Male Wistar rats, weighing 120–140 g, received intra-peritoneal injections of STZ (65 mg/kg in 100 mM sodium citrate buffer, pH 4.5). The use of these animas had been approved by Medical College of Georgia's Institutional Animal Care and Use Committee. Beginning 1 week after STZ injection, blood samples were obtained at each time point from the same rats following a 15-hour starvation period by pricking the tip of the rat tail and bleeding a drop (about 0.06 ml) of blood. The whole blood was obtained from each rat by heart puncture at the endpoint of the experiment. The blood-glucose levels were measured using a strip-operated blood glucose meter (Accu-Chek Advantage Blood Glucose Meter, Roche Diagnostics, Indianapolis, IN). Two animals were housed per cage in a controlled environment. All animals were allowed free access to water and food, except during the starvation periods just prior to blood glucose assessments.

All rice and control diets were manufactured as powdered feed by Harlan Teklad (Madison, WI). The control diet (AIN93G) was composed of cornstarch [39.7% (w/w)], α-cornstarch [13.2% (w/w)], casein [20.0% (w/w)], L-cysteine [0.3% (w/w)], sucrose [10% (w/w)], soybean oil [7.0% (w/w)], cellulose powder [5.0% (w/w)], mineral mix [3.5% (w/w)], vitamin mix [1.0% (w/w)], choline bicitrates [0.25% (w/w)], and_butylhydroquinone [0.0014% (w/w)]. Pre-germinated brown (PR), brown (BR), or white (WR) rice diets were produced by replacing cornstarch and α-cornstarch with pre-germinated brown, brown, or white rice, respectively (Table 1).

Male Wistar rats (n = 40) received intra-peritoneal injections of STZ and were maintained on AIN93G for 2 weeks. Only those animals whose blood glucose levels were > 400 mg/dl were considered diabetic. The diabetic rats (n = 31) were divided into four experimental groups: the DWR (n = 9), DBR(n = 9), and DPR (n = 9) groups received WR-, BR-, and PR-enriched diets, respectively, and DC group (n = 4) were fed AIN93G. Non-diabetic rats (n = 22) were also divided into four feeding groups: the WR (n = 6), BR (n = 6), and PR (n = 6)groups received WR-, BR-, and PR-enriched diets, respectively, while the C control group (n = 4) received AIN93G.

Because rats became moribund with severe diabetic deterioration initiating 3 weeks after STZ injection, the experimental endpoint was designated 3 weeks after initiation of the rice-enriched diet regimens (i.e., 5 weeks after STZ-injection). Food consumption was measured by weighing the feeding containers daily. Three weeks after initiation of the rice-enriched diets, tail-nerve conduction electrophysiological studies were performed, and then the rats were sacrificed. Whole serum samples were obtained for biochemical analyses as described below. The sciatic nerves were removed; the right sciatic nerve was utilized for morphometric studies while the left sciatic nerve was assayed for Na⁺/K⁺-ATPase activities.

Nerve-conduction velocities (NCVs, m/sec) were assessed in the rat tail nerve using a Nicolet VikingQuest EMG machine (Neurocaregroup, Madison, WI) according to the modified procedure of Anderson et al. 24. In brief, the nerves were stimulated using external digital ring electrodes with twisted wires (Medtronic Functional Diagnostics, Skovlunde, Denmark) instead of needle electrodes 25. The electrodes were placed in segments proximal (5 cm) and distal (2 cm) from the recording position (7 cm far from the joint of rat tail). Blood collection was from the tip of the tail that was distant and non-interactive with NCV measurement. NCV was evaluated from four different waves generated from electrical stimulations; each wave showed a reproducible pattern and the same amplitude level as the stimulator-voltage was increased. During each measurement, a constant surface temperature of the rat tail was maintained (34–35°C). NCV values represent an average value from 4 nerve conduction wave measurements per animal.

The right sciatic nerves were carefully dissected from their origin (5 mm distal to the gluteus maximus) through the distal branch point at the peroneal and tibial nerves in order to avoid stretching. These nerve sections were placed overnight in a fixative solution containing paraformaldehyde. After washing 3 times in cacodylate buffer (pH 7.2), the nerves were cross-sectioned into two pieces and embedded in epoxy resin (Poly/Bed 812, Polysciences Inc., Warrington, PA). Fascicle cross sections were stained with 1% toluidine blue and observed under an Axiophot photomicroscope equipped with an Axiocam (Carl Zeiss, Jena, Germany). Images were stored and analyzed using AxioVision. The total number of myelinated fibers in each fascicle was assessed by visual counting. In myelinated fibers, both the axonal and total fiber diameters were measured, based on the average of the major and minor diameters. Diameters of the myelinated fibers and axons, as well as the G ratio (axonal diameter/total fiber diameter, a measure of the degree of myelination), were obtained, and histograms of their distributions and the G ratios were constructed as percentages of size frequency.

The total lipid fractions from PR and BR bran (TLp and TLb, respectively) were prepared by extracting 5 g of bran twice with 30 ml and 20 ml of chloroform-methanol (1:1 (v/v) and 2:1 (v/v), respectively) using the procedure of Folch et al. 26. Lipid fractions from these two extractions were combined, evaporated, and used in in vitro enzyme assays.

Lipoprotein fractions were prepared from rat serum using previously published procedures 27. Briefly, freshly collected serum obtained from normal male Wistar rats (n = 4) was pooled and adjusted to a density of 1.3 g/ml with solid potassium bromide. A discontinuous density gradient was formed in a centrifuge tube by layering normal saline (3.5 ml, 1.006 g/ml) over the adjusted serum (1.5 ml, 1.3 g/ml). Lipoproteins were separated by ultracentrifugation (370,000 × g, 45 min, 4°C) in a TV865 rotor. Three major lipoprotein fractions (VLDL, LDL, and HDL) were collected and dialyzed overnight against PBS at 4°C.

Crude sciatic nerve membranes were prepared from each rat of the diet experiment using a previously published procedure 28. Briefly, the left sciatic nerve from rats fed WR-, BR-, or PR-enriched diets, or AIN93G, or from untreated male Wistar rats were homogenized in Polytron homogenizer in a cold iso-osmotic solution containing 250 mM sucrose, 10 mM HEPES-Tris buffer (pH 7.6), EDTA 2 mM, and PSMF 1 mM. The homogenates were centrifuged for 10 min at 1,500 × g at 4°C; the supernatants were collected and then centrifuged at 191,000 × g for 45 min at 4°C. After decanting the supernatant, the pellets were resuspended in 100 μl of 250 mM sucrose in 10 mM HEPES-Tris buffer (pH 7.6).

Na⁺/K⁺-ATPase activities were assayed as previously described 28. Briefly, a 0.2 ml assay medium containing 10 mM MgCl₂, 20 mM HEPES-Tris (pH 7.0), 120 mM NaCl, 30 mM KCl, 0.5 mg/ml of crude membrane preparations, and 25 mM [γ-³²P]ATP (100,000 cpm) were incubated at 37°C for 15 min, followed by the addition of 0.1 ml of activated carbon (0.1 mg/ml). Parallel assays were also performed in which 1 mM ouabain was added to the assay medium. Following the addition of activated carbon, the samples were centrifuged at 1,500 × g for 15 min at 4°C, the supernatants were collected, and the cpm of the inorganic ³²P radioactivity in the fractions was measured using a Beckman scintillation counter. Ouabain-sensitive Na⁺/K⁺ ATPase activity was calculated by subtracting the ouabain-sensitive activity from the Na⁺/K⁺-enhanced activity.

PON1 activity was measured in 96-well ELISA plates; each well contained 20 μl of rat serum (diluted 1:4 in Tris-HCl buffer) and 200 μl of substrate solution containing 5.5 mM paraoxoethyl and 2 mM CaCl₂ in 100 mM Tris-HCl buffer (pH 8.0). Generation of p-nitrophenol was monitored as a function of time at 25°C by recording the absorbance at 412 nm using a spectrophotometer. Enzyme activities were calculated using the molar absorption coefficient of p-nitrophenol (ε₄₁₂^M, 169,000 M^-1 cm^-1).

HTase activities in rat serum or in HDL fractions from normal male Wistar rats were measured with a commercial assay kit (Alfresa Auto HTLase; Alfresa Pharma Corp., Osaka, Japan) 23 that utilizes γ-thiobutyrolactone as the substrate. HTase hydrolyzes the lactone ring, generating free thiol, which reacts with 5, 5'-dithio-bis (2-nitrobenzoic acid). Generation of 5-thio-2-nitrobenzoic acid was monitored as a function of time by recording the absorbance at 450 nm using a spectrophotometer.

In vitro homocysteinylation of LDL was done according to the procedure of Vignini et al. 16. Briefly, an aliquot of LDL (100 μg) was resuspended in 10 mM PBS (pH 8.2) and incubated with homocysteine-thiolactone (100 μmol/L, Sigma, St. Louis, MO) and the indicated amount (0.1 to 1.0 μg) of a total lipid fraction (TLb or TLp) with gentle stirring at 37°C for 2 h. After incubation, the mixture was passed through a Bio-gel P-2 column equilibrated with 10 mM PBS (pH 8.2) to remove any unreacted HT.

The creatinine concentrations in rat serum samples were assayed by Jaffe's alkaline-picrate method as described by Adeoye et al. 29. Proteins were precipitated from 1 ml of serum by the addition of 1 ml 10% sodium tungstate in 1 ml of 0.67 M sulfuric acid, and the supernatants were collected;1.5 ml of the supernatants were mixed with 0.5 ml of 2.5 M NaOH and 0.5 ml of 0.04 M picric acid, and then incubated for 5 min at 25°C. The absorbance at 520 nm was measured in each sample against a blank that consisted of distilled water with the reagent as described above using a spectrophotometer.

Statistical analyses were performed for animal data using the GraphPad Prism 2.01 software package (GraphPad, San Diego, CA). Differences among diet groups with non-diabetic treatment or with diabetic treatment were analyzed using the one-way ANOVA (Table 2) test. Normalization of data distribution was given by Prism 2.01 before the comparison test. If the data were found to be parametric, they were analyzed by Tukey's multiple comparison test, followed by determination of differences vs. the control group after obtaining a significant ANOVA by the Dunnet's test. If the data were non-parametric, they were analyzed by the Kruskal-Wallis test of one way analysis of variance on rank, followed by Dunn's multiple comparison test. Differences among the diet treatment groups were considered significant if p < 0.01. For comparisons of size-frequency distributions, the difference of skewness was evaluated by the chi-square test for trend in the Prism 2.01 to evaluate the statistical difference of skewness. A value of p < 0.01 was considered a statistically significant difference.

In correlation analyses, Pearson correlation coefficients were calculated to identify potential associations of the HTase, PON1, NCV, and Na⁺/K⁺-ATPase activities. A value of p < 0.05 was considered a statistically significant difference. Data were presented as mean ± standard error of the mean (SEM).

The total mass of food consumed during the 3-week experimental diet period for each non-diabetic rat group was as follows: WR, 590 ± 31 g; BR, 611 ± 42 g; and PR, 631 ± 64 g. The total mass of food consumed during the same period for each diabetic rat group was as follows: DWR, 641 ± 58 g; DBR, 658 ± 30 g; and DPR 631 ± 53 g. The consumption rates were not significantly different among the normal and diabetic rats fed WR-, BR-, and PR-enriched diets (assessed by the Kruskal-Wallis test followed by Dunn's test).

The body-weight gain in the WR, BR, and PR groups was smooth, and no differences were observed among the three groups. Modest body-weight gain due to diabetes was observed in the DWR, DBR, and DPR groups, and considered to be a well-established phenomenon in insulin-deficiency caused by STZ treatment in the rat 30. There were also no statistically significant differences in body-weight gain between rats fed a rice-enriched diet and those fed the control AIN93G diet. The DPR group, however, had statistically significant body-weight gain as compared with the DC group (p < 0.01, Dunnett's test) but not with the DBR group, as shown in Table 2.

All the blood-glucose levels of the non-diabetic AIN93G- and rice-diet groups (C, PR, BR and WR) were within the normal range during the experimental period (Table 2). In contrast, the blood-glucose levels of the diabetic rats fed the AIN93G, WR-, BR-, or PR-enriched diets (i.e., DC, DWR, DBR, and DPR groups, respectively) were elevated. During the initial week of the rice-enriched diet regimen, all diabetic groups demonstrated high blood-glucose levels (Table 2); after 3 weeks of the rice-enriched diets, however, the blood glucose levels of the DPR group were significantly lower than those of the DWR and DBR groups (p < 0.001, Tukey's test), as shown in Table 2.

A positive correlation was established between values for NCV and Na⁺/K⁺-ATPase activities in non-diabetic and STZ-diabetic rats (Fig. 1). The data also include rats fed by a regular diet (AIN93G). This result demonstrates that decreased sciatic-nerve Na⁺/K⁺-ATPase activities are likely to result in electrophysiological and neurophysiological abnormalities. Comparison of the effects of the diet intakes on NCV and Na⁺/K⁺-ATPase activity is shown in Table 2, expressed as mean ± SEM. Diabetic rats fed the control diet (AIN93G) showed significantly decreased NCVs when compared with those in the non-diabetic rats. NCVs in the DPR group, on the other hand, were significantly faster than those observed in the DC, DWR or DBR group (p < 0.001, Tukey's test). Na⁺/K⁺-ATPase activities in the DPR group were also significantly higher than those seen in the DBR, DWR, and DC groups (p < 0.001, Tukey's test), as shown in Table 2.

Size-frequency distribution histograms for myelinated fiber diameter, myelinated axon diameter, and the G ratio in rat sciatic nerves were evaluated by the chi-square test for trend in the Prism 2.01, and compared among three groups: a non-diabetic group fed the AIN93G diet (C group); a diabetic group fed the AIN93G diet (DC group); and a diabetic group fed the PR diet (DPR group). There was no statistically significant difference between myelinated fiber distributions; one difference between axonal distributions for the C and DC groups, however, indicated that the STZ rat model of diabetes primarily affected axons. Since the G ratio is a reflection of the diameters of axons and myelinated fibers, these observations indicate that PR intake could prevent or correct myelinated axonal injury in the STZ model of diabetic neuropathy (data not shown).

Serum activities for creatinine or PON1 and HTase in the diabetic and non-diabetic animals are shown in Table 2 as means ± SEM. Given that the non-diabetic groups had similar weight gains and diet consumptions, we cannot attribute the changes in serum homocysteine concentrations to differences in vitamin or amino-acid intake in the rice diets. The decreased homocysteine levels in the diabetic rats indicate that increased creatinine levels may reflect diabetic complications such as nephropathy. The serum homocysteine level and PON1 activity in the DPR group sera were affected by STZ-induced diabetes, which was not the case in the DC, DWR and DBR groups (p < 0.001, Tukey's test) (Table 2).

To determine whether the PR-diet promoted amelioration of the nervous function by a specific active component of PR due to an HT-related mechanism, a correlation analysis was performed between activities of Na⁺/K⁺-ATPase and HTase (or PON1) for each group of Table 2. A positive correlation was established between activities of serum HTase and PON1 in all the diet-treated and the control groups (r = 0.819 to 0.992). On the other hand, a positive correlation between Na⁺/K⁺-ATPase and HTase activities was established only in the DPR (r = 0.913, p < 0.001) and PR groups (r = 0.878, p < 0.05). Other diet-treated groups showed no correlation between the two enzyme activities. Although the DBR group showed an improvement in diabetic neuropathy, there was no correlation between the two enzyme activities in DBR group (r = -0.217, non-significant) and BR group (r = -0.142, non-significant).

HT causes LDL deterioration, resulting in the development of diabetic complications as a vascular risk factor by HT-modified LDL. HTase is present in serum HDL and has an important role in the antioxidation of LDL. We hypothesized that the efficacy of a PR diet on diabetic neuropathy might be dependent on an HT-related mechanism. To test this possibility, an additional effect of rice bran lipid extracted fractions (TLp or TLb) on Na⁺/K⁺-ATPase activity was tested with HT-treated LDL. Changes in the Na⁺/K⁺-ATPase activity of normal rat sciatic nerve was studied by incubation of HT-treated LDL with HT and TLp or TLb (Fig. 2A). A significant decrease in Na⁺/K⁺-ATPase activity occurred in an in vitro system incubated with HT-modified LDL as compared with intact LDL. Incubation with TLp (0.1 μg to 10 μg) attenuated the inhibitory activity of HT-modified LDL on Na⁺/K⁺-ATPase activity (columns 7 to 9, Fig. 2A), whereas TLb had no effect on the inhibition of Na⁺/K⁺-ATPase activity by HT-modified LDL (columns 4 to 6, Fig. 2A). Neither TLp nor TLb alone produced any modification of Na⁺/K⁺-ATPase activity (data not shown). The effect of TLp suggests the presence of an inhibitory factor causing normalization of Na⁺/K⁺-ATPase activity that has been blocked by HT-modified LDL.

In the previous experiment, we showed that Na⁺/K⁺-ATPase activity was controlled by HT-modified LDL. The efficacy of PBR on diabetic neuropathy may be related to the HTase activity of HDL on an HT-related mechanism. To verify this possibility, the effect of rice bran lipid extracted fractions (TLp or TLb) on HTase activity was tested. HDL prepared from normal male Wistar rat serum was used as an enzyme source of HTase as described in the Methods section. HTase activity in the presence of TLp or TLb (0.1 to 5.0 μg) was compared. TLp treatment showed a statistically significant difference HTase activity as compared with treatment with TLb (Fig. 2B). TLp had a dose-dependent stimulatory effect (0.1 μg to 1.0 μg) on HTase activity (columns 6 to 8 in Fig. 2B), and more than 5.0 μg produced saturation of this effect. In contrast, TLb had no stimulating effect on HTase activity in the same dosage range (columns 2 to 4, Fig. 2B).

On the other hand, PON1 activity in HDL was assayed using paraoxyethyl as a substrate, and there was no change upon the addition of TLp in the above dosage range (data not shown). This finding suggests that other PON isozymes may be responsible for TLp-induced HTase activation. The HTase/PON1 activity ratio calculated for each HDL sample showed a dose-dependent elevation by TLp (Fig. 3). One- and 5 μg of TLp gave a statistically significant ratio increase as compared with lower amounts of TLp. This suggests that TLp has no effect on the activities of PON1, whereas it does have an effect on an isozyme such as PON3.