SP1-independent inhibition of FOXM1 by modified thiazolidinediones

Article abstract of DOI:10.1016/j.ejmech.2020.112902

This research article describes an approach to modify the thiazolidinedione scaffold to produce test drugs capable of binding to, and inhibit, the in vitro transcriptional activity of the oncogenic protein FOXM1. This approach allowed us to obtain FOXM1 i

Full text of DOI:10.1016/j.ejmech.2020.112902

European Journal of Medicinal Chemistry xxx (xxxx) xxx
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
SP1-independent inhibition of FOXM1 by modified thiazolidinediones
a
a,
Seyed Amirhossein Tabatabaei Dakhili , David J. Perez b, Keshav Gopal a,
ꢀ
Moinul Haque c d, John R. Ussher a, Khosrow Kashfi e , Carlos A. Velazquez-Martínez
a Faculty of Pharmacy and Pharmaceutical Sciences, University of Alberta, Edmonton, AB, Canada
,
,
f
a, *
ꢀ
b
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Unidad Radiofarmacia-Ciclotron, Division de Investigacion, Facultad de Medicina, Universidad Nacional Autonoma de Mexico, Mexico City, Mexico
c Department of Laboratory Medicine and Pathology, University of Alberta, Edmonton, AB, Canada
d Department of Oncology, University of Alberta, Edmonton, AB, Canada
e Department of Molecular, Cellular, & Biomedical Sciences, City University of New York School of Medicine, New York, USA
f Department of Molecular, Cellular and Biomedical Sciences, Sophie Davis School of Biomedical Education, City University of New York School of Medicine,
New York, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
This research article describes an approach to modify the thiazolidinedione scaffold to produce test drugs
capable of binding to, and inhibit, the in vitro transcriptional activity of the oncogenic protein FOXM1.
This approach allowed us to obtain FOXM1 inhibitors that bind directly to the FOXM1-DNA binding
domain without targeting the expression levels of Sp1, an upstream transcription factor protein known to
activate the expression of FOXM1. Briefly, we modified the chemical structure of the thiazolidinedione
scaffold present in anti-diabetic medications such as pioglitazone, rosiglitazone and the former anti-
diabetic drug troglitazone, because these drugs have been reported to exert inhibition of FOXM1 but
hit other targets as well. After the chemical synthesis of 11 derivatives possessing a modified thiazoli-
dinedione moiety, we screened all test compounds using in vitro protocols to measure their ability to (a)
dissociate a FOXM1-DNA complex (EMSA assay); (b) decrease the expression of FOXM1 in triple
negative-breast cancer cells (WB assay); (c) downregulate the expression of FOXM1 downstream targets
(luciferase reporter assays and qPCR); and inhibit the formation of colonies of MDA-MB-231 cancer cells
(colony formation assay). We also identified a potential binding mode associated with these compounds
in which compound TFI-10, one of the most active molecules, exerts binding interactions with Arg289,
Trp308, and His287. Unlike the parent drug, troglitazone, compound TFI-10 does not target the in vitro
expression of Sp1, suggesting that it is possible to design FOXM1 inhibitors with a better selectivity
profile.
Received 16 June 2020
Received in revised form
16 September 2020
Accepted 28 September 2020
Available online xxx
Keywords:
FOXM1
Transcription factors
Troglitazone
Thiazolidinediones.
Anticancer
MDA-MB-231
Breast cancer
© 2020 Elsevier Masson SAS. All rights reserved.
1. Introduction
the cell cycle, and it becomes transcriptionally active at the G2/M
phase [4]. However, it has been observed that differentiated cells
The Forkhead box M1 protein is a member of a large family of
transcription factors that share a unique wing helix DNA binding
domain [1], and it is one of the proteins responsible for maintaining
normal cell replication by promoting cell cycle progression [2]. In
normal cells, FOXM1 is expressed during the S phase; it induces the
G1 phase; and it regulates normal cell cycle by expressing a number
of G2/M-related genes including PIK1, CCNB2, and Nek2 [3]. From
an activation standpoint the FOXM1 protein undergoes multiple
phosphorylation reactions exerted by different kinases throughout
have marginal expression levels of FOXM1 [5].
In contrast to its role in normal cell proliferation FOXM1 is
overexpressed in cancer cells and it is an essential driving force in
tumorigenesis. FOXM1 causes genomic instability and unregulated
cell division in a wide variety of human cell carcinomas including
lung, oropharyngeal, melanoma, leukemia, pancreatic, and breast
tissue [6e12]. The overexpression of this transcription factor is also
linked to poor prognosis and resistance to chemotherapeutic
agents, which makes FOXM1 both a useful biomarker and a
promising drug target [13,14].
We previously proposed a “lead hopping” hypothesis describing
how three structurally different FOXM1 inhibitors bind to, and
disrupt, the interaction between the FOXM1 protein and its DNA
* Corresponding author.
E-mail address: velazque@ualberta.ca (C.A. Velazquez-Martínez).
ꢀ
https://doi.org/10.1016/j.ejmech.2020.112902
0223-5234/© 2020 Elsevier Masson SAS. All rights reserved.
ꢀ
Please cite this article as: S.A. Tabatabaei Dakhili, D.J. Perez, K. Gopal et al., SP1-independent inhibition of FOXM1 by modified
thiazolidinediones, European Journal of Medicinal Chemistry, https://doi.org/10.1016/j.ejmech.2020.112902

ꢀ
S.A. Tabatabaei Dakhili, D.J. Perez, K. Gopal et al.
European Journal of Medicinal Chemistry xxx (xxxx) xxx
binding site [15]. The three drugs involved in our prior publication
are the forkhead domain inhibitor-6 (FDI-6), thiostrepton (TSP),
and troglitazone (TGZ) (Fig. 1). In addition, we described a drug-
binding site at the interface of the FOXM1 protein and its DNA
binding domain, through (at least) two binding interactions: (1) a
good potency compared to the parent drugs.
To validate this hypothesis, we designed a series of molecules
possessing (a) the thiazolidinedione moiety; (b) a simpler chemical
scaffold based on TGZ; (c) eliminate the free rotation around the
methylene group (eCH2e) attached to the thiazolidinedione by
using an alkene (-CH ¼ ) instead. This replacement would, in the-
p
-sulfur interaction between the aromatic imidazole ring present
in His287 and an electron-deficient sulfur atom present in the
FOXM1 inhibitors; and (2) a halogen bonding interaction between
the 4-fluorophenyl ring in the FDI-6 and Arg297 in the protein.
Based on these two previously reported binding interactions, we
herein describe the design and biological screening of 11 thiazoli-
dinedione forkhead domain inhibitors (TFI-1 to TFI-11), similar to
the former antidiabetic drug TGZ. Even though thiazolidinediones
in general, and TGZ in particular, exert their antidiabetic activity by
ory, decrease the likelihood of the test drugs to bind to the PPARg
protein [18]. Consequently, we used a computer-based molecular
modeling (docking) protocol to examine a wide variety of func-
tional groups attached to a much simpler benzyloxybenzene ring
(Scheme 1). The best scores in the modeling protocol were
observed when (1) the R1 position contained a nitro group (-NO2),
trifluoromethyl (-CF3), or methyloxycarbonyl (-CO2CH3) ester; (2)
the R2 position contained a methoxy (-OCH3) or a nitro group.
Additionally, to increase the robustness of the structure-activity
relationship study, we explored the effect produced by varying
the position of the thiazolidinedione ring relative to the methoxy
and nitro groups on the benzene ring. The overall drug design
strategy is explained schematically in Fig. 2.
targeting
proliferator-activated receptor gamma, [PPAR
ecules possess heterocyclic electron-deficient sulfur atom
a
different transcription factor (the peroxisome
g
]), these drug mol-
a
required to exert binding interactions with the His287 residue in
FOXM1, and consequently, this feature makes the thiazolidinedione
a useful and promising scaffold to transition from an established
PPARg-dependent antidiabetic effect, to a FOXM1-dependent
anticancer activity.
3. Results and discussion
3.1. Chemistry
2. Drug design
Based on the modeling results described above, we synthesized
eleven compounds shown in Scheme 1 (TFI-1 to TFI-11) using the
Knoevenagel condensation reaction. We adapted several methods
described in the literature to prepare the intermediate compounds
TFI-1(int) to TFI-11(int) by using microwave assisted synthesis
[19e22]. Briefly, a 4-substituted benzylbromide was reacted with
the corresponding hydroxybenzaldehyde (2a-2c; 1 eq.) in ethanol,
under microwave-assisted heating. After ~12 h, we obtained a se-
ries of product intermediates labeled TFI-1(int) to TFI-11(int), in
excellent yields (80%). Then, based on reported protocols [19] we
reacted the thiazolidinedione (3 eq.) with piperidine in ethanol,
heating under microwave irradiation at 90 ꢀC for 3 h, to obtain the
target products (TFI-1 to TFI-11). All final compounds were the
corresponding Z isomers, based on the chemical shift, a previously
reported methine proton chemical shift [23].
The anti-diabetic mechanism of action exerted by thiazolidine-
diones does not appear to be directly correlated to their inhibitory
effect on FOXM1. However, Petrovic et al. report an interesting
study in which they propose an indirect mechanism of action by
which thiazolidinediones significantly decrease the expression of
FOXM1 [16], and consists on the inhibition of SP1, one of the up-
stream proteins that activates the expression of FOXM1 [17]. This
report is highly relevant to our investigation because, it proposes
indirect FOXM1 inhibition. Nevertheless, we propose that the
thiazolidinedione ring also binds to the FOXM1-DNA binding
domain, and structural modification of this moiety might produce
new molecules with relative selectivity toward FOXM1. More
importantly, we hypothesize that it may be possible to bypass the
Sp1-dependent FOXM1 inhibition while maintaining a relatively
Fig. 1. Current FOXM1 inhibitors. Troglitazone (A) and thisotrepton (B) are known to inhibit FOXM1 indirectly while FDI-6 (C) is the only known selective inhibitor of FOXM1.
2

ꢀ
S.A. Tabatabaei Dakhili, D.J. Perez, K. Gopal et al.
European Journal of Medicinal Chemistry xxx (xxxx) xxx
Scheme 1. Synthesis and final chemical structures of eleven thiazolidinedione forkhead domain inhibitors TFI-1 to TFI-11.
Fig. 2. The drug design strategy for making selective FOXM1 inhibitors. Besides incorporating the
p-sulfur interaction of the thiazolidinedione group, we also included halogen
bonding between Arg297 and fluorine atom based on our previous finding. Furthermore, the presence of unsaturation has been shown to make the compound PPAR-Gamma non-
agonist.
3.2. Western blot assay
showed significant activity except for the TFI-7 where the thiazo-
lidinedione was placed at R4 position (too far from His287). Ac-
cording to our modelling study, if the thiazolidinedione is at R4
To evaluate the potency of the test compounds, first we
measured the concentration-dependent effect produced by drug
molecules on the protein expression of FOXM1 using a triple
negative-breast cancer (TNBC) cell line (MDA-MB-231 cells) after a
48-h incubation period. Fig. 3A shows the results obtained in the
Western blot screening assay; we compared the potency of the test
compounds (TFI-1 to TFI-11) to that observed with three reference
molecules: (a) the parent TGZ, (b) the reported forkhead domain
inhibitor FDI-6, and (c) the antibiotic TSP. We observed that com-
pounds TFI-2, TFI-6, and TFI-10 significantly reduced (by more than
90%) protein expression of FOXM1 in MDA-MB-231 cells, while the
other derivatives showed a modest activity compared to the
reference compounds (80e98% inhibition).
position, it can’t make the
p -sulfur interaction with the His287
anymore. Of surprise, compounds carrying nitro groups at the po-
sition R1 were also active except for the TFI-4 where an additional
nitro group was also introduced in the R2 position. Addition of
methoxy group significantly improved the activity of TFI-6 in
compare to the TFI-1 where the change was not so expressive when
the same functional group was added to the R2 position of TFI-2. All
derivatives carrying methyl ester at R1 were inactive irrespective of
substitutions at the other sites.
3.3. EMSA and luciferase assay
As we were expecting and described in the drug design strategy,
most of the compounds having -CF3 group at the R1 position
It is well known in the literature that some FOXM1 inhibitors
exert their mechanism of action by either indirect inhibition of
3

ꢀ
S.A. Tabatabaei Dakhili, D.J. Perez, K. Gopal et al.
European Journal of Medicinal Chemistry xxx (xxxx) xxx
Fig. 3. A) Western blot analysis of FOXM1 protein levels decreased by the test compounds (40
mM; 48 h, MDA-MB-231 cell line). B) electromobility shift assay (EMSA) showing the
Ki value determined for the compound TFI-10 and the positive control TGZ; TFI-10 was almost twice as potent as TGZ in preventing the formation of the FOXM1-DNA complex. C)
Luciferase reporter assay of the FOXM1 luciferase reporter (the corresponding plasmid was transfected into 293-T cells and the relative level of luciferase activity was measured after
incubation of MDA-MB-231 cells with 40
values were calculated by one way ANOVA (ns ¼ not significant ¼ P > 0.05; * ¼ P ꢁ 0.05; ** ¼ P ꢁ 0.01; *** ¼ P ꢁ 0.001; **** ¼ P ꢁ 0.0001.
m
M of TFI-10 and TGZ, for 24 h). The average values in all experiments were calculated after three independent experiments (n ¼ 3); P
upstream FOXM1 activator proteins [24,25] or by increasing the
concentration of a negative regulator of the FOXM1 protein, which
in turn, is degraded by the proteasome [26,27]. Consequently, to
study the mechanism of action exerted by the TGZ derivatives we
selected compound TFI-10 to carry out complementary screening
assays, and determine if this molecule exerts dissociation of the
FOXM1-DNA binding domain using the cell-free electrophoretic
mobility shift assay (EMSA; Fig. 3B). The incubation of recombinant
FOXM1-DBD and its consensus DNA with the corresponding drug
molecules showed that both TGZ and TFI-10 disrupt the complex
formed by FOXM1 and DNA. Quantitatively speaking, as shown in
Fig. 3B, TGZ and TFI-10 inhibited the protein-DNA complex with a
Ki ¼ 51.9 nM and 21.4 nM, respectively. This data suggests that
compound TFI-10 is about twice as potent than TGZ (Fig. 3B),
suggesting a better inhibitory profile on FOXM1’s transcriptional
activity.
To evaluate this hypothesis, we incubated MDA-MB-231 cells in
the presence of increasing concentrations of test compounds, to
determine an IC50 values, using a firefly luciferase reporter assay
possessing the corresponding FOXM1 DNA binding sites. In this
regard, the relative decrease in FOXM1 transcriptional activity in
the presence of the test molecules (Fig. 3C) reproduced the pattern
that we previously observed with immunoblotting (Fig. 3A). Both
compounds (TFI-10 and TGZ) showed significant activity at both at
20 and 40 mM.
To confirm the results obtained in both the Western blot and the
luciferase assays described above, in the sense that any decrease in
the expression of the FOXM1 protein would correlate with higher
4

ꢀ
S.A. Tabatabaei Dakhili, D.J. Perez, K. Gopal et al.
European Journal of Medicinal Chemistry xxx (xxxx) xxx
inhibition of triple negative-breast cancer cell proliferation, we
measured the mRNA levels of FOXM1 its downstream target pro-
teins CDC25B and CCNB1. The expression of CCNB1 promotes cell
cycle progression (from G2 to M phase) and the expression of
CDC25 plays an important role in the cell’s M phase [28]. Consistent
with our previous protein immunoblot and EMSA experiments, TFI-
10 decreased the mRNA levels of the FOXM1 target genes to a
higher extent than that produced by TGZ (Fig. 4A).
In a previous report by Petrovic et al. [16] in which it described a
thiazolidinedione-induced decrease in the expression of FOXM1,
authors reported that this effect is mediated by the inhibition of the
transcription factor SP1. Furthermore, Petrovic et al. proposed a
new mechanism of action for thiazolidinediones in which this drug
class increases the proteasome-dependent degradation of the
transcription factor Sp1. Since this protein has a binding domain in
the FOXM1 promoter, the activation of Sp1 was associated with an
increase in the expression of FOXM1 in cancer cells. Nevertheless,
as we showed in Fig. 4B, we observed a significant decrease in the
expression levels of SP1 exerted by the parent molecule TGZ, but
this effect was not observed when we incubated MDA-MB-231 cells
with compound TFI-10, suggesting that TFI-10 inhibits the
expression of FOXM1 by a different (SP1-independent) mechanism
of action.
Thiazolidinediones (TZDs) are very well studied ligands of the
peroxisome proliferator activated receptor gamma (PPARg), but
this activity, according to Petrovic et al., is not involved in the in-
duction of FOXM1 [16].
Finally, it has been reported that TZDs exert a significant anti-
proliferative effect in many epithelial-derived cancer cells by G1
phase cell cycle arrest through accumulation of the CDK inhibitor
proteins p21Cip and p21Kip [29,30], but the investigation of these
targets goes beyond the scope of this paper and remains to be
explored. Another limitation of our investigation is the observation
that other PPARg-independent targets of thiazolidinediones are the
c-Jun N-terminal protein kinase and p38 [31], the early growth
response-1 (EGR1) [32], and the tumor suppressor protein p53 [33].
At this point we do not know if these proteins are targeted by our
drug molecule TFI-10 and this will be the subject of upcoming in-
vestigations by our group.
3.4. Cell colony formation assay
The sensitization of human carcinoma cells with a FOXM1 in-
hibitor to conventional chemotherapy (i.e. doxorubicin), via acti-
vation of apoptosis, has been studied and reported previously by
Halasi et al. [34]. Consequently, we decided to explore if our
Fig. 4. A) qPCR analysis of mRNA levels of human FOXM1 and its downstream targets including CDC25B and CCNB1. The drug TFI-10 (20
of FOXM1 target genes at after 24 h, but this effect was not observed with TGZ at the same concentration. B) Western blot assay measuring the expression of the transcription factors
SP1 and FOXM1; TGZ an TFI-10 (80
M) significantly decreased the level of FOXM1 after 24 h, but the level of SP1 was only affected by TGZ. No significance (ns) ¼ P > 0.05;
(*) ¼ P < 0.05; (**) ¼ P < 0.01; (***) ¼ P ꢁ 0.001; **** ¼ P ꢁ 0.0001.
mM) significantly decreased the mRNA level
m
5

ꢀ
S.A. Tabatabaei Dakhili, D.J. Perez, K. Gopal et al.
European Journal of Medicinal Chemistry xxx (xxxx) xxx
thiazolidinedione derivative, TFI-10, is capable of sensitizing hu-
man carcinoma cells to DNA-damaging agents. Briefly, cells were
initially treated with (a) 5 mM of TFI-10; (b) 50 nM of doxorubicin;
by TFI-10 were better that those observed for TZD, with binding
free of energies of ꢂ103.3 Kj/mol and ꢂ82.2 Kj/mol TGZ, respec-
tively (Table 1).
or (c) TFI-10 and doxorubicin. As shown in Fig. 5, in contrast to cells
treated either with TFI-10 or doxorubicin alone, cells incubated in
the presence of both drugs did not form any (detectable) colonies.
The results of this experiment agree with those reported by Halasi
et al. the inhibition of FOXM1 synergized with the doxorubicin-
induced DNA damage in cancer cells. We could speculate that this
synergistic effect could be due, at least in part, to inhibition of
FOXM1-induced expression of anti-apoptotic factors, such as JNK
and Bcl-2 [34].
4. Discussion and conclusions
In summary, this research article describes a new approach to
modify the thiazolidinedione scaffold to produce derivatives that
retain
a desirable FOXM1 inhibitory profile, via an SP1-
independent mechanism, paving the way for the design of novel
drug molecules with a good chance of exerting anticancer effects on
triple negative-breast cancer cells. We submit evidence demon-
strating the direct inhibition of the FOXM1-DNA complex (in vitro),
and to some extent, transcriptional inhibition of the FOXM1 protein
that results in the inhibition of downstream target proteins, as well
as significant decrease in cancer cell colony formation (in vitro). In
this report, we provide preliminary evidence showing that it might
be possible to modulate (and bypass) the SP1-dependent anti-
cancer mechanism of action exerted by thiazolidinediones, pro-
ducing relatively potent FOXM1 inhibitors that bind to the target
3.5. Docking and molecular dynamic simulation studies
Finally, to provide a fine-tuned justification for the observed
biological profile shown by compound TFI-10, we conducted a
computer-based molecular modeling and dynamic simulation
study based on a previously reported protocol described by our
group in which we proposed a drug-binding pocket at the interface
of the FOXM1-DBD. As shown in Fig. 6A, we observed a
p
-sulfur
site via (a)
halogen bonding, reported by our group previously. An additional
binding interaction ( stacking) observed with the test drug TFI-
10 is new to this type of scaffold and it adds to the list of binding
interactions for novel FOXM1 inhibitors.
One additional and very important contribution of this research
work is the fact that, despite showing a better selectivity profile,
compound TFI-10, as well as a wide variety of “direct” FOXM1 in-
hibitors (including the drugs FDI-6 [35] and TSP [27,36], show weak
inhibitory profiles as seen with their in vitro IC50 values, regardless
of the biological assay employed, the best of them in the low
p-sulfur interaction, (b) hydrogen bonding, and (c)
binding interaction between the thiazolidinedione ring present in
TFI-10 and the His287 residue, which agrees with our previous
report [15]. The same binding interaction was also observed be-
tween His287 and thiazolidinedione ring of the TGZ (Fig. 6B). In this
regard, the introduction of a methoxy group to TFI-10 contributed
to its increased binding affinity by exerting two hydrogen bonding
interactions with Arg286. Furthermore, we also observed an addi-
tional binding interaction involving the trifluoromethyl moiety and
Trp308, Ser306 and Arg297, providing additional evidence in sup-
port of a potential drug binding site at that location. The role of a
halogen bonding is not only relevant and significant for the thia-
zolidinediones reported in this paper, but also for structurally
different FOXM1 inhibitors such as the drug FDI-6 reported by
Gormally et al. [35].
We also carried out another MD simulation (20 ns) to evaluate
the stability of the test compounds inside the proposed binding
pocket. It was not surprising for us to see how TFI-10 showed an
improved stability inside the binding pocket between FOXM1 and
DNA, as quantified by the ligand positional Root Mean Square De-
viation (RMSD) (Fig. 6C). Besides that, upon binding, the RMSD
values (Fig. 6D) were also decreased, suggesting a higher binding
affinity by TFI-10. The side chain Root Mean Square Fluctuation
(RMSF) was also utilized to analyze which residues are interacting
with the test compounds as shown in Fig. 6E. Finally, the MMPBSA
binding free energy was also calculated during the last 5 ns of the
drug/FOXM1-DBD simulations. The binding interactions produced
p-p
(20e40 mM) range. To our knowledge, there has not been any re-
ports on FOXM1 inhibitors with a significant and potentially useful
and “clinically” relevant potency with IC50 values in the low nM
range. Therefore, it is reasonable to question whether targeting the
FOXM1 protein, at its DNA binding domain, or via inhibition of the
proteasome, or inhibiting any of its reported upstream positive
modulators, are promising strategies. It is difficult to predict if
direct inhibition of the FOXM1 protein will yield drugs potent
enough to be considered promising in therapy. Perhaps one of the
areas in which direct inhibition of FOXM1 could find a clinical
application is combination therapy. As shown in this report, and
several others, inhibiting the transcriptional activity of FOXM1
together with chemotherapeutic drugs which depend on a different
mechanism of action, seems to be a better strategy that may pro-
duce better outcomes, in terms of inhibiting cancer cell
proliferation.
Fig. 5. Colony formation assay using triple negative-breast cancer cells in the presence of test drugs TFI-10, doxorubicin, and a combination of both. Photographs represent the
results obtained in one of the three experiments carried out (n ¼ 3).
6

ꢀ
S.A. Tabatabaei Dakhili, D.J. Perez, K. Gopal et al.
European Journal of Medicinal Chemistry xxx (xxxx) xxx
Fig. 6. A) Schematic representation of the drug TFI-10 bound to the FOXM1-DBD showing the expected binding interactions based on a previously reported model; B) Graphical
representation of TGZ in complex with the FOXM1-DBD. In addition to the -sulfur interaction with His287, we observed -alkyl interactions with Val305, Leu259 and Lys260; C)
p
p
the calculated ligand positional RMSD for TFI-10 and TGZ suggests higher stability of TFI-10 in the binding pocket; D) RMSD of FOXM1-DBD (no ligand), TFI-10, and TGZ suggesting a
strong binding exerted by TFI-10; E) the calculated RMSF values for the FOXM1-DBD (no ligand), TFI-10, and TGZ suggesting a more prominent binding profile of TFI-10, compared to
TGZ.
7

ꢀ
S.A. Tabatabaei Dakhili, D.J. Perez, K. Gopal et al.
European Journal of Medicinal Chemistry xxx (xxxx) xxx
Table 1
MMPBSA binding free energy calculation for TFI-10 and TGZ during the last 5 ns of the simulation. The binding free energy was calculated by substracting the polar solvation
energy from solvent accessible surface area, electrostatic, and VdW energies.
Compd.
vdW energy
Electrostatic energy
Polar solvation energy
SASA energy
Binding free energy
TFI-10
TGZ
ꢂ130.8 7.4
ꢂ127.6 29.6
ꢂ24.7 5.8
ꢂ10.3 5.0
65.6 5.9
70.0 20.6
ꢂ13.4 0.7
ꢂ14.2 3.4
ꢂ103.3 6.4
ꢂ82.2 17.8
5. Experimental
5.1.1.3. 4-methoxy-3-[(4-nitrophenyl)methoxy]benzaldehyde
[TFI-
6
(int)]. 3-Hydroxy-4-methoxybenzaldehyde (100 mg, 0.66 mmol),
5.1. Chemistry
4-nitrobenzyl bromide (143 mg, 0.66 mmol) and K2CO3 (90 mg,
0.66 mmol), white crystals, 80% yield. 1H NMR (600 MHz, DMSO)
All the reagents and solvents were purchased from Sigma-
Aldrich and were used without further purification. All reactions
were monitored by thin-layer chromatography (RediSep® TLC
plates) and visualized using UV light. Melting points were
measured with an Electrothermal melting point apparatus (Ther-
mofisher, USA) and were uncorrected. 1H, 13C and 19F NMR spectra
were determined on a Bruker FT-600 MHz instrument (600 MHz,
150 MHz and 565 MHz, respectively) using DMSO‑d6 or MeOD as
d
9.83 (s, 1H), 8.28 (d, J ¼ 8.7 Hz, 2H), 7.73 (d, J ¼ 8.8 Hz, 2H), 7.61
(dd, J ¼ 8.2, 1.9 Hz, 1H), 7.48 (d, J ¼ 1.9 Hz, 1H), 7.24 (d, J ¼ 8.3 Hz,
1H), 5.35 (s, 2H), 3.91 (s, 3H). 13C NMR (151 MHz, DMSO)
191.30,
d
154.51, 147.76, 147.08, 144.65, 129.58, 128.27, 126.89, 123.66, 111.73,
111.18, 68.69, 56.09.
5.1.1.4. 4-methoxy-3-{[4-(trifluoromethyl)phenyl]methoxy}benzalde-
hyde [TFI-7(int)]. 3-Hydroxy-4-methoxybenzaldehyde (100 mg,
0.66 mmol), 4-(trifluoromethyl)benzyl bromide (158 mg, 0.66),
K2CO3 (90 mg, 0.66 mmol), white crystals, 90%. 1H NMR (600 MHz,
solvents and TMS as a reference. Chemical shifts (d) and coupling
constants (J) are expressed in parts per million and Hertz, respec-
tively. Signal multiplicity is expressed as s (singlet), d (doublet), t
(triplet), q (quartet), m (multiplet) and br (broad singlet). The
microwave-assisted synthesis was carried out using an Initiator
Reactor (Biotage).
DMSO)
d
9.83 (s, 1H), 7.78 (d, J ¼ 8.0 Hz, 2H), 7.69 (d, J ¼ 8.0 Hz, 2H),
7.60 (dd, J ¼ 8.2, 1.9 Hz, 1H), 7.49 (d, J ¼ 1.9 Hz, 1H), 7.22 (d,
J ¼ 8.3 Hz, 1H), 5.29 (s, 2H), 3.90 (s, 3H). 13C NMR (151 MHz, DMSO)
d
191.31,154.51,147.90,141.61, 129.59,128.74,128.53,128.32, 127.90,
126.94, 126.75, 125.39, 125.36, 125.14, 123.34, 121.54, 111.66, 111.60,
5.1.1. General procedure for the synthesis of intermediates [TFI-1(int)
111.09, 68.96, 56.04. 19F NMR (565 MHz, DMSO)
d
ꢂ60.96.
to TFI-11(int)
]
We adapted methods described in the literature to prepare the
intermediate compounds TFI-1(int) to TFI-11(int) by using microwave
assisted synthesis [19e22]. The corresponding benzylbromide (1a-
1c; 1 eq.), the hydroxybenzaldehyde (2a-2c; 1 eq.), and K2CO3 (1
eq.) were mixed in EtOH (5 mL) inside a microwave reactor vessel
and heated to 90 ꢀC for 12 h with magnetic stirring. The reactions
were monitored by TLC and, upon completion, the final products
precipitated as white solid crystals (insoluble in EtOH), which were
then filtered off and washed with hot water (5 mL), followed by a
second washing with hexane (3 mL); after drying the product in-
termediates we calculated their purity (about 95%) by 1H and 13C
NMR. The overall yields for these reactions was about 80%. In-
termediates TFI-1(int), TFI-3(int) [19,20,22], and TFI-10(int) [21], were
confirmed by 1H and 13C NMR which were in accordance to the
reported analytical data [23].
5.1.1.5. Methyl
[TFI-8(int)]. 3-Hydroxy-4-methoxybenzaldehyde
4-[(5-formyl-2-methoxyphenoxy)methyl]benzoate
(100 mg,
0.66 mmol), methyl 4-(bromomethyl) benzoate (152 mg,
0.66 mmol), K2CO3 (90 mg, 0.66 mmol), white crystal, 80%. (TFI-
8int). 1H NMR (600 MHz, DMSO)
d
9.82 (s, 1H), 7.99 (d, J ¼ 8.4 Hz,
2H), 7.63e7.56 (m, 3H), 7.47 (d, J ¼ 1.9 Hz, 1H), 7.22 (d, J ¼ 8.3 Hz,
1H), 5.28 (s, 2H), 3.90 (s, 3H), 3.86 (s, 3H). 13C NMR (151 MHz,
DMSO)
129.32, 129.08, 127.59, 127.57, 126.71, 111.66, 111.16, 69.20, 60.76,
56.05, 52.17, 39.52. 13C NMR (151 MHz, DMSO)
191.31, 166.01,
d 191.31, 166.01, 154.52, 147.93, 142.26, 129.57, 129.36,
d
154.52, 147.93, 142.26, 129.57, 129.36, 129.32, 129.08, 127.59, 127.57,
126.71, 111.66, 111.16, 69.20, 60.76, 56.05, 52.17.
5.1.1.6. 3-Methoxy-4-[(4-nitrophenyl)methoxy]benzaldehyde
(int)]. 4-Hydroxy-3-methoxybenzaldehyde (100 mg, 0.65 mmol),
4-nitrobenzyl bromide (142 mg, 0.65 mmol), K2CO3 (90 mg,
0.65 mmol), white crystals, 80%. 1H NMR (600 MHz, DMSO)
9.85
[TFI-
5.1.1.1. 3-nitro-4-[(4-nitrophenyl)methoxy]benzaldehyde [TFI-4(int)].
4-Hydroxy-3-nitrobenzaldehyde (100 mg, 0.598 mmol), 4-
nitrobenzyl bromide (129 mg, 0.598 mmol) and K2CO3 (83 mg,
0.598 mmol), white crystals, yield 80%. 1H NMR (600 MHz, DMSO)
9
d
(s, 1H), 8.28 (d, J ¼ 8.8 Hz, 2H), 7.73 (d, J ¼ 8.8 Hz, 2H), 7.55 (dd,
d
9.96 (s, 1H), 8.49 (d, J ¼ 2.1 Hz, 1H), 8.31 (d, J ¼ 8.8 Hz, 2H), 8.22
(dd, J ¼ 8.7, 2.1 Hz, 1H), 7.74 (d, J ¼ 8.9 Hz, 2H), 7.66 (d, J ¼ 8.7 Hz,
1H), 5.62 (s, 2H). 13C NMR (151 MHz, DMSO)
190.49, 154.85,
J ¼ 8.2, 1.9 Hz, 1H), 7.45 (d, J ¼ 1.8 Hz, 1H), 7.25 (d, J ¼ 8.3 Hz, 1H),
5.40 (s, 2H), 3.87 (s, 3H). 13C NMR (151 MHz, DMSO)
d 191.43,
d
152.64, 149.44, 147.15, 144.25, 130.17, 128.39, 125.75, 123.70, 112.84,
109.94, 68.76, 55.69.
147.30, 143.14, 139.56, 135.05, 133.33, 130.58, 129.25, 128.15, 126.74,
124.37, 123.87, 123.81, 116.03, 69.90.
5.1.1.2. Methyl 4-[(4-formyl-2-nitrophenoxy)methyl]benzoate [TFI-
5(int)]. 4-Hydroxy-3-nitrobenzaldehyde (100 mg, 0.598 mmol),
methyl 4-(bromomethyl) benzoate (137 mg, 0.598 mmol), K2CO3
(83 mg, 0.598 mmol), white crystals, yield 80%. 1H NMR (600 MHz,
5.1.1.7. Methyl
[TFI-11(int)]. 4-Hydroxy-3-methoxybenzaldehyde
4-[(4-formyl-2-methoxyphenyl)methoxy]benzoate
(100 mg,
0.65 mmol), methyl 4-(bromomethyl) benzoate (150 mg,
0.65 mmol), K2CO3 (90 mg, 0.65 mmol). 1H NMR (600 MHz, DMSO)
DMSO)
1H), 8.01 (d, J ¼ 8.3 Hz, 2H), 7.63 (dd, J ¼ 20.2, 8.6 Hz, 3H), 5.54 (s,
2H), 3.86 (s, 3H). 13C NMR (151 MHz, DMSO)
190.47, 165.93,
d
9.95 (s, 1H), 8.47 (d, J ¼ 2.0 Hz, 1H), 8.20 (dd, J ¼ 8.7, 2.1 Hz,
d
9.84 (s, 1H), 7.99 (d, J ¼ 8.2 Hz, 2H), 7.60 (d, J ¼ 8.1 Hz, 2H), 7.54
(dd, J ¼ 8.2, 1.9 Hz, 1H), 7.43 (d, J ¼ 1.8 Hz, 1H), 7.24 (d, J ¼ 8.3 Hz,
1H), 5.32 (s, 2H), 3.85 (s, 6H). 13C NMR (151 MHz, DMSO)
191.43,
d
d
154.99, 140.83, 139.59, 134.97, 129.47, 129.40, 129.12, 127.37, 126.67,
116.03, 70.42, 52.21.
166.00, 152.89, 149.44, 141.89, 130.04, 129.41, 129.21, 127.73, 127.71,
125.81, 112.78, 109.87, 69.32, 55.66, 52.19.
8

ꢀ
S.A. Tabatabaei Dakhili, D.J. Perez, K. Gopal et al.
European Journal of Medicinal Chemistry xxx (xxxx) xxx
5.1.2. General procedure for the synthesis of compounds TFI-1 - TFI-
11
(br, 1H), 8.19 (d, J ¼ 2.2 Hz, 1H), 8.00 (d, J ¼ 8.4 Hz, 2H), 7.85 (dd,
J ¼ 9.1, 2.3 Hz, 1H), 7.81 (s, 1H), 7.59 (t, J ¼ 8.5 Hz, 3H), 5.49 (s, 2H),
We adapted the methods previously reported [19] to use a
microwave-assisted synthesis protocol. The corresponding inter-
mediate (TFI-1(int) to TFI-11(int), 1eq.), thiazolidinedione (3 eq.),
piperidine (cat.), AcOH (cat.), and EtOH (5 mL), were mixed in a
microwave reactor vessel and heated at 90 ꢀC for 3 h, under mag-
netic stirring. Once completed, the reactions were allowed to cool
down to RT to yield a pale yellowish precipitate that was filtered off
and washed with cold EtOH (10 mL) and water (10 mL). The target
products were obtained with an overall yield of about 60%. In this
regard, compound TFI-2 [19], was confirmed by 1H and 13C NMR
which was in accordance with the reported data. All derivatives
were determined to be the corresponding Z isomers according to a
previously reported methine proton chemical shift [23].
3.86 (s, 3H). 13C NMR (151 MHz, DMSO)
d 167.49, 167.22, 165.94,
151.74, 141.03, 139.64, 135.18, 129.46, 129.35, 127.33, 126.87, 126.04,
123.93, 116.39, 70.17, 52.21. [M ꢂ H]-: m/z calc. 413.0 found 413.2 m/
z (100%).
5.1.2.6. (5Z)-5-({4-methoxy-3-[(4-nitrophenyl)methoxy]phenyl}
methylidene)-1,3-thiazolidine-2,4-dione (TFI-6).
Thiazolidinedione (61 mg, 0.52 mmol), TFI-6(int) (50 mg,
0.17 mmol), piperidine (cat.) and AcOH (cat.), pale yellowish crys-
tals, 60% yield, m.p. 200e202 ꢀC. 1H NMR (600 MHz, DMSO)
d 12.50
(br, 1H), 8.27 (d, J ¼ 8.8 Hz, 1H), 7.73 (d, J ¼ 8.8 Hz, 1H), 7.69 (s, 1H),
7.24e7.19 (m, 1H), 7.18 (d, J ¼ 9.0 Hz, 1H), 5.33 (s, 1H), 3.87 (s, 2H).
13C NMR (151 MHz, DMSO)
d 167.95, 167.37, 151.15, 147.39, 147.08,
144.70, 131.89, 128.28, 125.60, 124.56, 123.67, 120.73, 115.10, 112.56,
5.1.2.1. (5Z)-5-({4-[(4-nitrophenyl)methoxy]phenyl}methylidene)-
1,3-thiazolidine-2,4-dione (TFI-1). Thiazolidinedione (67 mg,
0.57 mmol), TFI-1(int) (50 mg, 0.19 mmol), piperidine (cat.) and
AcOH (cat.), pale yellowish crystals, 60% yield, m.p. 251e253 ꢀC 1H
68.78, 55.88. [M ꢂ H]-: m/z calc. 385.1 found 385.3 m/z (100%).
5.1.2.7. (5Z)-5-[(4-methoxy-3-{[4-(trifluoromethyl)phenyl]methoxy}
phenyl)methylidene]-1,3-thiazolidine-2,4-dione
(TFI-7).
NMR (600 MHz, DMSO)
7.78e7.69 (m, 3H), 7.58 (d, J ¼ 8.8 Hz, 2H), 7.20 (d, J ¼ 8.9 Hz, 2H),
5.37 (s, 2H). 13C NMR (151 MHz, DMSO)
168.02, 167.56, 159.58,
d
12.52 (br, 1H), 8.27 (d, J ¼ 8.8 Hz, 2H),
Thiazolidinedione (56.2 mg, 0.48 mmol), TFI-7(int) (50 mg,
0.16 mmol), piperidine (cat.) and AcOH (cat.), pale yellow crystals,
d
60% yield, m.p. 203e205 ꢀC. 1H NMR (600 MHz, DMSO)
d 12.50 (br,
147.12, 144.45, 132.12, 131.60, 128.37, 126.11, 123.69, 120.84, 115.76,
68.27. [M ꢂ H]-: m/z calc. 355.0 found 355.1 m/z (100%).
1H), 7.77 (d, J ¼ 8.2 Hz, 2H), 7.70 (s, 1H), 7.68 (d, J ¼ 8.1 Hz, 2H),
7.24e7.19 (m, 2H), 7.16 (d, J ¼ 8.3 Hz, 2H), 5.27 (s, 2H), 3.86 (s, 3H).
13C NMR (151 MHz, DMSO)
d 167.95, 167.39, 151.15, 147.53, 141.63,
5.1.2.2. (5Z)-5-[(4-{[4-(trifluoromethyl)phenyl]methoxy}phenyl)
methylidene]-1,3-thiazolidine-2,4-dione (TFI-2).
131.94,128.76, 128.55, 128.34, 128.13, 128.06, 126.94, 125.60,125.42,
125.40, 125.37, 125.35, 125.14, 124.47, 123.33, 121.53, 120.68, 114.99,
Thiazolidinedione (63 mg, 0.54 mmol), TFI-2(int) (50 mg,
0.18 mmol), piperidine (cat.) and AcOH (cat.), pale yellowish crys-
112.50, 69.04, 55.83. 19F NMR (565 MHz, DMSO)
d
ꢂ60.96. [M ꢂ H]-:
m/z calc. 408.1 found 408.3 m/z (100%).
tals, 60% yield, m.p. 202e203 ꢀC. 1H NMR (600 MHz, DMSO)
d 12.53
(br, 1H), 7.77 (d, J ¼ 8.2 Hz, 2H), 7.75 (s, 1H), 7.68 (d, J ¼ 8.1 Hz, 2H),
5.1.2.8. Methyl
methyl]-2-methoxyphenyl}methoxy)benzoate (TFI-8).
Thiazolidinedione (53 mg, 0.45 mmol), TFI-8(int) (50 mg,
0.15 mmol), piperidine (cat.) and AcOH (cat.), pale yellowish crys-
tals, 60% yield, m.p. 233e236 ꢀC. 1H NMR (600 MHz, DMSO). 1H
4-({5-[(Z)-(2,4-dioxo-1,3-thiazolidin-5-ylidene)
7.58 (d, J ¼ 8.8 Hz, 2H), 7.19 (d, J ¼ 8.8 Hz, 2H), 5.32 (s, 2H). 13C NMR
(151 MHz, DMSO)
d
167.97, 167.49, 159.72, 141.44, 132.11, 131.65,
128.56, 128.35, 128.13, 125.97, 125.43, 125.40, 120.69, 115.74, 68.54,
56.02. 19F NMR (565 MHz, DMSO)
378.0 found 378.2 m/z (100%).
d
ꢂ60.96. [M ꢂ H]-: m/z calc.
NMR (600 MHz, DMSO)
d
12.50 (br, 1H), 7.99 (d, J ¼ 8.4 Hz, 2H), 7.69
(s, 1H), 7.60 (d, J ¼ 8.5 Hz, 2H), 7.21 (dd, J ¼ 4.3, 2.4 Hz, 2H), 7.17 (d,
5.1.2.3. Methyl
methyl]phenyl}methoxy)benzoate
4-({4-[(Z)-(2,4-dioxo-1,3-thiazolidin-5-ylidene)
(TFI-3). Thiazolidinedione
J ¼ 9.0 Hz, 1H), 5.26 (s, 2H), 3.86 (s, J ¼ 3H), 3.85 (s, 3H). 13C NMR
(151 MHz, DMSO)
d 167.98, 167.41, 166.01, 165.50, 151.17, 147.58,
(64 mg, 0.55 mmol), TFI-3(int) (50 mg, 18.5 mmol), piperidine (cat.)
and AcOH (cat.). (TFI-3), pale yellowish crystals, 60% yield, m.p.
142.30, 142.22, 131.93, 129.38, 129.36, 129.34, 129.08, 127.59, 127.57,
127.55, 125.58, 124.49, 120.67, 115.00, 112.50, 69.28, 60.76, 55.84,
52.16. [M ꢂ H]-: m/z calc. 398.1 found 398.2 m/z (100%).
220e222 ꢀC. 1H NMR (600 MHz, DMSO)
d 12.53 (br, 1H), 7.99 (d,
J ¼ 8.4 Hz, 2H), 7.74 (s, 1H), 7.59 (dd, J ¼ 20.4, 8.6 Hz, 4H), 7.18 (d,
J ¼ 8.9 Hz, 2H), 5.30 (s, 2H), 3.86 (s, 3H). 13C NMR (151 MHz, DMSO)
5.1.2.9. (5Z)-5-({3-methoxy-4-[(4-nitrophenyl)methoxy]phenyl}
methylidene)-1,3-thiazolidine-2,4-dione (TFI-9).
d
167.97, 167.49, 165.98, 165.47, 159.79, 142.07, 141.99, 132.09, 131.67,
129.39, 129.35, 129.13, 127.62, 127.61, 125.92, 120.64, 115.74, 68.78,
60.77, 52.18. [M ꢂ H]-: m/z calc. 368.1 found 368.1 m/z (100%).
Thiazolidinedione (60 mg, 0.52 mmol), TFI-9(int) (50 mg,
0.17 mmol), piperidine (cat.) and AcOH (cat.), pale yellowish crys-
tals, 65% yield, m.p. 205e208 ꢀC. 1H NMR (600 MHz, DMSO)
d 12.53
5.1.2.4. (5Z)-5-({3-nitro-4-[(4-nitrophenyl)methoxy]phenyl}methyl-
(br, 1H), 8.27 (d, J ¼ 8.8 Hz, 2H), 7.75 (s, 1H), 7.72 (d, J ¼ 8.9 Hz, 2H),
idene)-1,3-thiazolidine-2,4-dione
(TFI-4). Thiazolidinedione
7.25 (d, J ¼ 2.0 Hz, 1H), 7.17 (dt, J ¼ 8.4, 5.1 Hz, 2H), 5.42 (s, 2H), 3.85
(59.73 mg, 0.51 mmol), TFI-4(int) (50 mg, 0.17 mmol), piperidine
(cat.) and AcOH (cat.), pale yellowish crystals, 60% yield, m.p.
(s, 3H). 13C NMR (151 MHz, DMSO)
d 167.97, 149.26, 149.22, 147.11,
144.52, 131.96, 128.35, 126.45, 123.69, 123.39, 113.78, 113.71, 68.68,
55.72. [M ꢂ H]-: m/z calc. 385.1 found 385.1 m/z (100%).
242e243 ꢀC. 1H NMR (600 MHz, DMSO)
d
12.68 (br, 1H), 8.30 (d,
J ¼ 8.8 Hz, 1H), 8.21 (d, J ¼ 2.2 Hz, 1H), 7.88 (dd, J ¼ 9.1, 2.3 Hz, 1H),
7.81 (s, 1H), 7.74 (d, J ¼ 8.9 Hz, 1H), 7.60 (d, J ¼ 9.0 Hz, 1H), 5.57 (s,
5.1.2.10. (5Z)-5-[(3-methoxy-4-{[4-(trifluoromethyl)phenoxy]
1H). 13C NMR (151 MHz, DMSO)
d
167.63, 151.55, 147.27, 143.36,
methyl}phenyl)methylidene]-1,3-thiazolidine-2,4-dione
(TFI-10).
139.63, 135.23, 129.15, 128.12, 126.89, 126.28, 123.80, 116.39, 69.65.
[M ꢂ H]: m/z calc. 400.0 found 400.3 m/z (100%).
Thiazolidinedione (56 mg, 0.48 mmol), TFI-10(int) (50 mg,
0.16 mmol), piperidine (cat.) and AcOH (cat.), pale yellowish crys-
tals, 65% yield, m.p. 201e203 ꢀC. 1H NMR (600 MHz, DMSO)
d 12.51
5.1.2.5. Methyl
methyl]-2
4-({4-[(Z)-(2,4-dioxo-1,3-thiazolidin-5-ylidene)
nitrophenoxy}methyl)benzoate (TFI-5).
(br, 1H), 7.78 (d, J ¼ 8.0 Hz, 2H), 7.74 (s, 1H), 7.67 (d, J ¼ 8.1 Hz, 2H),
7.24 (s, 1H), 7.18 (dd, J ¼ 22.6, 8.5 Hz, 2H), 5.30 (s, 2H), 3.84 (s, 3H).
Thiazolidinedione (56.2 mg, 0.51 mmol), TFI-5(int) (50 mg,
0.16 mmol), piperidine (cat.) and AcOH (cat.), pale yellowish crys-
13C NMR (151 MHz, DMSO)
d 167.97, 167.43, 149.39, 149.24, 141.49,
132.03,128.77,128.56,128.35,128.15,126.93,126.30,125.45,125.43,
125.40, 125.38, 125.13, 123.46, 123.32, 120.89, 113.67, 68.95, 55.68.
tals, 60% yield, m.p. 225e227 ꢀC. 1H NMR (600 MHz, DMSO)
d 12.67
9

ꢀ
S.A. Tabatabaei Dakhili, D.J. Perez, K. Gopal et al.
European Journal of Medicinal Chemistry xxx (xxxx) xxx
19F NMR (565 MHz, DMSO)
found 408.3 m/z (100%).
d
ꢂ60.95. [M ꢂ H]-: m/z calc. 408.1
(100
mg/mL) were selected and grown in LB media (100 mg/mL
ampicillin) at 37 ꢀC until they reached the optical density (OD600)
of 0.8 and then 1 mM isopropyl
b -D-1- thiogalactopyranoside
5.1.2.11. Methyl4-({4-[(Z)-(2,4-dioxo-1,3-thiazolidin-5-ylidene)
methyl]-2 methoxyphenoxy}methyl)benzoate (TFI-11).
(IPTG) was added. After 6 h incubation at 37 ꢀC, the supernatant
layer was purified using glutathione resin (GenScript, USA)
following the manufacturer’s instructions.
Thiazolidinedione (56 mg, 0.48 mmol), TFI-11(int) (50 mg,
0.16 mmol), piperidine (cat.) and AcOH (cat.), yellowish crystals,
60% yield, m.p. 212e214 ꢀC. 1H NMR (600 MHz, DMSO)
d 12.53 (br,
5.6. Electrophoretic mobility shift assay
1H), 7.99 (d, J ¼ 8.4 Hz, 2H), 7.75 (s, 1H), 7.59 (d, J ¼ 8.5 Hz, 2H), 7.24
(d, J ¼ 2.0 Hz, 1H), 7.16 (dt, J ¼ 8.6, 5.3 Hz, 2H), 5.28 (s, 2H), 3.86 (s,
For the protein-DNA complex dissociation experiments we used
recombinant FOXM1-DBD, which was incubated with the corre-
sponding compound for 1.5 h at RT [35]. Then its target DNA oligo
(Forward strand: 50-/IRD700/-AAACAAACAAACAATCAAACAAA-
CAAACAATC-30) was incubated with the mixture for 20 min in a
buffer solution composed of 20 mM Tris (pH 7.5), 100 mM KCl,
1 mM EDTA, 0.1 mM DTT, and 10% glycerol. The mixture was then
loaded on a non-denaturing 6% polyacrylamide gel for 30 min at
120 V. The Ki value was calculated based on the previously deter-
mined Kd (Data not shown) value using the following formula Eq.
(1):
3H), 3.84 (s, 3H).13C NMR (151 MHz, DMSO)
d 167.97, 167.42, 166.00,
149.47, 149.25, 142.13, 132.04, 129.38, 129.13, 127.71, 127.66, 126.26,
125.80, 123.45, 120.85, 113.70, 113.65, 112.78, 109.86, 69.30, 69.21,
55.68, 55.65, 52.18. [M ꢂ H]-: m/z calc. 398.1 found 398.2 m/z
(100%).
5.2. Cell culture
The MDA-MB-231 cancer cell line was a gift from Dr. Frank
Wuest (Cross Cancer Institute; Edmonton, Alberta, Canada). RPMI
media was supplemented with 10% fetal bovine serum (FBS) in a 5%
CO2 atmosphere at 37 ꢀC to grow and maintain the cells.
½Iꢄ
ꢀ
ꢁ
Ki ¼
Eq. (1)
ð½Iꢄ50
Þ
½Pꢄ0
þ
Kd
Kdþ1
5.3. Colony formation assay
MDA-MB-231 cells were plated at 3 ꢃ 105 Confluence and
treated with different concentration of compound TFI-10. After
24 h the cells were trypsinized, resuspended and counted. 750 cells
were then seeded and incubated for 10 days and media Renewed on
the 4th day. On the 10th day, media was removed from the petri
dishes and washed once with cold PBS. The colonies were stained
with 5 mL of 1% Crystal Violet for 30 min. The dishes were fixed by
methanol and rinsed with water three times, air-dried and the
colonies were counted using Imagej.
where: [I]50 ¼ IC50 of the inhibitor; [L]50 ¼ concentration of IR-
labeled DNA at 50% inhibition; [P] ¼ concentration of the FOXM1
protein; and Kd ¼ dissociation constant calculated from the initial
titration curve.
5.7. Luciferase reporter assay
The 6x-FOXM1 firefly luciferase reporter and its corresponding
backbone plasmid-pGL4.10 (i.e empty reporter) were gifts from Drs.
Carter J Barger and Adam R. Karpf. Briefly, MD-MB-231 cells were
transiently transfected with equal amounts of the 6x-FOXM1 and
empty luciferase reporter plasmids. Following 24 h incubation, cells
5.4. Western blot
6 well plates (3 ꢃ 105 confluency) were treated at different
concentrations for 24 h. Then, the cells were washed twice with ice
cold PBS before incubating and extracting with RIPA lysis buffer
(Thermofisher) for 30 min. The protein levels were quantified using
were treated with 40 mM of drug for 12 h. Cells were then washed
with cold PBS and lysed with Reporter Lysis Buffer (Promega).
Protein concentration was estimated, and equal amounts of total
protein from each lysate was analyzed for firefly luciferase activity
using the Luciferase Reporter Assay System (Promega). Assays were
performed in triplicates.
the Bradford assay prior to loading of 40 mg of protein into a 10%
SDS-PAGE (Sodium dodecyl sulfate Polyacrylamide gel electro-
phoresis). Upon completion of the run, the proteins were trans-
ferred from the gel to a nitrocellulose membrane (Thermofisher)
and blocked with 10% fat-free milk in TBST for 45 min. The corre-
sponding antibody was incubated with the membrane (1:1000
dilution) at 4 ꢀC overnight. Next day, the membrane was washed
with TBST before incubating the appropriate secondary antibody
for 45 min at RT. Then, the membrane was washed three times
(15 min total) with TBST and the conjugated protein bands were
visualized by adding the Chemiluminescence reagent (Thermo-
fisher) using ImageQuant™ LAS 4000 mini biomolecular imager
(GE Healthcare Life Sciences). The quantification was carried out for
5.8. Quantitative real time PCR (qPCR)
The RNeasy Mini Kit (Qiagen, Valencia, CA) was used to isolate
total RNA from cell lines. Trace DNA was removed on column
through DNA digestion with DNAse I (Invitrogen, Carlsbad, CA). 1 mg
of total RNA was used to prepare cDNA with the High-Capacity
cDNA Reverse Transcription Kit (Invitrogen, Carlsbad, CA). qPCR
reactions were prepared in 96-well plates (Invitrogen, Carlsbad, CA)
with the PowerUp™ SYBR™ Green Master Mix (Invitrogen, Carls-
bad, CA) according to manufacturer’s protocol. The Mastercycler®
ep Realplex system (Eppendorf, Hamburg, Germany) was used for
cycling and detecting amplification. Primers for all genes have been
validated in prior studies and were purchased from Invitrogen.
Sequences of the primers used in this study include: FOXM1: F-5-
all proteins relative to
monoclonal (mouse, sc-271746) antibody; SP1 monoclonal (mouse,
sc-420) antibody; -Actin monoclonal (mouse, sc-47778) anti-
b-Actin using ImageJ. We used a FOXM1
b
bodies and anti-mouse (sc-516102) and anti-rabbit (sc-2030) sec-
ondary antibodies from Santa Cruz Biotechnology.
CGTCGGCCACTGATTCTCAAA-3
TAGGTTC-3, CDC25B: F-5-CCTCCGAATCTTCTGATGCAG-3 & R- 5-
GCGTCTGATGGCAAACTGC-3, CCNB1: F-5-GTAATGTTGTA-
&
R-
5-GGCAGGGGATCTCT-
5.5. Protein expression and purification
GAGTTGGTGTCC-3 & R- 5-CATGGTGCACTTTCCTCCTT-3, GAPDH: F-
50GGTCTCCTCTGACTTCAACAGCG-3 & R- 5 ACCACCCTGTTGCTG-
TAGCCAA-3. qPCR cycling was setup as denaturation at 95 ꢀC for
10 min, followed by 40 cycles at 95 ꢀC for 15 s, annealing &
BL21 (DE3) competent cells where used to transform the PEX-N-
GST FOXM1-DBD plasmid (OriGene Technologies, USA). Next,
positive colonies on LB agar media supplemented with ampicillin
10

ꢀ
S.A. Tabatabaei Dakhili, D.J. Perez, K. Gopal et al.
European Journal of Medicinal Chemistry xxx (xxxx) xxx
extension at 60 ꢀC for 1 min. The relative gene expression was
determined using the DD-CT method. For each treatment group,
target gene expression was normalized to GAPDH expression fol-
lowed by normalization of target gene expression to the control
treatment.
Biochem. Biophys. Res. Commun. 500 (2018) 924e929, https://doi.org/
10.1016/j.bbrc.2018.04.201.
[4] M. Halasi, A.L. Gartel, Targeting FOXM1 in cancer, Biochem. Pharmacol. 85
(2013) 644e652, https://doi.org/10.1016/j.bcp.2012.10.013.
[5] M. Halasi, B. Pandit, M. Wang, V. Nogueira, N. Hay, A.L. Gartel, Combination of
oxidative stress and FOXM1 inhibitors induces apoptosis in cancer cells and
inhibits xenograft tumor growth, Am. J. Pathol. 183 (2013) 257e265, https://
doi.org/10.1016/j.ajpath.2013.03.012.
[6] I.M. Kim, T. Ackerson, S. Ramakrishna, M. Tretiakova, I.C. Wang, T. V Kalin,
M.L. Major, G.A. Gusarova, H.M. Yoder, R.H. Costa, V.V. Kalinichenko, The
Forkhead Box m1 transcription factor stimulates the proliferation of tumor
cells during development of lung cancer, Canc. Res. 66 (2006) 2153e2161,
https://doi.org/10.1158/0008-5472.CAN-05-3003.
[7] P.A. Madureira, R. Varshochi, D. Constantinidou, R.E. Francis, R.C. Coombes,
K.M. Yao, E.W. Lam, The Forkhead box M1 protein regulates the transcription
of the estrogen receptor alpha in breast cancer cells, J. Biol. Chem. 281 (2006)
25167e25176, https://doi.org/10.1074/jbc.M603906200.
[8] H. Wang, D.I. Hammoudeh, A. V Follis, B.E. Reese, J.S. Lazo, S.J. Metallo, E. V
Prochownik, Improved low molecular weight Myc-Max inhibitors, Mol. Canc.
Therapeut. 6 (2007) 2399e2408, https://doi.org/10.1158/1535-7163.MCT-07-
0005.
[9] S. Nakamura, I. Hirano, K. Okinaka, T. Takemura, D. Yokota, T. Ono, K. Shigeno,
K. Shibata, S. Fujisawa, K. Ohnishi, The FOXM1 transcriptional factor promotes
the proliferation of leukemia cells through modulation of cell cycle progres-
sion in acute myeloid leukemia, Carcinogenesis 31 (2010) 2012e2021, https://
doi.org/10.1093/carcin/bgq185.
[10] K.M. Huynh, J.W. Soh, R. Dash, D. Sarkar, P.B. Fisher, D. Kang, FOXM1
expression mediates growth suppression during terminal differentiation of
HO-1 human metastatic melanoma cells, J. Cell. Physiol. 226 (2011) 194e204,
https://doi.org/10.1002/jcp.22326.
[11] J.R. Janus, R.R. Laborde, A.J. Greenberg, V.W. Wang, W. Wei, A. Trier,
S.M. Olsen, E.J. Moore, K.D. Olsen, D.I. Smith, Linking expression of FOXM1,
CEP55 and HELLS to tumorigenesis in oropharyngeal squamous cell carci-
noma, Laryngoscope 121 (2011) 2598e2603, https://doi.org/10.1002/
lary.22379.
[12] X. Song, S.S. Fiati Kenston, J. Zhao, D. Yang, Y. Gu, Roles of FoxM1 in cell
regulation and breast cancer targeting therapy, Med. Oncol. 34 (2017) 41,
https://doi.org/10.1007/s12032-017-0888-3.
[13] D.W. Chan, S.Y. Yu, P.M. Chiu, K.M. Yao, V.W. Liu, A.N. Cheung, H.Y. Ngan,
Over-expression of FOXM1 transcription factor is associated with cervical
cancer progression and pathogenesis, J. Pathol. 215 (2008) 245e252, https://
doi.org/10.1002/path.2355.
5.9. Molecular modeling and molecular dynamic simulation
The crystal structure of FOXM1-DBD [37], was retrieved from
Protein data bank (PDB_ID: 3G73). Chimera v 1.10.12 [38], was used
to prepare the structure followed by the charge assignment of
protonated groups to pH 7 using PROPKA. MD simulation of the
FOXM1-DBD was performed using Gromacs 2018 package. We
solvated the protein complex in a box with TIP3P water model with
1 nm margin on each side. The system was further neutralized with
NaCl before performing the steepest descent minimization using
the AMBER99SB0ILDN force field. The system was gradually heated
to 300 K using Berendsen thermostat while the pressure was
maintained and equilibrated at 1 bar with Parrinello-Rahman
barostat respectively. Finally, while maintaining the periodic
boundary condition, 50 ns production run was performed. The
Particle Mesh Ewald (PME) summation and Lenard-Jones were
used to calculate the long range electrostatic and VDW interaction
respectively. Next, TFI-10 and TGZ were docked in the FOXM1-DBD
binding pocket (previously reported and validated) using Auto-
doktools and the AnteChamber Python Parser Interface (ACPYPE)
was used to generate their parameters. The binding free energy of
the TFI-10 and TGZ was calculated using G_MMPBSA as previously
reported. Root Mean Square Deviation (RMSD), Root Mean Square
Fluctuation (RMSF) and for the complex, Ligand positional RMSD
were calculated using the Gromacs package and the graphs were
plotted using Graphpad Prism 8. Schrodinger’s Maestro and Dis-
covery Studio Visualizer 32, were used to visualize and make the
figures.
[14] J.T. Xia, H. Wang, L.J. Liang, B.G. Peng, Z.F. Wu, L.Z. Chen, L. Xue, Z. Li, W. Li,
Overexpression of FOXM1 is associated with poor prognosis and clinico-
pathologic stage of pancreatic ductal adenocarcinoma, Pancreas 41 (2012)
629e635, https://doi.org/10.1097/MPA.0b013e31823bcef2.
ꢀ
[15] S.A. Tabatabaei-Dakhili, R. Aguayo-Ortiz, L. Domínguez, C.A. Velazquez-Mar-
Declaration of competing interest
tínez, Untying the knot of transcription factor druggability: molecular
modeling study of FOXM1 inhibitors, J. Mol. Graph. Model. 80 (2018)
197e210, https://doi.org/10.1016/j.jmgm.2018.01.009.
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
[16] V. Petrovic, R.H. Costa, L.F. Lau, P. Raychaudhuri, A.L. Tyner, Negative regula-
tion of the oncogenic transcription factor FoxM1 by thiazolidinediones and
mithramycin, Canc. Biol. Ther. 9 (2010) 1008e1016, https://doi.org/10.4161/
cbt.9.12.11710.
[17] X. Kong, L. Li, Z. Li, X. Le, C. Huang, Z. Jia, J. Cui, S. Huang, L. Wang, K. Xie,
Dysregulated expression of FOXM1 isoforms drives progression of pancreatic
cancer, Canc. Res. 73 (2013) 3987e3996, https://doi.org/10.1158/0008-
5472.CAN-12-3859.
Acknowledgements
Authors acknowledge the support from the Noujaim Fund for
Strategic Initiatives, Faculty of Pharmacy and Pharmaceutical Sci-
ences, University of Alberta. The National Council for Science and
Technology (CONACYT) provided a fellowship for D.J.P (CVU:
287012). This research was enabled in part by Compute Canada and
its regional partner WestGrid. The graphical abstract and Fig. 2
were created using the program provided by BioRender.
[18] J.-W. Huang, C.-W. Shiau, Y.-T. Yang, S.K. Kulp, K.-F. Chen, R.W. Brueggemeier,
C.L. Shapiro, C.-S. Chen, Peroxisome proliferator-activated receptor
g-inde-
pendent ablation of cyclin D1 by thiazolidinediones and their derivatives in
breast cancer cells, Mol. Pharmacol. 67 (2005) 1342e1348, https://doi.org/
10.1124/mol.104.007732.
[19] B.R. Bhattarai, B. Kafle, J.-S. Hwang, D. Khadka, S.-M. Lee, J.-S. Kang, S.W. Ham,
I.-O. Han, H. Park, H. Cho, Thiazolidinedione derivatives as PTP1B inhibitors
with antihyperglycemic and antiobesity effects, Bioorg. Med. Chem. Lett 19
(2009) 6161e6165, https://doi.org/10.1016/j.bmcl.2009.09.020.
[20] K. Bhooshan, A.N. G, K. Deegendra, P. Hwangseo, C. Hyeongjin, Isoxazol-5(4H)
one derivatives as PTP1B inhibitors showing an anti-obesity effect, Chem.
Asian J. 6 (2011) 2073e2079, https://doi.org/10.1002/asia.201100154.
[21] R. Carrasco-Gomez, S. Keppner-Witter, M. Hieke, L. Lange, G. Schneider,
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.ejmech.2020.112902o.
€
M. Schubert-Zsilavecz, E. Proschak, B. Spankuch, Vanillin-derived anti-
proliferative compounds influence Plk1 activity, Bioorg. Med. Chem. Lett 24
(2014) 5063e5069, https://doi.org/10.1016/j.bmcl.2014.09.015.
References
ꢁ
ꢁ
[22] M. Hrast, K. Rozman, M. Jukic, D. Patin, S. Gobec, M. Sova, Synthesis and
structureeactivity relationship study of novel quinazolinone-based inhibitors
of MurA, Bioorg. Med. Chem. Lett 27 (2017) 3529e3533, https://doi.org/
10.1016/j.bmcl.2017.05.064.
[1] D. Nandi, P.S. Cheema, N. Jaiswal, A. Nag, FoxM1: repurposing an oncogene as
a biomarker, Semin. Canc. Biol. 52 (2018) 74e84, https://doi.org/10.1016/
j.semcancer.2017.08.009.
[2] F.C. Kelleher, H. O’Sullivan, FOXM1 in sarcoma: role in cell cycle, pluripotency
genes and stem cell pathways, Oncotarget 7 (2016) 42792e42804, https://
doi.org/10.18632/oncotarget.8669.
[23] Y. Momose, K. Meguro, H. Ikeda, C. Hatanaka, S. Oi, T. Sohda, Studies on
antidiabetic agents. X. Synthesis and biological activities of pioglitazone and
related compounds, Chem. Pharm. Bull. 39 (1991) 1440e1445, https://doi.org/
10.1248/cpb.39.1440.
[24] S. Blagden, J. de Bono, Drugging cell cycle kinases in cancer therapy, Curr.
Drug Targets 6 (2005) 325e335, https://doi.org/10.2174/1389450053765824.
[25] J. Rader, M.R. Russell, L.S. Hart, M.S. Nakazawa, L.T. Belcastro, D. Martinez, Y. Li,
[3] N. Chai, H.H. Xie, J.P. Yin, K.D. Sa, Y. Guo, M. Wang, J. Liu, X.F. Zhang, X. Zhang,
H. Yin, Y.Z. Nie, K.C. Wu, A.G. Yang, R. Zhang, FOXM1 promotes proliferation in
human hepatocellular carcinoma cells by transcriptional activation of CCNB1,
11

ꢀ
S.A. Tabatabaei Dakhili, D.J. Perez, K. Gopal et al.
European Journal of Medicinal Chemistry xxx (xxxx) xxx
E.L. Carpenter, E.F. Attiyeh, S.J. Diskin, S. Kim, S. Parasuraman, G. Caponigro,
R.W. Schnepp, A.C. Wood, B. Pawel, K.A. Cole, J.M. Maris, Dual CDK4/CDK6
inhibition induces cell-cycle arrest and senescence in neuroblastoma, Clin.
Canc. Res. 19 (2013) 6173e6182, https://doi.org/10.1158/1078-0432.CCR-13-
1675.
proliferator-activated receptor gamma (PPAR gamma) ligand, selectively in-
duces the early growth response-1 gene independently of PPAR gamma. A
novel mechanism for its anti-tumorigenic activity, J. Biol. Chem. 278 (2003)
5845e5853, https://doi.org/10.1074/jbc.M208394200.
[33] T. Okura, M. Nakamura, Y. Takata, S. Watanabe, Y. Kitami, K. Hiwada, Trogli-
tazone induces apoptosis via the p53 and Gadd45 pathway in vascular smooth
muscle cells, Eur. J. Pharmacol. 407 (2000) 227e235, https://doi.org/10.1016/
s0014-2999(00)00758-5.
[26] A.L. Gartel, Mechanisms of apoptosis induced by anticancer compounds in
melanoma cells, Curr. Top. Med. Chem. 12 (2012) 50e52, https://doi.org/
10.2174/156802612798919196.
[27] A.L. Gartel, A new target for proteasome inhibitors: FoxM1, Expet Opin. Invest.
Drugs 19 (2010) 235e242, https://doi.org/10.1517/13543780903563364.
[28] E. Schmitt, R. Boutros, C. Froment, B. Monsarrat, B. Ducommun, C. Dozier,
CHK1 phosphorylates CDC25B during the cell cycle in the absence of DNA
damage, J. Cell Sci. 119 (2006) 4269e4275, https://doi.org/10.1242/jcs.03200.
[29] H. Koga, S. Sakisaka, M. Harada, T. Takagi, S. Hanada, E. Taniguchi,
T. Kawaguchi, K. Sasatomi, R. Kimura, O. Hashimoto, T. Ueno, H. Yano,
M. Kojiro, M. Sata, Involvement of p21(WAF1/Cip1), p27(Kip1), and
p18(INK4c) in troglitazone-induced cell-cycle arrest in human hepatoma cell
lines, Hepatology 33 (2001) 1087e1097, https://doi.org/10.1053/
jhep.2001.24024.
[30] S. Takeuchi, T. Okumura, W. Motomura, M. Nagamine, N. Takahashi, Y. Kohgo,
Troglitazone induces G1 arrest by p27(Kip1) induction that is mediated by
inhibition of proteasome in human gastric cancer cells, Japanse J. Cancer Res.
93 (2002) 774e782, https://doi.org/10.1111/j.1349-7006.2002.tb01319.x.
[31] M.A. Bae, B.J. Song, Critical role of c-Jun N-terminal protein kinase activation
in troglitazone-induced apoptosis of human HepG2 hepatoma cells, Mol.
Pharmacol. 63 (2003) 401e408, https://doi.org/10.1124/mol.63.2.401.
[34] M. Halasi, A.L. Gartel, Suppression of FOXM1 sensitizes human cancer cells to
cell death induced by DNA-damage, PloS One 7 (2012), e31761, https://
doi.org/10.1371/journal.pone.0031761.
[35] M. V Gormally, T.S. Dexheimer, G. Marsico, D.A. Sanders, C. Lowe, D. Matak-
Vinkovic, S. Michael, A. Jadhav, G. Rai, D.J. Maloney, A. Simeonov,
S. Balasubramanian, Suppression of the FOXM1 transcriptional programme
via novel small molecule inhibition, Nat. Commun. 5 (2014) 5165, https://
doi.org/10.1038/ncomms6165.
[36] A.L. Gartel, FoxM1 inhibitors as potential anticancer drugs, Expert Opin. Ther.
Targets 12 (2008) 663e665, https://doi.org/10.1517/14728222.12.6.663.
[37] D.R. Littler, M. Alvarez-Fernandez, A. Stein, R.G. Hibbert, T. Heidebrecht,
P. Aloy, R.H. Medema, A. Perrakis, Structure of the FoxM1 DNA-recognition
domain bound to
a promoter sequence, Nucleic Acids Res. 38 (2010)
4527e4538, https://doi.org/10.1093/nar/gkq194.
[38] E.F. Pettersen, T.D. Goddard, C.C. Huang, G.S. Couch, D.M. Greenblatt,
E.C. Meng, T.E. Ferrin, UCSF Chimerada visualization system for exploratory
research and analysis, J. Comput. Chem. 25 (2004) 1605e1612, https://
doi.org/10.1002/jcc.20084.
[32] S.J. Baek, L.C. Wilson, L.C. Hsi, T.E. Eling, Troglitazone,
a peroxisome
12