NSE/hPS2m Tg mice overexpressing human mutant PS2 (hPS2m, N141I), under the control of an NSE promoter were used in this study (19,20) and were obtained from the Department of Laboratory Animal Resources in Korea FAD. These Tg mice showed behavioral dysfunction, Aβ-42 deposition and the induction of caspase-3 and Cox-2 activities at 12 months of age. To analyze the protein level and insulin concentration, a total of 10 mice were used; five NSE/hPS2m Tg mice and five non-Tg littermates at 12 months of age. The mice were handled in a Pusan National University-Laboratory Animal Resources Center accredited by the Korea FDA in accordance with the USA NIH guidelines (accredited unit no. 231). The mice were maintained in a specified pathogen-free environment and were housed in cages under a strict light cycle (light period for 12 h and dark period for 12 h) and were given a standard irradiated chow diet (Purina Mills Inc., Milford, IN, USA) ad libitum.

For detection of the Aβ-42 level, the frozen brain of mice was sectioned with scissors and homogenized in Pro-Prep™ Protein Extraction Solution [Intron Biotechechnology Co. Ltd., Seongnam, Korea, (50 mM Tris (pH 7.4), 150 mM NaCl, 1 mM DTT, 0.5% NP-40, 1% Triton X-100, 1% deoxycholate, 0.1% SDS, proteinase inhibitor)] with a glass homogenizer. The homogenate mixture was centrifuged at 22,250 × g for 10 min at 4°C to eliminate the nuclei and unbroken cells. The protein prepared from the brain was separated by electrophoresis in a 4–20% SDS-PAGE gel for 3 h and transferred to nitrocellulose membranes for 2 h at 40 V. Each membrane was incubated separately with the primary anti-Aβ-42 antibody (MAB1560, 6E10; Chemicon International, Temecula, CA, USA) overnight at 4°C. The membranes were washed with the washing buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄ and 0.05% Tween-20) and incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG (1:1,000 dilution; Zymed, South San Francisco, CA, USA) at room temperature for 2 h. Blots were developed using a Chemiluminescence Reagent Plus kit (ECL; Amersham Pharmacia, Piscataway, NJ, USA).

Detection of γ-secretase activity was performed according to the manufacturer’s instructions (FP003; R&D System Inc., Wiesbaden, Germany). Frozen brain tissues were sectioned with scissors and homogenized in lysis solution and 1X cell extraction buffer with a glass homogenizer into 0.5–2.0 mg/ml final concentrations. The homogenate was separated by centrifugation at 22,250 × g for 15 min at 4°C and then supernatant was collected for the protein and enzyme assays. Protein was assayed by the BCA method (Pierce, Rockford, IL, USA) using an ELISA reader. For the determination of γ-secretase activity, the 50 μl of tissue lysate (25–200 μg) prepared from the brain extract was added to each well in a 96-well microplate in triplicate and then 2X reaction buffer (50 μl) was added. The substrate (5 μl) was added to each well and the plate was incubated in the dark at 37°C for 1–2 h. Fluorescence was measured at a wavelength of 335–355 nm using a Fluorescent Microplate Reader FL600 (Bio-Tek Instrument, Inc., Winooski, VT, USA).

Blood was collected from the abdominal vein of NSE/hPS2m Tg mice and non-Tg littermates and then incubated at room temperature for 30 min. The serum was separated by centrifugation at 890 × g for 15 min at 4°C. Serum insulin concentration was carried out as per the manufacturer’s instructions [Coat-A-Count Insulin kit; Diagnostic Products Corp., LA, CA, USA) using a gamma counter (Cobra 5010 Quantum, Cobra 5010 II; Packard Instrument Co., Inc., Meriden, CT, USA)]. The glucose concentration was assayed according to the manufacturer’s instructions by Glucose Kit using an Automatic Biochemical Analyzer (Hitachi 747, Tokyo, Japan).

The brain was harvested from NSE/hPS2m and non-Tg mice. Frozen brain was sectioned with scissors and homogenized in buffer A [10 mM Tris (pH 7.4), 1 mM EDTA, 250 mM Sucrose, proteinase inhibitor (pH 7.4)] with a glass homogenizer. The homogenate mixture was centrifuged at 900 × g for 10 min at 4°C to eliminate the unbroken cells. The supernatant was transferred to a new tube and centrifuged at 110,000 × g for 75 min at 4°C to collect the microsomal fraction. The pellet containing the microsomal fraction was resuspended in lysis buffer A containing 1% Triton X-100 for use in the western blot analysis.

The protein prepared from the brain tissues was separated by electrophoresis in a 4–20% SDS-PAGE gel for 3 h and transferred to nitrocellulose membranes for 2 h at 40 V. Each membrane was incubated separately with the primary, anti-insulin receptor α (sc-710; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), anti-insulin receptor β (sc-711; Santa Cruz Biotechnology, Inc.), anti-IGF receptor (I7151; Sigma-Aldrich, St. Louis, MO, USA), anti-Glut-1 (sc-7903; Santa Cruz Biotechnology, Inc.), anti-Glut-3 (sc7582; Santa Cruz Biotechnology, Inc.), anti-Glut-4 (400064; Calbiochem, Darmstadt, Hesse, Germany), anti-AKT (Ab3130; Chemicon International), anti-p-AKT (Ab3132; Chemicon International), anti-GSK (9332; Cell Signaling Technology, Inc., Boston, MA, USA), anti-p-GSK (9331; Cell Signaling Technology, Inc.), anti-Tau (T7194; Sigma-Aldrich) and anti-p-Tau (T8069; Sigma-Aldrich) antibodies overnight at 4°C. The membranes were washed with washing buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄, 0.05% Tween-20) and incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG (1:1,000 dilution; Zymed) at room temperature for 2 h. Blots were developed using a Chemiluminescence Reagent Plus kit (ECL; Amersham Pharmacia).

Initially, the 12-month-old NSE/hPS2m Tg mice were divided into three subgroups: No-treatment group [1X phosphate-buffered saline (PBS)], low-dose group (0.2 units of insulin), high-dose group (0.4 units of insulin). Insulin (Novolin N, 100 IU/ml; GreenCross, Yongin, Korea) was subcutaneously injected into NSE/hPS2m Tg mice in a volume of 0.2 ml dilution in 1X PBS solution for 4 weeks.

Tests for significance between non-Tg and Tg mice were performed using a one-way ANOVA test of variance (SPSS for Windows, Release 10.10, standard version; SPSS, Chicago, IL, USA). Post-hoc tests of variance (SPSS for Windows, Release 10.10, standard version) were used to determine significance between insulin treatment and non-treatment Tg groups. Values were reported as the mean ± standard deviation (SD). P<0.05 was considered to indicate a statistically significant difference.

To detect the phenotypes of AD resulting from the overexpression of the hPS2m transgene, the Aβ-42 peptide and activity of γ-secretase were quantified in total brain tissue of 12-month-old NSE/hPS2m Tg mice by western blot analysis and γ-secretase activity assay, respectively. As shown in Fig. 1A, the level of Aβ-42 peptide was significantly higher in the brain of the NSE/hPS2m Tg mice compared with that of the non-Tg littermates. Also, γ-secretase activity was significantly increased in the brain of the NSE/hPS2m Tg mice (Fig. 1C). Furthermore, the level of p-Tau (Thr 231 site) aggregated into the neurofibrillary tangle in the cytoplasm of particular pyramidal neurons was significantly increased in the brain of NSE/hPS2m Tg mice (Fig. 1B). These results suggest that NSE/hPS2m Tg mice aged 12-months possessed the AD-like phenotypes involving the deposition of Aβ-42, the increase of γ-secretase activity and Tau hyperphosphorylation.

To investigate whether cortical and hippocampal disorganization caused by Aβ-42 peptides affected the insulin and glucose levels, insulin and glucose concentrations were measured in the serum obtained from NSE/hPS2m Tg mice. In the Coat-A-Count Insulin kit analysis, the concentration of insulin was significantly lower in the NSE/hPS2m Tg mice (2.1 mg/dl) than that in the non-Tg littermates (6.2 mg/dl) (Fig. 1D). By contrast, the concentration of glucose was significantly higher in the NSE/hPS2m Tg mice (210 mg/dl) compared with that in the non-Tg littermates (120 mg/dl) (Fig. 1E). Therefore, changes of the insulin and glucose concentration in Tg mice indicated that the deposition of Aβ-42 peptide was directly associated with defects in the glucose metabolism.

To determine whether the decrease in the insulin concentration caused by Aβ-42 peptide affected its related receptor expression, the expression levels of the insulin receptor and IGF receptor protein were measured using western blot analysis from the membrane protein of the brain tissues. Results of the western blot analysis revealed that the expression of insulin receptors α and β to be significantly reduced in the NSE/hPS2m Tg mice compared with the non-Tg littermates (Fig. 2A and B). However, these analyses also revealed no differences in the expression level of IGF receptors in the brain between the NSE/hPS2m Tg mice and non-Tg littermates (Fig. 2C). These results suggest that the defect in glucose regulation was capable of significantly decreasing the level of insulin receptors α and β, although the level of IGF receptor was unaffected.

To test the hypothesis that the defect of the glucose metabolism in the AD model would alter the phosphorylation of signal protein in the insulin receptor signal pathway, the phosphorylation rate of AKT and GSK proteins in the brain of NSE/hPS2m Tg mice was detected. AKT2 is a key signaling molecule in the insulin signaling pathway that induces glucose transport (21). Glycogen synthase kinase 3 (GSK-3) involved in the Wnt signaling cascade may be inhibited following phosphorylation by AKT (22). The level of p-AKT protein involved in the downstream signal pathway of the insulin receptor was reduced in the brain of NSE/hPS2m Tg mice compare with non-Tg littermates (Fig. 3A). By contrast, the p-GSK-3β level was higher in the NSE/hPS2m Tg mice than the non-Tg littermates (Fig. 3B). Thus, this result suggests that the decrease in the insulin receptors α and β significantly induced the change of signal protein activation in the downstream signal pathway of the insulin receptor, respectively.

To determine whether the decrease of the glucose level affected the expression of glucose transporter, western blot analysis was performed to detect the expression level of the glucose transporter in the membrane protein of mice brains. Fig. 4 shows the differences in the distribution of various molecular mass forms of Glut-1 and -3 in total cerebral homogenate. Two discrete bands at 45- and 55-kDa are observed in total homogenate. The density of the 45- and 55-kDa band in Glut-1 was significantly reduced in NSE/hPS2m Tg mice compared with that of non-Tg littermates (Fig. 4A). The 45-kDa band in Glut-3 was significantly reduced in the NSE/hPS2m Tg mice when this level was compared with non-Tg littermates. However, the density of the 55-kDa band was not significantly different between the NSE/hPS2m Tg and control mice (Fig. 4B). By contrast, there was no difference in the Glut-4 protein level between the NSE/hPS2m Tg mice and non-Tg littermates (Fig. 4C). These results suggest that the defect of glucose regulation may be attributed to the decrease in the expression level of Glut-1 and -3 proteins, but not for Glut-4.

Previously, Boyt et al(23) suggested that glucose ingestion and the subsequent elevation of plasma concentration of glucose and insulin lead to a decrease in the level of amyloid precursor protein in plasma. Therefore, to examine whether treatment of insulin in the NSE/hPS2m Tg mice affected the concentration of the Aβ-42 peptide, the insulin was injected into 12-month-old Tg mice. In RIA and serum biochemical analysis, the insulin concentration was significantly higher in the serum of insulin-treated mice than that of non-treated Tg mice. However, the glucose concentration in the insulin-treated mice was 3–4-fold lower than in the non-treated Tg mice, respectively (Fig. 5A). Furthermore, the western blot analysis revealed that the level of Aβ-42 peptide in the brain of the insulin-treated mice was slightly reduced compare with that of non-treated mice (Fig. 5B). Altering the level of Aβ-42 peptide from insulin-treated mice as shown by western blot analysis, suggests the possibility that insulin treatment is directly associated with a decrease in the Aβ-42 peptide (Fig. 5C).