Discovery of 3-[(4,5,7-trifluorobenzothiazol-2-yl)methyl]indole-N-acetic acid (lidorestat) and congeners as highly potent and selective inhibitors of aldose reductase for treatment of chronic diabetic complications

Article abstract of DOI:10.1021/jm0492094

Recent efforts to identify treatments for chronic diabetic complications have resulted in the discovery of a novel series of highly potent and selective 3-[(benzothiazol-2-yl)methyl]indole-N-alkanoic acid aldose reductase inhibitors. The lead candidate, 3-[(4,5,7-trifluorobenzothiazol-2-yl)methyl]indole-N-acetic acid (lidorestat, 9) inhibits aldose reductase with an IC₅₀ of 5 nM, while being 5400 times less active against aldehyde reductase, a related enzyme involved in the detoxification of reactive aldehydes. It lowers nerve and lens sorbitol levels with ED₅₀'s of 1.9 and 4.5 mg/kg/d po, respectively, in the 5-day STZ-induced diabetic rat model. In a 3-month diabetic intervention model (1 month of diabetes followed by 2 months of drug treatment at 5 mg/kg/d po), it normalizes polyols and reduces the motor nerve conduction velocity deficit by 59% relative to diabetic controls. It has a favorable pharmacokinetic profile (F, 82%; t_1/2, 5.6 h; Vd, 0.694 L/kg) with good drug penetration in target tissues (C_max in sciatic nerve and eye are 2.36 and 1.45 μg equiv/g, respectively, when dosed with [¹⁴C] lidorestat at 10 mg/kg po).

Full text of DOI:10.1021/jm0492094

J. Med. Chem. 2005, 48, 3141-3152
3141
Discovery of 3-[(4,5,7-Trifluorobenzothiazol-2-yl)methyl]indole-N-acetic Acid
(Lidorestat) and Congeners as Highly Potent and Selective Inhibitors of Aldose
Reductase for Treatment of Chronic Diabetic Complications
Michael C. Van Zandt,*,† Michael L. Jones,† David E. Gunn,† Leo S. Geraci,† J. Howard Jones,†
Diane R. Sawicki,† Janet Sredy,† Jorge L. Jacot,† A. Thomas DiCioccio,† Tatiana Petrova,‡ Andre Mitschler,‡ and
Alberto D. Podjarny‡
The Institute for Diabetes Discovery, LLC, 23 Business Park Drive, Branford, Connecticut 06405, and Departement de Biologie
et Genomique Structurales, UMR 7104 du CNRS, IGBMC, 1 rue Laurent Fries, B.P. 163, 67404 Illkirch Cedex, France
Received September 30, 2004
Recent efforts to identify treatments for chronic diabetic complications have resulted in the
discovery of a novel series of highly potent and selective 3-[(benzothiazol-2-yl)methyl]indole-
N-alkanoic acid aldose reductase inhibitors. The lead candidate, 3-[(4,5,7-trifluorobenzothiazol-
2-yl)methyl]indole-N-acetic acid (lidorestat, 9) inhibits aldose reductase with an IC50 of 5 nM,
while being 5400 times less active against aldehyde reductase, a related enzyme involved in
the detoxification of reactive aldehydes. It lowers nerve and lens sorbitol levels with ED50’s of
1.9 and 4.5 mg/kg/d po, respectively, in the 5-day STZ-induced diabetic rat model. In a 3-month
diabetic intervention model (1 month of diabetes followed by 2 months of drug treatment at 5
mg/kg/d po), it normalizes polyols and reduces the motor nerve conduction velocity deficit by
59% relative to diabetic controls. It has a favorable pharmacokinetic profile (F, 82%; t1/2, 5.6 h;
Vd, 0.694 L/kg) with good drug penetration in target tissues (Cmax in sciatic nerve and eye are
2.36 and 1.45 µg equiv/g, respectively, when dosed with [14C]lidorestat at 10 mg/kg po).
Scheme 1
Introduction
Diabetes mellitus and its disabling complications,
which include blindness, renal failure, neuropathy, limb
amputation, myocardial infarction, and stroke, affect
some 17 million people in the United States with an
estimated cost of over 130 billion dollars annually.1
Through various clinical studies, including the Diabetes
Control and Complications Trial (DCCT) and the United
Kingdom Prospective Diabetes Study (UKPDS), the
development and progression of these complications in
type 1 and type 2 patients have been clearly linked to
elevated blood glucose levels.2 Although glucose is
preferentially metabolized through the glycolytic path-
way, during conditions of hyperglycemia, as observed
in diabetes mellitus, elevated blood glucose levels
saturate the normal pathways of glucose metabolism
and a dramatic increase in flux through the polyol
pathway results (Scheme 1). Glucose entering the polyol
pathway is reduced to sorbitol by aldose reductase
(ALR2) and NADPH. Sorbitol is subsequently oxidized
to fructose by sorbitol dehydrogenase and NAD.+ This
increased flux through the polyol pathway results in a
reduced ratio of NADPH to NADP+ and an increased
ratio of NADH to NAD.+ These changes, which alter the
reduction potential of the cell, are collectively termed
oxidative stress. The impact of oxidative stress has been
clearly demonstrated.3 It is linked to depleted intracel-
lular levels of reduced glutathione, increased nonenzy-
matic glycation, and activation of protein kinase C.
In addition to sorbitol and fructose,4 myo-inositol is
also an important sugar that is effected by flux through
the polyol pathway.5 As glucose and sorbitol levels are
increased, myoinositol levels are decreased. This de-
crease results in reduced production of diacylglycerol,
diminished Na+-K+-ATPase activity, and eventually a
reduced nerve conduction velocity.6 In addition to oxida-
tive stress, accumulation of sorbitol in certain cells
results in osmotic stress that can lead to cell swelling
and eventually cell death.7 Inhibitors of aldose reductase
can prevent these metabolic and biochemical changes,
which have been linked to the pathogenesis of diabetic
complications.8
Further supporting the role of ALR2 in the develop-
ment of diabetic complications, certain allelic variants
of the hALR2 gene associated with overexpression of
aldose reductase are well-correlated with increased risk
of developing diabetic complications.9 Similar conclu-
sions can be drawn from the ALR2 knockout mouse.
Even when severely diabetic, these animals are pro-
tected against decline in motor nerve conduction veloc-
ity.10 As a result, a high ALR2 level is now considered
an important risk factor for chronic diabetic com-
plications, including neuropathy, retinopathy, and
nephropathy.11
During the past 25 years at least 14 different potent
and highly efficacious aldose reductase inhibitors (ARIs)
have been identified and advanced through phase II
clinical development (Figure 1).12 Despite this tremen-
* To whom correspondence should be addressed. Fax: (203) 315-
4002.
ipd-discovery.com.
† The Institute for Diabetes Discovery.
‡ Departement de Biologie et Genomique Structurales.
Tel:
(203)
315-5951.
E-mail:
michael.vanzandt@
10.1021/jm0492094 CCC: $30.25 © 2005 American Chemical Society
Published on Web 03/30/2005

3142 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 9
Van Zandt et al.
Figure 1. Inhibitors of aldose reductase advancing to phase II clinical development.
Scheme 3a
Figure 2. Aldose reductase inhibitors with 4,5,7-trifluoroben-
zothiazole side chain.
Scheme 2a
a Reagents: (a) morpholine, K2CO3, DMSO, 80 °C (97%); (b)
phenol, K2CO3, DMSO, 80 °C (63%); (c) dimethylformamide
dimethyl acetal, pyrrolidine, 100 °C; (d) H2, Pd-C, ethanol, 55
psi (31%, 2 steps); (e) analogous to method set forth in Scheme 2.
demonstrated.14 It converts highly reactive 2-oxoalde-
hydes such as 3-deoxyglucosone and methyl glyoxal to
their corresponding nonreactive alcohols.15 These alde-
hydes, which are involved in the formation of various
protein cross-links and other advanced glycation end
products (AGEs), are formed as degradation products
from glucose and fructose.16 During conditions of pro-
longed hyperglycemia, high concentrations of these
aldehydes are formed, resulting in the development of
chronic diabetic complications and accelerated aging.17
a Reagents: (a) Ac2O, pyridine, 120 °C (82%); (b) P4S10, benzene
(88%); (c) NaH, toluene (96%); (d) 30% aqueous NaOH, ethylene
glycol, 125 °C (73%); (e) ethyl bromoacetate, NaH, acetonitrile
(49%); (f) trifluoroethanol, cat. BHT, 80 °C (59%); (g) aqueous
NaOH, ethanol (98%).
dous effort, none are currently marketed for worldwide
use. Although epalrestat is currently available in Japan,
and fidarestat and AS-3201 are in late-stage clinical
development, many of these candidates failed to gain
acceptance due to an inadequate therapeutic index.13
In some cases this toxicity may have resulted from a
lack of selectivity relative to aldehyde reductase (ALR1).
The physiological importance of ALR1 has been clearly
Our work in this area has focused primarily on
identifying new carboxylic acid-based templates with
significantly improved oral efficacy and selectivity rela-
tive to aldehyde reductase. In this paper we report our
work on a novel series of indole-N-acetic acids, including
our clinical candidate, 3-[(4,5,7-trifluorobenzothiazol-2-
yl)methyl]indole-N-acetic acid (lidorestat), compound 9.

Potent and Selective Inhibitors of Aldose Reductase
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 9 3143
Table 1. Physical and in Vitro Properties of Indoleacetic Acid ARIs
hALR2a
(nM)
HALRb
(nM)
no.
substituents
mp, °C
empirical formula
analysis
17
18
19
20
21
22
23
24
25
26
27
9
28
29
30
31
32
33
34
35
36
16
37
15
38
39
40
41
42
43
44
45
46
47
5′-CF3
233-234
248-249
238-240
265-267
158-160
170-172
208
C
19H13F3N2O2S
20H15F3N2O2S
19H12ClF3N2O2S
19H12F3N2O2SBr
18H13ClN2O2S
18H13FN2O2S
18H13FN2O2S
18H13FN2O2S
18H13FN2O2S
18H12F2N2O2S
18H12F2N2O2S
18H11F3N2O2S
C,H,N,S
C,H,N,S
NDc
99
100
130
52
660
34
9
5 600
10 200
4 900
5-CH3, 5′-CF3
C
C
C
C
C
C
C
C
C
C
C
5-Cl, 5′-CF3
6-Br, 5′-CF3
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
ND
3 400
6′-Cl
2 100
4′-F
11 000
6 400
5′-F
6′-F
203
1 000
55
560
690
5
4 500
7′-F
181-183
193-195
216-218
177-178
178-180
238-239
203-206
188-189
165-168
143-144
131-133
165-167
165-168
128-130
156-159
119-121
200-203
194-195
211-213
118-120
156-159
178-180
194-196
228-230
228-230
216-218
18 000
13 000
5 100
5′-F, 6′-F
4′-F, 6′-F
4′-F, 5′-F, 7′-F
27 000
13 000
16 000
14 000
11 000
21 000
4 700
2-CH3, 4′-F, 5′-F, 7′-F
2-Ph, 4′-F, 5′-F, 7′-F
4-Cl, 4′-F, 5′-F, 7′-F
5-Cl, 4′-F, 5′-F, 7′-F
5-F, 4′-F, 5′-F, 7′-F
5-Br, 4′-F, 5′-F, 7′-F
5-CH3, 4′-F, 5′-F, 7′-F
5-OCH3, 4′-F, 5′-F, 7′-F
5-OBn, 4′-F, 5′-F, 7′-F
5-OPh, 4′-F, 5′-F, 7′-F
5-Ph, 4′-F, 5′-F, 7′-F
5-morpholino, 4′-F, 5′-F, 7′-F
6-F, 4′-F, 5′-F, 7′-F
6-Cl, 4′-F, 5′-F, 7′-F
6-CH3, 4′-F, 5′-F, 7′-F
6-OCH3, 4′-F, 5′-F, 7′-F
6-Ph, 4′-F, 5′-F, 7′-F
6-morpholino, 4′-F, 5′-F, 7′-F
7-F, 4′-F, 5′-F, 7′-F
7-Cl, 4′-F, 5′-F, 7′-F
7-Br, 4′-F, 5′-F, 7′-F
7-CH3, 4′-F, 5′-F, 7′-F
C19 13
H
F3N2O2S
8
C
C
C
C
C
24H15F3N2O2S
100
11
10
11
13
8
18H10ClF3N2O2S
18H10ClF3N2O2S
18H10F4N2O2S
18H10BrF3N2O2S
C19 13
C19 13
H
F3N2O2S
F3N2O3S
34 000
21 000
4 800
H
8
C
C
C
C
C
C
25H17ClF3N2O3S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
ND
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
12
30
53
8
24H15F3N2O3S
11 000
10 000
41 000
7 500
24H15F3N2O2S
22H18F3N3O3S H2O
18H10F4N3O2S
7
8
18H10F3N2O2SCl
2 700
C19 13
C19 13
H
H
F3N2O2S
F3N2O3S·H2O
10
5
13 000
12 000
6 100
C
C
C
C
C
24H15F3N2O2S
25
15
7
22H18F3N3O3S
44 000
7 000
28H10F4N2O2S·H2O
18H10F3N2O2SCl
18H10F3N2O2SBr
9
19 000
18 000
21 000
14
6
C19H13F3N2O2S
a Recombinant human aldose reductase. b Recombinant human aldehyde reductase. c Not determined.
Inhibitor Design and Synthesis
fication with 2 N HCl gives the 2-aminothiophenol as
the hydrochloride salt (5). Although prepared in high
yield, this intermediate is highly unstable and com-
pletely dimerizes within 24 h. This oxidation can be
easily prevented, however, by addition of a free radical
inhibitor. With the addition of a catalytic amount of
BHT (a few crystals) this key intermediate is stable on
the benchtop for several months.
Indole-3-acetonitrile-N-acetic acid ethyl ester 7 was
prepared from indole-3-acetonitrile (6) by alkylation
with ethyl bromoacetate and sodium hydride. Substi-
tuted indole examples where the 3-acetonitrile substitu-
ent was not commercially available were conveniently
prepared from the corresponding indole via the gramine
intermediate using classical procedures.19 Condensation
of 2-aminothiophenol 5 with indole 3-acetonitrile 7 in
2,2,2-trifluoroethanol provided benzothiazole 8 in 59%
yield. Initial results using ethanol as solvent were
considerably less efficient due to partial esterification
of the cyano group. Finally, hydrolysis with aqueous
NaOH gives the target compound (9) in 29% overall
yield (six linear steps).
As we began work on our indole-based ARIs, we were
particularly intrigued by a series of inhibitors containing
a 4′,5′,7-trifluorobenzothiazole side chain (Figure 2).
Although these compounds do not appear to have been
developed clinically, published data suggests that the
side chain provides significant improvements in in vitro
and in vivo activity relative to the 5-trifluoromethyl-
benzothiazole or 2-fluoro-4-bromobenzyl used in previ-
ously developed ARIs.18 In our efforts to identify and
develop a new ARI with the potency, selectivity, and
safety required to proceed through clinical development,
we found that the same side chain was optimal in our
indole-N-acetic acid based series. The structure-activity
relationship for the series is described in the Results
and Discussion.
The target compounds were synthesized using meth-
ods illustrated in Schemes 2 and 3. The synthesis
of 3-[(4,5,7-trifluorobenzothiazol-2-yl)methyl]indole-N-
acetic acid (9) is described in Scheme 2. Here, tetrafluo-
roaniline 1 is acylated with acetic anhydride in pyridine
at 120 °C followed by treatment with P4S10 in benzene
to give thioamide 3 in good yield. Subsequent cyclization
of the thioamide using sodium hydride provides 2-
methylbenzothiazole 4. Hydrolysis of the heterocyclic
ring with aqueous sodium hydroxide followed by acidi-
Compounds 15 and 16, where the 5′-position on the
indole ring is substituted with a morpholino or phenoxy
group, were prepared using the chemistry illustrated
in Scheme 3. 4-Nitro-3-methylfluorobenzene 10 is treated

3144 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 9
Table 2. Physical and in Vitro Properties of Indoleacetic Acid ARIs
Van Zandt et al.
no.
A
B
mp, °C
empirical formula
analysis
hALR2a (nM)
hALR1b (nM)
9
48
49
50
51
-CH2-
-CH2-
177-178
>250
198-199
200-201
176-177
C18H11F3N2O2S
C18H9F3N2O3S
C19H13F3N2O3S
C19H13F3N2O3S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
C,H,N,S
5
1 600
8
27 000
-(CO)-
-CH2CH2-
-CH2-
-CH2-
NDc
-CH2-
-CH2CH2-
-CH(CH3)-
11 000
88 000
6 500
17
-CH2-
C19H13
F3N2O2S·H2O
163
a Recombinant human aldose reductase. b Recombinant human aldehyde reductase. c Not determined.
with morpholine or phenol with K2CO3 in DMSO to
provide substituted products 11 and 12, respectively.
Subsequent treatment with dimethylformaldehyde di-
methyl acetal followed by hydrogenation provides the
corresponding indole intermediates 13 and 14. Once
formed, the substituted indole intermediates were con-
verted to the target compounds using chemistry outlined
in Scheme 2.
Results and Discussion
The major objective of this program has been to
identify highly potent, selective, and efficacious aldose
reductase inhibitors for treatment of chronic diabetic
complications. All compounds in the program were
initially tested for potency against human aldose re-
ductase (hALR2) and selectivity relative to human
aldehyde reductase (hALR1). The results from these
experiments are listed in Tables 1 and 2. Selected
compounds with significant in vitro activity against
hALR2 were also tested in the streptozotocin (STZ)
Figure 3. Compound 9 bound to ALR2. Enzyme-inhibitor
complex of 9 bound in the active site of human aldose
reductase. Contacts shorter than 3.5 A are indicated. Note
short contacts of the inhibitor carboxylic acid head with
residues Tyr48, His110, and Trp111 and the perfect π-stacking
interaction of the benzothiazole side chain with the indole
moiety of Trp111.
induced diabetic rat model.
X-ray crystallography was used throughout the pro-
gram. The structures provided a clear understanding
of the important interactions that make up the enzyme-
inhibitor complexes. This information was used to focus
efforts on the specific fragments of the inhibitor tem-
plate that provided the greatest opportunities for opti-
mization of potency and selectivity relative to aldehyde
reductase.
The complexes were obtained using hALR2 expressed
in Escherichia coli and crystallized with the oxidized
form of the coenzyme â-NADP+ at pH 5 and 277 K.
Initial results with 9 provided a precise picture of the
inhibitor-enzyme complex (1.04 Å resolution). As il-
lustrated in Figure 3, the inhibitor is oriented in the
active site of hALR2 in a manner such that the hydro-
philic carboxylate head forms tight hydrogen bonds with
the OH of Tyr48 (2.72 Å), the NE2 of His110 (2.69 Å),
the O7N of NADP+ (3.41 Å), and the NE1 of Trp111
(2.95 Å). These interactions anchor the inhibitor deep
within the enzyme active site.
inhibitors with binding interactions in this “specificity
pocket” are generally highly selective for aldose reduc-
tase. The specific interactions within the specificity
pocket for 9 include a nearly perfect π-stacking interac-
tion of the benzothiazole side chain with the indole
moiety of Trp111 and H-bonding interactions between
the amide N-H of Leu300 and the benzothiazole N (3.42
Å) and the 4′-fluorine (3.38 Å).
Within this series, the indole, benzothiazole, and
carboxylic acid moieties were kept constant while the
various substituents and linkers were optimized. Since
selectivity relative to aldehyde reductase was a critical
program objective, initial efforts focused on optimization
of the benzothiazole ring system.22 On the basis of work
published for zopolrestat, SG-210, and GP-1447, various
examples with fluorine or trifluoromethyl substituents
were evaluated. The first examples with interesting in
vitro activity were those with the 5′-trifluoromethyl
substituent (17-20). This side chain provides modest
in vitro activity with IC50’s of approximately 100 nM.
In addition to the CF3 substituent, fluorine substituents
As this class of inhibitors binds to the hALR2 active
site, a conformational change occurs, opening a pocket
localized between Trp111 and Leu300.20 The specific
opening of the pocket varies to accommodate each
inhibitor, producing an “induced fit.” Since the residues
lining this pocket are not conserved in aldehyde reduc-
tase, the interactions are specific for ALR2.21 As a result,

Potent and Selective Inhibitors of Aldose Reductase
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 9 3145
Table 3. Activity of Lidorestat (9) Relative to Other Clinically
in various positions around the benzothiazole ring
provided an interesting structure-activity relationship.
As seen in examples 22-25, introduction of a fluorine
substituent in the 4′-, 5′-, and 7′-positions (22, 23, and
25, respectively) provides excellent in vitro activity, with
the 5′-position being most important, giving an IC50 of
9 nM. Unlike the 4′-, 5′-, and 7′-positions, compound 24,
with the 6′-fluoro substituent, is relatively inactive, with
an IC50 of 1000 nM. This same trend is seen in the
difluoro-substituted analogues. When the 6′-fluoro sub-
stituent is added to the 5′-fluoro (23) and 4′-fluoro (22)
analogues, giving the corresponding difluoro compounds
26 and 27, the in vitro activity is decreased by about
10-fold (IC50’s > 500 nM). Conversely, when the 4′-, 5′-,
and 7′-fluorine substituents are combined, the binding
affinity is enhanced, giving example 9 with an IC50 of 5
nM. This structure-activity relationship is consistent
with the well-defined interactions between the ben-
zothiazole moiety and the specificity pocket. Introducing
the 6′-fluorine into the crystal structure of 9 requires a
short contact (about 2 Å) with the OG1 of Thr113. This
unfavorable steric interaction results in the loss of
binding affinity.
In addition to the potency of these compounds, the
selectivity relative to aldehyde reductase is also addi-
tive. The 4′-, 5′-, and 7′-fluorine-substituted examples
(22, 23, and 25) have selectivity ratios (hALR1/hALR2)
of 710, 320, and 330, respectively, but when the sub-
stituents are combined together, the selectivity ratio for
compound 9 becomes 5400.
In contrast to the strong substitution effects seen on
the benzothiazole, substituents on the indole framework
have little effect. This is due primarily to the orientation
of the inhibitors in the active site. While the carboxylate
and benzothiazole side chains reach deep within the
active site, the space around the indole core is relatively
open.
While the 4′,5′,7′-trifluorobenzothiazole is kept con-
stant, introduction of a methyl at the 2-position gives
28, which has an IC50 of 8 nM. Even with the large
phenyl substituent in the 2-position (29), the IC50 only
increases to 100 nM. Examples with halogens or methyl
in the 4- or 5-position have IC50’s of about 10 nM (30,
4-cloro; 31, 5-cloro; 32, 5-floro; 33, 5-bromo; 34; 5-meth-
yl). Other alkoxy substituents in the 5-position, includ-
ing methoxy (35), benzyloxy (36), and phenoxy (16),
have IC50’s of 12, 30, and 53 nM, respectively. The
5-phenyl example (37) is somewhat less active, with an
IC50 of 53 nM, while the 5-morpholino is quite interest-
ing, with an IC50 of 8 nM and selectively ratio of >5000
(15). Similar results are observed for the corresponding
examples with substituents in the 6- and 7-positions
(38-47).
In addition to the benzothiazole and indole portions
of the inhibitor template, some modifications of the
linkers were also investigated. Oxidation of the indole-
benzothiazole linker A to a carbonyl (48) results in a
significant loss in activity (IC50 ) 1600 nM), while the
homologue (A ) -CH2CH2-, 49) remains very active,
with an IC50 of 8 nM. The loss of activity for example
48 is expected on the basis of the increased bond angle
required for the sp2 hybridized carbon atom. In this
configuration the substituted benzothiazole and car-
boxylate are unable to interact with their respective
Developed ARIsa
% inhibn of sorbitol accum
IC50
(10 mg/kg/d)c
group
hALR2a hALR1b
nerve
lens
tolrestat
13
9
1 940
38 600
9 900
27
80
7
9
zopolrestat
zenarestat
lidorestat (9)
8
64
2
5
27 000
100
77
a Recombinant human aldose reductase. b Recombinant human
aldehyde reductase. c Percent improvement of drug treated group
relative to difference between nondiabetic and diabetic control
groups.
Table 4. Nerve Function Assesment with Lidorestat (9) in
1-month Diabetic Prevention Modela
dose
(mg/kg/d)
group
MNCV
% imp.b
nondiabetic control
diabetic control
lidorestat (9)
-
-
47.73 ( 1.08c
41.15 ( 0.43
44.77 ( 0.69*
46.25 ( 0.94*
-
-
55
78
5
lidorestat (9)
10
a One month of diabetes following STZ injection. b Percent
improvement of drug-treated group relative to the difference
between nondiabetic and diabetic control groups. c All values are
mean ( SEM. *p < 0.05 as compared to diabetic control group.
pockets at the same time. In contrast, the homologue is
able to retain these binding interactions by simply
moving the indole moiety slightly. It is interesting to
note that the same homologue in the zopolrestat series
is about 100-fold less active.23 This likely results from
a reduced flexibility around the phthalazine ring.
Homologation of the methylene connecting the indole
moiety to the carboxylic acid (B ) -CH2CH2-, 50) is
also quite active, with an IC50 of 17 nM. In addition to
its good potency, this example is also quite selective,
with a selectivity ratio of over 5000. Introduction of a
methyl substituent to linker B (51) results a significant
loss of activity (IC50 ) 163 nM). The reduced activity
for this example arises from a steric interaction with
Trp20.
As the optimization program was developing, com-
pounds with good in vitro activity were selected for
further evaluation in vivo. In general, the streptozotocin
(STZ)-induced diabetic rat model was used to evaluate
the in vivo efficacy of compounds from this series.
Sorbitol concentrations in nerve and lens tissues were
used as the primary biochemical endpoints in the initial
5-day screening assay. In this model, where sorbitol
levels in the sciatic nerve and lens are increased 10 and
20-fold in the diabetic controls, lidorestat is highly
efficacious, normalizing this polyol with an ED50 of 1.9
and 4.5 mg/kg/d, respectively. When compared to other
clinically developed ARIs, lidorestat is one of the most
potent, selective, and efficacious (Table 3). With these
excellent results, lidorestat was subsequently tested in
more long-term studies to better define its pharmaco-
dynamic effects. Dose-response prevention and low-
dose intervention study designs were employed in which
the animals either received drug 4 days after STZ-
induction of diabetes (prevention) or were diabetic for
28 days prior to receiving daily drug treatment for the
two following months (intervention).
In the prevention study, after one-month of STZ-
induced diabetes, the untreated diabetic animals dem-

3146 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 9
Van Zandt et al.
Table 5. Biochemical and Nerve Function Assessment with Lidorestat (9) in 3-Month Diabetic Intervention Modela
sorbitol
(nmol/mg
of dry wt)
fructose
(nmol/mg
of dry wt)
myo-inositol
(nmol/mg
of dry wt)
group
MNCV
nondiabetic control
diabetic control
9 at 5.0 mg/kg/d
0.24 ( 0.02b
4.84 ( 0.69*
BQLc **
2.16 ( 0.11
20.10 ( 1.53*
1.62 ( 0.05**
9.90 ( 0.27
6.89 ( 0.18*
9.26 ( 0.26**
51.00 ( 0.79
44.58 ( 0.49
48.39 ( 0.42
% improvementd
100
100
79
59
a One month of diabetes followed by 2 months of drug treatment. b All values are mean ( SEM. *p < 0.05 as compared to nondiabetic
control group. **p < 0.05 as compared to untreated diabetic control group. c Below quantifiable level. d Percent improvement of drug-
treated group relative to the difference between nondiabetic and diabetic control groups.
Figure 4. Slit-lamp assessment of cataract formation with lidorestat. Slit-lamp evaluation of cataract formation in rat lens after
3 months of diabetes. Image A, nondiabetic control; image B, diabetic control; and image C, diabetic with lidorestat at 25 mg/kg/d
po. Results indicate that the drug-treated group is indistinguishable from the nondiabetic control group.
Table 6. Triglyceride Lowering with Lidorestat
onstrated a significant loss in motor nerve conduction
velocity (MNCV, 6.58 m/s relative to the control group).
Lidorestat, administered at 5 and 10 mg/kg/d, substan-
tially reduced the MNCV decline, resulting in an
improvement of 3.62 and 5.10 m/s, respectively (Table
4). This improvement corresponds to a 55% and 78%
recovery of the MNCV-deficit relative to the untreated
diabetic rats.
In the intervention study, at three-months, untreated
diabetic rats had a decline in MNCV of 6.42 m/s
compared to the nondiabetic control group (Table 5).
Diabetic animals treated for the last 2 months with
lidorestat lost only 2.61 m/s in MNCV. This represents
a 59% improvement relative to the diabetic control
group. In addition to MNCV, in the intervention study,
nerves were removed and analyzed for polyol and sugar
content.
plasma
triglycerides
(mg/dL)
glucose
(mg/dL)
group
control
nondiabetic
diabetic
9-treated
4.8 mg/kg/d
1.9 mg/kg/d
1.4 mg/kg/d
0.95 mg/kg/d
0.48 mg/kg/d
% red.a
72 ( 2b
439 ( 12**
81 ( 18
349 ( 78**
-
-
417 ( 12**
440 ( 16**
427 ( 26**
430 ( 13**
432 ( 16**
189 ( 34*
218 ( 27**
242 ( 58*
261 ( 81*
301 ( 102
60
49
40
33
18
a Percent reduction of plasma triglycerides in drug-treated group
relative to difference between nondiabetic and diabetic controls.
b All values are means ( SEM . **p < 0.01 as compared to
nondiabetic control group. *p < 0.05 as compared to nondiabetic
control group.
As expected, the concentrations (nmol/mg/dry wt)
of glucose, sorbitol, and fructose were significantly
increased and that of myo-inositol significantly de-
creased in the sciatic nerve of untreated diabetic rats
relative to the nondiabetic control group. Treatment
with lidorestat (5 mg/kg/d) in the intervention study
completely eliminated the accumulation of sorbitol and
fructose and corrected myo-inositol concentrations (79%)
without altering nerve glucose levels.
In the eye, activation of the polyol pathway and
corresponding increase of sorbitol concentration in the
lens epithelium is an early initiating event in catarac-
togenesis.24 This increase in lens sorbitol level has also
been shown to correlate with the rate of diabetic
cataract formation.25 Once formed inside the cell, sor-
bitol only slowly transverses the cell membrane, result-
ing in accumulation inside the cell. This results in
intracellular hypertonicity, followed by an influx of
water, which leads to cellular edema. Continued intra-
cellular hydration leads to vacole formation and loss of
membrane integrity, leading to irreversible lens opaci-
fication. Treatment with lidorestat exhibited a dose-
dependent prevention of cataract formation in STZ-
induced diabetic rats (5-25 mg/kg/d admixed in diet).
After 3 months, untreated diabetic control animals had
full nuclear cataracts formed while groups treated with
5 and 10 mg/kg/d exhibited only diffuse cloudiness and
loss of transparency. The 25 mg/kg/d group suffered no
loss of clarity and was indistinguishable from the
nondiabetic control group. Figure 4 provides a slit-lamp
evaluation of the rat lens at 3 months for the control
STZ-induced diabetic rat verses the same animals
treated with lidorestat at 25 mg/kg/d.
In addition to polyols, MNCV, and cararact formation,
lidorestat was also evaluated for its effect on plasma
triglyceride levels. In the 15-day STZ-induced diabetic
rat model, where blood glucose and plasma triglyceride
levels are increased 5.7- and 5.4-fold relative to the
nondiabetic controls (Table 6), treatment with lidorestat
at doses of 0.48-4.8 mg/kg/d resulted in a significant
dose-related reduction in triglyceride concentration (18-
60%).
These data suggests that in addition to its direct
beneficial effect on the peripheral nerve and lens,
lidorestat has the added advantage of lowering the

Potent and Selective Inhibitors of Aldose Reductase
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 9 3147
Table 7. Pharmacokinetics of [14C]Lidorestat in Rata
Cmax
(µg equiv/
mL)
AUC0-24h
(µg equiv/
h/mL)
ClT
(mL/
min/kg)
dose
(mg/kg/d)
Tmax
(h)
t1/2
(h)
Vd
(L/kg)
group
iv, fasted
10
6
6
0.08
3.50
5.00
24.0
4.18
6.09
5.04
5.61
5.18
97.1
43.5
72.8
0.694
-
-
1.70
2.24
1.37
po, fasted
po, nonfasted
a Study measures unchanged drug in plasma.
theory unless otherwise noted. Products from all reactions
were purified by either flash column chromatography or
medium-pressure liquid chromatography using silica gel 60
(230-400 mesh Kieselgel 60, EM Reagents) unless otherwise
indicated. Thin-layer chromatography using glass-backed silica
plates containing a fluorescent indicator (0.25 mm, Whatman,
Merck) was used to monitor reactions. The chromatograms
were visualized using ultraviolet illumination, exposure to
iodine vapors, or dipping in an aqueous potassium perman-
ganate solution. All starting materials were used without
further purification. All reactions were carried out under an
atmosphere of dried nitrogen.
General Procedures for the Synthesis of 3-(Benzothi-
azol-2-yl)methylindole-N-alkanoic Acid ARIs. 3-(Ben-
zothiazol-2-yl)methylindole-N-alkanoic acids 9 and 15-51
(Tables 1 and 2) were prepared from the desired 2-fluoroaniline
or corresponding commercially available substituted 2-ami-
nothiophenol and the indole moiety as outlined in Scheme 2
via the general experimental methods described for 9 below.
The desired indole moiety used in the synthesis was prepared
by the method outlined in Schemes 3 or purchased com-
mercially.
3-(4,5,7-Trifluorobenzothiazol-2-yl)methylindole-N-
acetic Acid (9). Step 1: Synthesis of 2,3,5,6-Tetrafluoro-
acetanilide (2). A solution of 2,3,5,6-tetrofluoroaniline (200
g, 1.21 mol) in anhydrous pyridine (103 mL, 1.27 mol) was
treated with acetic anhydride (120 mL, 1.27 mol) and heated
to 120 °C for 2 h. After cooling to room temperature, the
solution was poured into ice-cold water (500 mL). The resulting
precipitate was filtered, dissolved in ethyl acetate, dried over
MgSO4, filtered, and concentrated. The solid material was
washed with heptane (200 mL) and dried to give 2,3,5,6-
tetrafluoroacetanilide as a white crystalline solid (206 g,
82%): mp 136-137 °C; Rf 0.48 (50% ethyl acetate in heptane);
1H NMR (DMSO-d6, 300 MHz) δ 10.10 (s, 1 H), 7.87-7.74 (m,
1 H), 2.09 (s, 3 H). Anal. (C8H5F4NO) C, H, N.
Step 2: 2,3,5,6-Tetrafluorothioacetanilide (3). A flame-
dried, four-necked, 5000 mL round-bottomed flask was charged
with phosphorus pentasulfide (198 g, 0.45 mol) and diluted
with anhydrous benzene (3000 mL, 0.34 M). 2,3,5,6-Tetrafluo-
roacetanilide (185 g, 0.89 mol) was added in one portion and
the bright yellow suspension was heated to a gentle reflux for
3 h. The solution was cooled to 0 °C and filtered. The insoluble
material was washed with ether (2 × 250 mL) and the
combined filtrate was extracted with 10% aqueous NaOH (750
mL, 500 mL). After cooling the aqueous layer to 0 °C, it was
carefully acidified with concentrated HCl (pH 2-3). The
precipitated product was collected by filtration and washed
with water (500 mL). The yellow-orange material was dis-
solved in ethyl acetate (1000 mL), dried over MgSO4 and
activated charcoal (3 g), filtered through a short pad of silica
(50 g), and concentrated. The resulting solid was triturated
with heptane (500 mL) and filtered to give 2,3,5,6-tetrafluo-
rothioacetanilide (174.9 g, 88%): mp 103-104 °C; Rf 0.67 (50%
ethyl acetate in heptane); 1H NMR (DMSO-d6, 300 MHz) δ
11.20 (s, 1 H), 8.00-7.88 (m, 1 H), 2.66 (s, 3 H). Anal. (C8H5F4-
NS) C, H, N, S.
elevated levels of triglycerides that are commonly seen
in people with diabetes.
A pharmacokinetic evaluation of [14C]lidorestat was
conducted in rat using both intravenous and oral dosing
(Table 7). Following intravenous dosing with 10 mg/kg/d
in fasted animals, [14C]lidorestast has a terminal elimi-
nation half-life (t1/2) of 5.04 h with a volume of distribu-
tion and total body clearance of 0.694 L/kg and 1.70 mL/
min/kg, respectively. The oral bioavailability was 82%
with a Cmax of 4.18 µg equiv/mL and tmax of 3.50 h. The
t1/2 following oral administration was determined to be
5.61 h. In a separate tissue distribution study with [14C]-
lidorestat also dosed orally at 10 mg/kg in male rats,
the Cmax in sciatic nerve and eye were 2.36 and 1.45 µg
equiv/g, respectively. Unlike many other carboxylic acid
based aldose reductase inhibitors,4 the favorable phar-
macokinetic profile of lidorestat provides excellent
penetration in the hard-to-reach tissues where compli-
cation occur (nerve and lens). The details of this tissue
distribution and pharmacokinetic study will be pub-
lished separately.
Lidorestat has proceeded through early development
into a phase II clinical trial. Preliminary results in man
demonstrate complete normalization of sural nerve
sorbitol levels at all doses tested (25-400 mg/d). Mea-
sured sorbitol levels after 60 days were 0.00-0.09 nM/
mg of dry weight compared to 2.5-8.0 nM/mg of dry
weight that would be expected from untreated diabetic
patients.26 Although these results are preliminary, it is
clear lidorestat is highly efficacious in man and will
likely require a dose significantly less than 25 mg/d.
In conclusion, a novel series of 3-[(benzothiazol-2-yl)-
methyl]indole-N-acetic acid derivatives has been identi-
fied and optimized as highly potent and selective
inhibitors of aldose reductase. In vivo evaluation of
3-[(4,5,7-trifluorobenzothiazol-2-yl)methyl]indole-N-ace-
tic acid (lidorestat), compound 9, in the STZ-induced
diabetic rat model resulted in normalization of nerve
sugars (sorbitol, fructose, and myo-inositol), improve-
ment of MNCV, reduction of triglycerides, and preven-
tion of diabetic cataracts. Continued development of
these compounds may result in an important new
treatment for chronic diabetic complications.
Experimental Section
Chemistry. General Methods. Melting points were de-
termined in open capillary tubes on a Thomas-Hoover ap-
paratus and are uncorrected. Proton nuclear magnetic reso-
nance (1H NMR) spectra were determined on a Varian Gemini
2000 (300 MHz) instrument. Chemical shifts are provided in
parts per million (ppm) downfield from tetramethylsilane
(internal standard) with coupling constants in hertz (Hz).
Multiplicity is indicated by the following abbreviations: singlet
(s), doublet (d), triplet (t), quartet (q), multiplet (m), broad (br).
Mass spectra were recorded on a Perkin-Emer API 100.
Elemental analyses (C, H, N, S) were performed by Atlantic
Microlab, Inc. (Norcross, Georgia) and are within (0.4% of
Step 3: 4,5,7-Trifluoro-2-methylbenzothiazole (4). A
flame-dried, 5000 mL round-bottomed flask equipped with
overhead stirrer was charged with sodium hydride (15.9 g, 0.66
mol) and diluted with anhydrous toluene (3000 mL, 0.2 M).
The suspension was cooled to 0 °C, and treated with 2,3,5,6-
tetrafluorothioacetanilide (134 g, 0.60 mol) in one portion. The
solution was warmed to room temperature over 1 h and then

3148 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 9
Van Zandt et al.
heated to a gentle reflux. After 30 min, dimethylformamide
(400 mL) was carefully added and the mixture was stirred for
an additional 2 h. The solution was cooled to 0 °C and added
to ice-water (2000 mL). The solution was extracted with ethyl
acetate (1500 mL) and washed with saturated aqueous NaCl
(1000 mL). The organic layer was concentrated to dryness,
diluted with heptane, and successively washed with water (300
mL) and saturated aqueous NaCl (1000 mL). The organic layer
was dried over MgSO4, filtered, and concentrated to give 4,5,7-
trifluoro-2-methylbenzothiazole (116.8 g, 96%) as a light brown
solid: mp 91-92 °C; Rf 0.56 (30% ethyl acetate in heptane);
1H NMR (DMSO-d6, 300 MHz) δ 7.76-7.67 (m, 1 H), 2.87 (s,
3 H). Anal. (C8H4F3NS) C, H, N, S.
Step 4: 2-Amino-3,4,6-trifluorothiophenol Hydrochlo-
ride (5). A solution of 4,5,7-trifluoro-2-methylbenzothiazole
(25.0 g, 123 mmol) in ethylene glycol (310 mL, 0.4 M) and 30%
aqueous NaOH (310 mL, 0.4 M) was degassed using a nitrogen
stream then heated to a gentle reflux (125 °C) for 3 h. The
solution was cooled to 0 °C and acidified to pH 3-4 using
concentrated HCl (appox. 200 mL). The solution was extracted
with ether (750 mL) and washed with water (200 mL). The
organic layer was dried over Na2SO4, filtered, and treated with
2,2-di-tert-butyl-4-methylphenol (0.135 g, 0.5 mol %). After
concentrating to dryness, the crude product was dissolved in
anhydrous methanol (200 mL) and treated with an HCl
solution in 1,4-dioxane (37 mL, 4 N, 148 mmol). The resulting
mixture was concentrated to dryness, triturated with isopropyl
ether (100 mL), and filtered to give 2-amino-3,4,6-trifluo-
rothiophenol hydrochloride (19.3 g, 73%) as a light brown solid
that was used without further purification: mp 121-124 C;
Rf 0.43 (30% ethyl acetate in heptane). Anal. (C6H5ClF3NS)
C, H, N, S.
Step 7: 3-(4,5,7-Trifluorobenzothiazol-2-yl)methylin-
dole-N-acetic Acid (9). A solution of 3-(4,5,7-trifluoroben-
zothiazol-2-yl)methylindole-N-acetic acid ethyl ester (5.91 g,
14.6 mmol) in 1,2-dimethoxyethane (73 mL, 0.2 M) was cooled
to 0 °C and treated with aqueous NaOH (1.25 N, 58 mL, 73.1
mmol) in a dropwise manner over 15 min. After the addition
was complete, the solution was stirred for an additional 30
min, acidified to pH 3 with 2 N HCl, and concentrated under
reduced pressure. The residue was dissolved in ethyl acetate
(200 mL) and washed with saturated aqueous NaCl (30 mL).
The organic extract was dried over Na2SO4, filtered, and
concentrated. The resulting material was stirred as a suspen-
sion in heptane, filtered, and dried to give 3-(4,5,7-trifluo-
robenzothiazol-2-yl)methylindole-N-acetic acid (5.38 g, 98%)
as a pale yellow solid: mp 177-178 °C; Rf 0.44 (20% methanol
1
in dichloromethane); H NMR (DMSO-d6, 300 MHz) δ 7.74-
7.65 (m, 1 H), 7.53 (d, J ) 7.5 Hz, 1 H), 7.46 (s, 1 H), 7.40 (d,
J ) 8.1 Hz, 1 H), 7.15 (b t, J ) 6.9 Hz, 1 H), 7.03 (b t, J ) 7.2
Hz, 1 H), 5.03 (s, 2 H), 4.65 (s, 2 H); LRMS calcd for
C18H11F3N2O2S 376.4, found 375.0 (M - 1)-. Anal. (C18H11-
F3N2O2S) C, H, N, S.
Examples with a phenyl substituent coupled to the 5- and
6-positions of the indole ring (compounds 37 and 42, respec-
tively) were prepared from the corresponding bromides via a
Suzuki coupling.27 The following example is provided to
illustrate the general method.
Preparation of 6-Phenylindole. A solution of 6-bromoin-
dole (2.0 g, 10.20 mmol) in anhydrous toluene (20 mL) under
a nitrogen atmosphere was treated with Pd[P(Ph3)]4 (10 mol
%). After stirring for 30 min, a solution of phenylboronic acid
(1.87 g, 15.30 mmol) in anhydrous ethanol (10 mL) and
saturated aqueous NaHCO3 (6 mL) was added and the biphasic
mixture was heated to reflux for 24 h. After cooling to room
temperature, the mixture was added to a saturated aqueous
NaCl solution and extracted with ethyl acetate (2×). The
combined organic layers were dried over MgSO4, filtered, and
concentrated in vacuo. The resulting residue was purified by
flash column chromatography (50% dichloromethane in hep-
tane) to give the desired material as a white powder (900 mg,
45%): 1H NMR (DMSO-d6, 300 MHz) δ 11.15 (br s, 1H), 7.58-
7.66 (m, 4H), 7.41-7.47 (m, 2H), 7.36 (m, 1H), 7.26-7.31 (m,
2H), 6.42 (m, 1H).
Step 5: 3-Cyanomethylindole-N-acetic Acid, Ethyl
Ester (7). Under an atmosphere of nitrogen, a solution of
3-indolyl acetonitrile (25.0 g, 160 mmol) in dry acetonitrile (530
mL, 0.3 M) was treated with sodium hydride (95%, 4.2 g, 168
mmol) and stirred for 30 min. Ethyl bromoacetate (21.3 mL,
192 mmol) was added in a dropwise manner over 10 min and
the solution was stirred at room temperature for 16 h. After
concentrating under reduced pressure, the resulting residue
was dissolved in ethyl acetate and washed with saturated
aqueous NaCl. The organic extracts were dried over MgSO4,
filtered, and concentrated. The crude product was recrystal-
lized from heptane and ethyl acetate to give the target
compound as a white crystalline solid (19 g, 49%): mp 98-99
The 5-morpholino and 5-phenoxy examples (15 and 16) were
prepared from 4-fluoro-2-methylnitrobenzene 10 as illustrated
in Scheme 3 and described in detail below.
Preparation of 5-Morpholino-2-nitrotoluene (11). A
solution of 5-fluoro-2-nitrotoluene (5.11 g, 32.9 mmol), mor-
pholine (4.31 mL, 49.4 mmol), and K2CO3 (6.83 g, 49.4 mmol)
in anhydrous DMSO (80 mL) was heated to 80 °C for 24 h.
After cooling to room temperature, deionized water was added
and the resulting mixture was extracted with ethyl acetate (3
× 50 mL). The organic layer was washed with saturated
aqueous NaCl (100 mL), dried over MgSO4, filtered, and
concentrated. The remaining solid was triturated in heptane
(200 mL) and filtered to give the desired material (7.10 g, 97%)
as a yellow powder: Rf 0.40 (25% ethyl acetate in heptane):
1H NMR (DMSO-d6, 300 MHz) δ 7.96 (d, J ) 9.9 Hz, 1 H),
8.85-8.88 (m, 2 H), 3.70 (t, J ) 5.0 Hz, 4 H), 3.35 (t, J ) 5.0
Hz, 4 H), 2.53 (s, 3 H).
Preparation of 5-Morpholinoindole (13). Under an
atmosphere of nitrogen, a solution of 5-morpholinyl-2-nitro-
toluene (7.0 g, 31.5 mmol) in DMF (100 mL) was treated with
dimethylformamide dimethyl acetal (4.81 mL, 36.2 mmol) and
pyrrolidine (2.62 mL, 31.5 mL), and heated to 100 °C for 12 h.
After cooling to room temperature, the mixture was concen-
trated to give the crude enamine as a brick-red solid.
1
°C; Rf 0.29 (30% ethyl acetate in heptane); H NMR (DMSO-
d6, 300 MHz) δ 7.59 (dd, J1 ) 7.8 Hz, J2 ) 0.6 Hz, 1 H), 7.40
(dd, J1 ) 8.1 Hz, J2 ) 0.6 Hz, 1 H), 7.36 (s, 1 H), 7.18 (b t, J
) 7.2 Hz, 1 H), 7.10 (b t, J ) 7.2 Hz, 1 H), 5.12 (s, 2 H), 4.14
(q, J ) 7.2 Hz, 2 H), 4.06, (s, 2 H), 1.20 (t, J ) 7.2 Hz, 3 H);
LRMS calcd for C14H14N2O2 242.3, found 243.0 (M + 1)+. Anal.
(C14H14N2O2) C, H, N.
Step 6: 3-(4,5,7-Trifluorobenzothiazol-2-yl)methylin-
dole-N-acetic Acid, Ethyl Ester (8). Under a nitrogen
atmosphere, a solution of 3-acetonitrileindole-N-acetic acid,
ethyl ester (11.0 g, 45.4 mmol) and 2,6-di-tert-butyl-4-meth-
ylphenol (BHT, 10 mg) in anhydrous trifluoroethanol (90 mL,
0.5 M) was treated with 2-amino-3,4,6-trifluorothiophenol
hydrochloride (12.7 g, 59.0 mmol) and heated to a gentle reflux
for 16 h. After cooling to room temperature, the solution was
concentrated under reduced pressure, diluted with ethyl
acetate, and washed with 2 N HCl and saturated aqueous
NaCl. The organic layer was dried over MgSO4, filtered, and
concentrated. Purification by MPLC (10-50% ethyl acetate in
heptane, 23 mL/min, 150 min) to give 3-(4,5,7-trifluoroben-
zothiazol-2-yl)methylindole-N-acetic acid ethyl ester (10.8 g,
59%) as a white crystalline solid: mp 110-111 °C; Rf 0.41 (30%
ethyl acetate in heptane); 1H NMR (DMSO-d6, 300 MHz) δ
7.74-7.66 (m, 1 H), 7.54 (d, J ) 7.8 Hz, 1 H), 7.46 (s, 1 H),
7.40 (d, J ) 8.1 Hz, 1 H), 7.15 (br t, J ) 6.9 Hz, 1 H), 7.04 (br
t, J ) 7.8 Hz, 1 H), 5.14, s, 2 H), 4.66 (s, 2 H), 4.14 (q, J ) 7.2
Hz, 3 H); LRMS calcd for C20H15F3N2O2S 404.4, found 405.0
(M + 1)+. Anal. (C20H15F3N2O2S) C, H, N, S.
The intermediate enamine was dissolved in ethyl acetate
(200 mL) and added to a precharged Parr bottle containing
10% Pd/C (600 mg) in ethyl acetate (40 mL). The mixture was
hydrogenated on a Parr shaker at 55 psi for 2.5 h. After the
hydrogenation was complete, the reaction mixture was filtered
through a Celite plug and washed with ethyl acetate. The
resulting filtrate was concentrated and purified by flash
column chromatography (50% heptane in ethyl acetate) to give

Potent and Selective Inhibitors of Aldose Reductase
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 9 3149
2.0 g (31% over 2 steps) of the desired indole as a off-white
powder: Rf 0.30 (10% methanol in chloroform); 1H NMR
(DMSO-d6, 300 MHz) δ 10.77 (br s, 1 H), 7.24 (s, 1 H), 7.18-
7.20 (m, 1 H), 6.97 (d, J ) 1.8 Hz, 1 H), 6.81 (dd, J1 ) 8.7 Hz,
J2 ) 2.1 Hz, 1 H), 6.25 (dd, J1 ) 3.0 Hz, J2 ) 1.8 Hz, 1 H), 3.7
(t, J ) 4.50 Hz, 4 H), 2.96 (t, J ) 4.50 Hz, 4 H).
added to initiate the reaction. The microplate is placed in the
plate reader for 30 min at 340 nm and read at 2 min intervals.
The reaction contents in a final volume of 300 µL are 6.6%
w/v (NH4)2SO4, 33 mM NaH2PO4 (pH 6.6), 0.11 mM NADPH,
4.7 mM DL-glyceraldehyde, 0.59 µg of enzyme, 1% DMSO, and
compound. Percent inhibition is calculated on the basis of the
enzyme activity in the presence or absence of compound. The
IC50 is calculated using SAS release 6.10 and a four-parameter
logistic regression of data.
The inhibition of aldehyde reductase (ALR1) was assessed
using a standard spectrophotometric assay.31 The procedure
is similar to the method described above for determining ALR2
inhibition. The reaction contents in a final volume of 300 µL
are 33 mM NaH2PO4 (pH 6.6), 0.11 mM NADPH, 4.7 mM
DL-glyceraldehyde, 0.28 µg of enzyme, 1% DMSO, and com-
pound.
STZ-Induced Diabetic Rat Model (General). Male
Sprague-Dawley rats (Harlan, Indianapolis, IN), weighing
150 g, are allowed to acclimate for 1 week and placed on a
standard diet (7012, Harlan Teklad certified LM-485 mouse/
rat, Madison, WI). Water is supplied ad libitum. After a 24 h
fast, 30 mg/kg STZ (S01230 Sigma, St. Louis, MO) prepared
in 0.03 M citrate buffer, pH 4.5, is administered ip. Control
animals receive citrate buffer.
Preparation of 5-Phenoxy-2-nitrotoluene (12). A solu-
tion of phenol (12.16 g, 0.129 mol) in anhydrous DMSO was
treated with K2CO3 (17.88 g, 0.129 mol) and stirred at room
temperature for 15 min. 5-Fluoro-2-nitrotoluene (13.38 g, 0.086
mol) was added and the resultant mixture was heated to 80
°C for 12 h. After cooling to room temperature, the mixture
was poured into deionized water (100 mL) and extracted with
ethyl acetate (2 × 100 mL). The combined organic layers were
washed with a saturated aqueous NaCl, dried over MgSO4,
filtered, and concentrated. The resulting residue was purified
by flash column chromatography (0-10% ethyl acetate in
heptane) to give 5-phenoxy-2-nitrotoluene as a yellow crystal-
line solid (12.50 g, 63%). Rf 0.60 (15% ethyl acetate in heptane);
1H NMR (DMSO-d6, 300 MHz) δ 8.05 (d, J ) 9.0 Hz, 1 H),
7.44-7.47 (m, 2 H), 7.23-7.29 (m, 1 H), 7.12-7.16 (m, 2 H),
7.04 (d, J ) 2.7 Hz, 1 H), 6.90 (dd, J1 ) 9.0 Hz, J2 ) 2.7 Hz,
1 H), 2.51 (s, 3 H).
Preparation of 5-Phenoxyindole (14). A solution of
5-phenoxy-2-nitrotoluene (10.03 g, 0.0428 mol) in anhydrous
DMF was treated with N,N-dimethylformamide dimethyl
diacetal (6.73 mL, 0.0508 mol) and pyrrolidine (3.63 mL,
0.0438 mol) and heated to 110 °C for 2.5 h. After cooling to
room temperature, the mixture was diluted with ethyl acetate
(500 mL) and washed successively with deionized water (500
mL) and saturated aqueous NaCl. The remaining organic layer
was dried over MgSO4, filtered, and concentrated. The crude
intermediate was dissolved in glacial acetic acid (250 mL) and
warmed to 85 °C. Zinc ribbon (24.62 g, 0.377 mol) was added
to the solution over 30 min and the mixture was heated to
reflux for 4 h. After cooling to room temperature, the mixture
was neutralized with saturated aqueous NaHCO3 and ex-
tracted with diethyl ether (3 × 300 mL). The combined organic
layers were washed with saturated aqueous NaCl, dried over
MgSO4, filtered, and concentrated. The resulting residue was
purified by flash column chromatography (0-30% ethyl acetate
in heptane) to give the 5-phenoxyindole as a white crystalline
powder (3.1 g, 34%, 2 steps): Rf 0.50 (30% ethyl acetate in
Four days after receiving STZ, tail blood glucose is measured
using a One Touch II blood glucose meter (Lifescan, Inc.,
Johnson & Johnson, Milpitas, CA). Diabetic rats whose blood
glucose levels are g300 mg/dL are randomized according to
blood glucose levels into diabetic-control and diabetic-treat-
ment groups.
Five-Day STZ-Induced Diabetic Rat Screening Model.
Using the STZ-induced diabetic rat model (as described above),
each study includes nondiabetic controls (n ) 5), diabetic
controls (n ) 10), and diabetic rats for each drug treated group
(n ) 7). Beginning on day 4 after induction of diabetes with
STZ, the compound, suspended in a vehicle of 2% Tween 80
in saline, is administered po consecutively for 5 days. Non-
diabetic and diabetic controls receive vehicle during this
period. After the initial dose of compound, animals continue
to receive the standard diet for the remainder of the study.
After receiving the final dose of compound, the food is removed.
Four hours after the final dose is administered, the rats are
euthanized by asphyxiation with CO2 and sciatic nerves and
lens are removed, weighed, frozen on porcelain plates on dry
ice, and stored at -80 °C for sorbitol analysis.
1
heptane); H NMR (DMSO-d6, 300 MHz) δ 11.12 (br s, 1 H),
7.48 (s, 1 H), 7.30-7.38 (m, 1 H), 7.25-7.29 (m, 2 H), 7.17 (d,
J ) 2.7 Hz, 1 H), 6.89-7.02 (m, 1 H), 6.86-6.88 (m, 2 H), 6.80
(dd, J1 ) 8.7 Hz, J2 ) 2.4 Hz, 1 H), 6.37 (m, 1 H).
Tissue Sorbitol Analysis. Sciatic nerve sorbitol levels are
measured by a standard enzymatic method that uses sorbitol
dehydrogenase.32 A D-sorbitol kit purchased from Boehringer
Mannheim (670 057 Indianapolis, IN) was adapted to a
microplate format.
Pharmacology. Enzyme Preparation. Recombinant hu-
man aldose reductase (hALR2) with a His6-tag was overex-
pressed in E. coli using a T7-based expression system and
purified using a metal affinity column.28 The enzyme was
precipitated in 3.5 M ammonium sulfate, 10 mM Tris-HCl
buffer, and 1 mM DTT. The enzyme was centrifuged at 10 000
rpm for 10 min at 4 °C and the pellet was resuspended to an
approximate concentration of 10-15 mg of protein/mL in
(NH4)2SO4 in sodium phosphate, pH 6.6. Protein content was
determined using Coomassie Protein Assay reagent (Pierce,
Rockford, IL). Aliquots were made and the enzyme was stored
at -40 °C.
Recombinant human aldehyde reductase (hALR1), supplied
by Dr. K. Gabbay (Baylor College of Medicine, Houston, TX),
was prepared as previously described.29 The enzyme was stored
in 5 mM sodium phosphate, pH 7.4, and 0.1 mM EDTA at 4
°C. Protein content was determined using Coomassie Protein
Assay reagent (Pierce, Rockford, IL).
Enzymatic Assays. The method to examine the inhibition
of aldose reductase (ALR2) was similar to those methods used
by numerous other laboratories.30 Each assay is done in
triplicate. Enzymatic activity was measured by monitoring the
disappearance rate of NADPH. Compound or solvent is placed
in the microplate well to which buffer is added and mixed for
1 min. Enzyme is added to the well and incubated at 37 °C
and shaken for 10 min. DL-Glyceraldehyde and NADPH are
Tissues are homogenized in ice-cold 0.2% perchloric acid in
methanol using a DUALL tissue grinder and centrifuged at
10 000 rpm for 10 min at 4 °C. A sampling of the supernatant
is transferred to a microplate. A reaction mixture combining
buffer, diaphorase, NAD+, and iodonitrotetrazolium chloride
(INT) is added to each sample followed by an initial absorbance
reading at 490 nm. Sorbitol dehydrogenase (SDH) is added to
initiate the reaction and the plate is mixed for 90 s on an
orbital shaker. Plates are incubated at room temperature in
the dark for 30 min and the final absorbance reading at 490
nm is made. The reaction contents in a final volume of 255 µL
at pH 8.7 are 14 mM potassium phosphate, 94 mM trietha-
nolamine, 0.94% Triton X-100, 0.16 U diaphorase, 1.66 mM
NAD, 0.034 mM INT, and 0.82 U SDH.
Sciatic nerve sorbitol content is calculated as nmol of
sorbitol/mg of wet tissue weight by taking the difference in
the final and initial absorbance readings and comparing to a
sorbitol standard curve. Percent inhibition is calculated by
comparing treatment groups to both the nondiabetic and
diabetic control groups within the study as described by the
equation: (t - n)/(d - n) where n ) nondiabetic control, d )
diabetic control, and t ) drug-treated diabetic control. The
ED50 was calculated using Sigma plot, version 4.0, using a Hill
slope four-parameter logistic equation.

3150 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 9
Van Zandt et al.
Motor Nerve Conduction Velocity (MNCV). Using the
STZ-induced diabetic rat model (as described above), each
study includes nondiabetic controls, diabetic controls, and
diabetic rats for each drug treated group (prevention study, n
) 10 per group; intervention study, n ) 12 per group). MNCV
was measured in the sciatic/tibialis-interosseous system by
the method of Sharma.33 The rats are anesthetized with 4%
isoflurane in oxygen (1 L/min) for induction, 2.5% for controls,
and 2% for diabetics. After the initial anesthetization, flow is
reduced to 0.5 L/min for maintenance. Near-nerve/subcutane-
ous temperature was recorded and maintained at 37 °C using
a wire thermister and thermometer (YSI Tele-thermometer,
Model 4000A, Yellow Springs Instrument Co. Inc.), a Heat
Therapy System (Baxter K-MOD 100 pump and Duo Therm
pad, Baxter Healthcare Corp, Deerfield, IL), and a source of
radiant heat. The left sciatic/tibial nerve was stimulated at
the sciatic notch and Achilles tendon by square wave pulses
(duration ) 0.1 m/second, intensity ) 20 V) and compound
muscle action potentials were recorded from the interosseus
muscle using needle electrodes (Sapphire II, 4ME electro-
physiological system, Teca Corp, Pleasantville, NY). An aver-
age of 10 action potentials traces was recorded for each
stimulation point. The latencies to inflection (takeoff) of the
compound muscle action potential were measured for each
stimulation point, and the nerve length was measured after
dissection. MNCV ) nerve length (mm)/[sciatic notch latency
- Achilles tendon latency (m/s)]. When the measurements
were completed, rats were euthanized and both sciatic nerves
were removed, weighed, and frozen for polyol determination
by capillary gas chromatography. Data from all groups were
compared using one-way analysis of variance and the Tukey
post-test (Graph Pad Prism, version 2).
Cataract Assessment by Slit-Lamp Biomicroscopy.
Using the STZ-induced diabetic rat model (as described above),
each study includes nondiabetic controls, diabetic controls, and
diabetic rats for each drug treated group (n ) 8 per group).
Assessment of cataract progression was conducted by slit-lamp
biomicroscopy (American Optical Slit-lamp biomicroscope-
model 11665, SN AP5813) in a dark room. Two drops of 1%
mydriacyl (Alcon Inc., Ft. Worth, TX) were administered to
cause full dilation of the pupil. Rats were maintained anes-
theized by an inhalation system, which provided continuously
delivered isoflurane (Ohmeda Isotec 3, BOC Health Care,
Yorkshire, England). While under anesthesia, rats were
positioned in front of the slit-lamp on an adjustable height
platform.
kg dose of [14C]lidorestat, by either the intravenous or oral
route (n ) 6 per group). Blood was drawn at 0.083 (iv only),
0.25, 0.5, 1, 2, 4, 7, 12, and 24 h following each dose via a
surgically implanted jugular vein cannula. The corresponding
plasma was harvested and analyzed for total radioactivity and
unchanged drug using a scintillation counter and LC/MS/MS.
Pharmacokinetic parameters were calculated using model-
independent pharmacokinetics. Whole-body autoradiography
was used to determine drug concentrations in specific tissues.
X-ray Crystallography. Overproduction and Purifica-
tion of Recombinant Human ALR2. The ORF of the human
ALR2 gene (Accession GenBank/embl Data Bank Number
J05017) was amplified by PCR from its cDNA and cloned into
the T7 RNA polymerase-based vector pET15b (Novagen).
Expression of the (His)6-human ALR2 in the E. coli strain
BL21(DE3) (Novagen) is induced by IPTG (Euromedex) during
a 3 h culture at 37 °C. The pellet from a 4 L culture was
disrupted by sonication and centrifuged at 4 °C. The super-
natant was applied on a Talon metal affinity column (Clon-
tech). After thrombin cleavage of the hexahistidine extension,
the detagged protein was then loaded on a DEAE Sephadex
A-50 column (Pharmacia) and eluted with a NaCl gradient.
Crystallization Trials and Preliminary Crystallo-
graphic Analysis. All trials were carried out in Linbro 24-
well tissue culture plates (Flow Laboratories) using the
hanging-drop vapor diffusion method. The enzyme was dia-
lyzed against 50 mM ammonium citrate, pH 5.0. The drops of
a final volume of 12 mL contained 15 mg/mL of human ALR2
with 2 equiv of NADP+ (Sigma) and 5% of PEG 6000; they
were equilibrated against a well containing 120 mM am-
monium citrate, pH 5.0, 20% PEG 6000 and the same
concentration of NADP+ as above. Crystals were grown at 4
°C and reached the size of 0.5 × 0.4 × 0.4 mm.
Soaking with the Inhibitor. Crystals of the ALR2-
inhibitor complex were obtained by soaking of the native
crystals with a solution containing 2 mg/mL of inhibitor
dissolved in the mother liquor containing 120 mM ammonium
citrate, pH 5.0, and 25% of PEG 6000. The soaking time was
3 weeks.
Data Collection and Refinement. The diffraction analy-
sis was performed using a rotating anode laboratory source
and an image plate, and the data were treated with the HKL
package.36 The inhibitor-soaked crystals diffracted to 1.04 Å
and were isomorphous to the native crystals. The native
structure was used as a model for the molecular replacement;
after a rigid body minimization, the inhibitor was placed in
the electron density shown by a Fourier difference map. The
model of the complex was used as starting point of the
refinement, which included a rigid-body refinement step, a
slow-cooling step, a Powell minimization, and a temperature-
factor refinement. Water molecules were then placed in a
difference map and minimization was performed as above
without putting restraints on the water molecules; bulk solvent
correction was also applied. This was followed by anisotropic
refinement and H-atom addition. The programs O,37 X-PLOR,38
and SHELXL39 were used for model building and refinement.
In vivo photodocumentation was conducted by using a
reversal-coupling ring for reverse mounting of a 28 mm f/2.8
lens on a camera body (Pentax). The camera was placed on a
minitripod with the camera lens maintained at the largest
aperture. Color reversal film (AGFAchrome RSX 50) was used
to capture the images.
Triglyceride Determination.34 The assay used to dem-
onstrate the compounds ability to lower elevated serum
triglyceride levels is described as follows. Using the STZ-
induced diabetic rat model (as described above), the study
includes a nondiabetic control group (n ) 5), an untreated
diabetic control group (n ) 7), and drug treated diabetic group
(n ) 7). The daily dosages are administered by gavage as a
single dose of the test compound in 2% Tween 90 in saline
each day. The nondiabetic and diabetic control groups are
administered vehicle.
Diffraction experiments were conducted on one crystal at
the X-ray beam of the undulator line W32 at LURE. The
atomic coordinates and structure factors of the hALR2-
NADP+-9 (lidorestat) complex have been deposited in the
Protein Data Bank, accession code 1Z3N.
After the final dose, all groups of animals are fasted for 4 h
and anesthetized with CO2, and blood is collected by cardiac
puncture into EDTA tubes. The plasma is separated from the
red blood cells by centrifugation. Plasma triglyceride levels are
quantitated on an automated COBAS chemistry system utiliz-
ing the Roche reagent for triglycerides (Cat. #44119).35 Plasma
samples are diluted with 1:1 saline and run in quadruplicate.
Data are given as mean ( SEM. Statistical comparisons
between groups employed a one-tailed t-test and demonstrated
statistical significance with p > 0.01.
Acknowledgment. This work was supported, in
part, by the Centre National de la Recherche Scienti-
fique (CNRS), the Institut National de la Sante´ et de la
Recherche Me´dicale, and the Hoˆpital Universitaire de
Strasbourg (H.U.S.). We thank the personnel of the
LURE for their help in data collection.
Supporting Information Available: 1H NMR spectra
and combustion analysis. This material is available free of
charge via the Internet at http://pubs.acs.org.
Pharmacokinetics and Tissue Distribution. In an un-
paired study, 12 male Sprague Dawley rats received a 10 mg/

Potent and Selective Inhibitors of Aldose Reductase
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 9 3151
741-744. (b) Suzuki, K.; Koh, Y. H.; Mizuno, H.; Hamaoka, R.;
Taniguchi, N. Overexpression of aldehyde reductase protects
PC12 cells from the cytotoxicity of methylglyoxal or 3-deoxyglu-
cosone. J. Biochem. 1998, 123, 353-357.
References
(1) American Diabetes Association’s homepage. http://www.diabe-
tes.org.
(2) (a) 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. (b) The United Kingdom Prospective Diabetes Study
Group. Intensive blood-glucose control with sulfonyl-ureas or
insulin compared with conventional treatment and risk of
complications in patients with type 2 diabetes (UKPDS33).
Lancet 1998, 352, 837-853.
(15) (a) Takahashi, M.; Fujii, J.; Teshima, T.; Suzuki, K.; Shiba, T.;
Taniguchi, N. Identity of a major 3-deoxyglucosone-reducing
enzyme with aldehyde reductase in rat liver established by
amino acid sequencing and cDNA expression. Gene, 1993, 127,
249-253. (b) Kanazu, T.; Shinoda, M.; Nakayama, T.; Deyashiki,
Y.; Hara, A.; Sawada, H. Aldehyde reductase is a major protein
associated with 3-deoxyglucosone reductase activity in rat, pig
and human livers. Biochem. J. 1991, 279, 903-906.
(16) Hamada, Y.; Araki, N.; Koh, N.; Nakamura, J.; Horiuchi, S.;
Hotta, N. Rapid formation of advanced glycation end-products
by intermediate metabolites of glycolytic pathway and polyol
pathway. Biochem. Biophys. Res. Commun. 1996, 228, 539-543.
(17) (a) Oya, T.; Hattori, N.; Mizuno, Y.; Miyata, S.; Maeda, S.;
Osawa, T.; Uchida, K. Methylglyoxal Modification of Protein. J.
Biol. Chem. 1999, 274, 18492-18502. (b) Cavalot, F.; Anfossi,
G.; Russo, I.; Mularoni, E.; Massucco, P.; Mattiello, L.; Burzacca,
S.; Hahn, A.; Trovati, M. Nonenzymatic Glycation of Fibronectin
Impairs Adhesive and Proliferative Properties of Human Vas-
cular Smooth Muscle Cells. Metabolism 1996, 45, 285-292. (c)
Fu, M.; Wells-Knecht, K.; Blackledge, J. Lyons, T.; Thorpe, S.;
Baynes, J. Glycation, Glycoidation, and cross-Linking of Collagen
by Glucose. Diabetes 1994, 43, 676-683.
(3) (a) Textbook of Diabetes; Pickup, J. C., Williams, G., Eds.;
Blackwell Science Ltd: United Kingdom, 1997; Vol. 1, Chapter
42. (b) Williamson, J. R.; et al. Hyperglycemic pseudohypoxia
and diabetic complications. Diabetes 1993, 42, 801-813. (c)
Setter, S. M.; Campbell, R. K.; Cahoon, C. J. Biochemical
pathways for microvascular complications of diabetes mellitus.
Ann. Pharmacother. 2003, 37, 1858-1866.
(4) Mylari, B. L.; Armento. S. J.; Beebe, D. A.; Conn, E. L.; Coutcher,
J. B.; Dina, M. S.; O’Gorman, M. T.; Linhares, M. C.; Martin,
W. H.; Oates, P. J.; Tess, D. A.; Withbroe, G. J.; Zembrowski,
W. J. A highly selective, non-hydantoin, non-carboxylic acid
inhibitor of aldose reductase with potent oral activity in diabetic
rat models: 6-(5-Chloro-3-methylbenzofuran-2-sulfonyl)-2-H-
pyridazin-3-one. J. Med. Chem. 2003, 46, 2283-2286.
(5) Greene, D. A.; Sima, A. F. A.; Stevens, M. J.; Feldman, E. L.;
Lattimer, S. A.; Complications: Neuropathy, pathogenetic con-
siderations. Diabetes Care 1992, 15, 1902-1925.
(6) Green, D. A,; Lattimer, S. A.; Sima, A. A. F. Sorbitol,
pophoinosititides, and sodium-potassium-ATPase in the patho-
genesis of diabetic complications, N. Engl. J. Med. 1987, 316,
599-606.
(7) Lee, A.; Chung, S. K.; Chung, S. S. Demonstration that polyol
accumulation is responsible for diabetic cataract by use of
transgenic mice expressing the aldose reductase gene in the lens.
Proc. Natl. Acad. Sci. 1995, 92, 2780-2784.
(18) (a) Aotsuka, T.; Abe, N.; Fukushima, K.; Ashizawa, N.; Yoshida,
M. Benzothiazol-2-yl carboxylic acids with diverse spacers:
A
novel class of potent, orally active aldose reductase inhibitors.
Bioorg. Med. Chem Lett., 1997, 7, 1677-1682. (b) Aotsuka, T.;
Hosono, H.; Kurihara, T.; Nakamura, Y.; Matsui, T.; Kobayashi,
F. Chem. Pharm. Bull. 1994, 42, 1264-1271. (c) Matsui, T.;
Nakamura, Y.; Ishikawa, H.; Matsuura, A.; Kobayashi, F. Jpn.
J. Pharmacol. 1994, 64, 115.
(19) Somei, M.; Kizu, K.; Kunimoto, M.; Yamada, F. Chem. Pharm.
Bull. 1985, 33, 3996.
(20) Urzhumtsev, A.; Teˆte-Favier, F.; Mitschler, A.; Barbanton, J.;
Barth, P.; Urzhumtseva, 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 phar-
maceutically important inhibitors tolrestat and sorbinil. Struc-
ture 1997, 5, 601-612.
(8) (a) Oates, P. J.; Mylari, B. L. Aldose reductase inhibitors:
Therapeutic implications for diabetic complications. Exp. Opin.
Invest. Drugs 1999, 8, 1-15. (b) Ramasamy, R. Aldose reduc-
tase: A novel target for cardioprotective interventions. Current
Drug Targets 2003, 4, 625-632.
(21) (a) El-Kabbani, O.; Wilson, D. K.; Petrash, J. M.; Quiocho, F. A.
Structural features of the aldose reductase and aldehyde reduc-
tase inhibitor-binding sites. Mol. Vis. 1998, 4, 19-25. (b) 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. (c) 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 ter-
nary complex. Proteins 1997, 29, 186-192. (d) Sotriffer, C. A.;
Kramer, O.; Klebe, G. Probing flexibility and “induced-fit”
phenomena in aldose reductase by comparative crystal structure
analysis and molecular dynamics simulations. Proteins: Struct.,
Funct. Bioinformatics 2004, 56, 52-66. (e) Harrison, D. H. T.;
Bohren, K. M.; Petsko, G. A.; Ringe, D.; Gabbay, K. H. The
alrestatin double-decker: Binding of two inhibitor molecules to
human aldose reductase reveals a new specificity determinant.
Biochemistry 1997, 36, 16134-16140.
(22) (a) El-Kabbani, O.; Ruiz, F.; Darmanin, C.; Chung, R. P. Aldose
reductase structures: Implications for mechanism and inhibi-
tion. Cell. Mol. Life Sci. 2004, 61, 750-762. (b) Petrash, J. M.
All in the family: Aldose reductase and closely related aldo-keto
reductases. Cell. Mol. Life Sci. 2004, 61, 737-749.
(23) Mylari, B.; Larson, E.; Beyer, T.; Zembrowski, W.; Aldinger, C.;
Dee, M.; Siegel, T.; Singleton, D. Novel, potent aldose reductase
inhibitors: 3,4-Dihydro-4-oxo-3-[[5-(trifluoromethyl)-2-benzothi-
azoyl]methyl]-1-phthalazine-acetic acid (zopolrestat) and con-
geners. J. Med. Chem. 1991, 34, 108-122.
(24) (a) Nagata, M.; Hohman, T. C.; Nishimura, C.; Drea, C. M.;
Oliver, C.; Robison, W. G., Jr. Polyol and vacuole formation in
cultured canine lens epithelial cells. Exp. Eye Res. 1989, 48,
667-677. (b) Kinoshita, J. H. Aldose reductase in the diabetic
eye. XLIII Edward Jackson Memorial Lecture. Am. J. Ophthal-
mol. 1986, 102, 685-692. (c) Cogan, D. G.; Kinoshita, J. H.;
Kador, P. F.; Robison, W. G., Jr.; Datilis, M. B.; Cobo, L. M.;
Kupfer, C. Aldose reductase and complications of diabetes. Ann.
Intern. Med. 1984, 101, 82-91.
(25) Lerner, B. C.; Varma, S.; Richards, R. Polyol pathway metabo-
lites in human cataracts. Arch. Ophthalmol. 1984, 102, 917-
920.
(26) Values for drug-treated diabetic patients in study and typical
untreated diabetic patients (not in current study) provided by
Dr. Henry Powell, M.D., D.Sc., University of California, San
Diego.
(9) (a) Kao, Y.-L.; Donahue, K.; Chan, A.; Knight, J.; Silink, M. A
novel polymorphism in the aldose reductase gene promoter
region is strongly associated with diabetic retinopathy in
adolescents with Type I diabetes. Diabetes 1999, 48, 1338-1340.
(b) Heesom, A.; Hibberd, M.; Millward, A.; Demaine, A. Poly-
morphism in the 5′-end of the aldose reductase gene is strongly
associated with the development of diabetic nephropathy in type
I diabetes. Diabetes 1997, 46, 287-291. (c) Moczulski, D.; et al.
The role of aldose reductase gene in the susceptibility to diabetic
nephropathy in type II (non-insulin-dependent) diabetes melli-
tus. Diabetologia 1999, 42, 94-97. (d) Demaine, A.; Cross, D.;
Millward, A. Polymorphism of the aldose reductase gene and
susceptibility to retinopathy in type 1 diabetes mellitus. Invest.
Ophthalmol. Visual Sci. 2000, 41, 4064-4068. (e) Heesom, A.
E.; Millward, A.; Demaine, A. G. Susceptibility to diabetic
neuropathy in patients with insulin dependent diabetes mellitus
is associated with a polymorphism at the 5′-end of the aldose
reductase gene. J. Neurol., Neurosurg. Psychiat. 1998, 64, 213-
216. (f) Taverna, M. J. Genetics of diabetic complications:
Retinopathy. Ann. Endocrinol. 2004, 65 (Suppl. 1), 1S17-1S25.
(10) Ho, E. C. M.; Lam, K. S. L.; Chung, S. S. M. Aldose reductase
deficient mice are alleviated from depletion of GSH in the
peripheral nerve and MNCV deficit associated with diabetes.
Diabetes 2001, 50, 59.
(11) (a) Lajer, M.; Tarnow, L.; Fleckner, J.; Hansen, B. V.; Edwards,
D. G.; Parving, H.; Boel, E. Association of aldose reductase
gene Z+2 polymorphism with reduced susceptibility to diabetic
nephropathy in caucasian type 1 diabetic patients. Diabetic
medicine 2004, 21, 867-873. (b) Sivenius, K.; Pihlajamaki, J.;
Partanen, J.; Niskanen, L.; Laakso, M.; Uusitupa, M. Aldose
reductase gene polymorphisms and peripheral nerve function
in patients with type 2 diabetes. Diabetes Care 2004, 27, 2021-
2026.
(12) (a) Miyamoto, S. Recent advances in aldose reductase inhibi-
tors: Potential agents for the treatment of diabetic complica-
tions. Exp. Opin. Ther. Patents 2002, 12, 621-631. (b) Suzen,
A.; Buyukbingol, E. Recent studies of aldose reductase enzyme
inhibition for diabetic complications. Curr. Med. Chem. 2003,
10, 1329-1352.
(13) Hamada, Y.; Nakamura, J. Clinical potential of aldose reductase
inhibitors in diabetic neuropathy. Treatments Endocrinol. 2004,
3, 245-255.
(14) (a) Shinoda, M.; Mori, S.; Shintani, S.; Ishikura, S.; Hara, A.
Inhibition of Human Aldehyde Reductase by Drugs for Testing
the Function of Liver and Kidney. Biol. Pharm. Bull. 1999, 22,

3152 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 9
Van Zandt et al.
(27) Miyaura, N.; Suzuki, A. Palladium-catalyzed cross-coupling
reactions of organoboron compounds. Chem. Rev. 1995, 95,
2475-2483.
(28) Hayman, S.; Kinoshita, J. H. J. Biol. Chem. 1965, 240, 877-
880.
and Academic Press: New York, 1974; Vol. 3, pp 1323-1330.
(b) Malone, J. I.; Knox, G.; Benford, S.; Tedesco, T. A. Red cell
sorbitol. An indicator of diabetic control. Diabetes 1980, 29, 861-
864.
(33) Sharma, A. K.; Thomas, P. K. Peripheral nerve structure and
function in experimental diabetes. J. Neurol. Sci. 1974, 23, 1-15.
(34) Schnatz, J. D.; Furmaniak, J. M.; Chlouverakis, C. The origin
of hypertriglyceridema in streptozotocin diabetic rats. Diabeto-
logia 1972, 8, 125-130.
(35) Trinder, P. Determination of glucose in blood using glucose
oxidase with an alternative oxygen acceptor. Ann Clin Biochem.
1969, 6, 24.
(36) Otwinowski, Z.; Minor, W. Methods in Enzymology, Macromo-
lecular Crystallography; Carter, C. W., Jr., Sweet, R. M., Eds.;
Academic Press: New York, 1997; Vol. 276, Part A, pp 307-
326
(37) Jones, T. A.; Zou, J. Y.; Cowan, S. W.; Kjeldgaard, M. O.
Improved methods for building protein models in electron
density maps and the location of errors in these models. Acta
Crystallogr. 1991, A47, 110-119.
(38) Brunger, A. T. X-PLOR, version 3.1. A system for X-ray crystal-
lography and NMR. Yale University Press: New Haven, CT,
1992.
(39) Sheldrick, G. M.; Schneider, T. R. Shelxl-high-resolution refine-
ment. Methods Enzymol. 1997, 277, 319-343.
(29) Barski, O. A.; Gabbay, K. H.; Grimshaw, C. E.; Bohren, K. M.
Mechanism of human aldehyde reductase: Characterization of
the active site pocket. Biochemistry 1995, 34, 11264-11275.
(30) (a) Stribling, D.; Mirrlees, D. J.; Harrison, H. E.; Earl, D. C. N.
Properties of ICI 128,436 a novel aldose reductase inhibitor, and
its effects on diabetic complications in the rat. Metabolism 1985,
4, 336-344. (b) Ashizawa, N.; Yoshida, M.; Sugiyama, Y.;
Akaike, N.; Ohbayashi, S.; Aotsuka, T.; Abe, N.; Fukushima, K.;
Matsuura, A. Effects of a novel potent aldose reductase inhibitor,
GP-1447, on aldose reductase activity in vitro and on diabetic
neuropathy and cataract formation in rats. Jpn. J. Pharmacol.
1997, 73, 133-144. (c) Ehrig, T.; Bohren, K. M.; Prendergast,
G.; Gabbay, K. Mechanism of aldose reductase inhibition:
Binding of NADP+/NADPH and alrestatin-like inhibitors. Bio-
chemistry 1994, 33, 7157-7165.
(31) Barski, O. A.; Gabbay, K. H.; Grimshaw, C. E.; Bohren, K. M.
Mechanism of human aldehyde reductase: Characterization of
the active site pocket. Biochemistry 1995, 34, 11264-11275.
(32) (a) Bergmeyer, H. U.; Gruber, W.; Gutmann, I. Methods of
Enzymatic Analysis 2nd ed.; Bergmeyer, H. U., Ed.; Weinheim
JM0492094