Improvement of Motor Nerve Conduction Velocity in Diabetic Rats Requires Normalization of the Polyol Pathway Metabolites Flux

Home , Ranirestat, Zenarestat

Takafumi Matsumoto1,*, Yoshiyuki Ono2, Masuo Kurono3, Akemi Kuromiya4, Keiji Nakamura5, and Vera Bril6 1Pharmacology Research Laboratories, 4Genomic Science Laboratories, Dainippon Sumitomo Pharma Co., Ltd., 1-98, Kasugade-naka 3-chome, Konohana-ku, Osaka 554-0022, Japan 2International Business Management, Dainippon Sumitomo Pharma Co., Ltd., 5-51, Ebie 1-chome, Fukushima-ku, Osaka 553-0001, Japan 3Chemistry Research Laboratories, Dainippon Sumitomo Pharma Co., Ltd., Enoki 33-94, Suita, Osaka 564-0053, Japan 5Analytical Chemistry Center Nichiei Sangyo Co., Ltd., 1-98, Kasugade-naka 3-chome, Konohana-ku, Osaka 554-0022, Japan 6Department of Medicine, University of Toronto EN11-209, Toronto General Hospital, 200 Elizabeth St, Toronto, ON, M5G 2C4, Canada

Received July 4, 2008; Accepted December 8, 2008

Abstract. Using ranirestat, an aldose reductase (AR) inhibitor, we investigated the relationship between sorbitol and fructose levels in the sciatic nerve and motor nerve conduction velocity (MNCV) in streptozotocin (STZ)-treated diabetic rats. Ranirestat inhibited rat and recombinant human AR with similar IC50 values and equipotently prevented sorbitol accumulation in rat erythrocytes and sciatic nerves in vitro. One week after STZ administration, sorbitol levels in rat erythrocytes and sciatic nerves significantly increased while MNCV decreased. Oral administration of ranirestat (0.03 – 1.0 mg/kg per day) for 3 weeks dose-dependently decreased the elevated sorbitol and fructose levels in the rat sciatic nerves without affecting blood glucose level. Particularly, at doses of 0.1 mg/kg per day or higher, ranirestat normalized both sorbitol and fructose levels in the sciatic nerves of STZ-treated rats. Ranirestat (0.1 – 1.0 mg/kg per day) also improved the STZ-induced decrease in MNCV in a dose-dependent manner. This improvement correlated with the decrease of sorbitol and fructose levels in the rat sciatic nerves. These findings indicate that ranirestat improves MNCV via normalization of sorbitol and fructose accumulation in the sciatic nerve.

Keywords: ranirestat, aldose reductase inhibitor, sorbitol, fructose, motor nerve conduction velocity (MNCV)

Introduction diabetic conditions plays an important role in the patho- genesis of diabetic complications, including neuropathy In diabetes mellitus, hyperglycemia results in excess (1 – 3), retinopathy (4, 5), and nephropathy (6). In spite glucose being metabolized through the polyol pathway, of recent progress in the management of diabetic which involves activation of the enzyme aldose patients’ blood sugar levels, these complications continue reductase (AR). This enzyme catalyzes the reduction of to represent serious problems for diabetic patients. Thus, glucose to sorbitol, which is in turn oxidized to fructose as the incidence of diabetes increases, the occurrence of by sorbitol dehydrogenase. This results in the accumula- complications associated with it also increases. The tion of sorbitol and fructose in various tissues, and it is inference that increased sorbitol and/or fructose levels believed that activation of the polyol pathway under in various tissues is the main cause of microvascular diabetic complications is known as the “polyol pathway *Corresponding author. [email protected] theory”. However, it is still unclear whether this theory Published online in J-STAGE on February 10, 2009 (in advance) is correct or not. Previous studies have indicated that doi: 10.1254/jphs.08177FP activated aldose reductase and increased sorbitol levels

203 204 T Matsumoto et al

constant, and Kii is the intercept inhibition constant, that is, the dissociation constant between the isoform (inter- mediate during enzymatic reaction) and AR inhibitor after the substrate (glucose) binds to the enzyme (12).

Sorbitol accumulation in rat erythrocytes and sciatic nerves Sorbitol accumulation in rat erythrocytes was deter- Fig. 1. Chemical structure of ranirestat: (3R)-2'-(4-bromo-2- mined according to the method of Matsumoto et al. (10, fluorobenzyl)spiro[pyrrolidine-3,4'(1'H)-pyrrolo[1,2-a]pyrazine]- 11) with a slight modification. An aliquot of washed 1',2,3',5(2'H)-tetrone. erythrocytes was added to Krebs-Henseleit buffer containing glucose at a concentration of 90 or 500 mg/dl. in tissues may affect NADPH/NADP+ (nicotinamide These concentrations were chosen to mimic normal adenine dinucleotide phosphate/oxidized form) ratio and diabetic conditions in vitro, respectively. Krebs (7, 8), myo-inositol content, Na+ and K+–ATPase activity, Henseleit buffer consisted of 118 mM NaCl, 4.74 mM and oxidative stress (9). Therefore, we investigated in KCl, 1.21 mM MgSO4, 2.50 mM CaCl2, 0.92 mM the present study whether treatment with an aldose KH2PO4, and 25.0 mM NaHCO3 and was thoroughly reductase inhibitor, ranirestat, normalizes elevated equilibrated with a mixture of 95% O2 and 5% CO2. The sorbitol level in the sciatic nerve of STZ diabetic rats and assay mixture was first incubated at 37°C for 2 h and improves nerve function. Ranirestat {(3R)-2'-(4-bromo-2- then centrifuged at 1,050 × g for 10 min at 4°C. The fluorobenzyl)spiro[pyrrolidine-3,4'(1'H)-pyrrolo[1,2-a] sedimented erythrocytes were next suspended in 2 – 3 pyrazine]-1',2,3',5(2'H)-tetrone; Fig. 1} is a novel AR volumes of isotonic saline and centrifuged again at inhibitor developed in our laboratories and is now in 1,050 × g for 10 min to obtain washed packed cells. clinical development for the treatment of diabetic The washed packed cells were then added to 6% sensorimotor polyneuropathy (DSP). perchloric acid (PCA) and distilled water and centrifuged at 1,050 × g for 15 min at 4°C to remove any Materials and Methods formed proteins and other insoluble substances. An aliquot of the supernatant was neutralized with 2 M Preparation of AR KOH and left to stand for 10 min at 4°C to precipitate Rat lenses were collected under anesthesia with the formed KClO4. An aliquot of the new supernatant sodium pentobarbital (Abbott, Abbott Park, IL, USA), was then passed through a cation exchange resin column and lens AR was prepared and purified according to AG50W-X4 (9 × 25 mm) to remove fluorescent the method of Matsumoto et al. (10, 11). To compare materials. After discarding the initial aliquot of the the inhibitory effects of ranirestat on AR from different column eluate, the next 4-ml aliquot of the eluate was species, recombinant human aldose reductase (rhAR) collected and neutralized with 2 M K2CO3, and the was purchased from Wako Pure Chemical Industries, supernatant was used for the sorbitol assay. Ltd. (Osaka). Assuming that erythrocytes mean sorbitol levels after incubation in a medium containing glucose at 90 or Determination of AR activity 500 mg/dl were 0% or 100%, respectively, % inhibition AR activity was determined according to the method of sorbitol accumulation by ranirestat was calculated of Matsumoto et al. (10, 11). In brief, the decrease in based on the erythrocyte mean sorbitol level after absorbance of coenzyme NADPH due to its oxidation incubation in glucose at 500 mg/dl in the presence of into NADP was measured at 340 nm and taken as a ranirestat. Isolated rat sciatic nerves were incubated at measure of AR activity. Percent inhibition of AR activity 37°C for 4 h in Krebs-Henseleit buffer containing was calculated assuming AR activity in the absence of glucose at a final concentration of 90 or 500 mg/dl. the inhibitor as 100%. The concentration of AR inhibitor After incubation, the nerves were quickly washed with required to inhibit AR activity by 50% (IC50) was isotonic saline, blotted with filter paper, and weighed. estimated from the least-square regression line of the log Nerve tissue was then heated in distilled water (1 ml per concentration–response curve. Inhibition constant was 40 mg tissue) at 100°C for 2 min, added to 6% PCA calculated by fitting the apparent inhibition constant (1 ml of 6% PCA/10 mg tissue), and homogenized. (Ki') obtained at different glucose concentrations to the Next, the suspension was centrifuged at 1,050 × g for equation Ki' = Kii ([S] + Km) / [S], where [S] is the 15 min at 4°C, and the supernatant was neutralized with glucose concentration, Km is the glucose Michaelis 2M K2CO3 and used for the sorbitol assay. Normalized Polyol Pathway Metabolites and NCV 205

Sorbitol assay measured, and all STZ-treated rats were confirmed to A 1-ml aliquot of erythrocytes or sciatic nerves be hyperglycemic for the experimental period (plasma extract was mixed with 2 ml of 50 mM glycine buffer glucose level: 300 mg/dl or more). (pH 9.4) containing 2 mM NAD and 0.05 ml of Ranirestat was suspended in 0.5% tragacanth solution 25.6 U/ml sorbitol dehydrogenase and incubated at and administered orally to the diabetic rats once daily room temperature for 60 min. Fluorescence of the for 3 weeks (0.03 – 1.0 mg/kg per day). After the 3rd generated NADH was measured at an excitation week of ranirestat administration, MNCV in the sciatic wavelength of 366 nm and emission wavelength of nerve was determined in all groups. At 4 h after the final 452 nm using a spectrofluorometer (F-3000; Hitachi dose of ranirestat on Day 21 of treatment, blood samples High-Technologies Corporation, Tokyo). were collected from each rat. Rats’ sciatic nerves were Based on the difference in fluorescence intensity then removed under anesthesia (sodium pentobarbital; with or without SDH, sorbitol concentration in each Abbott, Abbott Park, IL, USA). The isolated sciatic extract was calculated from the calibration curve of nerves were promptly treated with distilled water (1 ml reference D-sorbitol. /40 mg wet weight) at 100°C for 2 min, homogenized in Erythrocyte sorbitol level was expressed in nmol/g 6% PCA (1 ml/10 mg wet weight), and centrifuged at hemoglobin (Hb). Hb content in the erythrocytes was 1,050 × g for 10 min at 4°C. Next, the supernatant was determined with Hemoglobin B-Test Wako (Wako) neutralized with 2 M KCO4 and cooled to precipitate based on the sodium lauryl sulfate–hemoglobin method. insoluble materials. After removing insoluble materials, Sciatic nerve sorbitol level was expressed in μmol/g ww the supernatant was used for measurement of sorbitol (wet weight). and fructose levels.

Inhibition of sorbitol accumulation Measurement of sorbitol and fructose in the sciatic Mean sorbitol levels in the erythrocytes and sciatic nerves nerves after incubation with low (90 mg/dl, low glucose Sorbitol and fructose levels in rat sciatic nerves were [LG]) or high (500 mg/dl, high glucose [HG]) concen- measured according to the method of Matsumoto (10, tration of glucose were taken as 0% or 100%, respec- 11) and expressed in μmol/g wet weight of tissue. To tively. Based on the measured sorbitol levels (TEST) in determine the percent inhibition of diabetes-induced the erythrocytes and sciatic nerves incubated in the high increase in sorbitol or fructose content at each dose (500 mg/dl) concentration of glucose in the presence of ranirestat, the difference in mean sorbitol content of ranirestat, % inhibition of sorbitol accumulation was between the non-diabetic control and diabetic control calculated by the following equation: groups was taken as the diabetes-induced increase in % Inhibition = (HG − TEST) / (HG − LG) × 100 sorbitol content and % inhibition of this increase was The inhibitory activity of ranirestat against sorbitol calculated based on sorbitol content in each of the accumulation was expressed as IC50, that is, the concen- ranirestat-treated diabetic groups using the following tration required for 50% inhibition. IC50 values were equation: % Inhibition = [1 − (AS − NC) / (DC − NC)] × estimated from the log concentration–inhibition curve. 100, where NC, DC, and AS indicate sorbitol contents (μmol/g wet weight) in the non-diabetic control group, In vivo experiment in STZ-treated diabetic rats diabetic control group, and the ranirestat-treated diabetic All experimental procedures were approved by the group, respectively. We calculated the % inhibition of Institutional Animal Care and Use Committee of diabetes-induced increase in fructose content in the Dainippon Pharma Co., Ltd. Male STD-Wistar rats same manner. aged about 12 weeks (body weight 260 – 290 g) were purchased from Japan SLC Co., Ltd. (Shizuoka) The rats Determination of MNCV were maintained in an animal room with controlled MNCV in the rat sciatic nerves was determined temperature (20°C–26°C) and relative humidity (40% – according to the method of Matsumoto et al. (10, 11). 70%) under a 12-h light/dark cycle (light on 0600 to Briefly, the muscle action potential of the rat haunches 1800 h) and allowed free access to pellet chow (CE-2; was recorded under inhalation anesthesia (fluothane; Clea Japan, Tokyo) and water. Diabetes was induced in Takeda Chemical Industries, Osaka). A constant body each rat by intravenous injection of STZ (30 mg/kg) into temperature (rectal temperature of 37.5°C – 38.5°C) was the tail vein. STZ (Sigma, St. Louis, MO, USA) was maintained by an animal body temperature controller dissolved in physiological saline containing 0.75 mM (ATB-1100; Nihon Kohden, Tokyo). The sciatic notch citrate buffer (pH 4.5) immediately before dosing. One of the right sciatic nerve was used as the proximal day after STZ injection, plasma glucose levels were stimulus point (S1) and the ankle region of the sciatic 206 T Matsumoto et al nerve was used as the distal stimulus point (S2). Using tively. These IC50 values were almost equal irrespective an electric stimulator (SEN-3301, Nihon Kohden), S1 of the source of the enzyme: either rat lens or human and S2 were stimulated by applying single rectangular recombinant AR. pulses (duration: 0.1 ms, voltage: supramaximal), and the evoked action potentials from the recording electrode Mode of ranirestat inhibition of recombinant human AR were traced on an oscilloscope (VC-11, Nihon Kohden) The mode of ranirestat inhibition of recombinant and recorded on an XY recorder (RW-21S; Rikadenki human AR (rhAR) was investigated by assessment of Kogyo, Tokyo). The proximal and distal latencies were enzyme reaction kinetics. Ranirestat and zenarestat, a measured as the time from stimulation of S1 and S2 to potent AR inhibitor used here as a positive control, the rise in action potential [t1 and t2 (ms)], respectively, uncompetitively inhibited rhAR. However, the affinity and the distance [d (mm)] between S1 and S2 was of ranirestat for rhAR was greater than that of zenarestat measured. MNCV was calculated by the following as indicated by an inhibition constant (kii = 0.38 nM) 5 formula: times lower than that of zenarestat (kii = 1.9 nM). MNCV (m/s) = d / (t1 − t2) Ranirestat inhibition of sorbitol accumulation in rat Statistical analyses erythrocytes and sciatic nerves Results are expressed as mean values ± S.E.M. After incubation in high concentration (500 mg/dl) of Statistical analyses were performed using SAS statistical glucose, sorbitol levels in rat erythrocytes were signifi- processing software (Statistical Analysis System; SAS cantly (P<0.01) higher than those after incubation in Institute, Tokyo). For comparison between the diabetic low concentration (90 mg/dl) of glucose (Tables 1 and control group and the non-diabetic control group, the 2). Ranirestat as well as the positive control zenarestat F-test to confirm homogeneity of variance and Student’s concentration-dependently inhibited sorbitol accumula- t-test were used. For multiple comparisons between the tion in rat erythrocytes and sciatic nerves incubated in diabetic control group and ranirestat-treated diabetic the high concentration (500 mg/dl) of glucose. The groups, homogeneity of variance was confirmed using potency of ranirestat inhibition of sorbitol accumulation Bartlett’s test followed by Dunnett’s multiple comparison was similar between rat erythrocytes and sciatic nerves test. Regression analysis was used for dose–response with IC50 values of 0.010 and 0.041 μM, respectively. relationships. In all cases, a P value less than 5% was On the other hand, the potency of zenarestat inhibition considered as statistically significant. of sorbitol accumulation in rat erythrocytes (IC50 = 0.044 nM) was more potent than that for inhibition of Results sorbitol accumulation in rat sciatic nerves (IC50 = 1.7 nM) (Table 2). Following repeated dosing, ranirestat Inhibitory effect of ranirestat on various ARs at a dose of 0.1 mg/kg per day essentially normalized Ranirestat inhibited rat lens AR and recombinant sciatic nerve sorbitol level in rats with elevated sorbitol human AR with IC50 values of 11 and 15 nM, respec- levels.

Table 1. Inhibitory effects of ranirestat on sorbitol accumulation in rat erythrocytes and sciatic nerves Sorbitol content Ranirestat Treatment (μM) rat erythrocytes rat sciatic nerves (nmol/g Hb) (μmol/g) Low glucose, 90 mg/dl — 29.3 ± 2.6 11.5 ± 4 High glucose, 500 mg/dl — 91.1 ± 2.9 626 ± 12 0.001 — 599 ± 18 0.004 75.8 ± 1.3 — High glucose + Ranirestat 0.01 — 479 ± 48 0.02 48.3 ± 2.1 — 0.1 16.1 ± 1.6 242 ± 19 1.0 — 29.6 ± 14

IC50 (μM) — 0.010 0.041

Rat erythrocytes and isolated sciatic nerves were incubated at 37°C in low (90 mg/dl) or high (500 mg/dl) concentration of glucose in the presence of ranirestat for 2 and 4 h, respectively, and then sorbitol levels were measured. Normalized Polyol Pathway Metabolites and NCV 207

Table 2. Inhibitory effects of zenarestat on sorbitol accumulation in rat erythrocytes and sciatic nerves Sorbitol content Zenarestat Treatment (μM) rat erythrocytes rat sciatic nerves (nmol/g Hb) (μmol/g)

Low glucose, 90 mg/dl — 39 ± 1.8 16 ± 12 High glucose, 500 mg/dl — 116 ± 6.7 1033 ± 81 0.01 105 ± 7.9 844 ± 118 High glucose + Zenarestat 0.1 42 ± 6.7 954 ± 62 130± 3.5 620 ± 89 10 — 172 ± 25

IC50 (μM) — 0.044 1.7 Rat erythrocytes and isolated sciatic nerves were incubated at 37°C in low (90 mg/dl) or high (500 mg/dl) concentration of glucose in the presence of zenarestat for 2 and 4 h, respectively, and then sorbitol levels were measured.

Table 3. Relationship between sorbitol content in rat erythrocytes and sciatic nerves and the period of hyperglycemia in STZ diabetic rats Days after STZ administration Origin Group 71228

Erythrocytes (nmol/g Hb) NDM 20.6 ± 1.5 33.8 ± 2.3 33.5 ± 2.4 DM 55.1 ± 8.2## 105.5 ± 5.4## 100.5 ± 5.4## Sciatic nerves (μmol/g) NDM 0.17 ± 0.02 0.17 ± 0.01 0.12 ± 0.02 DM 1.13 ± 0.06## 1.30 ± 0.07## 1.27 ± 0.07## Rats were made diabetic by administration of STZ. Sorbitol contents in rat erythrocytes and sciatic nerves were determined 7, 12, and 28 days after administration of vehicle or STZ solution. Sorbitol content in the non-diabetic control group (n = 10) and that in the diabetic control group (n = 11 – 12) are each expressed as the mean ± S.E.M. ##P<0.01: Significance of difference between the diabetic control group and the non-diabetic control group was examined by Student’s t-test.

Effect of hyperglycemia on sorbitol accumulation and 48.3 ± 0.72, and 51.4 ± 0.60 m/s, respectively. These on MNCV in rat sciatic nerves values in the sciatic nerves of the diabetic rats were Sorbitol accumulation and changes in MNCV in rat 42.9 ± 0.49, 38.5 ± 0.65, 40.5 ± 0.39, and 40.3 ± 0.30 sciatic nerve were sequentially measured. Sorbitol m/s, respectively (Fig. 2). contents in the erythrocytes of non diabetic and diabetic rats were measured on days 7, 12, and 28 after admin- Effect of ranirestat on MNCV in rats sciatic nerves istration of vehicle or STZ solution. Erythrocytes Six weeks after STZ injection, mean MNCV in the sorbitol levels in the non-diabetic rats on days 7, 12, and diabetic control group was 39.8 ± 0.5 m/s, which was 28 were similar (Table 3). These values in the diabetic about 81% of MNCV in the non-diabetic control group rats increased day by day, reaching a plateau level on (49.3 ± 0.6 m/s). Oral administration of ranirestat at day 12 (Table 3). Similarly, sorbitol contents in the 0.03, 0.1, 0.3, or 1 mg/kg per day for 3 weeks dose- sciatic nerves of non-diabetic and diabetic rats were dependently improved MNCV in the diabetic rats to measured on days 7, 12, and 28 after administration of 41.4 ± 0.4, 42.2 ± 0.5, 45.7 ± 0.5, and 47.1 ± 0.7 m/s, vehicle or STZ solution. Sorbitol levels in the sciatic respectively. Among all groups, MNCV in the rats nerves of the non-diabetic rats on days 7, 12, and 28 treated with ranirestat at 0.1, 0.3, or 1 mg/kg per day were similar (Table 3). These values in the diabetic rats was significantly higher than that in the diabetic control reached a plateau level on day 12 (Table 3). group (Fig. 3). P values ranged from 0.01 to 0.001. MNCV in the sciatic nerves of non diabetic and diabetic rats was measured on days 7, 14, 21, and 42 Inhibitory effects of ranirestat on sorbitol and fructose after administration of vehicle or STZ solution. MNCV accumulation in rat sciatic nerves values in the sciatic nerves of the non-diabetic rats on Sorbitol and fructose contents in the sciatic nerves in days 7, 14, 21, and 42 were 46.5 ± 0.64, 46.2 ± 0.43, the diabetic control group were about 6- and 4-fold, 208 T Matsumoto et al

Fig. 2. Time-dependent changes in MNCV after administration of STZ. Rats were made diabetic by administration of STZ. MNCV in the sciatic nerve was determined 1, 2, 3, and 6 weeks after STZ administration. Open circles show MNCV in the non-diabetic rat, and closed circles show MNCV values in the diabetic rats. ##P<0.01: significance of difference between the diabetic control group (n = 12) and the non-diabetic control group (n = 12) was examined by Student’s t-test.

Fig. 3. Effect of ranirestat on decreased MNCV in the sciatic nerves of diabetic rats. Rats were made diabetic by administration of STZ. Ranirestat was administered orally once a day for 3 weeks starting Fig. 4. Interrelationship between MNCV and sorbitol (A) and 3 weeks after induction of diabetes. MNCV in the non-diabetic = = fructose (B) contents in the sciatic nerves. Figures in parentheses control group (n 10), diabetic control group (n 11 – 12), and indicate the dose of ranirestat (mg/kg per day). Each value represents ranirestat-treated diabetic groups (n = 10 – 12) is expressed as the ± = ± < < the mean S.E.M. (n 10 – 12). P-values indicate linearity by the mean S.E.M. **P 0.01, ***P 0.001: significance of difference Lack of Fit test between MNCV and sorbitol or fructose content in between ranirestat-treated diabetic groups and the diabetic control the sciatic nerve. group was examined by Dunnett’s multiple comparison test; ###P<0.001: significance of difference between the diabetic control group and the non-diabetic control group was examined by Student’s t-test. respectively, higher than those in the non-diabetic contents were decreased by 101% and 89%, respectively control group. Treatment with ranirestat at 0.03, 0.1, (Fig. 4, Table 4). Finally, in the diabetic rats that 0.3, and 1 mg/kg per day dose-dependently decreased received ranirestat at 0.3 or 1 mg/kg per day, the doses sorbitol and fructose levels in the diabetic rats (Fig. 4, that produced an even greater improvement in MNCV Table 4). In the diabetic rats that received ranirestat at than the dose of 0.1 mg/kg per day, ranirestat completely 0.03 mg/kg per day, sorbitol and fructose contents were abolished diabetes-induced increases in sorbitol and decreased by 35% and 34%, respectively. In the diabetic fructose contents (Fig. 4, Table 4). Sorbitol and fructose rats treated with ranirestat at 0.1 mg/kg per day, the tissue levels in the diabetic rats after treatment with dose at which ranirestat produced a statistically signifi- ranirestat (0.3 or 1 mg/kg per day) were lower than cant improvement in MNCV, sorbitol and fructose those in the non-diabetic control group (Fig. 4, Table 4). Normalized Polyol Pathway Metabolites and NCV 209

Table 4. Effects of ranirestat on sorbitol and fructose levels in the sciatic nerves of diabetic rats Sorbitol Fructose Dose Experimental group n (mg/kg per day) (μmol/g wet weight) (μmol/g wet weight) [inhibition%] [inhibition%]

Diabetic control — 11 1.04 ± 0.07 [ — ] 5.72 ± 0.35 [ — ] Ranirestat-treated 0.03 10 0.74 ± 0.04*** [34 ± 5] 4.19 ± 0.13*** [35 ± 3] diabetic group 0.1 12 0.16 ± 0.02*** [101 ± 2] 1.87 ± 0.15*** [89 ± 3] 0.3 11 0.06 ± 0.01*** [113 ± 1] 1.02 ± 0.05*** [109 ± 1] 1120.01± 0.00*** [118 ± 0] 0.39 ± 0.04*** [123 ± 1]

Non-diabetic control — 10 0.17 ± 0.01### [—] 1.39± 0.06### [—]

Ranirestat was administered orally for 21 days. Sorbitol and fructose levels in the sciatic nerve were measured by fluorometric enzyme assay 4 h after the final dose of ranirestat. Values indicate the mean ± S.E.M. Significance of difference from the diabetic control group: ***P<0.001 (Dunnett’s multiple comparison test). ###P<0.001 (Student’s t-test).

Discussion and polyuria. In addition, on week 3 after STZ injection, MNCV in the diabetic control group decreased, and Earlier AR inhibitors such as sorbinil (13), tolrestat sorbitol and fructose in the rat sciatic nerves increased. (14), and ponalrestat (15) are known to inhibit accumu- Ranirestat given at doses of 0.1 – 1 mg/kg per day for lation of polyol metabolites in tissues. In Japan, 3 weeks improved STZ-induced decrease in MNCV and epalrestat, an AR inhibitor, is the only licensed drug for inhibited sorbitol accumulation in the sciatic nerve. diabetic sensorimotor polyneuropathy (DSP), although These findings indicate a negative correlation between more recent AR inhibitors such as zenarestat (16, 17) MNCV and sorbitol and fructose contents in the sciatic and fidarestat (18) have been reported to have beneficial nerve (P<0.01; r2 = 0.99 and 0.97, respectively; Fig. 4). clinical effects on DSP. Ranirestat is another recent AR It is therefore clear that ranirestat improvement of sciatic inhibitor that has been shown to have beneficial effects nerve MNCV in STZ-treated diabetic rats is closely on DSP after 12 weeks administration (19). A previous related to its inhibition of sorbitol and fructose accumu- study has shown that improvement of decreased MNCV lation. Furthermore, the results of the present study in STZ diabetic rats can be achieved by treatment with indicate that improvement of decreased MNCV in STZ fidarestat at a dose over 1 mg/kg per day, zenarestat at a diabetic rats requires almost complete normalization of dose of 50 mg/kg per day, or epalrestat at a dose of sorbitol and fructose levels in the sciatic nerve (19). 50 mg/kg per day (20). Meanwhile, the effective dose of Ranirestat exhibited potent inhibitory activity against ranirestat is over 0.1 mg/kg per day. We investigated AR from rat and recombinant human AR with IC50 whether normalization of elevated sorbitol and fructose values in the order of 10 nM. We believe that AR levels in the sciatic nerves of STZ diabetic rats following inhibitors need good tissue penetration (membrane treatment with ranirestat improves nerve function, as permeability) to achieve normalization of excess polyol measured by MNCV. pathway metabolites flux, thereby reversing slowed Sorbitol content in rat erythrocytes and sciatic nerves MNCV. In this study we found that ranirestat and increased time-dependently after administration of STZ, zenarestat inhibit human recombinant AR with similar reaching a plateau concentration 12 days after STZ inhibition constants. However, in a clinical study, the administration (Table 3). On the other hand, MNCV in dose of zenarestat required to improve decreased MNCV the sciatic nerves decreased after STZ administration, was higher than that of ranirestat (21, 22). Although reaching minimal values 2 weeks after STZ treatment many AR inhibitors have been described, only few (Fig. 2), and remained low for up to 2 weeks thereafter. have been shown to be effective in vivo. One possible These findings indicate that increase in polyol metabo- explanation is poor tissue penetration (23). When lites flux decreases MNCV in the sciatic nerve, support- isolated nerve tissues and erythrocytes were incubated ing the polyol pathway theory. Previous reports have in high concentration of glucose, sorbitol accumulated also shown the link between activated AR and diabetic in these tissues as it does in hyperglycemic animals in complications (1 – 3). vivo. Ranirestat inhibitory effect on sorbitol accumula- In the STZ-treated groups (diabetic control group, tion in rat erythrocytes was equipotent to that in rat ranirestat-treated diabetic groups), all rats showed sciatic nerves. However, the inhibitory effect of diabetic symptoms, including polyphagia, polydipsia, zenarestat on sorbitol accumulation was 40 times more 210 T Matsumoto et al potent in rat erythrocytes than in rat sciatic nerves. Diabetologia. 1996;39:172–182. Furthermore, ranirestat showed no difference in inhibi- 8 Cappiello M, Vilardo PG, Micheli V, Jacomelli G, Banditelli tion of AR from different species. S, Leverenz V, et al. Thiol disulfide exchange modulates the activity of aldose reductase in intact bovine lens as a response The results of this study in a rat model are consistent to oxidative stress. Exp Eye Res. 2000;70:795–803. with those of a clinical study in which nerve sorbitol 9 Oates PJ. Aldose reductase, still a compelling target for diabetic levels and sensory and motor NCV were measured neuropathy. Curr Drug Targets. 2008;9:14–36. (19, 22). In that study, mean nerve sorbitol level in the 10 Matsumoto T, Ono Y, Kurono M, Kuromiya A, Nakamura K, ranirestat 5 mg/day group was reduced by 65% Bril V. Ranirestat (AS-3201), a potent aldose reductase inhibitor, compared to the placebo group. In addition, in patients reduces sorbitol levels and improves motor nerve conduction velocity in streptozotocin-diabetic rats. J Pharmacol Sci. 2008; treated with ranirestat at 20 mg/day, mean nerve sorbitol 107:231–237. level was reduced by 84% reduction compared to the 11 Matsumoto T, Ono Y, Kuromiya A, Toyosawa K, Ueda Y, Bril placebo group. In patients who received ranirestat at V. Long-term treatment with ranirestat (AS-3201), a potent 5mg/day there were no statistically significant im- aldose reductase inhibitor, suppresses diabetic neuropathy and provements in motor or sensory NCV at 12 weeks. cataract formation in rats. J Pharmacol Sci. 2008;107:340–348. However, in patients who received ranirestat at 12 Kurono M, Fujiwara I, Yoshida K. Stereospecific interaction of a novel spirosuccinimide type aldose reductase inhibitor, AS- 20 mg/day there were statistically significant improve- 3201, with aldose reductase. Biochemistry. 2001;27:8216–8226. ments in sensory NCV in the right and left sural nerves 13 Sima AA, Bril V, Nathaniel V, McEwen TA, Brown MB, and proximal median sensory nerve, although not in Lattimer SA, et al. Regeneration and repair of myelinated fibers median or peroneal motor nerves, at 12 weeks. Velocity in sural-nerve biopsy specimens from patients with diabetic improvement in both the sural and proximal median neuropathy treated with sorbinil. N Engl J Med. 1988;319:548– sensory nerves exceeded 1 m/s, which is clinically 555. 14 Wrobel J, Millen J, Sredy J, Dietrich A, Kelly JM, Gorham BJ, significant (19, 22). To show the same efficacy, et al. Orally active aldose reductase inhibitors derived from zenarestat need to be used at much higher dose than bioisosteric substitutions on tolrestat. J Med Chem. 1989;32: ranirestat (15, 19, 22). 2493–2500. In conclusion, we have shown in this study that 15 Calcutt NA, McMurray HF, Moorhouse DF, Bache M, ranirestat has the desirable properties as an AR inhibitor, Parthasarathy S, Powell HC, et al. Inhibition of macrophage chemotaxis and peripheral nerve regeneration in normal and including potent inhibition of AR from different species hyperglycemic rats by the aldose reductase inhibitor Tolrestat. and dose-dependent improvement of slowed MNCV via Exp Neurol. 1994;128:226–232. normalization of sorbitol and fructose accumulation in 16 Greene DA, Arezzo JC, Brown MB. Effect of aldose reductase the sciatic nerve of STZ-treated diabetic rats. inhibition on nerve conduction and morphometry in diabetic neuropathy. Zenarestat Study Group. Neurology. 1999;53:580– 591. References 17 Greene DA, Stevens MJ, Feldman EL. Diabetic neuropathy: scope of the syndrome. Am J Med. 1999;107:2S–8S. 1 Obrosova IG, Ilnytska O, Lyzogubov VV, Pavlov IA, Mashtalir 18 Hotta N, Toyota T, Matsuoka K, Shigeta Y, Kikkawa R, N, Nadler JL, et al. High-fat diet induced neuropathy of pre- Kaneko T, et al. Clinical efficacy of fidarestat, a novel aldose diabetes and obesity: effects of “healthy” diet and aldose reductase inhibitor, for diabetic peripheral neuropathy: a 52- reductase inhibition. Diabetes. 2007;56:2598–2608. week multicenter placebo-controlled double-blind parallel group 2 Yagihashi S. Diabetic neuropathy. Nippon Rinsho. 2006;28 study. Diabetes Care. 2001;24:1776–1782. Suppl 3:155–160. 19 Bril V, Buchanan RA. Aldose reductase inhibition by AS-3201 3 Baba M, Kimura K, Suda T, Yagihashi S. Aomori Diabetic in sural nerve from patients with diabetic sensorimotor Study (ADNS) Group. Three-year inhibition of aldose reductase polyneuropathy. Diabetes Care. 2004;27:2369–2375. on development of symptomatic neuropathy in diabetic patients. 20 Pfeifer MA, Schumer MP, Gelber DA. Aldose reductase J Peripher Nerv Syst. 2006;11:176–178. inhibitors: the end of an era or the need for different trial 4 Lorenzi M. The polyol pathway as a mechanism for diabetic designs? Diabetes. 1997;46 Suppl 2:S82–S89. retinopathy: attractive, elusive, and resilient. Exp Diabetes Res. 21 Perkins BA, Bril V. Emerging therapies for diabetic neuropathy: 2007;2007:61038. a clinical overview. Curr Diabetes Rev. 2005;1:271–280. 5 Takamura Y, Tomomatsu T, Kubo E, Tsuzuki S, Akagi Y. Role 22 Bril V, Buchanan RA. Long-term effects of ranirestat (AS-3201) of the polyol pathway in high glucose-induced apoptosis of on peripheral nerve function in patients with diabetic sensori- retinal pericytes and proliferation of endothelial cells. Invest motor polyneuropathy. Diabetes Care. 2006;29:68–72. Ophthalmol Vis Sci. 2008;49:3216–3223. 23 Mizuno K, Kato N, Makino M, Suzuki T, Shindo M. Continuous 6 Schrijvers BF, De Vriese AS. Novel insights in the treatment of inhibition of excessive polyol pathway flux in peripheral nerves diabetic nephropathy. Acta Clin Belg. 2007;62:278–290. by aldose reductase inhibitor fidarestat leads to improvement of 7 Cameron NE, Cotter MA, Hohman TC. Interactions between diabetic neuropathy. J Diabetes Complications. 1999;13:141– essential fatty acid, prostanoid, polyol pathway and nitric oxide 150. mechanisms in the neurovascular deficit of diabetic rats.