Pyrido[1,2-a]pyrimidin-4-one derivatives as a novel class of selective aldose reductase inhibitors exhibiting antioxidant activity

Article abstract of DOI:10.1021/jm070398a

2-Phenyl-pyrido[1,2-a]pyrimidin-4-one derivatives bearing a phenol or a catechol moiety in position 2 were tested as aldose reductase (ALR2) inhibitors and exhibited activity levels in the micromolar/submicromolar range. Introduction of a hydroxy group in

Full text of DOI:10.1021/jm070398a

J. Med. Chem. 2007, 50, 4917-4927
4917
Pyrido[1,2-a]pyrimidin-4-one Derivatives as a Novel Class of Selective Aldose Reductase
Inhibitors Exhibiting Antioxidant Activity
Concettina La Motta,*,† Stefania Sartini,† Laura Mugnaini,† Francesca Simorini,† Sabrina Taliani,† Silvia Salerno,†
Anna Maria Marini,† Federico Da Settimo,† Antonio Lavecchia,*,§ Ettore Novellino,§ Miriam Cantore,‡ Paola Failli,‡ and
Mario Ciuffi‡
Dipartimento di Scienze Farmaceutiche, UniVersita` di Pisa, Via Bonanno 6, 56126 Pisa, Italy, Dipartimento di Chimica Farmaceutica e
Tossicologica, UniVersita` di Napoli “Federico II”, Via D. Montesano, 49, 80131 Napoli, Italy, Dipartimento di Farmacologia Preclinica e
Clinica, UniVersita` di Firenze, Viale Pieraccini 6, 50139 Firenze, Italy
ReceiVed April 4, 2007
2-Phenyl-pyrido[1,2-a]pyrimidin-4-one derivatives bearing a phenol or a catechol moiety in position 2 were
tested as aldose reductase (ALR2) inhibitors and exhibited activity levels in the micromolar/submicromolar
range. Introduction of a hydroxy group in position 6 or 9 gave an enhancement of the inhibitory potency
(compare 18, 19, 28, and 29 vs 13 and 14). Lengthening of the 2-side chain to benzyl determined a general
reduction in activity. The lack or the methylation of the phenol or catechol hydroxyls gave inactive (10-
12, 21, 22, 25-27) or scarcely active (15, 17, 20) compounds, thus demonstrating that the phenol or catechol
hydroxyls are involved in the enzyme pharmacophoric recognition. Moreover, all the pyridopyrimidinones
displayed significant antioxidant properties, with the best activity shown by the catechol derivatives. The
theoretical binding mode of the most active compounds obtained by docking simulations into the ALR2
crystal structure was fully consistent with the structure-activity relationships in the pyrido[1,2-a]pyrimidin-
4-one series.
Introduction
research on the prevention of diabetes itself is promising,8,9
preventing the onset of the chronic biochemical and functional
alterations occurring in response to diabetic hyperglycemia is
still unfruitful. Different clinical studies10,11 demonstrate that
close control of blood glucose is significantly effective in
reducing these alterations, but even an optimal control of blood
glucose could not prevent their onset, suggesting that alternative
strategies are needed.
Diabetes mellitus is a metabolic disorder characterized by
high levels of blood glucose, resulting from defects in insulin
production, insulin action, or both. It is recognized as a public
health problem, as it affects a significant portion of the
population worldwide and is rising to pandemic proportions.
According to epidemiological studies, approximately 135 million
adults worldwide were diagnosed with diabetes in 1995, and
this number is expected to rise to at least 300 million by 2025,
with a 122% overall increase in the worldwide prevalence.1-5
Diabetes is a lifelong condition that seriously affects a
person’s quality of life. It can be successfully controlled by the
administration of insulin and potent oral hypoglycemics, but it
still remains the cause of significant morbidity and mortality
due to a progressive development of disabling complications.
Patients with diabetes are at a higher risk for cardiovascular
events, including strokes, and show accelerated formation of
severe atherosclerotic lesions in peripheral, coronary, and
cerebral arteries. They quickly develop visual impairment and
blindness due to cataracts and severe retinopathy, as well as
terminal renal diseases and different forms of nervous system
damage, including an impaired sensation of pain and physical
disability.6,7
Experimental and clinical evidence demonstrates that the
pathogenic mechanism leading to diabetic complications is
causally linked to an increased activity of the enzyme aldose
reductase (alditol/NADP+ a oxidoreductase, EC 1.1.1.21, ALR2).
This is the first enzyme of the polyol pathway and catalyzes
the NADPH-dependent reduction of glucose to sorbitol, which
is then oxidized to fructose by sorbitol dehydrogenase (L-iditol/
NAD+, 5-oxidoreductase, EC 1.1.1.14, SD). As ALR2 has a
low substrate affinity for glucose, the conversion of glucose to
sorbitol is generally nonsignificant in normoglycemic conditions.
In fact, ALR2 must compete directly with the hexokinase of
the glycolytic pathway for the utilization of glucose, and as the
substrate affinity of hexokinase is greater than that of ALR2,
glucose is preferentially phosphorylated with ATP by this
enzyme. Under hyperglycemic conditions, hexokinase is rapidly
saturated and the polyol pathway becomes activated. Sorbitol
is formed more rapidly than it is converted to fructose, and its
polarity hinders its penetration through membranes and subse-
quent removal from tissues by diffusion. The resulting elevated
The high and rising incidence of diabetes, its impact on
morbidity and mortality, and its high human and economic costs
clearly establish the prevention of this disease as an urgent
priority for national and international health authorities. While
a Abbreviations: PPP, 2-phenylpyrido[1,2-a]pyrimidin-4-one; ALR2,
aldose reductase; ALR1, aldehyde reductase; ARI, aldose reductase inhibitor;
NADPH, â-nicotinamide adenine dinucleotide phosphate reduced form;
NADP+, â-nicotinamide adenine dinucleotide phosphate; NADH, â-nico-
tinamide adenine dinucleotide reduced form; NAD+, â-nicotinamide adenine
dinucleotide; SD, sorbitol dehydrogenase; ATP, adenosine triphosphate;
ROS, reactive oxygen species; PKC, protein kinase C; MAPK, mitogen-
activated protein kinase; PARP, poly(ADP-ribose)polymerase; HNE, hy-
droxynonenal; AKR, aldo-keto reductase; PPA, polyphosphoric acid;
TBARS, thiobarbituric acid reactive substances; MD, molecular dynamics;
SAR, structure-activity relationship; MDA, malondialdehyde.
* To whom all correspondence should be addressed. Tel.: (+)-
390502219593 (C.L.M.); (+)39081678613 (A.L.V.). Fax: (+)390502219605
(C.L.M.); (+)39081678613 (A.L.V.). E-mail: lamotta@farm.unipi.it (C.L.M.);
lavecchi@unina.it (A.L.V.).
† Dipartimento di Scienze Farmaceutiche, Universita` di Pisa.
§ Dipartimento di Chimica Farmaceutica e Tossicologica, Universita`
“Federico II” di Napoli.
‡ Dipartimento di Farmacologia Preclinica e Clinica, Universita` di
Firenze.
10.1021/jm070398a CCC: $37.00 © 2007 American Chemical Society
Published on Web 09/11/2007

4918 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 20
Chart 1. Aldose Reductase Inhibitors
La Motta et al.
intracellular concentration of sorbitol increases cellular osmo-
larity, which in turn initiates a cascade of events that lead to
the development of long-term diabetic complications.12-14
In addition to the osmotic imbalance, an increase in the
activity of the polyol pathway during hyperglycemia causes a
substantial imbalance in the free cytosolic coenzyme ratios
NADPH/NADP+ and NAD+/NADH. This alteration in the
redox state of pyridine nucleotides induces a state of pseudo-
hypoxia, which contributes to the onset of hyperglycemic
oxidative stress through the accumulation of reactive oxygen
species (ROS). ROS, in turn, trigger activation of downstream
mechanisms, namely, protein kinase C (PKC) isoforms, mito-
gen-activated protein kinases (MAPKs), and poly(ADP-ribose)-
polymerase (PARP), as well as the inflammatory cascade, which
sustains the pathogenesis of diabetic complications.15-19 Fur-
thermore, the increase in fructose levels connected with the
polyol pathway activation accelerates the development of these
complications, because fructose and its metabolites are almost
10 times more potent nonenzymatic glycation agents than
glucose.
Besides these biochemical evaluations, newer approaches
involving genetic analysis conducted on transgenic and knockout
mice provide unambiguous evidence of the role of ALR2 in
the development of diabetic complications. Recent studies also
clearly demonstrate a correlation between overexpression of the
human ALR2 gene and the likelihood of the development of
complications among diabetic patients.20
Inhibition of ALR2 is therefore a useful therapeutic strategy
to prevent the onset, or at least delay, the progression and the
severity of diabetic complications.
specificity (Chart 1). However, none of them but epalrestat are
currently marketed. Products that appear to be promising during
in vitro studies or in trials with animal models often fail to
proceed any further either because of undesirable side effects
or as a result of poor efficacy. These effects are mainly due to
a lack of selectivity for other enzymes, especially the closely
related aldehyde reductase (EC 1.1.1.2, ALR1). ALR1 is
ubiquitously present in all tissues and plays a detoxification role,
as it shows substrate specificity toward toxic aldehydes such
as hydroxynonenal (HNE), methyl glyoxale and 3-deoxyglu-
cosone, which arise in large quantities from pathological
conditions connected with oxidative stress, as in hyperglyce-
mia.21 Both ALR1 and ALR2 belong to the aldo-keto reductase
(AKR) superfamily, and thus, share a common functional role
and the highest structural homology, with 51% identity in their
amino acid sequences. The least conserved residues are located
at the C-terminal end of the proteins in a region lining the
hydrophobic pocket of the active site called the “specificity
pocket”. This is responsible for substrate and inhibitory specific-
ity in AKRs22 and can be usefully exploited for the design of
selective aldose reductase inhibitors (ARIs).
The limited efficacy of currently known ARIs is generally
attributed to pharmacokinetic problems due to their absorption,
distribution, metabolism, and excretion properties, but it may
also be related to the post-translational modification of ALR2
activity, which is caused by the hyperglycemia-induced oxida-
tive stress. This modification involves oxidation of a critical
active cysteine thiol (C298), which regulates both substrate and
inhibitor binding. The resulting oxidized form of ALR2 shows
an increase in Km for aldehyde substrates and a marked reduction
in sensitivity to ARIs. If a large fraction of the enzyme
undergoes conversion to the activated form, the therapeutic
Many structurally different compounds have been shown to
inhibit this enzyme with various degrees of efficacy and

SelectiVe Aldose Reductase Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 20 4919
effectiveness of such compounds is largely compromised.23,24
Therefore, the development of more specific, clinically effective
therapies to prevent long-term diabetic complications should
necessarily combine ARIs and antioxidants to keep the enzyme
in the reduced form. Furthermore, given the extensive data
showing the detoxification role of ALR2 against oxidative
stress,25 the concurrent administration of antioxidants can
counterbalance its inhibition.
Scheme 1. Synthesis of Pyrido[1,2-a]pyrimidine Derivatives
10-29
Intrigued by this challenge, we directed our well-known
interest in the ARI field26-29 toward the identification of a novel
chemical class capable of combining a good ALR2 inhibitory
efficacy with antioxidant properties.
A thorough survey of pertinent literature revealed that the
only compounds described with this dual activity are flavonoids.
Since the mid-1970s, many papers have appeared reporting
extensive structure-activity relationships (SARs) for naturally
occurring and synthetic flavonoids. However, none of them
provided compounds useful for clinical development. Sometimes
they lack appreciable dual efficacy, as compounds exerting
significant antioxidant properties show poor ALR2 inhibitory
activity or vice versa. In the majority of cases, however,
flavonoids possess an inadequate therapeutic index, as they show
a broad spectrum of activity.30-33
On the basis of these data, we focused our attention on the
2-phenyl-pyrido[1,2-a]pyrimidin-4-one (PPP) scaffold, which
is a bioisoster of the flavone nucleus, with the aim of developing
a novel class of potent and selective ALR2 inhibitors structurally
related to flavonoids endowed with antioxidant properties.
Herein we present the synthesis and the extensive biological
evaluation of a number of derivatives substituted in positions
2, 6, and 9 of the PPP nucleus. Docking simulations of the most
active compounds into the human ALR2 binding site were also
carried out to rationalize the SARs observed and to guide,
perspectively, the design of new analogues.
Chemistry
The synthesis of the title compounds was performed as
outlined in Scheme 1. Inhibitors 10-12, 15-17, 20-22, and
25-27 were obtained in a one-step reaction by condensing the
appropriate 2-aminopyridine 1-3 with the suitable â-keto ester
4-9 in the presence of polyphosphoric acid (PPA) at 100 °C,
following a procedure previously reported for the synthesis of
the 4-oxo-4H-pyrido[1,2-a]pyrimidine ring system.34
The starting â-keto esters 7-9, with the exception of the
commercially available ones 4-6, were prepared in accordance
with a literature procedure35 involving acylation of Meldrum’s
acid with the appropriate acyl chloride, followed by ethanolysis
of the resulting acylated intermediate.
showed any appreciable ALR2 inhibitory activity. In particular,
derivative 13, bearing a 4-hydroxyl group, inhibited ALR2, with
an IC50 value in the micromolar range (IC50 4.56 µM), thus
exhibiting a 2-fold gain in efficacy with respect to quercetin,
our starting reference compound. As the pKa value of 13 (pKa
) 7.60) approaches the pH value of the enzymatic assay, this
compound probably acts in its anionic form, as usually occurs
with the inhibitors of ALR2. The insertion of an additional
hydroxyl group, as in 14, determines an increase in the acidity
of the resulting compound (pKa ) 6.80) and a more favorable
interaction with the enzyme anion binding site. Actually, 14
shows a 7-fold increase in the inhibitory potency when compared
with 13, and a 12-fold increase with respect to the reference
standard quercetin, with an IC50 value in the submicromolar
range (IC50 0.67 µM). The lack of activity of compound 10,
which is devoid of hydroxy substituents, as well as that of the
methoxylated compounds 11 and 12, in which proton dissocia-
tion is prevented, corroborates the hypothesized interaction of
the PPP derivatives 13 and 14 with the enzyme active site.
To verify the importance of the hydroxy substitution on the
heterocyclic scaffold, we initially continued our studies by
synthesizing derivatives 15-19, which bear this group in
position 6 of the core. Consistently with the previously described
series, the preferred substitution patterns appeared to be the
4-hydroxy and the 3,4-dihydroxy ones. Compounds 18 and 19
proved to be extremely potent, with IC50 values of 0.42 µM
and 0.10 µM, respectively. Moreover, they both show an
appreciable increase in inhibitory potency, specifically of 11-
Inhibitors 13, 14, 18, 19, 23, 24, 28, and 29 were obtained in
good yields from the corresponding methoxy derivatives 11,
12, 16, 17, 21, 22, 26, and 27 by ether cleavage performed with
boron tribromide in a dichloromethane solution.36
Results and Discussion
Biological Evaluation. The aim of our research project was
to identify potent and selective ARIs endowed with antioxidant
properties. All the compounds synthesized were, therefore, tested
for efficacy against ALR2 and selectivity for ALR1. Their
antioxidant properties were also verified. Tables 1 and 2 list
the results of these biological evaluations.
We started our investigation by synthesizing the pyrido[1,2-
a]pyrimidin-4-one compound 10 and its derivatives, 11-14,
bearing methoxy or hydroxyl groups in position 3 or 4 of the
2-phenyl ring. In this series of compounds, only 13 and 14

4920 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 20
La Motta et al.
Table 1. ALR2 Inhibition Data of Pyrido[1,2-a]pyrimidine Derivatives 10-29
a
a
ALR2 IC50
ALR1 IC50
(µM)
N
R1
R2
R3
(µM)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
C6H5
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
>10 µM
>10 µM
C6H4-4-OCH3
C6H3-3,4-diOCH3
C6H4-4-OH
C6H3-3,4-diOH
C6H5
C6H4-4-OCH3
C6H3-3,4-diOCH3
C6H4-4-OH
C6H3-3,4-diOH
CH2-C6H5
CH2-C6H4-4-OCH3
CH2-C6H3-3,4-diOCH3
CH2-C6H4-4-OH
CH2-C6H3-3,4-diOH
C6H5
C6H4-4-OCH3
C6H3-3,4-diOCH3
C6H4-4-OH
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
>10 µM
2.32 (2.05-2.78)
4.56 (3.62-5.47)
0.67 (0.52-0.83)
10 (7.40-12.1)
0.74 (0.59-0.89)
100 (79.1-111.5)
0.42 (0.36-0.52)
0.10 (0.077-0.13)
55 (42.2-65.6)
>10 µM
H
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
H
>10 µM
44 (32.2-55.2)
12 (9.30-14.95)
>10 µM
OH
OH
OH
OH
OH
H
H
H
H
>10 µM
>10 µM
2.07 (1.59-2.49)
0.45 0.34-0.56)
7.81 (5.47-10.15)
27
28
29
C6H3-3,4-diOH
quercetin
a IC50 (95% CL) values represent the concentration required to produce 50% enzyme inhibition.
fold for 18 and 7-fold for 19, compared with the parent
compounds 13 and 14. Therefore, it seems reasonable to assume
that the hydroxy group on the pyridopyrimidine nucleus, which
is involved in a hydrogen bond with the adjacent carbonyl group,
widens the planar area of the inhibitor’s scaffold, increasing
the interaction with the lipophilic amino acid residues surround-
ing the ALR2 active site. Surprisingly, in this series, also
compounds 15-17 displayed an appreciable inhibitory activity.
Compound 15, which displays a much higher pKa value (pKa
) 12.0), is a weak acidic compound that binds to the active
site in an undissociated form, seeing that the hydroxy group on
the pyrido[1,2-a]pyrimidine nucleus is involved in the stable
hydrogen bond with the adjacent carbonyl group. This holds
true also for compounds 16 and 17, in which any proton
dissociation is prevented. Nevertheless, these three compounds
show IC50 values in the micromolar/submicromolar range,
suggesting that an alternative way of binding to the active site
of ALR2 should be achieved. While the insertion of a methoxy
lipophilic substituent, as in 16, increases the interaction with
the enzyme (16, IC50 ) 0.74 µM, vs 15, IC50 ) 10 µM), an
unit, as well as the wider size of the molecule, disturbs the
favorable conformation for interaction with the enzyme.
Finally, we synthesized compounds 25-29 to investigate the
role of the hydroxy group in position 9 of the heterocyclic
nucleus. These derivatives generally displayed parallel SARs
to those of the parent compounds 10-14, seeing that compounds
25-27, which are devoid of any phenolic group, are inactive.
Compound 25 is a weak acid (pKa ) 10.2) and, consequently,
it interacts with ALR2 in an undissociated form. This should
hold true also for products 26 and 27, which are devoid of any
efficacy. It therefore seems reasonable to hypothesize that no
alternative way of binding to the active site of ALR2 can be
achieved by these compounds. On the contrary, the phenolic
28 (IC50 ) 2.07 µM) and 29 (IC50 ) 0.45 µM) proved to be
the only compounds that possessed an inhibitory activity, like
their analogues 13 and 14. These results demonstrate once more
the role of the hydroxyl groups on the 2-phenyl ring in anchoring
the PPP derivatives to the catalytic site of ALR2.
All the synthesized compounds were assayed for their ability
to inhibit ALR1, but none of them showed any appreciable
inhibitory properties (IC50 > 10 µM, Table 1); all of them
proved to be completely selective inhibitors of ALR2.
additional methoxy group, as in 17, is detrimental (17, IC50
)
100 µM, vs 15, IC50 ) 10 µM), probably because a steric clash
takes place with the walls of the enzyme hydrophobic cleft.
To explore the effects of increasing the distance between the
2-phenyl ring and the pyridopyrimidine nucleus, a methylene
spacer was inserted between them to give derivatives 20-24.
This distance appeared to be a crucial element, as the lengthened
products were less active than the corresponding 15-19. Also
in this series, the presence of the hydroxy substituents, as in 23
(IC50 ) 44.0) and 24 (IC50 ) 12.0), conferred the best activity,
allowing an effective interaction with the enzyme. In addition,
the alternative binding mode hypothesized for 15-17 may be
shared by the unsubstituted derivative 20, which preserves ALR2
inhibitory efficacy (IC50 ) 55 µM). However, in this subseries,
the methoxy derivatives 21 and 22 display a lower level of
activity, probably because the free rotation of the methylene
The antioxidant properties of the newly synthesized com-
pounds were also investigated. Antioxidants can act at different
levels in the cellular oxidative pathway, showing multiple
mechanisms of action, including scavenging of, or prevention
of the formation of, free radicals. We chose to evaluate activities
of pyrido[1,2-a]pyrimidin-4-ones by examining their effects on
hydroxyl radical-dependent lipoperoxidation induced in rat brain
homogenate by the oxidant system Fe(III)/ascorbic acid, which
initiates the Fenton reaction. For a proper comparison, two
standard references were used: R-tocopherol, which is an
endogenous chain-breaking antioxidant37 that protects cells from
diverse actions of ROS by donating its hydrogen atom, and
deferoxamine, which is an iron chelator38 that decreases the

SelectiVe Aldose Reductase Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 20 4921
Table 2. Effects of Pyrido[1,2-a]pyrimidin-4-one Derivatives,
R-Tocopherol, or Deferoxamine on the Production of Thiobarbituric
Reactive Substances (TBARS) in Rat Brain Homogenate Treated with
the Generating Reactive Oxygen Species Fe(III)-Ascorbic Acida
pathway, an increased quantity of ROS can be produced. It is
well accepted that this redox imbalance can worsen diabetes-
induced tissue damage at different levels (kidney, eye, vascu-
lature). Furthermore, treatment with antioxidants can keep the
enzyme in the reduced form, thus preventing the development
of drug resistance following its oxidative modification.
Molecular Modeling. To get a better comprehension of the
high ALR2 inhibitory potency of the newly synthesized
compounds at a molecular level and to propose a binding mode
that explains the above-described SARs, docking experiments
and molecular dynamics (MD) simulations were performed.
Compounds 16-19, 24, 28, and 29 were docked into the binding
pocket of the human ALR2/NADP+/tolrestat complex (PDB
entry code 2FZD).39 Docking was carried out using the
automated docking program AutoDock,40 which allows torsional
flexibility in the ligand and incorporates an efficient Lamarckian
genetic search algorithm together with an empirical free energy
function.
Recent crystallographic22,41 and molecular modeling studies26-28
have shown that carboxylic acid inhibitors bind to the active
site of ALR2 with the acidic function interacting with Y48,
H110, and W111, which are three key residues in binding and
catalysis.42,43 Thus, for inhibitors 13 and 14, the acidic nature
of the hydroxyls in the 4 (pKa 7.60) or 3 and 4 positions (pKa
6.80) of the pendant phenyl ring is fundamental to the inhibition
of ALR2. Indeed, the lack of these hydroxyls, as well as their
methylation, which prevents dissociation, leads to completely
inactive (10-12, 21, 22, 25-27) or scarcely active (15, 17, 20)
compounds, with the exclusion of 16. These findings are a clear
indication that 4-hydroxy and 3,4-dihydroxy derivatives exert
their inhibitory activity toward ALR2 in their anionic dissociated
form and that the anionic form produced by dissociation of these
hydroxyls could resemble the carboxylate function of carboxylic
acid inhibitors. For this reason, inhibitors 18, 19, 24, 28, and
29 were considered as dissociated (anionic) in the docking and
MD simulations. Moreover, theoretical and experimental struc-
tures of several 5-hydroxyflavones44-46 revealed the presence
of an intramolecular H-bond, with formation of a six-membered
ring involving the 5-OH and 4-CdO groups. For this reason,
during the docking of compounds 16-19 and 24, we blocked
the torsions involved in this intramolecular bond to prevent the
loss of this interaction.
% of inhibition
of control value
100
10
N
R1
R2
R3
µM
µM
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
C6H5
H
H
H
H
H
22.4
10.8
C6H4-4-OCH3
C6H3-3,4-diOCH3
C6H4-4-OH
C6H3-3,4-diOH
C6H5
C6H4-4-OCH3
C6H3-3,4-diOCH3
C6H4-4-OH
C6H3-3,4-diOH
CH2-C6H5
H
H
H
H
H
H
H
H
H
H
H
H
H
H
15.1
18.8
23.7
71.6
78.1
82.8
64.7
42.4
85.9
45.7
42.4
45.1
51.4
89.9
8.3
n.a.
n.a.
H
26.6
41.6
50.4
28.5
21.2
66.5
19.3
18.9
25.8
11.5
56.7
5.90
3.40
1.20
11.8
48.2
OH
OH
OH
OH
OH
OH
OH
CH2-C6H4-4-OCH3
CH2-C6H3-3,4-diOCH3 OH
CH2-C6H4-4-OH
CH2-C6H3-3,4-diOH
C6H5
C6H4-4-OCH3
C6H3-3,4-diOCH3
C6H4-4-OH
OH
OH
H
H
H
OH 11.9
OH 11.5
OH 8.80
OH 43.2
OH 81.5
18.2
26
27
28
H
H
29
C6H3-3,4-diOH
R-tocopherol
deferoxamine
72.0
a Basal value (rat brain homogenate): TBARS 2.3 ( 0.39 nmol/10 mg
wet weight rat brain (n ) 18). Control (rat brain homogenate plus FeCl3/
ascorbic acid): TBARS 7.31 ( 1.15 nmol/10 mg wet weight rat brain (n
) 18, P < 0.001 vs basal value, paired Student’s t test). Values are means
( standard errors. Each value is the mean of at least two different
determinations performed in triplicate. Absolute values were calculated by
performing for each experiment a reference curve using 1,1,3,3-tetramethox-
ypropane.
concentration of oxidative radicals by inhibiting their iron-
catalyzed production.
All the PPP products exhibited significant antioxidant proper-
ties. Actually, they inhibited the production of thiobarbituric
acid reactive substances (TBARS), an index of lipid peroxida-
tion, with different degrees of efficacy when assayed at a final
concentration of 100 µM, as shown in Table 2. In particular,
compounds lacking the hydroxy group on the heterocyclic core,
namely, 10-13, showed appreciable antioxidant properties,
which became excellent for derivative 14, characterized by the
ortho-dihydroxy substitution on the 2-phenyl ring. The insertion
of the hydroxy group in position 6 of the pyridopyrimidine
scaffold, as in compounds 15-24, greatly enhanced antioxidant
potency, which still remained marked at 10 µM. Also in this
series, the presence of the catechol moiety, as in derivatives 19
and 24, gave the best activity. On the contrary, shifting the
hydroxy substituent to position 9 of the scaffold led to a weakly
active series, derivatives 25-27, with the exception of 28 and
29, in which the presence of the 2-phenyl ring hydroxy groups
restored excellent antioxidant properties.
The 50 independent docking runs performed for each ligand
usually converged to a small number of different clusters
(“clusters” of results differing by less than 1.5 Å rmsd).
Generally, the top clusters (i.e., those with the most favorable
∆Gbind) were also associated with the highest frequency of
occurrence, suggesting a good convergence of the search
algorithm. Docking results of the top two clusters, ranked
according to the total docking energy ∆Gbind, are summarized
in Table 3.
Docking of 18, 19, 28, and 29 into the ALR2 crystal structure
revealed a very clear preference for two prevailing positions in
the binding pocket; these are designated binding orientations
A and B, as shown in Figure 1. Interestingly, both results are
located in the active site and occupy the same spatial position
as the crystallized inhibitor tolrestat.
The top-ranking results, corresponding to orientation A, have
their dissociated 4-hydroxyls or 3,4-dihydroxyls placed within
the anion binding site, where they make charge-assisted H-bonds
with the Oη hydrogen of catalytic Y48, while the pyridopyri-
midine nucleus is hosted in the specificity pocket. In this
position, the six-membered ring intramolecular H-bond formed
between the 5-OH and the 4-CdO of the pyridopyrimidine
Interestingly, for two of these compounds, namely, 19 and
29, the high activity as ALR2 inhibitors showed a good
correlation with their high activity as ROS scavenging mol-
ecules. This additional propriety can ameliorate the pharmaco-
logical profile of ALR2 inhibitors, because in this metabolic

4922 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 20
Table 3. Results of 50 Independent Docking Runs for Each Liganda
La Motta et al.
binding mode of the ligand, thus decreasing the relative stability
of the complex. This may be the reason for the reduced ALR2
inhibitory activity of 17.
cmpd
cluster
orientation
A
focc
∆Gbind
16
1
2
1
2
1
2
1
2
1
2
13
5
38
5
23
10
38
12
27
19
-9.7
-9.1
As shown in Table 3, AutoDock unambiguously predicted a
preference for binding orientation A only for 19, whereas for
inhibitors 18, 28, and 29, the difference between the estimated
free energy of binding for the two orientations A and B was
lower than 0.5 kcal/mol. Thus, AutoDock was not able to
indicate the preferred orientation for ALR2.
18
19
28
29
A
B
A
B
A
B
A
B
-10.7
-10.5
-10.8
-10.2
-10.7
-10.4
-10.8
-10.5
As AutoDock only treats the ligand in a flexible way, an
exhaustive protocol of MD simulations was applied, both to
relax the inhibitor/ALR2/NADP+ complexes and to evaluate
the dynamic behavior of the systems at 300 K. We hypothesized
that this procedure might eventually lead us to assess the
feasibility of the binding modes identified by AutoDock. For
this purpose, 1.2 ns of MD simulations were carried out on the
complexes formed between ALR2 and the two prevailing
docking positions A and B found for 19 and 29, chosen as
representative members of the whole series.
a The number of results contained in the top two clusters is given by the
frequency of occurrence, focc; ∆Gbind is the estimated free energy of binding
for the top two clusters and is given in kcal/mol.
nucleus in 19 (Figure 1a) and the 9-OH group in 29 (Figure
1c) point toward the bottom of the specificity pocket. Interest-
ingly, the second-ranking results, which correspond to orienta-
tion B, resemble the top ones, except for the pyridopyrimidine
nucleus, which is found to rotate by 180 degrees so as to project
the six-membered ring intramolecular H-bond (Figure 1b) and
the 9-OH (Figure 1d) toward the top of the specificity pocket.
Two prevailing binding positions are also found regarding
inhibitor 16, as can be seen in Figure 2a, and the first (∆Gbind
) -9.7 kcal/mol, found 5 times out of 50) closely resembles
the above-described binding orientation A, with the OCH3
oxygen in the 4 position of the pendant phenyl ring involved in
two H-bonds with both Y48 Oη and H110 Nꢀ2 hydrogens, and
the pyridopyrimidine nucleus placed in the specificity pocket,
with the six-membered ring intramolecular H-bond pointing
toward the bottom of the pocket. A significant alternative to
Chemico-physical parameters, such as temperature, density,
potential, and total energy, monitored during the MD simula-
tions, reached stable values for all the complexes after a few
hundred picoseconds. Monitoring of the progression of the root-
mean-square deviation (rmsd) of the coordinates of the CR atoms
with respect to the initial structure revealed a highly unstable
behavior in the complexes formed between ALR2 and both B
orientations of 19 (5.7 Å) and 29 (11.7 Å), whereas this behavior
was much more stable in the complexes formed between ALR2
and both A orientations of 19 (1.1 Å) and 29 (1.2 Å). This
indicated that the overall architectures of the macromolecular
complexes were preserved for the whole duration of the
simulations. Next, the relative dynamic stability of the A and
B orientations of 19 and 29 was monitored by computing the
rmsd of the ligands relative to their initial docked position and
by monitoring enzyme/ligand hydrogen bond distance time
series. Interestingly, some representative interacting distances
between the inhibitors analyzed and the ALR2 residues were
much more stable for the A binding modes than the B ones. In
particular, the interaction between the dissociated 3,4-dihy-
droxyls of both 19 and 29 with the catalytic residues Y48, H110,
and W111 was far more stable for A orientations with respect
to B ones throughout the MD simulations. Binding position A
of the two inhibitors showed a tendency to fluctuate within the
binding site, remaining firmly anchored to certain ALR2 amino
acids. These results clearly show that orientation A was
dynamically more stable during the MD simulations. This was
further confirmed by analyzing the rmsd value of the ligand
atoms between the starting and the last MD conformation.
Actually, this was as small as 2.2 Å for 19 and 2.1 Å for 29.
this position is given by the second-ranked solution (∆Gbind
)
-9.1 kcal/mol, found 13 times out of 50), which places the
pyridopyrimidine nucleus in the ALR2 catalytic site and the
pendant phenyl ring inside the specificity pocket (Figure 2b).
This position allows the formation of two H-bonds: the first
between the OCH3 oxygen in the 4 position of the pendant
phenyl ring and the OH hydrogen of T113 and the second one
between the 4-CdO oxygen of the pyridopyrimidine nucleus
and the W20 Nꢀ1 hydrogen. These results are in agreement with
the SAR data, showing that although 16 is a weak acidic
compound, it is still able to inhibit ALR2, with an IC50 of 0.74
µM.
To elucidate the reasons for the slight decrease in ALR2
inhibitory activity of the lengthened compounds 20-24, the
docking of 24 was carried out. The molecule occupies the same
space as 19 in its position A (∆Gbind ) -10.9 kcal/mol, found
41 times out of 50), with the hydroxyl in position 4 of the
pendant phenyl ring involved in a H-bond with Y48 and the
pyridopyrimidine system located within the specificity pocket.
However, the degree of dissociation of the 4-hydroxyl of 24 is
reasonably lower than that of the corresponding hydroxyl of
19 as a result of the methylene spacer that breaks the conjugation
between the pyridopyrimidine moiety and the 3,4-dihydroxy
phenyl group. This effect decreases the Coulombic component
of the H-bond formed between 24 and Y48 and explains the
lower ALR2 inhibitory activity of 23 and 24 in comparison with
that of 18 and 19, which form a charge-reinforced H-bond with
the Y48 side chain.
By contrast, position B of the two inhibitors showed a
remarkably different behavior; during MD simulation, the
ligand/ALR2/NADP+ complexes turned out to be highly
unstable, as the ligands left the ALR2 binding site and moved
toward the bulk of the solvent already after a few hundred
picoseconds. All of the most important interactions with the
enzyme residues were lost, and the final rmsd value of the ligand
atoms between the starting and the last MD conformations were
equal to 6.7 Å for 19 and 10.7 Å for 29. In this connection, it
should be remembered that, during MD simulation, the ligand
was allowed to move freely, looking for its best position at the
ALR2 binding site, and that the ALR2 binding site is totally
exposed to the solvent. Indeed, we added a 20 Å water cap
around the binding site, and a consequence of this might have
been that position B of 19 and 29, in which the hydrophilic
In the case of the scarcely active inhibitor 17, the automated
docking calculations did not converge toward a single binding
position. From a visual inspection of the 17/ALR2 complex, it
seems clear that the presence of the two methyl groups in the
3 and 4 positions of the pendant phenyl ring increases the steric
hindrance inside the binding cavity and changes the optimal

SelectiVe Aldose Reductase Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 20 4923
Figure 1. Compounds 19 (yellow) and 29 (white) bound into the active site of human ALR2 represented as a cyan Connolly surface. Binding
mode A is depicted at the top for 19 (a) and at the bottom for 29 (c). Binding mode B is illustrated at the top for 19 (b) and at the bottom for 29
(d). Cofactor NADP+ is shown in magenta. The side chains of the key receptor residues in proximity of the docked ligands are highlighted and
labeled. Nonpolar hydrogens have been removed for clarity. Hydrogen bonds are represented by yellow lines.
Figure 2. Compound 16 (orange) bound into the active site of human ALR2 represented as a cyan Connolly surface. The top-ranked position is
depicted on the left (a) and the second-ranked one is illustrated on the right (b). Cofactor NADP+ is shown in magenta. The side chains of the key
enzyme residues in proximity of the docked ligands are highlighted and labeled. Nonpolar hydrogens have been removed for clarity. Hydrogen
bonds are represented by yellow lines.
5-OH and 8-OH point toward the bulk of the solvent, tended to
escape from the binding site, looking for a more convenient
environment in the water.
F115, F122, L124, V130, L300, C303, and W219. It is known
that when an inhibitor binds to the ALR2 active site, a
conformational change occurs, opening a pocket, designated the
specificity pocket, which is localized between F122, W111,
L300, and A299.22 The specific opening of the pocket varies to
accommodate each inhibitor, producing an “induced fit”.
Because the residues lining this pocket are not conserved in
ALR1, the interactions are specific for ALR2,47-49 and as a
result, inhibitors with binding interactions in this “specificity
pocket” are generally highly selective for ALR2.22 The finding
that 19 and 29 bind in the above-mentioned hydrophobic pocket
accounts for the ALR2 selectivity observed for these inhibitors.
Of particular interest are the unsuccessful attempts to dock the
above compounds into the active site of ALR1, due to steric
hindrances between the compounds and the nonconserved
residues P301 (L300 in ALR2), T113 (Y115 in ALR2), and
R312 (missing in ALR2) from the C-terminal loop, which
further explains the high ALR2 selectivity observed for this
series of inhibitors.
Figure 3 depicts the MD snapshots characterizing binding
mode A of inhibitors 19 and 29 into the ALR2 binding site. As
discussed above, the dissociated hydroxyl in the 4 position of
the pendant phenyl ring interacts with Y48 and W111 in the
19/ALR2/NADP+ complex (Figure 3A), while it interacts with
Y48 and H110 in the 29/ALR2/NADP+ complex (Figure 3B).
Moreover, the dissociated hydroxyl in position 3 of 19 is within
the H-bond distance of the C298 SH group. These results are
particularly important because they rationalize the need for
dissociated free hydroxyls in positions 3 and 4 of the pendant
phenyl ring revealed by SAR data. Furthermore, this cor-
roborates the hypothesis that the dissociated hydroxyls success-
fully resemble the carboxylate group of several ALR2 inhibitors
that have been reported to date.
The pyridopyrimidine system of both inhibitors gets trapped
tightly in a hydrophobic pocket formed by residues W79, W111,

4924 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 20
La Motta et al.
Figure 3. MD snapshots of 19/ALR2//NADP+ (top, A) and 29/ALR2//NADP+ (bottom, B) complexes. Only aminoacids located within a 4 Å
distance of the bound ligand are displayed and labeled. Carbon atoms of 19 and 29 are yellow and white, respectively. Cofactor NADP+ is shown
in magenta. The hydrogen bonds discussed in the text are depicted as dashed yellow lines.
appropriate â-keto ester (4-9, 1.50 mmol) in PPA (2.00 g) was
heated at 100 °C for 1 h while stirring with a glass stick. The thick
syrup thus obtained was slowly poured into crushed ice, and the
resulting suspension was neutralized with 5% aqueous sodium
hydroxide. The solid precipitate was collected by filtration, washed
with water, and recrystallized (Supporting Information, Tables 1
and 2).
General Procedure for the Synthesis of 2-Phenylpyrido[1,2-
a]pyrimidin-4-ones 13-14, 6-Hydroxy-2-phenylpyrido[1,2-a]-
pyrimidin-4-ones 18-19, 2-Benzylpyrido[1,2-a]pyrimidin-4-ones
23-24, 9-Hydroxy-2-phenylpyrido[1,2-a]pyrimidin-4-ones 28-
29. A solution of boron tribromide (0.47 mL, 5.00 mmol) in 3 mL
of anhydrous dichloromethane was added dropwise to a well-stirred
solution of the appropriate methoxy derivative (14-21, 1.00 mmol)
in 10 mL of anhydrous dichloromethane, cooled at -10 °C. Once
the addition was complete, the reaction mixture was allowed to
warm to room temperature and maintained under stirring until the
disappearance of the starting material (10-24 h, TLC analysis).
The suspension was then carefully poured into crushed ice, and
the solid precipitate was collected by filtration, washed with water,
and recrystallized (Supporting Information, Tables 1 and 2).
Experimental Section
Chemistry. Melting points were determined using a Reichert
Ko¨fler hot-stage apparatus and are uncorrected. Infrared spectra
were recorded with a FT-IR spectrometer Nicolet/Avatar in Nujol
mulls. Routine nuclear magnetic resonance spectra were recorded
in DMSO-d6 solution on a Varian Gemini 200 spectrometer
operating at 200 MHz. Mass spectra were obtained on a Hewlett-
Packard 5988 A spectrometer using a direct injection probe and an
electron beam energy of 70 eV. Evaporation was performed in
vacuo (rotary evaporator). Analytical TLC was carried out on Merck
0.2 mm precoated silica gel aluminum sheets (60 F-254). Elemental
analyses were performed by our Analytical Laboratory and agreed
with theoretical values to within (0.4%.
The pyridines 2-aminopyridine, 1, 2-amino-4-hydroxypyridine,
2, 2-amino-3-hydroxypyridine, 3, and the â-keto esters ethyl
benzoylacetate, 4, ethyl 4-methoxybenzoylacetate, 5, and ethyl 3,4-
dimethoxybenzoylacetate, 6, used to obtain the target inhibitors,
were from Aldrich and Fluka. The following compounds were
prepared in accordance with a reported procedure: ethyl 3-oxo-4-
phenylbutanoate,50 7, ethyl 4-(4-methoxyphenyl)-3-oxobutanoate,50
8, and ethyl 4-(3,4-dimethoxyphenyl)-3-oxobutanoate,51 9.
General Procedure for the Synthesis of 2-Phenylpyrido[1,2-
a]pyrimidin-4-ones 10-12, 6-Hydroxy-2-phenylpyrido[1,2-a]-
pyrimidin-4-ones 15-17, 2-Benzylpyrido[1,2-a]pyrimidin-4-ones
20-22, and 9-Hydroxy-2-phenylpyrido[1,2-a]pyrimidin-4-ones
25-27. A mixture of 2-aminopyridine (1-3, 1.00 mmol) and the
Biology. Materials and Methods. ALR2 and ALR1 were
obtained from Sprague-Dawley albino rats, 120-140 g, b.w.,
supplied by Harlan Nossan, Italy. To minimize cross-contamination
between ALR2 and ALR1 in the enzyme preparation, rat lens, in
which ALR2 is the predominant enzyme, and kidney, where ALR1

SelectiVe Aldose Reductase Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 20 4925
shows the highest concentration, were used for the isolation of
ALR2 and ALR1, respectively.
nation of the IC50 values was performed by linear regression analysis
of the log-dose response curve, which was generated using at least
four concentrations of the inhibitor causing an inhibition between
20% and 80%, with two replicates at each concentration. The 95%
confidence limits (95% CL) were calculated from t values for
n-2, where n is the total number of determinations.
Pyridine coenzyme, D,L-glyceraldehyde, sodium D-glucuronate,
quercetin, ascorbic acid, ferric chloride, thiobarbituric acid, 1,1,3,3-
tetramethoxypropane, R-tocopherol, and deferoxamine were from
Sigma-Aldrich. All other chemicals were of reagent grade.
Enzyme Preparation. Aldose Reductase. A purified rat lens
extract was prepared in accordance with the method of Hayman
and Kinoshita,52 with slight modifications. Lenses were quickly
removed from rats following euthanasia and were homogenized
(Glas-Potter) in three volumes of cold deionized water. The
homogenate was centrifuged at 12 000 rpm at 0-4 °C for 30 min.
Saturated ammonium sulfate solution was added to the supernatant
fraction to form a 40% solution, which was stirred for 30 min at
0-4 °C and then centrifuged at 12 000 rpm for 15 min. Following
this same procedure, the recovered supernatant was subsequently
fractionated with saturated ammonium sulfate solution using first
a 50% and then a 75% salt saturation. The precipitate recovered
from the 75% saturated fraction, possessing ALR2 activity, was
redissolved in 0.05 M NaCl and dialyzed overnight in 0.05 M NaCl.
The dialyzed material was used for the enzymatic assay.
Aldehyde Reductase. Rat kidney ALR1 was prepared in
accordance with a previously reported method.53 Kidneys were
quickly removed from normal killed rats and homogenized (Glas-
Potter) in three volumes of 10 mM sodium phosphate buffer, pH
7.2, containing 0.25 M sucrose, 2.0 mM EDTA dipotassium salt,
and 2.5 mM â-mercaptoethanol. The homogenate was centrifuged
at 12 000 rpm at 0-4 °C for 30 min and the supernatant was
subjected to a 40-75% ammonium sulfate fractionation, following
the same procedure previously described for ALR2. The precipitate
obtained from the 75% ammonium sulfate saturation, possessing
ALR1 activity, was redissolved in 50 volumes of 10 mM sodium
phosphate buffer, pH 7.2, containing 2.0 mM EDTA dipotassium
salt and 2.0 mM â-mercaptoethanol and was dialyzed overnight
using the same buffer. The dialyzed material was used in the
enzymatic assay.
Enzymatic Assays. The activity of the two test enzymes was
determined spectrophotometrically by monitoring the change in
absorbance at 340 nm, which is due to the oxidation of NADPH
catalyzed by ALR2 and ALR1. The change in pyridine coenzyme
concentration/min was determined using a Beckman DU-64 kinetics
software program (Solf Pack TM Module).
ALR2 activity was assayed at 30 °C in a reaction mixture
containing 0.25 mL of 10 mM d,L-glyceraldehyde, 0.25 mL of 0.104
mM NADPH, 0.25 mL of 0.1 M sodium phosphate buffer (pH 6.2),
0.1 mL of enzyme extract, and 0.15 mL of deionized water in a
total volume of 1 mL. All the above reagents, except D,L-
glyceraldehyde, were incubated at 30 °C for 10 min; the substrate
was then added to start the reaction, which was monitored for 5
min. Enzyme activity was calibrated by diluting the enzymatic
solution to obtain an average reaction rate of 0.011 ( 0.0010
absorbance units/min for the sample.
Thiobarbituric Acid Reagent Substance Assay. Freshly iso-
lated rat brain was homogenized in 10% (w/v) phosphate buffer,
pH 7.4. FeCl3 (20 µM) and 100 µM ascorbic acid (final concentra-
tions) were added to 100 µL of rat brain homogenate, in the absence
or presence of the test compounds or reference scavenger molecules,
and were diluted to 1 mL with phosphate buffer, pH 7.4. Samples
were incubated for 30 min at 37 °C in a water bath under slight
stirring, and were then mixed with 0.5 mL of thiobarbituric acid
(1% w/v in 0.05 N NaOH) and 0.5 mL of 25% v/v HCl. The
mixture was boiled for 10 min and, after cooling in an ice-cold
water bath, extraction was performed with 3 mL of n-butanol. After
liquid-phase separation (centrifugation at 2000 g for 10 min), the
absorbance of the pink organic phase was spectrophotometrically
evaluated at 532 nm and TBARS was expressed as nmoles of
malondialdehyde (MDA)/10 mg of rat brain tissue (wet weight),
using a curve with 1,1,3,3-tetramethoxypropane as the standard
reference.54
Computational Chemistry. Molecular modeling and graphics
manipulations were performed using the SYBYL55 and UCSF
CHIMERA software packages,56 running them on a Silicon
Graphics Tezro R16000 workstation. Model building of inhibitors
16-19, 28, and 29 was accomplished by using the TRIPOS force
field57 available within SYBYL. Energy minimizations and MD
simulations were performed with the AMBER 8.0 program,58 using
the parm99 force field.59
Docking Simulations. Docking was performed with version 3.05
of the program AutoDock,40 which combines a rapid energy
evaluation through precalculated grids of affinity potentials with a
variety of search algorithms to find suitable binding positions for
a ligand on a given protein. While the protein is required to be
rigid, the program allows torsional flexibility in the ligand.
Ligand Setup. Molecular models of compounds 16-19, 28, and
29 were constructed using standard bond lengths and bond angles
of the SYBYL fragment library. Hydroxyls in the 3 and 4 position
of the ligand pendant phenyl ring were taken as dissociated, on
account of the low pKa value of these groups. Geometry optimiza-
tions were realized with the SYBYL/MAXIMIN2 minimizer by
applying the BFGS (Broyden, Fletcher, Goldfarb and Shannon)
algorithm60 and setting an rms gradient of the forces acting on each
atom of 0.05 kcal/mol Å as the convergence criterion. Atomic
charges were assigned using the Gasteiger-Marsili formalism,61
which is the type of atomic charge used in calibrating the AutoDock
empirical free energy function.
Protein Setup. The crystal structure of human ALR2 complexed
with tolrestat (PDB entry code 2FZD),39 recovered from Brookhaven
Protein Database,62 was used in the docking experiments.
Although the inhibition assays on our compounds were conducted
on rat ALR2, the use of the human ALR2 crystal structure for
docking is justified by the following facts: (i) the crystal structure
of rat ALR2 is unknown; (ii) the human and rat sequences of this
enzyme are characterized by 81% identity and 89% homology;63
(iii) all active-site residues, including those of the specificity pocket,
are largely conserved across the ALR2 isoforms so far sequenced.
The structure was set up for docking as follows: polar hydrogens
were added by using the BIOPOLYMERS module within the
SYBYL program (residues Arg, Lys, Glu, and Asp were considered
ionized, while all His were considered to be neutral by default with
hydrogen atoms fixed at the Nꢀ2 atom), Kollman united atom partial
charges were assigned, and all waters were removed.
ALR1 activity was determined at 37 °C in a reaction mixture
containing 0.25 mL of 20 mM sodium D-glucuronate, 0.25 mL of
0.12 mM NADPH, 0.25 mL of dialyzed enzymatic solution, and
0.25 mL of 0.1 M sodium phosphate buffer (pH 7.2) in a total
volume of 1 mL. The enzyme activity was calibrated by diluting
the dialyzed enzymatic solution to obtain an average reaction rate
of 0.015 ( 0.0010 absorbance/min for the sample.
Enzymatic Inhibition. The inhibitory activity of the newly
synthesized compounds against ALR2 and ALR1 was assayed by
adding 0.1 mL of the inhibitor solution to the reaction mixture
described above. All the inhibitors were solubilized in water and
the solubility was facilitated by adjustment to a favorable pH. After
complete solution, the pH was readjusted to 7. To correct for the
nonenzymatic oxidation of NADPH and for absorption by the
compounds tested, a reference blank containing all the above assay
components except the substrate was prepared. The inhibitory effect
of the new derivatives was routinely estimated at a concentration
of 10-4 M. Those compounds found to be active were tested at
additional concentrations between 10-5 and 10-8 M. The determi-
AutoDock Docking. Docking of compounds 16-19, 28, and
29 to ALR2 was carried out using the empirical free energy function
and the Lamarckian genetic algorithm, applying a standard protocol
with an initial population of 50 randomly placed individuals, a
maximum number of 1.5 × 106 energy evaluations, a mutation rate
of 0.02, a crossover rate of 0.80, and an elitism value of 1.

4926 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 20
La Motta et al.
(3) Wild, S.; Roglic, G.; Green, A.; Sicree, R.; King, H. Global
Prevalence of Diabetes: Estimates for the Year 2000 and Projections
for 2030. Diabetes Care 2004, 27, 1047-1053.
(4) Zimmet, P. Z.; Alberti, K. G. M. M.; Shaw, J. Global and Societal
Implications of the Diabetic Epidemic. Nature 2001, 414, 782-787.
(5) American Diabetes Association. Economic costs of Diabetes in the
U.S. in 2002. Diabetes Care 2003, 26, 917-932.
(6) Brownlee, M. Biochemical and Molecular Cell Biology of Diabetic
Complications. Nature 2001, 414, 813-820.
(7) Altan, V. M. The Pharmacology of Diabetic Complications. Curr.
Med. Chem. 2003, 10, 1317-1327.
(8) Knowler, W. C.; Narayan, K. M. V.; Hanson, R. L.; Nelson, R. G.;
Bennett, P. H.; Tuomilehto, J.; Schersten, B.; Pettitt, D. J. Preventing
Non-Insulin-Dependent Diabetes Mellitus. Diabetes 1995, 22, 483-
488.
(9) The Diabetes Prevention Program Research Group. The Diabetes
Prevention Program: Design and Methods for a Clinical Trial in the
Prevention of Type 2 Diabetes. Diabetes Care 1999, 22, 623-634.
(10) The Diabetes Control and Complications Trial Research Group. The
Effect of Intensive Treatment of Diabetes on the Development and
Progression of Long-Term Complications in Insulin-Dependent
Diabetes Mellitus. N. Engl. J. Med. 1993, 329, 977-986.
(11) U. K. Prospective Diabetes Study Group. Intensive Blood-Glucose
Control with Sulfonylureas of Insulin Compared with Conventional
Treatment and Risk of Complications in Patients with Type 2
Diabetes (UKPDS 33). Lancet 1998, 352, 837-853.
Proportional selection was used, where the average of the worst
energy was calculated over a window of the previous 10 genera-
tions. For the local search, the so-called pseudo Solis and Wets
algorithm was applied, using a maximum of 300 iterations. The
probability of performing local search on an individual in the
population was 0.06, and the maximum number of consecutive
successes or failures before doubling or halving the local search
step size was 4.
Fifty independent docking runs were carried out for each ligand.
Results differing by less than 1 Å in positional rmsd were clustered
together and were represented by the result with the most favorable
free energy of binding (∆Gbind). Finally, the compounds were set
up for docking with the help of AutoTors, the main purpose of
which is to define the torsional degrees of freedom to be considered
during the docking process. All torsion angles for each compound
were considered flexible. Solvation parameters were added to the
final protein file by using the ADDSOL utility of AutoDock. The
grid maps representing the proteins in the actual docking process
were calculated with AutoGrid. The grids (one for each atom type
in the ligand plus one for electrostatic interactions) were chosen to
be sufficiently large to include not only the active site but also
significant portions of the surrounding surface. The dimensions of
the grids were thus 30 Å × 30 Å × 30 Å, with a spacing of 0.375
Å between the grid points.
MD Simulations. All MD simulations were performed using
the SANDER module in the AMBER suite of programs. An all-
atom representation of the system was used, employing the Cornell
et al.59 force field to assign parameters for the standard amino acids.
General AMBER force field (GAFF) parameters were assigned to
ligands and NADP+ cofactor, while the partial charges were
calculated using the AM1-BCC method as implemented in the
ANTECHAMBER suite of AMBER. 19/ALR2/NADP+ and 29/
ALR2/NADP+ complexes were solvated with a water cap (TIP3P)64
up to 30 Å from the inhibitor. A “belly” region of 15 Å was then
defined around the ligand. During the simulations, those “belly”
atoms that are less than 15 Å from the ligand were allowed to move,
while all atoms beyond 15 Å were held rigid.
The initial complexes were energy-minimized for 10 500 steps
using combined steepest descent and conjugate gradient methods
until a convergence value of 0.001 kcal/Å‚mol. Upon minimization,
the water cap was equilibrated for 10 picoseconds (ps) to allow
cavities in the protein to become solvated. The entire “belly” region
was then equilibrated for 100 ps with the system warmed to 300 K
over the initial 0.2 ps equilibrated at 300 K for another 25 ps. A
subsequent production run was performed at a constant temperature
of 300 K and a constant pressure of 1 atm, giving a total simulation
time of 1.2 ns. The time interval was set to 2 fs. The SHAKE
method65 was applied to constrain all of the covalent bonds
involving hydrogen atoms. During the simulation, nonbonded
cutoffs were set to 12 Å. Neither constraints nor restraints were
applied to the complexes during the MD simulations, allowing the
inhibitor molecule to move freely between the protein surface and
the bulk of the solvent. The CARNAL module of AMBER was
used to check some structural properties (rmsd, hydrogen bonds).
The hydrogen bond criterion was a maximum donor-acceptor
distance of 3.5 Å and a minimum donor-proton-acceptor angle
of 120°.
(12) Kador, P. F. The role of Aldose Reductase in the Development of
Diabetic Complications. Med. Res. ReV. 1998, 8, 325-352.
(13) Tomlinson, D. R.; Stevens, E. J.; Diemel, L. T. Aldose Reductase
Inhibitors and their Potential for the Treatment of Diabetic Complica-
tions. Trends Pharmacol. Sci. 1994, 15, 293-297.
(14) Yabe-Nishimura, C. Aldose Reductase in Glucose Toxicity:
A
Potential Target for the Prevention of Diabetic Complications.
Pharmacol. ReV. 1998, 50, 21-33.
(15) Williamson, J. R.; Chang, K.; Frangos, M.; Hasan, K. S.; Ido, Y.;
Kawamura, T.; Nyengaard, J. R.; van der Enden, M.; Kilo, C.; Tilton,
R. G. Hyperglycemic Pseudohypoxia and Diabetic Complications.
Diabetes 1993, 42, 801-813.
(16) Ido, Y.; Williamson, J. R. Hyperglycemic Cytosolic Reductive Stress
“Pseudohypoxia”: Implications for Diabetic Retinopathy. InVest.
Ophthalmol. Visual Sci. 1997, 38, 1467-1470.
(17) Purves, T.; Middlemas, A.; Agthon, S.; Jude, E. B.; Boulton, A. J.;
Fernyhough, P.; Tomlinson, D. R. A Role for Mitogen-Activated
Protein Kinases in the Etiology of Diabetic Neuropathy. FASEB J.
2001, 15, 2508-2514.
(18) Obrosova, I. G.; Li, F.; Abatan, O. I.; Forsell, M. A.; Komjati, K.;
Pacher, P.; Szabo, C.; Stevens, M. J. Role of Poly(ADP-ribose)-
polymerase Activation in Diabetic Neuropathy. Diabetes 2004, 53,
711-720.
(19) Joussen, A. M.; Poulaki, V.; Le, M. L.; Koizumi, K.; Esser, C.;
Janicki, H.; Scharaermeyer, U.; Kociok, N.; Fauser, S.; Kirchhof,
B.; Kern, T. S.; Adamis, A. P. A Central Role for Inflammation in
the Pathogenesis of Diabetic Retinopathy. FASEB J. 2004, 18, 1450-
1452.
(20) Chung, S. S.; Chung, S. K. Genetic Analysis of Aldose Reductase
in Diabetic Complications. Curr. Med. Chem. 2003, 10, 1375-1387.
(21) Carper, D.; Wiston, G.; Nishimura, C.; Graham, C.; Watanabe, K.;
Fuji, Y.; Hayaishi, H.; Hayaishi, O. A Superfamily of NADPH-
Dependent Reductases in Eukaryotes and Prokaryotes. Exp. Eye Res.
1988, 49, 377-388.
(22) Urzhumtsev, A.; Tete-Favier, F.; Mitscler, A.; Barbanton, J.; Barth,
P.; Urzhumgtseva, L.; Biellmann, J. F.; Podjarny, A. D.; Moras, D.
A “Specificity” Pocket Inferred from the Crystal Structures of the
Complexes of Aldose Reductase with the Pharmaceutically Important
Inhibitors Tolrestat and Sorbinil. Structure 1997, 5, 601-612.
(23) Cappiello, M.; Voltarelli, M.; Cecconi, I.; Vilardo, P. G.; Dal Monte,
M.; Marini, I.; Del Corso, A.; Wilson, D. K.; Quiocho, F. A.; Petrash,
J. M.; Mura, U. Specifically Targeted Modification of Human Aldose
Reductase by Physiological Disulfides. J. Biol. Chem. 1996, 271,
33539-33544.
Supporting Information Available: Tables 1 and 2, including
physical and spectral data of compounds 11, 13, 15-17, and 19-
29, and analytical data of compounds 11, 13, 15-17, and 19-29.
This material is available free of charge via the Internet at http://
pubs.acs.org.
(24) Grimshaw, C. E.; Lai, C. J. Oxidized Aldose Reductase: In Vivo
Factor, Not In Vitro Artifact. Arch. Biochem. Biophys. 1996, 327,
89-97.
(25) Srivastava, S. K.; Ramana, K. V.; Bhatnagar, A. Role of Aldose
Reductase and Oxidative Damage in Diabetes and the Consequent
Potential for Therapeutic Options. Endocr. ReV. 2005, 26, 380-392.
(26) Da Settimo, F.; Primofiore, G.; Da Settimo, A.; La Motta, C.; Taliani,
S.; Simorini, F.; Novellino, E.; Greco, G.; Lavecchia, A.; Boldrini,
E. [1,2,4]Triazino[4,3-a]benzimidazole Acetic Acid Derivatives: A
New Series of Selective Aldose Reductase Inhibitors. J. Med. Chem.
2001, 44, 4359-4369.
References
(1) King, H.; Aubert, R. E.; Herman, W. H. Global Burden of Diabetes,
1995-2025. Prevalence, Numerical Estimates, and Projection. Dia-
betes Care 1998, 21, 1414-1431.
(2) Boyle, J. P.; Honeycutt, A. A.; Venkat Narayan, K. M.; Hoerger, T.
J.; Geiss, L. S.; Chen, H.; Thompson, T. J. Projection of Diabetes
Burden Through 2025. Diabetes Care 2001, 24, 1936-1940.

SelectiVe Aldose Reductase Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 20 4927
(27) Da Settimo, F.; Primofiore, G.; Da Settimo, A.; La Motta, C.;
Simorini, F.; Novellino, E.; Greco, G.; Lavecchia, A.; Boldrini, E.
Novel, Highly Potent Aldose Reductase Inhibitors: Cyano-(2-oxo-
2,3-dihydroindol-3-yl)-acetic Acid Derivatives. J. Med. Chem. 2003,
46, 1419-1428.
(28) Da Settimo, F.; Primofiore, G.; La Motta, C.; Sartini, S.; Taliani, S.;
Simorini, F.; Marini, A. M.; Lavecchia, A.; Novellino, E.; Boldrini,
E. Naphtho[1,2-d]isothiazole Acetic Acid Derivatives as a Novel
Class of Selective Aldose Reductase Inhibitors. J. Med. Chem. 2005,
48, 6897-6907.
(29) Da Settimo, F.; Primofiore, G.; La Motta, C.; Salerno, S.; Novellino,
E.; Greco, G.; Lavecchia, A.; Laneri, S.; Boldrini, E. Spirohydantoin
Derivatives of Thiopyrano[2,3-b]pyridin-4(4H)-one as Potent In Vitro
and In Vivo Aldose Reductase Inhibitors. Bioorg. Med. Chem. 2005,
13, 491-499.
(30) Kawanishi, K.; Ueda, H.; Moriyasu, M. Aldose Reductase Inhibitors
from the Nature. Curr. Med. Chem. 2003, 10, 1353-1374.
(31) Costantino, L.; Rastelli, G.; Gamberini, M. C.; Vinson, J. A.; Bose,
P.; Iannone, A.; Staffieri, M.; Antolini, L.; Del Corso, A.; Mura, U.;
Albasini, A. 1-Benzopyran-4-one Antioxidants as Aldose Reductase
Inhibitors. J. Med. Chem. 1999, 42, 1881-1893.
(48) Rees-Milton, K. J.; Jia, Z.; Green, N. C.; Bhatia, M.; El-Kabbani,
O.; Flynn, T. G. Aldehyde Reductase: The Role of C-Terminal
Residues in Defining Substrate and Cofactor Specificities. Arch.
Biochem. Biophys. 1998, 355, 137-144.
(49) El-Kabbani, O.; Carper, D. A.; McGowan, M. H.; Devedjiev, Y.;
Rees-Milton, K. J.; Flynn, T. G. Studies on the Inhibitor-Binding
Site of Porcine Aldehyde Reductase: Crystal Structure of the
Holoenzyme-Inhibitor Ternary Complex. Proteins 1997, 29, 186-
192.
(50) Svenstrup, N.; Simonsen, K. B.; Thorup, N.; Brodersen, J.; Dehaen,
W.; Becher, J. A Pyrazole to Furan Rearrangement. Thermolysis of
5-Azido-4-formylpyrazoles. J. Org. Chem. 1999, 64, 2814-2820.
(51) Shin, H.; Choi, B. S.; Lee, K. K.; Choi, H-W.; Chang, J. H.; Lee, K.
W.; Nam, D. H.; Kim, N. S. Efficient Activation of Zinc: Application
of the Blaise Reaction to an Expedient Synthesis of a Statin
Intermediate. Synthesis 2004, 16, 2629-2632.
(52) Hayman, S.; Kinoshita, J. H. Isolation and Properties of Lens Aldose
Reductase. J. Biol. Chem. 1965, 240, 877-882.
(53) Ward, W. H. J.; Sennitt, C.M.; Ross, H.; Dingle, A.; Timmus, D.;
Mirrless, D. J.; Tuffin, D. P. Ponalrestat, A Potent and Specific
Inhibitor of Aldose Reductase. Biochem. Pharmacol. 1990, 39, 337-
346.
(54) Cantore, M.; Siano, S.; Coronnello, M.; Mazzetti, L.; Franchi-Micheli,
S.; Ciuffi, M.; Failli, P. Pirenoxine Prevents Oxidative Effects of
Argon Fluoride Excimer Laser Irradiation in Rabbit Corneas:
Biochemical, Histological, and Cytofluorimetric Evaluation. J. Pho-
tochem. Photobiol., B. 2005, 75, 34-42.
(55) SYBYL, Molecular Modeling System, version 7.1; TRIPOS Assoc.,
St. Louis, MO.
(56) Huang, C. C.; Couch, G. S.; Pettersen, E. F.; Ferrin, T. E. Chimera:
An Extensible Molecular Modeling Application Constructed Using
Standard Components. Pac. Symp. Biocomput. 1996, 1, 724 (http://
www.cgl.ucsf.edu/chimera).
(57) Vinter, J. G.; Davis, A.; Saunders, M. R. Strategic Approaches to
Drug Design. 1. An Integrated Software Framework for Molecular
Modelling. J. Comput.-Aided Mol. Des. 1987, 1, 31-55.
(58) Case, D. A.; Darden, T. A.; Cheatham, T. E., III; Simmerling, C. L.;
Wang, J.; Duke, R. E.; Luo, R.; Merz, K. M.; Wang, B.; Pearlman,
D. A.; Crowley, M.; Brozell, S.; Tsui, V.; Gohlke, H.; Mongan, J.;
Hornak, V.; Cui, G.; Beroza, P.; Schafmeister, C.; Caldwell, J. W.;
Ross, W. S.; Kollman, P. A. AMBER 8; University of California:
San Francisco, CA, 2004.
(59) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.,
Jr.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.;
Kollman, P. A. A Second Generation Force Field for the Simulation
of Proteins, Nucleic Acids, and Organic Molecules. J. Am. Chem.
Soc. 1995, 117, 5179-5197.
(60) Head, J.; Zerner, M. C. A Broyden-Fletcher-Goldfarb-Shannon
Optimization Procedure for Molecular Geometries. Chem. Phys. Lett.
1985, 122, 264-274.
(61) Gasteiger, J.; Marsili, M. Iterative Partial Equalization of Orbital
ElectronegativitysA Rapid Access to Atomic Charges. Tetrahedron
1980, 36, 3219-3228.
(62) Bernstein, F. C.; Koetzle, T. F.; Williams, G. J. B.; Meyer, E. F.,
Jr.; Brice, M. D.; Rodgers, J. R.; Kennard, O.; Shimanouchi, T.;
Tasumi, T. The Protein Data Bank: A Computer Based Archival
File for Macromolecular Structures. J. Mol. Biol. 1977, 112, 535-
542.
(63) Database searching (SWISS-PROT), sequence alignment, and analysis
of rat and human ALR2 sequences were carried out using FASTA
(Pearson, W. R. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 2444-2448)
and BLAST programs (Wang, S.; Pak, Y. J. Phys. Chem. B 2000,
104, 354-359).
(32) Lim, S. S.; Jung, S. H.; Ji, J.; Shin, K. H.; Keum, S. R. Synthesis of
Flavonoids and their Effects on Aldose Reductase and Sorbitol
Accumulation in Streptozotocin-Induced Diabetic Rat Tissues. J.
Pharm. Pharmacol. 2001, 53, 653-668.
(33) Matsuda, H.; Morikawa, T.; Toguchida, I.; Yoshikawa, M. Structural
Requirements of Flavonoids and Related Compounds for Aldose
Reductase Inhibitory Activity. Chem. Pharm. Bull. 2002, 50, 788-
795.
(34) Shur, M.; Israelstam, S. S. The Reaction of Amino Heterocycles with
Reactive Esters. I. 2-Aminopyridines. J. Org. Chem. 1968, 33, 3015-
3019.
(35) Oikawa, Y.; Sugano, K.; Yonemitsu, O. Meldrum’s Acid in Organic
Synthesis. 2. A General and Versatile Synthesis of â-Keto Esters. J.
Org. Chem. 1978, 43, 2087-2088.
(36) Felix, A. M. Cleavage of Protecting Groups with Boron Tribromide.
J. Org. Chem. 1974, 39, 1427-1429.
(37) Burton, G. W.; Ingold, K. U. Vitamin E as an In Vitro and In Vivo
Antioxidant. Ann. N.Y. Acad. Sci. 1989, 570, 7-22.
(38) Kontoghiorghes, G. J. New Chelation Therapies and Emerging
Chelating Drugs for the Treatment of Iron Overload. Expert Opin.
Emerging Drugs 2006, 11, 1-5.
(39) Steuber, H.; Zentgraf, M.; Gerlach, C.; Sotriffer, C. A.; Heine, A.;
Klebe, G. Expect the Unexpected or Caveat for Drug Designers:
Multiple Structure Determinations Using Aldose Reductase Crystals
Treated under Varying Soaking and Co-crystallisation Conditions.
J. Mol. Biol. 2006, 363, 174-187.
(40) Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart, W.
E.; Belew, R. K.; Olson, A. J. Automated Docking Using a
Lamarckian Genetic Algorithm and an Empirical Binding Free Energy
Function. J. Comput. Chem. 1998, 19, 1639-1662.
(41) Wilson, D. K.; Tarle, I.; Petrash, J. M.; Quiocho, F. A. Refined 1.8
A Structure of Human Aldose Reductase Complexed with the Potent
Inhibitor Zopolrestat. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 9847-
9851.
(42) Bohren, K. M.; Grimshaw, C. E.; Lai, C. J.; Harrison, D. H.; Ringe,
D.; Petsko, G. A.; Gabbay, K. H. Tyrosine-48 is the Proton Donor
and Histidine-110 Directs Substrate Stereochemical Selectivity in the
Reduction Reaction of Human Aldose Reductase: Enzyme Kinetics
and Crystal Structure of the Y48H Mutant Enzyme. Biochemistry
1994, 33, 2021-2032.
(43) Grimshaw, C. E.; Bohren, K. M.; Lai, C. J.; Gabbay, K. H. Human
Aldose Reductase: pK of Tyrosine 48 Reveals the Preferred
Ionization State for Catalysis and Inhibition. Biochemistry 1995, 34,
14374.
(44) Etter, M. C.; Lipkowska, Z.; Baer, S.; Barbara, P. F. The Crystal
Structures and Hydrogen-Bond Properties of Three 3-Hydroxy-
flavone Derivatives. J. Mol. Struct. 1986, 144, 155-167.
(45) Shoja, M. 5-Hydroxyflavone. Acta Crystallogr., Sect. C 1990, 46,
517-519.
(46) Meyer, M. Ab Initio Study of Flavonoids. Int. J. Quantum Chem.
2000, 76, 724-732.
(47) El-Kabbani, O.; Wilson, D. K.; Petrash, J. M.; Quiocho, F. A.
Structural Features of the Aldose Reductase and Aldehyde Reductase
Inhibitor-Binding Sites. Mol. Vis. 1998, 4, 19-25.
(64) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.;
Klein, M. L. Comparison of Simple Potential Functions for Simulat-
ing Liquid Water. J. Chem. Phys. 1983, 79, 926-935.
(65) Hayward, S.; Kitao, A.; Go, N. Harmonicity and Anharmonicity in
Protein Dynamics: A Normal-Mode Analysis and Principal Com-
ponent Analysis. Proteins 1995, 23, 177-186.
JM070398A