A total of 50 male specific-pathogen free (SPF) Sprague-Dawley rats (weighing, 150-170 g) at the age of 10 weeks, were purchased from Beijing Vital River Laboratory Animal Technology. The justification of male rats was based on the relatively longer distance between the urethral orifice and genitals compared with female rats. The animals were housed in individual cages under a controlled environment (12:12 h day/night cycle; temperature, 22±1°C) with free access to food and water. All experimental procedures followed the Guide for the Care and Use of Laboratory Animals. All protocols for animal treatment were approved by the Animal Ethics Committee of Hebei Medical University. The animals were allowed to acclimatize to the laboratory conditions for 2 weeks prior to the experiment.

Troxerutin was purchased from Shanghai Jingchun Biotechnology Co., Ltd. α-lipoic acid (LA) was purchased from Stada Arzneimittel AG.

The rat model of diabetes was established by a single intraperitoneal injection of STZ (Sigma-Aldrich; Merck) at doses of 60 mg/kg for 12-h fasted rats at the age of 12 weeks. STZ is a glucosamine-nitrosourea compound and alkylating agent that is particularly toxic to islet β-cells. STZ is widely used to induce type 1 diabetes in mice or rats by using a large dose on an empty stomach. A single intravenous injection of STZ into fasted rats at dose of 50-75 mg/kg body weight will cause the necrosis of β-cells followed by β-cell loss and the atrophy of islets (25). It will then lead to a decrease in insulin levels, hyperglycemia and an increase in serum free-fatty acids. STZ was dissolved in 0.1 mmol/l sodium citrate buffer (pH 4.4). The rats in the normal control group received an intraperitoneal injection of sodium citrate buffer. The blood glucose levels of the caudal vein were measured after 72 h using a portable glucose meter (ACCU-CHEK Performa, Roche Diagnostics). The rat model of diabetes was successfully established once the blood glucose levels exceeded 16.7 mmol/l for 2 consecutive days. Body weight and fasting blood glucose (FBG) levels of the rats were recorded once each week. At 4 weeks after the STZ injection, 40 rats with diabetic cognitive dysfunction and 10 normal control rats were well prepared and evaluated.

The diabetic rats were randomly divided into 4 groups (n=10 in each group) as follows: i) The diabetes control group (DC group), in which diabetic rats were left without any treatment; ii) diabetes and saline-treated group (DN group), in which diabetic rats were intraperitoneally injected with saline; iii) diabetes with LA treatment group (DL group), in which diabetic rats were injected intraperitoneally with LA at doses of 60 mg/kg/day; iv) diabetes with troxerutin treatment group (DT group), in which diabetic rats were injected intraperitoneally with troxerutin at doses of 150 mg/kg/day. Moreover, the normal control rats were injected intraperitoneally with physical saline at the same volume as the NC group (n=10). The treatment was administered once a day for 6 weeks.

Spatial learning and memory in each group were analyzed using the Morris water maze for 5 consecutive days following the 6-week intervention. The Morris water maze test is a convenient and classical method with which to assess cognitive function in rodents. The apparatus was composed of a circular water pool (150 cm in diameter and 60 cm high), containing various prominent visual cues, with opaque water kept at 24-26°C. A transparent platform was hidden 1 cm below the water surface in one quadrant. For each trial, the latency to escape from water of each rat was calculated. A probe test, in which the hidden platform was removed, was conducted after the last trial on training day 5. The probe test was performed and the rats were allowed to swim freely for 60 sec with the platform absent. The time spent in the target quadrant and the times of platform crossing were measured. All data were recorded using a visual tracking system (SuperMaze software, Shanghai Xinruan Information Technology, Co., Ltd.).

Rats were sacrificed following anesthesia with 10% chloral hydrate (300 mg/kg, intraperitoneal injection). The rats were anaesthetized within 5 min after the chloral hydrate intraperitoneal injection. The anesthetized states of the rats were evaluated by the reaction of rats to external stimuli, such as clamping the root of the rat's tail. There were no signs of peritonitis, pain or discomfort following the intraperitoneal injection. The rats were then sacrificed by intracardiac puncture. The death of rats was confirmed by cardiac arrest. The brains of the rats were carefully and rapidly removed and washed with cold physiological saline. The hippocampus was immediately separated from the cerebrum on an iced plate and then stored at −80°C for use in subsequent experiments. The hippo-campal tissue was made into a tissue homogenate in ice-cold physiological saline solution, and centrifuged (1,000 × g, 4°C, 10 min) to remove particulates. The obtained supernatant was assayed. Superoxide dismutase (SOD) activity, and the malondi-aldehyde (MDA) and glutathione (GSH) levels were measured using a SOD Activity Assay kit, MDA assay kit and GSH assay kit supplied by Nanjing Jiancheng Bio-engineering according to the manufacturer's instructions.

Total RNA was isolated from the hippocampal tissues using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.) and reverse transcribed into cDNA using the Takara RNA PCR kit (AMV) (Takara Biotechnology Co., Ltd.). Specific primers for the target transcripts listed in Table I were synthesized (Invitrogen; Thermo Fisher Scientific, Inc.). mRNA expression was quantified using the SYBR-Green PCR kit (SYBR^® Premix Ex Taq™ II; Takara Biotechnology Co., Ltd.). The real-time detection was performed using the ABI prism 7900 RT-PCR System (Applied Biosystems). The results were normalized to the β-actin expression levels and the quantities of gene expression were calculated by the 2^−ΔΔCq method (26). All experiments were performed in triplicate. The primers of the target genes are listed in Table I.

Hippocampal samples (100 mg) were homogenized and lysed on ice with cell lysis buffer (Ding Changsheng Biotechnology). The homogenates were centrifuged at 2,000 × g or 5 min at 4°C. The supernatants were then collected. Nuclear extraction reagent (Pierce; Thermo Fisher Scientific, Inc.) was used for the separation of the cytosolic extract (cytosol) and nuclear extract (nucleus). The protein concentration was determined using the Bradford assay (Pierce; Thermo Fisher Scientific, Inc.). Equal amounts of protein (20 µg) were separated by 12% SDS-polyacrylamide gel electrophoresis, and transferred onto PVDF membranes. The membranes were blocked with 5% non-fat milk for 2 h and incubated overnight at 4°C with primary antibodies (rabbit anti-HO-1, Epitomics, cat. no. 2322-1, 1:800; anti-gp91phox, Santa Cruz Biotechnology, Inc., cat. no. sc-130543, 1:500; anti-p47phox, EMD Millipore, cat. no. 07-502, 1:500; anti-p22phox, Santa Cruz Biotechnology, Inc., cat. no. sc-130550, 1:200; anti-NQO1, Epitomics, cat. no. 2558-1, 1:1,000; anti-Nrf2, Epitomics, cat. no. 2178-1, 1:500; anti-β-actin, Santa Cruz Biotechnology, Inc., cat. no. sc-69879, 1:2,000; anti-histone H3, Santa Cruz Biotechnology, Inc., cat. no. sc-517576, 1:1,000). The membranes were then incubated with goat anti-rabbit secondary antibody conjugated with horseradish peroxidase (cat. no. 7074, 1:15,000; Cell Signaling Technology, Inc.) for 1 h at room temperature. Immunoreactive proteins were detected using enhanced chemiluminescence (Applygen Technologies Inc.). The quantification of the detected bands was performed using ImageJ software 3.0 (NIH). The data were standardized using β-actin (internal control of cytosolic protein) or histone H3 (internal control of nuclear protein).

All data are expressed as the means ± SD. The data of the Morris water maze were analyzed by repeated measures ANOVA and Tukey's test as the post hoc test. The other data were analyzed by one-way ANOVA followed by the Tukey's test. All statistical analyses were performed using SPSS 23 software (SPSS, Inc.). A P-value <0.05 was considered to indicate a statistically significant difference.

Movement, drinking, eating, urine volume and the fur color of the NC rats were normal. By contrast, the diabetic rats exhibited polyuria, polydipsia, weight loss and a dull fur color. As shown in Table II, hyperglycemia persisted in the diabetic rats after the STZ injection. The rats in the DC group was characterized by significantly increased FBG levels and a reduced body weight compared with those in the NC group (F_(4,45)=130.22, P<0.05; F_(4,45)=33.59, P<0.05). The other diabetic groups exhibited no statistically significant differences in FBG and body weight compared with the DC group (P>0.05). Thus, troxerutin and LA treatment had no marked effect on the FBG level and body weight of the diabetic rats.

Repeated measures ANOVA was used to compare the data between each group (Fig. 1A). The escape latency of DC and DN group increased significantly compared with the NC group (P<0.01). The escape latency of the DT and DL group decreased significantly compared with the DC group (P<0.05). No significant differences were observed in escape latency between the rats treated with troxerutin and LA (P>0.05).

In a probe trial of the water maze test, the times of crossing the platform area (Fig. 1C) and percentage of time spent in the target quadrant (Fig. 1D) were significantly decreased in the DC rats compared with the NC rats. These performances were significantly improved by troxerutin or LA treatment (F_(4,45)=21.74, P<0.05; F_(4,45)=129.46, P<0.05 respectively), with no significant differences observed between the DT and DL groups (P>0.05). No significant differences were also observed in the swimming speed among the 5 groups (P>0.05, Fig. 1B), indicating that the motor deficits of rats did no contribute to the differences in escape latency, times of crossing and time spent in the target quadrant.

As shown in Fig. 2, an increased SOD activity (F_(4,25)=427.22, P<0.01) and GSH concentration (F_(4,25)=74.25, P<0.01), alongside reduced MDA (F_(4,25)=189.12, P<0.01) levels were observed in the DL and DT groups, compared with the DC and DN groups.

RT-qPCR and western blot analysis were performed to measure the expression of NOX subunits in the hippocampus, including gp91phox, p47phox and p22phox. As illustrated in Fig. 3E-G, STZ-induced diabetes increased the mRNA expression of gp91phox, p47phox and p22phox in the DC and DN groups, while troxerutin treatment attenuated the augmentation of gp91phox, p47phox and p22phox expression in the diabetic rats (F_(4,25)=12.18, P<0.01; F_(4,25)=7.44, P<0.05; F_(4,25)=137.25, P=0.01, respectively). In addition, the cytosolic protein fractions of gp91phox, p47phox and p22phox were significantly increased in the diabetic groups. Following troxerutin treatment, all these levels were decreased compared with the samples from the DC group (F_(4,25)=55.50, P<0.01; F_(4,25)=6.59, P<0.05; F_(4,25)=60.58, P<0.01) (Fig. 3A-D).

To determine the effects of troxerutin on the activation of the Nrf2/ARE signaling pathway in the model of diabetes-induced cognitive deficits, the hippocampal expression levels of Nrf2 and two well-known downstream phase II antioxidant enzymes, HO-1 and NQO1, were measured by western blot analsyis and RT-qPCR. As illustrated in Fig. 4H and I, STZ-induced diabetes significantly decreased the NQO1 and HO-1 mRNA expression levels. Troxerutin treatment resulted in a noticeably increased mRNA expression of NQO1 and HO-1 compared with the DC group (F_(4,25)=72.53, P<0.01; F_(4,25)=6.95, P<0.05, respectively).

As regards the cytosolic protein fractions, the NQO1 and HO-1 levels were markedly decreased in the DC group compared with those in the NC group. Troxerutin resulted in noticeably increased cytosolic protein fractions of NQO1 and HO-1 compared with those in the DC group (F_(4,25)=86.86, P<0.01; F_(4,25)=96, P<0.01) (Fig. 4A, E and F).

In addition, STZ-induced diabetes significantly inhibited the nuclear translocation of Nrf2. This inhibitory effect of Nrf2 was reversed by troxerutin. Consistent with the changes observed in Nrf2 cytosolic and nuclear protein expression (F_(4,25)=52.35, P<0.01; F_(4,25)=159.36, P<0.01) (Fig. 4A-D), the mRNA level of Nrf2 was markedly increased in the DT group compared with the DC group (F_(4,25)=69.62, P<0.01) (Fig. 4G). These results collectively demonstrate that troxerutin plays a protective role in the suppression of oxidative stress via the regulation of Nrf2/ARE signaling pathway. In Fig. 5, the schematic diagram demonstrates the mechanism of troxerutin in relieving the oxidative stress in hippocampus of diabetic rats.