Identification of new non-carboxylic acid containing inhibitors of aldose reductase

Article abstract of DOI:10.1016/j.bmc.2010.04.016

Non-carboxylic acid containing bioisosteres of (5-arylidene-2,4-dioxothiazolidin-3-yl)acetic acids, which are active as aldose reductase (ALR2) inhibitors, were designed by replacing the carboxylic group with the trifluoromethyl ketone moiety. The in vitr

Full text of DOI:10.1016/j.bmc.2010.04.016

Bioorganic & Medicinal Chemistry 18 (2010) 4049–4055
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Identification of new non-carboxylic acid containing inhibitors of aldose
reductase
a
a
b
b
Rosanna Maccari a, , Rosella Ciurleo , Marco Giglio , Mario Cappiello , Roberta Moschini ,
*
Antonella Del Corso b, Umberto Mura b, Rosaria Ottanà a
a Dipartimento Farmaco-chimico, Faculty of Pharmacy, University of Messina, Polo Universitario dell’Annunziata, 98168 Messina, Italy
b Department of Biology, Biochemistry Section, University of Pisa, via S. Zeno, 51, 56127 Pisa, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Non-carboxylic acid containing bioisosteres of (5-arylidene-2,4-dioxothiazolidin-3-yl)acetic acids, which
are active as aldose reductase (ALR2) inhibitors, were designed by replacing the carboxylic group with the
trifluoromethyl ketone moiety. The in vitro evaluation of the ALR2 inhibitory effects of these trifluoro-
methyl substituted derivatives led to the identification of two inhibitors effective at low micromolar
doses. It was further confirmed that a carboxylic chain on N-3 of the thiazolidinedione scaffold is a deter-
mining requisite to obtain the highest efficacy levels; however, it is not essential for the interaction with
the target enzyme and it can be replaced by different polar groups, thus obtaining less ionised or union-
ised inhibitors.
Received 24 February 2010
Accepted 6 April 2010
Available online 9 April 2010
Keywords:
Aldose reductase inhibitors
2,4-Thiazolidinediones
In vitro inhibition
Ó 2010 Elsevier Ltd. All rights reserved.
Diabetes complications
1. Introduction
fructose by sorbitol dehydrogenase with concomitant reduction of
NAD+. Under euglycaemic conditions, ALR2 does not metabolize
Diabetes mellitus (DM) is a very common metabolic disease
whose actual worldwide prevalence is estimated at about 250 mil-
lion and expected to rise to 400 million by 2030.1,2 Chronic hyper-
glycaemia is the typical metabolic disorder of DM. This condition,
through the activation of established biochemical mechanisms,
can induce tissue injury, mainly micro- and macrovascular damage
which can lead to atherosclerosis, coronary and cerebrovascular
pathologies, blindness, renal failure, even in the early stages of
the disease.3,4 These debilitating pathologies represent a serious,
and sometimes fatal, health threat for diabetic patients. Effective
drugs aimed to prevent or delay the onset and progression of dia-
betes complications are required and their identification still rep-
resents a challenge for the research.
Out of the biochemical mechanisms that have been recognized
as being implicated in the etiopathogenesis of hyperglycaemia-in-
duced tissue damage, such as increased intracellular formation of
advanced glycation-end products (AGEs), abnormal activation of
protein kinase C, increased flux of glucose through both the polyol
and the hexosamine pathways, the metabolization of glucose
through the polyol pathway plays an important role.5,6
significant amounts of glucose, due to its low affinity for this sub-
strate. In contrast, under hyperglycaemic conditions, up to about
one-third of the total metabolized amount of glucose is reduced
by ALR2 in tissues possessing insulin-independent uptake of glu-
cose, such as kidney, lens, retina, peripheral nerves. This leads to
intracellular accumulation of sorbitol with a consequent osmotic
imbalance, especially in the lens. In addition, increased fructose
levels contribute to protein glycation and formation of AGEs which,
in turn, can cause pathological changes in the functions of intracel-
lular proteins as well as in signalling pathways. At the same time,
NADPH and NAD+ deprivation causes changes in cellular redox
potentials, reduces the activity of other NADPH-dependent en-
zymes, such as nitric oxide synthase and glutathione reductase,
and induces intracellular oxidative stress. These metabolic and bio-
chemical changes result in inflammation, chronic vascular damage,
pro-atherogenic effects and decrease in nerve conduction veloc-
ity.5–12
It has been demonstrated that the increased flux of glucose
through the polyol pathway is implicated in the development of
diabetic cataracts, nephropathy, retinopathy, neuropathy, coronary
damage, atherosclerosis lesions. Moreover, ALR2 inhibitors (ARIs)
have been shown to be able to prevent or slow down the progres-
sion of certain long-term diabetic disorders.7,13–16 Therefore, ALR2
inhibition has been considered as an attractive strategy for thera-
peutic intervention to attenuate the pathogenetic consequences
of chronic hyperglycaemia associated with DM, independently of
glycaemic control.
Aldose reductase (EC 1.1.1.21, ALR2) is the first and rate-limit-
ing enzyme of the polyol pathway. It catalyses the NADPH-depen-
dent reduction of glucose to sorbitol which, in turn, is oxidized to
* Corresponding author. Tel.: +39 90 6766406; fax: +39 90 6766402.
E-mail address: rmaccari@pharma.unime.it (R. Maccari).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.04.016

4050
R. Maccari et al. / Bioorg. Med. Chem. 18 (2010) 4049–4055
In addition, there is compelling evidence that ALR2 upregula-
In the last few years, we have designed and evaluated numerous
5-arylidene-2,4-thiazolidinedione derivatives (series 1–5, Fig. 1) as
ARIs. Many of them have been shown to inhibit bovine lens ALR2
in vitro at low micromolar or submicromolar doses.25–28 Their
structures include features which are essential for ALR2 inhibition,
that is, (a) an acidic hydrogen and/or H-bond acceptor groups,
which can bind the positively charged polar recognition region of
the ALR2 active site formed by Tyr48, His110, Trp111 residues
and by the nicotinamide ring of cofactor NADP+, and (b) an aro-
matic moiety, which can establish hydrophobic interactions with
a lipophilic zone of the catalytic cleft localized among Trp111,
Thr113, Phe122, Ala299, Leu300, Ser302 and Cys303.20,22,29–31
The structure–activity relationships that we have detected so
far for these 2,4-TZDs have shown that the presence of the acetic
acid chain on N-3 (compounds 3) is related to the best ALR2 inhib-
itory effects with IC50 values in the submicromolar range.25–27 Sev-
eral N-unsubstituted analogues 1 were also found to be good ALR2
inhibitors, although they are from 10 to 100 times less potent than
corresponding carboxylic acids 3. In addition, few analogous
methyl esters 2, although devoid of any acidic functionality, were
shown to produce interesting in vitro ALR2 inhibitory effects, sim-
tion can also occur under euglycaemic conditions as a result of dif-
ferent factors and it may be involved in the etiology of other
pathologies that are not related to DM, such as inflammatory dis-
eases, cancer, renal insufficiency and mood disorders.10,17,18
Many ARIs have been reported in the last three decades, many
of which belonging to the classes of carboxylic acids (such as epal-
restat, lidorestat, tolrestat, zopolrestat, Fig. 1) and cyclic imides,
especially spirohydantoins (such as sorbinil, fidarestat, Fig. 1).
However, many of them proved to be clinically inadequate, mainly
due to adverse effects or low in vivo efficacy. Currently, epalrestat,
which has been approved in Japan for the management of diabetic
neuropathy, is the only ARI in therapy.18–23
In this context, 2,4-thiazolidinediones (2,4-TZDs) have received
attention as hydantoin bioisosteres potentially devoid of the phe-
nytoin-like hypersensitivity reactions that have been imputed to
the hydantoin ring and caused the withdrawn of sorbinil (Fig. 1)
from clinical trials.24 The discovery of glitazones as oral antidia-
betic drugs also encouraged the search for 2,4-TZDs that could be
useful for the treatment of both type 2 DM and its long-term
complications.19,22
O
O
OH
O
O
OH
S
N
OH
N
N
F
F
S
S
S
N
O
F
CF3
Tolrestat
Epalrestat
Lidorestat
O
O
H
O
OH
H
N
N
CF3
N
HN
HN
N
S
O
O
F
F
N
O
O
CONH2
O
Sorbinil
Zopolrestat
Fidarestat
O
O
S
O
S
R
N
N
1
2
3
4
5
R = H
R = CH2COOCH3
R = CH2COOH
R = CH2CONH2
R = CH2CONHOH
R
O
O
Ar
HO
2a R = OCH3 IC50 = 6.18 μM
3a R = OH
1-5
IC50 = 0.15 μM
IC50 = 6.52 μM
4a R = NH2
5a R = NHOH IC50 = 1.79 μM
O
O
O
O
S
N
NHOH
N
S
NHOH
O
O
R
5b R = OCH2C6H5 IC50 = 7.34 μM
5c R = OC6H5 IC50 = 3.67 μM
5d IC50 = 4.20 μM
Figure 1. Structures of some ARIs.

R. Maccari et al. / Bioorg. Med. Chem. 18 (2010) 4049–4055
4051
ilar to those of N-unsubstituted 2,4-TZDs 1.25–27 Among other non-
carboxylic acid containing 5-arylidene-2,4-thiazolidinedione
derivatives, the 2-[5-(4-hydroxybenzylidene)-2,4-dioxothiazoli-
din-3-yl]acetamide (4a, Fig. 1), the corresponding N-hydroxyaceta-
mide (5a, Fig. 1) and analogues 5b–d (Fig. 1) effectively inhibited
ALR2 with IC50 values in the low micromolar range.28
tion.26 Methoxy substituted analogues were also included for
comparison. Here we present the synthesis and evaluation of the
in vitro ALR2 inhibitory activity of these new non-carboxylic acid
containing derivatives.
2. Chemistry
Indeed, among ARIs, carboxylic acids are known as potent
in vitro inhibitors. However, their effectiveness has generally been
shown to decrease in vivo, probably because of their poor capabil-
ity to penetrate key target tissues, in particular peripheral
nerves.19–21 Thus, the identification of new low acidity ARIs, which
are mainly non-ionised at physiological pH values, would be a
desirable goal.32
In continuing our search for non-carboxylic acid containing 2,4-
TZDs, we selected the trifluoromethyl ketone moiety as a bioisoster
of the carboxylic group of acetic acids 3 to obtain new derivatives
in which the pharmacophoric elements that we previously defined
for 5-arylidene-2,4-thiazolidinediones (i.e., two H-bond acceptor
groups and a lipophilic region)26 were maintained (compounds 6,
Fig. 2). In most of derivatives 6, the 5-arylidene moiety contained
two aromatic rings or a phenyl ring bearing a hydroxy group, since
in the previously studied series of 5-arylidene-2,4-thiazolidinedi-
ones this was found to be a favourable feature for ALR2 inhibi-
The synthetic route which was followed to obtain the designed
trifluoromethyl substituted derivatives (Scheme 1) started from
the already reported synthesis of 2,4-TZDs 1, which were N-alkyl-
ated by the reaction with methyl bromoacetate to give esters 2.25,26
These latter were hydrolysed to acids 3 in acidic medium and, sub-
sequently, converted into the corresponding chlorides 7 by react-
ing with thionyl chloride. According to a reported procedure,33
the reaction of compounds 7 with trifluoroacetic anhydride and
pyridine provided 5-arylidene-3-(3,3,3-trifluoro-2-oxopropyl)-
2,4-thiazolidinediones 6 (Scheme 1). 5-Hydroxybenzylidene substi-
tuted derivatives 6i, j were obtained by the reaction of the corre-
sponding methoxybenzylidene analogues with boron tribromide
(Scheme 2).
Mass spectra of all compounds 6 showed, along with the molec-
ular peak, a diagnostic intense peak (which is often the base peak)
due to the loss of a fragment with m/z = 181 corresponding to
CON(CH2COCF3)CO. Fragmentation modes by loss of CF3 or COCF3
were also detected which, however, produced weak peaks (gener-
ally with relative abundance 62%).
O
O
It is well-known that trifluoromethyl ketones can be readily hy-
drated. In fact, analogously to what was observed by other authors
for different triflouromethyl ketones,33,34 compounds 6 were found
to exist in an equilibrium with the corresponding hydrates (60,
Scheme 1) under proper conditions.
N
CF3
S
O
Ar
6 a-j
In 1H NMR spectra recorded in CDCl3 solution, signals attribut-
able to hydrates 60, along with those due to related ketones 6, were
a
b
c
d
Ar = 3-OC6H5-C6H4
Ar = 4-OC6H5-C6H4
Ar = 4-OCH2C6H5-C6H4
Ar = 4-C6H5-C6H4
O
CH2COCF3
O
CH2COCF3
O
N
N
e
f
g
Ar = 1-Naphthyl
Ar = 2-Naphthyl
Ar = 3-OCH3-C6H4
S
S
O
BBr3, CH2Cl2
h
i
j
Ar = 4-OCH3-C6H4
Ar = 3-OH-C6H4
Ar = 4-OH-C6H4
CH3O
HO
6 g, h
6 i, j
Figure 2. General structure of the newly synthesised 2,4-TZDs.
Scheme 2. Synthesis of compounds 6i, j.
O
S
O
CH2COOCH3
O
S
CH2COOH
O
H
O
O
H
N
N
a
N
N
b
c
Ar CHO
+
S
S
O
O
Ar
Ar
Ar
1
2
3
d
O
O
S
CH2C(OH)2CF3
O
CH2COCl
O
O
CH2COCF3
O
N
e, f
f
N
N
S
S
Ar
Ar
Ar
7
6'
6
(a) C5H11N, C2H5OH, Δ; (b) BrCH2COOCH3, K2CO3, acetone, Δ; (c) AcOH, HCl, Δ; (d) SOCl2, Δ;
(e) (CF3CO)2O, pyridine, anhydrous Et2O, r.t.; (f) H2O.
Scheme 1. General method for the synthesis of 5-arylidene-3-(3,3,3-trifluoro-2-oxopropyl)-2,4-thiazolidinediones 6a–h.

4052
R. Maccari et al. / Bioorg. Med. Chem. 18 (2010) 4049–4055
present (ratios 6:60 ranged from 1:1.5 to 1:3 at 25 °C). Two singlets
at 4.25–4.30 ppm and 4.91–4.96 ppm were diagnostic, which are
attributable to the resonance of the methylene protons of hydrates
60 and ketones 6, respectively. In addition, two singlets, which were
due to the resonance of the two different 5-methylidene protons of
species 6 and 60, were present in the range from 8.00 to 8.05 ppm.
When 1H NMR spectra were recorded in hygroscopic DMSO-d6,
we observed only one set of signals, which can be attributed to hy-
drates 60. In this case, the resonance of the methylene protons gave
rise to a singlet at 3.88–3.94 ppm, whereas the signal attributable
to the methylidene proton appeared in the range between
7.82 ppm and 8.55 ppm.
regard to their inhibitory activity since, analogously to N-hydrox-
yacetamides
5 (calculated pKa values ranging from 8.64 to
8.75),28 compounds 60 could bind ALR2 in their deprotonated form
which can be stabilized by the positive charge of NADP+ and by
hydrogen bonds donated by the amino acid residues lining the po-
lar recognition region of the ALR2 active site.
Out of the tested compounds, the 5-(4-hydroxybenzylidene)
substituted derivative (6/60j) was the most active ARI, with an
IC50 value (3.0 lM) very similar to that of sorbinil (Table 1).
Although 20-fold less potent than the [5-(4-hydroxybenzylid-
ene)-2,4-dioxothiazolidin-3-yl]acetic acid (3a, Fig. 1),26 6/60j was
twice more active than amide 4a and almost as effective as N-
hydroxyacetamide 5a (Fig. 1).28 The 3-hydroxybenzylidene isomer
13C NMR spectra of compounds 6 were recorded in DMSO-d6
solution, with the exception of few cases, because of the generally
moderate solubility of these derivatives in CDCl3. 13C NMR spectra
(6/60i) also showed a reasonable IC50 value (12.31
lM), but it was
fourfold less effective (Table 1). The replacement of the hydroxy
group of both compounds 6/60i and 6/60j with a methoxy one led
to less active derivatives (6/60g and 6/60h, respectively) (Table 1)
and, in particular, the loss in potency was more marked for the
para substituted analogue than for the meta substituted isomer.
Compound 6/60b stood out of 2,4-TZDs 6/60a–f, which possess
two aromatic rings in their 5-arylidene moiety, with its IC50 value
in DMSO-d6 solution showed
a diagnostic quartet at 92.0–
92.4 ppm (JCF = 30–32 Hz) that was due to the resonance of C-2
of the dihydroxypropylic side chain of geminal diols 60, whereas
no signal attributable to the carbonyl group of corresponding com-
pounds 6 was detected. Another quartet was present at 121.2–
124.0 ppm (JCF = 277.5–288.0 Hz) which was due to the resonance
of the trifluoromethyl group (see Section 4). In the CDCl3 solution
of compound 6d, besides the signals attributable to hydrate 60d,
a quartet at 183 ppm (JCF = 37.5 Hz) was observed, which was
due to the resonance of the carbonyl group adjacent to the CF3
group of the ketone. In addition, two distinct signals were present
at 44.6 ppm and 45.0 ppm, which were originated by the reso-
nance of the methylene carbon atoms of compounds 6d and 60d.
Variable temperature NMR experiments in CDCl3 solution con-
firmed the existence of an equilibrium between compounds 6
and 60, which was more and more shifted towards the hydrate form
as temperature was increased whereas it was restored by cooling.
The configuration Z of the exocyclic C@C bond of compounds 6
was assigned on the analogy of that of previously reported thiazo-
lidinones, which was determined by means of X-ray diffractome-
try,25,35 as well as on the bases of the NMR data.
of 8.3 l
M (Table 1), whereas analogues 6/60a, d–f produced from
45% to 64% ALR2 inhibition at 25 lM dose. The 4-benzyloxybenzy-
lidene derivative (6/60c) strongly affected NADPH oxidation, mak-
ing impossible the exact evaluation of its IC50 value, which,
however, appeared to be higher than 9 lM.
On the whole, the tested trifluoromethyl derivatives 6/60 re-
sulted to be significantly more potent ARIs than acetamides 4,
whereas their inhibitory effects were generally less marked than
those of corresponding N-hydroxyacetamides 5. In addition, in
comparison with the corresponding N-unsubstituted 2,4-TZDs,25,26
compounds 6/60b, h and j, which bear different groups linked to an
oxygen atom in the position 4 of the benzylidene ring, were found
to be from 2.4-fold to 4.5-fold more effective ARIs, whereas their
meta substituted isomers 6/60a, g and i resulted to be less active.
Moreover, the comparison of the inhibitory potency of para
substituted analogues 6/60b, h and j showed an efficacy order 6/
60j > 6/60b > 6/60h. Thus, the replacement of the phenoxy group of
6/60b with a methoxy one proved to be detrimental since it halved
the inhibitory effectiveness (compound 6/60h), whereas, in con-
trast, its replacement with a hydroxy group provided a threefold
gain in potency (compound 6/60j). The removal of the oxygen
bridge in the para position of the 5-benzylidene moiety produced
a marked decrease in potency (compound 6/60d).
3. Results and discussion
Compounds 6 were evaluated for their ability to inhibit the
in vitro reduction of D,L-glyceraldehyde by highly purified ALR2
from bovine lenses, using sorbinil as a reference drug (Table 1).
On the basis of the above reported findings, it is reasonable to
expect that hydrates 60 are the predominant species under the
experimental conditions of the enzymatic assay as well as in the
biological environment. Moreover, given their calculated pKa val-
ues (7.94–7.95),36 geminal diols 60 are expected to be ionised for
22–23% at physiological pH values, in comparison with ketones 6
which are completely non-ionised. This aroused our interest with
These data are in accordance with the SARs that we previously
observed for 5-arylidene-2,4-thiazolidinediones active as ARIs.26,28
In fact, analogously to 6/60j, the 5-(4-hydroxybenzylidene) moiety
led to 2,4-TZDs of series 1–5 also endowed with good ALR2 inhib-
itory activity. The above-mentioned phenol derivative 4a (Fig. 1)
was the only compound out of previously evaluated acetamides
4 which showed an appreciable affinity towards ALR2.28 The corre-
sponding N-hydroxyamide (compound 5a, Fig. 1) was also found to
be the most effective inhibitor in the series of N-hydroxyaceta-
mides that we investigated.28 Analogously, in the class of (5-arylid-
ene-2,4-dioxothiazolidin-3-yl)acetic acids 3, 5-(4-hydroxybenzy-
lidene) substituted derivative 3a (Fig. 1) is one of the most potent
ARIs.26 A similar observation can be made for methyl esters 2,
among which the 4-hydroxybenzylidene analogue (2a, Fig. 1)
was one of few derivatives with a micromolar IC50 value.26 A rea-
sonable explanation for the appreciable affinity of these 5-(4-
hydroxybenzylidene) substituted derivatives towards ALR2
emerged from our recent docking results that retrieved a favour-
able binding mode in which the deprotonated phenol group inter-
acts with the positively charged polar recognition region of the
ALR2 catalytic centre, in particular with Tyr48 and His110,
whereas the side chain on N-3 and the carbonyl group at position
Table 1
In vitro bovine lens ALR2 inhibitory activity of compounds 6/60
a
Compd
Ar
IC50
6/60a
6/60b
6/60c
6/60d
6/60e
6/60f
3-OC6H5–C6H4
4-OC6H5–C6H4
4-OCH2C6H5–C6H4
4-C6H5–C6H4
1-Naphthyl
64%
8.3 (3.0–23.4)
N.D.
45%
52%
2-Naphthyl
59%
6/60g
6/60h
6/60i
3-OCH3–C6H4
4-OCH3–C6H4
3-OH–C6H4
23.3 (13.8–39.3)
17.4 (12.5–24.1)
12.31 (9.65–14.28)
3.0 (2.2–4.0)
2.0 (1.7–3.5)
6/60j
Sorbinil
4-OH–C6H4
N.D. = not detectable.
a
IC50 (lM) (95% CL) or %inhibition at 25 lM.

R. Maccari et al. / Bioorg. Med. Chem. 18 (2010) 4049–4055
4053
2 of the thiazolidinedione ring can establish hydrogen bonds with
Leu300, Leu301 and Ser302 in the lipophilic pocket of the enzyme.
Thus, these compounds appeared to bind ALR2 in a flipped pose
compared with the more generally expected one which, however,
was shown to be adopted by more lipophilic analogues devoid of
any acidic group in the 5-arylidene moiety.28 Therefore, it is rea-
sonable to assume that the wide lipophilic arylidene moiety of 4-
phenoxybenzylidene derivative 6/60b fits the lipophilic pocket of
the ALR2 active site, whereas the acetic chain on N-3 interacts with
the polar recognition region. Accordingly, the replacement of the
distal phenyl ring of 6/60b with a methyl group led to a less lipo-
philic analogue (6/60h) which can interact less effectively with
the lipophilic region of the active site.
The results of the in vitro testing here reported further con-
firmed the SAR picture that we have detected so far. In fact, the
presence of a carboxylic chain on N-3 of the thiazolidinedione scaf-
fold is a determining requisite to obtain the highest ALR2 inhibi-
tion levels. However, it is not essential for the inhibition of the
enzyme, provided that it is replaced by groups that can effectively
interact with the ALR2 polar recognition region, such as the N-
hydroxycarboxamide or trifluoromethyl ketone groups.
course was monitored by means of TLC. Then H2O (5 ml) was
added cautiously and the mixture was stirred until CO2 disap-
peared. The mixture was poured into H2O (70 ml) and extracted
with diethyl ether. The organic layer was dried (Na2SO4) and the
solvent removed under reduced pressure to provide compound 6.
1H and 13C NMR data reported below for all compounds 6 refer
to DMSO-d6 solutions, in which (as described above) only signals
due to the hydrate forms (60) were detected.
4.2.1. (Z)-5-[(3-Phenoxyphenyl)methylidene]-3-(3,3,3-trifluoro-
2-oxopropyl)-2,4-thiazolidinedione (6a)
Yield 34%; mp 93–95 °C 1H NMR (DMSO-d6): d 3.88 (s, 2H,
NCH2); 7.07–7.22 (m, 5H, CH arom); 7.36–7.58 (m, 4H, CH arom);
7.97 (s, 1H, CH methylidene). 13C NMR (DMSO-d6): d 45.3 (NCH2);
92.0 (q JCF = 30.75 Hz, C(OH)2); 119.4, 119.8, 120.7, 124.7, 125.3,
130.7, 131.6, 132.5 (CH arom + CH methylidene); 122.8 (C-5);
123.7 (q JCF = 285.0 Hz, CF3); 135.4, 156.3, 158.1 (Cq arom);
165.6, 167.0 (C@O). MS (EI), m/z (%): 407 (63) [M+], 226 (100). Anal.
Calcd for C19H12F3NO4S: C, 56.02; H, 2.97; N, 3.44. Found: C, 56.37;
H, 3.21; N, 3.46.
The in vitro evaluation of this series of 5-arylidene-3-(3,3,3-tri-
fluoro-2-oxopropyl)-2,4-thiazolidinediones as ARIs allowed the
identification of two new non-carboxylic acid containing 5-arylid-
ene-2,4-thiazolidinedione derivatives (6/60b and 6/60j) that are ac-
tive at low micromolar doses. This work is part of a project which is
targeted to identify new ARIs endowed with favourable pharmaco-
logical profiles. The best inhibitors selected among compounds of
series 1–6, which contain groups having different degrees of ioni-
zation on the critical portions of the 5-arylidene-2,4-thiazolidined-
ione skeleton, will be included in further studies to assess their
in vivo efficacy.
4.2.2. (Z)-5-[(4-Phenoxyphenyl)methylidene]-3-(3,3,3-trifluoro-
2-oxopropyl)-2,4-thiazolidinedione (6b)
Yield 23%; mp 128–131 °C 1H NMR (DMSO-d6): d 3.90 (s, 2H,
NCH2); 7.10–7.25 (m, 5H, CH arom); 7.44 (m, 2H, CH arom); 7.63
(m, 2H, CH arom); 7.89 (s, 1H, CH methylidene). 13C NMR
(DMSO-d6): d 45.5 (NCH2); 92.3 (q JCF = 31.0 Hz, C(OH)2); 119.1,
120.6, 125.4, 131.1, 132.9, 133.1 (CH arom + CH methylidene);
120.3 (C-5); 123.9 (q JCF = 285.0 Hz, CF3); 128.5, 155.9, 159.8 (Cq
arom); 166.2, 167.5 (C@O). MS (EI), m/z (%): 407 (52) [M+], 226
(100). Anal. Calcd for C19H12F3NO4S: C, 56.02; H, 2.97; N, 3.44.
Found: C, 56.21; H, 2.78; N, 3.51.
4. Experimental
4.1. Chemistry
4.2.3. (Z)-5-[(4-Benzyloxyphenyl)methylidene]-3-(3,3,3-triflu-
oro-2-oxopropyl)-2,4-thiazolidinedione (6c)
Yield 37%; mp 118–120 °C. 1H NMR (DMSO-d6): d 3.90 (s, 2H,
NCH2); 5.17 (s, 2H, OCH2); 7.17–7.49 (m, 7H, CH arom); 7.59 (m,
2H, CH arom); 7.86 (s, 1H, CH methylidene). 13C NMR (DMSO-
d6): d 45.5 (NCH2); 70.4 (OCH2); 92.4 (q JCF = 31.35 Hz, C(OH)2);
116.5, 117.8, 129.2, 132.9, 133.5, 134.7 (CH arom + CH methyli-
dene); 123.7 (q JCF = 285.0 Hz, CF3); 124.7 (C-5); 126.4, 155.1,
158.6 (Cq arom); 166.2, 167.4 (C@O). MS (EI), m/z (%): 421 (15)
[M+], 91 (100). Anal. Calcd for C20H14F3NO4S: C, 57.01; H, 3.35; N,
3.32. Found: C, 57.21; H, 3.45; N, 3.03.
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 with-
in 0,4% of theory. 1H and 13C NMR spectra were recorded on a
Varian 300 magnetic resonance spectrometer (300 MHz for 1H
and 75 MHz for 13C). Chemical shifts are given in d units (ppm) rel-
ative to internal standard Me4Si and refer to DMSO-d6 solutions.
Coupling constants (J) are given in hertz (Hz). 13C NMR spectra
were determined by Attached Proton Test (APT) experiments and
the resonances were always attributed by proton-carbon hetero-
nuclear chemical shift correlation. Mass spectra were recorded
on a Shimadzu GC17A/MSQP5050 spectrometer.
4.2.4. (Z)-5-[(4-Phenylphenyl)methylidene]-3-(3,3,3-trifluoro-2-
oxopropyl)-2,4-thiazolidinedione (6d)
Yield 30%; mp 184–186 °C. 1H NMR (DMSO-d6): d 3.92 (s, 2H,
NCH2); 7.30 (br s., 2H, 2 OH, exchangeable with D2O); 7.41–7.52
(m, 3H, CH arom); 7.71–7.76 (m, 4H, CH arom); 7.86 (m, 2H, CH
arom); 7.97 (s, 1H, CH methylidene). 13C NMR (DMSO-d6): d 45.3
(NCH2); 92.1 (q JCF = 30.75 Hz, C(OH)2); 121.6 (C-5); 123.7 (q
JCF = 287.2 Hz, CF3); 127.4, 128.0, 128.8, 129.6, 131.3, 132.8 (CH ar-
om + CH methylidene); 132.6, 139.3, 142.5 (Cq arom); 165.9, 167.2
(C@O). MS (EI), m/z (%): 391 (61) [M+], 210 (100). Anal. Calcd for
C19H12F3NO3S: C, 58.31; H, 3.09; N, 3.58. Found: C, 58.57; H,
2.89; N, 3.33.
Unless stated otherwise, all materials were obtained from com-
mercial suppliers and used without further purification.
4.2. General method for the synthesis of 5-arylidene-3-(3,3,3-
trifluoro-2-oxopropyl)-2,4-thiazolidinediones 6a–h
(5-Arylidene-2,4-dioxothiazolidin-3-yl)acetic acids 3 were ob-
tained according to a reported procedure.25,26 Acid 3 (1.4 mmol)
was dissolved in 5–6 ml of thionyl chloride and the reaction mix-
ture was refluxed for 1 h. The crude mixture was evaporated under
reduced pressure to dryness, providing pure compound 7. Acid
chloride 7 (1.5 mmol) was dissolved in anhydrous diethyl ether
(7 ml). The ether solution of 7 and anhydrous pyridine (0.95 g,
12 mmol) were added to a solution of trifluoroacetic anhydride
(1.9 g, 9 mmol) in anhydrous diethyl ether (7 ml). The reaction
mixture was stirred at room temperature for 2 h and the reaction
4.2.5. (Z)-5-(Naphthalen-1-ylmethylidene)-3-(3,3,3-trifluoro-2-
oxopropyl)-2,4-thiazolidinedione (6e)
Yield 35%; mp 144–145 °C. 1H NMR (DMSO-d6): d 3.94 (s, 2H,
NCH2); 7.62–7.72 (m, 4H, CH arom); 8.02–8.13 (m, 3H, CH arom);
8.55 (s, 1H, CH methylidene). 13C NMR (DMSO-d6): d 45.2 (NCH2);
92.0 (q JCF = 30.5 Hz, C(OH)2); 123.6 (q JCF = 277.5 Hz, CF3); 123.9,
126.1, 126.8, 127.3, 128.0, 129.4, 130.1, 131.4 (CH arom + CH
methylidene); 125.4 (C-5); 130.7, 131.6, 133.8 (Cq arom); 165.4,

4054
R. Maccari et al. / Bioorg. Med. Chem. 18 (2010) 4049–4055
166.2 (C@O). MS (EI), m/z (%): 365 (58) [M+], 184 (100). Anal. Calcd
for C17H10F3NO3S: C, 55.89; H, 2.76; N, 3.83. Found: C, 56.01; H,
3.00; N, 3.59.
fluoro-2-oxopropyl)-2,4-thiazolidinedione (6h). Yield 75%; mp
148–151 °C. 1H NMR (DMSO-d6): d 3.90 (s, 2H, NCH2); 6.91 (m,
2H, CH arom); 7.24 (br s, 2H, 2 OH, exchangeable with D2O);
7.48 (m, 2H, CH arom); 7.82 (s, 1H, CH methylidene); 10.37 (br s,
1H, OH, exchangeable with D2O). 13C NMR (DMSO-d6): d 45.2
(NCH2); 92.0 (q JCF = 30.0 Hz, C(OH)2); 117.1 (C-5); 116.9, 132.9,
133.7 (CH arom + CH methylidene); 123.7 (q JCF = 287.3 Hz, CF3);
124.4, 160.5 (Cq arom); 166.1, 167.5 (C@O). MS (EI), m/z (%): 331
(40) [M+], 150 (100). Anal. Calcd for C13H8F3NO4S: C, 47.13; H,
2.43; N, 4.23. Found: C, 46.91; H, 2.69; N, 3.98.
4.2.6. (Z)-5-(Naphthalen-2-ylmethylidene)-3-(3,3,3-trifluoro-2-
oxopropyl)-2,4-thiazolidinedione (6f)
Yield 34%; mp 183–186 °C. 1H NMR (DMSO-d6): d 3.92 (s, 2H,
NCH2); 7.25 (br s., 2H, 2 OH, exchangeable with D2O); 7.60–8.07
(m, 7H, CH arom); 8.22 (s, 1H, CH methylidene). 13C NMR
(DMSO-d6): d 45.3 (NCH2); 92.0 (q JCF = 30.75 Hz, C(OH)2); 122.0
(C-5); 123.7 (q JCF = 288 Hz, CF3); 126.4, 127.6, 128.2, 128.6,
129.4, 131.4, 132.3, 133.1 (CH arom + CH methylidene); 131.1,
132.2, 133.8 (Cq arom); 165.8, 167.8 (C@O). MS (EI), m/z (%): 365
(53) [M+], 184 (100). Anal. Calcd for C17H10F3NO3S: C, 55.89; H,
2.76; N, 3.83. Found: C, 56.12; H, 3.07; N, 3.48.
4.3. Determination of the in vitro aldose reductase inhibition
Bovine eyes were obtained from a local abattoir soon after
slaughtering; lenses were removed and kept frozen at ꢀ20 °C until
used. Aldose reductase was purified from bovine lens by an ionic
exchange chromatographic step on DE-52 and an affinity chro-
matographic step on Matrex Orange A.37 The final enzyme prepara-
tion (0.93 U/mg specific activity) was stored at 4 °C in 10 mM
sodium phosphate buffer pH 7.0 supplemented with 2 mM
dithiothreitol.
4.2.7. (Z)-5-[(3-Methoxyphenyl)methylidene]-3-(3,3,3-trifluoro-
2-oxopropyl)-2,4-thiazolidinedione (6g)
Yield 52%; mp 120–122 °C. 1H NMR (DMSO-d6): d 3.80 (s, 3H,
OCH3); 3.90 (s, 2H, NCH2); 7.07 (d J = 8.1 Hz, 1H, CH arom); 7.19–
7.25 (m, 2H, CH arom); 7.46 (dd J = 8.1 and 8.1 Hz, 1H, CH arom);
7.90 (s, 1H, CH methylidene). 13C NMR (DMSO-d6):
d
45.5
The enzyme activity was measured at 37 °C, using 4.67 mM D,L-
(NCH2); 56.1 (OCH3); 92.3 (q JCF = 30.75 Hz, C(OH)2); 116.2,
117.2, 122.5, 131.3, 133.4 (CH arom + CH methylidene); 121.2 (q
JCF = 277.5 Hz, CF3); 122.4 (C-5); 135.1, 160.4 (Cq arom); 166.1,
167.5 (C@O). MS (EI), m/z (%): 345 (58) [M+], 164 (100). Anal. Calcd
for C14H10F3NO4S: C, 48.70; H, 2.92; N, 4.06. Found: C, 48.39; H,
3.11; N, 3.96.
glyceraldehyde as substrate in 0.25 M sodium phosphate, pH 6.8,
0.38 M ammonium sulfate, 0.11 mM NADPH, and 0.5 mM EDTA.
One Unit of enzyme activity is the amount of enzyme that catalyses
the oxidation of
conditions.
1 lmol/min of NADPH in the above assay
The sensitivity of aldose reductase to different compounds was
tested in the above assay conditions in the presence of inhibitors
dissolved at proper concentration in DMSO. The concentration of
DMSO in the assay mixture was kept constant at 1% (v/v). IC50 val-
ues (the concentration of the inhibitor required to produce a 50%
inhibition of the enzyme catalyzed reaction) were determined by
non linear regression analysis by fitting the data to the equation
describing one site competition in a log dose–inhibition curve.
Each log dose–inhibition curve was generated using at least five
concentrations of inhibitor (causing an inhibition between 20%
and 80%) and each concentration was tested at least in triplicate.
The 95% confidence limits (95% CL) were calculated using Graph-
Pad Prism software.
4.2.8. (Z)-5-[(4-Methoxyphenyl)methylidene]-3-(3,3,3-trifluoro-
2-oxopropyl)-2,4-thiazolidinedione (6h)
Yield 35%; mp 112–113 °C. 1H NMR (DMSO-d6): d 3.82 (s, 3H,
OCH3); 3.91 (s, 2H, NCH2); 7.10 (m, 2H, CH arom); 7.22 (br s, 2H,
2 OH, exchangeable with D2O); 7.61 (m, 2H, CH arom); 7.88 (s,
1H, CH methylidene). 13C NMR (DMSO-d6): d 45.6 (NCH2); 56.4
(OCH3); 92.3 (q JCF = 30.0 Hz, C(OH)2); 115.8, 133.0, 133.5 (CH ar-
om + CH methylidene); 118.8 (C-5); 124.0 (q JCF = 287.7 Hz, CF3);
126.3, 159.7 (Cq arom); 166.3, 168.3 (C@O). MS (EI), m/z (%): 345
(59) [M+], 164 (100). Anal. Calcd for C14H10F3NO4S: C, 48.70; H,
2.92; N, 4.06. Found: C, 48.51; H, 3.17; N, 4.12.
4.2.9. (Z)-5-[(3-Hydroxyphenyl)methylidene]-3-(3,3,3-trifluoro-
2-oxopropyl)-2,4-thiazolidinedione (6i)
Acknowledgements
Boron tribromide (1.74 g, 7 mmol) was added to a solution of 6g
(0.4 g, 1.1 mmol) in dichloromethane on an ice bath and the reac-
tion mixture was stirred for 2 h. Then methanol was added and the
mixture was evaporated under reduced pressure to dryness. The
crude solid was dissolved in ethyl acetate and the solution was
washed with H2O, dried with anhydrous Na2SO4 and evaporated
under reduced pressure to provide pure compound 6i. Yield 84%;
mp 130–132 °C. 1H NMR (DMSO-d6): d 3.90 (s, 2H, NCH2); 6.88
(d J = 8.1 Hz, 1H, CH arom); 7.00 (s, 1H, CH arom); 7.06 (d
J = 7.8 Hz, 1H, CH arom); 7.25 (br s, 2H, 2 OH, exchangeable with
D2O); 7.34 (dd J = 8.1 and 7.8 Hz, 1H, CH arom); 7.82 (s, 1H, CH
methylidene); 9.87 (br s, 1H, OH, exchangeable with D2O). 13C
NMR (DMSO-d6): d 45.3 (NCH2); 92.1 (q JCF = 30.0 Hz, C(OH)2);
116.5, 118.4, 121.8, 130.9, 133.5 (CH arom + CH methylidene);
123.7 (q JCF = 287.3 Hz, CF3); 125.6 (C-5); 134.7, 158.5 (Cq arom);
165.9, 167.4 (C@O). MS (EI), m/z (%): 331 (41) [M+], 150 (100). Anal.
Calcd for C13H8F3NO4S: C, 47.13; H, 2.43; N, 4.23. Found: C, 46.87;
H, 2.78; N, 4.02.
This work was supported in part by University of Messina (PRA
2006) and in part by University of Pisa. We are indebted to Dr. G.
Pasqualetti and Dr. R. Di Sacco (veterinary staff of Consorzio Macel-
li S. Miniato, Pisa) for their valuable cooperation in bovine lens
collection.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2010.04.016.
References and notes
1. Zimmet, P.; Alberti, K. G. M. M.; Shaw, J. Nature 2001, 414, 782.
2. Wild, S.; Roglic, G.; Green, A.; Sicree, R.; King, H. Diabetes Care 2004, 27, 1047.
3. Porte, D., Jr.; Schwartz, M. W. Science 1996, 272, 699.
4. Stratton, I. M.; Adler, A. I.; Neil, H. A. W.; Matthews, D. R.; Manley, S. E.; Cull, C.
A.; Hadden, D.; Turner, R. C.; Holman, R. R. BMJ 2000, 321, 405.
5. Brownlee, M. Nature 2001, 414, 813.
6. Brownlee, M. Diabetes 2005, 54, 1615.
7. Kador, P. F. Med. Res. Rev. 1988, 8, 325.
8. Yabe-Nishimura, C. Pharmacol. Rev. 1998, 50, 21.
4.2.10. (Z)-5-[(4-Hydroxyphenyl)methylidene]-3-(3,3,3-trifluoro-
2-oxopropyl)-2,4-thiazolidinedione (6j)
The procedure was the same as reported for compound 6i, start-
ing from (Z)-5-[(4-methoxyphenyl)methylidene]-3-(3,3,3-tri-
9. Srivastava, S. K.; Ramana, K. V.; Bhatnagar, A. Endocr. Rev. 2005, 26, 380.
10. Ramana, K. V.; Srivastava, S. K. Cytokine 2006, 36, 115.
11. Gleissner, C. A.; Galkina, E.; Nadler, J. L.; Ley, K. Drug Discovery Today: Dis. Mech.
2007, 4, 131.

R. Maccari et al. / Bioorg. Med. Chem. 18 (2010) 4049–4055
4055
12. Gleissner, C. A.; Sanders, J. M.; Nadler, J.; Ley, K. Arterioscler. Thromb. Vasc. Biol.
2008, 28, 1137.
26. Maccari, R.; Ottanà, R.; Curinga, C.; Vigorita, M. G.; Rakowitz, D.; Steindl, T.;
Langer, T. Bioorg. Med. Chem. 2005, 13, 2809.
13. Vinik, A. I.; Park, T. S.; Stansberry, K. B.; Pittenger, G. L. Diabetologia 2000, 43,
957.
27. Maccari, R.; Ottanà, R.; Ciurleo, R.; Vigorita, M. G.; Rakowitz, D.; Steindl, T.;
Langer, T. Bioorg. Med. Chem. Lett. 2007, 17, 3886.
14. Hotta, N.; Toyota, T.; Matsuoka, K.; Shigeta, Y.; Kikkawa, R.; Kaneko, T.;
Takahashi, A.; Sugimura, K.; Koike, Y.; Ishii, J.; Sakamoto, N. Diabetes Care 2001,
24, 1776.
15. Miwa, K.; Nakamura, J.; Hamada, Y.; Naruse, K.; Nakashima, E.; Kato, K.;
Kasuya, Y.; Yasuda, Y.; Kamiya, H.; Hotta, N. Diabetes Res. Clin. Pract. 2003, 60, 1.
16. Ramirez, M. A.; Borja, N. L. Pharmacotherapy 2008, 28, 646.
17. Tammali, R.; Ramana, K. V.; Srivastava, S. K. Cancer Lett. 2007, 252, 299.
18. Alexiou, P.; Pegklidou, K.; Chatzopoulou, M.; Nicolaou, I.; Demopoulos, V. J.
Curr. Med. Chem. 2009, 16, 734.
19. Costantino, L.; Rastelli, G.; Cignarella, G.; Vianello, P.; Barlocco, D. Expert Opin.
Ther. Patents 1997, 7, 843.
20. Costantino, L.; Rastelli, G.; Vianello, P.; Cignarella, G.; Barlocco, D. Med. Res. Rev.
1999, 19, 3.
21. Costantino, L.; Rastelli, G.; Gamberini, M. C.; Barlocco, D. Expert Opin. Ther.
Patents 2000, 10, 1245.
22. Miyamoto, S. Chem. -Bio. Inform. J. 2002, 2, 74.
23. Suzen, S.; Buyukbingol, E. Curr. Med. Chem. 2003, 10, 1329.
24. Lipinski, C. A.; Aldinger, C. E.; Beyer, T. A.; Bordner, J.; Burdi, D. F.; Bussolotti, D.
L.; Inskeep, P. B.; Siegel, T. W. J. Med. Chem. 1992, 35, 2169.
25. Bruno, G.; Costantino, L.; Curinga, C.; Maccari, R.; Monforte, F.; Nicolò, F.;
Ottanà, R.; Vigorita, M. G. Bioorg. Med. Chem. 2002, 10, 1077.
28. Maccari, R.; Ottanà, R.; Ciurleo, R.; Rakowitz, D.; Matuszczak, B.; Laggner, C.;
Langer, T. Bioorg. Med. Chem. 2008, 16, 5840.
29. Urzhumtsev, A.; Tête-Favier, F.; Mitschler, A.; Barbanton, J.; Barth, P.;
Urzhumtseva, L.; Biellmann, J. F.; Podjarny, A.; Moras, D. Structure 1997, 5,
601.
30. El-Kabbani, O.; Darmanin, C.; Schneider, T. R.; Hazemann, I.; Ruiz, F.; Oka, M.;
Joachimiak, A.; Schulze-Briese, C.; Tomizaki, T.; Mitschler, A.; Podjarny, A.
Proteins 2004, 55, 805.
31. Steuber, H.; Heine, A.; Podjarny, A.; Klebe, G. J. Mol. Biol. 2008, 379, 991.
32. Mylari, B. L.; Armento, S. J.; Beebe, D. A.; Conn, E. L.; Coutcher, J. B.; Dina, M. S.;
O’Gorman, M. T.; Linhares, M. C.; Martin, W. H.; Oates, P. J.; Tess, D. A.;
Withbroe, G. J.; Zembrowski, W. J. J. Med. Chem. 2005, 48, 6326.
33. Boivin, J.; El Kaim, L.; Zard, S. Z. Tetrahedron 1995, 51, 2573.
34. Kruchok, I. S.; Gerus, I. I.; Kukhar, V. P. Tetrahedron 2000, 56, 6533.
35. Ottanà, R.; Maccari, R.; Barreca, M. L.; Bruno, G.; Rotondo, A.; Rossi, A.;
Chiricosta, G.; Di Paola, R.; Sautebin, L.; Cuzzocrea, S.; Vigorita, M. G. Bioorg.
Med. Chem. 2005, 13, 4243.
36. Marvin Calculator Plugins, http://www.chemaxon.com.
37. Del Corso, A.; Barsacchi, D.; Giannessi, M.; Tozzi, M. G.; Camici, M.;
Houben, J. L.; Zandomeneghi, M.; Mura, U. Arch. Biochem. Biophys. 1990,
283, 512.