Rationally designed hecogenin thiosemicarbazone analogs as novel MEK inhibitors for the control of breast malignancies

Article abstract of DOI:10.1016/j.bmc.2017.09.033

Natural products have documented oncology success history as valuable scaffolds for selective target modulation. Herein, the sapogenin hecogenin (1) was screened for its anti-breast cancer inhibitory capacity using in vitro assays, including proliferation, cytotoxicity, migration, invasion assays, and Western blotting. The results identified 1 as a propitious hit with modest activities attributed to the concurrent down-regulation of mitogen activated protein kinase kinase/extracellular signal-regulated kinase (MEK) distinctive downstream effectors. Guided by in silico 3D-structural insights of MAPK kinase domain, an extension strategy was adopted at 1's C-3 and C-12 aimed at the design of novel hecogenin-based analogs with improved target binding affinity. Thirty-three analogs were prepared and tested, among which hecogenin 12-(3′-methylphenyl thiosemicarbazone) (30) displayed the most potent selective anticancer effects. Analog 30 demonstrated antiproliferative, antimigratory and anti-invasive activities at low μM level, compared to the negligible effect on the non-tumorigenic MCF-10A mammary epithelial cells. Durable regression of breast tumor xenografts in athymic nude mice was observed after treatments with 30, compared to its parent hecogenin at the same dose regimen, confirmed the hit-to-lead promotion of this analog. Hecogenin-12-thiosemicarbazones, represented by 30, is a novel MEK inhibitory lead class to control breast neoplasms.

Full text of DOI:10.1016/j.bmc.2017.09.033

Accepted Manuscript
Rationally designed hecogenin thiosemicarbazone analogs as novel MEK in-
hibitors for the control of breast malignancies
Heba E. Elsayed, Hassan Y. Ebrahim, Eman G. Haggag, Amel M. Kamal, Khalid
A. El Sayed
PII:
S0968-0896(17)31493-1
https://doi.org/10.1016/j.bmc.2017.09.033
BMC 13993
DOI:
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
20 July 2017
16 September 2017
21 September 2017
Please cite this article as: Elsayed, H.E., Ebrahim, H.Y., Haggag, E.G., Kamal, A.M., El Sayed, K.A., Rationally
designed hecogenin thiosemicarbazone analogs as novel MEK inhibitors for the control of breast malignancies,
Bioorganic & Medicinal Chemistry (2017), doi: https://doi.org/10.1016/j.bmc.2017.09.033
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. 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.

Article
Rationally designed hecogenin thiosemicarbazone analogs as novel MEK
inhibitors for the control of breast malignancies
Heba E. Elsayeda,b, Hassan Y. Ebrahima, Eman G. Haggagb, Amel M. Kamalb, Khalid A. El
Sayeda*
aDepartment of Basic Pharmaceutical Sciences, School of Pharmacy, University of Louisiana at
Monroe, Monroe, Louisiana, USA, 71201.
bDepartment of Pharmacognosy, Faculty of Pharmacy, Helwan University, Helwan, Cairo,
Egypt, 11795.
*Correspondence: Professor Khalid El Sayed, Department of Basic Pharmaceutical Sciences,
School of Pharmacy, University of Louisiana at Monroe, 1800 Bienville Drive, Monroe,
Louisiana 71201, USA. Phone: +1-318-342-1725; Fax: +1-318-342-1737; E-mail:
elsayed@ulm.edu
1

Abstract
Natural products have documented oncology success history as valuable scaffolds for selective
target modulation. Herein, the sapogenin hecogenin (1) was screened for its anti-breast cancer
inhibitory capacity using in vitro assays, including proliferation, cytotoxicity, migration, invasion
assays, and western blotting. The results identified 1 as a propitious hit with modest activities
attributed to the concurrent down regulation of mitogen activated protein kinase
kinase/extracellular signal-regulated kinase (MEK) distinctive downstream effectors. Guided by
in silico 3D-structural insights of MAPK kinase domain, an extension-strategy was adopted at 1's
C-3 and C-12 aimed at the design of novel hecogenin-based analogs with improved target
binding affinity. Thirty-three analogs were prepared and tested, among which hecogenin 12-(3'-
methylphenyl thiosemicarbazone) (30) displayed the most potent selective anticancer effects.
Analog 30 demonstrated antiproliferative, antimigratory and anti-invasive activities at low μM
level, compared to the negligible effect on the non-tumorigenic MCF-10A mammary epithelial
cells. Durable regression of breast tumor xenografts in athymic nude mice was observed after
treatments with 30, compared to its parent hecogenin at the same dose regimen, confirmed the
hit-to-lead promotion of this analog. Hecogenin-12-thiosemicarbazones, represented by 30, is a
novel MEK inhibitory lead class to control breast neoplasms.
Keywords: Antiproliferative, Breast cancer, Hecogenin, MEK inhibitor, Rational design,
Thiosemicarbazones.
2

1. Introduction
Steroidal sapogenins encompass a structurally diverse class of natural products that
possess non-polar, 27-carboncore skeleton.1 They have received considerable attention as
potential precursors for semisynthesis of various pharmaceutically valuable derivatives, as
cortisone and other related corticosteroids.1 Spirostanes represent a distinct class of steroidal
sapogenins with a spiroketal side chain.2 Hecogenin (1) is a ubiquitous spirostane abundant in
plants of the genus Agave (sisal, Agavaceae) and perhaps the only known naturally occurring 12-
oxo-spirostane steroidal sapogenin. Manifold biological activities have been reported for 1,
particularly anticancer.2-5 Sapogenin 1 has been subjected to various chemical transformations
aimed at optimizing its anti-neoplastic activity. Examples of 1’s prior semisynthetic
manipulation are the cleavage or modifications of thespirostanic side-chain to cholestanic
analogs with selective antitumor activity.6,7 Moreover, many aliphatic esters of hecogenin have
been semisynthesized and exploited in this regard; particularly hecogenin acetate. Hecogenin
acetate was identified as a suitable and economical starting material for multiple complex
anticancer entities. Fortunately, it is incorporated in the convergent synthesis of the
cephalostatins, ritterazines and hippuristanol.8,9 Reportedly, hecogenin acetate modulated various
key cell signaling events as extracellular signal-regulated kinases (ERK1/2) phosphorylation and
matrix metalloproteinase-2 (MMP-2) production as well as interference with reactive oxygen
species (ROS) production.10
Thiosemicarbazones deemed an outstanding pharmacophoric core, with self-
complementary hydrogen bonding motif in diverse antiprotozoal, antiviral, antibacterial and
antitumorentities.11-14 The antitumor activity of thiosemicarbazone-containing structures perhaps
is due to the moon-lightening mechanisms, such as inhibiting ribonucleotide reductase (RR),
3

ROS production, inhibiting topoisomerase II and more recently inhibiting multidrug resistance
protein 1 (MDR1).15-17 The immense significance of thiosemicarbazones as potential
pharmacophoric moiety in drug design is documented.
Cancer is among the main health concerns worldwide and one of the primary medicinal
chemistry and pharmacologytargets.18 Breast cancer is the most common malignancy affecting
females worldwide.18 Moreover, breast cancer is considered the leading cause of disability-
adjusted life-years globally attributed to lack of effective early diagnosis and selective treatment
options.19 Chemotherapy is still the established forefront treatment alternative for many cancer
patients, however selectivity and off-target side effects are major concerns.19 Intent transition
from cytotoxic chemotherapy to molecularly-targeted cancer therapy resulted in astonishing
successful therapies, with significant safety and improved response rates.20 Blockade of protein
phosphorylation is amongst the clinically validated auspicious targets in cancer control. Protein
phosphorylation is a decisive reversible post-translational mechanism, which is controlled
through various protein kinases that are capable to phosphorylate on tyrosine (Tyr), serine (Ser),
or threonine (Thr). Dual-specificity kinases are those that overall belong to Ser/Thr kinase
group.21 Protein kinase inhibitors are molecules aimed at protein phosphorylation inhibition and
interrogate specific intracellular signaling. Imatinib, an inhibitor of the oncogenic kinase BCR-
Abelson murine leukemia viral oncogene homolog 1 (Abl1) in chronic myelogenous leukemia,
was the first protein kinase-targeting anticancer drug followed by several other examples.22
Rat sarcoma small GTPase/rapidly accelerated fibrosarcoma kinase/mitogen activated
protein kinase kinase/extracellular signal-regulated kinase (RAS/RAF/MEK/ERK) is an
organized conserved eukaryotic signaling pathway.23 The signaling module is the driver behind
paramount cellular processes such as gene transcription, cellular proliferation, and survival.24,25
4

The Ser/Thr kinases MEK1/2 specifically phosphorylate and activate ERK1/2.26,27 The activated
ERK is dimerized thereafter, regulating various targets in the cytosol and translocate to the
nucleus where it phosphorylates a variety of transcription factors regulating gene expression. The
clinical trial out comes of selumetinib (AZD6244), a potent selective MEK1 inhibitor, was
reported in 2008 and showed significant improvements in patients with advanced cancers.28 On
the other hand, in case of patients with advanced melanoma harboring N-RAS/B-RAF mutations,
binimetinib (MEK162) was endorsed as a contemporary MEK1/2 inhibitor.29 Recently, FDA
approved the first pan-MEK1/2 inhibitor, trametinib (Mekinist™), for the treatment of metastatic
melanoma.30 However, despite the considerable progress on new molecularly targeted therapies,
discovery of more potent, targeted MEK lead inhibitors is still a dire therapeutic need.
This study reports the rational design of semisynthetic hecogenin analogs with improved
in vitro anti-breast cancer activities through targeting MEK kinase domain. Rational design
targeted C-3 and C-12 ester, semicarbazone and thiosemicarbazone hybrid extensions to improve
MEK kinase domain binding affinity, anti-breast cancer activity and selectivity. The
thiosemicarbazone 30 reigned supreme in in vitro and in vivo anticancer potencies and had no
toxic effects on the non-tumorigenic mammary epithelial MCF-10A cell growth. Western blot
analysis revealed that 30's anticancer effect is, at least in part, mediated by the blockade of MEK
activation with subsequent inhibition of the downstream mitogenic signaling. 30 markedly
reduced tumor growth in orthotopic breast cancer xenograft model compared to its parent 1
without overt toxicity signs. Collectively, the present findings suggest the hecogenin-12-
thiosemicarbazones, represented by 30, as a promising MEK inhibitory lead entity with a
potential to control breast malignancies associated with aberrant MEK activity.
5

2. Results and discussion
Hecogenin (1), a ubiquitous spirostanic steroidal sapogenin in sisal wastes, has been
previously described for its anticancer activity.3 Therefore, 1 was screened in vitro for ability to
inhibit breast cancer proliferation, migration, and invasion. The antiproliferative effect of 1 was
assessed against a panel of six human mammary carcinoma cell lines endowed by diverse
molecular and phenotypic characteristics. For instance, estrogen receptor (ERα) is expressed in
the luminal A-type MCF-7 and T-47D cell lines; while SKBR-3 and the luminal B-type BT-474
are human epidermal growth factor receptor 2 (HER2) overexpressing ones. On the other hand,
the claudin low MDA-MB-231 and the basal MDA-MB-468 were negative for the triple
receptors; ERα, HER2 and progesterone receptor (PR) and so alternatively named, triple negative
breast cancer (TNBC) cell lines.31 Sapogenin 1 exhibited modest antiproliferative activity with
IC50values in the mid μM level (Figure 1). Thus, it was reasonable to hypothesize that 1 target a
common pathway(s) involved in the proliferation and survival of breast cancer cells.
Hecogenin-3-O-acetate’s ability to modulate the extracellular signal-regulated kinases
(ERK1/2) phosphorylation was previously reported and evidenced by Western blot analysis.10
ERK is the only downstream substrate for the non-receptor hydrophilic protein mitogen-
activated protein kinase/extracellular signal-regulated kinase (MEK). MEK1/2 phosphorylate,
activate and spark ERK1/2-mediated oncogenic proliferation and differentiation pathways.26,27
A
computational docking experiment was accomplished using Glide 5.8 module32 in standard mode
to explore the virtual binding modes of 1 and hence explore potential future rational design of
more potent semisynthetic analogs. Careful inspection of the prepared dual-specific MEK protein
kinase revealed its resemblance to alternative protein kinases in having a multifunctional, small
N-terminal lobe, protein kinase cleft and large C-terminal lobe comprising several conserved α-
6

helices and β-strands.33 Sapogenin 1 was anchored within the generated grid constraints of MEK
cavity with a notable docking score of -7.9.
Figure 1. Effect of sapogenin 1 treatments on the proliferation of multiple human breast cancer
cell lines. Displayed is the mean percent cell viability at different concentrations of 1 after 72 h
incubation. The red dashed line showed the calculated IC50 ±SEM for each cell line.
The space-filling model assumed 1 in an extended spatial conformation, consistent with the
general horizontal dimensions of the pocket, leaving behind the C-3 hydroxyl and C-12 ketone
enough spaces to accommodate structure extensions (Figure 2). Interestingly, 1 displayed
favorable H-bond bridge by the C-3 OH group with the gatekeeper Met146 carbonyl backbone
allocated in the hinge region, while the rear pyran oxygen displayed a H-bond interaction with
Lys192 terminal amino in the catalytic loop. Accordingly, 1 was hypothesized as a potential
MEK inhibitory scaffold amenable for semisynthetic optimizations at C-3 hydroxyl and C-12
ketone owing to the vacant room prevailed behind.
7

27
(A)
21
17
25
23
O
O
18
22
20
12
19
O
H
14
9
H
1
8
H
H
3
6
HO
1
H
(B)
(C)
Figure 2. Overview of the binding of 1 in green (ball and stick cartoon or CPK format) with
MEK kinase domain (PDB: 3EQF) represented as a ribbon diagram. The protein is oriented as
for the canonical kinase domain, with the N-terminal lobe at the top and the C-terminal lobe at
the bottom. The three figures are from a similar vantage point, displayed in space-filling model
the good fitting of 1 within the identified grid. 1 Displayed favorable H-bond interactions with
residues Met146 and Lys192, while enough space around C-3 hydroxyl and C-12 ketone likely
to accommodate subsequent structure extensions.
The first optimization direction of 1 aimed at C-3 spacer domain via esterification of C-3
secondary alcohol. This was implemented through base-catalyzed esterification reaction of 1
with various aromatic acid chlorides (Scheme 1), affording the corresponding ester analogs 2-9.
8

27
O
O
O
21
O
H
25
O
H
O
18
H
22
12
20
12
O
O
i
12
H
17
H
O
19
16
H
H
14
H
O
O
H
H
1
H
3
H
H
10
3
H
H
O
1'
5
3
7
7'
5'
3'
1'
O
H
2'
HO
H
6'
1
3'
5'
O
H
2-4, 6-9
5
R1
Scheme 1. General semisynthetic route to hecogenin ester analogs 2-9. Reagents and conditions:
(i) Acid chloride, pyridine, DMAP, RT, overnight.
Purified products were fully characterized by 1D and 2D NMR spectroscopy and mass
spectrometry. For instance, 3-O-hecogenin benzoyl ester (2) displayed a downfield shift of H-3
to δ 4.90 (+1.3 ppm), along with its corresponding carbon C-3 by +4.0 ppm (δ 75.0), compared
to the parent 1, confirming the C-3 esterification.
The human triple negative breast cancer (TNBC) MDA-MB-231 cells are best known for
their mutant RAS and RAF proteins and showed high levels of phosphorylated ERK (p-ERK)
expression, indicating active MAPK pathway.34 Additionally, MDA-MB-231 cells were optimally
sensitive to 1’s treatments in MTT proliferation assay (Figure 1). Thus, semisynthesized esters
were evaluated for their antiproliferative potency on the surveyed panel of breast
adenocarcinoma cell lines, while their SAR was configured according to their activity against
MDA-MB-231 cells. Results demonstrated that ester analogs showed inconsiderable
improvement in 1's antiproliferative activity (Table 1). This may be attributed, at least in-part, to
potential hydrophobic interactions with the proximity amino acids that replaced the free C-3 OH
interactions, which may stipulate the significance of the C-3 OH group to anchorage at the target
kinase’s hinge region. Subsequent optimizations maintained free C-3 OH group and targeted C-
12 spacer in 1 to probe additional binding interactions. An extension strategy was adopted using
a hydrophilic linker capable to interplay H-bond bridges with the target, meanwhile come to an
9

end with lipophilic functionalized substituent directed towards the target hydrophobic cleft.
Hence, a cross-coupling reaction using various commercially available phenyl semicarbazides
and phenyl thiosemicarbazides appropriately functionalized on the aromatic ring was
successfully achieved (Scheme 2).
H
7'
3'
H
27
N
N
1'
5'
2'
21
Y
O
O
O
25
N
H
O
R2
18 H
X
22
20
12
12
i/ii
17
H
O
19
16
H
14
H
H
1
H
H
3
10
H
H
5
7
3
HO
H10 34
HO
-
1
H
Scheme 2. Semisynthesis of hecogenin phenylsemicarbazone 10-14 and hecogenin
thiosemicarbazones 15-34. Reagents and conditions: (i) Semicarbazide.HCl, EtOH, Pyridine,
RT, 6 h. (ii) Semicarbazide/thiosemicarbazide, EtOH, RT, overnight.
The aimed products were purified, isolated in moderate to excellent yields and elucidated
using 1D and 2D NMR spectroscopy and mass spectrometry. Careful inspection of the 13C-
PENDANT NMR spectrum of hecogenin-12-(phenyl semicarbazone) (10) and hecogenin-12-
(phenyl thiosemicarbazone) (15) revealed the replacement of 1's C-12 ketone signal (δC 213.4)
with new ones at δC 158.5 or 159.9, corresponding to their new C-12 imine carbonyls,
respectively. The new imine carbonyl carbons showed 3J HMBC cross interactions with the C-18
methyl proton singlets (δH 0.98-1.00). Moreover, the integrity of urea or thiourea fragment was
based on their C-1` carbonyl or thiocarbonyl at δC 154.6 or δC176.3, respectively. Meanwhile,
the broad singlet signals at δ 8.50-9.50 were ascribable to the two aza-methines of the urea
fragment, manifesting 2J HMBC correlations with the C-1` carbonyl and thiocarbonyl or the C-2`
quaternary aromatic carbons. In addition, one of the aza-methines showed a 3J HMBC
correlation with the C-12 imine carbonyl. While E and Z-geometrical isomers are possible for the
10

non-symmetric semicarbazone and thiosemicarbazone products, it is interesting to note that
evidence for such isomeric mixtures was not observed by 1D NMR experiments, suggested the
presence of only one isomer. The 1H-1H NOESY spectrum indicated the exclusive oxime double
bond E-geometry orientation in 30, based on the cross peak interactions between the hydrazine
NH singlet (δ 8.70) and H2-11 methylene proton at δ 1.92. This assignment was in agreement
with the literature, which the preference to the less sterically hindered E-isomer product of such
reactions.35 Thus, the same double bond geometry was also assumed in rest of analogs.
Analogs 10-34 were evaluated for their antiproliferative activities against the TNBC
MDA-MB-231 cell line. In fact, the improved activity of hecogenin-12-phenylsemicarbazone
(10), compared to the parent1 (IC50 = 9.2μM) encouraged additional optimizations. Various
analogs with substituted phenyl ring that featured different steric, hydrophobic, and electronic
properties were synthesized and tested merely to perceive better activity. For instance, the
electron withdrawing hybrids represented by analogs 11-13; m-chlorophenyl, p-chlorophenyl and
m-trifluoromethyl phenyl semicarbazones, respectively, (13) were semisynthesized and tested.
The chlorinated analogs declared ahead improvement in potency compared to10 (Table 1), even
though the positional effect of the Cl atom seems to have a significant role in the activity. The
para-positioned chlorinated analog declared better potency than its meta-conger (Table 1). On
the other hand, unexpectedly the meta-substituted analog 13, bearing trifluoromethyl phenyl
moiety showed a significant activity decrease (IC50 = 14.0 μM), drawing the attention to the
position-activity discriminatory effects. Interestingly, while examining the urea linker optimal
span, the tosyl semicarbazone analog 14 revealed significantly reduced activity compared to 10
(Table 1), suggesting the four-atom hydrazine carbothioamide hydrophilic linker is optimal for
possible for target affinity and subsequent cellular activity. This was also supported by the
11

literature use of "soft" donor atoms, as sulfur in thiosemicarbazone scaffolds for good anticancer
activity.36 Hecogenin-12-phenyl thiosemicarbazone (15) exhibited the expected improved
potency (IC50 = 7.8 µM), compared to its phenyl semicarbazone counterpart 10, proofing the
activity preference for thiosemicarbazones. The impact of electronic effects of different chloro,
fluoro, nitro, and trifluoromethyl phenyl-substituted thiosemicarbazones 16-25 was then studied.
The 3`-trifluoromethyl-bearing analog 23 exhibited the most attained potency of this group (IC50
= 2.1 µM). On the other hand, the electron-donating functionalities were also explored through
synthesizing and biological testing of methoxy (26-27), phenoxy (28), ethyl (29), methyl (30-33)
and methylenedioxy (34) phenyl thiosemicarbazone analogs. Notably, the 3`-methyl analog 30
exhibited the utmost potency against a panel of breast cancer cell lines (Table 1 and Figure 3),
spotlighting the final requirement of an optimal mono-substituted, hydrophobic, sterically
tolerable phenyl substituent. Hence, analog 30 was selected for a comprehensive activity study.
Figure 3. Effect of analog 30 treatments on multiple human breast cancer cell lines and the
human non-tumorigenic mammary epithelial cells MCF-10A. Displayed is the mean percent cell
12

viability at different concentrations of 30 after treatment incubation period. The red dashed line
showed calculated IC50 ±SEM of 30 for each cell line.
Molecular docking of 30 within the prepared MEK kinase domain showed that it is quite
closely overlaps with its parent1, with better extension towards the hydrophobic pocket,
affording an excellent docking score of -9.0. The imine nitrogen displayed a H-bond bridge with
the amino acid Ser194, meanwhile the E-configuration enabled the substituted phenyl thio-
semicarbazone fragment to be well-fitted within a hydrophobic pocket clustered with Val127,
Ile141, Cys207, Leu74 and Met143 residues, generating an excellent hydrophobic interaction.
Interestingly, the distal phenyl moiety in 30 displayed parallel-displaced stacking with Phe209
allocated within the Asp-Phe-Gly (DFG) motif, while the ortho-methyl functionality
predominantly contributed a hydrophobic interaction with Cys207-methylene. The thiourea
linker formed a H-bond interaction with the Lys97’s positively charged terminal amino group
(Figure 4).
O
H
H
N
O
7'
3'
O
O
N
1'
H
H
5'
O
2'
N
Y
H
12
12
O
R2
X
12
O
H
H
O
H
H
H
O
O
H
H
H
H
H
H
3
H
3
3
7'
O
1'
1'
O
5'
3'
2'
3'
H
HO
H
H
5'
6'
10-34
O
2-4, 6-9
5
R1
Table 1. In vitro antiproliferative activity (IC50 (μM) ±SE) of compounds 2-34 against a panel
of human breast cancer cell lines.
Cpd
IC50 (μM±SEM))
No. R1
R2
-
X
-
Y
-
MDA-MB-
MCF-7
BT-474
SKBr3
T-47D
231
468
-
28.7±1.8 35.2±2.0 31.0±2.1 38.2±2.1 33.3±2.6 36.6±2.0
1
13

H
-
-
-
-
-
-
-
-
-
-
-
-
-
-
29.1±2.0 34.2±2.2 30.3±2.3 36.3±1.9 33.4±2.0 37.8±3.1
2
4'-Cl
-
-
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
>40.0
3
3'-OCH3
-
-
4
-
-
-
5
4'-CF3
-
-
6
5'-F
-
-
27.0±1.3 33.4±1.8 28.0±1.4 30.0±1.9 36.0±2.0 29.0±1.9
25.7±1.4 32.6±1.9 33.2±1.8 35.0±1.7 34.7±2.3 >40.0
7
5'-CN
-
-
8
5'-NO2
-
-
25.3±1.7 >40.0
9.2±1.1 11.9±0.9 10.4±0.7 11.9±0.4 12.0±0.7 13.5±0.6
10.0±1.3 10.9±0.7 9.7±0.9
8.5±0.9 11.3±0.8 9.1±1.0
32.4±1.2 >40.0
34.9±1.9 >40.0
9
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
H
O
O
O
O
O
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
4'-Cl
12.0±0.9 9.8±0.8
9.8±0.7 9.0±0.9
13.0±0.9
12.0±0.7
5'-Cl
4'-CF3
5'-CH3
H
14.0±1.2 15.3±1.0 13.8±1.1 19.2±1.5 15.7±0.9 20.9±1.0
31.0±1.9 >40.0
SO2 28.0±1.5 36.4±1.8 29.3±2.0 >40.0
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
7.8±0.7
5.5±0.9
8.4±0.7
6.2±0.5
8.0±0.5
5.9±0.8
9.3±0.6
6.5±0.6
9.1±0.8
5.7±0.7
10.3±1.4
7.0±0.6
3'-Cl
5'-Cl
11.5±0.8 12.8±1.2 12.0±0.9 11.8±0.9 14.8±1.2 15.1±1.1
3',5'-di-Cl
3',7'-di-Cl
5'-F
7.5±0.7
6.0±0.9
5.0±0.8
4.3±0.8
7.3±0.9
2.1±0.4
3.2±0.5
7.3±0.8
6.8±0.7
9.5±0.9
8.0±0.8
7.3±0.8
4.2±0.7
6.1±0.9
8.0±1.0
2.7±0.5
3.9±0.2
8.2±0.7
7.5±0.9
7.9±0.6
6.4±0.8
4.0±0.6
5.0±0.7
7.5±0.9
2.2±0.3
3.7±0.2
8.0±0.5
7.1±0.4
9.0±0.9
6.9±0.7
6.3±0.8
6.9±0.9
8.8±0.7
9.0±0.8
4.5±0.7
5.5±0.6
9.5±1.0
10.4±1.1
5.9±0.5
9.0±1.3
11.5±1.2
2.4±0.4
3.8±0.2
11.0±1.3
10.2±1.0
3'-NO2
5'-NO2
3'-CF3
5'-CF3
4',6'-di-CF3
3'-OCH3
5'-OCH3
5'-O-ph
3'- CH2CH3
3'-CH3
5'-CH3
3',5'-di-CH3
3',7'-di-CH3
3',5'-
10.3±1.0 7.9±0.9
3.1±0.6
4.4±0.6
2.1±0.3
4.0±0.5
10.8±1.0 8.6±0.9
8.8±0.6 7.5±0.8
11.2±2.0 10.1±1.4 10.9±0.9 11.5±1.3 14.2±0.8
28.1±1.7 30.8±2.0 27.0±1.6 31.2±1.8 29.0±2.1 >40
3.8±0.8
5.8±0.4
5.0±0.7
2.1±0.2
3.7±0.9
3.0±0.6
2.5±0.2
5.3±0.4
8.3±0.7
2.6±0.4
4.5±0.9
3.7±0.7
3.2±0.4
7.5±0.9
4.7±0.4
2.3±0.4
3.8±0.5
4.2±0.3
2.9±0.5
6.2±0.8
8.8±1.1
3.0±0.4
6.1±0.8
5.2±0.4
4.2±0.7
7.8±0.9
1.9 ±0.3 2.4±0.2
3.1±0.7
2.7±0.4
2.3±0.3
5.5±0.8
4.0±0.8
3.8±0.4
2.8±0.5
5.9±0.9
(-O-CH2-O-)
Green highlight: parent (hecogenin); Violet highlights: parent of each optimization group; Pink
highlight: the most active semisynthetic analog.
14

(A)
H
N
H
7'
N
(B)
1'
O
2'
N
5'
H
4'
8'
S
12
O
H
H
H
H
3
30
HO
H
(D)
(C)
Figure 4. Binding mode of analog 30 at the MEK kinase domain (PDB: 3EQF). A. Chemical
structure of 30. B. Ribbon diagram of the binding pose of 30 (maroon ball and stick cartoon) at
the MEK kinase domain. This space-filling model revealed the good fitting of 30 within the
proximity of Phe209 and Cys207 residues satisfying the C-12 extension tactic. C. Binding pose
of 30 within the MEK kinase domain using CPK format. D. Space filling model showing the
protein target canonical kinase domain, with the N-terminal lobe at the top and the C-terminal
lobe at the bottom. All parts are from a similar vantage point.
To add further justification to the docking results, the native ligand of the MEK crystal
structure (PDB: 3EQF) was docked into its ATP binding site, applying the same parameters
which have been used for docking analog 30 (Figure 5). The bound conformation of co-
crystallized ligand was obtained with a good root mean square displacement (RMSD) of 0.2 Å,
suggesting the robustness of the docking experiments. In the investigated MEK crystal structure
PDB 3EQF, 30 demonstrated partial overlay with almost the same interactions (Figure 5).
15

Together, in silico data suggested the potential of 30 as a propitious MEK inhibitory hit
appropriate for further validation.
Figure 5. (a) Overview of MEK domain (PDB: 3EQF) represented as a ribbon diagram in
complex with 1 in green and 30 in maroon with ball and stick cartoon format. (B) Superposition
of 1 in green and 30 in maroon with ball and stick cartoon format. (C) Superposition of K252A
(MEK ATP-competitive inhibitor in orange) with 30 in maroon.
Many therapies, while killing the bulk of cancerous cells, may ultimately fail to
discriminate normal healthy cells from their purported malignant cells. Accordingly, anticancer
selectivity of 30 was evaluated using the human mammary epithelial cell line MCF-10A. These
cells are immortalized and non-tumorigenic with some features of normal breast epithelium,
including lack of anchorage-independent growth and dependence on growth factors and
hormones for proliferation and survival.37 Cells were exposed to various concentrations of 30 as
stated per experimental section. The results showed that 30's various treatments (1-20 μM)
displayed non-significant effect on MCF-10A viability, compared to their respective vehicle
control-treated group, reinforcing the 30's selective anticancer effect towards malignant breast
cells.
Measurement of intracellular contents leakage through impaired plasma membrane has
been widely used to assess chemical cytotoxicity.38 Lactate dehydrogenase (LDH) is a soluble
cytoplasmic enzyme allocated in almost all cells and is released into extracellular space in
16

respond to plasma membrane destruction.38 Therefore; the detection of LDH in the culture
medium can be used as a marker for cytotoxicity. Herein, the Cayman’s LDH cytotoxicity assay
kit was used to evaluate the ability of 30 to induce the LDH release in MDA-MB-231 cells in
comparison to its parent 1. Four different doses of tested compounds were used to assess targeted
cell death. Sapogenin 1 showed 27.8 % cytotoxicity at the maximum tested dose (100 μM),
which is more than three-fold its reported antiproliferative IC50. Analog 30 achieved 5.9%
cytotoxicity at the maximum tested dose (10 μM), which is more than five-fold its reported
antiproliferative IC50 (Table 2). This clearly suggested that 30 did not show significant
cytotoxicity up to several-fold its antiproliferative IC50 values and therefore its effect is mainly
due cytostatic, which is among distinctive features for targeted therapies.
Table 2. Percent cytotoxicity of 1 and 30 evaluated by LDH release from MDA-MB-231 cells.
Hecogenin (1)
25 50
0.04% 2.72% 10.3% 27.8%
Triton X
Analog 30
1.25 2.5
0.74% 1.27% 1.80% 5.94%
10%
75
100
5.0
10.0
Conc. (μM)
100%
%Cytotoxicity
Although cancer cell proliferation is often regarded as the most important aspect of
cancer progression, however, other key aspects play a crucial role in cancer progression and
metastasis, including migration and invasion.39 Hence, interfering with both processes could
have positive impacts on patient survival. The inhibitory effects of 1 on the migration and
invasion capacities of the highly metastatic MDA-MB-231 cells using wound-healing (WHA)
and CultreCoat® invasion assays, respectively, was probed at various non-toxic doses. In WHA,
after the confluent cancerous cells exposed to a treatment period of 24 h, 1 significantly
repressed the MDA-MB-231 cells migration towards the denuded zone in a direct function of
increasing efficacy with treatment concentrations (Figure 6B). Treatment with 30 resulted in a
dose-dependent cell migration inhibition, with a calculated IC50 of 2.5 µM (Figure 6B).
17

(A)
(B)
Figure 6. Effects of 1 and 30 on the TNBC MDA-MB-231 cells migration capacity using WHA.
(A) Representative microscopic images of created wounds at zero time and 24 h post-incubation
with vehicle as negative control or various treatment concentrations. (B) Dose response curve of
1 and 30 treatments versus percent wound closure.
18

On the other hand, 1 and 30 effects on MDA-MB-231 cells invasion across the
extracellular matrix was probed at four different doses. 24 h post plating, most of seeded cells
within the control group invaded the coating extract, thus the upper surface comprised minimal
cell density. Analog 30 significantly inhibited cancerous cells invasion, with maximum percent
invasion inhibition calculated as 92.4% at 5μM (maximum tested dose) indicated by higher cell
density at the upper surface of the invasion chamber (Fig. 7). On the other hand, treatment with
1at the maximum tested dose (80 μM) significantly inhibited cell invasion, with 86.2% invasion
inhibition (Fig. 7). The calculated IC50for 30 was 2.2µM, compared to 37.1 μM for the parent 1.
(A)
(B)
Figure 7. Effects of analog 30, compared to its parent 1 against the TNBC MDA-MB-231 cells
invasive capacity using CultreCoat® cell invasion assay. (A) Representative microscopic images
of upper chambers (non-invasive cell density) with the vehicle, 1 or 30. (B) Dose response effect
of 1 and 30 treatments versus average percent cell invasion.
Invasive breast cancer subtypes, including TNBC MDA-MB-231 cells, are associated
with augmented aberrant cellular proliferation and invasiveness pathways, just as the mitogen-
19

activated protein kinases (MAPKs).34 MAPK regulates diverse cellular programs that
coordinately regulate cell proliferation, differentiation, motility, and survival. The wide range of
functions regulated by the MAPK is mediated through phosphorylation of several downstream
substrates, including members of a family of protein kinases termed MAPK-activated protein
kinases (MAPKAPKs) as mitogen- and stress-activated kinases (MSKs) and MAPK-interacting
kinases (MNKs).40,41
Accordingly, to confirm the MEK inhibitory activity of 30, Western blot analyses was
implemented using MDA-MB-231 cells lysate, following administration of two different doses
of each of 1 and 30, along with a vehicle control treatment. The phosphorylation levels of the
specific MEK downstream effector, ERK (MAPK) was significantly reduced in both 1 and 30
treatment groups, compared to vehicle control treatment (Figure 8B). Interestingly, MSK
activation was attenuated due to ERK-phosphorylation inhibition, which subsequently, at least
in-part, hampered the proliferation, survival, motility and invasiveness of breast cancer cells.
Collectively, immunoblot results strongly support the ability of 30 to inhibit MEK catalytic
activity and thus interfering with MAPK cellular-related functions.
Significant in vitro activity of analog 30 against MDA-MB-231 cells (Table 1) motivated
subsequent in vivo study is to assess its anticancer efficacy concurrently with its parent 1 in a
pertinent breast cancer xenograft model. Athymic nude Foxn1nu/Foxn1+ mice orthotopically-
transplanted with MDA-MB-231/GFP cells were used. Five days after latency period after tumor
cells implantation, solid tumors became palpable. Mice then were intraperitoneally administrated
equivalent 10 mg/kg/3X week doses 1 (group I) or 30 (group II), while control group received
vehicle only. Treatments were continued for twenty-eight days. Mice were observed and body
weights were monitored per dosing for afterwards preliminary toxicity signs. Both treatments
20

were endurable without overt signs of neither toxicity nor significant body weight change,
compared to the vehicle-control group. Consistent with the in vitro lack of toxicity against the
non-tumorigenic MCF10A mammary epithelial cells, nude mice showed good tolerance to the
both treatments. Tumor growth was monitored and quantified. On the basis of caliper
measurements, treatment groups manifested significant tumors growth attenuation by the end of
the study, compared to the vehicle control group. Hecogenin (1) induced 57.4% tumor growth
inhibition while analog 30 treatments induced 78.1% tumor inhibition (Table 3 and Figure 9).
Ultimately, 30 significantly suppressed cell proliferation and attenuate tumor growth in this
orthotopic model of TNBC and therefore qualified as a lead compound.
(A)
(B)
Figure 8. (A) Overview of the MAPK pathway. (B) Western blot analyses of in vitro vehicle
control and 30 treatments in the TNBC MDA-MB-231 cells. Data show downregulation of
activated MAPK kinase and downstream effectors, ERK and MSK in both treatment groups
compared to vehicle control.
21

Figure 9. In vivo anticancer activities of 1 and 30 against MDA-MB-231/GFP breast cancer cells
in athymic nude mouse xenograft model. (A) Effects of 1 and 30 treatments on the mean tumor
volume throughout the study compared to vehicle control group. (B) Percent tumor growth (TG)
in treatment groups, compared to vehicle control group calculated at the study end. (C) Pictorial
representation of mice from control and treatment groups at the study end.
Table 3. In vivo anticancer activity of 1 and 30 against the TNBC MDA-MB-231 cells in nude
mouse xenograft model.
Group
Initial tumor
volume (TV0)
51.3 ±1.9
49.4 ±2.7
50.2 ±1.4
Final tumor
volume (TV0)
1837.0 ±281.8
810.7 ±140.2
441.3 ±88.8
Dose
% Growth
(mg/kg/day) inhibition
-
10 mg
10 mg
Control
Group I
Group II
-
57.36%
78.09%
Conclusions
Rational design of natural products-based semisynthetic analogs is a versatile strategy to
optimize parent bioactive hits, improve their lead-like and druggability characters, and promote
22

them as future clinical candidates. This study identified the natural sapogenin hecogenin (1) and
its rationally modified semisynthetic analogs as potential MAPK pathway inhibitors for the
control of breast malignancies. The thiosemicarbazone analog 30 showed a robust antitumor
efficacy that was correlated with the inhibition of MAPK kinase signaling in highly proliferating
malignant cells and attenuate their growth and invasiveness. The in vivo data reinforced the in
vitro cellular conclusions and promoted30 to the lead rank, validating its potential for future use
to control breast malignancies with dysregulated MAPK pathway.
Experimental section
General experimental procedures
Thin-layer chromatography analysis was carried on aluminum pre-coated Si gel 60 F254
TLC plates (EMD Chemicals Inc., Gibbstown, NJ, USA), while Si gel 60 (230-400 mesh,
Natland International Corporation, Morrisville, NC, USA) was used for column chromatography.
UV light and freshly prepared p-anisaldehyde–methanol–acetic acid–sulfuric acid (2:170:20:10
v/v/v/v) used as a spray reagent for chemical visualization.1H and 13C NMR spectra were
obtained on a JEOL Eclipse ECS-400 NMR spectrometer (Boston, MA, USA) operated at 400
and 100 MHz, respectively. Delta™ NMR Data Processing Software (JEOL Inc., MA, USA)
was used for analysis and spectral processing with chemical shifts (δ) reported in parts per
million (ppm) relative to internal residual solvent signals (7.26 and 77.1 ppm for CDCl3; 2.04
and 29.8, 206.3 ppm for acetone-d6 in 1H and 13C NMR spectra, respectively). NMR data
assignments (Tables SI 1-15) were based on PENDANT, 1H-1H COSY, 1H-13C HMQC or
HMBC experiments. The following abbreviations were used to assign proton signal
multiplicities: s (singlet), d (doublet), t (triplet), m (multiplet), dd (doublet of doublet). The ESI-
MS carried out on AB Sciex-3200 QTRAP LC/MS/MS system (Applied Biosystems, Foster
23

City, CA) using Analyst version 1.4.1 software (MDS Sciex; Toronto, Canada). Analytes were
ionized using electrospray ionization (ESI) interfaced with standard turbo V ion source.
Chemicals, reagents, and antibodies
All chemicals were purchased from Sigma-Aldrich (St. Louis, MO), unless otherwise
stated. The steroidal sapogenin, hecogenin, (purity >90%; Catalog# 230015; CAS # 467-55-0)
was purchased from Natland International Incorporation (NC, USA). All reactions were
conducted in a well-ventilated fume hood using Teflon-coated magnetic stirrer bars at room
temperature. All solvents and reagents were purchased from Sigma-Aldrich (or Alfa Aesar and
used without further purification. All yields reported are isolated yields. All antibodies were
purchased from Cell Signaling Technologies (Beverly, MA) and used at a dilution of 1:1000.
Chemical synthesis
General procedure for preparation of 3-O-hecogenin esters
Hecogenin (40 mg, 0.093 mmol), Et3N (80 μL, 0.79 mmol) and DMAP (98 mg, 0.79
mmol) were dissolved in 3 mL dry pyridine. The mixture was stirred at room temperature for 20
min followed by gradual addition of 1.5 equivalents of different acid chlorides. The reaction was
set under continuous stirring overnight. Thereafter, water (10 mL) was added to quench the
reaction followed by shaking with saturated solution of NaHCO3 (10 mL x 3). Finally, the
aqueous phase was extracted with EtOAc (3 x 5 mL) and the organic layers were pooled,
concentrated, and dried over anhydrous Na2SO4. The crude mixture was purified using normal
phase Si gel 60 column chromatography eluted with n-hexanes-acetone 8:2 in isocratic mode to
afford analogs 2-9.
3β-O-(Benzoyl)-5α,25R-spirostan-12-one (2)
24

Following the abovementioned esterification procedure using benzoyl chloride afforded 2
in a 60% yield as a white amorphous powder. TLC: n-hexane-acetone, 7:3; ESI-MS m/z: 535.3
[M+H]+, calcd for C34H47O5. 1H and 13C NMR data; see Tables SI1 and SI10 (Supplementary
Information).
3β-O-(3'-Methoxybenzoyl)-5α,25R-spirostan-12-one (3)
General esterification procedure using 2-methoxybenzoyl chloride afforded 3 in a 70%
yield as a white amorphous powder. TLC: n-hexane-acetone, 7:3; ESI-MS m/z: 565.3 [M+H]+,
calcd for C35H49O6. 1H and 13C NMR data; see Tables SI1 and SI10 (Supplementary
Information).
3β-O-(3'-Chlorobenzoyl)-5α,25R-spirostan-12-one (4)
General esterification procedure using 2-chlorobenzoyl chloride afforded 4 in a 60%
yield as a white amorphous powder. TLC: n-hexane-acetone, 7:3; ESI-MS m/z: 569.5 [M+H]+,
calcd for C34H46ClO5. 1H and 13C NMR data; see Tables SI1 and SI10 (Supplementary
Information).
3β-O-(Furoyl)-5α,25R-spirostan-12-one (5)
General esterification procedure using 2-furoyl chloride afforded 5 in 71% yield as a
white amorphous powder, TLC: n-hexane-acetone, 7:3; ESI-MS m/z: 525.3 [M+H]+ calcd for
C32H45O6. 1H and 13C NMR data; see Tables SI2 and SI11 (Supplementary Information).
3β-O-(4'-Trifluoromethybenzoyl)-5α,25R-spirostan-12-one (6)
General esterification procedure using 3-trifluoromethyl benzoyl chloride afforded 6 in a
55% yield as a white amorphous powder. TLC: n-hexane-acetone, 7:3; ESI-MS m/z: 603.3
[M+H]+, calcd for C35H46F3O5. 1H and 13C NMR data; see Tables SI1 and SI10 (Supplementary
Information).
25

3β-O-(5'-Fluorobenzoyl)-5α,25R-spirostan-12-one (7)
General esterification procedure using 4-fluorobenzoyl chloride afforded 7 in 35% yield
as white amorphous powder, TLC: n-hexane:acetone, 7:3; ESI-MS m/z: 553.7 [M+H]+ calcd for
C34H46FO5. 1H and 13C NMR data; see Tables SI2 and SI10 (Supplementary Information).
3β-O-(5'-Cyanobenzoyl)-5α,25R-spirostan-12-one (8)
General esterification procedure using 4-cyanobenzoyl chloride afforded 8 in 68% yield
as white amorphous powder, TLC: n-hexane-acetone, 7:3; ESI-MS m/z: 560.3 [M+H]+ calcd for
C35H46NO5. 1H and 13C NMR data; see Tables SI2 and SI10(Supplementary Information).
3β-O-(5'-Nitrobenzoyl)-5α,25R-spirostan-12-one (9)
General esterification procedure using 4-nitrobenzoyl chloride afforded9in a 63% yield as
a white amorphous powder, TLC: n-hexane-acetone, 7:3; ESI-MS m/z: 580.3 [M+H]+ calcd for
C34H46NO7. 1H and 13C NMR data; see Tables SI2 and SI11 (Supplementary Information).
Preparation of 3-O-hecogenin-12-phenyl semicarbazones and phenylthiosemicarbazones
General procedure A
A solution of hecogenin (20 mg, 0.046 mmol) in dry EtOH (5 mL) and pyridine (0.5 mL)
mixture, was charged with three equivalents of appropriate functionalized phenyl
semicarbazide.HCl or phenyl thiosemicarbazide.HCl. The resulting mixture was stirred overnight
at RT. The mixture was diluted with ice-cold water and the precipitate formed was filtered and
washed with ice cold water (2 x 5 mL) after reaction completion. The combined solid was
lyophilized and then solubilized in EtOAc. The crude product was purified by flash column
chromatography using Si gel 60 eluted with gradients of n-hexanes-EtOAc (8:2 to 6:4), to afford
the corresponding hecogenin 12-semicarbazones or thiosemicarbazones.
3β-Hydroxy-5α,25R-spirostan-12-(phenyl-semicarbazone) (10)
26

General condensation procedure A, using 4-(phenyl)-3-semicarbazide.HCl afforded 10 in
30% yield as white amorphous powder, TLC: n-hexane-acetone-NH3, 7:3:0.5; ESI-MS m/z:564.7
[M+H]+ calcd for C34H50N3O4. 1H and 13C NMR data; see Tables SI3 and SI11 (Supplementary
Information).
3β-Hydroxy-5α,25R-spirostan-12-(4'-chloroyphenyl semicarbazone) (11)
General reaction procedure A using 4-(3-chlorophenyl)-3-semicarbazide.HCl afforded 11
in 46% yield as a white amorphous powder, TLC: n-hexane-acetone-NH3, 7:3:0.5; ESI-MS m/z:
598.4 [M+H+], calcd for C34H49ClN3O4. 1H and 13C NMR data; see Tables SI3 and SI11
(Supplementary Information).
3β-Hydroxy-5α,25R-spirostan-12-(5'-chloroyphenyl semicarbazone) (12)
General reaction procedure A using 4-(4-chlorophenyl)-3-semicarbazide.HCl afforded 12
in 40% yield as a white amorphous powder, TLC: n-hexane-acetone-NH3, 7:3:0.5; ESI-MS m/z:
598.4 [M+H+] calcd for C34H49ClN3O4. 1H and 13C NMR data; see Tables SI3 and SI11
(Supplementary Information).
3β-Hydroxy-5α,25R-spirostan-12-(4'-trifluoromethylphenyl semicarbazone) (13)
General reaction procedure A using 4-(3-trifluoromethylphenyl)-3-semicarbazide.HCl
afforded 13 in 37% yield as a white amorphous powder, TLC: n-hexane-acetone-NH3, 7:3:0.5;
ESI-MS m/z: 632.4 [M+H]+ calcd for C35H49F3N3O4. 1H and 13C NMR data; see Tables SI3 and
SI11 (Supplementary Information).
3β-Hydroxy-5α,25R-spirostan-12-[(5'-methylphenylsulphonyl)semicarbazone] (14)
General reaction procedure A using 4-(4-methylphenylsulfonyl)-3-semicarbazide.HCl
afforded 14 in 52% yield as a white amorphous powder, TLC: n-hexane-acetone-NH3, 7:3:0.5;
27

ESI-MS m/z: 642.6 [M+H+] calcd for C35H52N3O6S. 1H and 13C NMR data; see Tables SI3 and
SI12 (Supplementary Information).
General procedure B
Hecogenin (20 mg, 0.046 mmol) was dissolved in dry EtOH (5 mL) and pyridine (0.5
mL) mixtures, and then charged with three equivalents of appropriate phenyl semicarbazide or
phenyl thiosemicarbazide. The resulting mixture was stirred overnight at room temperature.
Water was added to quench the reaction and the product was extracted with EtOAc (3 x 5L) after
reaction completion. The combined organic layers dried were dried over anhydrous Na2SO4 and
concentrated under vacuum. The crude residue was purified using flash column chromatography
over silica gel 60, eluted with gradients of n-hexanes-EtOAc-NH4OH (8:2:0.5 to 6:4:0.5), to
afford the corresponding hecogenin 12-phenyl semicarbazones or phenyl thiosemicarbazones.
3β-Hydroxy-5α,25R-spirostan-12-(phenyl thiosemicarbazone) (15)
General reaction procedure Busing 4-phenyl thiosemicarbazide, afforded 15 in 78% yield
as a white amorphous powder, TLC: n-hexane-acetone-NH3, 7:3:0.5; ESI-MS m/z: 580.3
1
[M+H]+ calcd for C34H50N3O3S. H and 13C NMR data; see Tables SI4 and SI12 (Supplementary
Information).
3β-Hydroxy-5α,25R-spirostan-12-(3'-chloroyphenyl thiosemicarbazone) (16)
General reaction procedure Busing 4-(2-chlorophenyl)-3-thiosemicarbazide afforded 16
in 65% yield as a white amorphous powder, TLC: n-hexane-acetone-NH3, 7:3:0.5; ESI-MS m/z:
614.5 [M+H+] calcd for C34H49 ClN3O3S.1H and 13C NMR data; see Tables SI4 and SI12
(Supplementary Information).
3β-Hydroxy-5α,25R-spirostan-12-(5'-chlorophenyl thiosemicarbazone) (17)
28

General reaction procedure B using 4-(4-chlorophenyl)-3-thiosemicarbazide afforded 17
in 46% yield as a white amorphous powder, TLC: n-hexane-acetone-NH3, 7:3:0.5; ESI-MS m/z:
614.5 [M+H+] calcd for C34H49ClN3O3S. 1H and 13C NMR data; see Tables SI4 and SI12
(Supplementary Information).
3β-Hydroxy-5α,25R-spirostan-12-(3',5'-dichlorophenyl thiosemicarbazone) (18)
General reaction procedure Busing 4-(2,4-dichlorophenyl)-3-thiosemicarbazide afforded
18 in 58% yield as a white amorphous powder, TLC: n-hexane-acetone-NH3, 7:3:0.5; ESI-MS
m/z: 648.3 [M+H+] calcd for C34H48Cl2N3O3S. 1H and 13C NMR data; see Tables SI5 and SI12
(Supplementary Information).
3β-Hydroxy-5α,25R-spirostan-12-(3',7'-dichlorphenyl thiosemicarbazone) (19)
General reaction procedure Busing 4-(2,6-dichlorophenyl)-3-thiosemicarbazide afforded
19 in 45% yield as a white amorphous powder, TLC: n-hexane-acetone-NH3, 7:3:0.5; ESI-MS
m/z: 648.3 [M+H+] calcd for C34H48Cl2N3O3S. 1H and 13C NMR data; see Tables SI5 and SI12
(Supplementary Information).
3β-Hydroxy-5α,25R-spirostan-12-(5'-fluorophenyl thiosemicarbazone) (20)
General reaction procedure B using 4-(4-flourophenyl)-3-thiosemicarbazide afforded 20
in 30% yield as a white amorphous powder, TLC: n-hexane-acetone-NH3, 7:3:0.5; ESI-MS m/z:
598.3 [M+H]+ calcd for C34H49FN3O3S. 1H and 13C NMR data; see Tables SI5 and SI12
(Supplementary Information).
3β-Hydroxy-5α,25R-spirostan-12-(3'-nitrophenyl thiosemicarbazone) (21)
General reaction procedure Busing 4-(2-nitrophenyl)-3-thiosemicarbazide afforded 21 in
72% yield as a yellow amorphous powder, TLC: Rf=0.25 (n-hexane-acetone-NH3, 7:3:0.5); ESI-
29