Exploring quercetin and luteolin derivatives as antiangiogenic agents

Article abstract of DOI:10.1016/j.ejmech.2015.04.056

The formation of new blood vessels from the pre-existing vasculature (angiogenesis) is a crucial stage in cancer progression and, indeed, angiogenesis inhibitors are now used as anticancer agents, clinically. Here we have explored the potential of flavonoid derivatives as antiangiogenic agents. Specifically, we have synthesised methoxy and 4-thio derivatives of the natural flavones quercetin and luteolin, two of which (4-thio quercetin and 4-thio luteolin) had never been previously reported. Seven of these compounds showed significant (p < 0.05) antiangiogenic activity in an in vitro scratch assay. Their activity ranged from an 86% inhibition of the vascular endothelium growth factor (VEGF)-stimulated migration (observed for methoxyquercetin at 10 μM and for luteolin at 1 μM) to a 36% inhibition (for thiomethoxy quercetin at 10 μM). Western blotting studies showed that most (4 out of 7) compounds inhibited phosphorylation of the VEGF receptor-2 (VEGFR2), suggesting that the antiangiogenic activity was due to an interference with the VEGF/VEGFR2 pathway. Molecular modelling studies looking at the affinity of our compounds towards VEGFR and/or VEGF confirmed this hypothesis, and indeed the compound with the highest antiangiogenic activity (methoxyquercetin) showed the highest affinity towards VEGFR and VEGF. As reports from others have suggested that structurally similar compounds can elicit biological responses via a non-specific, promiscuous membrane perturbation, potential interactions of the active compounds with a model lipid bilayer were assessed via DSC. Luteolin and its derivatives did not perturb the model membrane even at concentrations 10 times higher than the biologically active concentration and only subtle interactions were observed for quercetin and its derivatives. Finally, cytotoxicity assessment of these flavonoid derivatives against MCF-7 breast cancer cells demonstrated also a direct anticancer activity albeit at generally higher concentrations than those required for an antiangiogenic effect (10 fold higher for the methoxy analogues). Taken together these results show promise for flavonoid derivatives as antiangiogenic agents.

Full text of DOI:10.1016/j.ejmech.2015.04.056

Accepted Manuscript
Exploring quercetin and luteolin derivatives as antiangiogenic agents
Divyashree Ravishankar, Kimberly A. Watson, Samuel Y. Boateng, Rebecca J.
Green, Francesca Greco, Helen. M.I. Osborn
PII:
S0223-5234(15)30020-9
10.1016/j.ejmech.2015.04.056
EJMECH 7871
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 9 February 2015
Revised Date: 24 April 2015
Accepted Date: 27 April 2015
Please cite this article as: D. Ravishankar, K.A. Watson, S.Y. Boateng, R.J. Green, F. Greco, H.M.I.
Osborn, Exploring quercetin and luteolin derivatives as antiangiogenic agents, European Journal of
Medicinal Chemistry (2015), doi: 10.1016/j.ejmech.2015.04.056.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical abstract

ACCEPTED MANUSCRIPT
Exploring quercetin and luteolin derivatives as antiangiogenic
agents
Divyashree Ravishankara, Kimberly A. Watsonb, Samuel Y. Boatengb, Rebecca J. Greena,
Francesca Grecoa*, Helen. M. I. Osborna*.
aSchool of Pharmacy, University of Reading, Whiteknights, RG6 6AD, Berkshire, UK.
bInstitute of Cardiovascular and Metabolic Research, School of Biological Sciences, University
of Reading, Whiteknights, RG6 6AD, Berkshire, UK.
ABSTRACT: The formation of new blood vessels from the pre-existing vasculature
(angiogenesis) is a crucial stage in cancer progression and, indeed, angiogenesis inhibitors are
now used as anticancer agents, clinically. Here we have explored the potential of flavonoid
derivatives as antiangiogenic agents. Specifically, we have synthesised methoxy and 4-thio
derivatives of the natural flavones quercetin and luteolin, two of which (4-thio quercetin and 4-
thio luteolin) had never been previously reported. Seven of these compounds showed significant
(P<0.05) antiangiogenic activity in an in vitro scratch assay. Their activity ranged from an 86%
inhibition of the vascular endothelium growth factor (VEGF)-stimulated migration (observed for
methoxyquercetin at 10 µM and for luteolin at 1 µM) to a 36% inhibition (for thiomethoxy
quercetin at 10 µM). Western blotting studies showed that most (4 out of 7) compounds
inhibited phosphorylation of the VEGF receptor-2 (VEGFR2), suggesting that the
antiangiogenic activity was due to an interference with the VEGF/VEGFR2 pathway. Molecular
modelling studies looking at the affinity of our compounds towards VEGFR and/or VEGF
confirmed this hypothesis, and indeed the compound with the highest antiangiogenic activity
(methoxyquercetin) showed the highest affinity towards VEGFR and VEGF. As reports from
others have suggested that structurally similar compounds can elicit biological responses via a
non-specific, promiscuous membrane perturbation, potential interactions of the active
compounds with a model lipid bilayer were assessed via DSC. Luteolin and its derivatives did
not perturb the model membrane even at concentrations 10 times higher than the biologically
1

ACCEPTED MANUSCRIPT
active concentration and only subtle interactions were observed for quercetin and its derivatives.
Finally, cytotoxicity assessment of these flavonoid derivatives against MCF-7 breast cancer cells
demonstrated also a direct anticancer activity albeit at generally higher concentrations than those
required for an antiangiogenic effect (10 fold higher for the methoxy analogues). Taken together
these results show promise for flavonoid derivatives as antiangiogenic agents.
KEYWORDS: Angiogenesis, flavones, quercetin, luteolin, antiangiogenic therapy
1 INTRODUCTION
Angiogenesis is a physiological process involving the formation of new blood vessels from pre-
existing blood vessels. It is a complex process characterized by the proliferation, migration,
sprouting and elongation of endothelial cells [1], and plays an important role during cell
reproduction and organ development, as well as in wound healing processes. Under pathological
conditions, angiogenesis is poorly regulated and contributes to the pathogenesis of various
diseases such as rheumatoid arthritis, inflammation, psoriasis, degenerative eye conditions and,
of particular relevance to our study, cancer [2]. Sustained angiogenesis is one of the central
hallmarks of cancer [3,4] and has been validated as a key target for cancer therapy [5], as
evidenced by several antiangiogenic agents having entered clinical practice in the last decade
(for example, bevacizumab (Avastin® for several metastatic cancers) and sunitinib
maleate (SU11248, Sutent® for renal cell carcinoma and gastrointestinal stromal tumour)).
Flavonoids are polyphenolic compounds that have attracted ongoing and recent interest due to
their potential anticancer properties. They are very well known for their antiproliferative
activities against various cell lines [6] with some flavonoids such as Flavopiridol [7–9], silibinin
(also known as silybin) [10], quercetin [11] and its derivative QC12 [12] having progressed to
various stages of clinical trials. Interestingly, recent studies have highlighted that two common
dietary flavonoids (quercetin and luteolin) inhibit angiogenesis in vitro and in vivo [13,14] at
2

ACCEPTED MANUSCRIPT
concentrations 10-40 µM. Also, other structurally similar flavones have been reported to possess
antiangiogenic properties. In order to develop and optimise the flavone scaffolds into a
pharmaceutical agent, a better understanding of the structure-activity relationships (SARs) is
crucial. Hence, in this study, the antiangiogenic properties of a panel of small molecules,
comprised of quercetin and luteolin, as well as their rationally designed and synthesised
derivatives, are probed to allow a better understanding of the structure activity relationships of
flavonoids as antiangiogenic agents.
Many antiangiogenic agents exhibit their action by interfering with the vascular endothelial
growth factor (VEGF) and the VEGF receptor-2 (VEGFR2) [15]. Therefore, the compounds
reported in our study were evaluated for their potential to inhibit the VEGF-stimulated effects in
an in vitro model comprised of endothelial cells. The molecular mechanism of action of these
derivatives was investigated using Western blotting and additionally in silico studies were
conducted to probe the interactions of this panel of molecules with VEGFR2 and VEGF.
Additionally, we probed the interactions of these synthesised compounds with a model lipid
bilayer in order to determine the possible contribution of a membrane perturbation effect
towards the observed biological activity. Finally, as flavonoid derivatives are known to display
direct anticancer activity against tumour cells, cytotoxicity studies against a breast cancer cell
line (MCF-7) and a chemotherapy resistant sub-strain (MCF-7/DX) were also carried out. This
cytotoxicity assessment was carried out to investigate how variations in the chemical structure
of the flavonoid derivatives affected their activity, but also to establish the concentrations
required to elicit direct anticancer activity compared to those required for antiangiogenic effects.
Indeed, a widely reported problem associated with antiangiogenic drugs that target endothelial
proliferation is their inhibition of the proliferation of other cell types, which leads to undesirable
cell toxicity; molecules with high selectivity towards VEGF/VEGFR2 with low cytotoxicity are
3

ACCEPTED MANUSCRIPT
considered to be ideal candidates for antiangiogenic therapy as well as for combination therapy
[16].
2 RESULTS AND DISCUSSION
This paper reports the synthesis of methoxy and 4-thio derivatives of quercetin and luteolin,
which includes two new derivatives (4 and 12) and explores how subtle variations in the
chemical structure of flavonoids affect their biological activity, particularly in relation to their
antiangiogenic activity.
2.1 Synthesis
In order to gain a better understanding of SARs, a rational panel of compounds (2-4 and 9-12)
comprised of flavonoids with differing numbers of hydroxyl groups (quercetin versus luteolin),
different lipophilicities (hydroxyl versus methylated derivatives), and different key functional
groups (4-C=O versus 4-C=S) was synthesised in good yields (ranging from 20% to 30% overall
yield). The derivatives of quercetin (2-4) and luteolin (9-12) were prepared using the general
synthetic approaches outlined in Scheme-1 and 2 respectively.
Synthesis of quercetin derivatives: The polymethylated derivative of quercetin, namely
methoxyquercetin (2) was obtained by the methylation of quercetin (1) using dimethyl sulfate in
the presence of 15% KOH, in 60% yield. The compound methoxyquercetin (2) was treated with
Lawesson’s reagent in anhydrous toluene to obtain the 4-thio derivative, thiomethoxyquercetin
(3) in 65% yield. The thiomethoxyquercetin (3) was demethylated using BBr3 in anhydrous
dichloromethane at 40 oC for 18 h to afford thioquercetin (4) in 55% yield (Scheme-1).
Synthesis of luteolin derivatives: The first step in the synthesis of luteolin (10) and its
derivatives was esterification of 2-hydroxy-4,6-dimethoxyacetophenone (5) with 3,4-dimethoxy
benzoyl chloride (6) in pyridine and this was achieved in 60% yield. The obtained ester (7) was
4

ACCEPTED MANUSCRIPT
converted to a 1,3-diketone (8) via a Baker-Venkataramann rearrangement in the presence of
KOH in pyridine in approximately 75% yield. This 1,3-diketone (8) underwent cyclo-
dehydration in the presence of an acid to give the methoxyluteolin (9) in 70% yield. The
hydroxy flavone, luteolin (10), was obtained by demethylation of methoxyluteolin using BBr3 in
o
anhydrous dichloromethane at 40 C for 18 h in 59% yield. The 4-thiomethoxy derivative of
luteolin (11) was obtained in 61% yield by treating the methoxyluteolin (9) with Lawesson’s
reagent. Thioluteolin (12) was successfully synthesised by the demethylation of
o
thiomethoxyluteolin (11) using BBr3 in anhydrous dichloromethane at 40 C for 18 h in 55%
yield (Scheme-2). The identities of these synthesised compounds were confirmed by 1H NMR,
13C NMR and IR spectroscopy, and mass spectrometry. Also, the purity of these compounds
were determined to be >95% pure by HPLC.
2.2 Antiangiogenic studies
2.2.1 Scratch assay
Following the synthesis of the quercetin and luteolin derivatives (2-4 and 9-12, respectively),
their antiangiogenic properties were assessed in vitro. The antiangiogenic potencies of these
compounds were assessed at the cellular level using human umbilical endothelial vascular cells
(HUVECs) using a well-established in vitro model [17,18]. The migration of endothelial cell is a
key step in the process of angiogenesis [19]. HUVECs show very high rates of migration when
stimulated with the chemokine VEGF. Therefore, the antiangiogenic potencies of these
derivatives were assessed through their ability to inhibit the VEGF-induced migration of
HUVECs using an in vitro scratch/wound healing assay [20].
The effects of quercetin, luteolin and their derivatives on the inhibition of the chemotactic
motility of HUVECs are shown in Figures-2 and 3. All compounds were initially screened at 10
µM concentration. However, changes in the cell morphology were observed in the presence of
5

ACCEPTED MANUSCRIPT
luteolin (Figure-2) as well as the 4-thio derivatives of quercetin (4) and luteolin (12) (data not
shown) at this concentration, which indicates an undesirable level of toxicity. Hence,
compounds 10, 4 and 12, along with quercetin (for comparison), were tested at a lower (1 µM)
concentration. Figure 1 shows the effects of quercetin and its derivatives on VEGF-induced cell
migration. Quercetin at 10 µM exhibited 74% inhibition of VEGF-induced cell migration, and
this is in agreement with previous reports[13]. Also, when tested at 1 µM concentration,
quercetin still retained antiangiogenic properties with 54% inhibition of VEGF-stimulated
migration. The newly synthesised 4-thio derivative of quercetin (4) displayed 47% inhibitory
effect at 1 µM concentration, which is comparable to that of quercetin at the same concentration.
Interestingly, the methoxy derivative of quercetin (2) at 10 µM concentration showed 86%
inhibition of VEGF-stimulated migration whereas its 4-thio analogue (3) at 10 µM concentration
displayed 63% inhibition. Among quercetin and its derivatives, the inhibitory effect observed is
in the order of compound-2≥compound-1>compound-3>compound-4. Figure 2 shows the
profile for luteolin and its derivatives. Luteolin at 10 µM displayed undesirable morphological
changes, however such observation has not been reported in the previous study [14]. At 1 µM
concentration, luteolin exhibited 86% inhibition of VEGF-stimulated migration. The new 4-thio
derivative of luteolin (12) showed statistically significant inhibition of VEGF-stimulated
migration at 1 µM. The methoxy derivative of luteolin (9) and its 4-thio analogue (11) at their
10 µM concentration exhibited 58% and 36% inhibition of VEGF-stimulated migration
respectively. Among luteolin and its derivatives, the inhibitory effect observed is in the order of
compound-10>compound-9>compound-11>compound-12.
2.2.2 Western blotting
Having observed the inhibition of VEGF induced cell migration in the presence of quercetin
and the luteolin derivatives, their possible mechanisms of angiogenesis inhibition were
elucidated using Western blotting. VEGF signalling via the phosphorylation of VEGFR2
6

ACCEPTED MANUSCRIPT
promotes cell growth and survival [21]. Tyr1175 is a major phosphorylation site of VEGFR2
and its phosphorylation triggers a downstream signalling cascade, which ultimately results in
increased endothelial cell proliferation, migration and vascular permeability [22]. Thus,
quercetin, luteolin and their derivatives that showed significant inhibition of the VEGF
stimulated cell migration in the in vitro scratch assay were investigated for their effects on
VEGFR2 autophosphorylation at Tyr1175 site in HUVEC’s.
In general, the results obtained from the Western blotting analysis (Figure-4) mirror the results
obtained with the scratch assay. VEGF (10 ng/mL) [+ve control] significantly promoted the
levels of phosphorylated VEGFR2 (P-VEGFR2) in comparison to the –ve control without
VEGF stimulation. In the case of quercetin and its derivatives, quercetin (1) at both 10 µM and
1 µM, its methoxy derivative (2) and thiomethoxy derivative (3) at 10 µM all showed significant
reduction in the levels of P-VEGFR2, whereas, such reduction in the levels of P-VEGFR2 was
not displayed by the 4-thio derivative of quercetin (4) at 10 µM concentration. As mentioned
earlier, the order of inhibitory effect correlated well with the scratch assay results, i.e.
compound-2≥compound-1>compound-3>compound-4.
In the case of luteolin and its
derivatives only luteolin at 1 µM concentration displayed significant reduction in the P-
VEGFR2 levels while its methoxy and 4-thiomethoxy derivatives (9 and 11) did not show
significant inhibition on the VEGF-induced autophosphorylation of VEGFR2. These results
suggest that compounds-1, 2, 3 and 10 mediates their antiangiogenic effects via the suppression
of VEGFR2 phosphorylation.
2.3 Molecular docking
Molecules that interfere with the VEGF/VEGFR signalling pathway typically work by either
blocking the VEGF ligand directly or by inhibiting its receptor VEGFR2. Hence, molecular
docking studies were conducted to elucidate the potential molecular interactions between
7

ACCEPTED MANUSCRIPT
quercetin and luteolin derivatives and both VEGFR2 and VEGF. The derivatised molecules
were docked sequentially into the respective binding sites, using the X-ray crystallographic
structures of VEGFR2 receptor (PBD code-1YWN) and VEGF (PBD code-1VPP). The docking
procedure was first validated by re-docking the extracted co-crystallised ligand of VEGFR2 (4-
amino-furo[2,3-d]pyrimidine) into the prepared protein to be used for docking. The RMSD
between the docked conformation generated by the program and the native co-crystallized
ligand conformation was found to be 0.10 Å, well within the 2 Å grid spacing used for docking,
indicating that the docking procedure to be used was reliable and valid. Furthermore, the
interactions between the docked ligand and the receptor mimicked those observed in the crystal
structure of the same.
Interactions with VEGFR2: The results of ligand-directed docking into the ATP-binding site of
VEGFR2 are summarised in Table-S1 (see supporting information) in decreasing order of the
ligand binding affinities, as determined by their docking scores. The binding modes of
quercetin, luteolin and their derivatives with VEGFR2 are depicted in Figure-5. The quercetin
and luteolin derivatives have shown potential interactions with the key residues, Glu-883, Lys-
866 and Cys-917 of VEGFR2 through hydrogen bonding. The docking scores suggest that
compounds 2, 10 and 3 have greater binding affinity towards VEGFR2.
Interactions with VEGF: The results of automated docking procedure with VEGF are
summarised in Table-S2 (see supporting information) in decreasing order of the ligand binding
affinities, as determined by their docking scores. The binding modes of quercetin, luteolin and
their derivatives with VEGF are depicted in Figure-6. The quercetin and luteolin derivatives
have also shown potential interactions with the key residues, Leu-32V, Gly-59V and Asp-34W
residues of VEGF through hydrogen bonding. Compounds 9, 2 and 11 showed greater binding
affinity towards VEGF.
8

ACCEPTED MANUSCRIPT
Among these compounds, the methoxy derivative of quercetin (compound-2) has shown higher
binding affinities towards both VEGFR2 and VEGF. This compound was found binding to the
ATP binding site of VEGFR2 through the formation of hydrogen bonds to the ‘hinge’ region
residues Cys-917-N (2.98 Å), Cys-917-O (2.92 Å), Glu-883-O (2.61 Å) and Lys-866-N (2.94 Å)
residue of the hydrophobic pocket, and also through hydrophobic interactions (Figure-5). The
major interactions of compound-2 with VEGF are Leu-32V-N (2.98 Å), Gly-59V-N (2.79, 2.91
Å), and Asp-34W-N (3.13 Å) (Figure-6). In general, the analysis of molecular docking results
revealed that compound-2 possibly acts as an inhibitor of VEGFR2 and its binding potency
relies on strong hydrophobic interactions between the compound and the binding site of the
receptor. Also, the docking studies suggested that the higher binding affinities of compound-2
towards both VEGFR2 and VEGF could possibly be the reason for its antiangiogenic activity as
observed in both the scratch assay and Western blotting. Similarly, the weaker antiangiogenic
activities of the 4-thio derivatives of quercetin and luteolin could be because of their poor
binding affinities towards VEGFR2 and VEGF. This could be further verified in future studies
by utilising a cell free protein kinase assay of VEGFR2.
2.4 Membrane perturbation (probed by DSC)
A recent paper by Ingólfsson et al. has provided new insight into how the biological effects of
flavonoids might be explained [23]. The authors suggest that many of the biological activities
triggered by flavonoids (more specifically, activation of certain membrane proteins) might, in
fact, be the result of flavonoid-induced perturbation of biological membranes, rather than a
direct ligand-receptor effect. The authors do not exclude the possibility of direct binding, but
recommend ruling out bilayer effects first. Therefore, the 6 compounds (1-3 and 9-11), which
showed VEGF/VEGFR mediated antiangiogenic activity (in the scratch assay and Western
blotting) in our study, were further investigated for a possible lipid interaction (bilayer effect)
with a dipalmitoylphosphatidylcholine bilayer [DPPC (lipid)] model using differential scanning
9

ACCEPTED MANUSCRIPT
calorimetry (DSC). In order to validate our system, genistein, which has been reported to perturb
biological membranes by Ingolfsson et al. and has a structural similarity with our panel of
compounds, was used as a positive control.
Dipalmitoylphosphatidylcholine (DPPC) alone (i.e. in the absence of genistein or any
compounds) showed a pre transition (Tp) centered at 36.25 °C and a main transition (Tm) at
41.31 °C. These data for the pure lipid system are in good agreement with previous reports
[24,25] (Figure-7). Addition of genistein to the DPPC system showed perturbation of this lipid
bilayer, as evidenced by a significant lowering of the Tm (from 41.23 °C to 40.89 °C (at 60 ꢀM)
and to 40.51 °C (at 100 ꢀM)), as reported in Table-1 and Figure-6. No changes were observed
with the ꢁH of the main transition temperature (ꢁHm) peak, but the Tp peak broadened at 10 ꢀM
and became undetectable at 60 ꢀM and 100 ꢀM. The lowest concentration of genistein at which
a membrane effect was observed (60 ꢀM ) was in line with the concentrations at which
genistein’s bilayer effect had been reported by Ingolfsson et al [23], which indicated that the
experimental system was fit for purpose and could be used to assess the potential bilayer effect
of our compounds.
Quercetin and its derivatives: Unlike genistein, quercetin and its derivatives did not affect the
Tm at the concentration for which antiangiogenic activity was reported (10 ꢀM) (see supporting
information) and, indeed, no significant effect was evident at 10 times that concentration (100
ꢀM) (see Table-1 and Figure-6). No change was observed for ꢁHm but the Tp peak was affected
(Tp was lowered along with the peak broadening). Others have suggested that the Tp is highly
sensitive to the presence of other molecules in the phospholipids and that changes in the pre-
transition of DSC profile cannot be attributed to any specific molecular changes [26]. Several
studies have been reported on the interactions of quercetin with bilayer lipid membranes [27,28],
mainly the study by Sinha et al. in which the quercetin was reported to affect biological
10

ACCEPTED MANUSCRIPT
membranes was primarily observed at concentrations 250 fold higher than those for which we
are reporting activity[29].
Luteolin and its derivatives: When luteolin (10), methoxy luteolin (9) and thiomethoxy luteolin
(11) were analysed, no changes in the DSC profile of DPPC was observed at the biologically
active concentrations nor at 10 times that concentration.
Taken together, these results indicate that the antiangiogenic activity observed for the
compounds studied herein is likely to be mediated via an interference with the VEGF/VEGFR
signaling pathway, as evidenced by the Western blotting and computational studies, rather than
being the result of membrane perturbation. This statement is certainly true for luteolin and its
derivatives, for which no interactions were observed. The subtle membrane effect observed for
quercetin and its derivatives does not appear to be sufficiently significant to explain the
biological effects and dismiss a specific interaction with the VEGF/VEGFR system.
2.5 Antiproliferative activity
The flavonoids are known to exhibit a direct antiproliferative effect on various tumour cells [6]
including hormone dependent cancers such as MCF-7 breast cancer cell line. Therefore, we
considered MCF-7 as an appropriate model to assess the antiproliferative activity. Hence, in this
study, the direct inhibitory effects of quercetin and luteolin derivatives were determined against
the breast cancer cell line MCF-7 (wild type) and its doxorubicin-resistant cell line MCF-7/DX
to compare their concentrations that are required to elicit the direct anticancer activity and the
antiangiogenic activity. Cell viabilities were determined in the presence and the absence of
quercetin, luteolin and their derivatives by MTT assay [30] after 72 h incubation. The
cytotoxicity profiles and the IC50 values of these compounds are shown in Figure-8 and Table-2
respectively.
11

ACCEPTED MANUSCRIPT
Among quercetin and its derivatives, only compound-1 was found to be antiproliferative with
IC50 values of 13.7 ± 0.61 µM and 80.4 ± 1.79 µM against MCF-7 cell line and MCF-7/DX cell
line respectively. The methoxy and the 4-thio derivative, compounds-2 and 4 respectively
displayed their antiproliferative activity at concentrations greater than 100 µM, whereas the 4-
thiomethoxy derivative, compound-3 did not demonstrate any cytotoxicity below 250 µM
concentration. Thus, the antiproliferative activities of quercetin and its derivatives are in the
order of compound-1>compound-4>compound-2>compound-3. Among luteolin and its
derivatives, compound-10 and 12 were found to be active with IC50 values of 21.6 ± 0.81 µM
and 27.3 ± 1.55 µM respectively against MCF-7 cell line and 81.1 ± 1.79 µM and 75.9 ± 3.08
µM respectively against MCF-7/DX cell line. However, compounds 9 and 11 did not
demonstrate any antiproliferative activity below 250 µM concentration. Therefore, the
antiproliferative activities of luteolin and its derivatives are in the order of compound-
10≥compound-12>compound-9≡compound-11. In general, the antiproliferative activities of
these compounds suggest that the methoxy and the thiomethoxy derivatives are 10-15 fold less
active than their parent compounds.
2.6 Considerations of the structure-activity relationships (SARs)
Comparison of the antiangiogenic activities and antiproliferative activities of quercetin, luteolin
and their derivatives has led to identification of the key structural features that are specific for an
enhanced and a selective antiangiogenic activity. Comparison of the antiangiogenic activities of
underivatised quercetin (1, bearing 5 hydroxyls, 4-C=O) with underivatised luteolin (10, bearing
4 hydroxyls, 4-C=O) indicates that luteolin without the C-3 hydroxyl has a higher
antiangiogenic potential (86% versus 54% inhibition of VEGF stimulated migration at 1 µM
concentration). This observation is consistent with the Western blotting analysis, which showed
increased inhibition of autophosphorylation of VEGFR2 in presence of luteolin (1 µM).
Docking studies also suggested greater binding affinity of luteolin (docking score-6.87) with
12

ACCEPTED MANUSCRIPT
VEGFR2 than quercetin (docking score-6.56). These observations suggest that the C-3 hydroxyl
is not important for the antiangiogenic activity. However, for their methoxy derivatives (-OMe
and 4-C=O), it was observed that the methoxy derivative of quercetin containing a C-3 methoxy
group has a greater potency to inhibit VEGF stimulated migration (86% at 10 µM) through the
inhibition of VEGF mediated autophosphorylation at Tyr-1175 site than the methoxy analogue
of luteolin (9) as well as the hydroxy analogues quercetin and luteolin. Docking studies also
indicated the highest binding affinity of compound-2 with both VEGFR2 and VEGF, and also it
was observed that the enhanced antiangiogenic activity of compound-2 is due to the enhanced
hydrophobic interactions formed by of C-3 -OMe group with the VEGFR2. This demonstrates
that the presence of a C-3 -OMe group is beneficial for the antiangiogenic activity.
Similarly, comparison of the antiangiogenic activities of compounds containing the 4-C=O
functionality (1, 2, 9 and 10) with the derivatives containing the 4-C=S functionality (3, 4, 11
and 12) indicated that replacement of the 4-C=O group with a 4-C=S resulted in a weaker in
vitro inhibition of VEGF-stimulated migration, especially for 4-C=S compounds with the free
hydroxyl groups (4 and 12). Hence, conversion of the 4-C=O to 4-C=S group is detrimental to
the antiangiogenic activities of both quercetin and luteolin. This observation is further supported
by the Western blotting studies that illustrate very poor inhibition of autophosphorylation of
VEGFR2, and also by the molecular docking studies that show weaker interactions with
VEGFR2 and VEGF, by these 4-C=S derivatised compounds. Collectively, these results suggest
that the 4-C=O moiety is crucial for antiangiogenic activity.
From the comparison of the antiproliferative and the antiangiogenic activities of quercetin,
luteolin and their derivatives, it is interesting to note that the derivatives with higher
antiangiogenic properties were generally less cytotoxic than those with lower antiangiogenic
properties. This would be an attractive feature of these derivatives for use in future clinical
13

ACCEPTED MANUSCRIPT
development as they could potentially be used for their anti-angiogenic properties at
concentrations significantly lower than those required for a direct anticancer activity. Thus
comparison of the antiproliferative activities of quercetin and luteolin with their methoxy, 4-
thiomethoxy and 4-thiohydroxy derivatives illustrates that methylation of the free hydroxyls and
replacement of the 4-C=O group with a 4-C=S group decreases the antiproliferative activity. In
the case of quercetin, methylation of the hydroxyl group resulted in a 10-18 fold decrease in the
antiproliferative potency. A similar trend was observed in the case of luteolin where an
approximately 11 fold decline in antiproliferative activity was observed for its methoxy
derivative. Replacement of the 4-C=O group with a 4-C=S group in the case of quercetin has
also shown a detrimental impact on the antiproliferative activity with the 4-thio derivative
showing about 8 fold less reactivity than quercetin. However, in the case of luteolin, the
conversion of the 4-C=O group to the 4-C=S group has not impacted on the antiproliferative
property with both luteolin and its 4-thio derivative exhibiting similar cytotoxicity profiles. This
suggests that methylation of the free hydroxyl groups in flavonoids allows selective
antiangiogenic activity. Figure-9 presents the SAR observed for the quercetin and luteolin
derivatives.
Overall, methoxyquercetin (2) and luteolin (10) possess higher potency to prevent VEGF driven
HUVEC migration at a 10 and 20 fold lower concentration respectively than the concentration
that is required to directly inhibit the in vitro proliferation of tumour cells. This could be
attributed to the high degree of selectivity of this compound for the inhibition of VEGFR2
versus a range of other kinases [31].
As the clinical outcomes of the current antiangiogenic therapy and preclinical studies have
suggested the need for co-administration of the inhibitors with chemotherapeutics, for continual
constraint of angiogenesis [32–34], molecules with an appropriate balance of high potency
14

ACCEPTED MANUSCRIPT
against VEGFR2 and minimal toxicity are preferred. Therefore, with the view of developing
antiangiogenic agents for cancer therapeutics and combination therapy, our present work
suggests compound-2 and compound-10 as potential leads for further optimisation and
development.
3 CONCLUSION
In conclusion, we have synthesised the methoxy and 4-thio derivatives of quercetin and luteolin
and evaluated their resulting antiangiogenic and antiproliferative potencies, alongside quercetin
and luteolin. As part of this programme we have demonstrated the impact of subtle structural
modifications such as methylation and 4-thio conversion on the antiangiogenic and
antiproliferative properties of quercetin and luteolin. Comparison of their antiangiogenic
activities has led to identification of key structural features required for the antiangiogenic
activities. It was found that the 4-C=O functionality is crucial for the antiangiogenic activity and
also methylation of the free hydroxyl groups provides the benefit of potential antiangiogenic
activity with low cytotoxicity. Also, analysis of bilayer membrane interactions of the active
compounds suggests that bilayer effects do not play a significant role in the biological
mechanism of action at the concentrations tested, especially for luteolin and its derivatives.
Among quercetin, luteolin and their derivatives, compound-2 and compound-10 were found to
potentially inhibit the in vitro VEGF induced cell migration with low cytotoxicity. In addition,
the molecular docking studies indicated that compound-2 and compound-10 possess high
binding affinities towards VEGF and VEGFR2. Overall, compound-2 and compound-10 exhibit
an appropriate balance of high antiangiogenic activity and minimal toxicity. Therefore, they
have been identified as the most promising flavonoid templates from our programme of study
for further development as antiangiogenic agents.
15

ACCEPTED MANUSCRIPT
4 EXPERIMENTAL SECTION
4.1 Materials and Methods
Chemicals and Reagents: All chemicals were purchased from Sigma-Aldrich unless specified.
2-hydroxy-4,6-dimethoxyacetophenone was purchased from Alfa Aesar, UK and quercetin was
purchased from Sigma-Aldrich, UK. Solvents were purchased from Fisher Scientific, UK and
1
Sigma Aldrich, UK. H-NMR spectra were recorded in either deuterated chloroform (CDCl3),
deutereated dimethyl sulfoxide (DMSO-d6) using a Bruker DPX 400 (400 MHz) spectrometer.
Mass spectrometry data were recorded on a Thermo Fisher LTQ Orbitrap XL instrument. A
HPLC (Agilent-1100 series) equipped with Column: C-18 (ACE-221-2546; 4.6 mm X 250 mm;
5 µm); 1.0 mL/min flow rate; 20 min gradient of 1% acetic acid in acetonitrile [organic mobile
phase (A)] and 1% acetic acid in water [aqueous mobile phase (B)] from 5% A at 0 min, 20% A
at 2min, 40% A at 6 min, 60% A at 8 min, 95% A at 10 min, 5% A at 20 min, at 258 nm was
used to determine the purity of all the synthesised compounds.
Cell line and Culture: HUVECs were purchased from ECACC and cultured in EGM-2 medium.
MCF-7 (ER +ve breast cancer cell line, wild type) and MCF-7/DX (ER +ve breast cancer cell
line, doxorubicin resistant type) were provided by Tenovus centre for Cancer research (Cardiff,
UK). Both MCF-7 and MCF-7/DX cells were cultured in RPMI 1640 supplemented with 5%
fetal bovine serum. All the culture media and other reagents were obtained from Lonza, UK.
Bovine serum albumin (BSA) was purchased from Sigma Aldrich, UK. Recombinant human
VEGFR-A165 was purchased from Peprotech, UK. The primary antibodies directed against
phosphorylated tyrosine-1175 site in the VEGFR2 (KDR) was purchased from Cell signalling,
UK. The primary antibody directed against total VEGFR2 was purchased from Santa Cruz
Biotechnology, UK and the primary antibody directed against actin was purchased from Abcam,
UK. Horseradish peroxidase-conjugated secondary antibody was purchased from Abcam, UK.
Enhanced chemiluminescences (ECL) detection solutions were purchased from Bio-Rad, UK.
16

ACCEPTED MANUSCRIPT
PVDF membrane was purchased from Bio-Rad, UK. Reagents for Phosphate buffered saline
(PBS) for cell culture were purchased from Sigma Aldrich, UK. PBS (pH 7.4) was freshly
prepared in lab and solution pH was checked before used. Scratch assay images were captured
using 1.3M microscope digital eyepiece camera. ImageJ software was used to quantify the cell
migration.
The stock solutions of the test compounds (10 mg/mL) were prepared in sterile DMSO and
ethanol (1:1 v/v). These stocks were then appropriately diluted with the complete culture
medium and the ethanol and DMSO levels were maintained below 1 % in the test
concentrations.
Statistics analysis: Statistical analysis was carried out against the control groupby one-way
ANOVA followed by Bonferroni’s post hoc test using GraphpadPrism 6. Statistical significance
value was set at p<0.05.
17

ACCEPTED MANUSCRIPT
4.1.1 Synthesis of methoxy and 4-thio methoxy quercetin (2-4)
4.1.1.1 Synthesis of 2-(3,4-dimethoxyphenyl)-3,5,7-trimethoxy-4H-chromen-4-one (2)
Quercetin (1) (1g, 3.30 mmol) was dissolved in KOH (15%, 10 mL) at ambient temperature.
Dimethyl sulfate (2.29 g, 1.7 mL, 18.15 mmol, 5.5 eq., d=1.33 g/mL) was added slowly over 10
min and the reaction was stirred at ambient temperature for 1 h. After 1 h, further dimethyl
sulfate (5.5 eq.) and KOH (15%, 10 mL) were added and the mixture was heated to 90 oC. After
3 h, further dimethyl sulfate (5.5 eq.) and KOH (15%, 10 mL) were added and the reaction was
o
stirred overnight at 90 C. KOH (15%, 15 mL) was then added and the reaction was stirred for
another 5 h at 90 oC. The reaction was cooled to room temperature, and acidified to pH 5 by the
addition of2 M HCl (20 mL). The precipitate was filtered and washed with water. The 2-(3,4-
dimethoxyphenyl)-3,5,7-trimethoxy-4H-chromen-4-one (2) was obtained as an off white solid in
a pure form after purification by column chromatography [Solvent system: EtOAc (100%)].
o
o
Yield: 60%; m.p: 155-7 C (lit [35]-150-52 C);1H NMR: (DMSO-d6,400 MHz) δ 3.88, 3.90,
3.96 (15H, s, -OCH3), 6.34 (1H, s, H-6), 6.50 (1H, s, H-8), 6.97 (1H, d, J= 8.0 Hz, H-5’), 7.69
13
(1H, d, J= 8.0 Hz, H- 6’), 7.71 (1H, s, H-2’); C NMR: (DMSO-d6, 100 MHz) δ 56.3 (4 x-
OCH3), 59.9 (-OCH3), 92.3 (C8), 95.8 (C6), 109.4 (C10), 110.7 (C2’), 111.5 (C5’), 121.6 (C6’),
123.7 (C1’), 141.3 (C3), 148.8 (C4’), 150.8 (C3’), 152.6 (C2), 158.9 (C9), 161.3 (C5), 164.1
(C7), 174.0 (C=O); IR νmax[cm-1]: 1651(C=O, ν, s), 1291 (-OCH3, ν, s), 1097 (C-O, ν, m); m/z
(FTMS+ESI): M+H (C20H21O7) requires 373.1282, found 373.1277. HPLC purity: 96.1%
4.1.1.2 Synthesis of 2-(3,4-dimethoxyphenyl)-3,5,7-trimethoxy-4H-chromene-4-thione (3)
2-(3,4-Dimethoxyphenyl)-3,5,7-trimethoxy-4H-chromen-4-one (2) (0.20 g, 0.53 mmol) was
dissolved in anhydrous toluene (1 mL), and Lawesson’s reagent (0.21, 0.53 mmol, 1 eq.) was
o
added and the reaction was heated to 110 C for 4 h. The reaction mixture was cooled to room
temperature and concentrated in vacuo. The residue was purified by column chromatography
18

ACCEPTED MANUSCRIPT
[Solvent system: CHCl3 (100 %)] to obtain the pure 2-(3,4-Dimethoxyphenyl)-3,5,7-trimethoxy-
4H-chromene-4-thione (3) as a brown solid.
o
o
1
Yield: 65%; m.p: 168-70 C (lit [36]-170 C); H NMR: (CDCl3, 400 MHz) δ 3.72 (3H, s,
OCH3), 3.95, 3.93, 3.07, 4.03 (12H, s, 4 x -OCH3), 6.42 (1H, s, H-6), 6.54 (1H, s, H-8), 7.00
(1H, d, J= 8.5 Hz, H-5’), 7.78 (1H, d, J= 8.5 Hz, H-6’), 7.86 (1H, s, H-2’); 13C NMR: (CDCl3,
100 MHz) δ 56.0 (-OCH3), 58.7 (-OCH3), 92.3 (C8), 96.7 (C6), 110.6 (C2’), 111.6 (C5’), 116.8
(C10), 122.0 (C6’), 123.2 (C1’), 146.3 (C3), 148.6 (C4’), 149.6 (C3’), 151.2 (C2), 154.4 (C9),
161.6 (C7), 163.5 (C5), 194.0 (C=S); IR νmax[cm-1]: 1351 (C=S, ν, m), 1220 (-OCH3, ν, m)
1143 (C-O, ν, m); m/z (FTMS+ESI): M+H (C20H21O6S) requires 389.1053, found 389.1055.
HPLC purity: 96.1%
4.1.1.3 Synthesis of 2-(3,4-dihydroxyphenyl)-3,5,7-trihydroxy-4H-chromene-4-thione (4)
2-(3,4-Dimethoxyphenyl)-3,5,7-trimethoxy-4H-chromene-4-thione (3) (0.10 g, 0.257 mmol)
was dissolved in anhydrous DCM (1 mL). BBr3 (0.35 g, 0.135 mL, 1.413 mmol, 5.5 eq., d=2.60
o
g/mL) was added and the reaction was stirred at 40 C overnight. The reaction was cooled to
room temperature and diluted with water (2 mL). The pH was adjusted to 6 with 5% Na2HPO4
and the precipitated product was filtered to afford the crude product. The pure product, 2-(3,4-
dimethoxyphenyl)-3,5,7-trihydroxy-4H-chromene-4-thione (4), was obtained as a bright red
solid after purification by column chromatography [Solvent system: EtOAc (100 %)].
o
1
Yield: 55%; m.p: 268-70 C (decomposed); H NMR: (CDCl3, 400 MHz) 6.34 (1H, s, H-6),
6.54 (1H, s, H-8), 6.92 (1H, d, J = 8.5 Hz, H-5’), 7.71 (1H, d, J = 8.5 Hz, H-6’), 7.76 (1H, s, H-
2’), 8.66 (1H, s, OH), 9.44 (1H, s, OH), 9.91 (1H, s, OH), 11.07 (1H, s, OH), 13.06 (1H, s, OH);
13C NMR: (CDCl3, 100 MHz) δ 93.8 (C8), 100.4 (C6), 111.3 (C10), 115.7 (5’), 115.8 (C2’),
120.7(C6’), 121.6 (C1’), 142.4 (C3’), 142.7 (C4’), 145.3 (C9), 149.1 (C7), 152.5 (C2), 160.1
(C5), 163.5 (C3), 181.6 (C=S); IR νmax[cm-1]: 1267 (C=S, ν, m), 1147(C-O, ν, m), 3084 (OH,
19

ACCEPTED MANUSCRIPT
w, b); m/z (FTMS+ESI): M+H (C15H11O6S) requires 319.0271, found 319.0266. HPLC purity:
99.4%
4.1.2 Synthesis of luteolin, methoxy luteolin and 4-thiomethoxy luteolin (9-12)
4.1.2.1 2-Acetyl-3, 5-dimethoxyphenyl 3, 4-dimethoxy benzoate (7)
1-(2-Hydroxy-4, 6-dimethoxyphenyl) ethanone (5) (1 g, 5.09 mmol) was dissolved in pyridine
(20 mL) and heated to 50 oC. DBU (1 % v/v of pyridine) was added and the mixture was stirred
at 50 oC for 30 min. 3,4-Dimethoxy benzoyl chloride (6) (1.53 g, 7.63 mmol, 1.5 eq.) was added
slowly over 15 min, the mixture was heated to 75 oC and stirred for 1 h. The reaction was cooled
to room temperature, acidified to pH 5 with 2 M HCl and extracted with EtOAc (2 x 20 mL).
The organic layers were combined, dried over anhydrous magnesium sulfate and concentrated in
vacuo to yield the desiredproduct. The pure compound was obtained as a white solid after
purification by column chromatography [Hexane: CHCl3: EtOAc (3:6:1 v/v/v)].
o
1
Yield: 60%; m.p: 151-53 C; H NMR: (CDCl3,,400 MHz) δ 2.47 (3H, s, -CH3) 3.82 (3H, s, -
OCH3), 3.84 (3H, s, -OCH3), 3.86 (3H, s, -OCH3), 3.96 (3H, s, -OCH3), 6.37 (1H, s, H-4), 6.40
13
(1H, s, H-6), 6.91 (1H, d, J= 8 Hz, H-6’), 7.61 (1H, s, H-2’), 7.80 (1H, d, J= 7 Hz, H-5’); C
NMR: (CDCl3, 100 MHz) δ 29.4 (CH3), 56.0 (-OCH3), 60.3 (-OCH3), 109.7 (C4), 111.1 (C5’),
112.1 (C2’), 120.6 (C6), 124.3 (C5, C6’), 126.1 (C1’), 140.9 (C2), 143.5 (C3’), 153.5 (C3),
156.7 (C4’), 163.7 (COO-Ar), 195.5 (COCH3); IR νmax [cm-1]: 1713 (C=O, ν, m), 1102 (O-
C=O, ν, m), 1689 (C=O, ν, m), 1246 (-OCH3, ν, s); m/z (FTMS+ESI): M+Na (C19H20O7Na)
requires 383.1101, found 383.1102.
4.1.2.2 1-(2-Hydroxy-4,6-dimethoxyphenyl)-3-phenyl propane-1, 3-dione (8)
The 2-acetyl-3, 5-dimethoxyphenyl 3, 4-dimethoxy benzoate (7) (1.0 g, 2.7 mmol) was
o
dissolved in pyridine (15 mL) and heated to 50 C for 1 h. Anhydrous potassium hydroxide
[powdered] (0.32g, 5.4 mmol, 2 eq.) was added and the reaction was heated to 75 oC . After 1 h,
20

ACCEPTED MANUSCRIPT
the reaction mixture was cooled to room temperature, acidified to pH 5 with 2M HCl and
extracted with EtOAc (2 x 5 mL). The organic layers were combined, dried over anhydrous
magnesium sulfate and concentrated in vacuo to yield the crude 1,3-diketone. The crude 1,3-
diketone was treated with glacial acetic acid to remove the pyridine and the yellow compound
was used for the next step without purification.
1
Yield: 75%; H NMR: (CDCl3, 400 MHz) [The compound exists as a mixture of keto-enol
tautomers] δ 3.78 (2H, s, -CH2), 3.89-3.98 (12H, 4 x -OCH3), 5.98 (1 H, s, H-5), 5.99 (1 H, s,
H-3), 6.02 (1H, s, =CH of enol form), 6.93 (1H, d, J= 8.5 Hz, H-6’), 7.27 (1H, d, J= 7.0 Hz, H-
5’), 7.44 (1H, s, H-2’), 13.44 (1H, s, OH), 13.72 (1H, s, OH of enol form).
4.1.2.3 2-(3,4-Dimethoxyphenyl)-5,7-dimethoxy-4H-chomen-4-one (9)
1-(2-Hydroxy-4,6-dimethoxyphenyl)-3-phenyl propane-1, 3-dione (8) (0.75g, 2.08 mmol) was
o
dissolved in acetic acid (6 mL) and heated to 90 C for 1 h. Concentrated sulfuric acid (0.06
o
mL, 1% v/v of acetic acid) was added and the reaction was heated at 110 C for 1 h. The
reaction mixture was cooled to room temperature, diluted with water (10 mL) and the aqueous
layer was extracted with EtOAc (2 x 5 mL). The organic layers were combined, dried over
anhydrous magnesium sulfate and concentrated in vacuo to obtain the crude 2-(3,4-
dimethoxyphenyl)-5,7-dimethoxy-4H-chomen-4-one. The pure compound was obtained as a
white solid after purification by column chromatography [Hexane: CHCl3: EtOAc (3:6:1 v/v/v)].
Yield: 70%; m.p: 192-93 oC (lit [37]-190-94 oC); 1H NMR: (DMSO-d6, 400 MHz) δ 3.82 (3H,
s, -OCH3) 3.83 (3H, s, -OCH3), 3.87 (3H, s, -OCH3), 3.89 (3H, s, -OCH3), 6.49 (1H, s, H-3),
6.76 (1H, s, H-6), 6.86 (1H, s, H-8), 7.01 (1H, d, J= 8.5 Hz, H-6’), 7.52 (1H, s, H-2’), 7.63 (1H,
13
d, J= 8.0 Hz, H-5’); C NMR: (DMSO-d6, 100 MHz) δ 55.9 (-OCH3), 93.4 (C8), 96.3 (C6),
107.0 (C3’), 107.9 (C10), 108.7 (C6’), 111.5 (C3’), 119.2 (C2’), 122.9 (C1’), 149.1 (C4’), 151.5
(C5’), 159.0 (C9), 159.3 (C5), 160.4 (C2), 163.9 (C7), 175.7 (C=O); IR νmax [cm-1]: 1644
21

ACCEPTED MANUSCRIPT
(C=O, ν, m), 1252 (-OCH3, ν, m) 1138 (C-O, ν, m); m/z (FTMS+ESI):M+H (C19H19O6)
requires 343.1176, found 343.1176. HPLC purity: 99.5%
4.1.2.4 2-(3,4-Dihydroxyphenyl)-5,7-dihydroxy-4H-chomen-4-one (10)
2-(3,4-Dimethoxyphenyl)-5,7-dimethoxy-4H-chomen-4-one (9) (100 mg, 0.292 mmol) was
dissolved in anhydrous DCM (1 mL). BBr3 (329.32 mg, 0.126 mL, 1.314 mmol, 4.5 eq., d=2.60
o
g/mL) was added and stirred at 40 C overnight. The reaction was cooled to room temperature
and diluted with water (2 mL). The pH was adjusted to 6 with 5% Na2HPO4 and the aqueous
layer was extracted with n-butanol (2 x 5 mL). The organic layer was washed with brine, dried
over anhydrous magnesium sulfate and concentrated in vacuo. The crude product was purified
by column chromatography [Solvent system: EtOAc (100 %)] to afford 2-(3, 4-
dihydroxyphenyl)-5,7-dihydroxy-4H-chomen-4-one (10) as a pale yellow solid.
o
o
1
Yield: 59%; m.p:328-29 C (lit [38]-326-28 C); H NMR: (DMSO-d6, 400 MHz) δ 6.25 (1H,
s, H-3), 6.50 (1H, s, H-6), 6.72 (1H, s, H-8), 6.95 (1H, d, J= 8.0 Hz, H-6’), 7.45 (1H, s, H-2’),
7.46 (1H, d, J= 8.0 Hz, H-5’), 9.48 (1H, s, OH), 9.98 (1H, s, OH), 10.90 (1H, s, OH), 13.02 (1H,
13
s, OH) ; C NMR: (DMSO-d6, 100 MHz) δ 93.7 (C8), 98.8 (C6), 102.6 (C3), 103.6 (C10),
113.2 (C2’), 115.9 (C5’), 118.9 (C6’), 121.4 (C1’), 145.6 (C3’), 149.6 (C4’), 157.2 (C9), 161.4
(C5), 163.8 (C2), (164.0, C7), 181.6 (C=O); IR νmax [cm-1]: 1600 (C=O, ν, s), 1160 (C-O, ν, m),
3236 (OH, w, b); IR νmax [cm-1]: 1600 (C=O, ν, s), 1160 (C-O, ν, m), 3236 (OH, w, b); m/z
(FTMS+ESI): M+H (C15H11O4) requires 287.0550, found 287.0554. HPLC purity: 99.4%
4.1.2.5 Synthesis of 2-(3,4-dimethoxyphenyl)-5,7-dimethoxy-4H-chromene-4-thione (11)
2-(3,4-Dimethoxyphenyl)-5,7-dimethoxy-4H-chomen-4-one (9) (0.25 g, 0.73 mmol) was
dissolved in anhydrous toluene (1 mL), Lawesson’s reagent (0.30 g, 0.73 mmol, 1 eq.) was
o
added and the reaction was heated to 110 C for 4 h. The reaction mixture was cooled to room
temperature and concentrated in vacuo. The residue was purified by column chromatography
22

ACCEPTED MANUSCRIPT
[Solvent system: Hexane: CHCl3: EtOAc (3:6:1 v/v/v)] to obtain the pure compound as a green
solid.
o
o
Yield: 61%; m.p: 204-5 C (lit [36]-203 C); 1H NMR: (DMSO-d6, 400 MHz) δ 3.82 (3H, s, -
OCH3) 3.86 (3H, s, -OCH3), 3.90 (3H, s, -OCH3), 3.94 (3H, s, -OCH3), 6.58 (1H, s, H-2’), 6.94
(1H, s, H-8), 7.13 (1H, d, J= 8.0 Hz, H-5’), 7.53 (1H, s, H-3), 7.55 (1H, s, H-6), 7.69 (1H, d, J=
8.0 Hz, H-6’); 13C NMR: (DMSO-d6, 100 MHz) δ 53.6 (-OCH3), 53.7 (-OCH3), 53.8 (-OCH3),
54.0 (-OCH3), 91.2 (C8), 94.8 (C6), 107.0 (C2’), 109.8 (C5’), 114.3 (C10), 117.8 (C3), 118.1
(C6’), 120.2 (C1’), 147.0 (C4’), 147.2 (C3’), 149.7 (C2), 152.9 (C9), 158.7 (C7), 161.4 (C5),
196.7 (C=S); IR νmax[cm-1]: 1316 (C=S, ν, m), 1254 (-OCH3, ν, m) 1140 (C-O, ν, m); m/z
(FTMS+ESI): M+H (C19H19O5S) requires 359.0948, found 359.0948. HPLC purity: 95.7%
4.1.2.6 Synthesis of 2-(3, 4-dihydroxyphenyl)-5, 7-dihydroxy-4H-chromene-4-thione (12)
2-(3,4-Dimethoxyphenyl)-5,7-dimethoxy-4H-chromene-4-thione (11) (0.15 g, 0.418 mmol) was
dissolved in anhydrous DCM (1 mL). BBr3 (0.47 g, 0.181 mL, 1.86 mmol, 4.45 eq., d=2.60
o
g/mL) was added and the reaction was stirred at 40 C overnight. The reaction was cooled to
room temperature and diluted with water (2 mL). The pH was adjusted to 6 with 5% Na2HPO4
and the precipitated product was filtered and purified by column chromatography [Solvent
system: EtOAc (100 %)] to obtain the pure compound as a brown solid.
o
1
Yield: 55%; m.p: 269-70 C; H NMR: (DMSO-d6, 400 MHz) δ 6.29 (1H, s, H-2’), 6.51 (1H,
s, H-8), 6.89 (1H, d, J= 8.0 Hz, H-5’), 7.31 (1H, s, H-3), 7.43 (1H, s, H-6), 7.49 (1H, d, J= 8.0
Hz, H-6’), 9.80 (1H, s, OH), 10.12 (1H, s, OH), 11.20 (1H, s, OH), 13.67 (1H, s, OH); 13C
NMR: (DMSO-d6, 100 MHz) δ 94.4 (C8), 100.5 (C6), 111.6 (C10), 113.5 (C2’), 115.7 (C5’),
116.2 (C3), 119.7 (C1’), 120.1 (C6’), 145.9 (C3’), 150.3 (C4’), 153.9 (C2), 155.1 (C9), 161.7
(C7), 164.0 (C5), 194.5 (C=S); IR νmax [cm-1]: 1444 (C=S, ν, m), 1146 (C-O, ν, m), 3293 (OH,
23

ACCEPTED MANUSCRIPT
w, b); m/z (FTMS+ESI): M+H (C15H11O5S) requires 303.0322, found 303.0326. HPLC purity:
99.3%
4.1.3 In vitro cell migration (Scratch assay)
HUVECs were seeded in 12-well plates at 3 x 104 cells/well and cultured for 48 h. After 48 h,
they were serum-starved (0.1% serum) for 24 h to inactivate the cell proliferation. A scratch was
performed on the cell monolayers using a 100 ꢀL pipette tip (diameter). Cells were then washed
twice with PBS and then treated with medium containing VEGF (10 ng/mL) and either
quercetin, luteolin, or one of the synthesised derivatives (at 10 ꢀM or 1 ꢀM). Images of the
scratches were taken immediately after performing the scratch (t = 0 h) and at 12 h (t = 12 h).
The area not covered by the cells was quantified using ImageJ software. Data were normalized
to the presence or absence of VEGF (VEGF and VEGF-free medium were set at 100% and 0%,
respectively).
4.1.4 Cell viability (MTT assay)
MCF-7 and MCF-7/DX cells were seeded at a density of 4 x 104 into 96 well plates and
incubated to allow the attachment for 24 h. After 24 h, the cells were treated with either
quercetin, luteolin or one of the synthesised derivatives at range of concentrations (0 to 250 µM)
for 67 h. After 67 h, 20 µL of MTT (5mg/mL) solution in PBS was added to each well and the
cells were incubated for 5 h. The purple crystals formed were dissolved in 100 µL of DMSO and
the plates were read at 570 nm using a SPECTRA max UV spectrometer (Bio-Rad).The data
represented are the mean of the three individual experiments.
24

ACCEPTED MANUSCRIPT
4.1.5 Western blotting
HUVECs were cultured in 6-well plates at 10x 104 cells/well and cultured for 48 h. After 48
h, they were serum-starved (0.1% serum) for 24 h to inhibit cell proliferation. Cells were then
pretreated with or without quercetin, luteolin or one of their derivatives for 1 h, followed by
stimulation with 10 ng/mL of VEGF for 15 min (for VEGFR2 activation). Cells were washed
twice with ice-cold PBS and lysed using 100 ꢀL radioimmunoprecipitation assay (RIPA) lysis
buffer with 1% protease inhibitor and phosphatase inhibitor. The amount of protein present
in the cell lysate was assessed using the DC (Bio-Rad) protein assay developed from Lowry’s
protocol for protein assessment. Proteins (15 µg) were separated by SDS-PAGE and then
transferred to PVDF membranes. Membranes were blocked with 3% fat-free dry milk in PBT
buffer and then probed for PTyr-1175-VEGFR2 (1:1000) or total VEGFR2 (1:1000) and actin
dilution. Protein bands were detected by incubating with horseradish peroxidase-conjugated
antibodies (1:10,000) and visualised with enhanced chemiluminescence reagent using
ImageQuant LAS 4000 (GE Healthcare). Gelbands were standardised to total protein loading
relative to actin by using a Gel-Pro analyzer6.0 software as previously described [39].
4.1.6 Molecular docking
Docking validation-To validate the accuracy of the docking procedure to be used, the original
ligands were extracted from the coordinate files (taken from the Protein Data Bank PDB), and
then docked again into the corresponding crystal structure of the proteins, using the automated
docking procedure in the program Surflex-Dock (SFXC) [40], as provided by SYBYL-X2.1.
The resulting ligand conformation from the docking procedure was compared with the ligand
conformation as found in the actual crystal structure of the complex. Comparative structural
orientation of the ligand was calculated as the root mean square deviation (RSMD) between the
docked ligand and the ligand as found in the crystal structure, using the programme LSQKAB,
as provided in the CCP4 [41] suite. If the root mean square deviation (RMSD) value between
25

ACCEPTED MANUSCRIPT
the real and the best-scored conformations is equal to or less than 2.0 Å (representing the grid
spacing used for the docking procedure), then the docking process was considered successful
[42].
Docking procedure- Docking studies were performed using the program Surflex-Dock (SFXC)
as provided by Sybyl-X 2.1. The X-ray crystallographic structures of human VEGFR2 receptor
complexed with 4-amino-furo[2,3-d]pyrimidine (PDB code1YWN, 1.71 Å resolution) [43] and
human VEGF protein complexed with receptor blocking peptide (PDB code1VPP, 1.90 Å
resolution) [44] were retrieved from the Protein Data Bank. The protein structures were prepared
for docking using the Biopolymer Structure Preparation Tool with the implemented default
settings provided in the SYBYL programme suite. Hydrogens were added to the protein
structures in idealised geometries and an overall energy minimisation of each protein was
performed using the MMFF94 force field, employing a conjugate gradient algorithm [45] with a
convergence criterion of 0.5 kcal/mol A−1and up to 5000 iterations. Finally, before the docking
run, all water molecules were removed and the ligand 4-amino-furo[2,3-d]pyrimidine was
extracted from the coordinate file of the VEGFR2 receptor (1YWN). The protomol, representing
the ligand binding groove, was generated using a ligand directed method for VEGFR2, which
allows the docking of ligands into predefined sites, as defined by occupancy of any co-
crystallised ligand at the site of interest. In the case of VEGF, as the crystal structure was not
bound to any ligand, the protomol was generated using the automatic docking procedure.
The Surflex-X docking algorithm docks a given ligand to a receptor using a flexible ligand and a
semi-flexible receptor; in this case the peptides were allowed to be fully flexible while the
receptor was semi-flexible. This allows for optimisation of potentially favourable molecular
interactions, such as those defined by hydrogen bond and van der Waal forces. The docking
results yield a docking score, which takes into consideration entropic, polar, hydrophobic,
26

ACCEPTED MANUSCRIPT
repulsive and desolvation factors. The docking scores were expressed in –log10 (Kd) units to
represent binding affinities, where Kd is the dissociation constant. The free energy of binding of
the ligand to the protein was extrapolated from equation (1).
Free energy of binding = RTlogeKd (1)
The docking results were visualised using the program PyMOL [46,47] and the molecular
interactions of the docked ligands were analysed by the programme CONTACTS, as provided in
the CCP4 suite of programs [41],[48],[49]. Potential hydrogen bonds were assigned if the
distance between two electronegative atoms was less than 3.3 Å, whereas any separation greater
than 3.3 Å, but less than 4.5 Å, was considered a van der Waal interaction.
4.1.7 DSC (Membrane perturbation effect)
DSC measurements were performed using the differential scanning calorimeter nano-DSC (TA
instruments, UK). The samples were degassed before being loaded into the reference and sample
cells. A scan rate of 1.5 °C/min over the temperature range 20-70 °C was employed. Data were
analysed using nano analyser software provided by TA instrument.
Liposome preparation: Model membranes [multilamellar vesicles (MLVs)] were prepared using
dipalmitoylphosphatidylcholine (DPPC) lipid. DPPC was dissolved in chloroform and the
solvent was evaporated under vacuum to obtain a lipid film. The resulted lipid film was
hydrated with an appropriate volume of 20 mM sodium phosphate buffer at pH 7.0 at room
temperature, followed by 3-4 freeze-thawing cycles. For DSC experiments, samples were
prepared by mixing the lipid and drug solutions to obtain 0.2 mg/mL of lipid and 10 or 100 ꢀM
of compound-1 to 3 and 9 to 11.
ASSOCIATED CONTENT
27

ACCEPTED MANUSCRIPT
1
Supporting Information. H NMR spectra of compounds 2-4 and 9-12 and DSC profiles of
compounds 1-3 and 9-12 are provided here. This material is available at xxx.
AUTHOR INFORMATION
Corresponding Authors
*Helen. M. I. Osborn.
Tel: +44 (0)118 378-7338,
E-mail address: h.m.i.osborn@reading.ac.uk.
*Francesca Greco
Tel: +44 (0) 118 378-8244
E-mail address: f.greco@reading.ac.uk
Present Addresses
Reading School of Pharmacy, University of Reading, Whiteknights, Reading, UK, RG6 6AD.
ACKNOWLEDGMENT
Financial support from the Felix Trust and funding from the Royal Society (grant award
reference RG110203) are gratefully acknowledged. We thank Amit Kumar Rajora and Olga
Florek for their help with Western blotting and DSC experiments respectively. We also thank
the University of Reading for the provision of the Chemical Analysis Facility.
ABBREVIATIONS
VEGF-Vascular endothelial growth factor; PDB-Protein data bank; SDS-PAGE- Sodium
dodecyl sulphate-Polyacrylamide gel electrophoresis; PVDF-Polyvinylidene fluoride, MMFF94-
Merck molecular force field; MTT-3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide; NMR-Nuclear magnetic resonance; IR-Infrared; FTMS-Fourier transform mass
spectroscopy; ESI-Electron spray ionisation.
28