Galloyl esters of trans-stilbenes are inhibitors of FASN with anticancer activity on non-small cell lung cancer cells

Article abstract of DOI:10.1016/j.ejmech.2019.111597

Fatty acid synthase (FASN) is a lipogenic enzyme that is selectively upregulated in malignant cells. There is growing consensus on the oncogenicity of FASN-driven lipogenesis and the potential of FASN as a druggable target in cancer. Here, we report the synthesis and FASN inhibitory activities of two novel galloyl esters of trans-stilbene EC1 and EC5. Inhibition of FASN was accompanied by a loss in AKT activation and profound apoptosis in several non-small cell lung cancer (NSCLC) cells at the growth inhibitory concentrations of EC1 and EC5. Both FASN and phospho-AKT levels were concurrently downregulated. However, addition of a lipid concentrate to the treated cells reinstated cell viability and reversed the loss of FASN and AKT protein levels, thus recapitulating the causal relationship between FASN inhibition and the loss in cell viability.

Full text of DOI:10.1016/j.ejmech.2019.111597

Journal Pre-proof
Galloyl esters of trans-stilbenes are inhibitors of FASN with anticancer activity on non-
small cell lung cancer cells
Yu-Jia Tan, Azhar Ali, Sheng-Yang Tee, Jun-Ting Teo, Xi Yu, Mei-Lin Go, Yulin Lam
PII:
S0223-5234(19)30731-7
DOI:
https://doi.org/10.1016/j.ejmech.2019.111597
EJMECH 111597
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 4 June 2019
Revised Date: 17 July 2019
Accepted Date: 5 August 2019
Please cite this article as: Y.-J. Tan, A. Ali, S.-Y. Tee, J.-T. Teo, X. Yu, M.-L. Go, Y. Lam, Galloyl esters
of trans-stilbenes are inhibitors of FASN with anticancer activity on non-small cell lung cancer cells,
European Journal of Medicinal Chemistry (2019), doi: https://doi.org/10.1016/j.ejmech.2019.111597.
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 Masson SAS.

Galloyl esters of trans-stilbenes are inhibitors of FASN with anticancer activity on non-
small cell lung cancer cells
a #
b #
a
b
c
c
Yu-Jia Tan , Azhar Ali , Sheng-Yang Tee , Jun-Ting Teo , Xi Yu , Mei-Lin Go *, Yulin
Lam *
a
a
Department of Chemistry, National University of Singapore, 3 Science Drive 3, 117543,
b
Singapore. Cancer Science Institute of Singapore, Yong Loo Lin School of Medicine, 14
c
Medical Drive, 117599, Singapore; Department of Pharmacy, National University of Singapore,
1
8 Science Drive 4, 117543, Singapore.
#
Co-First Authors; * Co-Corresponding authors.
Abstract
Fatty acid synthase (FASN) is a lipogenic enzyme that is selectively upregulated in malignant
cells. There is growing consensus on the oncogenicity of FASN driven lipogenesis and the
potential of FASN as a druggable target in cancer. Here, we report the synthesis and FASN
inhibitory activities of two novel galloyl esters of trans-stilbene EC1 and EC5. Inhibition of FASN
was accompanied by a loss in AKT activation and profound apoptosis in several non-small cell
lung cancer (NSCLC) cells at the growth inhibitory concentrations of EC1 and EC5. Both FASN
and phospho-AKT levels were concurrently downregulated. However, addition of a lipid
concentrate to the treated cells reinstated cell viability and reversed the loss of FASN and AKT
protein levels, thus recapitulating the causal relationship between FASN inhibition and the loss
in cell viability.
Keywords FASN inhibitors, anticancer, galloyl esters of trans-stilbenes, Non-small cell lung
cancer, EGCG
1
. Introduction
The multifunctional enzyme fatty acid synthase (FASN) catalyzes the final steps in the de novo
synthesis of fatty acids starting from acetyl-CoA, malonyl-CoA and NADPH [1–3]. The main
product of this process is the 16-C fatty acid palmitate which is a major component of cell
membranes, a precursor in the synthesis of structural lipids, a substrate in post-translational
palmitoylation of membrane proteins and an important energy source when incorporated into
triglycerides [4]. Healthy cells have low levels of FASN because their fatty acid requirements
are largely met by dietary sources and not by de novo lipogenesis [5]. In contrast, malignant
cells exhibit unusually high levels of FASN activity, even when dietary lipids are present in
abundance [6,7]. Upregulated FASN activity is an early and pervasive hallmark of malignancies
1

that is inextricably linked to disease progression and poor prognosis [8–10]. Elevated FASN
levels confer growth and survival benefits to cancer cells. Ostensibly, FASN signaling is
correlated to the expression and activation of oncogenic proteins that direct malignant
transformations [11–13]. Overexpression of the FASN gene has been shown to activate
HER1/HER2 tyrosine kinase receptors in human breast epithelial cells, underscoring a bona fide
oncogenic role for FASN [14]. Furthermore, RNA interference silencing of the FASN gene
induced apoptosis in prostate cancer cells [15], synergistically enhanced the cytotoxicity of
paclitaxel on breast cancer cells [16] and led to stable reversion of the malignant phenotype
[
17]. Collectively, these findings recapitulate a compelling role for FASN as a viable therapeutic
target for cancer drug discovery [18,19].
First generation FASN inhibitors comprised several natural products such as cerulenin,
thiolactomycin and (-)-epigallocatechin 3-gallate (EGCG) [20–22]. The green tea flavonoid
EGCG is a gallate ester of epigallocatechin. Structurally it is a diastereomeric hydroxylated
chroman with polyphenolic side chains at positions 2 (trihydroxyphenyl) and 3 (galloyloxy)
(
Figure 1A). EGCG is associated with a diversity of pharmacological properties which include
neuroprotective, anticancer, cardiovascular and antimicrobial activities [23–26]. The pleiotropic
profile of EGCG may be linked to the abundance of phenolic groups in its structure [27]. Ortho-
disubstituted phenolic hydroxyl (OH) groups in EGCG are efficient redox cyclers and have dual
functions as pro-oxidants and antioxidants [28,29]. In addition, polyphenolic groups impart
amphiphilicity to the molecule, predisposing it to non-specific perturbation of membrane bilayers
and consequential changes to membrane protein function [30]. Chemical reactivity also resides
in the ester linkage of EGCG which may react with nucleophilic serine/threonine OH groups in
enzymes to bring about inhibition. Notwithstanding, the inhibition of FASN by EGCG alludes to a
specific interaction that is characterized by well-defined binding parameters [31]. The reaction
kinetics pointed to an initial rapid and reversible binding to the ketoacyl reductase enzyme in the
FASN complex, followed by slow irreversible inactivation [31]. EGCG was however a weak
FASN inhibitor with an IC50 of 52 µM [31] and this has prompted attempts at understanding the
structure activity requirements of the EGCG template with the aim of enhancing FASN inhibition
[
32,33]. To date, the most successful analogs were derived by replacing the chroman scaffold of
EGCG with other ring systems [34]. Turrado and coworkers tested five scaffolds (cyclohexyl,
phenyl, naphthalenyl, tetrahydronaphthalenyl, biphenyl) and found significant FASN inhibitory
activity in only two - biphenyl and naphthalenyl – as exemplified by compounds 1 and 2 (Figure
1
B) [34]. These findings hint at untapped opportunities for enhancing inhibitory potency through
scaffold replacement in the EGCG template. Here, we confirmed the suitability of trans-stilbene
2

as an incipient scaffold with FASN inhibitory properties. Inhibition by the most promising analogs
disrupted the FASN-AKT signaling axis to bring about apoptotic cell death in non-small cell lung
cancer (NSCLC) cells.
Figure 1. Chemical structures of (A) EGCG; (B) FASN inhibitors biphenyl 1 and naphthalenyl 2 [30]; (C) EGC, ECG
and resveratrol.
2
. Chemistry
Of the three identifiable structural features in EGCG (galloyloxy, trihydroxyphenyl, hydroxylated
chroman), only the galloyloxy side chain is essential for inhibition. Thus, epigallocatechin (EGC,
Figure 1C) which lacks this entity was a significantly weaker inhibitor than EGCG [35]. The role
of the trihydroxyphenyl ring is less well-defined. Epicatechin gallate (ECG, Figure 1C) has one
less hydroxyl (OH) group on the phenyl ring than EGCG but is comparable in terms of inhibitory
activity [35]. As for the chroman ring, it acts primarily as a scaffold for positioning the other two
entities at a specific distance from each other and to keep them in a near planar arrangement
[
33]. In our efforts to identify other scaffolds that could serve as superior substitutes to chroman,
we selected trans-stilbene as a viable alternative, prompted mainly by its presence in the
polyphenol resveratrol (Figure 1C) which has FASN inhibitory activity exceeding that of several
naturally occurring flavones [33]. Accordingly, we designed stilbene analogs EC1-EC5, EC10-
1
2 bearing the galloyloxy side chains found in EGCG (Figure 2). Modifications in these analogs
3

centered on (i) varying the number of phenolic OH groups on these side chains from three in
EC1 to two (EC2) and one (EC3); (ii) positioning the galloyloxy side chains at the meta-para
(
3/4’) positions in EC1 and para-para (4/4’) in EC5; (iii) reducing the number of galloyloxy side
chains from two (EC1, EC5) to one (EC10, EC11) and (iv) undertaking phenolic OH to methoxy
conversions in EC4, EC7 and EC12. We also modified the previously reported “active” biphenyl
scaffold of 1 by inserting a methylene residue between the phenyl rings to give the
conformationally more flexible diphenylmethyl scaffold which is found in EC6-EC8 (Figure 2). In
the same way, we replaced the rigid naphthalene in 2 with the more flexible
tetrahydronaphthalenyl ring in EC9. We also explored a 1-phenylnaphthalene analog EC13
which was substituted with only one galloyloxy side chain. The structures of EC9 and EC13 are
shown in Figure 2.
4

R1
R2
R3
OH
O
OH
OH
O
O
O
O
O
O
HO
HO
O
OH
R1
R3
R2
R1
R2
R3
Compound
EC5
EC1
EC2
OH
OH
H
OH
OH
OH
H
EC3
EC4
OH
OMe
H
OMe
OMe
OH
OH
OH
O
O
O
O
O
HO
HO
OH
OH
R
R
R
R
O
O
O
O
O
O
HO
HO
OH
OH
O
R
R
OH
Compound
R
EC8
EC9
EC6
EC7
OH
OMe
O
O
O
O
R
R
HO
HO
O
O
R
OH
HO
OH
OH
EC10
Compound
R
EC13
EC11
EC12
OH
OMe
Figure 2. Chemical structures of EC analogs EC1 - EC13.
The synthesis of the EC analogs was achieved by a convergent approach in which the two main
fragments (OH-bearing scaffolds, substituted benzoic acids) were separately synthesized and
condensed. Except for the OH-bearing diphenylmethanes 13, 14 and tetrahydronaphthalene 15
(found in EC6-EC8 and EC9 respectively) which were commercially available, the hydroxylated
ring structures of the remaining compounds were obtained by synthesis.
The trans-stilbene scaffold was obtained by Heck coupling of a substituted styrene and
substituted iodobenzene in the presence of catalytic palladium (II) acetate. The synthesis of
trans-3,4-dihydroxystilbene 5a is shown in Scheme 1. Briefly, 3-hydroxybenzaldehyde was
5

reacted with methyltriphenylphosphonium bromide and n-butyllithium in a Wittig reaction to give
-hydroxystyene 3. The OH groups in 3 and 4-iodophenol were acetylated to give 4a and 5a
3
respectively, followed by Heck coupling of 4a and 5a to give the diacetoxystilbene 6a. The
protecting acetyl groups in 6a were then removed by alkaline hydrolysis to give
3
,4’-dihydroxystilbene 7a which is the core scaffold in EC1-EC4. The same reaction sequence
was followed to give 3-hydroxystilbene 7b (from 4a and 5b) and 4-hydroxystilbene 7c (from 4b
and 5a) which are the scaffolds in EC10 and EC11, EC12 respectively. 4,4’-Dihydroxystilbene 8
which is found in EC5 was prepared by Perkin condensation of 4-hydroxybenzaldehyde and 4-
hydroxyphenylacetic acid in a microwave-assisted reaction (Scheme 2). To obtain the 1-(4-
hydroxyphenyl)naphthalene scaffold of EC13, 1-naphthaleneboronic acid and 4-bromoanisole
were subjected to Suzuki coupling to give 1-(4-methoxyphenyl)naphthalene 9 which on
demethylation yielded the desired phenolic analog 10 (Scheme 3).
Scheme 1. Reagents and conditions: (a) CH
3 3 2 3 2 2
PPh Br, n-BuLi, THF, 0 °C to r.t., 4 h (b) Ac O, DMAP, Et N, CH Cl ,
r.t., 1.5 h (c) Pd(OAc) , PPh , Et N, DMF, 75 °C, 18 h, 50-66% (d) LiOH, THF, 100 °C, 4 h, 85-96%.
2
3
3
Scheme 2. Reagents and conditions: Piperidine, 1-methylimidazole, PEG 4600, MW, 160 °C, 25 min, 3%.
6

2 2 3 2 2
Scheme 3. Reagents and conditions: (a) Pd(OAc) , Ba(OH) , 95% aq. EtOH, r.t., 17 h, 74% (b) BBr , CH Cl , 0 °C to
r.t., 12 h, 96%.
To minimize side reactions in the final esterification step, the phenolic OH groups in 11a-c were
reacted with tert-butyldimethylsilyl chloride (TBSCl) to give the corresponding silyloxy analogs
1
2a-c (Scheme 4).
2
Scheme 4. Reagents and conditions: (i) TBSCl, imidazole, DMF, r.t., 24 h (ii) AcOH/H O (3:1), THF, r.t., 24 h,
7
9-88% over 2 steps.
Scheme 5 depicts the final carbodiimide-catalyzed esterification of the phenolic/alcoholic OH
groups in the various ring structures with the silyloxy benzoic acids. Removal of the protecting
silyl groups yielded the final compounds EC1-EC3, EC5, EC6, EC8 and EC9 bearing two
galloyloxy side chains. The methoxylated analogs EC4 and EC7 were obtained by directly
esterifying trimethoxybenzoic acid 12d with stilbene 7a and diphenylmethane 13. The mono-
galloyl analogs EC10, EC11, EC12 and EC13 were synthesized in a similar manner as depicted
in Scheme 6.
7

2 2 4
Scheme 5. Reagents and conditions: (a) EDCI or DCC, DMAP, CH Cl , r.t., 12 h, 51-85% (b) TBAF or NH F,
THF/MeOH, r.t., 6 h, 38-73%.
8

O
R4
R5
OH
HO
Backbone
+
R6
7
b, 7c, 10
4
5
6
1
1
2a R , R , R = OTBS
4
5
6
2d R , R , R = OMe
HO
R
a
7
7
b R = 3-OH
c R = 4-OH
10
R4
R6
R1
R3
O
O
O
R5
R2
b
O
Backbone
Backbone
R4
R5
R6
Backbone
R1
R2
R3
Backbone
R
R
1
6h
3-OH
4-OH
OTBS OTBS OTBS
OTBS OTBS OTBS
EC10 3-OH
EC11 4-OH
OH
OH
OH
OH
OH
OH
1
6i
R
R
EC12 4-OH
OMe
OMe
OMe
HO
HO
1
6j
OTBS OTBS OTBS
EC13
OH
OH
OH
2 2 4
Scheme 6. Reagents and conditions: (a) EDCI or DCC, DMAP, CH Cl , r.t., 12 h, 57-63% (b) TBAF or NH F,
THF/MeOH, r.t., 6 h, 43-88%.
3
3
. Results and Discussion
.1. The galloyl stilbenes EC1 and EC5 potently inhibited cell proliferation across a panel
of FASN expressing NSCLC cell lines
EC1-EC13 were initially assessed for growth inhibition on the FASN expressing NSCLC H1703
cell line at a single concentration of 25 µM (Figure 3). Growth inhibition was determined after 72
h using the tetrazolium-based MTS assay and only compounds that inhibited growth by more
than 50% inhibition were shortlisted as “actives”. Based on this cut-off, all the methoxylated
analogs (EC4, EC7, EC12) were excluded, as were the monosubstituted analogs EC10 and
EC13.
9

100
*
**
***
***
50 % inhibition
50
***
***
***
***
***
***
***
***
0
Figure 3. Viability of NSCLC H1703 cells on treatment with EC1-EC13 at a single concentration of 25 µM. Data
represents the mean ± standard deviation (SD) of n = 3 independent determinations. Significant difference from
control was analysed by unpaired t test using GraphPad Prism 5 and indicated by an asterisk (*p<0.05, **p<0.001,
***p<0.0001).
The shortlisted actives were then evaluated for their growth inhibitory concentrations (IC ) on
50
H1703 cells, alongside EGCG and the reported biphenyl 1 (Table 1). Three structure-activity
trends were noted from these results. First, among the four scaffolds examined, trans-stilbene
emerged as the most promising. It is found in the two most potent analogs EC1 and EC5 (IC50
1
µM) which gratifyingly, also exceeded 1 in terms of potency. Second, potency is conditional on
di-substitution of the template with galloyloxy side chains. The requirement for di-substitution
alluded to a possible regioisomeric preference but this was not apparent in EC1 and EC5 which
were equipotent in spite of differences in the positioning of the galloyloxy entities (meta-para in
EC1, para-para in EC5). On the other hand, regioselectivity was clearly demonstrated in the
diphenylmethyl analogs EC6 and EC8. Here, there was a distinct preference for a para-para
location (EC6) as compared to ortho-ortho (EC8), implicating close proximity of the galloyloxy
side chains as a structural liability. Yet another instance of regioselectivity was apparent in the
monosubstituted stilbenes EC10 and EC11. Here, the meta-substituted EC10 was eliminated as
an inactive in the initial single concentration screen but its para-substituted counterpart (EC11)
retained modest potency (IC50 11 µM). Lastly, all three OH groups on the galloyloxy side chain
were required for optimal growth inhibition. Thus, reduction in OH groups from three (EC1) to
two (EC2) and then one (EC3) led to progressive losses in activity.
1
0

To augment these structure-activity trends, the EC compounds were tested on two additional
FASN expressing NSCLC cell lines H1299 and H1650 (Table 1). In spite of variations in IC50
values, the rank order of each analog was broadly similar across the three cell lines. Notably,
EC1 and EC5 were still the most potent members although they were now comparable to 1 on
H1299 (in contrast to H1703) and only EC5 ranked above 1 on H1650. Similarly, EC3 and EC8
were consistently flagged out for sub-optimal growth inhibition on all cell lines. Taken together,
the structure-activity trends observed in H1703 corroborated with those observed on the other
cell lines.
Table 1. IC50 values and selective indices (EC1, EC5, EGCG only) of compounds on NSCLC cells and non-malignant
lung epithelial SAEC cells.
a,b
IC50 (µM)
Compound
H1703
H1299
H1650
PC-9
A549
SAEC
EC1
1.28 ± 0.02
3.1 ± 0.3
(2.2)
12.7 ± 0.9
5.0 ± 0.5
2.3 ± 0.1
6.8 ± 0.6
c
(SI )
(5.3)
(0.5)
(1.4)
(3.0)
d
d
d
EC2
EC3
8.9 ± 0.8
17.2 ± 1.2
11.0 ± 0.7
37.2 ± 2.0
14.1 ± 0.8
40.2 ± 2.0
-
-
-
-
-
d
d
d
-
EC5
0.97 ± 0.10
3.3 ± 0.1
(1.5)
2.5 ± 0.1
(2.0)
2.3 ± 0.1
(2.2)
2.1 ± 0.1
(2.4)
5.0 ± 1.0
c
(
SI )
(5.2)
d
d
d
EC6
EC8
EC9
6.2 ± 0.4
39.3 ± 1.2
6.5 ± 0.4
11.2 ± 0.1
8.6 ± 0.3
25.9 ± 2.1
10.2 ± 0.3
8.5 ± 0.4
49.7 ± 3.1
122 ± 19
36.0 ± 2.6
16.0 ± 1.2
-
-
-
d
d
d
-
-
-
-
-
-
d
d
d
d
d
d
EC11
-
-
-
EGCG
17.1 ± 1.5
20.6 ± 1.4
24.8 ± 1.9
77.4 ± 1.1
185 ± 5
(0.6)
110 ± 14
c
(SI )
(6.4)
(5.3)
(4.4)
(1.4)
d
d
d
1
5.8 ± 0.2
4.2 ± 0.4
7.4 ± 0.4
-
-
-
a
IC50: Concentration of compound required to reduce cell viability by 50%. Evaluated using MTS assay after 72 h at
7 °C, 5% CO . Data represents mean ± SD values from n = 3 independent experiments respectively. The NSCLC
2
b
3
cell lines were FASN expressing as revealed by immunoblotting (Supplementary Information, Figure S1). In contrast,
c
d
non-malignant SAEC cells did not express FASN. Selective index (SI) = IC50 SAEC /IC50 NSCLC. - : Not determined.
We noted with interest that some analogs (EC1, EC6, EC8 and EC9) were strikingly less potent
on H1650 than on H1703 and H1299. The anomalous responses to H1650 may be linked to the
absence of the tumor suppressor PTEN in this cell line [36]. PTEN moderates AKT signaling
which promotes cell growth and survival [36–38]. Loss of PTEN would thus desensitize cells to
cell death. Intriguingly, EC5 maintained the same level of growth inhibition on H1650 as the two
other cell lines, suggesting that it may have circumvented the loss of PTEN in H1650 by
activating other cell death pathways.
EC1 and EC5 were additionally tested on FASN-expressing NSCLC PC-9 and A549 cells (Table
1
). PC-9 cells harbor the clinically relevant EGCR delE746-A750 mutation which renders them
sensitive to tyrosine kinase inhibitors (TKI) such as gefitinib [39,40]. In contrast, H1703 and
1
1

A549 possess wild-type EGFR and were less sensitive to gefitinib [39,40]. Curiously, these
differences in gefinitib-sensitivity did not appear to affect the susceptibility of these cells to EC1
and EC5 as seen from the narrow 2-4 fold variation in their growth inhibitory concentrations on
these cell lines (H1703, A549, PC-9). These findings were significant in view of the close
functional relationship between mutated EGFR and FASN in NSCLC [39]. Briefly, mutated
EGFR relies on FASN-mediated palmitoylation to mediate TKI resistance in NSCLC. Ali et al.
showed that pharmacological inhibition of FASN by orlistat impeded the palmitoylation of EGFR,
induced EGFR ubiquitination, abrogated EGFR mutant signaling and suppressed the
proliferation of TKI-resistant EGFR-mutant NSCLC cells in culture and in vivo [39]. Viewed in
this context, the equipotent activities of EC1 and EC5 on NSCLC cells harboring wild type or
mutated EGFR intimated their potential as pan-NSCLC growth inhibitors whose cell killing
properties were ostensibly independent of the cellular EGFR status.
In summary, EC1 and EC5 were identified as the most promising analogs in this series,
inhibiting the growth of NSCLC cells (with the exception of EC1 on H1650) at potencies that
were confined to a narrow IC50 window (1-5 µM). To determine if their cytotoxicities were
selectively directed towards malignant cells, the compounds were evaluated on non-malignant
lung epithelial SAEC cells which do not express FASN (Supplementary Information, Figure S1)
Disappointingly, both EC1 and EC5 exhibited low IC50 values (< 10 µM) on the non-malignant
cells although they were significantly more cytotoxic (> 5-fold) on some (H1703) but not all of
the tested NSCLC cells (Table 1). Clearly, the susceptibility of the SAEC cells to EC1 and EC5,
in spite of the absence of FASN, hinted at the ability of these compounds to intercept other
vulnerable cellular targets besides FASN. We speculate that the polyphenolic nature of these
compounds may be contributory in this regard. Notably, ortho-disubstituted phenolic groups as
those found in EC1 and EC5 are capable of generating free radicals in the course of redox
cycling [28]. Besides, phenolic groups are widely reputed to be non-specific membrane
disruptors [30]. That said, we found the selective cytotoxicities of EGCG to be surprisingly
comparable or even lower than those of EC1 and EC5 (Table 1) even though it is reputed to be
relatively non-toxic [41]. Clearly, other correlates of toxicity would be required to confirm the
selective activities of these compounds.
3
.2. The galloyl stilbenes EC1 and EC5 inhibited NSCLC cell proliferation by inducing
apoptosis
Having shown that EC1 and EC5 potently suppressed proliferation of NSCLC cells, we asked if
growth inhibition led to apoptotic cell death. To investigate, NSCLC cells (PC-9, H1703) were
1
2

incubated with EC1 and EC5 (5 µM) for 72 h, double stained with the Annexin V-FITC conjugate
and propidium iodide (PI) and analyzed for apoptosis by fluorescence-activated cell sorting
(
FACS). EGCG and EC4 (the inactive methoxylated analog of EC1) were similarly tested at
higher concentrations of 100 µM in view of their weaker growth inhibitory effects on both cell
lines. Figure 4A shows the distribution of treated cells based on their staining responses to the
fluorescent probes. Cells in early apoptosis were selectively stained by the Annexin V conjugate
and appeared in the lower right quadrant of Figure 4A. Cells in late apoptosis were double
stained by both Annexin V and PI and localized in the upper right quadrant of the same figure.
Early and late apoptotic cells were collectively expressed as a percentage of total cells in Figure
4B. We found significant apoptosis in cells treated with EC1 and EC5, with most cells
concentrated in the late apoptotic stage after 72 h. As anticipated, no cell death was observed
among EC4-treated NSCLC cells. EGCG induced significant levels of late apoptosis in H1703
cells but was less apoptogenic in PC-9. These observations may reflect its greater potency on
H1703 (IC50 17 µM) as compared to PC-9 (IC50 77 µM).
A
H1703
PC-9
Vehicle
EC1, 5 µM
EC5, 5 µM
Vehicle
EC1, 5 µM
EC5, 5 µM
EC4, 100 µM
EGCG, 100 µM
EC4, 5 µM
EC4, 100 µM
EGCG, 100 µM
EC4, 5 µM
Annexin V
H1703
**
PC-9
B
*
100
100
***
***
50
*
50
*
*
*
0
0
Figure 4. (A) Distribution of FACS-sorted, compound-treated H1703 and PC-9 cells that were double-stained with
Annexin V and PI. Cells were sorted into 4 categories according to their staining characteristics: viable (lower left
quadrant), early apoptosis (lower right quadrant), late apoptosis (upper right quadrant) and necrotic (upper left
1
3

quadrant). (B) Percentage of compound-treated cells in early and late apoptosis. Data represents mean ± SD for n =
independent determinations. Significant difference from control cells (vehicle) was analysed by unpaired t test using
GraphPad Prism 5 and indicated by an asterisk (*p<0.05, **p<0.001, ***p<0.0001).
3
3
.3. The galloyloxy stilbenes EC1 and EC5 are FASN inhibitors.
As EC1 and EC5 were designed to retain the structural features found in FASN inhibitors
EGCG and 1), we proceeded to assess their FASN inhibitory properties. To obtain the FASN
(
protein, FASN plasmids were transfected in lysates derived from HeLa cells using a commercial
protein expression kit (1-Step Human In Vitro Protein Expression Kit, ThermoFisher Scientific).
The success of the transcription/translation process was confirmed by the presence of FASN in
immunoblots prepared from the lysates (Supplementary Information, Figure S3). To determine
the loss in FASN activity, lysates were incubated with the test compound for 2 h, after which
acetyl-CoA, malonyl-CoA and NADPH were added. The consumption of NADPH in the course
of reaction would result in a decrease in absorbance at 340 nm but this fall would be less
pronounced if the enzyme was inhibited. The compounds tested were EC1, EC5 and EC4.
Positive controls were EGCG, 1 and orlistat. All the compounds were tested at 10 µM, with the
exception of EGCG, which was tested at 150 µM as well.
As seen from Figure 5, the FASN inhibitory activity of orlistat exceeded that of the other
compounds while EC1, EC5 and 1 had comparable potencies. Biphenyl 1 was previously
identified as a potent FASN inhibitor in SK-Br3 breast cancer cells, with 90% inhibition at 4 µM
[
34]. In our hands, 1 exhibited only modest inhibition (20% at 10 µM), possibly due to
differences in the assay protocol. Even so, we were encouraged to note that the FASN inhibitory
potencies of EC1 and EC5 compared favorably to that of 1. These results implied that trans-
stilbene is a viable scaffold for FASN inhibition, equivalent to biphenyl and superior to the
chroman ring found in EGCG. However, we noted that inhibitory activity required the presence
of polyphenolic groups on the stilbene template. Thus, EC4 in which the phenolic OH groups
were methoxylated, was devoid of inhibitory activity. Tellingly, the polyphenolic EGCG had
negligible inhibitory activity at 10 µM and required a much higher concentration (150 µM) to
achieve the same level of inhibition as EC1 and EC5. We posited that the close proximity of the
two polyphenolic side chains in EGCG may have adversely affected FASN inhibitory activity
[
33]. If so, the extended configuration of trans-stilbene has served to advantageously space out
the side chains to avoid such an occurrence.
1
4

80
60
40
20
0
-20
Figure 5. Inhibition of FASN by test compounds EC1, EC4, EC5, EGCG, biphenyl
1 and orlistat. Inhibition was
determined from the decrease in NADPH absorbance in treated versus untreated HeLa cell lysates over 10 min.
Compounds were tested at 10 µM to be aligned to orlistat which induced 50-60% inhibition at this concentration. Data
represents mean ± SD of n = 3 independent determinations.
We noted with interest that EC1 and EC5 were more soluble than 1 when determined under
similar conditions (pH 7.4, 25 °C, 24 h). The kinetic solubilities of EC1 and EC5 were 6 µM and
2
.5 µM respectively, which exceeded that of 1 (0.4 µM) by 6-15 fold. We posited that the greater
solubilities of EC1 and EC5 may be a consequence of diminished intermolecular packing within
the extended framework of the trans-stilbene scaffold.
3
.4. Inhibition of FASN is causal to the growth inhibitory activities of EC1 and EC5
Having shown that EC1 and EC5 inhibited FASN, we asked if FASN inhibition was causal to the
growth inhibitory activities of these compounds. We reasoned that if this was true, reversing the
lipid deprived status of the FASN-inhibited cells should restore viability. Hence, we
supplemented EC1/EC5-treated PC-9 cells with a commercial lipid mixture (CD Lipid
concentrate, Gibco) which contained palmitate and other essential fatty acids for 72 h, after
which cell viability was assessed by the tetrazolium MTS assay. Figure 6A shows that cell
viability improved upon addition of lipids to EC1/EC5-treated PC-9 cells. Similarly, lipid
supplementation restored the viability of PC-9 cells treated with EGCG at 100 µM. The
experiment was repeated on H1703 cells treated with EC1 and EC5 at 1 µM. Again, viability of
cells was enhanced by lipid supplementation (Figure 6B).
1
5

A
PC-9
EGCG, 100 µM
EC1, 2 µM
EC5, 2 µM
1
00
***
1
00
100
80
60
40
20
0
**
**
8
0
8
6
4
2
0
0
0
0
0
60
40
20
0
Vehicle
0
2
Vehicle
0
2
Vehicle
0
2
Exogenous lipid concentration (% v/v)
Exogenous lipid concentration (% v/v)
Exogenous lipid concentration (% v/v)
B
H1703
EC1, 1 µM
EC5, 1 µM
1
00
100
80
60
40
20
0
**
**
8
0
0
0
0
0
6
4
2
Vehicle
0
2
Vehicle
0
2
Exogenous lipid concentration (% v/v)
Exogenous lipid concentration (% v/v)
Figure 6. Viability of (A) PC-9 and (B) H1703 cells treated with test compound only and test compound + lipid
concentrate (2% v/v). Compounds tested were EC1 EC5 and EGCG (on PC-9 only). Compounds were tested at
,
concentrations that fall within the range of their IC50 values. Cell viability was determined by the MTS assay, after 72
h incubation. Data represents mean ± SD for n = 3 independent determinations. Significant difference between the
different treatments were analysed by unpaired t test using GraphPad Prism 5 and indicated by an asterisk
(***p<0.0001, ** p<0.01).
3
.5. Inhibition of FASN by EC1 and EC5 intercepted the FASN-AKT feedback loop in
NSCLC cells
The PI3K/AKT signaling cascade has been implicated in the regulation of FASN expression in
several malignant cell lines [42–45]. Specifically, Wang et al. [43] demonstrated a mutually
reinforcing relationship between AKT activation and FASN expression in ovarian carcinoma
cells. Briefly, constitutive activation of AKT increases FASN expression and enhances the
biosynthesis of palmitate and other fatty acids. Incorporation of these fatty acids into membrane
phospholipids ensures an optimal lipid balance within the membrane which is required for the
functional integrity of AKT and other membrane embedded proteins. Pharmacological inhibition
of FASN disrupts the positive feedback loop between AKT activation and FASN expression with
two consequential outcomes. First, FASN inhibition would sensitize cells to apoptotic cell death
by inactivating AKT and nullifying its role in regulating cell survival [46]. Second, inhibition would
reduce the supply of palmitate and other essential fatty acids which would in turn alter
1
6

membrane composition, adversely affect the functional activity of AKT and consequently lead to
suppression of FASN.
We have already shown that EC1 and EC5 induced apoptosis in NSCLC cells at their growth
inhibitory concentrations (Figure 4) but have yet to demonstrate the ensuing losses in AKT and
FASN that would be incurred as a consequence of inhibition. Hence, levels of these proteins
were probed in PC-9 and H1703 cells exposed to EC1 and EC5 at 1 µM and 5 µM for 48 h. As
seen in Figure 7, dose dependent suppression of FASN was observed on both cell lines. In
addition, there were concurrent losses in total AKT and phospho-AKT levels. Tellingly, no
change was observed in the levels of these proteins in cells treated with the inactive analog
EC4 which did not inhibit FASN or induce apoptotic cell death.
PC-9
H1703
EC1
EC5
EC4
50 100
EC1
EC5
EC4
VEH
1
5
VEH
1
5
VEH
5
VEH
1
5
VEH
1
5
VEH 5 50 100 µM kDa
FASN
270
FASN/GAPDH
1
0.86 0.65
0.58 0.12
0.53 0.29
1
1
0.92 0.49
1.04 0.18
0.88 0.20
1
1
1
0.99 1.01 1.04
0.99 1.01 0.98
1.02 0.95 0.88
1
1
0.39 0.08
0.52 0.11
1
0.11 0.04
0.34 0.01
0.09 0.001
1
1
1.05 1.12 1.05
1.09 0.99 1.00
0.97 0.82 0.73
AKT
60
AKT/GAPDH
1
1
phospho-AKT
60
pAKT/GAPDH
1
1
1
0.57 0.04
1
1
GAPDH
37
Figure 7. Representative western blots (n = 3) showing concentration-dependent losses in the levels of FASN, AKT
and phospho-AKT S473 in PC-9 and H1703 cells treated with test compound (EC1 EC5 EC4) or media only (VEH)
,
,
for 48 h. EC1 and EC5 were tested at 1 µM and 5 µM. EC4 was tested at 5, 50 and 100 µM. Densitometric analysis
was performed on ImageJ and normalised against the untreated control (VEH).
Having shown that lipid supplementation increased the viability of EC1/EC5 treated PC-9 and
H1703 cells (Figure 6), we reasoned that if EC1 and EC5 disrupted the FASN-AKT axis,
compound-induced suppression of AKT and FASN should be restored to pre-inhibition levels on
addition of lipids. Hence, we monitored the AKT/FASN levels in PC-9 and H1703 cells treated
with EC1 and EC5 in the presence of the lipid concentrate. The results in Figures 8A and 8B
reveal that the presence of lipids negated the suppressive effects of the test compound on
FASN/AKT/phospho-AKT and restored levels to that found in control cells. Thus the strong
rebound in protein levels dovetailed with the positive effects of lipid supplementation on cell
viability and recapitulated the disruptive effects of EC1 and EC5 on the FASN-AKT axis as a
consequence of their FASN inhibitory activity.
1
7

PC-9
H1703
A
EC1, 2 µM
EC5, 2 µM
EGCG, 100 µM
EC1, 1 µM
EC5, 1 µM
FASN
FASN/GAPDH
1
0.06 0.62
1
0.02 0.55
1
0.06 0.89
1
0.38 1.64
1 0.24 0.74
GAPDH
+
-
-
+
-
-
+
-
-
+
-
-
+
-
-
VEH
-
+
+
+
-
+
+
-
+
-
+
-
+
-
+
-
+
-
+
EC1/EC5/EGCG
+
-
+
-
+
+
+
+
+
+
+
1% exogenous lipid concentrate
PC-9
H1703
B
EC1, 2 µM
EC5, 2 µM EGCG, 100 µM
EC1, 1 µM
EC5, 1 µM
phospho-AKT
pAKT/AKT
1
0.12 0.87
1
0.13 0.57
1
0.14 0.67
1
0.16 0.61
1 0.04 0.62
AKT
+
-
-
-
+
-
-
+
-
-
-
+
-
-
+
-
-
VEH
+
+
-
+
+
+
+
-
+
+
-
+
+
EC1/EC5/EGCG
+
-
+
+
-
+
+
-
+
+
-
+
+
-
+
1% exogenous lipid concentrate
Figure 8. Representative western blots showing the recovery of (A) FASN; (B) AKT and phospho-AKT S473 levels in
PC-9 and H1703 cells treated with test compound (EC1 EC5) + Lipid concentrate (1% v/v) or media only (VEH) for
8 h. EGCG (100 µM) was tested on PC-9 cells. Densitometric analysis was carried out with ImageJ and normalised
,
4
against the untreated control (vehicle).
Conclusion
Two trans-stilbene esters bearing polyphenolic galloyloxy side chains (EC1, EC5) were
identified for the first time as inhibitors of FASN. The FASN inhibitory activities of both
compounds were comparable to that of EGCG at their growth inhibitory concentrations,
implicating trans-stilbene as a viable FASN inhibitory scaffold. With its extended configuration
and limited flexibility, there are more options available for positioning the essential polyphenolic
side chains on trans-stilbene while still complying with the putative optimal distance between the
said groups. Consequently trans-stilbene has the potential of generating more “actives” than
other scaffolds that do not possess its stereochemical characteristics. Future work should focus
on exploring other substituted trans-stilbenes, such as H-bonding amino as a surrogate of the
phenolic hydroxyl and ring structures like indole/benzimidazole which are isosteric to catechols
[
47].
1
8

Mechanistically, we have shown that the FASN inhibitory properties of EC1 and EC5 disrupted
the mutually reinforcing signaling pathway between FASN expression and AKT activation.
Corroboratory evidence of this disruption is seen from the significant levels of apoptosis and
diminished FASN/phospho-AKT protein levels in EC1/EC5 treated cells. It is further affirmed by
the “rescue” afforded by lipid supplementation which reversed the loss in cell viability and
reinstated FASN/phospho-AKT expression to pre-inhibition levels.
Notwithstanding, FASN inhibition may not be the sole means by which EC1 and EC5 induced
cell death. The structure-activity requirement for polyphenolic groups taken together with their
multi-targeting potential hint at the involvement of other biological targets besides FASN. The
susceptibility of the non-FASN expressing SEAC cells to EC1 and EC5, as well as the
contrasting growth inhibitory responses of EC1 and EC5 in PTEN-deficient H1650 cell line are
cases in point. From the latter, we deduced that EC1 was more dependent on the FASN-AKT
signaling axis as the loss in PTEN in H1650 led to significant losses in its growth inhibitory
activity. In contrast, EC5 was as potent on H1650 as on other PTEN-bearing cell lines, possibly
due to its ability to engage other cell death pathways. This raises the interesting possibility that
EC5 may be better positioned than EC1 to act against resistant NSCLC cells in view of its
pleiotropy. Future work should be directed towards identifying these additional pathways
intercepted by EC5 with the aim of understanding their functional correlations with FASN and
their feasibility as co-therapeutic targets together with FASN for the treatment of NSCLC.
Materials and Methods
Chemistry
Commercially available reagents were purchased from Sigma Aldrich, Alfa Aesar and TCI
Chemicals and used without further purification. Solvents (hexane, ethyl acetate,
dichloromethane, methanol) were distilled before use. Moisture-sensitive reactions were
conducted under N2 environment using commercially obtained anhydrous solvents and
activated molecular sieves. Thin layer chromatography (TLC) was performed using Merck silica
gel 60 F254 pre-coated aluminum plates. Flash column chromatography was carried out with
1
13
silica gel (Merck, 230 – 400 mesh). H and C NMR spectra were recorded on Bruker Avance
5
00 (AV500) and Bruker Avance 400 (AV400) instruments at 298K. Chemical shifts were
1
reported in parts per million (ppm) on the δ scale using residual protio-solvent signals ( H NMR,
13
CDCl δ 7.26, DMSO-d δ 2.50, MeOH-d δ 3.31; C NMR, CDCl δ 77.0, DMSO-d δ 39.5,
3
6
4
3
6
MeOH-d δ 49.0, Acetone-d δ 29.8) as internal references. Mass spectra were determined by
4
6
1
9

high resolution mass spectrometry (HRMS) electrospray ionization (ESI) or atmospheric
pressure chemical ionisation (APCI). Final compounds were purified to ≥ 95% as determined by
reverse-phase HPLC using Zorbax Eclipse XDB-C18 column (5
solvent systems (Supplementary Information, S5).
µm, 150 mm × 4.6 mm) on two
3
-Vinylphenyl acetate (4a)
The Wittig synthesis of intermediate 3 was carried out following a reported method [48]. Briefly,
methyltriphenylphosphonium bromide (14.6 g, 2.5 eq.) was dissolved in dry THF (140 mL) and
cooled to 0 °C under N . To the stirred mixture was added 2M n-butyllithium (20.4 mL, 2.5 eq.)
2
dropwise, stirring was continued for 1 h at 0 °C after which the reaction was cooled to -78 °C.
A solution of 3-hydroxybenzaldehyde (2 g, 1 eq.) in dry THF (2.4 mL) was then added dropwise
to the stirred mixture. Thereafter the reaction mixture was sequentially warmed to room
temperature, stirred for 4 h, cooled to 0 °C and quenched with saturated NH Cl. The aqueous
4
layer was extracted with diethyl ether, dried with anhydrous Na SO , filtered and evaporated
2
4
under vacuum to give crude 3-vinylphenol 3. The latter was purified by column chromatography
using hexane: ethyl acetate (8:1) as eluting solvent. The subsequent acetylation of 3 was
carried out following a reported method [49]. 3 was dissolved in DCM (30 mL), 4-
dimethylaminopyridine (DMAP) (0.115 g, 0.05 eq.) and triethylamine (TEA) (2.11 mL, 0.8 eq.)
were added, followed by dropwise addition of acetic anhydride (2.14 mL, 1.2 eq.) dissolved in
o
DCM (5 mL). The reaction mixture was then stirred at room temperature (25 C) for 1.5 h and
diluted with water. The aqueous layer was extracted several times with DCM, the combined
organic layer was washed with brine, dried with anhydrous Na SO , filtered after which the
2
4
organic solvent was removed under reduced pressure. The product was then purified by column
chromatography with hexane: ethyl acetate (20:1) as eluting solvent. The acetylated product 4a
1
was obtained as an oil in 93% yield (2.47 g). H NMR (400 MHz, CDCl ) δ 7.21 (t, J = 7.8 Hz,
3
1
1
H), 7.15 (dt, J = 7.7, 1.1 Hz, 1H), 7.05 (s, 1H), 6.89 (dd, J = 7.9, 1.0 Hz, 1H), 6.64 – 6.52 (m,
13
H), 5.64 (d, J = 17.6 Hz, 1H), 5.17 (d, J = 10.9 Hz, 1H), 2.15 (s, 3H). C NMR (101 MHz,
) δ 169.1, 150.8, 139.0, 135.8, 129.2, 123.6, 120.7, 118.9, 114.6, 20.8. HRMS (APCI,
M+H): m/z calculated for [C H O ]: 163.0754; found 163.0750.
CDCl
3
10
11
2
4
-Iodophenyl acetate (5a)
The acetylation of 5a was carried out as described for 3. DMAP (0.056 g, 0.05 eq.) and TEA
1.01 mL, 0.8 eq.) were added to a stirred solution of 4-iodophenol (2 g, 1 eq.) in DCM (20 mL).
This was followed by dropwise addition of acetic anhydride (1.03 mL, 1.2 eq.) in DCM (5 mL),
(
2
0

followed by stirring at room temperature for 1.5 h and then dilution with water. The same workup
method was followed to give the crude product 5a which was purified by column
chromatography with hexane: ethyl acetate (20:1). 5a was used for the next step of the
synthesis without further characterization.
General procedure for the synthesis of acetylated stilbenes (6a-c) [50]
The respective styrenes (4a or 4b) (1 eq.) and iodoarenes (5a or 5b) (1 eq.) were dissolved in
dry DMF. The solution was purged with N
triphenylphosphine (0.045 eq.) were then added and the reaction was heated with stirring to 80
C for 18 h. Thereafter, it was quenched with 1M HCl at 0 °C, extracted with diethyl ether and
2
for 10 min. Pd(OAc)
2
(0.01 eq.) and
°
the combined organic layer worked up in the usual way. The crude product was purified by
column chromatography.
General procedure for the synthesis of hydroxylated stilbenes (7a-c)
To a stirred solution of the acetylated stilbene (6a-c) in THF was added 1M LiOH. The reaction
was stirred at 95°C for 4 h after which 1M HCl was added to quench the reaction. The aqueous
layer was extracted with ethyl acetate and the combined organic layer worked up in the usual
way to give the crude product which was purified by column chromatography.
(
E)-4,4'-(Ethene-1,2-diyl)diphenol (8)
8
was synthesized following a reported method with modifications [51]. 4-Hydroxybenzaldehyde
(
0.15 g, 1 eq.) and 4-hydroxyphenylacetic acid (0.243 g, 1.3 eq.) were stirred in the presence of
piperidine (0.90 mL, 7.5 eq.) and 1-methylimidazole (0.73 mL, 7.5 eq.) in polyethylene glycol
600 (0.75 g) as solvent, under microwave conditions at 160 °C for 25 min. On cooling, the
4
reaction was acidified with 1M HCl and extracted with ethyl acetate. The combined organic layer
was worked up in the usual way to give crude 8 which was purified by column chromatography
1
using hexane: ethyl acetate (8:1). 8 was obtained as a solid in 3% (0.009 g) yield. H NMR (400
1
3
MHz, DMSO) δ 9.50 (s, 2H), 7.36 (d, J = 8.5 Hz, 4H), 6.91 (s, 2H), 6.78 (d, J = 8.4 Hz, 4H).
NMR (101 MHz, DMSO) δ 156.8, 128.7, 127.4, 125.3, 115.6. HRMS (ESI, M-H): m/z calculated
]: 211.0765; found 211.0769.
C
11 2
for [C14H O
1
-(4-Methoxyphenyl)naphthalene (9)
9
was synthesized following a reported method with modifications [52]. 1-Naphthaleneboronic
acid (0.2 g, 1 eq.) and 4-bromoanisole (0.16 mL, 1.1 eq.) were stirred in the presence of
Pd(OAc) (0.011 g, 0.042 eq.) and Ba(OH) (0.55 g, 1.5 eq.) in 95% ethanol (14 mL) for 17 h at
2
2
2
1

room temperature. The reaction mixture was then quenched with 1M HCl, extracted with DCM
and worked up in the usual way. The crude product was purified by column chromatography
1
with hexane: DCM (100:1) to give 9 as a solid (0.20 g, 74%). H NMR (400 MHz, CDCl3) δ 7.83
(
d, J = 8.3 Hz, 1H), 7.79 (d, J = 7.9 Hz, 1H), 7.73 (d, J = 8.2 Hz, 1H), 7.39 (ddd, J = 9.7, 8.1, 4.2
13
Hz, 2H), 7.34 – 7.28 (m, 4H), 6.95 – 6.89 (m, 2H), 3.78 (s, 3H). C NMR (101 MHz, CDCl3) δ
1
5
58.9, 139.9, 133.8, 133.1, 131.8, 131.1, 128.2, 127.3, 126.9, 126.0, 125.9, 125.7, 125.4, 113.7,
5.3. HRMS (APCI, M+H): m/z calculated for [C17H15O]: 235.1117; found 235.1119.
4
-(Naphthalen-1-yl)phenol (10)
Boron tribromide (1M in DCM) (2.56 mL, 3 eq.) was added dropwise to a stirred solution of 9
o
(
0.2 g, 1 eq.) in DCM (8 mL) at 0 C. The reaction was then warmed to room temperature,
stirred for 12 h, quenched with water and extracted with DCM. The combined organic layer was
worked up in the usual way to give the crude product which was purified by column
chromatography using hexane: ethyl acetate (10:1). 10 was obtained as a solid in 96% (0.18g)
1
yield. H NMR (500 MHz, CDCl ) δ 7.79 (d, J = 8.4 Hz, 1H), 7.74 (d, J = 8.1 Hz, 1H), 7.68 (d, J
3
1
3
=
8.2 Hz, 1H), 7.33 (dd, J = 15.0, 7.1 Hz, 2H), 7.29 – 7.19 (m, 4H), 6.81 (d, J = 8.3 Hz, 2H).
C
NMR (126 MHz, CDCl3) δ 154.7, 139.7, 133.8, 133.3, 131.7, 131.3, 128.2, 127.3, 126.9, 125.9,
1
2
25.9, 125.7, 125.4, 115.2. HRMS (ESI, M-H): m/z calculated for [C H O]: 219.0815; found
16 11
19.0811.
General procedure for synthesis of silyloxy benzoic acids (12a-c)
Reported methods were followed with some modifications [53,54]. The respective commercially
available benzoic acid (11a-c) (1 eq.) was dissolved in DMF, tert-butyldimethylsilyl chloride
(
TBSCl) (1.4 eq. per TBS group added) and imidazole (2.5 eq. per TBS group added) were
added and the reaction mixture stirred for 24 h at room temperature. The reaction was then
diluted with diethyl ether, washed with water, and the organic layer worked up in the usual way.
The residue was dissolved in THF, a solution of acetic acid: water (3:1) was added and stirred
(
24 h, room temperature) to selectively effect hydrolysis of the silyloxy ester. The reaction
mixture was diluted with water, extracted with ethyl acetate and the combined organic layer
worked up in the usual way to give the desired silyloxybenzoic acid (12a-c).
General procedure for synthesis of disubstituted silyloxy esters 16a-g, EC4, EC7
Procedure A was used for the synthesis of 16a, 16c, EC4, 16d and EC7. The respective diol (1
eq.)
and
substituted
benzoic
acid
(2.4
eq.)
were
dissolved
in
DCM.
2
2

N,N'-Dicyclohexylcarbodiimide (DCC) (3.4 eq.) and DMAP (0.2 eq.) were added and the
reaction mixture stirred for 12 h at room temperature. The solvent was removed under reduced
pressure and the compound purified by column chromatography.
Procedure B was used for the synthesis of 16b and 16e-g. The respective diol (1 eq.) and
substituted benzoic acid (4 eq.) were dissolved in DCM. N-(3-Dimethylaminopropyl)-N’-
ethylcarbodiimide hydrochloride (EDCI) (4 eq.) and DMAP (4 eq.) were added, the mixture
stirred for 12 h at room temperature and DCM was subsequently added to stop the reaction.
The organic layer was washed with NaHCO and worked up in the usual way to give the crude
3
product which was purified by column chromatography.
General procedure for synthesis of monosubstituted silyloxy esters 16h-j, EC12
The respective acid (12a or 12d) (2 eq.), EDCI (2 eq.) and DMAP (2 eq.) were dissolved in DCM
and cooled to 0 °C. The respective alcohol (1 eq) dissolved in DCM was added dropwise and
stirring was continued for 12h at room temperature. The reaction mixture was diluted with DCM
and worked up as described in the preceding paragraph (Procedure B).
General procedure for the desilylation of 16a-j to hydroxylated esters EC1-3, EC5-6, EC8-11,
EC13
Procedure A was used for the synthesis of EC11 and EC13. The silylated ester (16i and 16j) (1
eq.) was dissolved in THF. 1M tetrabutylammonium fluoride (TBAF) (4 eq.) was added to the
reaction mixture, stirred for 4 h at room temperature, and diluted with ethyl acetate. The organic
layer was washed with water and worked up in the usual way. The crude product was purified
by column chromatography.
Procedure B was used for the synthesis of EC6, EC9 and EC10.The silylated ester (16e, 16g
and 16h) (1 eq.) was dissolved in THF. 1M TBAF (5 eq.) and NH F (5 eq.) were added to the
4
reaction mixture, stirred for 4 h at room temperature and diluted with ethyl acetate. The organic
layer was washed with water and worked up in the usual way. The crude product was purified
by column chromatography.
Procedure C was used for the synthesis of EC1-3, EC5 and EC8. The silylated ester (16a-d and
1
6f) (1 eq.) was dissolved in THF/methanol (ratios stated in Supplementary Information, S7),
NH F (1.5 eq. per TBS group) was added to the reaction mixture, stirred for 4 h at room
4
temperature (unless otherwise stated in Supplementary Information, S7) and diluted with ethyl
2
3

acetate. The organic layer was washed with water and worked up in the usual way. The crude
product was purified by column chromatography.
General Biology
Human NSCLC cells (H1299, H1650, H1703, PC-9, A549) and SAEC cell lines were purchased
from the American Type Culture Collection (ATCC). H1299, H1650, H1703 and PC-9 cells were
maintained in RPMI 1640 (HyClone), supplemented with 10% fetal bovine serum (FBS)
(
HyClone) and 1% penicillin/streptomycin, while A549 cells were maintained in Dulbecco’s
Modified Eagle Medium (DMEM) with high glucose (HyClone) and supplemented with 10% FBS,
% penicillin/streptomycin and 1 mM sodium pyruvate. SAEC cells were maintained in Bronchial
1
Epithelial Cell Growth Medium (BEGM), supplemented with growth factors and supplements
from BEGM Bronchial Epithelial SingleQuots Kit (Lonza). All cell lines were grown at 37°C in a
humidified incubator with 5% CO . CellTiter 96® AQueous One Solution Cell Proliferation Assay
2
(
MTS) reagent (#G3582) was purchased from Promega (Madison, WI, USA). FITC Annexin V
Apoptosis Detection kit (#556547) was obtained from BD Biosciences (Franklin Lakes, NJ).
Human FASN plasmids (#OHu24141D) were prepared by GenScript (Piscataway, NJ). 1-Step
Human Coupled IVT Kit - DNA (#88882) was obtained from Thermo Scientific. RIPA buffer
(
#89900) was obtained from Pierce and cOmplete Mini EDTA-free Protease Inhibitor Cocktail
TM
Tablets (#4693159001) were obtained from Merck. Pierce
BCA Protein Assay Kit was
obtained from ThermoFischer Scientific. Phospho-AKT Ser473 (#9271, 1:1,000) and total AKT
#9272, 1:1,000) antibodies were obtained from Cell Signaling Technology Inc. (Danvers, MA,
(
USA), while FASN (#20140, 1:2,000) and GAPDH (#25778, 1: 2, 000) antibodies were obtained
from Santa Cruz Biotechnologies Inc. (Santa Cruz, CA, USA). PVDF membranes (#88520) were
obtained from Thermo Scientific. WesternBright ECL detection kit (K-12045-D50) was obtained
from Advansta (San Jose, CA). Malonyl-CoA (#M1762), acetyl-CoA (#A2056) and NADPH were
purchased from Sigma Aldrich. Chemically defined (CD) Lipid concentrate (#11905031) was
obtained from Gibco.
Determination of cell viability
Cell viability was assessed using CellTitre 96® Aqueous One Solution (Promega, Madison, WI)
containing the tetrazolium salt 3-(4.5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-
3
sulfophenyl)-2H-tetrazolium (MTS). Cells were seeded at densities of 3 × 10 cells/100
µL
3
6
(
except H1650 and SEAC which were seeded at 3.5 x 10 cells / 100
µ
L and 4 x 10 cells/100
µ
L respectively) in 96-well format and incubated overnight for attachment to occur. Stock
solutions of test compounds were prepared in DMSO and serially diluted with media to give
2
4

solutions covering a range of concentrations. Aliquots (50 µL) were then added to the wells to
give the desired final concentrations. DMSO was kept at 0.1% v/v per well. The treated cells
were incubated for 72 h (37 °C. 5% CO ). The plates were then centrifuged at 500g for 5 min,
2
supernatant were removed from the wells and replaced with fresh aliquots of media (150
µL).
MTS reagent (20 L) was added to each well and the plates were then read at 490 nm on a
µ
microplate reader (Tecan) after 2 h incubation at room temperature. Cell viability at each
concentration was determined from the following equation:
Meanꢀabsorbanceꢁꢂꢃꢄꢂꢅꢆꢇ
%
ꢀCellꢀviability =ꢀ
ꢀxꢀ100%
Meanꢀabsorbanceꢁꢂꢆꢈꢉꢂꢊ
Where Mean absorbancecompound and Mean absorbancecontrol were the average readings (from 4
technical repeats) of wells containing treated cells at a known concentration of test compound
and cells not exposed to test compound respectively. Readings were corrected for control wells
containing media only. Growth inhibitory IC50 (concentration of test compound required to inhibit
cell growth by 50% relative to control untreated cells) was determined by plotting % viability
against logarithmic concentration of test compound (GraphPad Prism 5).The experiments were
repeated on 3 separate occasions to give mean IC . Statistical analysis was done using
50
GraphPad Prism 5 (GraphPad Software, San Diego, CA).
Determination of apoptosis
Apoptosis was monitored using the FITC Annexin V Apoptosis Detection kit (#556547) (BD
5
Biosciences, Franklin Lakes, NJ). H1703 and PC-9 cells were prepared at densities of 1.5 × 10
cells/2 mL medium in 6-well plates and allowed to attach overnight. The cells were incubated
with EC1 and EC5 at 5 µM, EC4 at 5 µM and 100 µM, EGCG at 100 uM or 0.1 % DMSO in
media (control) for 48 h (37 °C, 5% CO ). After this time, cells were harvested, rinsed with ice-
2
cold PBS and cell pellets were resuspended in a mixture containing 1× Annexin V binding
buffer, Annexin V and propidium iodide following manufacturer’s protocol. The Becton-Dickinson
LSR II FACS analyzer was used for the detection and FACSAria FACS sorter was used for
analyses (BD Biosciences, Franklin Lakes, NJ). Experiments were done in triplicates. Statistical
analysis was done using GraphPad Prism 5 (GraphPad Software, San Diego, CA).
Inhibition of FASN
The method was modified from a protocol described by Bays et al. [55] The 1-Step Human
Coupled IVT Kit - DNA (#88882, Thermo Scientific) was used to obtain FASN proteins. Briefly,
1
.2 µg of FASN plasmid (#OHu24141D, GenScript) was used for the in vitro translation of FASN
2
5

protein in accordance to manufacturer’s guidelines. The freshly translated protein was
immediately diluted with potassium phosphate buffer (100 mM, pH 6.5, containing 1 mM EDTA)
in the ratio 3 parts lysate to 7 parts buffer and an aliquot (10
35
µ
L) was added to test compound
o
(
µ
L) contained in a 96-well plate and incubated for 2 h at 37 C, 5% CO
2
. The final
concentration of test compound per well was 10 µM. Thereafter, an aliquot (45 µL) of the
NADPH activity buffer containing acetyl Co-A (55 µM) and NADPH (330 µM) prepared in the
above mentioned phosphate buffer was added. The FASN catalytic reaction was started by
adding an aliquot (10 µL) of malonyl-CoA (500 µM) in phosphate buffer to the compound-lysate
mixture in the well. Absorbances at 340 nm were recorded on a microplate reader (Tecan) at
the point of addition (0 min) and 10 min later. FASN inhibition was determined using the
expression:
Meanꢀabsorbanceꢁꢂꢃꢄꢂꢅꢆꢇꢀꢌꢈꢀꢍꢎꢀꢃꢏꢆ −ꢀMeanꢀabsorbanceꢁꢂꢃꢄꢂꢅꢆꢇꢀꢌꢈꢀꢎꢀꢃꢏꢆ
%
ꢀFASNꢀinhibition = ꢋ1 −ꢀ
ꢀꢐ xꢀ100%
Meanꢀabsorbanceꢁꢂꢆꢈꢉꢂꢊꢀꢌꢈꢀꢍꢎꢀꢃꢏꢆ −ꢀMeanꢀabsorbanceꢁꢂꢆꢈꢉꢂꢊꢀꢌꢈꢀꢎꢀꢃꢏꢆ
Experiments were done in triplicates.
Protein analysis using western immunoblotting
5
H1703 and PC-9 cells were seeded at densities of 8 × 10 cells/plate in 100 mm Petri dishes
and incubated overnight for attachment. They were then treated with test compound at their
indicated concentrations for 48h or 72 h, after which cells were harvested and lysed directly in
RIPA buffer (#89900, Pierce) with protease inhibitors (#4693159001, Merck). Proteins were
TM
quantified using Pierce BCA Protein Assay Kit (ThermoFischer Scientific). Approximately 20-
5
0 µg of proteins were loaded into wells of 8% SDS-PAGE gels and electrophoresed until the
bromophenol blue dye reaches the bottom of gel. Proteins were then transferred onto PVDF
membranes (#88520, Thermo Scientific). Membranes were blocked with 5% w/v milk in TBS-T
(for AKT and phospho-AKT) or PBS-T (for FASN and GAPDH), and then incubated with the
required primary and peroxidase conjugated secondary antibodies. Protein bands were
detected using the WesternBright ECL detection kit (K-12045-D50, Advansta). Densitometric
analysis was done using ImageJ.
Lipid rescue assay
3
5
PC-9 cells were seeded at densities of 3 × 10 cells/100 µL (for cell viability assay) or 1 × 10
cells/2 mL in RPMI 1640 medium and allowed to attach overnight. Cells were treated with test
compound as described earlier for the determination of cell viability except that after 6 h of
incubation (37 °C, 5% CO ), 1× chemically defined lipid concentrate (#11905031, Gibco) was
2
2
6

added to the treated cells to give a final concentration of 2 % v/v in the well. Incubation was
continued for another 72 h after which the cells were procressed for cell viability determinations
or Western analysis. All experiments were done in triplicates. Statistical analysis was done
using GraphPad Prism 5 (GraphPad Software, San Diego, CA).
Determination of solubility
Determination of aqueous solubility was carried out on MultiScreen Solubility filter plates
(
(
Millipore MSSLBPC10) from Millipore Corporation (MA, USA). The manufacturer’s protocol
PC2445EN00, Millipore Corporation) was followed.
Acknowledgements
The authors gratefully acknowledge funding support from Ministry of Education, Republic of
Singapore and National University of Singapore (ML Go R148000234 114; YL Lam
R143000681112, AA R713000216720), Ministry of Education for provision of the National
University of Singapore President Graduate Fellowship to YJ Tan and the NUS Drug
Development Unit for carrying out solubility determinations on the test compounds.
References
[
[
[
[
[
1]
2]
3]
4]
5]
S.J. Wakil, Fatty Acid Synthase, A Proficient Multifunctional Enzyme, Biochemistry. 28
1989) 4523–4530. doi:10.1021/bi00437a001.
(
T. Maier, S. Jenni, N. Ban, Architecture of Mammalian Fatty Acid Synthase at 4.5 Å
Resolution, Science. 311 (2006) 1258–1262. doi:10.1126/science.1123248.
T. Maier, M. Leibundgut, N. Ban, The Crystal Structure of a Mammalian Fatty Acid
Synthase, Science. 321 (2008) 1315–1322. doi:10.1126/science.1161269.
S.S. Chirala, S.J. Wakil, Structure and Function of Animal Fatty Acid Synthase, Lipids. 39
(
2004) 1045–1053. doi:10.1007/s11745-004-1329-9.
L. Weiss, G.E. Hoffmann, R. Schreiber, H. Andres, E. Fuchs, E. Körber, H.J. Kolb, Fatty-
Acid Biosynthesis in Man, a Pathway of Minor Importance. Purification, Optimal Assay
Conditions, and Organ Distribution of Fatty-Acid Synthase, Biol. Chem. Hoppe. Seyler.
3
67 (1986) 905–912. doi:10.1515/bchm3.1986.367.2.905.
[
6]
G. Medes, A. Thomas, S. Weinhouse, Metabolism of Neoplastic Tissue. IV. A Study of
2
7

Lipid Synthesis in Neoplastic Tissue Slices in Vitro, Cancer Res. 13 (1953) 27–29.
[
7]
K.E. Hopperton, R.E. Duncan, R.P. Bazinet, M.C. Archer, Fatty acid synthase plays a role
in cancer metabolism beyond providing fatty acids for phospholipid synthesis or
sustaining elevations in glycolytic activity, Exp. Cell Res. 320 (2014) 302–310.
doi:10.1016/j.yexcr.2013.10.016.
[
[
[
8]
9]
M.S. Shurbaji, J.H. Kalbfleisch, T.S. Thurmond, Immunohistochemical Detection of a
Fatty Acid Synthase (OA-519) as a Predictor of Progression of Prostate Cancer, Hum.
Pathol. 27 (1996) 917–921. doi:10.1016/S0046-8177(96)90218-X.
P.L. Alò, P. Visca, G. Trombetta, A. Mangoni, L. Lenti, S. Monaco, C. Botti, D.E. Serpieri,
U. Di Tondo, Fatty Acid Synthase (FAS) Predictive Strength in Poorly Differentiated Early
Breast Carcinomas, Tumori. 85 (1999) 35–40. doi:10.1177/030089169908500108.
10] V. Jensen, M. Ladekarl, P. Holm-Nielsen, F. Melsen, F.B. Soerensen, The Prognostic
Value of Oncogenic Antigen 519 (OA-519) Expression and Proliferative Activity Detected
By Antibody MIB-1 in Node-Negative Breast Cancer, J. Pathol. 176 (1995) 343–352.
doi:10.1002/path.1711760405.
[
[
11] J.A. Menendez, B.P. Oza, E. Atlas, V.A. Verma, I. Mehmi, R. Lupu, Inhibition of tumor-
associated fatty acid synthase activity antagonizes estradiol- and tamoxifen-induced
agonist transactivation of estrogen receptor (ER) in human endometrial adenocarcinoma
cells, Oncogene. 23 (2004) 4945–4958. doi:10.1038/sj.onc.1207476.
12] J.A. Menendez, L. Vellon, I. Mehmi, B.P. Oza, S. Ropero, R. Colomer, R. Lupu, Inhibition
of fatty acid synthase (FAS) suppresses HER2/neu (erbB-2) oncogene overexpression in
cancer cells, Proc. Natl. Acad. Sci. USA. 101 (2004) 10715–10720.
doi:10.1073/pnas.0403390101.
[
[
13] J.A. Menendez, R. Lupu, Fatty acid synthase regulates estrogen receptor-
α signaling in
breast cancer cells, Oncogenesis. 6 (2017) e299. doi:10.1038/oncsis.2017.4.
14] A. Vazquez-Martin, R. Colomer, J. Brunet, R. Lupu, J.A. Menendez, Overexpression of
fatty acid synthase gene activates HER1/HER2 tyrosine kinase receptors in human
breast epithelial cells, Cell Prolif. 41 (2008) 59–85. doi:10.1111/j.1365-
2
184.2007.00498.x.
[
15] E. De Schrijver, K. Brusselmans, W. Heyns, G. Verhoeven, J. V. Swinnen, RNA
2
8