanol in presence
of catalytic amount of trifluoroacetic acid.
H
O
R2
O
B
O
C
O
C
R2
B
NH2
N
NH2NH2.H2O
O
9a-q
N
N
R1
A
R1
R1
H
H
A
A
MeOH
0 °C to reflux
EtOH, reflux
TFA (Cat.)
7a-i
8a-i
10-46
Scheme 1. Synthesis of N-acyl hydrazones 10-46.
2.2. SAR study of acyl hydrazone derivatives

Synthesized compounds were screened in a cell-based (HEK293 cells) co-transfection assay using
ERRE reporter construct and pcDNA3.1 ERRγ or pcDNA3.1 ERR. Luciferase activity was
measured using the Dual-Glo luciferase reporter assay system (Promega). The values indicated
represent the means ± S.E. from four independently transfected wells. EC50 values for all described
compounds were obtained by titration in an eleven-point dose response format. All compounds
were screened against GSK4716 (6) as a reference compound for agonist activity. Compounds
XCT790 (3), which is a potent and specific inverse agonist for ERR, and 4-hydroxytamoxifen
(4-OHT), which is a nonselective inverse agonist for ERRγ, were used as positive controls to
identify inverse agonism activity.
Zuercher et al. [23] reported that the hydroxyl group or analogous hydrogen bond donor (HBD)
group on ring A is required to obtain ERRγ activity. Recently, Lin et al. [24] showed that removing
or repositioning this hydroxyl group led to significant decrease or even complete loss of ERRγ
activity. 4-Isopropyl and 4-diethylamino groups were the optimal substituents at ring B to maintain
ERRγ activity.
We started probing the SAR of GSK4716 (6) by removing both substituents at both rings A and
B, which led to complete loss of activity against both ERRα and ERRγ (i.e. 10). Then, we
investigated the SAR of ring B while ring A was unsubstituted phenyl. ERRγ activity was regained
upon introduction of 4-dimethylamino (i.e. 11) or 4-chloro (i.e. 12), but no ERRα activity was
observed. Compound 12 was >8-fold more potent than 11. While hydroxyl group at the ortho-
position of ring B was not tolerated (i.e. 13) and the compound was inactive against both ERRα
and ERRγ, compound 14 with hydroxyl group at the para-position showed weak activity against
both isoforms and was more selective toward ERRγ (>3-fold) (Table 1).

A methoxy group at para-position of ring B (i.e. 15) enhanced the activity toward both ERRα and
ERRγ but remained more selective toward ERRγ (≈3-fold). Compound 15 was ≈ 2.66-fold more
potent toward ERRα than 14. Introduction of nitro group at the ortho-position of ring B (i.e. 16)
increased the potency toward ERRα > 6-fold if compared to 15 and ≈17-fold if compared to 14.
Compound 16 remained more selective toward ERRγ over ERRα (1.6-fold). Moving the nitro
group from ortho -position to meta- (i.e. 17) or para-position (i.e. 18) led to the loss of ERRα
activity and significant decrease in ERRγ activity (Table 1). Docking of compound 16 in both
ERRα and ERRγ resulted in a similar binding pose in both receptors. The ligand is involved in
hydrophobic interactions with surrounding residues in the ligand binding pocket (LBP). The
aromatic ring B makes π-π interactions with Tyr326 in ERRγ and π-π interactions with Phe328 in
ERRα. The nitro group in this binding pose is oriented towards the inside of the cavity of ERRα,
and involved in cation- π interactions with Phe328 which correspond to Ala272 in ERRγ (Fig. 2).
We then begin investigating the SAR of ring A while ring B remained fixed as an unsubstituted
phenyl. When we introduced the hydroxyl group at ortho-position, we obtained compound 19 that
was active toward both ERRα (EC50 = 0.619 µM) and ERRγ (EC50 = 0.468 µM) with significant
enhancement in selectivity toward ERRα (EC50 α/γ = 1.32). Evans et al. [25] showed that
compound DY162, which is an analogue of 19 with hydroxyl group at para-position of ring A,
was ERRγ agonist with no activity toward ERRα.

A.
B.
Figure 2. Computational docking of 16 in ERRα (A) and ERRγ (B). Hydrogen bonds are shown
as green and blue dashed lines and hydrophobic interactions are demonstrated with purple and
orange dashed lines.
Table 1. In vitro ERRα and ERRγ agonistic activities of 10-24.
ERRα
EC50 (µM) EC50 (µM)
ERRγ
Comp.
Structure
EC50 α/γ ɑLog Pa
O
GSK4716
(6)
C
N
2.410
i.a.
0.067
i.a.
35.97
4.23
2.56
3.14
N
H
HO
O
C
N
10
11
-
-
N
H
N
O
C
i.a.
1.704
N
N
H
Cl
O
C
N
N
12
13
14
15
16
i.a.
0.197
i.a.
-
3.21
3.11
3.08
2.66
2.62
N
H
O
C
i.a.
-
N
H
OH
OH
O
O
C
N
7.267
2.733
0.430
2.219
0.919
0.268
3.27
2.97
1.60
N
H
O
C
N
N
N
H
O
C
N
H
NO2

O
C
N
17
18
19
20
21
9.464
i.a.
6.446
1.839
0.468
0.790
3.987
1.47
-
2.62
2.62
3.12
3.21
3.23
N
H
NO2
NO2
O
C
N
N
N
H
OH
O
C
0.619
1.714
i.a.
1.32
2.17
-
N
H
O
C
N
N
N
H
Cl
O
C
N
H
Br
O
C
N
N
22
23
24
i.a.
1.830
1.595
0.194
-
2.63
2.78
2.82
H
O2N
O
C
N
i.a.
-
N
H
O
C
N
0.209
1.07
N
H
ALOGPS 2.1, molecular property prediction [26]. i.a. = inactive
Substituting ring A with chlorine at para-position led to identification of compound 20, which was
weakly active against ERRα (EC50 =1.714 µM) and moderately potent against ERRγ (EC50 =0.790
µM). Bromine (i.e. 21) or nitro (i.e. 22) groups were not tolerated at the same position. While both
21 and 22 maintained moderate ERRγ activity, they completely lost ERRα activity. Methyl
substitution at ortho-position (i.e. 23) showed a similar pattern with moderate activity toward
ERRγ (EC50 = 1.595 µM) and no activity toward ERRα. Surprisingly, methyl substituent at para-
position (i.e. 24) enhanced the potency toward both ERRα (EC50 = 0.209 µM) and ERRγ (EC50 =
0.194 µM) significantly. Compound 24 was equipotent toward both isoforms (EC50 α/γ = 1.07)
(Table 1 and Fig. 3).

GSK4716 (6)
24
2.5
2.0
1.5
1.0
2.5
2.0
1.5
1.0
ERR α
ERR β
ERR γ
ERR α
ERR β
ERR γ
0.01
0.1
1
10
100
0.01
0.1
1
10
100
Log uM
Log uM
Figure 3. A cell-based co-transfection assay using ERR responsive luciferase reporter for
GSK4716 (6) and 24.
Molecular modeling of 24 in the active site of ERRγ show that ring B is making π-π interactions
with Tyr326 and the NH makes a hydrogen bond with the carbonyl oxygen of the same residue
(Fig. 4B). In ERRα, the predicted binding pose is a slightly tilted from the binding pose in ERRγ
and the compound is predicted to make π-π interactions with Phe328 on helix 3 (H3) (Fig. 4A).
A.
B.
Figure 4. Computational docking of 24 in ERRα (A) and ERRγ (B). Hydrogen bonds are shown
as blue dotted lines and π-π interactions are shown as broken purpled dashed lines.
To further probe the SAR of this scaffold, we decided to assess the effect of having two different
substituents at both rings A and B on the activity and selectivity toward ERRα and ERRγ. When
we introduced methyl group in ortho-position of ring A in 11, we obtained compound 25. Like 11,
compound 25 was inactive against ERRα and the potency toward ERRγ dropped more than 4-fold.
Introducing chlorine to the para-position of 12 gave compound 26, which was inactive against
both ERR isoforms. Interestingly, introducing hydroxyl group to the ortho-position of either ring

A (i.e. 19, Table 1) or ring B (i.e. 13, Table 1) showed a very distinct pharmacological behavior.
While 19 was moderately potent against both isoforms with slightly higher selectivity toward
ERRγ, 13 was completely inactive against both isoforms. This further supports the importance of
having a hydrogen bond donor (HBD) on ring A, especially in absence of any substituents at ring
B, for the compound to have affinity toward ERRs. Introducing two hydroxyl groups to the ortho-
position of both rings gave compound 27, which was more potent toward ERRα (EC50 =0.378µM)
than ERRγ (EC50 = 1.645 µM). Compound 27 is >4-fold more selective toward ERRα over ERRγ
(EC50 α/γ = 0.23). This compound is expected to have poor pharmacokinetic properties because it
is susceptible to metabolic clearance via glucuronidation or sulfonation of the two hydroxyl
groups.
A.
B.
Figure 5. Computational docking of 27 in ERRα (A) and ERRγ (B). Hydrogen bonds are shown
as blue dotted lines and π-π interactions are shown as broken cyan dashed lines.
In ERRγ, 27 is predicted to make π-π interactions with Tyr326 that lies on the β-sheets region and
hydrogen bonding interactions with Glu247. However, docking this compound in ERRα resulted
in a docking pose where the molecule is slightly tilted to interact with Phe328 (Figure 5). It is
possible that this interaction contributes to selectivity of this compound towards ERRα. The
hydroxyl group on ring A is predicted to make hydrogen bonding interactions with Glu331 in

ERRα. The amide group is making hydrogen bonding interaction with the backbone of Ty326 in
ERRγ and Phe328 in ERRα (Figure 5).
The activity toward ERRα was lost completely when we replaced the hydroxyl group on ring A of
27 with chlorine atom in 28. Compound 28 preserved activity toward ERRγ. On the contrary, 29
with methyl group in ortho-position of ring A was active toward both ERRα (EC50 = 0.453 µM)
and ERRγ (EC50 = 0.254 µM). Ring B of 29 is predicted to make π-π interactions with Tyr326 in
ERRγ. In ERRα, the same group is predicted to make π-π interactions with two aromatic side
chains; Phe328 and Phe382. Moreover, the NH group of 29 is predicted to make hydrogen bonding
interactions with the backbone of Tyr328 in ERRγ and Phe382 in ERRα (Figure 6).
A.
B.
Figure 6. Computational docking of 29 in ERRα (A) and ERRγ (B). Hydrogen bonds are shown
as green lines and hydrophobic interactions are shown as purple lines.
Moving hydroxyl group on ring B of 29 from ortho-position to para-position gave compound 30,
which was inactive toward both isoforms. Introducing chlorine or bromine in para-position of ring
A while maintaining hydroxyl group in ortho-position of ring B, gave compounds 31 and 32,
respectively. These two compounds showed weak activity toward ERRα and ERRγ but were more
selective toward ERRγ. Compound 33 with methyl group in the para-position of ring A showed
good activity toward both ERRα (EC50 = 0.483 µM) and ERRγ (EC50 = 0.235 µM). Attempt to
replace ring A with benzyl group gave compound 34 that was completely inactive against both

isoforms. Compound 34 has no affinity toward ERRs most probably due to the extra flexibility
gained by having methylene linker between ring A and the carbonyl group.
Compound 35 possess bromine in the para-position of ring A and methoxy in the para-position of
ring B and showed good activity toward ERRα (EC50 = 0.540 µM) and ERRγ (EC50 = 0.398 µM)
with good subtype selectivity (EC50 α/γ = 1.36). The addition of bromine to ring A enhanced the
subtype selectivity >2-fold (cf. 15; EC50 α/γ = 2.97). However, in absence of methoxy group,
compound 21 was inactive toward ERRα and very weak toward ERRγ. Compound 35
demonstrates the influence of additive effect and the interplay of substituents at ring A and ring B.
Therefore, it is feasible to develop isoform selective modulators for ERR subtypes by having the
right combination of substituents on both rings.
Both 18 (4-NO2 on ring B) and 22 (4-NO2 on ring A) were inactive toward ERRα and weakly
active toward ERRγ. Introducing the two nitro groups simultaneously at both rings in compound
36 led to a switch in activity toward ERRα from agonism to inverse agonism. Compound 36
showed a weak decrease in the basal activity of ERRα (IC50 = 8.116 µM) compared to the ERRα
potent inverse agonist XCT790 (3) (IC50 = 0.175 µM). Moreover, 36 was inactive against ERRγ.
Since ERRs are constitutively active receptors, ligands that decrease their constitutive activity are
identified as inverse agonists. Although the observed inverse agonism activity of this compound
was very weak, it demonstrates that two electron-withdrawing groups at para-position of both rings
are not tolerated. The compound lost activity toward both isoforms upon moving the nitro group
on ring B from para-position (i.e. 36) to meta-position (i.e. 37). The agonistic activity was retained
upon moving the nitro group on ring B from meta-position in 37 to ortho-position in 38 (EC50 =
0.646 µM for ERRα, and 0.260 µM for ERRγ). Comparing 38 with nitro group in the para-position
of ring A to 16 that lacks nitro group on ring A, we noticed a slight decrease in the potency of 38

against ERRα (≈1.5-fold) and insignificant change in potency against ERRγ. Moreover, subtype
selectivity decreased by ≈1.5-fold (EC50 α/γ = 1.6 for 16 vs 2.48 for 38). This decrease in activity
and subtype selectivity toward ERRα is most likely due to the smaller size of the ligand binding
pocket of ERRα, which prefers smaller ligands.
Table 2. In vitro ERRα and ERRγ agonistic activities of 25-38.
ERRα EC50 ERRγ EC50 EC50
Comp.
Structure
ɑLog Pa
(µM)
(µM)
α/γ
O
C
GSK4716
(6)
N
2.410
0.067
35.97
4.23
N
H
HO
N
O
25
26
i.a.
i.a.
7.268
i.a.
-
-
3.18
3.84
C
N
H
N
Cl
O
C
N
N
H
Cl
OH
O
C
N
27
28
29
30
0.378
i.a.
1.645
1.42
0.254
i.a.
0.23
2.79
3.79
3.35
3.35
N
H
OH
Cl
O
C
N
N
-
N
H
OH
OH
O
C
0.453
i.a.
1.78
-
N
H
OH
O
C
N
N
H
O
C
N
N
N
H
31
32
33
3.347
2.054
0.483
2.804
0.6748
0.235
1.19
3.04
2.06
3.76
3.86
3.37
OH
Cl
Br
O
C
N
H
OH
O
C
N
N
H
OH

O
C
N
34
35
i.a.
i.a.
-
2.70
3.32
N
H
O
O
C
N
0.540
0.398
1.36
N
H
Br
NO2
O
8.116
Inverse
agonist
C
N
N
N
36
i.a.
-
3.13
N
H
O2N
O2N
O2N
O
C
N
H
NO2
37
38
i.a.
i.a.
-
3.10
3.10
O
C
N
H
0.646
0.260
2.48
NO2
Finally, we decided to explore different substituents in lieu of ring B while maintaining the ring A
as an unsubstituted phenyl ring (Table 3). Neither small alkyl substituent (i.e. 39) nor phenylallyl
analogs (i.e. 41 and 42) were tolerated and these derivatives were inactive toward both ERR
isoforms. Compound 42 with thiophene ring as a bioisostere of phenyl ring in 10 showed weak
agonistic effect against both ERR isoforms and good subtype selectivity (EC50 α/γ = 1.78). Despite
the weak agonistic activity of 42, replacement of phenyl ring with its bioisostere (i.e. thiophene)
has changed the profile of 10 from completely inactive compound to an active compound (i.e. 42)
with acceptable isoform selectivity. The corresponding furan analogue (i.e. 43) was inactive
toward both isoforms. The 5-methyl substituted furan analogue (i.e. 44), has weak ERRγ activity
(EC50 = 2.701 µM) and no ERRα activity. The quinolin-2-ylmethylene benzohydrazide 45 was
more potent against ERRα (EC50 = 0.487 µM) over ERRγ (EC50 = 0.836 µM). Surprisingly, the
quinolin-4-ylmethylene benzohydrazide 46 showed weak inverse agonism behavior against ERRα
(IC50 = 4.578 µM) in comparison to the ERRα potent inverse agonist XCT790 (3) (IC50 = 0.175
µM).

Table 3. In vitro ERRα and ERRγ agonistic activities of 39-46.
ERRα
EC50 (µM) (µM)
ERRγ EC50
Comp.
Structure
EC50 α/γ ɑLog Pa
O
GSK4716
(6)
C
N
2.410
0.067
35.97
4.23
N
H
HO
O
C
N
39
40
N
H
i.a.
i.a.
i.a.
i.a.
-
-
1.50
3.42
O
C
N
N
H
N
O
C
41
i.a.
i.a.
-
3.81
N
N
N
N
H
O
C
S
42
43
2.158
i.a.
1.213
i.a.
1.78
-
2.79
2.25
N
H
O
C
O
N
H
O
C
O
N
N
N
44
i.a.
2.701
-
2.54
N
H
O
C
45
46
0.487
0.836
i.a.
0.58
-
2.84
2.54
N
H
N
O
C
N
4.578
Inverse
agonist
N
H
Our SAR data of developed agonists supported by molecular modelling show that removing the
hydroxyl group at para-position of ring A in GSK4716 (6), which forms H-bond with Glu384 and
avoiding using hydrogen bond donors at the same position reduced the selectivity toward ERRγ
and enhanced the selectivity toward ERRα. The best functionalities at ring A were unsubstituted
phenyl ring and phenyl ring substituted with Cl, Br, NO2, Me (at para-position) or OH (at ortho-
position). The best functionalities at ring B were unsubstituted phenyl ring, phenyl ring substituted

with OH or NO2 (at ortho-position), OMe (at para-position), and thiophene or quinoline
heterocycles. Any spacer between ring A and carbonyl group or between ring B and azomethine
group tend to abolish the activity.
Although ERRβ is expressed in low levels in adult tissues and its postnatal expression is highly
restricted, we have tested compounds 24, 27, 29, 33, and 42 against ERRβ to address the activity
of our synthesized agonists against this isoform (Table 4). All of our agonists activated ERRβ, but
there was no specific pattern corresponding to how it activate the other isoforms. Compounds 24
(Fig. 3) and 33 were more active toward ERRβ than ERRα and ERRγ. Unlike 24 and 33, compound
27 was more active toward ERRα than the two other isoforms (≈2-folds higher than ERRβ and
more than 4-folds higher than ERRγ). Compound 29 was more active toward ERRα and ERRγ
than ERRβ. The activity of 42 toward ERRβ was almost equal to ERRγ and 1.8-folds higher than
ERRα.
Table 4. In vitro ERRβ agonistic activities of 24, 27, 29, 33, and 42.
GSK4716 (6)
24
27
29
33
42
ERRβ EC50 (µM)
0.280
0.162
0.772
0.583
0.123
1.193
2.3. Molecular Modeling
We performed docking of selected compounds in the ligand binding pocket of ERRα and ERRγ.
Molecular docking in ERRγ was performed using the ERRγ bound agonist conformation (PDBID:
2GPP) [27]. There is no reported agonist bound conformation of ERRα and molecular docking of
GSK4716 (6) to the agonist pocket in apo ERRα X-ray structure (PDID: 1XB7) was not possible
initially because the pocket was occupied with the side chains of amino acid residues Phe328,
Arg372 and Glu331 and there was not sufficient room to accommodate the ligand [28]. The amino
acid residues in both pockets are similar and the agonist GSK4716 (6) is expected to bind in similar

manner in both receptors with slight differences. In order to obtain an ERRα structure suitable for
docking, we performed manual modeling of ERRα bound with the agonist GSK4716 (6) followed
by refinement of the ligand bound complex by performing local optimization using
Schrodinger/Prime [29]. The protein-ligand complex was validated using molecular dynamics
simulations for 500 ns (data will be published in due course). The refined ERRα-ligand complex
was used as the receptor in the docking calculations.
Molecular modelling of compounds 16, 24, 27, and 29 in the LBP of ERRα show that they have
favorable binding modes where they make π-π interactions with Phe328, which corresponds to
Ala272 in ERRγ. Moreover, the compounds tend to adopt a tilted confirmation to fit in the much
smaller binding pocket of ERRα and make π-π interactions with Phe328 that originally occupies
the available space in this binding pocket.
2.4. Gene Expression
ERRs are the major transcriptional regulators that regulate key genes involved in energy
metabolism within cells and tissues of high-energy demand. They regulate genes involved in
carbohydrate and energy metabolism, lipid synthesis and fatty acid oxidation, mitochondrial
activity and oxidative phosphorylation [20]. The agonistic effect of the best dual agonists (i.e. 16,
24, 27, 29, 35, 38, and 42) was validated by measuring the changes in gene expression of four ERR
target genes; PGC-1α, PGC-1β, CPT1α and PDK4 (Figure 7) in mouse myoblast cell line C2C12
at 1 µM of tested compounds. The peroxisome proliferator-activated receptor-γ coactivator 1α
(PGC-1α) and the peroxisome proliferator-activated receptor-γ coactivator 1β (PGC-1β) regulate
several metabolic pathways and their activity is of great importance to achieve maximal
mitochondrial biogenesis.

PGC1
PGC1
2.0
1.5
1.0
0.5
0.0
2.0
1.5
1.0
0.5
0.0
**
*
*
*
*
**
*
6
6
4
2
O
5
6
6
4
2
1
1
2
4
O
5
3
S
1
1
2
4
7
4
3
S
7
M
4
M
D
D
K
S
K
S
G
G
CPT1
PDK4
2.5
2.0
1.5
1.0
0.5
0.0
2.0
1.5
1.0
0.5
0.0
**
**
**
***
*
***
6
6
1
4
2
2
4
O
6
6
4
2
5
3
O
5
1
7
1
1
2
4
S
M
3
S
7
4
4
M
D
K
D
K
S
S
G
G
Figure 7. Normalized mRNA expression of PGC-1α, PGC-1β, CPT1α and PDK4 with 16, 24,
35, and 42 in muscle cells.
Compounds 24 and 42 were found to increase the mRNA levels of both PGC-1α and PGC-1β,
while compound 35 increased the mRNA levels of PGC-1α only. Carnitine palmitoyltransferase
1α (CPT1α) catalyzes the first step of long-chain fatty acid import into mitochondria, and its
activity is believed to be rate limiting for β-oxidation of fatty acids. Compound 16 significantly
increased the expression of CPT1α. The pyruvate dehydrogenase kinase isozyme 4 (PDK4) play
an important role in energy homeostasis and catalyzes the conversion of pyruvate to CoA. The
transcription of PDK4 is controlled by both ERRα and ERRγ in hepatic cells and muscle.
Compounds 16, 24, 35, and 42 potently upregulated PDK4 transcription.
2.5. Cytotoxicity assay
Cell viability was assessed using FITC Annexin V/ Dead cell Apoptosis kit (Invitrogen) for flow
cytometry according to the manufacturer's manual. Compounds 24, 27, 29, and 42 were tested at
1 and 10 µM. No signs of toxicity were observed for active compounds (Fig. 8).

Viability
20
15
10
5
0
O
M
M
M
M
µ
M
M
M
M
S
µ
µ
µ
µ
µ
µ
µ
0
1
0
1
0
1
0
1
M
1
1
1
1
2
4
9
7
2
D
4
2
2
4
2
9
7
2
4
2
2
Figure 8. Cell viability of 24, 27, 29, and 42.
2.6. In Silico ADME Evaluation
In order to evaluate the drug-likeness and to to predict the absorption, distribution, metabolism,
elimination and toxicity (ADMET) of some pf the best identified modulators, we have calculated
a set of molecular descriptors using QikProp program [30]. All evaluated compounds complied
with Lipinski’s Rule of Five and showed the desired molecular weight (MW), lipophilicity (log
P), number of hydrogen bond donors (HBD) and number of hydrogen bond acceptors (HBA)
(Table 5). Compliance with Lipinski’s Rule of Five indicate that these molecules have drug-like
properties [31]. These compounds were also found to comply with Jorgensen’s rule of 3 (LogSwat,
BIPCaco-2, and number of primary metabolites) and are predicted to have orally bioavailability
[32,33]. These compounds are predicted to have good aqueous solubility, cell permeability, and
oral bioavailability (Table 5). The calculated apparent MDCK cell permeability, which mimic the
BBB further suggests that these compounds have good cell permeability. Moreover, the estimated
plasma-protein binding to human serum albumin (logKHSA) of these compounds are within the
recommended range, which means that these compounds are predicted to circulate freely within
the blood stream and can achieve high efficiency.
Table 5. Calculated molecular descriptors for prediction of ADME properties for 16, 24, 27, 29,
35, and 42.

Descriptors/Properties
aMW (Da)
bHBD
cHBA
dPSA
eLogP o/w fLogSwat
(moles/liter)
-2.0 to 6.5 -6.5 to 0.5
Range for 95% of known
Drugs
16
24
27
29
35
42
130.0 to
725.0
269.26
238.29
256.26
254.29
333.18
230.28
g%Human
oral
0 to 6
2.0 to 20.0 7.0 to
200.0
3.5
2.5
3
3.25
3.25
2.5
iBIPCaco-2
(nm/sec)
1
1
2
2
1
1
93.04
49.78
89.93
68.23
57.98
48.35
jMDCK
2.81
-3.93
3.85
-4.46
2.52
-3.51
3.08
-4.24
4.06
-4.95
2.90
k#metab
-3.34
lLogKHSA
Properties
hLog
B/B
absorption
<25% is
poor
Range for 95% of known
Drugs
-3.0 to
1.2
<25 poor,
>500 great >500 great
<25 poor,
1.0 to 8.0
-1.5 to 1.5
16
24
27
29
35
42
88
100
87
100
100
100
-1.26
-0.34
-1.25
-0.67
-0.23
-0.29
319
2230
353
1170
2230
1891
144
1177
160
586
3120
1330
1
1
2
2
1
1
0.13
0.34
0.02
0.11
0.34
0.05
aMolar weight; bNumber of hydrogen bonds donated by the molecule; cNumber of hydrogen bonds accepted by the
molecule, dPolar surface area;eLogarithm of partitioning coefficient between n-octanol and water phases; fLogarithm
of aqueous solubility; gPredicted human oral absorption on a 0 -100 % scale, based on a multiple linear regression
model; hLogarithm of predicted blood/brain barrier partition coefficient; iPredicted apparent Caco-2 cell membrane
j
k
permeability in Boehringer–Ingelheim scale; Predicted apparent MDCK cell permeability; Number of likely
metabolic reactions; lLogarithm of predicted binding constant to human serum albumin.
3. Conclusions
We were able to modulate the estrogen-related receptors subtype selectivity by converting an
ERRβ/γ specific agonist, GSK4716 (6), into ERRα/β/γ pan agonists. Molecular modeling shows
that the new pan agonists have favorable binding modes in both isoforms but have different
conformation in the LBP of ERRα. These ligands induce a new conformation in the cavity of
ERRα where Phe328 is slightly displaced to accommodate the ligand and make π-π interactions
with one of the two phenyl rings. The identified ligands increased the expression of ERR target
genes, PGC-1α, PGC-1β, CPT1α and PDK4 in mouse myoblast cell line C2C12. Compounds 24,
27, 29, and 42 did not show signs of toxicity when tested at 1 and 10 µM. We identified the
important structural requirements to enhance the selectivity toward ERRα and our future goal is to

develop ERRα specific agonists to be used as chemical tools to validate ERRα as a putative target
for treatment of neurodegenerative diseases like Alzheimer Disease.
4. Experimental
4.1. Chemistry
All starting materials were purchased from commercial suppliers and used without further
purification. The purities of the final compounds were characterized by high-performance liquid
chromatography (LC/MS) using a gradient elution program (Ascentis Express Peptide C18
column, acetonitrile/water 5/95/95/5, 5 min, 0.05%trifluoracetic acid) and UV detection (254 nM).
The purities of final compounds were 95% or greater. NMR spectra was recorded on a Bruker
NMR 400 MHz Avance III spectrometer operating at 400 MHz for 1H NMR and 100 MHz for 13C
NMR. Chemical shifts are given in part per million (ppm) relative to the deuterated solvent residual
peak, coupling constants J are given in Hertz.
4.1.1. General procedure for the preparation of hydrazides 8
Carboxylic acids 1a-g (5 mmol) were heated under reflux in methanol (5 mL) for 2 h in presence
of conc. H2SO4 (catalytic amount) with continuous stirring. The reaction was monitored by TLC
till the acids were fully converted to the corresponding esters [Eluent: EtOAc/Hexanes (1:4)]. The
reaction mixture was allowed to cool down to room temperature and hydrazine monohydrate 80%
(20 mmol, 0.96 mL) was added slowly in an ice bath. The reaction was then warmed to room
temperature and heated under reflux for another 1-2 h and followed by TLC till formation of
hydrazide. The reaction mixture was kept at refrigerator till the product precipitated. All
hydrazides were isolated in quantitative yields and in a pure form. Their melting points were in
full agreement with the literature melting points of the same compounds.

4.1.2. General procedure for the preparation of hydrazones 10-46
A mixture of the appropriate hydrazides 8 (1 mmol) and the appropriate aldehyde 4a-j (1.1 mmol)
was refluxed in ethanol until the condensation was complete (monitored by TLC, 1-3 hours). The
reaction was allowed to cool to room temperature and the formed precipitate was filtered and
washed with cold ethanol. The purity of most compounds was ≈98% and in few cases the obtained
solid was recrystallized from ethanol to give the desired product. All acyl hydraones reported here
are known compounds their identity were confirmed by comparing their melting points and mass
to literature (See supporting information). 1H and 13C NMR spectra of most active compounds are
described here.
4.1.2.1. (E)-N'-(2-Nitrobenzylidene)benzohydrazide (16) [34]. Pale yellow crystals, mp 222-224
°C, 89% yield. 1H NMR (400 MHz, DMSO-d6) δ 12.21 (s, 1H), 8.88 (s, 1H), 8.14 (d, J = 7.3 Hz,
1H), 8.08 (d, J = 8.1 Hz, 1H), 7.83 (d, J = 7.2 Hz, 2H), 7.83 (t, J = 7.3 Hz, 1H), 7.68 (t, J = 7.6 Hz,
13
1H), 7.63 – 7.50 (m, 3H); C NMR (101 MHz, DMSO-d6) δ 163.3, 148.3, 142.9, 133.7, 133.0,
132.0, 130.7, 128.8, 128.5, 127.9, 127.7, 124.7. LC/MS m/z: 270 [M + H]+.
4.1.2.2. (E)-N'-Benzylidene-4-methylbenzohydrazide (24) [35]. White needles, mp 235–236 °C,
90% yield. 1H NMR (400 MHz, DMSO-d6) δ 11.78 (s, 1H), 8.47 (s, 1H), 7.84 (d, J = 7.7 Hz, 2H),
13
7.73 (d, J = 6.3 Hz, 2H), 7.50 – 7.41 (m, 3H), 7.33 (d, J = 7.9 Hz, 2H), 2.38 (s, 3H); C NMR
(101 MHz, DMSO-d6) δ 162.9, 147.5, 141.8, 134.4, 130.5, 129.0, 128.8, 127.6, 127.0, 21.0.
LC/MS m/z: 239 [M + H]+.
4.1.2.3. (E)-2-Hydroxy-N'-(2-hydroxybenzylidene)benzohydrazide (27)[36]. White microcrystals,
mp 265-267 °C, 85% yield. 1H NMR (400 MHz, DMSO-d6) δ 12.03 (s, 1H), 11.78 (s, 1H), 11.20
(s, 1H), 8.69 (s, 1H), 7.90 (d, J = 7.1 Hz, 1H), 7.57 (d, J = 6.9 Hz, 1H), 7.46 (t, J = 7.7 Hz, 1H),
13
7.32 (t, J = 8.2 Hz, 1H), 7.05 – 6.72 (m, 4H); C NMR (101 MHz, DMSO-d6) δ 164.5, 159.0,

157.50, 149.0, 134.0, 131.6, 129.5, 128.6, 119.4, 119.0, 118.6, 117.3, 116.5, 115.6. LC/MS m/z:
257 [M + H]+.
4.1.2.4. (E)-N'-(2-Hydroxybenzylidene)-2-methylbenzohydrazide (29) [37]. White crystals, mp
163-165 °C, 50% yield. 1H NMR (400 MHz, DMSO, two rotamers 4:1) δ 11.99 (s, 0.8H), 11.87
(s, 0.2H), 11.22 (s, 0.8H), 9.84 (s, 0.2H),8.50 (s, 0.8H), 8.27 (s, 0.2H), 7.53 – 7.26 (m, 6H), 6.95
– 6.76 (m, 2H), 2.40 (s, 2.4H), 2.25 (s, 0.6H); 13C NMR (101 MHz, DMSO) δ 164.9, 157.4, 147.9,
136.1, 134.6, 131.4, 130.7, 130.2, 129.5, 127.6, 126.7, 125.7, 125.5, 119.4, 118.6, 116.4, 19.4.
LC/MS m/z: 255 [M + H]+.
4.1.2.5. (E)-N'-(2-Hydroxybenzylidene)-4-methylbenzohydrazide (33) [38]. Yellow solid, mp 199–
201 °C, 83% yield. 1H NMR (400 MHz, DMSO-d6) δ 12.01 (s, 1H), 11.30 (s, 1H), 8.60 (s, 1H),
7.82 (d, J = 7.9 Hz, 2H), 7.50 (d, J = 7.5 Hz, 1H), 7.35-7.25 (m, 3H), 6.89 (t, J = 8.1 Hz, 2H), 3.30
(s, 3H); 13C NMR (101 MHz, DMSO-d6) δ 162.6, 157.5, 148.1, 142.1, 131.3, 129.9, 129.6, 129.1,
127.7, 119.3, 118.7, 116.4, 21.1. LC/MS m/z: 255 [M + H]+.
4.1.2.6. (E)-4-Bromo-N'-(4-methoxybenzylidene)benzohydrazide (35) [39]. White microcrystals,
mp 129-131 °C, 89% yield. 1H NMR (400 MHz, DMSO-d6) δ 11.78 (s, 1H), 8.39 (s, 1H), 7.86 (d,
J = 8.0 Hz, 2H), 7.71 (d, J = 8.0 Hz, 2H), 7.68 (d, J = 8.3 Hz, 2H), 7.02 (d, J = 8.3 Hz, 2H), 3.81
(s, 3H); 13C NMR (101 MHz, DMSO-d6) δ 162.0, 160.9, 148.0, 132.6, 131.5, 129.7, 128.8, 126.8,
125.4, 114.4, 55.3. LC/MS m/z: 333 [M+H+].
4.1.2.7. (E)-4-Nitro-N'-(2-nitrobenzylidene)benzohydrazide (38) [40]. White solid, mp 260-262
°C, 91% yield. 1H NMR (400 MHz, DMSO-d6) δ 12.47 (s, 1H), 8.90 (s, 1H), 8.38 (d, J = 8.6 Hz,
2H), 8.28 – 8.05 (m, 4H), 7.84 (t, J = 7.6 Hz, 1H), 7.71 (t, J = 7.6 Hz, 1H); 13C NMR (101 MHz,

DMSO-d6) δ 161.7, 149.4, 148.3, 144.2, 138.6, 133.8, 131.0, 129.3, 128.5, 128.1, 124.7, 123.7.
LC/MS m/z: 315 [M + H]+.
4.1.2.8. (E)-N'-(Thiophen-2-ylmethylene)benzohydrazide (42) [37]. White crystals, mp 207–209
°C, 91% yield. 1H NMR (400 MHz, DMSO-d6) δ 11.77 (s, 1H), 8.64 (s, 1H), 7.86 (d, J = 7.5 Hz,
2H), 7.79 (d, J = 7.1 Hz, 1H), 7.57 – 7.38 (m, 4H), 7.11 (dd, J = 5.0, 3.6 Hz, 1H); 13C NMR (101
MHz, DMSO-d6) δ 166.3, 143.4, 133.8, 131.4, 129.4, 128.9, 128.7, 128.3, 128.0, 127.4. LC/MS
m/z: 231 [M + H]+.
4.2. Molecular Modeling
Docking in ERRα and ERRγ was performed using the induced-fit docking module in Schrodinger
to represent ligand-induced conformational changes [41-43]. Each ligand was initially docked
using a softened potential (van der Waals radii scaling) of the ligand and the active site side chains.
The scaling factors to soften the potentials of the receptors and ligands were set to 0.5 and the
maximum number of poses was set to 20. The side chains of the residue within 5.0 A of the ligand
are refined along with the docked ligand with Prime using a continuum solvation based molecular
mechanics model, via rotamer-based library optimizations of the amino acids side chain
conformations [29]. The ligand is re-docked, using the standard precision algorithm implemented
in Glide into the induced-fit receptor structures and re-minimized, and the final poses are rank
ordered by the scoring function given by a linear combination of the Prime energy and GlideScore
referred to as the IFDScore [42,44]. Images were made using Discovery Studio [45].
4.3. Cell culture:
C2C12 cells, mouse myoblast cell line, were maintained in Dulbecco's modified Eagle's medium
(DMEM) supplemented with 10% FBS and 1% L-Glutamine. The C2C12 cells were plated in 6

well dishes (Corning) with two and a half million cells per well and grew for 24 hours until
confluence is reached. The cells were then treated with either the drug in DMSO, or DMSO as the
control at a concentration of 1μM or 10 μM. All groups were tested in triplicates.
4.4. Cell Viability assay
C2C12 cells were treated for 24h with 1μM or 10μM of different compounds. After treatment,
cells were harvested and stained for live/dead cells using FITC Annexin V/ Dead Apoptosis Kit
with FITC annexin V and PI, for Flow Cytometry (Invitrogen) according the manufacturer
instruction. Cells were then analyzed by flow cytometry using BD Accuri c6 flow cytometer (BD
Biosciences).
4.5. Gene expression
Cells were treated for 24h with 10μM of different compounds. After treatment, cells were
harvested and total RNA was extracted using RNeasy Mini Kit according the manufacturer
instruction. Total RNA were quantified using NanoDrop and 1μg of total RNA was reverse
transcript using qScript cDNA Synthesis Kit (QuantaBio) according the manufacturer instruction.
Gene expression was assessed by qPCR using Sybr Select Master Mix (Applied Biotechnologies).
All samples were run in duplicates and the analysis was completed by determining ΔΔCt values.
The reference gene used was 36b4, a ribosomal protein gene.
Acknowledgments
We would like to thank Support & Development of Scientific Research Center (SDSRC) at Benha
University for financial support for MS and SG. This work was partially supported by the National
Institute on Aging of the National Institutes of Health under Award Number R21AG065657 (to

B.E.) and National Institute of Arthritis and Musculoskeletal and Skin Diseases under Award
Number R01AR069280 (to T.B.).

References
[1]
[2]
V. Giguère, N. Yang, P. Segui, R.M. Evans, Identification of a new class of steroid
hormone receptors, Nature. 331 (1988) 91–94. doi:10.1038/331091a0.
F. Chen, Q. Zhang, T. McDonald, M.J. Davidoff, W. Bailey, C. Bai, Q. Liu, C.T. Caskey,
Identification of two hERR2-related novel nuclear receptors utilizing bioinformatics and
inverse PCR, Gene. 228 (1999) 101–109. doi:10.1016/S0378-1119(98)00619-2.
[3]
J.D. Eudy, S. Yao, M.D. Weston, M. Ma-Edmonds, C.B. Talmadge, J.J. Cheng, W.J.
Kimberling, J. Sumegi, Isolation of a gene encoding a novel member of the nuclear
receptor superfamily from the critical region of usher syndrome type IIa at 1q41,
Genomics. 50 (1998) 382–384. doi:10.1006/geno.1998.5345.
[4]
[5]
H. Hong, L. Yang, M.R. Stallcup, Hormone-independent transcriptional activation and
coactivator binding by novel orphan nuclear receptor ERR3, J. Biol. Chem. 274 (1999)
22618–22626. doi:10.1074/jbc.274.32.22618.
D.J. Heard, H. Vissing, P.L. Norby, J. Holloway, Human ERRγ, a Third Member of the
Estrogen Receptor-Related Receptor (ERR) Subfamily of Orphan Nuclear Receptors:
Tissue-Specific Isoforms Are Expressed during Development and in the Adult, Mol.
Endocrinol. 14 (2000) 382–392. doi:10.1210/mend.14.3.0431.
[6]
J.-M. Schlaeppi, B. Fournier, F. Bitsch, M. Geiser, A. Schilb, J. Kallen, V. Riou, A.
Strauss, I. Filipuzzi, A. Graham, Evidence for Ligand-independent Transcriptional
Activation of the Human Estrogen-related Receptor α (ERRα), J. Biol. Chem. 279 (2004)
49330–49337. doi:10.1074/jbc.m407999200.
[7]
[8]
J. Wang, F. Fang, Z. Huang, Y. Wang, C. Wong, Kaempferol is an estrogen-related
receptor α and γ inverse agonist, FEBS Lett. 583 (2009) 643–647.
doi:10.1016/j.febslet.2009.01.030.
W.C. Stevens, D.W. Sapp, B.B. Busch, S. Zhou, R. Mohan, R. Martin, R.A. Horlick, P.
Ordentlich, Identification of a Selective Inverse Agonist for the Orphan Nuclear Receptor
Estrogen-Related Receptor α, J. Med. Chem. 47 (2004) 5593–5596.
doi:10.1021/jm049334f.
[9]
N. Kersual, D. Chalbos, J.-M. Vanacker, S. Bianco, O. Lanvin, Potentiation of ICI182,780
(Fulvestrant)-induced Estrogen Receptor-α Degradation by the Estrogen Receptor-related
Receptor-α Inverse Agonist XCT790, J. Biol. Chem. 282 (2007) 28328–28334.
doi:10.1074/jbc.m704295200.
[10] J. Wang, Y. Wang, C. Wong, Oestrogen-related receptor alpha inverse agonist XCT-790
arrests A549 lung cancer cell population growth by inducing mitochondrial reactive
oxygen species production, Cell Prolif. 43 (2010) 103–113. doi:10.1111/j.1365-
2184.2009.00659.x.
[11] F. Wu, J. Wang, Y. Wang, T.T. Kwok, S.K. Kong, C. Wong, Estrogen-related receptor α
(ERRα) inverse agonist XCT-790 induces cell death in chemotherapeutic resistant cancer
cells, Chem. Biol. Interact. 181 (2009) 236–242. doi:10.1016/j.cbi.2009.05.008.

[12] L. Chen, C. Wong, Estrogen-Related Receptor α Inverse Agonist Enhances Basal Glucose
Uptake in Myotubes through Reactive Oxygen Species, Biol. Pharm. Bull. 32 (2009)
1199–1203. doi:10.1248/bpb.32.1199.
[13] Y. Nie, C. Wong, Suppressing the activity of ERRα in 3T3-L1 adipocytes reduces
mitochondrial biogenesis but enhances glycolysis and basal glucose uptake, J. Cell. Mol.
Med. 13 (2009) 3051–3060. doi:10.1111/j.1582-4934.2008.00382.x.
[14] W.J. Zuercher, R.A. Stein, S.M. Hyatt, T.M. Willson, L.A. Orband-Miller, E.L. Lockamy,
D.P. McDonnell, A.B. Miller, On the Intractability of Estrogen-Related Receptor α as a
Target for Activation by Small Molecules, J. Med. Chem. 50 (2007) 6722–6724.
doi:10.1021/jm7012387.
[15] M. Suetsugi, L. Su, K. Karlsberg, Y.-C. Yuan, S. Chen, Flavone and isoflavone
phytoestrogens are agonists of estrogen-related receptors., Mol. Cancer Res. 1 (2003)
981–91. http://www.ncbi.nlm.nih.gov/pubmed/14638870.
[16] L. Peng, X. Gao, L. Duan, X. Ren, D. Wu, K. Ding, Identification of pyrido[1,2-
a]pyrimidine-4-ones as new molecules improving the transcriptional functions of
estrogen-related receptor α, J. Med. Chem. 54 (2011) 7729–7733. doi:10.1021/jm200976s.
[17] K. Griffett, L.A. Solt, B.-E.M. El-Gendy, T.M. Kamenecka, T.P. Burris, A liver-selective
LXR inverse agonist that suppresses hepatic steatosis, ACS Chem. Biol. 8 (2013) 559–
567. doi:10.1021/cb300541g.
[18] C.A. Flaveny, K. Griffett, B.-E.M. El-Gendy, M. Kazantzis, M. Sengupta, A.L. Amelio,
A. Chatterjee, J. Walker, L.A. Solt, T.M. Kamenecka, T.P. Burris, Broad Anti-tumor
Activity of a Small Molecule that Selectively Targets the Warburg Effect and
Lipogenesis, Cancer Cell. 28 (2015) 42–56. doi:10.1016/j.ccell.2015.05.007.
[19] B.-E.M. El-Gendy, S.S. Goher, L.S. Hegazy, M.M.H. Arief, T.P. Burris, Recent Advances
in the Medicinal Chemistry of Liver X Receptors, J. Med. Chem. 61 (2018) 10935–10956.
doi:10.1021/acs.jmedchem.8b00045.
[20] S.S. Goher, K. Griffett, L. Hegazy, M. Elagawany, M.M.H. Arief, A. Avdagic, S.
Banerjee, T.P. Burris, B. Elgendy, Development of novel liver X receptor modulators
based on a 1,2,4-triazole scaffold, Bioorganic Med. Chem. Lett. 29 (2019) 449–453.
doi:10.1016/j.bmcl.2018.12.025.
[21] P.M. Khan, B.E.D.M. El-Gendy, N. Kumar, R. Garcia-Ordonez, L. Lin, C.H. Ruiz, M.D.
Cameron, P.R. Griffin, T.M. Kamenecka, Small molecule amides as potent ROR-γ
selective modulators, Bioorganic Med. Chem. Lett. 23 (2013) 532–536.
doi:10.1016/j.bmcl.2012.11.025.
[22] S. Banerjee, Y. Wang, L.A. Solt, K. Griffett, M. Kazantzis, A. Amador, B.M. El-Gendy,
S. Huitron-Resendiz, A.J. Roberts, Y. Shin, T.M. Kamenecka, T.P. Burris,
Pharmacological targeting of the mammalian clock regulates sleep architecture and
emotional behaviour, Nat. Commun. 5 (2014). doi:10.1038/ncomms6759.
[23] T.M. Willson, J.L. Collins, A.B. Miller, W.J. Zuercher, L.A. Orband-Miller, E.Y.H. Chao,
S. Gaillard, D.P. McDonnell, D.G. Jones, B.G. Shearer, Identification and