Structure-activity relationships and molecular modelling of new 5-arylidene-4-thiazolidinone derivatives as aldose reductase inhibitors and potential anti-inflammatory agents

Article abstract of DOI:10.1016/j.ejmech.2014.05.003

A series of 5-(carbamoylmethoxy)benzylidene-2-oxo/thioxo-4-thiazolidinone derivatives (6-9) were synthesized as inhibitors of aldose reductase (AR), enzyme which plays a crucial role in the development of diabetes complications as well as in the inflammat

Full text of DOI:10.1016/j.ejmech.2014.05.003

European Journal of Medicinal Chemistry 81 (2014) 1e14
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Structureeactivity relationships and molecular modelling of new
5-arylidene-4-thiazolidinone derivatives as aldose reductase
inhibitors and potential anti-inflammatory agents
,
,
,
c
Rosanna Maccari a 1, Rosa Maria Vitale b 1, Rosaria Ottanà a , Marco Rocchiccioli ,
Agostino Marrazzo d, Venera Cardile e, Adriana Carol Eleonora Graziano e, Pietro Amodeo b,
Umberto Mura c, Antonella Del Corso c
*
a Dipartimento di Scienze del Farmaco e dei Prodotti per la Salute, University of Messina, Polo Universitario dell’Annunziata, 98168 Messina, Italy
b Istituto di Chimica Biomolecolare (ICB) e CNR, Comprensorio Olivetti, Ed. 70, Via Campi Flegrei 34, 80078 Pozzuoli, Napoli, Italy
c Dipartimento di Biologia, Unità di Biochimica, University of Pisa, Via S. Zeno 51, 56126 Pisa, Italy
d Department of Drug Sciences, Medicinal Chemistry Section, University of Catania, Viale A. Doria 6, 95125 Catania, Italy
e Department of Bio-medical Sciences, Section of Physiology, University of Catania, V.le A. Doria 6, 95125 Catania, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of 5-(carbamoylmethoxy)benzylidene-2-oxo/thioxo-4-thiazolidinone derivatives (6e9) were
synthesized as inhibitors of aldose reductase (AR), enzyme which plays a crucial role in the development
of diabetes complications as well as in the inflammatory processes associated both to diabetes mellitus
and to other pathologies.
In vitro inhibitory activity indicated that compounds 6e9aed were generally good AR inhibitors.
Acetic acid derivatives 8aed and 9aed were shown to be the best enzyme inhibitors among the tested
compounds endowed with significant inhibitory ability levels reaching submicromolar IC50 values.
Moreover, some representative AR inhibitors (7a, 7c, 9a, 9c, 9d) were assayed in cultures of human
keratinocytes in order to evaluate their capability to reduce NF-kB activation and iNOS expression.
Compound 9c proved to be the best derivative endowed with both interesting AR inhibitory effectiveness
and ability to reduce NF-kB activation and iNOS expression.
Received 23 December 2013
Received in revised form
28 April 2014
Accepted 1 May 2014
Available online 4 May 2014
Keywords:
Aldose reductase
Enzyme inhibitors
Anti-inflammatory agents
Molecular modelling
5-Arylidene-4-thiazolidinone derivatives
5-(carbamoylmethoxy)benzylidene-2-oxo/
thioxo-4-thiazolidinone derivatives
Molecular docking and molecular dynamics simulations were undertaken to investigate the binding
modes of selected compounds into the active site of AR in order to rationalize the inhibitory effectiveness
of these derivatives.
Ó 2014 Elsevier Masson SAS. All rights reserved.
1. Introduction
was established that AR plays a crucial role in the development of
chronic diabetic complications, such as neuropathy, nephropathy,
Aldose reductase (AR) is an aldo-keto reductase present in the
cytoplasm of most human cells. During the last three decades, a
great deal of attention has been focused on this enzyme, since it
cataracts, retinopathy, accelerated atherosclerosis and increased
cardiovascular risk. Therefore, this enzyme has been assumed as
an attractive target to prevent or treat these invalidating pa-
thologies which are a serious health threat for diabetic patients
[1e3]
AR is the first and rate-limiting enzyme of the polyol pathway in
which it catalyses the NADPH-dependent reduction of glucose to
sorbitol. Subsequently, sorbitol dehydrogenase converts sorbitol
into fructose with concomitant reduction of NADþ. Under normal
glycaemic conditions, glucose is predominantly converted to
glucose-6-phosphate by hexokinase and then enters the glycolytic
pathway, whereas it is metabolized through the polyol pathway
only in small amount, because AR has a relatively low affinity for
this substrate. The major physiological role of AR appears to be the
Abbreviations: AGEs, advanced glycation-end products; AR, aldose reductase;
ARI, aldose reductase inhibitor; COX-2, cyclooxygenase-2; DHN, 1,4-
dihydroxynonene; DM, diabetes mellitus; GS-HNE, glutathione-4-hydroxy-trans-
2-nonenal; HNE, 4-hydroxy-trans-2-nonenale; IFN-
nuclear factor-kB; NOS, nitric oxide synthase; PKC, protein kinase C; ROS, radical
oxygen species; TNF- , tumour necrosis factor-alfa.
g, interferon-gamma; NF-kB,
a
* Corresponding author.
E-mail address: rottana@unime.it (R. Ottanà).
These Authors contributed equally to this work.
1
http://dx.doi.org/10.1016/j.ejmech.2014.05.003
0223-5234/Ó 2014 Elsevier Masson SAS. All rights reserved.

2
R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
removal of potentially toxic aldehydes generated during lipid per-
oxidation, such as 4-hydroxy-trans-2-nonenal (HNE), and their
glutathione adducts (GS-HNE), which are reduced more efficiently
than glucose by AR [4,5].
apoptosis, cell proliferation and angiogenesis, resulting in the
synthesis of cytokines, growth factors, tumour necrosis factor-alfa
(TNF-a), matrix metalloproteases, cyclooxygenase-2 (COX-2) and
inducible NOS (iNOS) [4,5,11e13]. The latter can be induced by
other transcription factors in addition to NF-kB [14]. Both AR-
derived PKC activation and production of AGEs also activate NF-
In contrast, under hyperglycaemic conditions, such as in dia-
betes mellitus (DM), up to one third of the total metabolized
amount of glucose is shunted to the polyol pathway, in tissues
possessing insulin-independent uptake of glucose, such as kidney,
lens, retina, peripheral nerves [1]. The increased flux of glucose
through the polyol pathway triggers multiple pathogenic mecha-
nisms which underlie the development of degenerative diabetic
complications. In fact, the intracellular accumulation of sorbitol,
which is unable to cross cellular membranes, leads to osmotic
imbalance, cell swelling and membrane permeability changes,
mainly in the lens. Moreover, NADPH and NADþ deprivation is
responsible for alterations in cellular redox potentials and,
concomitantly, impairs the activity of enzymes such as nitric oxide
synthase (NOS) and glutathione reductase. On the whole, these
alterations result in cellular oxidative stress, as a consequence of
the imbalance between increased production of radical oxygen
species (ROS) and reduced intracellular antioxidant defense [1,4].
In addition, the increase in the levels of glycating agents such as
fructose (end-product of the polyol pathway), dihydroxyacetone
phosphate, glyceraldehyde-3-phosphate and methylglyoxal
(deriving from increased glycolitic flux and fatty acid oxidation),
enhances the intracellular formation of advanced glycation-end
products (AGEs) [1]. AGEs in turn can cause pathological alter-
ations in the functions of intracellular proteins and lead to further
generation of ROS. Moreover, dihydroxyacetone phosphate can be
converted to diacylglycerol which also contributes to increased
oxidative stress through the abnormal activation of protein kinase C
(PKC) [1].
All these AR-mediated biochemical alterations originate
vascular damage and decreased nerve conduction velocity that
result in microvascular and macrovascular pathologies as well as
neuropathy. On this basis, the inhibition of AR may be an optimal
strategy to counteract the development of chronic pathologies
associated to DM, by blocking the activity of this central enzyme
that mediates hyperglycemia-induced oxidative stress and conse-
quent tissue and vascular alterations. Indeed, studies in animal
models and in humans showed that AR inhibitors (ARIs) can pre-
vent or slow down the progression of certain DM-associated dis-
orders [6e8].
Unfortunately, despite the wide number of ARIs reported
over the last three decades, clinical trials often provided
disappointing results due mainly to pharmacokinetic draw-
backs, adverse side effects or low in vivo efficacy. Currently,
epalrestat (Fig. 1) is the only ARI available for therapy, whereas
several other inhibitors, such as fidarestat (Fig. 1), have entered
clinical trials [2,3,9,10].
k
B [1].
Accordingly, it was ascertained that AR inhibition prevents
multiple inflammatory pathways [15e17]. Chronic, often sub-
clinical, inflammation is a key component of both DM and its
complications and, in particular, is closely related to microvascular
and macrovascular damage. Thus, these findings further highlight
the crucial role that AR inhibition could play in the management of
DM-associated chronic pathologies.
AR was also shown to be overexpressed in different types of
human cancers, such as colon, lung, breast tumours, and to play a
critical role in mediating ROS-induced carcinogenesis [18]. In
addition, it was demonstrated that AR can mediate myocardial
ischaemic injury both in diabetic and normoglycemic animals
[19,20].
On the whole, these findings suggest that ARIs could offer new
therapeutic strategies for the treatment of various pathologies,
besides DM and its complications, such as rheumatoid arthritis,
sepsis, asthma, uveitis, atherosclerosis [11,16].
At the same time, new approaches have been stimulated to test
ARIs in vivo by designing new trails suitable to assess the efficacy of
these agents in preventing or counteracting chronic inflammation
and vascular damage [4,21].
All these findings prompted us to pursue our search for new
ARIs, which in the last few years allowed us to identify 5-
arylidene-4-thiazolidinone derivatives (including compounds
1e4, Fig. 1) as potent in vitro ARIs, some of which also endowed
with antioxidant properties [22e28]. In our previous studies, we
ascertained that the presence of an acetic chain on N-3 of the
thiazolidinone ring is related to the highest levels of AR inhibi-
tory activity, whereas the substituents inserted on the 5-
arylidene moiety can modulate the effectiveness of these ARIs
by interacting with the so-called “specificity pocket” of the AR
active site [29]. In particular, the insertion of a carboxymethoxy
chain on the 5-benzylidene ring was generally shown to be
beneficial for activity by critically intervening in the binding to
AR [23].
In addition, we found that the replacement of the carbonyl
group in position 2 of the 2,4-thiazolidinedione scaffold of these
previously reported ARIs with the bioisoster thiocarbonyl group
can improve inhibitory efficacy [27].
In continuing our search, herein we report the synthesis,
structureeactivity relationships and docking studies of 5-(carba-
moylmethoxy)benzylidene-2-oxo/thioxo-4-thiazolidinone
de-
rivatives (6e9, Fig. 2). They were designed by assuming the
carbamoylmethoxy group as a non ionisable isoster of the carbox-
ymethoxy chain. This replacement may improve the cellular
permeability and, at the same time, favour the binding to AR active
site via hydrogen bonds, as suggested by our previous docking
Recently, advances in the understanding of the pathophysio-
logical actions of AR have resulted in a renewed interest in the
search for more efficacious and safer ARIs. In particular, it was
established that AR is critically involved in inflammatory pro-
cesses under both normoglycemic and diabetic conditions. This
enzyme, which is overexpressed under various oxidative condi-
tions, intervenes in multiple signalling pathways leading to
inflammation and tissue degeneration. In fact, ROS determine
lipid peroxidation and production of hydrophobic lipid-derived
aldehydes, such as HNE and its glutathione adduct GS-HNE,
which are reduced by AR by originating 1,4-dihydroxynonene
(DHN) and GS-DHN [4,5]. This latter has been identified as a
pivotal mediator of oxidative stress-induced activation of nuclear
transcription factor-kappa B (NF-kB), which in turn regulates the
transcription of numerous genes involved in inflammation,
studies
performed
with
analogous
2-(5-arylidene-2,4-
dioxothiazolidin-3-yl)acetamides (4, Fig. 1) [24].
Compounds 6e9 were evaluated as in vitro inhibitors of AR
isolated from bovine lens. This enzyme displays a sequence ho-
mology of more than 86% with human AR [30]. In order to assess
their anti-inflammatory potential, they were also assayed in cul-
tures of human keratinocytes by evaluating their capability to
reduce NF-kB activation and iNOS expression.
Molecular docking simulations into the AR active site were
carried out in order to gain new insight concerning the possible
binding modes of this class of ARIs.

R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
3
Fig. 1. Structures of some 4-thiazolidinone derivatives 1 [23], 2, 3 [27], 4 [24] active as AR inhibitors and of some known ARIs.
2. Results and discussion
and arylaldehydes 5aed in refluxing ethanol in the presence of
piperidine as base (Scheme 1). The reaction between commercial 2-
thioxo-4-thiazolidinone and suitable aldehyde (5aed) in glacial
acetic acid and in the presence of sodium acetate at reflux provided
2.1. Chemistry
Compounds 6aed were synthesized in good yield via Knoeve-
pure
2-[(4-oxo-2-thioxothiazolidin-5-ylidenemethyl)phenoxy]
nagel condensation starting from commercial 2,4-thiazolidinedione
acetamides (7aed) in good yield (Scheme 2).

4
R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
X
S
R
O
N
R'
R''
6-9 a-d
R’
H
R’’
OCH2CONH
6 R = H, X = O
7 R = H, X = S
a
b
c
OCH2CONH H
OCH3
OCH2CONH OCH3
8 R = CH2COOH, X = O
9 R = CH2COOH, X = S
OCH2CONH
d
Fig. 2. Structures of compounds 6e9.
Although commercially available, 2-(formylaryloxy)acetamides
The multi-step synthesis of (5-arylidene-2,4-dioxothiazolidin-
3-yl)acetic acids 8aed is depicted in Scheme 1. A mixture of com-
mercial 2,4-thiazolidinedione and methyl bromoacetate in the
presence of potassium carbonate in boiling acetone afforded (2,4-
dioxothiazolidin-3-yl)acetic acid methyl ester which was
(5aed) were prepared by reaction of the appropriate hydrox-
ybenzaldehyde and 2-chloroacetamide in refluxing acetonitrile in
the presence of potassium carbonate (Scheme 1) (see
Supplementary data).
CHO
CHO
i
R''
OH
R''
R'
5a-d
O
H
N
S
O
ii
R'
R''
6a-d
O
H
N
S
O
iii
O
COOH
O
COOH
N
N
ii
S
O
S
O
R'
R''
8a-d
Scheme 1. Synthesis of compounds 6aed and 8aed. Reagents and conditions: i) ClCH2CONH2, K2CO3, CH3CN,
acetone, ; b) CH3COOH, HCl,
D; ii) Ar-CHO (5aed), C5H11N, EtOH, D; iii) a) BrCH2COOCH3, K2CO3,
D
D.

R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
5
acids 8 and 9 the methylene carbon on N-3 resonated at 42.2e42.3
and 45.0e45.1, respectively.
R
CHO
S
R
S
N
N
On the basis of previously acquired X-ray diffraction analysis
[28,31] as well as NMR spectroscopy data of analogous derivatives
[22e28,31], the newly synthesized compounds 6e9 were obtained
solely as Z diastereoisomers. In fact, NMR spectra showed in all
cases only one set of signals and no isomerization at the C]C
exocyclic double bond was subsequently observed.
i
S
+
O
S
O
R''
R'
R'
5a-d
R''
7a-d R = H
9a-d R = CH2COOH
2.2. Aldose reductase in vitro inhibition
The in vitro AR inhibitory activity of compounds 6e9 aed (Ta-
bles 1 and 2) was evaluated by using highly purified AR from
bovine lenses [32]. Sorbinil and epalrestat (Fig.1) were employed as
reference drugs.
Scheme 2. Synthesis of compounds 7aed and 9aed. Reagents and conditions: i)
CH3COONa, CH3COOH,
D.
All tested compounds 6e9 were shown to be generally inter-
esting AR inhibitors reaching up to submicromolar IC50 values.
Entirely consistent with previous SARs outlined for this class of ARIs
[22e28], (5-arylidene-2,4-dioxothiazolidin-3-yl)acetic acids (8ae
d) and corresponding 2-thioxo isosters 9aed were showed to be
the best ARIs among the tested compounds.
hydrolysed to corresponding acetic acid 10 by refluxing in a
mixture of glacial acetic acid and HCl [22]. Finally a refluxing
mixture of compound 10 and suitable aldehyde (5aed) in the
presence of piperidine in ethanol afforded compounds 8ae
d (Scheme 1).
(5-Arylidene-4-oxo-2-thioxothiazolidin-3-yl)acetic acids (9ae
d) were obtained by the reaction of commercial (4-oxo-2-
thioxothiazolidin-3-yl)acetic acid and appropriate aldehyde (5ae
d) in refluxing glacial acetic acid and sodium acetate mixture
(Scheme 2).
The structure of all synthesized compounds was unambiguously
assigned on the basis of analytical and 1H and 13C NMR spectros-
copy data. In the 1H NMR spectra, the presence of one singlet at
7.55e7.88 ppm attributable to the CH methylidene proton proved to
be significant in assigning the structure of compounds 6e9. In
addition the resonance of the methylene protons of the carba-
Moreover (5-arylidene-4-oxo-2-thioxothiazolidin-3-yl)acetic
acids (9aed) were even more active than both reference drugs by
inhibiting the enzyme with IC50 values ranging from 0.11
0.13 M (Table 2).
mM to
m
The acetic acid group on N-3 once again proved to play a critical
role in the binding of compounds 8 and 9 to the AR catalytic site by
resulting in the reduction of IC50 values up to one hundred times in
comparison with N-unsubstituted compounds 6 and 7. Among
acetic acid inhibitors (compounds 8 and 9) the differences of IC50
values were negligible, except for compound 8a which showed to
be twice less active than analogues 8bed. Compounds 6 and 7
produced level of enzyme inhibition in a wider range between
8.65 mM and 33.9 mM. These results highlight the pivotal role of the
acetic group that thwarts the effects exerted by the substituents on
5-arylidene moiety.
Moreover the enzymatic inhibition results suggested that the 2-
thioxo-4-thiazolidinone core (compounds 7aed and 9aed) may
establish more favourable interactions with the enzyme in com-
parison to the isosteric 2,4-thiazolidinedione one (compounds 6ae
moylmethoxy group resulted in
a singlet ranging between
4.48 ppm and 4.73 ppm. In the 1H NMR spectra of compounds 8 and
9 a diagnostic singlet appears attributable to the methylene protons
on N-3 resonating at chemical shifts ranging between 4.34 and
4.55 ppm. In 13C NMR spectra of compounds 6e9, the methylene
carbon of the carbamoylmethoxy group gave rise to a signal in the
range from 66.6 to 67.6 ppm. Besides, in 13C NMR spectra of acetic
Table 1
In vitro bovine lens AR inhibitory activity of compounds 6aed and 7aed.
Table 2
In vitro bovine lens AR inhibitory activity of compounds 8aed and 9aed.
Compd
X
R0
R00
IC50 (mM)
Compd
X
R0
R00
IC50 (mM)
6a
6b
6c
6d
7a
7b
7c
7d
O
O
O
O
S
S
s
S
H
OCH2CONH2
H
OCH2CONH2
OCH3
OCH2CONH2
H
OCH2CONH2
OCH3
21.5 (14e29)
40%a
22.6 (12.1e33.1)
33.9 (21.9e45.9)
8.65 (5.70e11.6)
34%a
9.20 (7.70e10.70)
11.6 (5.23e18.0)
2.0 (1.70e3.50)
0.17 (0.033e0.16)
OCH2CONH2
OCH3
OCH2CONH2
H
OCH2CONH2
OCH3
OCH2CONH2
8a
8b
8c
8d
9a
9b
9c
9d
O
O
O
O
S
S
S
S
H
OCH2CONH2
H
OCH2CONH2
OCH3
OCH2CONH2
H
OCH2CONH2
OCH3
1.01 (0.71e1.43)
0.56 (0.41e0.76)
0.52 (0.39e0.67)
0.50 (0.43e0.58)
0.13 (0.07e0.18)
0.13 (0.06e0.21)
0.13 (0.09e0.17)
0.11 (0.08e0.15)
2.0 (1.70e3.50)
0.17 (0.03e0.16)
OCH2CONH2
OCH3
OCH2CONH2
H
OCH2CONH2
OCH3
OCH2CONH2
Sorbinil
Epalrestat
Sorbinil
Epalrestat
a
% inhibition at 50 mM.

6
R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
d and 8aed), being equal the substituent in the 5-arylidene moiety
(Tables 1 and 2).
thiazolidinone scaffold, exhibited the worst inhibitory properties
among all studied compounds, but, upon introduction of a methoxy
group in position 3 of the 5-arylidene ring, corresponding com-
pound 7d substantially recovered the activity of series 7. In this view,
the comparison between the docking properties of 7b and 7d also
provided an assessmentof the abilityof the computational approach
in discriminating between “good” and “bad” inhibitors.
The removal of the substituent on N-3 stresses the role played
by the 5-arylidene moiety in the interaction with the enzyme.
Comparing the inhibitory effectiveness of compounds 6 and 7
demonstrated that the carbamoylmethoxy group plays a critical
role when inserted in the meta position. In fact compounds 6a, 6c,
7a and 7c showed to be better inhibitors than the corresponding
para substituted analogues (6b, 6d, 7b and 7d).
The introduction of the methoxy group also resulted to be
favourable particularly in the meta position. Compounds 6d and 7d,
3-methoxy-4-carbamoylmethoxybenzylidene substituted (IC50
AR binding site comprises two sub-pockets, the so-called
“anionic binding site” encompassing the residues involved in
catalysis (Tyr48, Lys77, and His110) along with the nicotinamide
moiety of the cofactor NADPþ, and the “specificity pocket”, formed
by Trp111, Phe122, Trp219, Cys298, Ala299, Leu300, Ser302 and
Cys303 [29]. Since the “specificity pocket” can adopt different
conformations depending on the shape of the bound ligand, mainly
due to the flexibility of the Cys298eCys303 region, the three most
diverging X-ray AR structures (from complexes with Fidarestat,
Tolrestat and IDD594, respectively, Fig. 1) were used in molecular
docking calculations. MD simulations of selected complexes (see
details in the experimental section) were then performed both to
assess the overall stability of the binding modes and to “merge” the
different protein starting conformations, allowing protein rear-
rangements. The emerging features from a compared analysis of all
the selected and refined complex structures allowed the classifi-
cation of the binding modes in families and subfamilies, based on
the occurrence of:
33.9
with compounds 6b and 7b, 4-carbamoylmethoxybenzylidene
substituted (40% and 34% inhibition at 50 M, respectively),
mM and 11.6 mM, respectively), were more active in comparison
m
whereas the effect of the methoxy group in the para position
(compounds 6c and 7c) was showed to be indifferent on enzymatic
affinity (Table 1).
The replacement of 3/4-carboxymethoxy group (compounds
1aed, Fig. 1) [23] with the non ionisable carbamoylmethoxy one
(compounds 8aed) on 5-arylidene moiety resulted generally in
reduction of inhibitory activity (Table 2). Although compounds 8ae
d exerted interesting AR inhibition at submicromolar dose, they
proved to be generally up to five times less active than carbox-
ymethoxy isosters 1aed.
Starting
methoxybenzylidene)-2,4-dioxothiazolidin-3-yl]acetic acid (1e,
IC50 ¼ 0.48 M), the insertion of the carboxymethoxy group in para
position was favourable for the AR inhibition providing compound
1d endowed with IC50 ¼ 0.26 M (Fig. 1) [23], while the corre-
sponding carbamoyl substituted analogue maintained the same
level of activity (compound 8d IC50 ¼ 0.50 M).
from
the
previously
reported
[5-(3-
1) For all ligands, two main docking modes substantially inde-
pendent on enzyme conformations and dependent on the in-
hibitor part that points toward the anionic binding site, either
the thiazolidinone ring (hereinafter referred to as RING mode),
or the amide group (AMD mode);
2) For compounds 7 and 9, featuring a thiocarbonyl group in po-
sition 2 and a carbonyl group in position 4, two ring orientations,
with the carbonylic oxygen atom pointing alternatively inward
(RING-Oin) or outward (RING-Oout) respect to the enzyme
binding pocket;
3) For those ligands such as 7a and 6a bearing the carbamoylme-
thoxy substituent in the meta position of the 5-arylidene ring,
two different aryl group orientations (observed after MD, since
docking provided somewhat intermediate orientations),
referred to as AMD-m1 and AMD-m2, while in 7b, where the
amide group is present in para position, only a single orienta-
tion, referred to as AMD-p, was detected;
4) In all AMD poses, two possible amide group rotameric states,
giving rise to two different H-bond patterns between the polar
atoms of 3-carbamoylmethoxy group and protein Tyr48/His110/
Trp111 side-chains, featuring contacts either of inhibitor ether
oxygen with Trp111 and inhibitor carbonyl oxygen with Tyr48
(AMD-x-1, where x ¼ m1, m2 or p), or of inhibitor carbamoyl
oxygen with protein Tyr48 and ligand carbamoyl hydrogen with
protein His110 (AMD-x-2).
m
m
m
Comparing the activity of 4/3-hydroxybenzylidene analogues 1f
and 1g (Fig. 1) with the corresponding carboxymethoxy (1c and 1d)
and carbamoylmethoxy (8c and 8d) derivatives highlights that the
introduction of the carboxymethoxy group on the 5-arylidene
moiety (1c, 1d) is advantageous for the interaction with the
enzyme determining a reduction of the IC50 values of about two
times [23], while its replacement with the carbamoylmethylene
group is indifferent on inhibitory effectiveness (8c, 8d).
In the series of 2-thioxo-4-thiazolidinone derivatives 7 and 9,
the introduction of the carbamoylmethoxy group in the para or
meta position of previously reported compounds 2, 3a, 3b (Fig. 1)
[27] was either indifferent or unfavourable for the inhibitory
effectiveness. In fact the inhibitory effectiveness of compounds 2
(IC50 ¼ 0.15
m
M) and 9d (IC50 ¼ 0.11
mM) was comparable while,
among N-unsubstituted analogues (compounds 7), the introduc-
tion of the carbamoylmethoxy group was detrimental for the
enzyme affinity (cf. compounds 3a and 7d).
2.3. Molecular modelling
A combined approach of molecular docking and molecular dy-
namics (MD) simulations was undertaken to investigate the binding
modes of compounds 6, 7 and 9 into the active site of AR in order to
rationalize the inhibitory effectiveness of these series of compounds
which bear the same substituents on the 5-arylidene moiety but
differ each other either for the presence of a carboxylic acid group on
N-3 (9 vs. 7), or for the nature of the heterocyclic ring (6 vs. 7).
Docking calculations were performed on the best inhibitors among
2-thioxo-4-thiazolidinones 7 and 9, either containing the acid acetic
group in position 3 (compound 9d) or N-unsubstituted (compound
7a), and among N-unsubstituted 2,4-thiazolidinediones 6 (com-
pound 6a). Compounds 7b and 7d were also evaluated since com-
pound 7b, despite containing the favourable 2-thioxo-4-
Among N-unsubstituted inhibitors (6a, 7a, and 7b) subsequent
MD simulations were performed on both RING and AMD modes.
However, for compounds 7, although both RING-Oin and RING-
Oout orientations were simulated, only the former will be dis-
cussed, since it gave rise to H-bonds with the active site not
observed in the latter. For compound 9d only the RING mode was
simulated, since the carboxylate moiety in position 3 of 2-thioxo4-
thiazolidinone ring clearly can not form any ionic interaction with
the active site when in the AMD orientation.
2.3.1. ARe7a complex
A MD trajectory analysis of ARe7a complexes shows that both
RING and AMD orientations are stable over the entire simulated

R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
7
Table 3
period (20ns). In the RING pose the ligand formed stable H-bonds
with His110 and Trp111, through NeH and the carbonyl group
respectively. The exocyclic sulphur atom was involved in van der
Waals interactions with Val47, Tyr48 and Trp20. The aryl group
formed stable H-bond interactions through both amide and ether
oxygens with backbone NH atoms of Ala299 and Leu300, respec-
tively, thus anchoring the arylidene group to the Cys298eLeu301
loop of the specificity pocket. In this orientation, the benzylidene
ring lay between Trp219 and Leu300 side-chains. During MD
simulation a water molecule got trapped and became resident in
the anionic binding site, by forming a stable H-bond with Tyr48
(Fig. 3A). The presence of a resident water molecule was still
observed after a MD resubmission with a different random seed,
thus suggesting that it may represent a characteristic feature of this
system. The interaction energy between the resident water mole-
cule and the proteineligand complex, calculated on average on the
last 10 ns of trajectory, was ꢀ13.9 ꢁ 1.9 kcal molꢀ1, thus contrib-
uting appreciably to the overall complex stabilization (Table 3).
Mean total AReligand interaction energies (kcal molꢀ1) calculated over last 10 ns of
MD simulations for different orientations of 9d, 7a, 7b, 7d and 6a in their complexes
with AR.a
Ligand
Orientation
Mean proteineligand interaction energy (kcal molꢀ1
)
6a
RING
ꢀ51.95 (ꢁ6.24)
ꢀ41.08 (ꢁ4.27)
ꢀ52.04 (ꢁ3.25)
ꢀ50.92 (ꢁ3.97)
ꢀ13.94 (ꢁ 1.87) water
ꢀ42.77 (ꢁ3.28)
ꢀ54.27 (ꢁ3.05)
ꢀ41.20 (ꢁ3.80)
ꢀ40.03 (ꢁ3.36)
ꢀ46.55 (ꢁ7.39)
ꢀ46.20 (ꢁ4.84)
ꢀ120.98 (ꢁ9.17)
ꢀ117.36 (ꢁ9.60)
AMD-m1
AMD-m2
RING
7a
AMD-m1
AMD-m2
AMD-p-1
AMD-p-2
RING
AMD-p
RING-Oin
RING-Oout
7b
7d
9d
a
In parentheses the corresponding rmsd values are shown. For the RING orien-
tation of ARe7a the total interaction energy of a resident water molecule with AR
and ligand (italicized) is also reported. Herein, AR also includes NADPþ
.
In the AMD orientation, the two different arrangements of the
m-substituted aryl group exhibited considerable energetic differ-
ences, AMD-m1 being less stable than AMD-m2 (Table 3). In fact,
while in AMD-m1 compound 7a only formed hydrophobic contacts
with Leu300 and Cys298 side-chains and a
2-thioxo-4-thiazolidinone ring and Phe122 side-chain, in AMD-m2,
in addition to a -stacking of the heterocycle with Trp219 and to
p-stacking between the
p
hydrophobic contacts of the 5-arylidene group with Trp20 and
Leu300 side-chains, the esocyclic oxygen atom in position 4
engaged stable H-bonds with the backbone NH of Ala299 and
Leu300 (Fig. 4A). On the contrary, proteineligand interaction en-
ergies did not exhibit appreciable differences between AMD-x-1
and AMD-x-2, since the interactions of the amide group with the
protein were almost isoenergetic between the two modes and the
amide orientation itself did not affect other interactions of in-
hibitors (data not shown). During MD simulations AMD-x-1/AMD-
x-2 interconversion was also observed.
2.3.2. ARe7b complex
Differently from what observed for ARe7a complex, the RING
orientation of compound 7b was not stable during MD simulation,
where a lost of stabilizing H-bonds in the anionic binding site was
observed (Fig. 5). The simulation was repeated using another ligand
RING pose and also in this case the complex dissociated during MD
(data not shown). Conversely, both AMD poses remained stable
over the same simulated period (with their aryl group adopting the
AMD-p orientation) resembling 7a AMD-m2 rather than AMD-m1
one because of the observed
p-stacking of the heterocycle with
Trp219 (Fig. 4B). However, this pose was less stable than the cor-
responding AMD-m2 observed for 7a since the different position of
the functional group in the aryl moiety did not allow the occurrence
of stabilizing H-bonds with the protein (Table 3).
2.3.3. ARe7d complex
Compound 7d was stable in both RING and AMD orientations. In
the RING pose, the aryl group formed alternatively p-stacking with
Phe122 or with Trp219, where the methoxy group engaged H-bond
interactions with backbone NH atoms of Ala299 or Leu300, simi-
larly to the carbamoyl group of 7a (Fig. 3B). In AMD orientation, the
aryl group of 7d adopted the AMD-p orientation already described
for 7b, but with additional hydrophobic contacts formed by the
methoxy group with Trp20, Val47 and Phe122 that provided an
additional complex stabilization of almost 5 kcal molꢀ1 over 7b
(Fig. 4C).
Fig. 3. Views of the binding site in modelled AReligand complexes after 20 ns of MD.
The RING modes for 7a (A), 7d (B) and 6a (C) are shown using a beige ribbon plus
selected (within 5 A from ligand) stick sidechain representation for AR, pink sticks for
ꢀ
NADPþ and cyan ball-and-stick for ligands. For all molecules only polar hydrogens are
shown. Heteroatoms and polar hydrogens are coloured according to a standard scheme
(red for oxygen, blue for nitrogen, yellow for sulphur and white for hydrogen). Inter-
molecular H-bonds are shown as green dotted-lines. (For interpretation of the refer-
ences to colour in this figure legend, the reader is referred to the web version of this
article.)

8
R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
Fig. 4. Views of the binding site in modelled AReligand complexes after 20 ns of MD. The AMD modes for 7a (A), 7b (B), 7d (C) and 6a (D) are shown using a beige ribbon plus
ꢀ
selected (within 5 A from ligand) stick sidechain representation for AR, pink sticks for NADPþ and cyan ball-and-stick for ligands. For all molecules only polar hydrogens are shown.
Heteroatoms and polar hydrogens are coloured according to a standard scheme (red for oxygen, blue for nitrogen, yellow for sulphur and white for hydrogen). Intermolecular H-
bonds are shown as green dotted-lines. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
2.3.4. ARe6a complex
bond pattern involving the 3-carbamoylmethoxy group. Instead,
the AMD orientations in ARe6a complex are similar to those
already described for 7a (Fig. 4D).
Trajectory analysis of RING mode for ARe6a complex showed a
different arrangement of the 2,4-thiazolidinedione moiety in
comparison with the 2-thioxo-4-thiazolidinone one of 7a. Besides
the H-bonds with His110 and Trp111, shared by both compounds,
the exocyclic oxygen atom in position 2 of 6a formed an additional
H-bond with Tyr48 side-chain, which, in ARe7a complex, is H-
bonded to a resident water molecule (Fig. 3C). To exclude that 7a
can adopt this binding mode, a new MD simulation of ARe7a
complex was carried out, by modelling the binding mode of com-
pound 7a on the final configuration of the ARe6a complex, but also
in this case the model evolved toward the arrangement described
above for ARe7a. As a result of the different heterocycle arrange-
ments, the arylidene moiety is more flexible in ARe6a complex
than in ARe7a, oscillating between the interaction with Phe122
and the loop region Ala299-Leu301, in the absence of a stable H-
2.3.5. ARe9d complex
As expected, 9d, the representative inhibitor selected among
compounds 9aed, bearing a negatively-charged carboxylate group
in position 3, exhibited an interaction energy almost twice than
that of the neutral N-unsubstituted analogue (7d). The ionizable
group stably anchored the ligand in the anionic binding site,
through stable H-bond and ionic interaction, respectively with
Tyr48/His110 side-chains and the positively charged nicotinamide
moiety of the cofactor. This interaction is, in principle, fully
compatible with both RING-Oin and RING-Oout orientations of the
2-thioxo-4-thiazolidinone scaffold in the binding pocket. However,
as shown in Table 3, RING-Oin of ARe9d complex was more stable
by 3 kcal molꢀ1 than RING-Oout, because of the different orienta-
tion of the arylidene moiety, which in RING-Oin formed a persistent
H-bond through the ether oxygen of the 4-carbamoylmethoxy
group with NH backbone atom of Leu301. This interaction, in
turn, stabilized a network of hydrophobic interactions involving the
3-methoxy group of the inhibitor and Ala299, Leu301 and Trp219
side-chains of AR (Fig. 6A). In RING-Oout the aforementioned stable
H-bond was missing, replaced by a set of labile concurrent H-bonds,
which, by increasing local positional fluctuations, also negatively
affected the persistence of the hydrophobic interactions (Fig. 6B).
The most evident general trend among the different outcomes
obtained by combining the results of our computational approach
with the corresponding experimental activity data is the over-
whelming contribution of the carboxylic group of compounds 8 and
9 to AR binding. In fact, on one hand only negligible differences in
potency were observed in experimental data within compounds
9aed, and on the other hand MD simulations showed that the
stability of the different ARe9d binding modes is relatively unaf-
fected by the different orientation and interaction patterns of other
functional groups of the inhibitor.
Conversely, both 5-arylidene-2-thioxo-4-thiazolidinones (7)
and 5-arylidene-2,4-thiazolidinediones (6), which are lacking in
the carboxylic acid group on N-3, are very sensitive to both the
nature and position of the substituents on the aryl group, showing
similar trends within each series. For neutral compounds 6 and 7
the two main binding mode classes, both the so-called RING and
Fig. 5. Time-dependence of H-bond distances along MD simulation of the ARe7b RING
binding mode. Distances (in A) are plotted as a function of time (ns) for 7b (N3)-
AR(H110:Nε) (red line) and 7b (O4)-AR(W111:Nε1) (blue line) atom pairs. (For inter-
pretation of the references to colour in this figure legend, the reader is referred to the
web version of this article.)
ꢀ

R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
9
the interaction patterns and energies of the RING-mode complexes
of 7a and 6a were compared in detail, it emerged that the lack of a
H-bond was more than compensated by the extra-stabilization
provided by the stable interaction of a water molecule with Tyr48
(H-bond) and the inhibitor (favourable electrostatic and van der
Waals interactions). However, a quantitative comparison between
predicted AR affinities of 7a and 6a is beyond the scope of the
present study, since it would require both the precise evaluation of
polarization effects upon oxygen/sulphur replacement, and the
comparison of binding free energies between a binary and a ternary
complex.
2.4. Antiinflammatory activity
In order to evaluate the effectiveness of this series of ARIs on
inflammatory pathways, 2-thioxo-4-thiazolidinone derivatives 7a,
7c, 9a, 9c, and 9d were selected as representative inhibitors among
either acetic acid derivatives (compounds 9) or N-unsubstituted
analogues (compounds 7).
The anti-inflammatory activity was evaluated in vitro on the
normal human keratinocyte cell line NCTC 2544 exposed to IFN-
g
and histamine, this latter used to strongly increase the IFN-
induced keratinocyte activation [33,34].
g-
The tested compounds did not interfere significantly with
cellular viability at the tested doses, maintaining the capability of
these cells to metabolize tetrazolium salts (data not shown).
As above described, NF-kB is highly activated at sites of
inflammation in diverse diseases and can induce transcription of
proinflammatory cytokines, chemokines, adhesion molecules,
MMPs, COX-2, and iNOS. In this study the levels of NF-kB and iNOS
proteins were determined [35] and reported in Figs. 7 and 8.
Compounds 9a, 9c, and 9d can reduce the activation of NF-kB
demonstrating to inhibit NF-kB nuclear translocation in stimu-
lated cells compared to the control (Fig. 7). On the contrary, N-
unsubstituted compound 7a appeared to be ineffective, whereas
compound 7c reduced only slightly the activation of NF-kB.
Therefore, the presence of the acetic acid group was also shown
to be favourable for the reduction of NF-kB activation. Interestingly,
Fig. 6. Ligandeprotein interactions in ARe9d complex models. The LigPlotþ [47]
representation, enriched with the NADPþeligand ionic interaction (violet dotted line
between ꢁ charge signs), is shown for Oin (A) and Oout (B) binding modes. Ligand and
protein residues involved in H-bonds (green dotted lines) are shown as ball-and-lines
(with violet and brown bonds for 9d and protein residues, respectively), while protein
residues and ligand atoms forming hydrophobic contacts are surrounded by purple
spoked arcs. (For interpretation of the references to colour in this figure legend, the
reader is referred to the web version of this article.)
AMD orientations, were generally accessible in term of stability
shown during MD simulations and protein-inhibitor energy inter-
action values, except for 7b. The RING binding mode of this latter
compound, which exhibited the lowest experimental activity
within the series, was not stable already on a nanosecond time-
scale. Compound 7d, which shares with 7b the carbamoylmethoxy
group in the para position of the arylidene group, was able to
recover binding in the RING mode through the presence of the
methoxy group in meta position, which mimicked part of the sta-
bilizing interactions occurring in the corresponding complex model
of 7a, the most active inhibitor within the N-unsubstituted series.
Thus, the comparable inhibitory effectiveness of compounds 7a
and 7d, significantly higher than that of 7b, can easily be ratio-
nalized in terms of a stronger binding provided by RING modes
(plus an eventual minor contribution from AMD) in comparison
with the weaker binding associated to the sole predicted AMD
mode for 7b.
The slight differences in potency observed between compounds
7a and 7d can also be explained on the basis of the overwhelming
contribution to binding provided by RING modes, since their pre-
dicted complex models shared substantially the same interaction
pattern in AMD mode while differed in RING orientation.
The replacement of an oxygen atom with sulphur promoted a
rearrangement of 2-thioxo-4-thiazolidinone ring in the anionic
binding site that prevented the formation of a direct H-bond with
Tyr48, instead occurring in the 2,4-thiazolidinedione series. When
Fig. 7. Effects of 2-thioxo-4-thiazolidinone derivatives 9a, 9c, 9d, 7a, and 7c on NF-kB
activation induced by interferon-g (I) plus histamine (H) on human keratinocytes NCTC
2544 determined by Western blot analysis. Data show the relative expression
(mean ꢁ SEM) of NF-kB calculated as arbitrary densitometric units (A.D.U.) collected
from three independent experiments. *p < 0.05 compared to interferon-
g (I) plus
histamine (H)-induced NF-kB.

10
R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
thiazolidinone
series
in
comparison
with
the
2,4-
thiazolidinedione one could be provided by polarization effects
involving the sulphur atom in the former series. However an ac-
curate evaluation of this energy term is beyond the scope of the
present work, since it requires extensive quantomechanic (QM) and
hybrid QM-molecular mechanics (MM) calculations, with a careful
definition of the QM part of the system and of the QM/MM
boundary, also keeping into account the observed occurrence of a
resident water molecule only in the 7a complex.
SARs combined with modelling studies indicated that carboxylic
acid derivatives 8 and 9 have a greater binding affinity with the
enzyme than N-unsubstituted counterpart 6 and 7 independent of
the nature of the group in position 2. Comparing the AR inhibitory
activity of N-unsubstituted compounds 6 and 7 and the corre-
sponding N-3 acetic acid substituted analogues (compounds 8 and
9) highlights the central role carried out by the ionisable carboxylic
group that can stably anchor the ligand in the anionic binding site
of the enzyme.
Interestingly, compounds 7a, 7c, 9a, 9c and 9d also exhibited
anti-inflammatory activity by reducing NF-kB activation and/or
iNOS expression in cultures of human keratinocytes exposed to
Fig. 8. Effects of derivatives 9a, 9c, 9d, 7a, and 7c on iNOS expression induced by
IFN-
g and histamine. In particular, compounds 9a and 9c were
interferon-g (I) plus histamine (H) on human keratinocytes NCTC 2544 determined by
Western blot analysis. Data show the relative expression (mean ꢁ SEM) of iNOS
shown to interfere in the cellular inflammatory pathways by
intervening in NF-kB activation, which is an upstream event in the
inflammatory cascade. Among the selected compounds, {5-[(3-
carbamoylmethoxy-4-methoxyphenyl)methylidene]-4-oxo-2-
thioxothiazolidin-3-yl}acetic acid (9c) stood out for its sub-
micromolar AR inhibitory efficacy joined to the ability to reduce
both NF-kB activation and iNOS expression. Its interesting biolog-
ical profile makes it worthy to be considered as a new lead com-
pound. The goal of further optimization will be supported by the
molecular modelling study which provided a rationalization of the
observed trend in experimental affinity data of this class of ARIs
and also suggested criteria to direct future drug design. Moreover,
in view of the recently proposed new approach in the inhibition of
AR [36] it will be relevant to assess the ability of these compounds
to act as differential AR inhibitors.
calculated as arbitrary densitometric units (A.D.U.) collected from three independent
experiments. *p < 0.05 compared to interferon-g(I) plus histamine (H)-induced iNOS.
compounds 9a, 9c and 9d, at 10 mM dose, showed activity compa-
rable to that of sorbinil, used as reference drug (Fig. 7).
As shown in Fig. 8, iNOS expression levels by human keratino-
cyte cells in the culture medium decreased in a dose-dependent
manner upon the treatment with the selected compounds, except
for compound 9a that resulted to be nearly ineffective. At the
highest dose (20 mM), compounds 7a, 7c and 9c were shown to
reduce the iNOS expression up to reaching the control level.
In our experimental model compound 9c displays the most
marked anti-inflammatory activity being able, at the dose of 10 mM,
to significantly reduce both NF-kB activation and iNOS expression.
Similar considerations apply to compound 7c, even though the
effect on NF-kB activation is less evident. More puzzling is the
behaviour of compounds 9a, 7a, and 9d, which are able to affect
only NF-kB activation or only iNOS expression, respectively. Thus,
overall an anti-inflammatory activity can be envisaged for all the
tested ARIs. Moreover, it is important to note that the inhibition of
4. Experimental section
4.1. Chemistry
Melting points were recorded on a Kofler hot-stage apparatus
and are uncorrected. TLC controls were carried out on precoated
silica gel plates (F 254 Merck). Elemental analyses (C, H, N), deter-
mined by means of a C. Erba mod. 1106 elem. Analyzer, were within
ꢁ0.4% of the theoretical values. 1H and 13C NMR spectra were
routinely recorded at a constant temperature of 27C on a Varian
Inova-200 spectrometer in CDCl3 or DMSO-d6 solution; chemical
the IFN-
g
and histamine-stimulated production of pro-
inflammatory molecules in keratinocytes NCTC 2544 by 7a, 7c,
9a, 9c, and 9d was not due to cytotoxicity, as assessed by MTT assay.
However, for a complete clarification of their mechanism of action,
further study are needed, in particular by evaluating the contri-
bution of the AR-dependent pathway to the overall inflammatory
signalling in the adopted experimental model.
shifts
d are expressed in ppm with reference to TMS as an internal
standard. Coupling constants (J) are given in hertz (Hz). 13C NMR
spectra were determined by Attached Proton Test (APT) experi-
ments and the resonances were always attributed by proton-
carbon heteronuclear chemical shift correlation.
3. Conclusions
In this paper we report the synthesis, biological activity, and
SARs of a new series of 4-thiazolidinone derivatives which differ by
substitutions at N-3 and C-2 positions as well as at 5-arylidene
moiety.
In vitro AR inhibitory activity indicated that the 2-thioxo-4-
thiazolidinone core is more favourable for the AR affinity in com-
parison with the 2,4-thiazolidinedione isosteric ring.
Unless stated otherwise, all materials were obtained from
commercial suppliers and used without further purification.
General procedure for the synthesis of 2-(formylaryloxy)acet-
amides 5aed, their chemicalephysical 1H and 13C data are reported
in the Supplementary material.
Molecular docking studies suggested a different binding mode
in the RING orientation between the two series, with a resident
water molecule contributing to an extra-stabilization in the case of
the 4-oxo-2-thioxothiazolidine core. A potentially relevant contri-
bution to the better activity exhibited by the 2-thioxo-4-
4.2. General procedure for the synthesis of 2-[(2,4-dioxothiazolidin-
5-ylidenemethyl)phenoxy]acetamides 6aed
A mixture of 2,4-thiazolidinedione (2.3 g, 20 mmol), appropriate
aldehyde 5 (20 mmol), piperidine (1.4 g,16 mmol) and EtOH (80 ml)

R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
11
was refluxed for 24e27 h. The reaction mixture was poured into
H2O acidified with AcOH to give a crude solid which was recrys-
tallized from methanol to give pure acetamide 6.
(C12H10N2O3S2) calcd: C 48.97, H 3.42, N 9.52; found: C 48.76, H
3.55, N 9.34.
4.3.2. 2-[4-(4-Oxo-2-thioxothiazolidin-5-ylidenemethyl)phenoxy]
4.2.1. 2-[3-(2,4-Dioxothiazolidin-5ylidenemethyl)phenoxy]
acetamide (7b)
acetamide (6a)
Yield 72%; mp 292 ꢂC dec; 1H NMR (DMSO-d6):
d 4.52 (s, 2H,
Yield 46%; mp 272e275 ꢂC; 1H NMR (DMSO-d6):
d
4.48 (s, 2H,
CH2O); 7.10 (m, 2H, CH arom); 7.40 (brs, 1H, NHCO); 7.56 (m, 2H, CH
arom); 7.57 (brs, 1H, NHCO); 7.59 (s, 1H, CH methylidene); 13.73
CH2O); 7.05 (dd, J1 ¼ 8.0, J2 ¼ 2.0 Hz, 1H, CH arom); 7.14 (s, 1H, CH
arom); 7.19 (d, J ¼ 8.5 Hz,1H, CH arom); 7.39 (brs,1H, NH); 7.58 (brs,
1, NH); 7.44 (t, J ¼ 8.0 Hz, 1H, CH arom); 7.72 (s, 1H, CH methyl-
(brs, 1H, NH). 13C NMR (DMSO-d6):
d 66.6 (CH2); 115.7, 122.6, 125.9,
131.7 (CH arom, CH methylidene, 5-C); 132.5, 159.8 (Cq arom);
169.3, 169.4 (CO); 195.5 (CS). Anal. (C12H10N2O3S2) calcd: C 48.97, H
3.42, N 9.52; found: C 48.75, H 3.62, N 9.45.
idene); 12.59 (brs,1H, NH). 13C NMR (DMSO-d6):
d 66.7 (CH2); 115.6,
116.8, 122.8, 124.3, 130.3, 131.3 (CH arom, CH methylidene, 5-C);
134.4,158.2 (Cq arom); 167.6, 168.0, 169.6 (CO). Anal. (C12H10N2O4S)
calcd: C 51.79, H 3.62, N 10.07; found: C 51.69, H 3.47, N 9.97.
4.3.3. 2-[2-Methoxy-5-(4-oxo-2-thioxothiazolidin-5-
ylidenemethyl)phenoxy]acetamide (7c)
4.2.2. 2-[4-(2,4-Dioxothiazolidin-5-ylidenemethyl)phenoxy]
Yield 69%; mp 258 ꢂC dec. 1H NMR (DMSO-d6):
d 3.86 (s, 3H,
acetamide (6b)
CH3); 4.50 (s, 2H, CH2O); 7.08 (d, J ¼ 1.5 Hz, 1H, CH arom); 7.16 (d,
J ¼ 8.5 Hz, CH arom); 7.25 (dd, J1 ¼ 8.5, J2 ¼ 1.5 Hz, 1H, CH arom);
7.40 (2 brs, 2H, NH2); 7.55 (s, 1H, CH methylidene); 13.73 (brs, 1H,
Yield 48%; mp 295 ꢂC dec.; 1H NMR (DMSO-d6):
d
4.51 (s, 2H,
CH2O); 7.09 (m, 2H, CH arom); 7.39 (brs,1H, NHCO); 7.55 (m, 3H, CH
arom, NHCO); 7.74 (s, 1H, CH methylidene); 12.50 (brs, 1H, NH). 13
NMR (DMSO-d6): 66.6 (CH2); 115.5,120.6,131.7,131.9 (CH arom, 5-
C, CH methylidene); 126.0, 159.4 (Cq arom); 167.4, 167.9, 169.3 (CO).
Anal. (C12H10N2O4S) calcd: C 51.79, H 3.62, N 10.07; found: C 51.65,
H 3.49, N 10.17.
C
NH). 13C NMR (DMSO-d6):
d 55.8 (CH3); 67.6 (CH2); 112.6, 115.1,
d
122.5, 125.5, 125.9 (CH arom, CH methylidene, 5-C); 131.9, 147.6,
151.5 (Cq arom); 169.4, 169.5 (CO); 195.5 (CS). Anal. (C13H12N2O4S2)
calcd: C 48.14, H 3.73, N 8.64; found: C 48.05, H 3.62, N 8.55.
4.3.4. 2-[2-Methoxy-4-(4-oxo-2-thioxothiazolidin-5-
4.2.3. 2-[5-(2,4-Dioxothiazolidin-5-ylidenemethyl)-2-
ylidenemethyl)phenoxy]acetamide (7d)
methoxyphenoxy]acetamide (6c)
Yield 44%; mp 275 ꢂC dec. 1H NMR (DMSO-d6):
d 3.84 (s, 3H,
Yield 54%; mp 258 ꢂC dec; 1H NMR (DMSO-d6):
d
3.85 (s, 3H,
CH3); 4.53 (s, 2H, CH2O); 7.02 (d, J ¼ 8.0 Hz, 1H, CH arom); 7.16 (dd,
J1 ¼8.0, J2 ¼ 2.0 Hz, CH arom); 7.20 (d, J ¼ 2.0 Hz,1H, CH arom); 7.36,
7.39 (2 brs, 2H, NH2); 7.60 (s, 1H, CH methylidene); 13.75 (brs, 1H,
CH3); 4.49 (s, 2H, CH2O); 7.08 (d, J ¼ 2.0 Hz, 1H, CH arom); 7.15 (d,
J ¼ 8.5 Hz, CH arom); 7.23 (dd, J1 ¼ 8.5, J2 ¼ 2.0 Hz, 1H, CH arom);
7.40 (2 brs, 2H, NH2); 7.68 (s, 1H, CH methylidene); 12.48 (brs, 1H,
NH). 13C NMR (DMSO-d6):
d 55.7 (CH3); 67.3 (CH2); 113.8, 113.9,
NH). 13C NMR (DMSO-d6):
d
55.8 (CH3); 67.6 (CH2); 112.5, 114.8,
122.8, 124.1, 126.4 (CH arom, CH methylidene, 5-C); 132.0, 149.2,
149.7 (Cq arom); 169.3, 169.4 (CO); 195.5 (CS). Anal. (C13H12N2O4S2)
calcd: C 48.14, H 3.73, N 8.64; found: C 47.95, H 3.65, N 8.39.
120.6, 125.1, 125.5 (CH arom, CH methylidene, 5-C); 131.8, 147.6,
151.1 (Cq arom); 167.3, 168.0, 169.5 (CO). Anal. (C13H12N2O5S) calcd:
C 50.64, H 3.92, N 9.09; found: C 50.39, H 3.81, N 8.91.
4.4. General procedure for the synthesis of (5-arylidene-2,4-
4.2.4. 2-[4-(2,4-Dioxothiazolidin-5-ylidenemethyl)-2-
dioxothiazolidin-3-yl)acetic acids 8aed
methoxyphenoxy]acetamide (6d)
Yield 47%; mp 274 ꢂC dec; 1H NMR (DMSO-d6):
d
3.83 (s, 3H,
A mixture of 2,4-thiazolidinedione (1 g, 8.5 mmol), methyl
bromoacetate (2.6 g, 17 mmol) and potassium carbonate (2.35 g,
17 mmol) was refluxed for 24 h; then the solvent was evaporated
under reduced pressure. The residue was washed with methanol to
provide (2,4-dioxothiazolidin-3-yl)acetic acid methyl ester as oil. A
mixture of (2,4-dioxothiazolidin-3-yl)acetic acid methyl ester
(0.8 g, 4.23 mmol) glacial AcOH (17 ml) and HCl 2 N (4.2 ml) was
refluxed for 90 min. After evaporation in vacuo, the crude mixture
was refluxed again with AcOH (17 ml) and HCl (4.2 ml) for 90 min.
After evaporation to dryness in vacuo, the crude oil was washed
with H2O and then with ethanol to provide pure (2,4-
dioxothiazolidin-3-yl)acetic acid (10) as oil [22]. A mixture of
(2,4-dioxothiazolidin-3-yl)acetic acid (10) (0.5 g, 2.86 mmol),
appropriate aldehyde 5 (2.86 mmol), piperidine (0.19 g, 2.29 mmol)
and EtOH (50 ml) was refluxed for 24 h. The reaction mixture was
poured into H2O acidified with AcOH to give a crude solid which
was recrystallized from methanol to give pure acid 8.
CH3); 4.52 (s, 2H, CH2O); 7.01 (d, J ¼ 8.5 Hz, CH); 7.14 (dd, J1 ¼ 8.5,
J2 ¼ 2.0 Hz, 1H, CH arom); 7.22 (d, J ¼ 2.0 Hz, 1H, CH arom); 7.35,
7.38 (2 brs, 2H, NH2); 7.74 (s, 1H, CH methylidene); 12.51 (brs, 1H,
NH). 13C NMR (DMSO-d6):
d 55.7 (CH3); 67.3 (CH2); 113.7, 113.8,
121.0, 123.3, 126.5 (CH arom, CH methylidene, 5-C); 132.0, 149.1,
149.3 (Cq arom); 167.3, 167.9, 169.4 (CO). Anal. (C13H12N2O5S) calcd:
C 50.64, H 3.92, N 9.09; found: C 50.44, H 3.85, N 8.95.
4.3. General procedure for the synthesis of 2-[(4-oxo-2-
thioxothiazolidin-5-ylidenemethyl)phenoxy]acetamides 7aed
A
mixture of 2-thioxo-4-thiazolidinone (0.8 g,
6 mmoli),
appropriate aldehyde (6 mmol), sodium acetate (1.82 g,
5
22.2 mmol) and glacial acetic acid (15 ml) was refluxed for 4e6 h.
The reaction mixture was poured into H2O and filtered off. The solid
was recrystallized from methanol to give pure acetamide 7.
4.3.1. 2-[3-(4-Oxo-2-thioxothiazolidin-5-ylidenemethyl)phenoxy]
4.4.1. {5-[(3-Carbamoylmethoxyphenyl)methylidene]-2,4-
acetamide (7a)
dioxothiazolidin-3-yl}acetic acid (8a)
Yield 79%; mp 272e275 ꢂC; 1H NMR (DMSO-d6):
d
4.49 (s, 2H,
Yield 58%; mp 251e253 ꢂC; 1H NMR (DMSO-d6):
d 4.37 (s, 2H,
CH2O); 7.08 (dd, J1 ¼ 8.0, J2 ¼ 2.0 Hz, 1H, CH arom); 7.13 (m, 1H, CH
arom); 7.19 (d, J ¼ 8.0 Hz, 1H, CH arom); 7.40 (brs, 1H, CONH); 7.45
(t, J ¼ 8.0 Hz, 1H, CH arom); 7.58 (s, 1H, CH methylidene); 7.60 (brs,
NCH2); 4.49 (s, 2H, OCH2); 7.08e7.49 (m, 4H, CH arom); 7.59 (brs,
2H, CONH2); 7.94 (s, 1H, CH methylidene). 13C NMR (DMSO-d6):
d
42.3 (NCH2); 66.7 (OCH2); 115.8, 117.4, 123.0, 130.5, 133.6 (CH
1H, CONH); 13.82 (brs, 1H, NH). 13C NMR (DMSO-d6):
d
66.7 (CH2);
arom, CH methylidene); 121.2 (C-5); 134.0, 158.2 (Cq arom); 164.9,
166.9, 167.9, 169.5 (CO). Anal. (C14H12N2O6S) calcd: C 50.00, H 3.60,
N 8.33; found: C 49.78, H 3.44, N 8.11.
116.0, 117.2, 123.4, 125.9, 130.5, 131.3 (CH arom, CH methylidene, 5-
C); 134.2, 158.2 (Cq arom); 169.3, 169.5 (CO); 195.6 (CS). Anal.

12
R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
4.4.2. {5-[(4-Carbamoylmethoxyphenyl)methylidene]-2,4-
4.5.3. {5-[(3-Carbamoylmethoxy-4-methoxyphenyl)methylidene]-
dioxothiazolidin-3-yl}acetic acid (8b)
4-oxo-2-thioxothiazolidin-3-yl}acetic acid (9c)
Yield 53%; mp 275e276 ꢂC; 1H NMR (DMSO-d6):
d
4.34 (s, 2H,
Yield 68%; mp 261e263 ꢂC; 1H NMR (DMSO-d6):
d 3.87 (s, 3H,
NCH2) 4.53 (s, 2H, OCH2); 7.11 (m, 2H, CH arom); 7.62 (m, 2H, CH
CH3); 4.52 (s, 2H, NCH2); 4.71 (s, 2H, OCH2); 7.15 (d, J ¼ 2.0 Hz, 1H,
CH arom); 7.19 (d, J ¼ 8.5 Hz, 1H, CH arom); 7.33 (dd, J1 ¼ 8.5,
J2 ¼ 2.0 Hz, 1H, CH arom); 7.42 (brs, 2H, CONH2); 7.79 (s, 1H, CH
arom); 7.41 (brs, 2H, CONH2); 7.94 (s, 1H, CH methylidene). 13C NMR
(DMSO-d6):
d 42.3 (NCH2); 67.0 (OCH2); 114.5, 124.9, 133.9 (CH
arom, CH methylidene), 117.9, 125.9, 149.5 (Cq arom, 5-C); 165.0,
167.0, 167.5, 169.4 (CO). Anal. (C14H12N2O6S) calcd: C 50.00, H 3.60,
N 8.33; found: C 49.81, H 3.50, N 8.27.
methylidene). 13C NMR (DMSO-d6):
d 45.1 (NCH2); 55.9 (CH3); 67.6
(OCH2); 112.7, 115.3, 118.8, 125.3, 126.4 (CH arom, CH methylidene,
5-C); 134.1, 147.7, 151.9 (Cq arom); 166.4, 167.3, 169.5 (CO); 193.1
(CS). Anal. (C15H14N2O6S2) calcd: C 47.11, H 3.69, N 7.33; found: C
46.96, H 3.53, N 7.19.
4.4.3. {5-[(3-Carbamoylmethoxy-4-methoxyphenyl)methylidene]-
2,4-dioxothiazolidin-3-yl}acetic acid (8c)
4.5.4. {5-[(4-Carbamoylmethoxy-3-methoxyphenyl)methylidene]-
Yield 51%; mp 235e237 ꢂC; 1H NMR (DMSO-d6):
d 3.85 (s, 3H,
4-oxo-2-thioxothiazolidin-3-yl}acetic acid (9d)
Yield 62%; mp 238e240 ꢂC; 1H NMR (DMSO-d6):
d 3.86 (s, 3H,
CH3); 4.35 (s, 2H, NCH2); 4.50 (s, 2H, OCH2); 7.13e7.31 (m, 3H, CH
arom); 7.43 (brs, 2H, CONH2); 7.88 (s,1H, CH methylidene). 13C NMR
CH3); 4.55 (s, 2H, NCH2); 4.70 (s, 2H, OCH2); 7.04 (d, J ¼ 8.0 Hz, 1H,
CH arom); 7.24 (dd, J1 ¼ 8.5, J2 ¼ 2 Hz, 1H, CH arom); 7.28 (d,
J ¼ 2.0 Hz, 1H, CH arom); 7.37, 7.39 (2 brs, 2H, CONH2); 7.84 (s, 1H,
(DMSO-d6):
d 42.2 (NCH2); 55.8 (CH3); 67.6 (OCH2); 112.6, 114.9,
125.5, 133.9 (CH arom, CH methylidene); 117.7, 125.2, 147.6, 151.5
(Cq arom, 5-C); 165.1, 167.0, 168.0, 169.5 (CO). Anal. (C15H14N2O7S)
calcd: C 49.18, H 3.85, N 7.65; found: C 49.29, H 3.67, N 7.39.
CH methylidene). 13C NMR (DMSO-d6):
d 45.0 (NCH2); 55.7 (CH3);
67.2 (OCH2); 113.8, 114.2, 119.1, 124.6, 126.3 (CH arom, CH methyl-
idene, 5-C); 134.3, 149.3, 150.1 (Cq arom); 166.3, 167.3, 169.2 (CO);
193.1 (CS). Anal. (C15H14N2O6S2) calcd: C 47.11, H 3.69, N 7.33;
found: C 46.89, H 3.48, N 7.19.
4.4.4. {5-[(4-Carbamoylmethoxy-3-methoxyphenyl)methylidene]-
2,4-oxothiazolidin-3-yl}acetic acid (8d)
Yield 50%; mp 240e242 ꢂC; 1H NMR (DMSO-d6):
d 3.86 (s, 3H,
4.6. Aldose reductase inhibition assay
CH3), 4.38 (s, 2H, NCH2) 4.55 (s, 2H, OCH2); 7.02e7.41 (m, 5H, CH
arom, CONH2); 7.95 (s, 1H, CH methylidene). 13C NMR (DMSO-d6):
AR was purified to electrophoretic homogeneity from bovine
lenses as previously described [32]. The purified enzyme (specific
activity 1.2 U/mg) was stored at 4 ꢂC in 10 mM sodium phosphate
buffer pH 7.0 containing 2 mM dithiothreitol. The enzymatic ac-
tivity was measured at 37 ꢂC following the decrease of absorbance
at 340 nm. The assay mixture (finale volume 0.7 ml) contained, in
d
42.3 (NCH2), 55.7 (CH3), 67.3 (OCH2), 113.8, 114.0, 123.6, 134.0 (CH
arom, CH methylidene), 118.1, 126.3, 149.2, 149.7 (Cq arom, 5-C),
165.1, 167.0, 168.0, 169.3 (CO). Anal. (C15H14N2O7S) calcd: C 49.18, H
3.85, N 7.65; found: C 49.31, H 3.72, N 7.78.
0.25 M sodium phosphate buffer pH 6.8, 4.67 mM D,L-glyceralde-
hyde, 0.11 mM NADPH, 0.38 M ammonium sulphate and 0.5 mM
EDTA. One Unit of enzyme activity refers to the amount of enzyme
4.5. General procedure for the synthesis of (5-arylidene-4-oxo-2-
thioxothiazolidin-3-yl)acetic acids 9aed
that catalyses the oxidation of 1
conditions.
mmol/min of NADPH in the above
A mixture of (4-oxo-2-thioxothiazolidin-3-yl)acetic acid (1.15 g,
6 mmol), appropriate aldehyde 5 (6 mmol), sodium acetate (3.69 g,
45 mmol) and glacial acetic acid (30 ml) was refluxed for 2e5 h. The
reaction mixture was poured into H2O and filtered off. The solid
was recrystallized from methanol to give pure acid 9.
Compounds tested as ARIs were dissolved at proper concen-
trations in DMSO and added to the above described assay mixture
containing 5 mU of purified AR; the DMSO concentration in all the
assays was kept constant at 1% (v/v). IC50 values (the concentration
of compound required to determine a 50% inhibition of the AR
activity) was determined by non linear regression analysis by
fitting the data to the equation describing one site competition in a
log doseeinhibition curve. For each curve, at least five different
concentrations of inhibitor were analysed. The 95% confidence
limits (95% C) were calculated using GraphPad Prism software.
4.5.1. {5-[(3-Carbamoylmethoxyphenyl)methylidene]-4-oxo-2-
thioxothiazolidin-3-yl}acetic acid (9a)
Yield 75%; mp 254e256 ꢂC; 1H NMR (DMSO-d6):
d 4.51 (s, 2H,
NCH2) 4.74 (s, 2H, OCH2); 7.13 (dd J1 ¼8.0, J2 ¼ 2.0 Hz,1H, CH arom);
7.22 (m, 1H, CH arom); 7.27 (d J ¼ 8.0 Hz, 1H, CH arom); 7.41 (brs,
1H, CONH); 7.49 (t J ¼ 8.0 Hz, 1H, CH arom); 7.62 (brs, 1H, CONH).
7.84 (s, 1H, CH methylidene). 13C NMR (DMSO-d6):
d 45.0 (NCH2);
4.7. Molecular docking e methods
66.7 (OCH2); 116.3, 117.7, 122.3, 123.6, 130.6, 133.7 (CH arom, CH
methylidene, 5-C); 134.0, 158.3 (Cq arom); 166.3, 167.2, 169.5 (CO);
193.3 (CS). Anal. (C14H12N2O5S2) calcd: C 47.72, H 3.43, N 7.95;
found: C 47.59, H 3.35, N 7.79.
Docking studies were performed with AutoDockVina 1.1.2 [37].
The crystallographic structures of AR (PDB entries 1PWM, 1US0 and
2FZD) and ligands were processed with AutoDock Tools (ADT)
package version 1.5.6rc1 [38] to merge non polar hydrogens,
calculate Gasteiger charges and select rotatable side-chain bonds.
ꢀ
4.5.2. {5-[(4-Carbamoylmethoxyphenyl)methylidene]-4-oxo-2-
Grid dimensions of 22.5 ꢃ 22.5 ꢃ 22.5 A, centred in the ligand
thioxothiazolidin-3-yl}acetic acid (9b)
binding pocket, were used for docking evaluation. Flexibility was
allowed for all rotatable bond of docked ligands and selected res-
idue side-chains in protein binding site, namely Trp20 and Leu300.
For each target structure two docking settings were used, one with
fully-rigid protein and the other with flexible Trp20 and Leu300
side-chains, thus leading to a total of six runs (three-protein scaf-
folds by two side-chain flexibility options) per each selected com-
pound. The best ranking inhibitor/enzyme complexes obtained in
each docking run were analysed by visual inspection, discarding
Yield 64%; mp 275e276 ꢂC; 1H NMR (DMSO-d6):
d 4.54 (s, 2H,
NCH2); 4.72 (s, 2H, OCH2); 7.12 (m, 2H, CH arom); 7.41 (brs, 1H,
CONH); 7.59 (brs, 1H, CONH); 7.65 (m, 2H, CH arom); 7.84 (s, 1H, CH
methylidene). 13C NMR (DMSO-d6):
d 45.0 (NCH2); 66.7 (OCH2);
115.8, 132.9 (CH arom); 118.8, 125.8, 133.9, 160.2 (5-C, CH methyl-
idene, Cq arom); 166.4, 167.3, 169.2 (CO); 193.1 (CS). Anal.
(C14H12N2O5S2) calcd: C 47.72, H 3.43, N 7.95; found: C 47.45, H 3.29,
N 7.66.

R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
13
from the subsequent MD refinement either the poses not predicting
4.8.3. Western blot
direct favourable interactions (i.e. H-bonds) with at least one res-
idue in the anionic binding site, or those poses that, after classifi-
cation in families (see molecular modelling section), exhibited only
a subset of favourable interactions of the closest subfamily.
Representative complexes for each ligand were completed by
addition of all hydrogen atoms and underwent energy minimiza-
tion and then molecular dynamics (MD) simulations with ACEMD
program [39], using ff12SB version of AMBER force field [40] for the
protein and gaff parameters [41] for the ligands. Whenever sub-
stantial rearrangement during the MD refinement significantly
affected the AReinhibitor interaction pattern, the new ligand
orientation was considered as representative of a new pose sub-
family. Thus, the corresponding complex structures were built and
refined also for all the other inhibitors that could form appreciable
interactions according to the new binding mode.
The expression of NF-kB and iNOS was evaluated by Western
blot analysis. Briefly, the untreated and treated NCTC 2544 cells
were washed twice with ice-cold PBS and resuspended with lysis
buffer (M-PERÒ Mammalian Protein Extraction Reagent, Thermo
scientific, PIERCE Biotechnology) supplemented with a cocktail of
protease inhibitor (complete, Mini, Protease Inhibitor Cocktail
Tablets, Roche) according to manufacturer’s instructions. Fifty mi-
crograms of total protein, present in the supernatant, was loaded on
each lane and separated by 4e12% Novex Bis-Tris gel electropho-
resis (NuPAGE, Invitrogen, Italy). Proteins were then transferred to
nitrocellulose membranes (Invitrogen, Italy) in a wet system. The
transfer of proteins was verified by staining the nitrocellulose
membranes with Ponceau S and the Novex Bis-Tris gel with Bril-
liant blue R. The membranes were blocked in Tris buffered saline
containing 0.01% Tween-20 (TBST) and 5% non-fat dry milk at 4 ꢂC
overnight. Polyclonal antieNFekB (subunit 50 kDa; nuclear acti-
vated form of NF-kB) (AB1602, Merck Millipore) (1:50,000 dilu-
tion), -NOS2 (N-20, sc-651, Santa Cruz Biotechnology) (1:300
Ligand geometries were fully optimized using GAMESS program
[42] at the HartreeeFock level with STO-3G basis set. The partial
atomic charges for the ligands were derived using the RESP pro-
cedure of restrained fit to the HF/6-31G*/STO-3G electrostatic po-
tential [43].
dilution), and -a tubulin (T9026; SigmaeAldrich) (1:5000 dilution)
antibodies were diluted in TBST and the membranes incubated for
2 h at room temperature. Antibodies were detected with horse-
radish peroxidase conjugated secondary antibody using the
enhanced chemiluminescence detection Supersignal West Pico
Chemiluminescent Substrate (Pierce Chemical Co., Rockford, IL).
The signal intensity of primary antibody binding was quantitatively
analysed with ImageJ software and was normalized to a loading
To perform molecular dynamics (MD) simulation in solvent,
each complex was confined in water boxes using tleap module of
AmberTools12 program [44] with a minimum distance between
ꢀ
solute and box surfaces of 10 A, filled with TIP3P water molecules
and counterions (Naþ) to neutralize the system. The solvated sys-
tems were then energy minimized through 1000 steps with solute
atoms restrained to their starting positions using a force constant of
control a-tubulin. Values were expressed as arbitrary densitometric
units (A.D.U.) corresponding (proportional) to signal intensity.
ꢀꢀ1
10 kcal molꢀ1
A
prior to MD simulations. After this, the molecules
underwent 150 ps at 300 K gradually decreasing the force constant
for positional restraints on solute atoms from 100 to
4.8.4. Statistical analysis
ꢀꢀ1
1 kcal molꢀ1
A
at constant pressure. Production MD simulations
Each experiment was repeated at least three times in triplicate
and the mean ꢁ SEM for each value was calculated. Statistical
analysis of results [Student’s t-test for paired and unpaired data;
variance analysis (ANOVA)] was performed using the statistical
software package SYSTAT, version 11 (Systat Inc., Evanston IL, USA).
A difference was considered significant at p < 0.05.
were carried out at constant temperature (300 K) and pressure
(1 atm) for 20 up to 25 ns with a time-step of 1 fs. Bonds involving
hydrogens were constrained using the SHAKE algorithm [45]. Par-
ticle Mesh Ewald method was used to evaluate electrostatic in-
teractions with a cutoff of 9.0 A; the same cutoff was used for van
der Waals interactions.
ꢀ
Acknowledgements
4.8. Evaluation of anti-inflammatory activity in vitro
Work supported in part by University of Messina.
4.8.1. Cell cultures and treatments
Appendix A. Supplementary data
Keratinocyte cell line NCTC 2544 was provided by Interlab Cell
Line Collection (Genoa, Italy) and routinely maintained in Mini-
mum Essential Medium (MEM) (SigmaeAldrich, Italy) supple-
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2014.05.003.
mented with 10% foetal calf serum, 100 U/ml penicillin, and 100 mg/
ml streptomycin, and incubated at 37 ꢂC in a humidified, 95% air/5%
CO2 atmosphere. The medium was changed every 2e3 days. For
experiments, 24 h before the cells were trypsinized, counted in a
haemocytometer, and plated in 100 mm Petri-dishes (1.5 ꢃ 106/
10 ml/well for Western blot). Experimental keratinocytes were
References
[1] M. Brownlee, Biochemistry and molecular cell biology of diabetic complica-
tions, Nature 414 (2001) 813e820.
[2] L. Costantino, G. Rastelli, M.C. Gamberini, D. Barlocco, Pharmacological ap-
proaches to the treatment of diabetic complications, Exp. Opin. Ther. Patents
10 (2000) 1245e1262.
[3] P. Alexiou, K. Pegklidou, M. Chatzopoulou, I. Nicolaou, V.J. Demopoulos, Aldose
reductase enzyme and its implication to major health problems of the 21th
century, Curr. Med. Chem. 16 (2009) 734e752.
[4] S.K. Srivastava, K.V. Ramana, A. Bhatnagar, Role of aldose reductase and
oxidative damage in diabetes and the consequent potential for therapeutic
options, Endocr. Rev. 26 (2005) 380e392.
stimulated or not (untreated controls) with 200 U/ml of IFN-g and
10ꢀ4 M of histamine, used to reproduce the mechanisms involved
in the pathogenesis of inflammatory processes, in the presence of
different concentrations (10 and 20 mM) of the compounds 9a, 9c,
9d, 7a, and 7c. After 48 h, each sample was tested for expression of
NF-kB and iNOS.
[5] A. Del Corso, M. Cappiello, U. Mura, From a dull enzyme to something else:
facts and perspectives regarding aldose reductase, Curr. Med. Chem. 15 (2008)
1452e1461.
[6] N. Hotta, Y. Akanuma, R. Kawamori, K. Matsuoka, Y. Oka, M. Shichiri, T. Toyota,
M. Nakashima, I. Yoshimura, N. Sakamoto, Y. Shigeta, Long-term clinical ef-
fects of Epalrestat, an aldose reductase inhibitor, on diabetic peripheral neu-
ropathy, Diabetes Care 29 (2006) 1538e1544.
4.8.2. Cell viability assay
The cytotoxic effect of the tested substances was evaluated by a
cell viability test based on the cleavage of 3-(4,5-dimethylthiazol-2-
yl)-2.5-diphenyltetrazolium bromide (MTT) by mitochondrial de-
hydrogenases of metabolically active cells [46].
[7] N. Hotta, T. Toyota, K. Matsuoka, Y. Shigeta, R. Kikkawa, T. Kaneko,
A. Takahashi, K. Sugimura, Y. Koike, J. Ishii, N. Sakamoto, Clinical efficacy of

14
R. Maccari et al. / European Journal of Medicinal Chemistry 81 (2014) 1e14
Fidarestat, a novel aldose reductase inhibitor, for diabetic peripheral neu-
ropathy, Diabetes Care 24 (2001) 1776e1782.
[29] A. Urzhumtsev, F. Tete-Favier, A. Mitschler, J. Barbanton, P. Barth,
L. Urzhumtseva, J.F. Biellmann, A.D. Podjarny, D. Moras, A “specificity” pocket
inferred from the crystal structures of the complexes of aldose reductase with
the pharmaceutically important inhibitors tolrestat and sorbinil, Structure 5
(1997) 601e612.
[30] S.Z. Schade, S.L. Early, T.R. Williams, F.J. Kezdy, R.L. Heinrikson, C.E. Grimshaw,
C.C. Dioughty, Sequence analysis of bovine lens aldose reductase, J. Biol. Chem.
265 (1990) 3628e3635.
[31] R. Ottanà, R. Maccari, M.L. Barreca, G. Bruno, A. Rotondo, A. Rossi, G. Chiricosta,
R. Di Paola, L. Sautebin, S. Cuzzocrea, M.G. Vigorita, 5-Arylidene-2-imino-4-
thiazolidinones: design and synthesis of novel anti-inflammatory agents,
Bioorg. Med. Chem. 13 (2005) 4243e4252.
[32] A. Del Corso, D. Barsacchi, M. Giannessi, M.G. Tozzi, M. Camici, J.L. Houben,
M. Zandomeneghi, U. Mura, Bovine lens aldose reductase: tight binding of the
pyridine coenzyme, Arch. Biochem. Biophys. 283 (1990) 512e518.
[33] H. Koizomi, A. Ohkawara, H2 histamine receptor-mediated increase in intra-
cellular Ca2þ in cultured human keratinocytes, J. Dermatol. Sci. 21 (1999)
137e142.
[34] N. Kanda, S. Watanabe, Histamine enhances the production of nerve growth
factor in human keratinocytes, J. Invest. Dermatol. 121 (2003) 570e577.
[35] V. Cardile, M. Libra, S. Caggia, G. Frasca, K. Umezawa, F. Stivala,
M.C. Mazzarino, Y. Bevelacqua, M. Coco, G. Malaponte, Dehydrox-
ymethylepoxyquinomicin, a novel nuclear factor-kappaB inhibitor, prevents
inflammatory injury induced by interferon-gamma and histamine in NCTC
2544 keratinocytes, Clin. Exp. Pharmacol. Physiol. 37 (2010) 679e683.
[36] A. Del Corso, F. Balestri, E. Di Bugno, R. Moschini, M. Cappiello, S. Sartini, C. La
Motta, F. Da Settimo, U. Mura, A new approach to control the enigmatic ac-
tivity of aldose reductase, PLoS One 8 (2013) e74076.
[8] R. Maccari, R. Ottanà, 2,4-Thiazolidinedione and 2-thioxo-4-thiazolidinone
derivatives as aldose reductase inhibitors in the search for drugs to prevent
long-term diabetes complications, in: M. Stefek (Ed.), Advances in Molecular
Mechanisms and Pharmacology of Diabetic Complications, Transworld
Research Network, Kerala, 2010, pp. 219e245.
[9] S. Miyamoto, Molecular modeling and structure-based drug discovery studies
of aldose reductase inhibitors, Chem.-Bio. Inform. J. 2 (2002) 74e85.
[10] S. Suzen, E. Buyukbingol, Recent studies of aldose reductase enzyme inhibi-
tion for diabetic complications, Curr. Med. Chem. 10 (2003) 1329e1352.
[11] S.K. Srivastava, U.C.S. Yadav, A.B.M. Reddy, A. Saxena, R. Tammali, M. Shoeb,
N.H. Ansari, A. Bhatnagar, M.J. Petrash, S. Srivastava, K.V. Ramana, Aldose
reductase inhibition suppresses oxidative stress-induced inflammatory dis-
orders, Chem. Biol. Interact. 191 (2011) 330e338.
[12] S. Miyamoto, I.M. Verma, Rel/NF-kappa B/I kappa B story, Adv. Cancer Res. 66
(1995) 255e292.
[13] A.S. Baldwin Jr., The NF-kappa B and I kappa B proteins: new discoveries and
insights, Annu. Rev. Immunol. 14 (1996) 649e683.
[14] X. Teng, H. Zhang, C. Snead, J.D. Catravas, Molecular mechanism of iNOS in-
duction by IL-1b and INF-g in rat aortic smooth muscle cells, Am. J. Physiol.
Cell Physiol. 282 (2002) C144eC152.
[15] K.V. Ramana, S.K. Srivastava, Aldose reductase: a novel therapeutic target for
inflammatory pathologies, Int. J. Biochem. Cell Biol. 42 (2010) 17e20.
[16] K.V. Ramana, Aldose reductase: new insights for an old enzyme, BioMol.
Concept 2 (2011) 103e114.
[17] K.V. Ramana, S.K. Srivastava, Mediation of aldose reductase in
lipopolysaccharide-induced inflammatory signals in mouse peritoneal mac-
rophages, Cytokine 36 (2006) 115e122.
[37] O. Trott, A.J. Olson, AutoDock Vina: improving the speed and accuracy of
[18] M. Shoeb, K.V. Ramana, S.K. Srivastava, Aldose reductase inhibition enhances
TRAIL-induced human colon cancer cell apoptosis through AKT/FOXO3a-
dependent upregulation of death receptors, Free Radic. Biol. Med. 63 (2013)
280e290.
[19] N.-H. Son, R. Ananthakrishnan, S. Yu, R.S. Khan, H. Jiang, R. Ji, H. Akashi, Q. Li,
K. O’Shea, S. Homma, I.J. Goldberg, R. Ramasamy, Cardiomyocyte aldose
reductase causes heart failure and impairs recovery from ischemia, PLoS One
7 (2012) e46549.
docking with a new scoring function, efficient optimization and multi-
threading, J. Comput. Chem. 31 (2010) 455e461.
[38] G.M. Morris, R. Huey, W. Lindstrom, M.F. Sanner, R.K. Belew, D.S. Goodsell,
A.J. Olson, Autodock4 and AutoDockTools4: automated docking with selective
receptor flexibility, J. Comput. Chem. 30 (2009) 2785e2791.
[39] M. Harvey, G. Giupponi, G. De Fabritiis, ACEMD: accelerated molecular dy-
namics simulations in the microseconds timescale, J. Chem. Theory Comput. 5
(2009) 1632e1640.
[20] Y.C. Hwang, M. Kaneko, S. Bakr, H. Liao, Y. Lu, E.R. Lewis, S. Yan, S. Ii,
M. Itakura, L. Rui, H. Skopicki, S. Homma, A.M. Schmidt, P.J. Oates, M. Szabolcs,
R. Ramasamy, Central role for aldose reductase pathway in myocardial
ischemic injury, FASEB J. 18 (2004) 1192e1199.
[40] V. Hornak, R. Abel, A. Okur, B. Strockbine, A. Roitberg, B. Strockbine,
A. Roitberg, C. Simmerling, Comparison of multiple AMBER force fields and
development of improved protein backbone parameters, Proteins 65 (2006)
712e725.
[21] C.A. Gleissner, E. Galkina, J.L. Nadler, K. Ley, Mechanisms by which diabetes
increases cardiovascular disease, Drug Discov. Today Dis. Mech. 4 (2007) 131e
140.
[41] J. Wang, R.M. Wolf, J.W. Caldwell, P.A. Kollman, D.A. Case, Development and
testing of a general AMBER force field, J. Comput. Chem. 25 (2004) 1157e
1174.
[22] R. Maccari, R. Ottanà, C. Curinga, M.G. Vigorita, D. Rakowitz, T. Steindl,
T. Langer, Structureeactivity relationships and molecular modelling of 5-
arylidene-2,4-thiazolidinediones active as aldose reductase inhibitors, Bio-
org. Med. Chem. 13 (2005) 2809e2823.
[42] M.W. Schmidt, K.K. Baldridge, J.A. Boatz, S.T. Elbert, M.S. Gordon, J.H. Jensen,
S. Koseki, N. Matsunaga, K.A. Nguyen, S. Su, T.L. Windus, M. Dupuis,
J.A. Montgomery Jr., General atomic and molecular electronic structure sys-
tem, J. Comput. Chem. 14 (1993) 1347e1363.
[23] R. Maccari, R. Ottanà, R. Ciurleo, M.G. Vigorita, D. Rakowitz, T. Steindl, T. Langer,
Evaluation of in vitro aldose redutase inhibitory activity of 5-arylidene-2,4-
thiazolidinediones, Bioorg. Med. Chem. Lett. 17 (2007) 3886e3893.
[24] R. Maccari, R. Ottanà, R. Ciurleo, D. Rakowitz, B. Matuszczak, C. Laggner,
T. Langer, Synthesis, induced-fit docking investigations, and in vitro aldose
reductase inhibitory activity of non-carboxylic acid containing 2,4-
thiazolidinedione derivatives, Bioorg. Med. Chem. 16 (2008) 5840e5852.
[25] R. Maccari, R. Ciurleo, M. Giglio, M. Cappiello, R. Moschini, A. Del Corso,
U. Mura, R. Ottanà, Identification of new non-carboxylic acid containing in-
hibitors of aldose reductase, Bioorg. Med. Chem. 18 (2010) 4049e4055.
[26] R. Ottanà, R. Maccari, M. Giglio, A. Del Corso, M. Cappiello, U. Mura,
S. Cosconati, L. Marinelli, E. Novellino, S. Sartini, C. La Motta, F. Da Settimo,
Identification of 5-arylidene-4-thiazolidinone derivatives endowed with dual
activity as aldose reductase inhibitors and antioxidant agents for the treat-
ment of diabetic complications, Eur. J. Med. Chem. 46 (2011) 2797e2806.
[27] R. Maccari, A. Del Corso, M. Giglio, R. Moschini, U. Mura, R. Ottanà, In vitro
evaluation of 5-arylidene-2-thioxo-4-thiazolidinones active as aldose reduc-
tase inhibitors, Bioorg. Med. Chem. Lett. 21 (2011) 200e203.
[43] C.I. Bayly, P. Cieplak, W.D. Cornell, P.A. Kollman, A well-behaved electro-
static potential based method using charge restraints for determining
atom-centered charges: the RESP model, J. Phys. Chem. 97 (1993) 10269e
10280.
[44] D.A. Case, T.A. Darden, T.E. Cheatham III, C.L. Simmerling, J. Wang, R.E. Duke,
R. Luo, R.C. Walker, W. Zhang, K.M. Merz, B. Roberts, S. Hayek, A. Roitberg,
G. Seabra, J. Swails, A.W. Goetz, I. Kolossváry, K.F. Wong, F. Paesani, J. Vanicek,
R.M. Wolf, J. Liu, X. Wu, S.R. Brozell, T. Steinbrecher, H. Gohlke, Q. Cai, X. Ye,
J. Wang, M.-J. Hsieh, G. Cui, D.R. Roe, D.H. Mathews, M.G. Seetin, R. Salomon-
Ferrer, C. Sagui, V. Babin, T. Luchko, S. Gusarov, A. Kovalenko, P.A. Kollman,
AMBER 12, University of California, San Francisco, 2012.
[45] J.-P. Ryckaert, G. Ciccotti, H.J.C. Berendsen, Numerical integration of the Car-
tesian equations of motion of a system with constraints: molecular dynamics
of n-alkanes, J. Comput. Phys. 23 (1977) 327e341.
[46] T. Mosmann, Rapid colorimetric assay for cellular growth and survival:
application to proliferation and cytotoxicity assays, J. Immunol. Methods 65
(1983) 55e63.
[47] R.A. Laskowski, M.B. Swindells, LigPlotþ: multiple ligand-protein interaction
[28] G. Bruno, L. Costantino, C. Curinga, R. Maccari, F. Monforte, F. Nicolò, R. Ottanà,
M.G. Vigorita, Synthesis and aldose reductase inhibitory activity of 5-
arylidene-2,4-thiazolidinediones, Bioorg. Med. Chem. 10 (2002) 1077e1084.
diagrams for drug discovery, J. Chem. Inf. Model. 51 (2011) 2778e2786.