Investigation of new 2-aryl substituted Benzothiopyrano[4,3-d[pyrimidines as kinase inhibitors targeting vascular endothelial growth factor receptor 2

Article abstract of DOI:10.1016/j.ejmech.2015.08.027

Vascular Endothelial Growth Factor (VEGF) pathway has emerged as one of the most important positive modulators of Angiogenesis, a central process implicated in tumour growth and metastatic dissemination. This led to the design and development of anti-VEGF monoclonal antibodies and small-molecule ATP-competitive VEGFR-inhibitors. In this study, we describe the synthesis and the biological evaluation of novel 2-aryl substituted benzothiopyrano-fused pyrimidines 1a-i, 2a-i and 3a-i. The ability of the compounds to target the VEGF pathway was determined in vitro exploiting the compounds' antiproliferative efficacy against HUVEC cells. The VEGFR-2 inhibition was confirmed by enzymatic assays on recombinant human kinase insert domain receptor (KDR), by cell-based phospho-VEGFR-2 inhibition assays, and by ex vivo rat aortic ring tests. The selectivity profile of the best performing derivatives belonging to series 2 was further explored combining modeling studies and additional assays in a panel of human cell lines and other kinases.

Full text of DOI:10.1016/j.ejmech.2015.08.027

Accepted Manuscript
Investigation of New 2-Aryl Substituted Benzothiopyrano[4,3-d]pyrimidines as Kinase
Inhibitors targeting Vascular Endothelial Growth Factor Receptor 2
Silvia Salerno, Anna Maria Marini, Giacomo Fornaciari, Francesca Simorini,
Concettina La Motta, Sabrina Taliani, Stefania Sartini, Federico Da Settimo, Aída
Nelly García-Argáez, Ornella Gia, Sandro Cosconati, Ettore Novellino, Pilar D’Ocon,
Anna Fioravanti, Paola Orlandi, Guido Bocci, Lisa Dalla Via
PII:
S0223-5234(15)30211-7
10.1016/j.ejmech.2015.08.027
EJMECH 8063
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 5 June 2015
Revised Date: 28 July 2015
Accepted Date: 11 August 2015
Please cite this article as: S. Salerno, A.M. Marini, G. Fornaciari, F. Simorini, C. La Motta, S.
Taliani, S. Sartini, F. Da Settimo, A.N. García-Argáez, O. Gia, S. Cosconati, E. Novellino, P.
D’Ocon, A. Fioravanti, P. Orlandi, G. Bocci, L.D. Via, Investigation of New 2-Aryl Substituted
Benzothiopyrano[4,3-d]pyrimidines as Kinase Inhibitors targeting Vascular Endothelial Growth Factor
Receptor 2, European Journal of Medicinal Chemistry (2015), doi: 10.1016/j.ejmech.2015.08.027.
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
Investigation of New 2-Aryl Substituted Benzothiopyrano[4,3-d]pyrimidines as
Kinase Inhibitors targeting Vascular Endothelial Growth Factor Receptor 2.
1
1*
1
1
Silvia Salerno, Anna Maria Marini, Giacomo Fornaciari, Francesca Simorini,
1
1
1
1
Concettina La Motta, Sabrina Taliani, Stefania Sartini, Federico Da Settimo, Aída
2
2
3
4
5
Nelly García-Argáez, Ornella Gia, Sandro Cosconati, Ettore Novellino, Pilar D’Ocon,
6
6
6
2
Anna Fioravanti, Paola Orlandi, Guido Bocci, and Lisa Dalla Via .
Dipartimento di Farmacia, Università di Pisa, Via Bonanno 6, I-56126 Pisa, Italy;
Dipartimento di Scienze del Farmaco, Università di Padova, Via Marzolo 5, I-35131
Padova, Italy; DiSTABiF, Seconda Università di Napoli, Via Vivaldi 43, I-81100 Caserta,
Italy; Dipartimento di Farmacia, Università di Napoli Federico II, Via D. Montesano, 40,
I-80131 Napoli, Italy; Departamento de Farmacologia, Universitat de València Av.
Vicente Andrés s/n 46100 Burjassot, València, Spain; Dipartimento di Medicina Clinica e
Sperimentale, Università di Pisa, Scuola Medica - Via Roma 5, I-56126, Pisa, Italy.
*
To whom correspondence should be addressed. Phone: +39 050 2219555 Fax: +39 050
219605. E-mail: marini@unipi.it.
2
1
Dipartimento di Farmacia, Università di Pisa.
2
3
4
5
6
Dipartimento di Scienze del Farmaco, Università di Padova.
DiSTABiF, Seconda Università di Napoli.
Dipartimento di Farmacia, Università di Napoli Federico II.
Departamento de Farmacologia, Universitat de València.
Dipartimento di Medicina Clinica e Sperimentale, Università di Pisa.
1

ACCEPTED MANUSCRIPT
Abstract. Vascular Endothelial Growth Factor (VEGF) pathway has emerged as one of the
most important positive modulators of Angiogenesis, a central process implicated in
tumour growth and metastatic dissemination. This led to the design and development of
anti-VEGF monoclonal antibodies and small-molecule ATP-competitive VEGFR-
inhibitors. In this study, we describe the synthesis and the biological evaluation of novel 2-
aryl substituted benzothiopyrano-fused pyrimidines 1a-i, 2a-i and 3a-i. The ability of the
compounds to target the Vascular Endothelial Growth Factor (VEGF) pathway was
determined in vitro exploiting the compounds’ antiproliferative efficacy against HUVEC
cells. The VEGFR-2 inhibition was confirmed by enzymatic assays on recombinant human
kinase insert domain receptor (KDR), by cell-based phospho-VEGFR-2 inhibition assays,
and by ex vivo rat aortic ring tests. The selectivity profile of the best performing derivatives
belonging to series 2 was further explored combining modelling studies and additional
assays in a panel of human cell lines and other kinases.
Keywords: Tumor angiogenesis, Receptor Tyrosine Kinases, VEGFR, Kinase Inhibitors,
Benzothiopyranopirimidines
2

ACCEPTED MANUSCRIPT
1
. Introduction
Angiogenesis is a physiological process that leads to the formation of new capillaries
sprouting or splitting from pre-existing vessels, a requirement necessary for growth and
proliferation of solid tumors. [1]. Accordingly, inhibition of this process may reduce
vascularization, resulting in antitumor therapies that are supposed to be less toxic for
chronic administration, in comparison with conventional chemotherapeutic treatments, and
able to maximize the results, while suppressing tumour metastasis [2,3].
The angiogenic process is orchestrated via the expression of a range of cell surface and
intracellular proangiogenic factors, including the vascular endothelial growth factor
(VEGF) family (VEGF-A, VEGF-B, VEGF-C, VEGF-D, VEGF-E). The various members
of the VEGF family have overlapping abilities to interact with a set of highly homologous
tyrosine kinase receptors (TKRs): VEGFR-1 (Flt-1) and VEGFR-2 (human KDR or
murine Flk-1), largely expressed in endothelial cells, and primarily involved in both
physiological and pathological angiogenesis. [4,5]. Although the precise role of VEGFR-1
function is still emerging, it seems to act primarily as a decoy receptor, modulating the
availability of VEGF for VEGFR-2, which is the principal receptor, through which VEGF
exerts its responses in endothelial cells.
In tumour cells, VEGFR-2 is strongly autophosphorylated by increased expression of
VEGFs and mediates a serial of mitogenic and survival downstream pathways, including
MAPK, P13K or Akt, which are usually hyperactivated in several malignances [6-8].
From the improved understanding of angiogenesis mechanisms in tumorigenesis, the
therapeutic strategies blocking VEGF/VEGFR-2 signalling systems became a promising
and well-validated approach for suppression of pathological neovascularisation and,
consequently, for the treatment of both solid tumours and haematological malignancies [9,
3

ACCEPTED MANUSCRIPT
1
0]. Agents currently developed against VEGF or its receptors include the humanized anti-
VEGF monoclonal antibody Bevacizumab, which targets extracellular VEGF. By means
of receptor-based virtual screening studies, a significant number of small molecules
VEGFR-2 inhibitors, which compete with ATP binding in the catalytic domain of tyrosine
kinase, has been developed, leading to the identification of several classes of VEGFR
inhibitors, such as Semaxanib (SU5416), Vandetanib (ZD6474), Vatalanib (PTK787), and
AAL993 (ZK260253), (Chart 1).
Chart 1. Structures of VEGFR Tyrosine Kinase inhibitors.
The most performing leads, designed as ATP competitive compounds, proved to firmly
bind to the catalytic site of the proteins, exploiting hydrophobic interactions, strengthened
by one or more hydrogen bonding groups, often heteroatoms. Moreover, other substituents,
connected by different spacer chains, including amino and/or amide groups, can protrude
toward two other distinct areas of the site: the so called “solvent accessible region” and
“buried region” [11, 12].
Among the wide series of thoroughly described chemical structures [13, 14], we drew our
attention on those containing, as privileged moiety, a pyrimidine ring, either isolated or
fused to form different heterocyclic structures, (Pyrrolo-pyrimidines, Pyrazolo-
4

ACCEPTED MANUSCRIPT
pyrimidines, Furo- and Thieno-pyrimidines, Pyrimido-pyrimidinones, Pyrido-pyrimidones)
that have been reported as interesting VEGFR inhibitors (Chart 2).
O
CH3
NH2
N
N
N
N
N
N
HN
N
O
H N
N
N
2
N
N
N
S
N
Pyrazolo-pyrimidine
Furo-pyrimidine
H
AEE788 (Pyrrolo-pyrimidine)
Cl
O
N
NH
N
N
Cl
R1
HN
N
N
O
HN
S
HN
R
N
N
O
N
R
N
OH
Pyrimido-pyrimidines
TKI-28 (Pyrido-pyrimidinone)
Thieno-pyrimidine
Chart 2. VEGFR inhibitors containing both pendant or fused pyrimidines.
The research program of our group has always been devoted to the preparation and the
evaluation of polycyclic chromophores extensively studied as new antiproliferative agents
[
15-17]. In the last years, some studies on the benzothiopyranopyrimidine system gave rise
to the disclosure of novel compounds endowed with interesting pharmacological properties
18,19]; in particular the benzothiopyranopyrimidine A (Chart 3) showed an appreciable
cytotoxic activity on HL-60 cell line (GI = 7.15 µM) [20].
[
5
0
5

ACCEPTED MANUSCRIPT
Z
R
N
S
N
N
S
N
X
H CO
3
I
X = H, OCH3, Cl; Z = -, NH, NHCH2;
A
R = H, OCH , Cl
3
Chart 3. Structures of benzothiopyranopyrimidines.
Focusing on these results and moving from the structures of some reported anilino
substituted pyrimidines [21], we designed the series of benzothiopyrano-fused pyrimidines
of general formula I, (Chart 3), characterized by an aniline moiety inserted in the 2-
position of the tricyclic system. Moreover, with the aim to perform critical SAR studies
and to confirm the role of the NH group, the analogue 2-benzylamino and 2-phenyl
derivatives have been also synthesized.
It was hypothesized that these new products might share the pharmacophoric requirements
suitable for an anchoring to the ATP binding site, in particular the aniline NH group and
the pyrimidine N and N atoms, that might be implicated in H-bonding interactions with
1
3
critical residues of the catalytic cleft [13,14].
The ability of the new benzothiopyrano-pyrimidines I to inhibit the kinase activity of the
VEGFR-2 was determined by a biochemical assay. The effect of the compounds on
viability of human umbilical vein endothelial cells (HUVECs) was evaluated by means of
an inhibition growth assay. The capacity of the most interesting compounds to inhibit the
enzyme catalyzed phosphorylation processes, in HUVECs, as well as the ability to
interfere with the angiogenic process in vitro on rat aortic rings, has been assayed. The
antiproliferative properties of all compounds have been further investigated on a panel of
6

ACCEPTED MANUSCRIPT
human tumor cell lines, and the angio-kinase selectivity profile was assessed against a set
of 11 human kinases. Finally, molecular docking calculations allowed clarifying at
molecular level the interaction pattern established by the compounds with the KDR domain
of VEGFR-2.
2
. Results and discussion
2
.1. Chemistry
The synthetic procedure leading to the target 2-aryl-substituted-5H-benzothiopyrano
[
4,3-d]pyrimidine derivatives 1a-i, 2a-i and 3a-i takes advantage by the use as “synthetic
tool” of the 7-substituted benzothiopyrane moiety, which was demonstrated to be a
versatile intermediate [20].
The 1,3-bielectrophile reactivity of the opportune intermediate 7-substituted-3-
dimethylaminomethylene derivatives 4a-c [20,22-24] in the reaction with the appropriate
binucleophile amidines 5a-c or 6a-c in an ethanolic refluxing solution, in the presence of
sodium ethoxide [25], allowed to easily obtain the desired 2-phenyl substituted pyrimidines
1
a-i and their benzylamino substituted analogues 3a-i, as reported in Scheme 1. P-
substituted benzamidine hydrochlorides 5b-c, even commercially available, were difficult
to find; so they were obtained with good yields by reaction of the appropriate benzonitrile
derivative with a methanol solution of sodium methoxide and ammonium chloride, as
reported in literature [26-28]. The p-substituted benzylguanidines 6a-c were synthesized by
condensation of methylthiopseudourea sulfate with the appropriate benzylamine in
aqueous solution, following already described procedures [29-31].
7

ACCEPTED MANUSCRIPT
R
R
CH NH
2
2
CN
R = H, OCH3, Cl
ii
R = OCH3, Cl
i
R
R
O
S
CH3
CHN
CH3
X
HN
HN
HN
NH2 HCl
a-c
4a-c
R
5
NH2 H SO
2 4
HN
6
a-c
iii
iii
N
N
R
N
S
N
X
S
X
3
a-i
1
a-i
Scheme 1. Reagents and conditions: i) MeONa, MeOH, NH Cl; ii) methylthiopseudourea
4
sulphate, H O, r.t. for 7 hours, 100 °C for 16 hours; iii) EtONa, refluxing ethanol.
2
The crude target compounds were obtained as unique products (tlc analysis) in good
overall yields, and purified by crystallization.
The preparation of compounds 2a-i was accomplished by means of a different synthetic
route. [32] 3-Dimethylaminomethylene derivatives 4a-c were allowed to react with
guanidine hydrochloride to obtain the 7-substituted-2-aminopyrimidino derivatives 7a-c,
which, by reacting under nitrogen atmosphere with the appropriately p-substituted
phenyliodide in anhydrous dioxane, in the presence of K CO , N,N'-
2
3
dimethylethylendiamine and CuI, afforded the crude target compounds 2a-i which were
8

ACCEPTED MANUSCRIPT
purified by flash chromatography, using petroleum ether 60-80 °C/ethyl acetate 70:30 as
the eluting system (Scheme 2).
R
NH2
O
N
S
N
CH3
CH3
CHN
NH
i
ii
N
S
N
X
S
X
7a-c
4
a-c
7
7
7
a: X = H
b: X = OCH3
c: X = Cl
X
4
4
4
a: X = H
b: X = OCH3
c: X = Cl
2
a-i
Scheme 2. Reagents and conditions: i) guanidine hydrochloride, EtONa, refluxing ethanol;
ii) p-substitutedphenyliodide, K CO , N,N'-dimethylethylendiamine, CuI, dioxane.
2
3
The purity of the target compounds was assessed by TLC analysis and by physicochemical
1
13
properties, analytical and H NMR and C NMR spectral data, which were in agreement
with the proposed structures and with other previously reported results. (Experimental
protocols for details).
2
2
.2. Biological evaluation
.2.1. KDR inhibition assay
The ability of the new 2-aryl-substituted benzothiopyrano-pyrimidines 1a-i, 2a-i and
a-i in inhibiting the kinase activity of the VEGFR-2 was determined by a biochemical
3
assay performed with a recombinant human kinase insert domain receptor (KDR), using
Semaxanib (SU5414) as reference standard. Despite its moderate potency, Semaxanib
9

ACCEPTED MANUSCRIPT
shows a useful selectivity in inhibiting VEGFR-1 and -2 tyrosine kinases and VEGF-
dependent endothelial proliferation, as compared with novel more potent but multitargeted
receptor tyrosine kinase inhibitors, like Sunitinib or Sorafenib [33]. As reported in Table 1,
listing activities expressed as inhibition percentage at 7 µM concentration, the results
obtained highlighted generally good to high values of KDR inhibition for the most part of
the 2-anilino substituted derivatives (compounds of series 2), in some cases even higher
with respect to that exerted by Semaxanib. Both the addition of a methylene spacer
(compounds of series 3) and the deletion of the amino group (compounds of series 1)
reduced the inhibitory efficacy against the target protein, thus clearly indicating a crucial
role for the length of the side chain and the NH group.
Table 1. KDR Enzyme Inhibitory Activity and HUVEC Cell Growth Inhibition of
Derivatives 1a-i, 2a-i and 3a-i.
Z
R
N
S
N
X
a
b
c
%
IC50 (µM)
GI50 (µM)
Compd.
Z
X
R
KDR
KDR
HUVEC
-
-
H
H
H
OCH
<20
31
4.0
1
a
1
b
3
3
3
3
12
11.3
2.9
1
c
-
H
Cl
H
<20
<20
<20
<20
<20
41
1d
-
OCH
OCH
OCH
Cl
3
3
3
1
e
-
-
OCH
Cl
4.8
1
f
>20
2.3
1g
-
H
1h
-
Cl
OCH
Cl
4.7
1
i
-
Cl
<20
54
15.3
2.4
2a
NH
NH
NH
H
H
8.2
2.7
2b
H
OCH
Cl
67
0.45
7.3
2
c
H
<20
>50
1
0

ACCEPTED MANUSCRIPT
2
d
NH
NH
OCH
OCH
OCH
Cl
3
3
3
H
OCH
Cl
35
39
>50
17.5
>50
>50
>50
5.6
4.3
0.74
>20
>20
16.4
0.35
>20
>20
12.3
>20
>20
>20
15.1
12
2
e
3
3
3
3
3
2
f
NH
<20
<20
<20
61
2
g
NH
H
2
h
NH
Cl
OCH
Cl
2
i
NH
Cl
3
a
NHCH
NHCH
NHCH
NHCH
NHCH
NHCH
NHCH
NHCH
NHCH
2
2
2
2
2
2
2
2
2
H
H
OCH
Cl
<20
<20
<20
<20
<20
<20
<20
35
3
b
H
3
c
H
3
d
OCH
OCH
OCH
Cl
3
3
3
H
3
e
OCH
Cl
3
f
3
g
H
3
h
Cl
OCH
Cl
3
i
Cl
<20
17
Semaxanib
40
12.9
13.6
a
b
c
Percentage of kinase inhibition obtained at 7 µM of test compound.
IC values represent the concentration that induces the 50% reduction in enzyme activity.
5
0
GI values represent the concentration (µM) of compound able to produce 50% cell death
5
0
with respect to the control culture.
Moreover, an important role of the substituents on both the chromophore and the pendant
phenyl moiety, position 8 (X) and 4' (R), was observed. In the case of the 2-anilino-
substituted derivatives 2, when the distal phenyl ring is unsubstituted (2a, 2d, 2g), or bears
a p-methoxy group (2b, 2e, 2h), the inhibition percentages progressively decrease when
methoxy or, to a greater extent, chlorine substituents are present at 8-position of the
chromophore. Conversely, when a pendant p-chloro phenyl group is present (2c, 2f, 2i)
only the insertion of a chlorine in the 8-position, as in compound 2i, allowed the
occurrence of inhibitory capacity. Accordingly, a significant inhibitory efficacy has
displayed by compounds 2a, 2b, 2e and 2i; with 2b and 2i showing the highest percentage
values in the whole subseries.
1
1

ACCEPTED MANUSCRIPT
Within the series of the 2-phenyl (1a-i) or 2-benzylamino (3a-i) substituted derivatives, the
8
-chloro-4'-methoxy substituted analogues 1h and 3h turned out to be the most active
inhibitors with percentage values lower or comparable to those of the reference drug.
The activity of the best performing 2-anilino-substituted derivatives was also calculated as
IC (Table 1). The results confirm for 2a, 2b, 2e and 2i an interesting capacity to inhibit
5
0
the enzyme, with IC values ranging from 2.7 to 17.5 µM, comparable (2e) or even lower
5
0
(2a, 2b and 2i) with respect to that obtained for the reference compound.
2
.2.2. Evaluation on HUVEC cells
The ability of new derivatives to inhibit the phosphorylation activity of VEGFR-2,
prompted us to test their effect on human umbilical vein endothelial cells (HUVECs)
growth by means of an in vitro assay. The results, expressed as GI values, are showed in
5
0
Table 1, Semaxanib was used as reference compound.
After a 72-h exposure, the ability to exert antiproliferative effects on HUVEC cells was
shown by the majority of 2-phenyl substituted pyrimidines 1a-i showing GI values, in the
5
0
micromolar range (from 2.3 to 15.3 µM), comparable or lower to that of Semaxanib. The
most part of 2-anilino pyrimidines 2a-i is able to exert a significant cytotoxicity and,
interestingly inside this subseries, derivatives 2b, 2e and 2i show even submicromolar GI50
values (0.35-0.74 µM). In contrast, the 2-benzylamino substituted derivatives 3a-i always
exhibited low or null cytotoxic capacities in these cells. Significantly, by comparing the
antiproliferative effects exerted by analogous compounds, bearing the same aryl side
group, a certain role seems to be played by the substituents in position 8 (X) and 4' (R).
Within the sub-series 1a-i, the most active compounds are 1a, 1d, 1g followed by 1b, 1e,
1
h, bearing the unsubstituted (R = H) and the methoxy substituted (R = OCH ) pendant
3
phenyl group, respectively. Conversely, the presence of a chlorine (R = Cl), as in
1
2

ACCEPTED MANUSCRIPT
compounds 1c, 1f and 1i, generally reduces the activity. As regards the role of the X group
in the 8-position, a similar profile is displayed within each group of analogues, since the
introduction of methoxy (1d, 1e versus 1a, 1b) or chlorine (1g, 1h versus 1a, 1b) increases
cytotoxic ability, but in the p-chloro subseries (1f, 1i versus 1c). The nature of the
substituents X and R also modulates the antiproliferative effects in the group of 2-anilino
substituted pyrimidines 2a-i. Actually, cytotoxic potency is usually enhanced by the
insertion of a methoxy electron donating group on the phenyl ring (R = OCH , compounds
3
2
b, 2e versus 2a, 2d) or it is lowered by a chlorine (R = Cl, compounds 2c, 2f versus 2a,
d). The presence of the halogen in the 8-position (X = Cl) on the heterocyclic scaffold
2
gives rise to contrasting results. While 8-chloro substituted pyrimidines have null or scarce
activity when R = H (compound 2g) or R = OCH (compound 2h), the 4'-chloro analogue
3
2
i (R = Cl) is surprisingly the most effective to inhibit HUVEC cells growth of the whole
series. The less performing was sub-series 3, actually benzilamino-pyrimidines 3a-i that
displayed comparable and low cytotoxic potencies, independently of the presence of a
hydrogen, a methoxy group or a chlorine in the 8 (X) and/or 4' (R) positions. In this latter
series of analogues the increase in the distance between the phenyl ring and the amine
linker through a methylene spacer seems accountable for the almost disappearance of
cytotoxicity, and the dependence of the cellular effects on the other groups appears
responsible only for slightly different biological activities.
Overall, the above results confirm the important role played by the NH spacer group, given
that only the 2-anilino pyrimidines 2a-i showed both significant KDR inhibition potency
and HUVECs antiproliferative activity, likely correlating the VEGF/VEGFR signal system
with their cellular activity.
1
3

ACCEPTED MANUSCRIPT
2
.2.3. Inhibition of VEGFR-2 phosphorylation in endothelial cells.
The efficacy observed on HUVEC cells was further investigated by evaluating the
ability of the most cytotoxic derivatives, namely 2b, 2e, and 2i, to inhibit the
phosphorylation of the target receptor on whole cells [34]. After exposure to 2b (fig. 1A),
2
e (fig. 1B), and 2i (fig. 1C) at different concentrations, the amount of the phosphorylated
form of VEGFR-2 in HUVEC cell lysates was significantly reduced at the highest
concentration (2.5 µM).
Figure 1. Inhibition of VEGFR-2 phosphorylation by 2b (A), 2e (B) and 2i (C)
compounds in HUVEC cells after 72 h of treatment. pVEGFR-2 concentrations were
measured by ELISA kits and they were normalized to total VEGFR-2 protein
concentration. Columns and bars, mean values ± SD, respectively. *P<0.05 vs. vehicle-
treated controls.
1
4

ACCEPTED MANUSCRIPT
In particular, in samples treated with 2b and 2i the amount of phosphorylated form of
VEGFR-2 significantly decreased in a concentration-dependent manner, starting at the
experimental GI of cell proliferation (figs. 1B and C).
5
0
2
.2.4. Ex vivo anti-angiogenic activity evaluation
Angiogenesis assays are essential for the identification of potential pro-angiogenic
agents as well as for the screening of pharmacological inhibitors. Among them, the ex vivo
rat aortic ring" assay, initially developed by Nicosia and Ottinetti [35], bridges the gap
“
between in vivo and in vitro models. Actually, by using intact vascular explants, the
environment in which angiogenesis takes place results more accurately reproduced than
that with isolated endothelial cells [36]. Considering that the organization of endothelial
cells in capillary-like structures is a process that better occurs in a 3-dimensional
extracellular matrix, the culture of rat aortic explants in a matrix of this type results one of
the most effective ways for generating new capillaries in vitro.
In order to assess their capacity to inhibit microvessel sprouting in an angiogenic process,
the selected 2-anilino substituted derivatives 2b and 2i have been tested in the ex-vivo “rat
aortic ring assay”. Aortic explants were incubated with compounds 2b and 2i, at different
concentrations (100 nM, 1 µM, 5 µM and 10 µM) and the activity was compared with that
exerted by the reference drug Semaxanib, at the same concentrations, (figs. 2 and 3).
1
5

ACCEPTED MANUSCRIPT
Figure 2. Representative images of the inhibition of microvessel sprouting by compounds
2
b, and 2i at [100 nM], [1 µM], [5 µM], [10 µM], compared with DMSO and Semaxanib
at the same concentrations. Images were visualized using a Nikon fluorescence microscope
Eclipse E800 equipped with Nikon camera, with the 4x objective, after fluorescence cell
th
staining at day 7 of the experiment. All experiments have been conducted in duplicate and
at least on two different animals.
1
6

ACCEPTED MANUSCRIPT
Figure 3. Comparison of the angiogenic growth (Y = Length, µms) of rat aortic explants
cultured with compounds 2b and 2i at different concentrations (X = 100nM ;1ꢀM; 5ꢀM;
1
0ꢀM). DMSO is used as control and Semaxanib as reference compound. Data were
calculated at day 7th of the experiment and represent the mean of n = 2-4
Interestingly, both selected compounds inhibit the angiogenic sprouting in aortic ring
assay, and the use of different concentrations allowed us to identify the range in which the
inhibition took place. In particular, the benzothiopyrano derivatives 2b and 2i, seemed to
inhibit the process at concentrations ranging from 1 to 5µM, showing a profile similar to
that of the reference standard Semaxanib.
2
.2.5. Evaluation on HeLa, A-431 and MSTO-211H tumour cell lines
The high structural similarity of kinase ATP binding pocket has made the development
of selective inhibitors extremely challenging. Actually, most of the developed angio-kinase
inhibitors also hit a broad range of additional protein kinases; however, a significant
1
7

ACCEPTED MANUSCRIPT
capacity to affect simultaneously different mitogenic patways, is often associated with
serious side effects and toxicities [37, 38]. In this regard, we studied the cell proliferation
profile of the benzothiopyrano collection of compounds 1, 2 and 3 against a panel of
human tumour cell lines expressing different kinases: HeLa (cervix adenocarcinoma), A-
4
31 (epidermoid carcinoma) and MSTO-211H (biphasic mesothelioma). As reported in
Table 2, the cell growth evaluation once again revealed a crucial role for the length of the
side chain and the presence of the NH group, as evident cytotoxic effects have been
observed for the most part of 2-anilino substituted derivatives 2, showing GI values in the
5
0
micromolar range. In addition, a notable effect might also be attributable to the p-methoxy
substituent on the pendant phenyl ring (R = OCH ), being 2b and 2e the most effective
3
derivatives. The p-chlorine (R = Cl) substitution seems to be detrimental, with the
exception of 2i in which, accordingly with the above discussed results, a good activity was
present by introducing a chlorine atom on the 8-position (X = Cl).
Table 2. Human Tumour Cell Growth Inhibition of Derivatives 1a-i, 2a-i and 3a-i.
a
GI50 (µM)
Compd.
HeLa
A-431
MSTO-211H
1
a
>20
>20
>20
>20
>20
>20
1
b
1
c
>20
>20
>20
>20
1
d
6.5±1.2
>20
11.9±1.1
>20
1
e
>20
1
f
>20
>20
>20
1
g
>20
>20
>20
1
h
>20
>20
>20
1
i
>20
>20
>20
2
a
10.2±2.1
0.7±0.5
>20
11.7±1.3
2
b
0.96±0.25
1.08±0.04
2
c
>20
>20
1
5.6±2.2
20.1
11.5±1.5
15.0±1.6
2
d
1
8

ACCEPTED MANUSCRIPT
2
e
0.83±0.14
>20
0.96±0.2
>20
1.08±0.2
>20
2
f
2
g
>20
>20
>20
2
h
2.1±0.2
1.5±0.4
>20
5.8±0.4
2.2±0.4
>20
>20
1.6±0.3
>20
2
i
3
a
3
b
>20
>20
>20
3
c
>20
>20
>20
3
d
>20
>20
>20
3
e
>20
>20
>20
3
f
>20
>20
>20
3
g
>20
>20
>20
3
h
>20
>20
>20
3
i
>20
>20
>20
a
GI values represent the concentration (µM) of compound able to produce 50% cell
5
0
death with respect to the control culture.
Between the others subseries, 1a-i and 3a-i, only derivative 1d shows a detectable
antiproliferative effect, which appears more significant in A-431 cells.
2
.2.6. EGFR inhibition assay
Important evidences point out numerous cross-talks between activated epidermal
growth factor receptor (EGFR) largely expressed on A-431 cells and VEGFR [34, 39],
moreover a dose-dependent decrease in VEGF expression in EGFR-responsive A-431 cells
was indicated for different classes of EGFR targeting agents [40]. Accordingly, the
capability of compounds 1d and 2a-i to bind EGFR was assayed. The inhibitory effect of
the test compounds was routinely estimated at 100 µM, following a previously reported
protocol [41, 42]. Compounds 2a-f displayed null or low percentage values of kinase
inhibition, while compounds 1d and 2g-i showed a certain capacity to reduce the kinase
activity. These latter compounds, whose percentage values of inhibition were higher than
1
9

ACCEPTED MANUSCRIPT
8
0%, have been tested at additional concentrations calculating the IC values, exploiting
5
0
gefitinib as reference standard (Table 3). Significantly, the shortened derivative 1d
displayed an IC value of 4.72 µM, in accordance with its antiproliferative activity on A-
5
0
4
31 cell line.
Table 3. EGFR Inhibition Data of Derivatives 1d and 2a-2i.
Compd.
EGFR
EGFR
a
b
%
IC50 (µM)
1
d
100
50
4.72
d
2
a
n.d.
d
2
b
< 50
< 50
57
n.d.
d
2
c
n.d.
d
2
d
n.d.
d
2
e
< 50
n.d.
c
d
2
f
n. a.
n.d.
2
g
100
93
21.18
21.56
45.70
0.022
2
h
2
i
81
Gefitinib
a
Percentage of kinase inhibition obtained at 100 µM test compound.
b
IC values represent the concentration that induces the 50% reduction in enzyme activity.
5
0
c
d
not active, not determined
2
.2.7. Kinases inhibition assays
The high inhibitory effect of 2b and 2i on VEGFR phosphorylation and of 1d on
EGFR kinase activity, prompt us to deepen their kinase selectivity profile. In particular,
their effect has been assessed against a set of 10 human kinases (Akt1/PKBα, AurA/Aur2
2
0

ACCEPTED MANUSCRIPT
kinase, CDC2/CDK1, ERK1, FAK, MEK1/MAP2K1, PDGFRβ kinase, RAF-1 kinase,
ROCK1, Src kinase), which are reported to be mainly involved in cell proliferation. No
inhibitory activity was detected for 1d and 2i, when assayed at 10 µM concentrations
(Figs. 4a and 4c), thus excluding any involvement of the tested kinases in the effects of
these derivatives on cell viability. Only compound 2b might appear a multi-target inhibitor
of tyrosine kinases, inhibiting besides VEGFR-2, also Aurora kinase and CDC2/CDK1 at
significant extent (Fig. 4b).
Figure 4a. Kinases Inhibitory Activities of compound 1d.
2
1

ACCEPTED MANUSCRIPT
Figure 4b. Kinases Inhibitory Activities of compound 2b.
Figure 4c. Kinases Inhibitory Activities of compound 2i.
2
.3. Molecular Modelling
To better understand the reasons behind the VEGFR-2 inhibitory activity of the
compounds described above, molecular docking studies were performed on compound 2b.
It displays the highest, although not selective, VEGFR-2 kinase activity inhibitory property
(
IC50 2.7 µM), high antiproliferative activity on both HUVECs and tumour cell lines
2
2

ACCEPTED MANUSCRIPT
(HeLa, A-431 and MSTO-211H), as well as evident anti-angiogenic properties in the ex
vivo rat aortic assay.
These calculations were attained using the AD4 software [43, 44]. This choice was dictated
by results of a recent work in which this program was demonstrated to be the best in
reproducing the experimental (X-Ray) binding pose for different kinases in cross-docking
experiments among 17 different docking procedures [45]. To date, more than 30 crystal
structures of VEGFR-2 in complex with several inhibitors have been reported. In this
inspection, we decided to consider only the structures having a resolution below 2Å
(
1YWN [46], 2P2H [47], 2XIR, 3BE2 [48], 3EWH [49], 3VHE [50], 3VNT [51], 4AG8
52], 4ASE [53], 3VO3 [54]) in which 1YWN displayed the so-called A-loop (activation
[
loop) in its open conformation while all the others had the same protein region in its closed
form. Among these, the structure 4AG8 was used, as it demonstrated to be the best
performing one in AD4 cross-docking experiments. In particular, employing this structure,
AD4 was able to reproduce the X-ray conformation of all the analysed co-crystal ligand
with an average root-mean squared deviation (rmsd) of 0.89 Å (see Table 4).
2
3

ACCEPTED MANUSCRIPT
Table 4. Rmsd values resulting from cross-docking experiments on the 9 considered
VEGFR-2 structures.
Ligands
Average
1
YWN 2P2H 2XIR 3BE2 3EWH
3VHE
3VNT
4AG8
4ASE
rmsd
5.92
5.62
5.37
4.68
3.06
2.41
2.99
0.95
2.01
1
YWN
0.66
1.57
9.55
5.19
6.76
10.37
1.90
2.98
2.79
0.85
0.66
0.58
0.54
0.83
1.48
11.82
1.26
0.90
3.64
1.54
1.92
3.06
1.16
1.44
2.33
1.03
1.82
0.94
0.95
0.97
0.96
5.60
1.40
1.72
3.03
2.79
0.83
1.16
0.69
0.60
6.72
8.72
2.01
6.86
3.19
0.64
0.43
0.58
0.39
10.66
11.98
1.96
12.83
2.83
0.94
0.46
0.42
0.59
11.32
8.43
2.19
3.03
2.91
0.79
0.58
0.69
0.27
2
P2H
XIR
BE2
2
12.39
3.11
1.17
1.45
10.38
1.17
0.89
11.99
7.97
3
3
EWH
9.12
3
VHE
VNT
11.56
10.75
1.49
3
4
AG8
ASE
4
10.74
Interestingly, the 200 independent docking runs of 2b within the VEGFR-2 kinase active
site converged toward a single cluster solution (i.e. binding conformation differing by less
than 1.5 Å) for which a binding free energy (ꢁGAD4) of -7.62 kcal/mol was predicted (Fig.
5
).
2
4

ACCEPTED MANUSCRIPT
Figure 5. Binding mode of compound 2b, within the VEGFR-2 active site. The ligand is
represented as green sticks while the protein as orange sticks and ribbons. H-bonds are
depicted as dashed yellow lines. The picture was rendered with the UCSF Chimera
software. [55]
In this conformation, the benzothiopyranopyrimidine scaffold is inserted in the enzyme
active site with its nitrogen in position 3 and the adjacent exocyclic NH in position 2
establishing a double H-bond with the backbone NH and CO of Cys919 in the enzyme
hinge region, respectively. These interactions explain why 2b and its closely related
congeners are effective VEGFR-2 inhibitors and why the 2-aryl analogues (1a-i), by
lacking the double H-bond interaction generally display lower inhibition against the
enzyme (Table 1). Moreover, the lipophilic benzothiopyranopyrimidine ring is fitting in
the hydrophobic ATP cleft where favorable contacts are established with Leu1035,
Val899, Val916 (gate-keeper residue), Val848, and Leu840. The limited extension of this
cleft would not allow to easily host bulky substituents in position 8 of the chromophore
without a subsequent relocation of the ligand. Indeed, this relocation would result in the
2
5

ACCEPTED MANUSCRIPT
loss of the double H-bond interaction with the enzyme’s hinge region residues, thus
explaining the lower inhibition potencies of compounds 2d-2h. On the other hand, the
presence of the positively charged K868, within the same cleft, should negatively affect the
ligand/enzyme
recognition
suggesting
that
modifications
of
the
benzothiopyranopyrimidine scaffold, aimed at minimizing these unfavourable contacts,
should enhance the inhibitory potencies against the enzyme. We are now testing this
hypothesis and results will be reported in due of course. In the binding site, Phe1047 is
placed between the benzothiopyrane aromatic ring and the aniline moiety in position 2 of
the tricyclic system establishing with them charge-transfer interactions. In this respect, it
could be speculated that the double substitution in position 8 (X) and 4' (R) with the
electron withdrawing chlorine atom would enhance the strength of the aforementioned
interaction thereby explaining why 2i is still so efficient as VEGFR-2 kinase inhibitor.
The aniline substituent is partially solvent exposed and this should explain why its
substitution with the longer 2-benzylamino group would result in weaker VEGFR-2
inhibitors (3a-i). As depicted in Figure 5, the p-methoxy substituent of 2b is pointing
towards Asn923 residue and in principle additional H-bond interactions could be
established through the rearrangement of this residue.
3
. Conclusions
Activation of the “angiogenic switch” is considered one of the most critical factors for
tumour growth and progression, and numerous small molecule inhibitors of the VEGF-
VEGFR pathways have been proven to be an attractive approach for anticancer drug
discovery.
2
6

ACCEPTED MANUSCRIPT
In this work, we developed novel antiangiogenic agents based on the benzothiopyrano[4,3-
d]pyrimidine scaffold, characterized by a pendant aryl moiety. In particular, the synthesis
of new benzothiopyrano[4,3-d]pyrimidines carrying phenyl (compounds 1a-i), phenyl-
amino (compounds 2a-i) and benzyl-amino (compounds 3a-i) side groups in the 2-position
of the heterocyclic system is described. Of these derivatives, compounds 2a-i emerged as
the most interesting, revealing that the presence of the anilino side moiety is required to
obtain a significant effect against the target proteins.
Compounds 2a-i displayed a promising inhibitory profile on both human kinase receptor
KDR and endothelial HUVEC cells, with values in the low micromolar range. The kinase
inhibitory activities of the best performing compounds 2b and 2i was further investigated
by measuring the formation of newly phosphorylated substrates via a cell-based phospho-
VEGFR-2 inhibition assay. Moreover, their capacity to inhibit the generation of new
capillaries was confirmed by means of the ex-vivo “rat aortic ring assay”.
To determine the compounds’ selectivity profile, we evaluated the cytotoxic effects of 1a-i,
2
a-i and 3a-i against three human tumour cell lines. As a general trend, only few
compounds, belonging to the subseries 2, retained cytotoxic activity, albeit with about a 2-
fold increase in GI values with respect to HUVECs. Hence, in order to investigate the
5
0
involvement of other structurally related kinases in the cellular activity, we assayed the
capacity of compounds 2b and 2i to interfere with a panel of eleven additional human
kinases, (EGFR, Akt1/PKBα, AurA/Aur2, CDC2/CDK1, ERK1, FAK, MEK1/MAP2K1,
PDGFRβ, RAF-1, ROCK1 and Src kinase). By this study, the most selective derivative
seems to be 2i, conversely, the best performing compound 2b exhibited a certain activity
also against Aurora kinase and CDC2/CDK1, probably due to the high structural
similarities shared by the VEGFR-2 ATP binding pocket with that of other kinases.
2
7

ACCEPTED MANUSCRIPT
Molecular docking calculations allowed understanding the interaction pattern established
by the compounds and the KDR domain of VEGFR-2. The model, exploited for compound
2
b, showed that the benzothiopyranopyrimidine scaffold could interact with the enzyme
active site, with its nitrogen in position 3 and the adjacent aniline NH in position 2
establishing a double H-bond with the backbone NH and CO of Cys919, in the enzyme
hinge region, respectively; thus corroborating the crucial role of the 2-anilino side moiety.
In conclusion, the ability to exert evident inhibitory effects against VEGFR-2, both in vitro
and in endothelial cells, together with noteworthy cytotoxicity on HUVECs and ex vivo
antiangiogenic activity, render the new 2-anilino substituted heteropolycyclic moiety an
interesting structure inside the field of VEGFR-2-targeted antitumor agents. Thus, further
structural optimization of this promising system is in progress in our laboratories, with the
purpose of obtaining new antiangiogenic active candidates.
4
4
4
. Experimental
.1. Chemistry
.1.1. General
The uncorrected melting points were determined using a Reichert Köfler hot-stage
1
apparatus. H NMR spectra were recorded on a Varian Gemini 200 spectrometer in
1
13
dimethyl-d sulfoxide solution. or on a Bruker AVANCE 400 ( H, 400 MHz, C, 100
6
MHz) in DMSO-d6 (internal standard tetramethylsilane). The coupling constants are given
in Hertz. Elemental analyses were performed by our Analytical Laboratory and were
within ± 0.4%. Magnesium sulphate was used as the drying agent. Evaporations were made
in vacuo (rotating evaporator). Analytical TLC have been carried out on Merck 0.2 mm
2
8