Evaluation of in vitro aldose redutase inhibitory activity of 5-arylidene-2,4-thiazolidinediones

Article abstract of DOI:10.1016/j.bmcl.2007.04.109

A number of 5-arylidene-2,4-thiazolidinediones containing a hydroxy or a carboxymethoxy group in their 5-benzylidene moiety have been synthesised and evaluated as in vitro aldose reductase (ALR2) inhibitors. Most of them exhibited strong inhibitory activi

Full text of DOI:10.1016/j.bmcl.2007.04.109

Bioorganic & Medicinal Chemistry Letters 17 (2007) 3886–3893
Evaluation of in vitro aldose redutase inhibitory activity
of 5-arylidene-2,4-thiazolidinediones
a
a
a
Rosanna Maccari,a, Rosaria Ottana, Rosella Ciurleo, Maria Gabriella Vigorita,
*
`
Dietmar Rakowitz,b Theodora Steindlb and Thierry Langerb
a
`
`
Dipartimento Farmaco-chimico, Facolta di Farmacia, Universita di Messina, Viale SS. Annunziata, 98168 Messina, Italy
bInstitute of Pharmacy, Department of Pharmaceutical Chemistry, University of Innsbruck, Innrain 52a, A-6020 Innsbruck, Austria
Received 26 March 2007; revised 30 April 2007; accepted 30 April 2007
Available online 6 May 2007
Abstract—A number of 5-arylidene-2,4-thiazolidinediones containing a hydroxy or a carboxymethoxy group in their 5-benzylidene
moiety have been synthesised and evaluated as in vitro aldose reductase (ALR2) inhibitors. Most of them exhibited strong inhibitory
activity, with IC50 values in the range between 0.20 and 0.70 lM. Molecular docking simulations into the ALR2 active site high-
lighted that the phenolic or carboxylic substituents of the 5-benzylidene moiety can favourably interact, in alternative poses, either
with amino acid residues lining the lipophilic pocket of the enzyme, such as Leu300, or with the positively charged recognition
region of the ALR2 active site.
Ó 2007 Elsevier Ltd. All rights reserved.
Diabetes mellitus (DM) is a common chronic metabolic
disease characterized by hyperglycaemia and various
metabolic imbalances; its prevalence is about 6% world-
wide and the number of cases, presently estimated at
more than 150 million, is predicted to double by
2025.1–3
The increased flux of glucose through the polyol path-
way, which occurs under conditions of hyperglycaemia
in tissues possessing insulin-independent glucose trans-
port (retina, lenses, kidney, nerve), has been recognized
to play a major role in the development of diabetic com-
plications.6,7,9–15
DM is always associated with serious long-term compli-
cations, such as neuropathy, nephropathy, retinopathy,
cataracts, accelerated atherosclerosis and increased risk
of myocardial infarction, stroke and limb amputation,
which negatively affect the quality of life and life expec-
tancy of diabetic patients.4,5 These pathologies are the
consequences of microvascular and macrovascular dam-
age induced, at least in part, by hyperglycaemia.6–10 De-
spite the availability of different insulin preparations and
oral antidiabetic drugs, glycaemic levels cannot be com-
pletely normalized in diabetic patients and, thus, the on-
set of long-term complications is unavoidable.11
Therefore, the prevention and control of these patholo-
gies is still a challenging problem in the therapeutic
treatment of diabetes.
Aldose reductase (EC 1.1.1.21, ALR2) catalyses the
NADPH-dependent reduction of glucose to sorbitol in
the first step of this metabolic pathway. It has received
great attention as a target enzyme in the search for drugs
able to prevent or delay the onset and progression of
diabetic complications, independently of glycaemic
levels.
Over the last three decades, numerous ALR2 inhibitors
(ARIs) have been identified; most of them belong to
either carboxylic acid (such as epalrestat; Fig. 1) or
hydantoin (such as sorbinil, fidarestat; Fig. 1) classes
of compounds.9,11,13,14,16 However, many of the clini-
cally tested ARIs proved to be inadequate as drug can-
didates because of adverse pharmacokinetics, toxic side-
effects or low efficacy. At present, epalrestat is the only
ARI available on the market.9,11,17
In this context, 2,4-thiazolidinediones (2,4-TZDs) have
aroused interest as hydantoin bioisosteres that are
potentially devoid of the hypersensitivity reactions
related to the hydantoin ring, which caused the
Keywords: 2,4-Thiazolidinediones; Aldose reductase; Diabetes melli-
tus; Molecular docking.
*
Corresponding author. Tel.: +39 90 6766406; fax: +39 90 355613;
e-mail: rmaccari@pharma.unime.it
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.04.109

R. Maccari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3886–3893
3887
F
F
O
N
S
H
O
H
O
N
N
HN
HN
N
COOH
O
F
O
F
N
F
S
S
COOH
O
CONH2
O
Sorbinil
Fidarestat
Lidorestat
Epalrestat
CH2COOH
O
O
S
CH2COOH
O
O
S
O
R
O
R
N
N
N
S
N
S
HO
O
O
Ar
Ar
HO
1-3
4-6
3a
3b
1, 4 R = H
2, 5 R = CH2COOCH3
3, 6 R = CH2COOH
Figure 1.
withdrawal of sorbinil from clinical trials. In particular,
several 2,4-TZDs have been reported or patented be-
cause they were recognized as having an antihypergly-
caemic effect and ALR2 inhibitory activity.9,11,13
Moreover, the insertion of an additional aromatic ring
or an H-bond donor group in this moiety enhanced
ALR2 inhibitory effect. In particular, the presence of a
hydroxy group, especially in the para position of the
5-benzylidene ring, proved to be very favourable.19 In
fact, [5-(4-hydroxybenzylidene)-2,4-dioxothyazolidin-3-
yl]acetic acid (3a, Fig. 1) was shown to be more potent
(IC50 = 0.15 lM) than its 3-hydroxybenzylidene substi-
tuted isomer 3b (Fig. 1) (IC50 = 0.66 lM), which, how-
ever, also displayed appreciable inhibitory activity.19
Over the last few years we have synthesised numerous
2,4-thiazolidinediones (1–6) (Fig. 1) designed and
in vitro assayed as ARIs.18–20 They possess the struc-
tural requisites essential for ALR2 inhibition: (a) an
acidic hydrogen and 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 cofac-
tor NADP+; (b) an aromatic moiety, which can establish
hydrophobic interactions with a lipophilic contact zone
of the catalytic cleft lined by Leu300 and
Trp111.9,11,13,14,16,21–23
On the basis of these findings, we continued our search
aimed at amplifying the structure–activity relationships
of 2,4-TZDs active as ARIs and at identifying new
active analogues. In particular, starting from the prom-
ising activity of compounds 3a and 3b, we synthesised
and evaluated 5-arylidene analogues 7–14 (Scheme 1),
which retained the pharmacophoric elements previously
identified for 5-arylidene-2,4-thiazolidinediones active as
ARIs.19 Compounds 7–14 bore a hydroxy, a methoxy
and/or a carboxymethoxy group in their 5-benzylidene
moiety. Such substituents are potentially able to en-
hance enzyme/inhibitor complexes stability. This can
be achieved via further hydrogen bonds and/or electro-
static interactions with amino acid residues in the lipo-
philic pocket of the ALR2 active site, analogous to
what has recently been reported for fidarestat.24 In fact,
the carbamoyl group of fidarestat has been shown to be
hydrogen bonded to the main-chain nitrogen atom of
Leu300, which lines part of this lipophilic pocket, and
it was suggested that this interaction was responsible
for the high affinity and selectivity of fidarestat for
ALR2.24 Moreover, the substitution of this carbamoyl
moiety with a carboxylic group enhanced the predicted
binding energy of the enzyme/inhibitor complex.25
Most of the 2,4-TZDs 1–6 were shown to be effective
in vitro ARIs.18–20 In particular, several N-unsubstitut-
ed derivatives 1 and 4 exhibited appreciable activity at
micromolar doses. The insertion of an acetic chain on
N-3 of the thiazolidinedione ring (compounds 3 and 6)
gave a 10–100-fold gain in potency and led to the highest
inhibition levels. In addition, certain methyl esters 2 and
5, although devoid of any acidic functionality, showed
inhibitory properties similar to those of N-unsubstituted
analogues.
The inhibitory potency of 2,4-TZDs 1–6 was also
shown to be significantly influenced by the lipophilic
aryl moiety in the position 5 and by its substitution
pattern. The saturation of the exocyclic C,C-double
bond in 5 generally brought about a moderate de-
crease in activity and we found that most of the tested
5-benzylidene substituted 2,4-TZDs (1–3) were more
effective ARIs than the corresponding 5-benzyl ana-
logues (4–6).20
Among 2,4-TZDs 7–14, compounds 7 and 8 were
already known in the literature26 and 9, 11 and 13 are

3888
R. Maccari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3886–3893
CHO
R'
O
O
S
O
S
CH2COOCH3
O
O
S
H
O
CH2COOH
O
H
O
b
N
a
N
N
c
N
+
S
R
R'
R'
R'
9, 14 (35-84%)
R
R
R
7, 8 (60-68%)
7 R = OCH3; R' = OH
8 R = OH; R' = OCH3
9 R = OCH3;
10 R = OH;
11 R = H;
R' = OH
R' = OCH3
R' = OCH2COOH
d
12 R = OCH2COOH; R' = H
13 R = OCH3;
R' = OCH2COOH
14 R = OCH2COOH; R' = OCH3
O
O
CH2COOH
O
O
CH2COOCH3
O
H
b
N
N
c
N
S
S
S
O
Scheme 1. Reagents and conditions: (a) C5H11N, C2H5OH, D; (b) BrCH2COOCH3, K2CO3, acetone; (c) AcOH, HCl, D; (d) 3-hydroxy-4-
methoxybenzaldehyde or 4-hydroxy-3-methoxybenzaldehyde, C5H11N, C2H5OH, D.
commercially available products. However, to the best
of our knowledge, their ALR2 inhibitory activity has
never been reported.
10 started from the synthesis of (2,4-dioxothiazolidin-
3-yl)acetic acid19 followed by Knoevenagel condensa-
tion with the corresponding benzaldehydes (Scheme 1).
Analytical and spectroscopic data (1H and 13C NMR)
confirmed the structures assigned to compounds 7–14.28
5-Arylidene-2,4-thiazolidinediones 7 and 8 were synthes-
ised, following a known procedure,18,19 by Knoevenagel
condensation of commercially available 2,4-thiazolidin-
edione with suitable aldehydes (Scheme 1). The treat-
ment of 7 and 8 with methyl bromoacetate, in acetone
in the presence of potassium carbonate as base, pro-
duced N/O-alkylation; after acidic hydrolysis, only acids
13 and 14 were obtained with good yields.27 Acids 11
and 12 were synthesised, following the same procedure,
starting from 5-(4-hydroxybenzylidene)-2,4-thiazolidin-
edione and its 3-hydroxybenzylidene isomer. In order
to prevent O-alkylation, the preparation of acids 9 and
The ALR2 inhibitory activities of 2,4-TZDs 7–14 were
evaluated in vitro using partially purified ALR2 from
bovine lenses; sorbinil was used as a reference drug
(Table 1).29–31
5-(4-Hydroxy-3-methoxybenzylidene)-2,4-thiazolidined-
ione (7) displayed moderate ALR2 inhibitory activity, in
the micromolar range (IC50 = 11.8 lM), similar to that
of the 5-(3-methoxybenzylidene) and 5-(4-hydroxyben-
zylidene) monosubstituted analogues previously
Table 1. In vitro bovine lenses ALR2 inhibitory activity of 2,4-thiazolidinediones 7–14
O
S
R''
O
N
R'
R
a
Compound
R
R0
R00
IC50
3ab
H
OH
OH
H
CH2COOH
CH2COOH
H
0.15 (0.10–0.22)
0.66 (0.48–0.91)
11.8 (9.64–14.4)
33% (12.5 lM)
0.70 (0.68–0.72)
0.56 (0.46–0.69)
0.60 (0.40–0.90)
0.23 (0.092–0.57)
0.26 (0.10–0.70)
0.20 (0.096–0.43)
1.41 (1.12–1.79)
3bb
7
OCH3
OH
OH
8
OCH3
OH
OCH3
H
9
10
OCH3
OH
CH2COOH
CH2COOH
CH2COOH
CH2COOH
CH2COOH
CH2COOH
11
12
H
OCH2COOH
H
OCH2COOH
OCH3
OCH2COOH
13
14
OCH2COOH
OCH3
Sorbinil
a IC50 (lM) (95% C.L.) or % inhibition at the given concentration.
b Ref. 19.

R. Maccari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3886–3893
3889
reported (IC50 = 13.28 lM and 8.96 lM, respec-
tively).18,19 In contrast, its isomer 8 was less effective,
producing 33% inhibition at 12.5 lM dose (Table 1); it
also displayed lower activity than previously assayed
The removal of the methoxy group of 13 and 14 led
to different results. In fact, [5-(3-carboxymethoxyben-
zylidene)-2,4-dioxothiazolidin-3-yl]acetic acid (12) dis-
played an inhibitory effect equal to that of the
parent compound (14); moreover, it was shown to
be 3-fold more active than its 5-(3-hydroxybenzylid-
ene) analogue 3b (Table 1). In contrast, [5-(4-carboxy-
methoxybenzylidene)-2,4-dioxothiazolidin-3-yl]acetic acid
(11), although displaying appreciable ALR2 inhibitory
activity (IC50 = 0.60 lM), proved to be about
twice less potent than the parent compound (13); it
was also shown to be about 3-fold less active than
its isomer 12 and 4-fold less effective than
5-(4-hydroxybenzylidene) substituted analogue 3a
(Table 1).
5-(3-hydroxybenzylidene)-2,4-thiazolidinedione (IC50
10.7 lM).19
=
Acids 9–14 were shown to be potent in vitro ARIs, dis-
playing IC50 values in the range between 0.20 and
0.70 lM (Table 1).
As expected, compounds 9 and 10 were appreciably
more effective ARIs than the corresponding N-unsubsti-
tuted analogues (7 and 8). Once again these data demon-
strate that the presence of a carboxylic anionic head on
N-3 is an important structural requisite to produce high
levels of enzyme inhibition. However, acid 9 proved to
be less active than 5-(4-hydroxybenzylidene) substituted
analogue 3a (Table 1).
Indeed, 5-(3-carboxymethoxybenzylidene) substituted
acids 12 and 14 proved to be the most active ARIs out
of the 2,4-TZDs here reported, followed by [5-(4-carb-
oxymethoxy-3-methoxybenzylidene)-2,4-dioxothiazolidin-
3-yl]acetic acid (13) (Table 1). The latter also exhibited
potency twice as high as the 5-(3-methoxybenzylidene)
substituted analogue (IC50 = 0.48 lM), which we had
previously evaluated.18
The replacement of the hydroxy group in compounds 9
and 10 with a carboxymethoxy one led to derivatives 13
and 14 and gave a 3-fold gain in ALR2 inhibitory po-
tency (Table 1).
Figure 2. Docking pose of compound 14 within the binding site of ALR2 (PDB entry 1Z3N) showing possible interactions with the protein. Colour
coding in LigandScout: red, hydrogen bond acceptor; yellow, hydrophobic features; magenta, ring aromatic feature.

3890
R. Maccari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3886–3893
Figure 3. Docking pose of compound 10 within the binding site of ALR2 (PDB entry 1Z3N) showing possible interactions with the protein. Colour
coding see Figure 2.
A docking study was performed in order to investigate
how 2,4-TZDs 7–14 fit into the active site of ALR2
and to visualise interaction possibilities with the
enzyme.32
The most active compound 14 was frequently found in
poses fulfilling the typical interaction pattern also
shown by lidorestat. The software LigandScout,38
a
tool for automatic identification of interaction possi-
bilities and structure-based pharmacophore generation
from protein–ligand complexes, was used to investi-
gate and visualise the docked ligands within the bind-
ing site. Figure 2 shows an example of a good docking
pose of thiazolidinedione 14. The substitution pattern
of the benzene ring is well suited for making the typ-
ical hydrogen bonds with active site residues Tyr48
and Trp111. In addition, the carboxylic acid group
is positioned facing the basic His110, enabling
charge–charge interactions and probably also hydro-
gen bonds (Fig. 2). The latter are not visualised in
LigandScout because of bad angles, but can be
expected in reality, considering that the protein is trea-
ted as rigid during the docking process. As can be
seen from this example, an additional interaction with
Trp20 was sometimes found. Moreover, hydrogen
bonds to Leu300 were also achieved by two acceptor
functions of the ligand (carbonyl oxygen atoms
of the thiazolidinedione ring and the second carbox-
ylic acid moiety) (Fig. 2). Lipophilic parts of the
ligand, especially the benzene ring, are surrounded
PDB (Protein Brookhaven Databank)33 entry 1N3Z,
˚
which shows a resolution of about 1 A and has recently
been released, was chosen for the docking study.34 Its li-
gand, lidorestat (Fig. 1), shows the typical binding char-
acteristics shared by many ALR2 inhibitors,
thiazolidinediones included. These characteristics com-
prise: chemical functions enabling hydrogen bonds to
Tyr48, His110 and Trp111; an acidic function for charge
interactions; lipophilic areas for contacts with hydro-
phobic parts of the pocket and preferably also a hydro-
gen bonding possibility to Leu300.
Minimised structures35 of compounds 7, 10, 12, 13 and
14 were flexibly docked into the binding site using the
LigandFit module in the Cerius2 software package.36,37
Good poses of the five investigated new ARIs showed
that the ligands either satisfied interactions in the polar
part of the site very well (Tyr48, His110, Trp111), or
interacted predominantly in the lipophilic region
(Leu300), or both.

R. Maccari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3886–3893
3891
by hydrophobic binding site residues. In the proximity
˚
of the benzene ring (3–5.5 A) the aromatic rings of
Supplementary data
Trp20, Trp219, Phe122, Trp111 and Trp79 can be
found. Furthermore, looking at the overlay of com-
pound 14 with lidorestat demonstrated that their posi-
tions in the site were very similar when spatial
occupation and positioning of the chemical functions
for interaction were considered.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmcl.2007.04.109.
References and notes
Other docking poses also showed the ligands in 180 de-
gree rotated orientations, that is, the orientations we
originally expected on the basis of our previous results.19
However, especially for the larger structures (13 and 14),
these poses were found less frequently and the resultant
interaction possibilities were usually more limited than
the ones from the reversed poses. For example, only
one out of the 20 docking poses for ligand 14 was orien-
tated in such a way. In general the LigandFit dock_-
1. Zimmet, P.; Alberti, K. G. M. M.; Shaw, J. Nature 2001,
414, 782.
2. Zimmet, P. Diabetes Metab. 2003, 29, 6S9.
3. Diamond, J. Nature 2003, 423, 599.
4. Weissman, P. N. Am. J. Cardiol. 2002, 90, 42G.
5. Girach, A.; Vignati, L. J. Diabet. Compl. 2006, 20, 228.
6. Brownlee, M. Nature 2001, 414, 813.
7. Oyama, T.; Miyasita, Y.; Watanabe, H.; Shirai, K.
Diabetes Res. Clin. Pract. 2006, 73, 227.
8. Porte, D., Jr.; Schwartz, M. W. Science 1996, 272, 699.
9. Costantino, L.; Rastelli, G.; Vianello, P.; Cignarella, G.;
Barlocco, D. Med. Res. Rev. 1999, 19, 3.
10. Wolffenbuttel, B. H.; van Haeften, T. W. Drugs 1995, 50,
263.
11. Costantino, L.; Rastelli, G.; Gamberini, M. C.; Barlocco,
D. Exp. Opin. Ther. Patents 2000, 10, 1245.
12. Kador, P. F. Med. Res. Rev. 1988, 8, 325.
13. Costantino, L.; Rastelli, G.; Cignarella, G.; Vianello, P.;
Barlocco, D. Exp. Opin. Ther. Patents 1997, 7, 843.
14. Miyamoto, S. Chem.-Bio. Inform. J. 2002, 2, 74.
15. Suzen, S.; Buyukbingol, E. Curr. Med. Chem. 2003, 10,
1329.
score,
a
measurement
of
the
non-bonded
intermolecular interactions between the ligand and the
protein, was also lower for this alternative pose with a
value of 56.4. In comparison, most of the other retrieved
poses, which showed orientations as in Figure 2, ob-
tained higher dock-scores starting with a value of 71.6.
Good docking results with the thiazolidinedione part
facing the inner polar side of the binding cavity were
found, for example, for 5-(3-hydroxy-4-methoxybenzy-
lidene) substituted compound 10 (Fig. 3), where hydro-
gen bonds to Trp20, Leu 300 and p–p stacking with
Trp111 can be seen.
16. Steuber, H.; Heine, A.; Klebe, G. J. Mol. Biol. 2007, 368,
618.
In conclusion, this work allowed us to identify a number
of other 5-arylidene-2,4-thiazolidinediones active as
in vitro ARIs and to extend the structure–activity rela-
tionships of this class of ALR2 inhibitors.
17. Castan˜er, J.; Prous, J. Drugs Fut. 1987, 12, 336.
18. Bruno, G.; Costantino, L.; Curinga, C.; Maccari, R.;
`
Monforte, F.; Nicolo, F.; Ottana, R.; Vigorita, M. G.
Bioorg. Med. Chem. 2002, 10, 1077.
`
`
19. Maccari, R.; Ottana, R.; Curinga, C.; Vigorita, M. G.;
Rakowitz, D.; Steindl, T.; Langer, T. Bioorg. Med. Chem.
2005, 13, 2809.
`
20. Rakowitz, D.; Maccari, R.; Ottana, R.; Vigorita, M. G.
Bioorg. Med. Chem. 2006, 14, 567.
21. Urzhumtsev, A.; Tete-Favier, F.; Mitschler, A.; Barban-
ton, J.; Barth, P.; Urzhumtseva, L.; Biellmann, J. F.;
Podjarny, A.; Moras, D. Structure 1997, 5, 601.
22. El-Kabbani, O.; Wilson, D. K.; Petrash, J. M.; Quiocho,
F. A. Mol. Vis. 1998, 4, 19.
23. Lee, Y. S.; Hodoscek, M.; Kador, P. F.; Sugiyama, K.
Chem.-Biol. Interact. 2003, 143-144, 307.
24. El-Kabbani, O.; Darmanin, C.; Oka, M.; Schulze-Briese,
C.; Tomizaki, T.; Hazemann, I.; Mitschler, A.; Podjarny,
A. J. Med. Chem. 2004, 47, 4530.
25. Darmanin, C.; Chevreux, G.; Potier, N.; Van Dorsselaer,
A.; Hazemann, I.; Podjarny, A.; El-Kabbani, O. Bioorg.
Med. Chem. 2004, 12, 3797.
26. (a) Kaarsholm, N. C.; Olsen, H. B.; Madsen, P.; Oesterg-
aard, S.; Jakobsen, P.; Moeller, T. T. PCT Int. Appl. 2006,
WO 2006082245 A1 20060810; (b) Olsen, H. B.; Kaars-
holm, N. C.; Madsen, P.; Balschmidt, P. PCT Int. Appl.
2006, WO 2006005683 A1 20060119; (c) Morita, H.; Mori,
H. Jpn. Kokai Tokkyo Koho, 1996, JP 08277279 A2
19961022; (d) Kumar, B. R. P.; Karvekar, M. D.;
Adhikary, L.; Nanjan, M. J.; Suresh, B. . J. Heterocycl.
Chem. 2006, 43, 897; (e) Cetenko, W. A.; Connor, D. T.;
Sorenson, R. J.; Unangst, P. C.; Stabler, S. R. Eur. Pat.
Appl. 1989, EP 343643, CAN112:235298.
In particular, the introduction of a hydroxy group in the
para position of the 5-benzylidene ring appeared to be
beneficial. The addition of a methoxy group in position
3 or the displacement of the 4-hydroxy group to position
3 of the benzylidene ring usually led to less active ana-
logues. Moreover, substituting the hydroxy group with
a carboxymethoxy one generally enhanced activity.
These results suggested that the phenolic or carboxylic
substituents of the 5-benzylidene moiety played a bene-
ficial role in the binding to ALR2. In fact, molecular
docking simulations into the active site of the enzyme
highlighted that these groups can favourably interact,
in alternative poses, either with amino acid residues of
the C-terminal loop in the lipophilic pocket of the en-
zyme, such as Leu300, or with the positively charged
recognition region of the ALR2 active site, lined by
Tyr48, His110 and Trp111; in the latter case, hydrogen
bonds to Leu300 can be achieved by the acetic chain
on N-3 and the carbonyl oxygen atoms of the thiazolid-
inedione ring.
Taking into account that the interactions between these
2,4-TZDs and Leu300 could be responsible for a selec-
tive inhibition of ALR2 over other aldo–keto reduc-
tases, the evaluation of their inhibitory effects towards
related enzymes, such as aldehyde reductase, will be
the subject of a future note.
27. A mixture of 7 or 8 (2.51 g, 10 mmol), methyl bromoac-
etate (6.12 g, 40 mmol) and potassium carbonate (5.53 g,

3892
R. Maccari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3886–3893
40 mmol) in acetone (150 ml) was refluxed for 24 h. The
[5-(4-Carboxymethoxy-3-methoxybenzylidene)-2,4-dioxo-
thiazolidin-3- yl]acetic acid (13). Yield 67%; mp 256 °C. 1H
NMR (DMSO-d6): d 3.86 (s, 3H, OCH3); 4.39 (s, 2H,
NCH2); 4.81 (s, 2H, OCH2); 7.04 (d, J = 9 Hz, 1H, CH
arom); 7.22 (d, J = 9 Hz, 1H, CH arom); 7.30 (s, 1H, CH
arom); 7.96 (s, 1H, CH); 13C NMR (DMSO-d6): d 42.7
(NCH2); 56.2 (OCH3); 65.4 (OCH2); 113.8, 114.6, 124.1
(CH arom); 118.5 (5-C); 134.7 (CH methylidene); 126.6,
149.6, 150.1 (Cq arom); 165.6, 167.5, 168.5, 170.2 (CO).
Anal. (C15H13NO8S). Calcd.: C, 49.05%; H, 3.54%; N,
3.81%. Found: C, 49.27%; H, 3.73%; N, 3.75%.
[5-(3-Carboxymethoxy-4-methoxybenzylidene)-2,4-dioxo-
thiazolidin-3- yl]acetic acid (14). Yield 84%; mp 246 °C. 1H
NMR (DMSO-d6): d 3.85 (s, 3H, OCH3); 4.36 (s, 2H,
NCH2); 4.74 (s, 2H, OCH2); 7.14–7.18 (m, 2H, CH arom);
7.28 (dd, J = 8.7 and 2.4 Hz, 1H, CH arom); 7.91 (s, 1H,
CH); 13C NMR (DMSO-d6): d 42.0 (NCH2); 55.5 (OCH3);
64.7 (OCH2); 112.3, 114.2, 124.8 (CH arom); 117.4 (5-C);
133.8 (CH methylidene); 125.0, 147.1, 151.0 (Cq
arom);165.1, 166.8, 167.7, 169.7 (CO). Anal.
(C15H13NO8S). Calcd.: C, 49.05%; H, 3.54%; N, 3.81%.
Found: C, 48.91%; H, 3.35%; N, 4.01%.
solvent was evaporated under reduced pressure. The solid
residue was washed with H2O and then refluxed in glacial
AcOH (40 ml) and HCl 12N (10 ml) for 2 h. After
evaporation in vacuo, the residue was refluxed again with
AcOH (40 ml) and HCl (10 ml) for 2 h. After evaporation
to dryness in vacuo, the crude solid was washed with H2O
and recrystallized from ethanol providing pure carboxylic
acid 13 or 14.
28. 5-(4-Hydroxy-3-methoxybenzylidene)-2,4-thiazolidinedione
(7). Yield 68%; mp 217 °C. 1H NMR (DMSO-d6): d
3.84 (s, 3H, OCH3); 6.95 (d, J = 8.1 Hz, 1H, CH arom);
7.09 (d, J = 8.1 Hz, 1H, CH arom); 7.19 (s, 1H, CH arom);
7.74 (s, 1H, CH); 9.97 and 12.42 (2 br s exchangeable with
D2O, 2H, OH and NH); 13C NMR (DMSO-d6): d 56.1
(OCH3); 114.6, 116.7, 124.6 (CH arom); 119.7 (5-C); 133.1
(CH methylidene); 124.9, 148.5, 149.9 (Cq arom); 167.9,
168.5 (CO). Anal. (C11H9NO4S). Calcd.: C, 52.59%; H,
3.59%; N, 5.58%. Found: C, 52.38%; H, 3.70%; N, 5.39%.
5-(3-Hydroxy-4-methoxybenzylidene)-2,4-thiazolidinedione
(8). Yield 60%; mp 260 °C. 1H NMR (DMSO-d6): d
3.82 (s, 3H, OCH3); 7.01–7.07 (m, 3H, CH arom); 7.62 (s,
1H, CH); 9.49 and 11.95 (2 br s exchangeable with D2O,
2H, OH and NH); 13C NMR (DMSO-d6): d 56.2 (OCH3);
113.0, 116.4, 123.9 (CH arom); 120.6 (5-C); 132.8
(CH methylidene); 126.2, 147.5, 150.6 (Cq arom); 168.0,
168.5 (CO). Anal. (C11H9NO4S). Calcd.: C, 52.59%; H,
3.59%; N, 5.58%. Found: C, 52.65%; H, 3.48%; N, 5.73%.
[5-(4-Hydroxy-3-methoxybenzylidene)-2,4-dioxothiazolidin-
29. Isolation of ALR2 from calf lenses and purification of this
enzyme have been previously described in ref. 30. IC50
values were determined from least squares analyses of the
linear portion of the log dose-inhibition curves by using
CalcuSyn software. Each curve was generated using at
least three concentrations of the tested compounds (added
as a solution in DMSO; final concentration of DMSO in
the incubation mixture was 1%) causing an inhibition
between 20% and 80%, with two replicates at each
concentration. For details concerning the assay procedure,
see Ref. 19.
30. Costantino, L.; Rastelli, G.; Vescovini, K.; Cignarella, G.;
Vianello, P.; Del Corso, A.; Cappiello, M.; Mura, U.;
Barlocco, D. J. Med. Chem. 1996, 39, 4396.
31. Chou, T.-C.; Hayball, M. P. CalcuSyn Software Version
1.1.1.; Biosoft: Cambridge, UK, 1996.
1
3-yl]acetic acid (9). Yield 35%; mp187–189 °C. H NMR
(DMSO-d6): d 3.84 (s, 3H, OCH3); 4.38 (s, 2H, NCH2);
6.98 (d, J = 8.1 Hz, 1H, CH arom); 7.18 (d, J = 8.1 Hz,
1H, CH arom); 7.22 (s, 1H, CH arom); 7.88 (s, 1H, CH);
10.07 (br s exchangeable with D2O, 1H, OH); 13C NMR
(DMSO-d6): d 44.9 (NCH2); 56.4 (OCH3); 115.1, 117.0,
125.0 (CH arom); 117.7 (5-C); 134.4 (CH methylidene);
124.9, 148.8, 150.7 (Cq arom); 166.3, 168.0, 169.5 (CO).
Anal. (C13H11NO6S) Calcd.: C, 50.49%; H, 3.56%;
N, 4.53%. Found: C, 50.67%; H, 3.70%; N, 4.38%.
32. Docking calculations were performed on a Pentium IV
2.8 GHz PC running Fedora Core 4 using the LigandFit
module in the Cerius2 software package, version 4.11 and
the cff1.02 force field. The software tool Catalyst version
4.11 was employed for ligand building, minimisation and
export as mol files. In order to select a well-suited protein–
inhibitor complex for docking from the PDB, which
contains about 50 entries for ALR2, several PDB search
criteria were applied to filter the numerous structures: EC
number 1.1.1.21, keyword Homo sapiens, existence of
[5-(3-Hydroxy-4-methoxybenzylidene)-2,4-dioxothiazolidin-
3-yl]acetic acid (10). Yield 40%; mp 215–218 °C. 1H NMR
(DMSO-d6): d 3.84 (s, 3H, OCH3); 4.36 (s, 2H, NCH2);
7.07–7.17 (m, 3H, CH arom); 7.84 (s, 1H, CH); 10.05 (br s
exchangeable with D2O, 1H, OH); 13C NMR (DMSO-d6):
d 42.8 (NCH2); 56.4 (OCH3); 113.0, 116.3, 124.4 (CH
arom); 117.6 (5-C); 134.9 (CH methylidene); 126.0, 147.5,
151.0 (Cq arom); 166.4, 167.6, 168.3 (CO). Anal.
(C13H11NO6S). Calcd.: C, 50.49%; H, 3.56%; N, 4.53%.
Found: C, 50.31%; H, 3.29%; N, 4.68%.
˚
ligands, X-ray resolution from 0 to 2 A and deposition
[5-(4-Carboxymethoxybenzylidene)-2,4-dioxothiazolidin-
3-yl]acetic acid (11). Yield 78%; mp 288 °C. 1H NMR
(DMSO-d6): d 4.36 (s, 2H, NCH2); 4.74 (s, 2H, OCH2);
6.96 (m, 2H, CH arom); 7.40 (m, 2H, CH arom); 7.78 (s,
1H, CH); 13C NMR (DMSO-d6): d 42.7 (NCH2); 65.1
(OCH2); 116.0, 132.8 (CH arom); 118.2 (5-C); 134.2 (CH
methylidene); 126.2, 160.3 (Cq arom); 165.6, 167.5, 168.5,
170.2 (CO). Anal. (C14H11NO7S). Calcd.: C, 49.85%; H,
3.26%; N, 4.15%. Found: C, 50.01%; H, 3.38%; N, 4.02%.
[5-(3-Carboxymethoxybenzylidene)-2,4-dioxothiazolidin-3-
yl]acetic acid (12). Yield 68%; mp 290 °C. 1H NMR
(DMSO-d6): d 4.41 (s, 2H, NCH2); 4.78 (s, 2H, OCH2);
7.10 (d, J = 9 Hz, 1H, CH arom); 7.22 (s, 1H, CH arom);
7.28 (d, J = 9 Hz, 1H, CH arom); 7.50 (dd, J = 9 and 9 Hz,
1H, CH arom); 8.00 (s, 1H, CH); 13C NMR (DMSO-d6): d
42.8 (NCH2); 65.0 (OCH2); 116.3, 117.8, 123.1, 131.0 (CH
arom); 121.7 (5-C); 134.2 (CH methylidene); 134.6, 158.7
(Cq arom); 165.4, 167.3, 168.4, 170.4 (CO). Anal.
(C14H11NO7S). Calcd.: C, 49.85%; H, 3.26%; N, 4.15%.
Found: C, 49.78%; H, 3.16%; N, 4.27%.
between 2004 and 2007. The remaining entries were
inspected for the existence of typical ALR2 inhibitor
binding characteristics. Protein preparation of complex
1Z3N for docking in the Cerius2 software package
included the elimination of the inhibitor and of the
solvent molecules (keeping the cofactor as part of the
protein structure), the addition of hydrogen atoms from
templates upon protein import and definition of the suitable
tautomeric form of active site His110 for hydrogen bonding
to acceptor functions in the ligands. Ligands were processed
for docking by defining bond orders and charges correctly,
that is, setting the partial charges on both oxygen atoms of
the carboxylic acids to a value of ꢀ1/2. Site definition
resulted from the co-crystallized ligand. Flexible docking
runs were carried out without using interaction filters. In the
docking preferences the number of Monte Carlo trials was
set to a fixed value of 5000 and the maximum number of
saved poses to 20, in the energy preferences the force field
selection was changed to CFF. Default settings were kept
for the remaining site definition and docking parameters.

R. Maccari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3886–3893
3893
The LigandScout software version 1.02 served as a tool for
visualisation of the interactions of the docked ligands with
the protein.
35. Catalyst, version 4.11; Accelrys Inc: San Diego, CA, 2005.
www.accelrys.com.
36. Venkatachalam, C. M.; Jiang, X.; Oldfield, T.;
Waldman, M. J. Mol. Graph. Model 2003, 21,
289.
33. Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.;
Bhat, T. N.; Weissig, H.; Shindyalov, I. N.; Bourne, P. E.
Nucleic Acids Res. 2000, 28, 235.
34. Van Zandt, M. C.; Jones, M. L.; Gunn, D. E.; Geraci, L.
S.; Jones, J. H.; Sawicki, D. R.; Sredy, J.; Jacot, J. L.;
DiCioccio, A. T.; Petrova, T.; Mitschler, A.; Podjarny, A.
D. J. Med. Chem. 2005, 48, 3141.
37. Wolber, G.; Langer, T. J. Chem. Inf. Comput. Sci. 2005,
45, 160, Available from: Inte:Ligand GmbH, www.inteli-
gand.com/ligandscout.
38. Cerius2, version 4.11, http.//www.accelrys.com; 4.8 ed.;
Accelrys Inc.: San Diego.