11-β-hydroxysterols as possible endogenous stimulators of mitochondrial biogenesis as inferred from epicatechin molecular mimicry

Article abstract of DOI:10.1016/j.phrs.2019.104540

Currently, there is great interest in identifying endogenous (i.e. physiological) stimulators of mitochondrial biogenesis (MB), in particular, those that may mediate the effects of exercise. The molecular size of the cacao flavanols (epicatechin and catechin) highly resembles that of sterols and epicatechin has been reported to activate cells surface receptors leading to the stimulation of MB in endothelial and skeletal muscle cells translating into enhanced exercise capacity. We therefore hypothesize, that epicatechin may be acting as a structural mimic of an as yet unknown sterol capable of stimulating MB. We developed a new synthetic process for obtaining enantiomerically pure preparations of (-)-epicatechin and (+)-epicatechin. Applying spatial analytics and molecular modeling, we found that the two isoforms of epicatechin, (-) and (+), have a structural resemblance to 11-β-hydroxypregnenolone, a sterol with no previously described biological activity. As reported in this proof-of-concept study performed in primary cultures of endothelial and muscle cells, 11-β-hydroxypregnenolone is one of the most potent inducers of MB as significant activity can be detected at femtomolar levels. The relative potency of (-)/(+)-epicatechin isoforms and on inducing MB correlates with their degree of spatial homology towards the 11-β-hydroxypregnenolone. On the basis of these results, the detailed in vivo characterization of the potential for these sterols to act as endogenous modulators of MB is warranted.

Full text of DOI:10.1016/j.phrs.2019.104540

Journal Pre-proof
11-␤-hydroxysterols as possible endogenous stimulators of mitochondrial
biogenesis as inferred from epicatechin molecular mimicry
Sundeep Dugar, Francisco Villarreal, Frank H. Hollinger, Dinesh
Mahajan, Israel Ramirez-Sanchez, Aldo Moreno-Ulloa, Guillermo
Ceballos, George Schreiner
PII:
S1043-6618(19)30761-3
DOI:
https://doi.org/10.1016/j.phrs.2019.104540
YPHRS 104540
Reference:
To appear in:
Pharmacological Research
Received Date:
Revised Date:
Accepted Date:
30 April 2019
5 November 2019
9 November 2019
Please cite this article as: Dugar S, Villarreal F, Hollinger FH, Mahajan D, Ramirez-Sanchez I,
Moreno-Ulloa A, Ceballos G, Schreiner G, 11-␤-hydroxysterols as possible endogenous
stimulators of mitochondrial biogenesis as inferred from epicatechin molecular mimicry,
Pharmacological Research (2019), doi: https://doi.org/10.1016/j.phrs.2019.104540
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.
© 2019 Published by Elsevier.

YPHRS_2019_722_R2
11--hydroxysterols as possible endogenous stimulators of
mitochondrial biogenesis as inferred from epicatechin molecular
mimicry
Sundeep Dugar*,1,2 Francisco Villarreal,3,4 Frank H. Hollinger,2 Dinesh Mahajan,5 Israel Ramirez-
Sanchez, 3,6 Aldo Moreno-Ulloa, 3,7 Guillermo Ceballos,6 and George Schreiner 1
1 Cardero Therapeutics, Los Altos Hills, CA, USA
2 Sphaera Pharma Pte. Ltd., Singapore
3 University of California, San Diego
4 VA San Diego Healthcare System
5 Sphaera Pharma Pvt. Ltd., IMT Manesar, India
6 Escuela Superior de Medicina, Instituto Politécnico Nacional, Mexico
7 Departamento de Innovación Biomédica, Centro de Investigación Científica y de Educación
Superior de Ensenada (CICESE), Baja California, México
*Corresponding Author:
Dr. Sundeep Dugar
Cardero Therapeutics, Inc.
12774 Leander Drive
Los Altos Hills, CA
dugar.s@carderorx.com
AUTHOR INFORMATION
1

Corresponding Author
Sundeep Dugar, 12774 Leander Drive, Los Altos Hills, CA
Graphical abstract
ABSTRACT
Currently, there is great interest in identifying endogenous (i.e. physiological) stimulators of
mitochondrial biogenesis (MB), in particular, those that may mediate the effects of exercise. The
molecular structure of the cacao flavanols (epicatechin and catechin) highly resembles that of
sterols and epicatechin, has been reported to activate cells surface receptors leading to
the stimulation of MB in endothelial and skeletal muscle cells translating into enhanced exercise
2

capacity. We therefore hypothesize, that epicatechin may be active as a structural mimic of an as
yet unknown sterol capable of stimulating MB. We developed a new synthetic process for
obtaining enantiomerically pure preparations of (-)-epicatechin and (+)-epicatechin. Applying
spatial analytics and molecular modeling, we found that the two isoforms of epicatechin, (-)
and (+), structurally resemble 11--hydroxypregnenolone, a sterol with no previously described
biological activity. As reported in this proof-of-concept study performed in primary cultures of
endothelial and muscle cells, 11--hydroxypregnenolone is one of the most potent inducers of MB
as significant activity can be detected at femtomolar levels. The relative potency of (-)/(+)-
epicatechin isoforms and (+)-catechin on inducing MB correlates with their spatial homology
degree towards the 11--hydroxysterol structure. On the basis of these results, the detailed in
vivo characterization of the potential for these sterols to act as endogenous modulators of MB
is warranted.
Keywords: epicatechin, 11--hydroxypregnenolone, mitochondrial biogenesis, steroids
INTRODUCTION
Our published work has demonstrated that the cacao flavanol (-)-epicatechin (Figure 1) triggers
cellular responses including mitochondrial biogenesis (MB) via the activation of cell surface
receptors (1-4). (-)-Epicatechin is a plant metabolite that belongs to the family of flavonoids
(specifically to subclass of flavanols) and in principle, “specific” receptors for this molecule should
not be present in mammalian cells, unless an endogenous entity responsible for their
physiological activation were to exist and be able to activate downstream responses.
Recently, we reported that (-)-epicatechin is superior to other structurally similar flavonoids such
as (+)-catechin (Figure 1, a stereoisomer) in activating upstream signaling pathways associated
with MB (5). Additionally, the stereoisomer (+)-epicatechin was found to be substantially more
3

active than (-)-epicatechin in stimulating MB, as determined by the protein levels of electron
transport chain complexes II and IV, and mitochondria related endpoints both in cultured cells and
in treated mice (6, 7). These findings suggest highly selective molecular interactions between the
receptors and the flavanols.
Currently, there is great interest in identifying endogenous (i.e. physiological) stimulators of MB,
in particular, those that may mediate the effects of exercise. The molecular structure of flavanols
highly resembles that of sterols and this class of molecules is also known to activate various cells
surface receptors leading to a range of physiological responses (8-10).
Figure 1. Flavanol Structures
Figure 1. Structures of the flavanols used in this study.
We therefore hypothesize that (-)/(+)-epicatechins, are structural homologs and functional mimics
of a sterol molecule(s) with a hitherto endogenous capacity to activate MB.
The objective of this study was to implement molecular modeling/design studies of putative
physiological sterols, select and synthesize the best candidates and using endothelial and/or
muscle cells in culture, compare their capacity to active MB vs. (+)-catechin and (-)/(+)-
epicatechins.
4

MATERIALS AND METHODS
Chemical synthesis reactives
Unless otherwise noted, commercially available materials were used without further purification
for chemistry synthesis. Air sensitive reactions were carried out under nitrogen atmosphere.
Anhydrous solvents and all key starting materials were obtained from Sigma Aldrich. Thin layer
chromatography (TLC) was carried out using 0.2 mm Kieselgel F254 (Merck) silica plates and
compounds viewed under an ultraviolet lamp or by exposing to anisaldehyde solution. NMR
spectra were recorded on a Bruker 300 MHz spectrometer using CDCl3 or DMSO-D6 as solvent.
Chemical shifts (δ) are quoted in parts per million (ppm) relative to internal solvent reference.
Coupling constants are given in Hz and chemical shifts are reported in δ values in ppm. Data are
reported as followed: chemical shift, multiplicity (s = singlet, s br = broad singlet, d = doublet, t =
triplet, dd = double doublet, dt = double triplet, ddd = double double doublet, m = multiplet),
coupling constants (Hz) and integration.
Synthesis
Scheme 1. Synthesis of (+)/(-)-epicatechin from quercetin.
5

A novel short and commercially scalable synthesis of epicatechin was developed starting from a
natural flavonol (11), i.e. quercetin, which is found in abundance in nature. Benzylated quercetin
2 was synthesized from quercetin (1) in high yield (>85%) using benzyl bromide and potassium
carbonate. Treatment of 2 with lithium aluminum hydride in methyl-tert-butyl ether (MTBE) gave
a mixture of two regio-isomers of chromenes 3 with total conversion of 90% and in a ratio of 70:30
of -4H-chromene and -2H-chromene. This mixture of chromenes afforded racemic epicatechin
as the major product (75% conversion) when subjected to one pot hydrogenation/ hydrogenolysis
using H2, Pd/C. This procedure was also found suitable for a kilogram scale synthesis of racemic
epicatechin. The pure enantiomers of epicatechin were obtained using a chiral preparative HPLC
method using a CHIRALPAK® IC column with hexanes/ethanol as mobile in enantiomeric purity
of 99.6% for (-)-epicatechin and 99.8 % for (+)-epicatechin. Absolute configuration of the
enantiomers was assigned based on comparison with the retention time of the commercially
available natural (-)-epicatechin.
Scheme 2. 11--hydroxypregnenolone (11--OHP) was synthesized from commercially available
11--hydroxyprogesterone.
6

11--OHP was synthesized from commercially available 11--hydroxyprogesterone as shown in
Scheme 2. Treatment of 11--hydroxyprogesterone with CrO3 in acetic acid leads to formation
intermediate 6 which was subjected to a two-step protection for differential protection of 3-keto
and distal acetyl carbonyl leading to intermediate 8. The selective reduction of 8 with a sodium
borohydride at 0º C followed by a dilute sulfuric acid treatment provided reasonably pure form of
11-β-OHP, 10, as a white solid. This can be further purified using silica gel column
chromatography. The product was characterized by 1H NMR and LCMS spectroscopic
techniques. The 11--hydroxyl position of the pregnenolone can be assigned using characteristic
signal and shift of corresponding proton present at the 11 position of the steroid.
Synthesis of 2-[3, 4-bis(phenylmethoxy)phenyl]-3, 5, 7-tris (phenylmethoxy)chromen-4-one, 2
In a 3-neck, 1 L round bottomed flask fitted with a mechanical stirrer, quercetin dehydrate 1, (25
g) was added, followed by addition of acetone (80 mL) and N-methylpyrrolidone (40 mL). Solid
anhydrous potassium carbonate (81.5 g, 8 eq.) and tetra-butylammonium bromide (1.19 g, 0.05
eq.) was then added followed by benzyl chloride (52 mL, 6.2 eq.) under inert atmosphere, at 25-
30°C with constant but slow stirring to maintain effective mixing. The resulting reaction mixture
was refluxed at 75°C with constant stirring (180-200 RPM) for 24 hrs. After completion of the
reaction, the reaction mixture was cooled to room temperature. Ice cold water (500 mL) was slowly
added to the reaction mixture yielding a pale brown precipitate. The reaction mass was stirred for
an additional 30 min at room temperature followed by filtration using a Buchner funnel. The solid
product was washed with an additional 100 mL of ice-cold water. The wet cake was semi-dried
under vacuum. The semi-dried cake was treated with methanol (300 mL) and stirred for 5 min
7

followed by filtration. The solid mass was washed with an additional 100 mL of methanol and dried
under vacuum at ambient temperature to yield 2 (54.0 g, 87% yield) as an off-white solid.
Analytical Data: ESIMS: 753[M+1]; 1H NMR (300 MHz, CDCl3): δ (ppm) 7.72 (d, 1 H, Ar-H), 7.60-
7.62 (d, 1 H, Ar-H), 7.14-7.50 (m, 25 H, 5-Ar-H), 6.93 (d, 1 H, Ar-H), 6.52 (d, 1 H, Ar-H), 6.45 (d,
1 H, Ar-H), 5.24-5.28 (s, 4 H, 2 x CH2), 5.08-5.09 (s, 4 H, 2 x CH2), 4.96 (s, 2 H, CH2); 13C NMR
(75 MHz, CDCl3): δ 158.64, 157.93, 157.04, 153.89, 151.87, 150.99, 149.72, 148.89, 148.51,
148.18, 148.13, 137.69, 137.26, 137.16, 137.09, 136.95, 136.70, 136.62, 136.24, 133.47, 131.93,
128.49, 128.47, 128.43, 128.33, 128.25, 127.92, 127.77, 127.62, 127.54, 127.49, 127.43, 127.26,
127.15, 127.08, 126.40, 120.32, 120.19, 114.21, 114.14, 113.72, 105.14, 101.73, 95.36, 94.79,
94.31, 93.92, 89.34, 70.94, 70.25, 70.07, 69.93, 69.56, 22.
Synthesis of mixture of 2-[3,4-bis(phenylmethoxy)phenyl]-3,5,7-tris(phenylmethoxy)- 2H-1-
benzopyran and 2-[3,4-bis(phenylmethoxy)phenyl]-3,5,7-tris(phenylmethoxy)- 4H-1-benzopyran,
3
8

A 3-necked 1000 mL round bottomed flask, was charged with 2 (25 g, 1 eq) under an inert
atmosphere. MTBE (200 mL) was added as the solvent and the reaction mixture was stirred at
room temperature followed by addition of lithium aluminum hydride powder (1.32 g, 1.1 eq.) under
inert atmosphere at 25–30°C. The resulting mixture was stirred and heated to 55-65°C for 4h. The
reaction mixture was cooled to 0 – 5°C and quenched with slow addition of 0.1N HCl (500 mL, 20
vol.). Quenching of the reaction led to formation of an off-white solid mass. The liquid layer
(remaining solvent and water) was removed from the reaction mixture by simple decantation. The
remaining solid mass was washed with ice cold water (300 mL) by stirring for an additional 30
min. This slurry was further decanted leading to removal of the water from the flask, leaving behind
a crude solid product. The crude product was triturated with MeOH (200 mL). The resulting off
white solid was filtered using a Buchner funnel. The wet cake obtained was dried under vacuum
at room temperature to yield a mixture of 3a/3b (24 g, 98% yield) as an off-white solid.
Analytical Data: MS (ESI) m/z [M+1]: 739; 1H NMR (300 MHz, CDCl3): δ (ppm) 7.53-7.54 (d, 1 H,
J = 2.1 Hz), 7.26-7.46 (m, 25 H), 6.91-6.94 (d, 1 H, J = 9 Hz), 6.25-6.91 (2H, m), 5.190 (1 H , s),
4.982-5.133 (6 H, m), 4.735 (2 H, s), 3.62 (s, 2H). Ratio of 4H:2H is 65:35 based on 1H NMR.
1
NMR data of pure 3a ; H NMR (300 MHz, CDCl3): δ (ppm) 7.29-7.44 (m, 24 H), 7.21-7.26 (m,
2H), 7.02 (d, 1 H, J = 1.8 Hz), 6.91-6.94 (dd, 1 H, J = 6.6, 1.89 Hz), 6.83(d, 1 H, J = 8.4 Hz), 6.22
(d, 1 H, J = 2.1 Hz), 6.13 (d, 1 H, J = 2.1 Hz), 6.01 (s, 1H), 5.6 (s, 1H), 5.10 (s, 2H), 5.06 (s, 1H),
5.04 (s, 2H), 4.93 (s, 2H), 4.87 (d, 2H, J = 16.2 ); 13C NMR (75 MHz, CDCl3): δ 162.91, 158.10,
154.16, 151.01, 149.89, 148.97, 148.22, 137.30, 137.19, 137.12, 136.88, 136.34, 132.13, 128.57,
128.49, 128.44, 128.42, 127.97, 127.94, 127.76, 127.68, 127.63, 127.53, 127.41, 127.38, 127.21,
120.41, 114.47, 113.83, 105.21, 95.47, 94.50, 89.38, 77.23, 71.23, 71.09, 70.44, 70.06, 69.63.
NMR data of pure 3b; 1H NMR (300 MHz, CDCl3): δ (ppm) 7.53-7.54 (d, 1 H, J = 2.1 Hz), 7.26-
7.46 (m, 25 H), 6.91-6.94 (d, 1 H, J = 9 Hz), 6.25-6.91 (m, 2H), 5.19 (s, 1H), 4.98-5.13 (m, 6H),
4.73 (s, 2H), 3.62 (s, 2H), 13C NMR (75 MHz, CDCl3): δ 158.78, 157.18, 148.27, 137.40, 137.30,
9

136.84, 133.61, 128.63, 128.47, 128.39, 128.06, 127.91, 127.76, 127.68, 127.57, 127.40, 127.28,
127.22, 126.54, 120.46, 114.29, 101.89, 94.94, 94.06, 71.09, 70.22, 22.54.
Synthesis of trans-(+/-)-2-(3,4-Dihydroxyphenyl)chroman-3,5,7-triol, 4
A mixture of compounds 3a/3b (5g) in glacial acetic acid (50 mL) was hydrogenated in a Parr-
shaker in the presence of 10% Pd(OH)2 on activated carbon (2.00 g), under hydrogen pressure
(70 psi) at ambient temperature for 24 hrs. The hydrogen pressure was maintained at 70 psi for
the duration of the reaction. After 24 hrs, the reaction completion was confirmed by TLC (10%
MeOH in DCM) and the reaction mixture filtered through a Celite bed. The reaction flask and celite
was washed with water (100mL) and the combined filtrate was concentrated under vacuum on a
rotary evaporator to dryness. The crude product mixture was dissolved in water (15mL), the
resulting solution was refrigerated and the product crystalized as a brown solid. The product was
filtered and washed with a small amount of ice cold water (5mL) and dried under vacuum to yield
4 (0.8 g, 41% yield).
Analytical Data: MS (ESI) m/z [M+1]: 291; 1H NMR (300 MHz, DMSO-d6): δ (ppm) 9.11 (s, 1H),
8.90 (s, 1H), 8.81 (s, 1H), 8.72 (s, 1H), 6.88 (s, 1H), 6.65 (s, 2H), 5.89 (d, 1H, J=2.1 Hz), 5.70 (d,
1H, J=2.4 Hz), 4.72 (s, 1H), 4.66 (d, 1H, J=4.5 Hz), 3.98 (m, 1H), 2.63-2.71 (dd, 1H, J=4.2, 16.2)
2.44-2.49 (dd, 1H, J=3.3, 15.6); 13C NMR (75 MHz, CDCl3) δ 156.6, 156.2, 155.9, 144.5, 144.3,
130.8, 117.9, 114.4, 113.8, 98.6, 94.9, 94.4, 78.4, 66.0, 27.8.
Chiral resolution of trans-(+/-)-2-(3, 4-Dihydroxyphenyl) chroman-3, 5, 7-triol, 4
10

The racemic mixture 4 (0.200 g) was dissolved in methanol and separated on a preparative HPLC
using a CHIRAL PAK®IC (250 X 20) mm column at ambient temperature, using hexanes/ethanol
(60/40) as the mobile phase with a flow rate of 18 mL/ minute. (+)-(2S,3S)-2-(3,4-
Dihydroxyphenyl)chroman-3,5,7-triol (0.085g, 85% yield, 99.6% ee) eluted first at 4.7 minute
followed by (-)-(2R,3R)-2-(3,4-Dihydroxyphenyl)chroman-3,5,7-triol (0.084 g, 84% yield, 99.8%
ee).
Synthesis of 17-acetyl-10,13-dimethyl-1, 6, 7, 8, 9, 10, 12, 13, 14, 15, 16, 17-dodecahydro-2H-
cyclopenta[a]-Phenanthrene-3,11-dione, 6
A solution of chromium trioxide (3.3 g, 33.3 mmol) in water (4 mL) was added dropwise to a
solution of 5 (10 g, 30.3 mmol) in acetic acid (300 ml). The reaction was stirred for 1.5 hr and
quenched by the addition of methanol (30 mL), stirring for 0.5 hr. The mixture was concentrated
and dissolved in ethyl acetate, washed with saturated sodium bicarbonate and brine solution,
dried over sodium sulfate and concentrated to give 6 as an off white solid (9.0 g, 90% yield). MS
(ESI) m/z (M+1): 329; 1H NMR (300 MHz, CDCl3): δ 5.73 (s, 1H), 2.77 (ddd, 2H, J = 14.0, 5.2, 3.4
Hz), 2.66 (d, 1H, J = 12.2 Hz), 2.57 – 2.45 (m, 2H), 2.38 (dd, 1H, J = 5.2, 1.8 Hz), 2.33 (dd, 2H,
J = 6.2, 2.8 Hz), 2.29 –
2.21 (m, 2H), 2.11 (s, 3H), 2.04 – 1.72 (m, 6H), 1.68 (m, 2H), 1.41 (s, 3H), 1.27 (dd, 1H, J = 7.6,
5.7 Hz), 0.64 (s, 3H),
Synthesis
of
(8S,9S,10R,13S,14S,17S)-17-acetyl-10,13-dimethyl-11-oxo-
2,7,8,9,10,11,12,13,14,15,16,17-dodecahydro-1H-cyclopenta[a]phenanthren-3-yl acetate, 7
11

A solution of perchloric acid (10 mL) in ethyl acetate (10 ml) was added to a solution of acetic
anhydride (4.8 mL) in ethyl acetate (30 ml) and the final volume is brought up to 50 ml. Compound
6 (500 mg, 1.5 mmol) was added and the reaction was stirred for 20 minutes then reaction mass
was washed with saturated sodium bicarbonate and water. The organic layer was dried over
sodium sulfate and concentrated. The resulting oil was taken in methanol (10 ml), treated with
pyridine (0.5 ml) and cooled to 0 °C. Solid so formed was filtered and dried under vacuum to
1
obtain 7 as a white solid (500 mg, 89% yield). MS (ESI) m/z (M+1): 371; H NMR (300 MHz,
CDCl3): δ 5.67 (d, 1H, J = 2.0 Hz), 5.36 (d, 1H, J = 3.1 Hz), 2.80 – 2.58 (m, 3H), 2.51 (d, 2H, J =
12.6 Hz), 2.38 – 2.21 (m, 2H), 2.13 (s, 3H), 2.12 (s, 3H), 2.04 – 1.90 (m, 4H), 1.89 – 1.74 (m, 3H),
1.46 – 1.31 (m, 1H), 1.31 – 1.21 (m, 1H), 1.18 (s, 3H), 0.62 (s, 3H).
Synthesis of (8S,9S,10R,13S,14S,17S)-10,13-dimethyl-17-(2-methyl-1,3-dioxolan-2-yl)-11-oxo-
2,7,8,9,10,11,12,13,14,15,16,17-dodecahydro-1H-cyclopenta[a]phenanthren-3-yl acetate, 8
A mixture of 7 (500 mg, 1.35 mmol) and ethylene glycol (10 ml) in toluene (100 ml) was heated at
140 °C. After 5 mins p-toluenesulfonic acid hydrate (30 mg) was added and the reaction mixture
was refluxed during which the water produced was removed for 1.5 hr. The reaction was
monitored by TLC until all starting material was consumed. The mixture was brought to RT, diluted
with ethyl acetate, washed with saturated sodium bicarbonate and brine. The organic layer was
12

dried over sodium sulfate. The organic extracts were concentrated to obtain 8 as an off -white
1
crystalline solid (420 mg, 75% yield). MS (ESI) m/z (M+1): 415; H NMR (300 MHz, CDCl3): δ
5.67 (s, 1H), 5.35 (s, 1H), 4.02 – 3.92 (m, 2H), 3.92 – 3.80 (m, 2H), 2.69 (ddd, 3H, J = 14.7, 6.6,
2.4 Hz), 2.59 – 2.42 (m, 2H), 2.39 – 2.23 (m, 2H), 2.12 (dd, 5H, J = 6.1, 1.6 Hz), 2.06 (d, 2H, J =
7.0 Hz), 1.84 (ddd, 5H, J = 24.0, 15.0, 5.6 Hz), 1.56 (s, 3H), 1.53 (s, 1H), 1.29 – 1.17 (m, 7H),
0.76 (s, 1H), 0.74 (s, 1H), 0.67 – 0.55 (m, 1H).
Synthesis of (3S,8S,9S,10R,11S,13S,14S,17S)-10,13-dimethyl-17-(2-methyl-1,3-dioxolan-2-yl)-
2,3,4,7,8,9,10,11,12,13,14,15,16,17-tetradecahydro-1H-cyclopenta[a]- phenanthrene-3,11-diol, 9
An ice cooled solution of 8 (1.59 g) in 7:3 ethanol/THF (125 ml) was added dropwise to an ice-
cold suspension of sodium borohydride (2.5 g) in 80% ethanol (50 ml). The reaction was stirred
in an ice bath for 8 h, after which it was stirred overnight allowing the ice bath to slowly melt. The
reaction was quenched with 10% acetic acid and concentrated until a solid formed. The solid was
filtered, washed with water and dried to constant weight to yield 9 as a white solid (797 mg, 55%
yield). Asecond crop was obtained by further concentrating the mother liquor (228 mg, 16% yield).
MS (ESI) m/z (M+1): 377.8; 1H NMR (300 MHz, CDCl3) δ 5.24 (s, 1H), 4.34 (s, 1H), 3.97 (dt, 3H,
J = 7.7, 6.4 Hz), 3.91 – 3.84 (m, 2H), 3.49 (bs, 1H), 2.63 (s, 1H), 2.26 (t, 2H, J = 11.7 Hz), 2.02-
2.18 (m, 3H), 1.87 (dd, 2H, J = 13.7, 6.4 Hz), 1.74 (d, 4H, J = 5.5 Hz), 1.59 (m, 6H), 1.27 (s, 3H),
1.25 (3H), 1.02 (3H), 0.86 (m 1H).
13

Synthesis
of
1-((3S,8S,9S,10R,11S,13S,14S,17S)-3,11-dihydroxy-10,13-dimethyl-
2,3,4,7,8,9,10,11,12,13,14,15,16,17-tetradecahydro-1H-cyclopenta[a]phenanthren-17-yl)ethan-
1-one, 10
9 (570 mg, 1.51 mmol) was dissolved in acetone (40 ml) to which 4N sulfuric acid (2 mL) was
added. The resulting mixture was stirred at 0 °C for 90 min. The reaction mixture was concentrated
by rotavapor until a solid appeared, washed with water and extracted with ethyl acetate. The
organic layers were dried over anhydrous Na2SO4 and concentrated and crude product which
was purified by column chromatography (100-200 mesh, 2% methanol/dichloromethane) to yield
1
10 as a white solid (350 mg, 69.5% yield). HPLC purity; 98.2% ; MS (ESI) m/z (M+1): 333.7; H
NMR (300 MHz, DMSO-d6): δ 0.72 (s, 3H), 0.93 (dd, 1H, J = 11.6 and 3.6 Hz), 1.04 - 1.16 (m,
2H), 1.17 (s, 3H), 1.13- 1.24 (m, 1H), 1.38 (q, 1H, J = 12.8 Hz), 1.46 – 1.82 (m, 6H), 1.94 (dt, 1H,
9.6, J = 6.4 and 3.2 Hz), 1.99 – 2.13 (m, 4H), 2.05 (s, 3H), 2.18 (dd, 2H, J = 10.8 and 2.8 Hz),
3.19 – 3.29 (m, 1H), 4.10 (d, 1H, J = 3.2 Hz), 4.24 (t, 1H, J = 3.2 Hz), 4.57 (d, 1H, J = 4.4 Hz),
5.11 (br s, 1H). 13C NMR (75 MHz, CDCl3) δ 210.23, 141.15, 119.68, 70.20, 67.20, 63.54, 63.44,
58.20, 53.13, 42.63, 42.52, 40.45, 36.25, 36.14, 36.03, 31.49, 30.78, 30.68, 30.41, 27.88, 27.78,
23.78, 21.89, 21.78, 21.68, 14.93.
Computational methods
Molecules were constructed and superimposed using Discovery Studio 4.1.0.14169 (Accelrys
Software Inc.). Several conformations were considered for each molecule. Each conformer was
minimized using the Dreiding Minimization option in DS 4.1 until a converged energy was
reached. The resulting low energy conformations were then superimposed using the key
14

pharmacophores based on the lead molecule, 11--OHP, the two hydroxy groups, and the
carbonyl of the ketone while also attempting to maximize the hydrophobic overlap of the core
frameworks
Cell culture
Primary bovine coronary artery endothelial cells (BCAEC) were purchased from Cell Applications,
Inc. (San Diego, CA, USA) and mouse C2C12 myoblast were obtained from the American Type
Culture Collection (ATCC, VA, USA). Both cell types were grown with Dulbecco’s Modified Eagle’s
Media (DMEM) (1 g/L D-glucose, GIBCO, Waltham, MA, USA) supplemented with 10% fetal
bovine serum (FBS) (Omega Scientific, Tarzana, CA, USA) and 1% antibiotic and antimitotic
solution (PSF) (Thermo Fisher Scientific, Inc, Waltham, MA, USA) in a humidified atmosphere at
37oC and 5 % CO2 as previously described (2, 3). Cells were used at passages between 9 and
12. All compounds were solubilized in 10% dimethylsulfoxide (DMSO) for subsequent dilution
using DMEM. The night before treatment and when cells reached ~80% confluence the medium
was changed to a treatment medium consisting of DMEM (1g/L D-glucose) without phenol red
supplemented with 1% FBS, 1% PSF, and 1% nonessential amino acids (Thermo Fisher
Scientific, Inc, Waltham, MA, USA). The following day just prior to treatment cells were typically
>95% confluent. Cells were treated for 48 h with compounds and media with compounds was
replaced every 24 h. Control cells were treated with vehicle (DMSO <1% v/v).
In-cell ELISA
Endothelial and C2C12 cells were grown in 96 well plates. Each well was seeded with 30,000
cells and allowed to grow in standard medium until 80% confluent. Complex II (SDH-A, a subunit
of the succinate dehydrogenase) and IV (COX-I, a subunit of the cytochrome c oxidase) protein
content was evaluated in adherent cells using the MitoBiogenesis In-Cell ELISA kit (colorimetric)
from Abcam (Cambridge, MA, USA) following the manufacturer’s instructions. In brief, after
treatment, media was then removed, and cells were fixed for 20 min with 4% formaldehyde
15

followed by washes with phosphate buffer solution (PBS). Next, cells were exposed to 0.5% acetic
acid for 5 min to block endogenous alkaline phosphatase activity followed by washes with PBS
and exposure to a permeabilization buffer for 30 min. Cells were then exposed to a blocking
solution and incubated with primary antibodies overnight at 4oC, followed by rinsing with wash
buffer and incubation with secondary antibodies labeled with alkaline phosphatase and
horseradish peroxidase for 2 h. Developing solutions (alkaline phosphatase for complex II and
horseradish peroxidase for complex IV) were then added and kinetic colorimetric signals
measured in a spectrophotometer (Quant, Bio-Tek Instruments, Inc.) using 405 nm for complex II
and 600 nm for IV. Total absorbance values are normalized against cell number counting by Green
Janus staining.
Western blotting
Extraction of proteins, electrophoresis, and immunoblotting was performed as described
previously (2, 3). In brief, a total of 40 µg of protein was separated in 4-15% SDS-PAGE gels (Bio-
Rad, Hercules, CA, USA), electrotransferred to PVDF membranes (EMD Millipore, Bedford, MA,
USA), protein epitopes detected by primary antibodies (mouse anti-mitofilin [2E4AD5], mouse
anti-VDAC1/porin [20B12AF2], rabbit anti-NRF2 [ab92946], and rabbit anti-mtTFA [ab138351]
from Abcam, Cambridge, MA, USA; except for rabbit anti-GAPDH [D16H11] that was obtained
from Cell Signaling, Danvers, MA, USA) and HRP-conjugated secondary (Anti-mouse [7076] and
anti-rabbit [7074] IgG, HRP-linked antibodies from Cell Signaling, Danvers, MA, USA) antibodies
and protein abundance analyzed by chemiluminescence (ECL kit, Amersham Biosciences,
Piscataway, NJ, USA). The band intensities were digitally quantified by the ImageJ software
(version 1.8.0).
Citrate synthase activity
16

Twenty micrograms of protein were used for citrate synthase activity (ScienceCell-Research
Laboratories, Carlsbad, CA, USA) assessment following the manufacturer’s instructions as
described previously (1).
ATP levels
In brief, C2C12 were seeded onto 96-well plates and allowed to grow until they reached ~80%
confluence. A commercial kit was used to quantify ATP levels (Cayman Chemical, Inc. #700410).
As a positive control, a subset of cells was treated with 0.5 mM AICAR (12). After drug treatment,
cells were washed with cold PBS and harvested using a lysis buffer. Cell lysates were diluted in
sample buffer and used to measure luciferase based bioluminescense using a plate reader.
Immunofluorescence
C2C12 myoblasts were seeded onto glass slides (8-well NuncTM Lab-TekTM II Chamber SlideTM)
pre-coated with 0.1% gelatin, and cultured in standard medium until ~60% confluent as previously
described, but with some modifications (3). In brief, cells were treated with either vehicle (DMSO)
or 11--OHP (0.00001, 0.0001, 0.001 nM) for 48 h. When ready for visualization, myoblasts were
washed 3x with HBSS and then fixed with cold (-20oC) methanol for 10 min. Next, cells were
washed 3x with PBS and blocked with 3% BSA in PBS solution for 3 h. Slides were incubated
with a rabbit anti-VDAC1/porin monoclonal antibody (1:100, 3% BSAin PBS) (Abcam, Cambridge,
MA, USA [EPR10852(B)]) overnight at 4oC and thereafter washed 5x with PBS. Goat Alexa Fluor-
488 labeled anti-rabbit (1:400 in PBS) was then used as secondary antibody for 1 h at RT and
washed 5x with PBS. As a negative control, cells were incubated only with secondary antibody to
assess for non-specific binding. Finally, slides were washed with distilled water thoroughly (to
remove salts), dried vertically for 10 min, and framed on FLUOROSHIELD mounting medium with
DAPI (Abcam, Cambridge, MA, USA [ab104139)]. Cells were imaged (by a subject unaware of
treatment allocation within the slides) with an Olympus FV1000 Inverted Confocal IX81
Microscope at 800x800 pixels with a 40x objective lens, respectively. Subsequently, another
17

blinded subject calculated the corrected total cell fluorescence (CTCF) tool from Image J software
(version 1.8.0) using the following equation: CTCF=Integrated density-(Area of selected cell x
mean fluorescence of background readings). At least 40 myoblasts per field were recorded by
triplicate. For illustrative purposes, images were also collected at 1024x1024 pixels with a 60x
objective lens.
Data analysis
At least three independent cell culture experiments, each in duplicate, were performed. Results
are expressed as the mean ± standard error of the mean (SEM) or median and interquartile range.
Statistical analysis was performed by either using one-way analysis of variance (ANOVA) followed
by Tukey’s post hoc test or Mann-Whitney U test (as appropriate) using GraphPad Prism version
7.0 (GraphPad Software Inc., San Diego, CA, USA). A p-value of <0.05 was considered to be
statistically significant. For the calculation of EC50 values, data was normalized against baseline
(vehicle treated cells) and estimations derived using GraphPad.
18

RESULTS AND DISCUSSION
Modeling studies
The preliminary analysis and comparison of the geometric and spatial arrangement of
epicatechin(s) and sterol(s) scaffolds, identified pregnenolone (the sterol precursor of all steroid
hormones) as a candidate class of molecules to consider. Table 1, contrasts the depth, volume
and surface area of pregnenolone vs. (-)- and (+)-epicatechin indicating that both families of
scaffolds are similar with respect to their dimensions.
Epicatechin as well as sterol scaffolds, are highly planar and are capable of presenting axially
oriented hydroxyl group(s). Assuming some form of structural homology between an axial
hydroxylated epicatechin and an unknown sterol, and given the comparable spatial characteristics
shared by epicatechins and pregnenolone as the prototype sterol, hydroxylated pregnenolones in
particular, 11--OHP emerged as the strongest candidate structure to align against the
epicatechins (Table 1).
Table 1. Spatial characteristics of pregnenolone, (-)-epicatechin, (+)-epicatechin and 11--
hydroxypregnenolone (11--OHP).
Dimensions
Pregnenolone (-)-Epicatechin (+)-Epicatechin
11--OHP
length x width
x height
12.7 x 5 x 3.4
11.8 x 5.7 x 3.4
11.7 x 5.4 x 3.3
12.8 x 5.6 x 3.4
(Armstrong=A)
Volume (A3)
Area (A2)
333.8
294.6
234.1
246.0
246.1
252.1
342.6
299.2
As shown in Figure 2, using the Discovery Studio molecular modeling software, we implemented
quantitative, three-dimensional topological homology analyses focused on the three compounds,
19

(-)-epicatechin, (+)-epicatechin and 11--OHP (Figure 2A). The modeling demonstrated that
there was a reasonable, but not a good fit when 11--OHP was superimposed with (-)-epicatechin
(Figure 2B). Keeping the two axial hydroxyl groups, C3 on (-)-epicatechin and C11 on
pregnenolone, overlaid one finds that the pharmacophore 3 is partially aligned, pharmacophore
2 is not, though the molecular volumes fit well. Interestingly, (+)-epicatechin and 11--OHP
showed a significantly better overlay and fit, based on the structural overlap of all three
pharmacophores 1, 2 and 3, (Figure 2C). Molecular shape and volume also show a better overlay
between (+)-epicatechin and 11--OHP than the one observed for (-)-epicatechin. It should be
noted that this improved fit required the inversion and rotation of (+)-epicatechin compared to the
(-)-epicatechin leading to a reverse alignment of the pharmacophores 1 and 3, while the alignment
of the key axial hydroxyl (pharmacophore 2) group orientation is unchanged.
20

Figure 2. Molecular structures of ligands (A) and overlay of (-)-epicatechin and 11--
hydroxypregnenolone (B) and (+)-epicatechin and 11--hydroxypregnenolone (C). Numbers
denote de hydroxyl pharmacophores being compared among compounds.
The modeling experiment would predict that if the effects of (-)-epicatechin are attributable to the
partial structural alignment observed, then 11--OHP, as the endogenous species, should be
more effective in inducing MB and further, that both (+)-epicatechin and 11--OHP should be
more potent than (-)-epicatechin. (+)-Catechin with its C3-hydroxyl group in a non-axial
orientation, exhibited no overlap at the pharmacophore positions and thus, should be relatively
inactive. The importance of the axial orientation of the hydroxyl group in 11--OHP was further
explored in modeling studies, which predicted that the orientation of the hydroxyl group of 11--
OHP in the -orientation (i.e. mimicking (+)-catechin) should yield a loss of MB activity relative to
11--OH.
Effects of ligands on mitochondrial endpoints
As a surrogate marker of mitochondrial biogenesis (MB), we quantified the protein levels of
mitochondria-specific electron transport chain (ETC) complexes II and IV, determined by an in-situ
ELISA, after compound treatment. All values generated were normalized for 50,000 cells to eliminate
artificial increases in ETC complex proteins due to cell proliferation. Figure 3 shows the dose
dependent stimulatory effect on the expression levels ETC complexes II and IV by (-)-epicatechin.
(+)-Catechin had no effect on ETC complex II at all doses tested and only showed an increase in
ETC complex IV at the highest dose tested (100 nM). In contrast, (+)-epicatechin was found to be
significantly more active, demonstrating stimulatory effects on ETC complexes II and IV at
concentrations as low as 0.00001 nM, approximately 1000X lower than the concentration of (-)-
21

epicatechin (0.01 nM) required for a significant effect, an unusually high eudismic ratio. The fact that
(+)-catechin only induces significant increases of ETC complex IV protein levels at 100 nM and has
an apparent lesser effect on ETC complex II levels suggests that the biology of epicatechins is
associated with the stereochemistry of the hydroxyl group at C3 of the flavanol with a preferred axial
orientation.
Figure 3. Effects of (-)/(+)-epicatechin ((-)Epi/(+)-Epi) and (+)-catechin on bovine coronary artery
endothelial cell mitochondria complex II and IV protein levels as determined by ELISA (n=3, *p<0.05
by ANOVA followed by the Tukey’s posthoc test).
The capacity of 11--OHP, to induce MB (Figure 4), measured as increased expression of ETC II
and IV, shows that 11--OHP is as potent as (+)-epicatechin and both are, as predicted, substantially
more potent than (-)-epicatechin (Figure 2). When the epicatechin isoforms and two
hydroxypregnenolone isoforms were compared with respect to simulation of MB, and their EC50
22

calculated, the order of potency was 11--OHP = (+)-epicatechin > (-)-epicatechin >> 11--
hydroxypregnenolone (Table 2). Differences in potency between compounds (Table 2) are
noteworthy, and confirm that the orientation of the hydroxyl group in epicatechin vs. catechin and
between 11-- vs. 11--hydroxypregnenolone appear to be critical to the induction of MB.
An interesting observation from the MB assay results is that the activity profile of both (+)-epicatechin
and 11--OHP demonstrate a bell-shaped (biphasic) dose-response relationship, (Figures 3 and 4).
Apossible explanation for such a unique dose response profile could be that these molecules express
their biology through the initiation of a complex between the ligand and more than one protein target
simultaneously, suggesting that the activity of these compounds may require the formation of a variety
of protein complexes facilitated by the ligand, some with internal homeostatic properties (13). We are
currently designing molecules and experiments to further test this hypothesis.
Figure 4. Effect of 11--hydroxypregnenolone on bovine coronary artery endothelial cell mitochondria
complex II and IV protein levels as determined by ELISA (n=3, *p<0.05 by ANOVA followed by the
Tukey’s posthoc test).
C2C12 cells, which are an established mouse muscle cell line precursor, have been used as a model
to examine the effects that compound treatment has on relevant endpoints. For example, the capacity
23

of AICAR (an AMP analog that activates AMPK) to stimulate mitochondrial biogenesis markers and
ATP levels in skeletal muscle has been explored (14, 15). We used C2C12 cells to determine the
effects that 48 h treatment with either (+)-catechin, (-)-epicatechin, (+)-epicatechin (at 0.1 nM) or
AICAR (0.5 mM) has on ATP levels. Results are presented in Figure 5 and indicate that while (+)-
catechin was unable to significantly stimulate ATP, (-)-epicatechin led to a significant ~50%
stimulation in ATP levels that further increased to ~100% with either (+)-epicatechin or AICAR.
Noteworthy, is the much lower concentration required for 0.1 nM (+)-epicatechin to reach a
comparable (p > 0.05) increase in ATP levels as that of 0.5 mM AICAR.
Figure 5. Effects of ligands treatment on ATP levels in C2C12 cells. C2C12 cells were treated with
either vehicle or 0.1 nM of compounds for 48 h. As a positive control, AICAR (0.5 mM) was used.
ATP levels were determined by a bioluminescent luciferase based assay. (Cayman Chemical, Inc.).
24

Three to five experiments were performed in triplicate, p-value determined by ANOVA with statistical
significance established at p<0.05.
Using the same cell system, we then explored the capacity of 11--OHP to act as a potential
physiological stimulator of MB. Using 11--OHP doses identified from experiments using endothelial
cells as effective, C2C12 cells were treated 48 h with either 1, 0.1 or 0.01 pM of the compound.
Figure 6, reports on the effects that treatment with 11--OHP has on key stimulators of MB such as
TFAM (16) (Figure 6A) and NRF2 (17) (Figure 6B) and on porin (18) (Figure 6C, structural
component of mitochondrial membrane pore) and mitofilin (Figure 6D, cristae component). As per
Western blot results, an up to 25% stimulation of relative protein levels was noted with 0.01-0.1 pM
11--OHP.
25

Figure 6. Effects of 11-β-hydroxypregnenolone treatment on regulators of mitochondrial biogenesis
(TFAM, NRF2) (16, 17) and organelle components (porin, mitofilin) (18) in C2C12 cells. C2C12 cells
were treated with either vehicle or 0.01, 0.1 or 1 pM of the steroid compound for 48 h. Relative protein
levels were assessed by Western blots. Values were normalized by those of glyceraldehyde 3-
phosphate dehydrogenase (GAPDH). Three independent experiments were performed in triplicate,
p-value determined by ANOVA with statistical significance established at p<0.05.
Figure 7, further documents the capacity of 11--OHP to stimulate MB as per significant increases
in complex II (Figure 7A) and IV (Figure 7B) protein levels (by ELISA) effects that were corroborated
at the functional level by the use of a citrate synthase activity assay (19) (Figure 8A). Experiments
26

were also performed to determine the effects of 11--OHP on C2C12 ATP levels (Figure 8B). Results
demonstrate a concentration-response like effect, whereby 1 pM of 11--OHP increased ATP levels
by 70% vs. 100% noted with 0.5 mM of AICAR.
Figure 7. Effects of 11-β-hydroxypregnenolone treatment on mitochondrial complex II and IV protein
levels in C2C12 cells. C2C12 cells were treated with either vehicle or 0.01, 0.1 or 1 pM of the steroid
compound for 48 h. Protein levels were quantified by an ELISA (mitochondrial biogenesis kit) with
values normalized by cell count. Three to five experiments were performed in triplicate, p-value
determined by ANOVA with statistical significance established at p<0.05.
As noted in Figure 9, immunofluorescence analysis, using confocal microscopy, of C2C12 cells
treated with 1 pM 11--OHP (highest effect) also suggests an increase in organelle mass. No
significant background signals were detected in the negative control (no primary antibody, data not
shown).
27

Figure 8. Effects of 11-β-hydroxypregnenolone treatment on (A) citrate synthase activity
(mitochondrial function) and (B) ATP levels in C2C12 cells. C2C12 cells were treated with either
vehicle or 0.01, 0.1 or 1 pM of the steroid compound for 48 h. Citrate synthase activity was determined
by a colorimetric kit (ScienceCell-Research Laboratories). ATP levels were determined by a
bioluminescent luciferase based assay. (Cayman Chemical, Inc.). Three to five experiments were
performed in triplicate, p-value determined by ANOVA with statistical significance established at
p<0.05.
Figure 9. Effects of 11-β-hydroxypregnenolone treatment on C2C12 cell mitochondrial mass. C2C12
myoblast grown on 8-well plates (NuncTM Lab-TekTM II Chamber SlideTM) were treated with either
vehicle or 1 pM of the steroid compound for 48 h. Upper panel, representative immufluorescence
images of cells treated with DMSO (vehicle, <1% v/v) and 1 pM 11-β-hydroxypregnenolone. Bottom
panel, quantification of fluorescence levels. Imaging of mitochondria was performed by cell
immunostaining using a primary rabbit anti-VDAC1/porin monoclonal antibody and secondary goat
anti-rabbit Alexa Fluor-488 labeled. Fluorescence levels per myoblast were assessed by analysis of
the corrected total cell fluorescence (CTCF) tool from ImageJ software (version 1.8.0) using the
following equation: CTCF=Integrated density-(Area of selected cell x mean fluorescence of
background readings). At least 40 myoblasts per field were recorded. Data are expressed as median
and interquartile range.**** p-value <0.0001 vs. 0 nM (vehicle), as analyzed by the Mann-Whitney U
test. Images taken using a 60x objective lens. Scale bar, 50 µm. DIC, differential interference
contrast; 11- β-OHP, 11-β-hydroxypregnenolone.
28

Altogether, results from C2C12 cells support those derived from endothelial cell experiments and
suggest that treatment of cells types relevant to muscle structure/function (such as endothelial and
muscle) likely trigger coordinated responses that may lead to improved muscle performance as seen
with epicatechin.
To our knowledge this is the first report of activity of 11--OHP in any biological system. The fact that
11--OHP can be effective in vitro at concentrations as low as 10-15 M and has specific structural and
stereochemical requirements for its activity suggests, highly specific target engagement with a
potency exceeding, by several orders of magnitude, that reported for other putative inducers of MB
(20). 11--OHP is not recognized as a member of the classical pituitary-adrenal steroid class. As a
hydroxysterol, it could be a member of the neurosteroid class of sterols (21). First described in the
central nervous system but now known to be more systemically produced, neurosteroids reflect
peripheral synthesis of sterols via modification by cytochrome p450 enzymes, independent of the
pituitary-adrenal axis responsible for classical steroidogenesis (22, 23). Mitochondria are the locus
for CYP450 11A1, the enzyme that converts cholesterol to pregnenolone. 11--hydroxylase and
CYP450 11B1 are also located in the mitochondria and are known to be responsible for the specific
hydroxylation of steroids at the C-11 with the hydroxyl group in an axial orientation. Nearly all other
enzymes implicated in steroidogenesis are located outside the mitochondria, in the endoplasmic
reticulum (24). Given the relative simplicity of 11--OHP, it is interesting to consider that mitochondria
themselves might be the source of this novel sterol, inasmuch as they express the two enzymes that
could plausibly produce this novel hydroxylated pregnenolone.
Table 2. Stimulation mitochondrial electron transport chain (ETC) complexes II and IV. EC50
values were calculated using the ascending part of the dose response curve.
29