AL-1 was synthesized and purified in our laboratory 16. Andro, LA, DMSO and glibenclamide were purchased from Alfa Aesar (War Hill, MA, USA). Alloxan, leupeptin, luminol were purchased from Sigma-Aldrich Corp. (St Louis, MO, USA). pNF-κB-luc, PRL-TK plasmid and dual luciferase reporter (DLR) assay kits were purchase from Promega Corp. (Madison, WI, USA). Lipofectamine 2000 and Opti-MEM medium were purchased from Invitrogen Corp. (Carlsbad, CA, USA). Mouse IL-1β and IFN-γ were purchased from PeproTech (Rocky Hill, NJ, USA). Polyclone anti-GLUT4 antibody was purchased from Chemicon International Inc. (Temecula, CA, USA). Polyclone anti-insulin antibody, ployclone anti-β-actin antibody and HRP-conjugated goat anti-rabbit antibody were purchased from Beijing Biosynthesis Biotechnology Co. Ltd. (Beijing, China).

Female BALB/c mice, aged 6–8 weeks (18–22 g), were obtained from the Experimental Animal Center of Guangdong Province, China (SPF grade). Mice were housed in an animal room with 12 h light and 12 h dark, and were maintained on standard pelleted diet with water ad libitum. After fasting for 18 h, mice were injected via the tail vein with a single dose of 60 mg/kg alloxan (Sigma-Aldrich), freshly dissolved in 0.9% saline. Diabetes in mice was identified by polydipsia, polyuria and by measuring fasting serum glucose levels 72 h after injection of alloxan. Mice with a blood glucose level above 16.7 mM were used for experiments.

Diabetic mice were randomly divided into 6 groups of 6 mice. The first group was given vehicle (20% DMSO in distilled water) as a diabetic control group; the 2nd, 3rd and 4th groups were given AL-1 at doses of 20, 40 and 80 mg/kg, respectively; the 5th group was given Andro at 50 mg/kg (equal molar dose to 80 mg/kg AL-1); the 6th group was given glibenclamide at 1.2 mg/kg as a positive control. And 6 non-diabetic mice received vehicle as a normal control group. On the 4th day after alloxan administration, fasting (12–14 h) blood glucose levels were measured using a complete blood glucose monitoring system (Model: SureStep, LifeScan, Johson-Johson Co., Shanghai, China). AL-1, Andro, glibenclamide and vehicle were given by intragastric administration once daily for 6 days, respectively. On the evening of day 6, all mice were fasted overnight (12–14 h), and the following morning, after blood glucose of all groups was measured, animals were killed by decapitation. Blood was collected by drainage from the retroorbital venous plexus and kept on ice. Pancreas and soleus muscle were removed and immediately frozen at -80°C for various assays. Clotted blood samples were centrifuged at 3,000 × g for 15 min to obtain serum. The levels of serum insulin were determined by chemiluminescent immunoassay using a commercially available kit (Beijing Atom HighTech Co., Ltd., Beijing, China).

Pancreatic tissues were collected and placed in fixative (40 g/l formaldehyde solution in 0.1 M PBS) overnight, and was washed with 0.1 M PBS, then paraffin embedded, sectioned (2 μm), and stained with hematoxylin and eosin. For immunostaining studies, rabbit anti-mouse insulin antibody (1:50; Beijing Biosynthesis Biotechnology Co. Ltd.) was incubated with the sample sections for 3 h at 37°C. Horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG antibody (1:200; Beijing Biosynthesis Biotechnology Co. Ltd.) was used for 3, 3'-diaminobenzidine (DAB) coloration. Area of pancreatic islet was analyzed using Olypus analySIS image analysis software (Olympus Optical Co., Tokyo, Japan).

GLUT4 protein extract was prepared as described in Takeuchi et al. 17 with modifications. Briefly, soleus muscles were homogenized in an ice-cold buffer containing 20 mM HEPES, 250 mM sucrose, 2 mM EGTA, 0.2 mM phenylmethylsulfonyl fluoride (PMSF), and 1 μM leupeptin (Sigma-Aldrich) at pH 7.4. Nuclei and unbroken cells were removed by centrifugation at 2,000 × g for 10 min. Total membrane fraction was prepared by centrifugation of the supernatant in a super-speed centrifuge at 190,000 × g for 1 h at 4°C. The membrane pellets were re-suspended in homogenization buffer and stored at -80°C. Immunoblotting was performed using polyclonal anti-GLUT4 antibody (1:2,000 dilution; Chemicon) at 4°C overnight, and polyclonal anti-actin antibody (1:500 dilution; Beijing Biosynthesis Biotechnology Co. Ltd.) was used as an inter-control. After washing with TBS-T, the blots were incubated for 1 h at room temperature with HRP-conjugated goat anti-rabbit antibodies (1:2,000 dilution; Beijing Biosynthesis Biotechnology Co. Ltd.), and were detected using ECL Plus (PIERCE, Rockford, IL, USA).

RIN-m cell is an insulinoma cell line derived from a rat islet cell tumor 18. Cells were purchased from the American Type Culture Collection and grown at 37°C in a humidified 5% CO₂ atmosphere in DMEM (Gibco/BRL, Grand Island, NY, USA) supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 units/ml of penicillin, and 100 μg/ml of streptomycin.

RIN-m (5 × 10⁴ cells/ml, 100 μl/well) were plated in 96-well plates. After incubation for 24 h, cells were pretreated with Andro, LA and AL-1 for 1 h. An equal volume of 1% DMSO was added as a vehicle control (DMSO final concentration to 0.1%). Then, 500 μM H₂O₂ were added, and the cells were incubated for another 24 h to induce cell injury. Viability of cultured cells was determined by MTT assay.

Luminol chemiluminescence (CL) was used to evaluate intracellular oxidant production. RIN-m cells were planted in 96-well plates and cultured in DMEM containing 10% fetal bovine serum and 450 mg/dl glucose. When cells reached the loose confluent layer, medium was replaced with DMEM containing 1% FBS and 100 mg/dl glucose for 24 h. The cells were then exposed to 100, 275 and 450 mg/dl glucose or 0.1, 1 and 10 μM glibenclamide under the presence of 100 mg/dl glucose for 2 h or pretreated with Andro, LA and AL-1 at a concentration of 1 μM for 1 h and exposed to 450 mg/dl glucose or 1 μM glibenclamide for another 2 h. After treatment, 1 mM luminol (in DMSO) was added to the cells (final concentration of 50 μM). The time luminol was added was recorded as time "0", and relative luminescence units (RLU) were measured for 10 s every 2 min for a total of 30 min on a luminometer (TECAN, Männedorf, Switzerland). The areas under the chemiluminescence curves (AUC_CL) measured from time "0" to 30 min after adding luminol were calculated using an Orange software (OriginLab, Jersey, NJ, USA).

RIN-m cells (1 × 10⁵ cells/ml, 400 μl/well) in growth medium (high glucose DMEM containing 10% FBS) were plated in a 24-well plate, and were incubated for 24 h. Plasmid pNF-κB-luc and PRL-TK (Promega) in a ratio of 50:1 were co-transfected into RIN-m cells as described by the transfection guideline of lipofectamine 2000 (Invitrogen), and cultured in Opti-MEM medium (Invitrogen) for 4 h. Then medium was changed with the growth medium, and the cells were cultured for another 12 h. Andro, LA, AL-1 or vehicle control (DMSO final concentration to 0.1%) was added (final concentration: 1 μM) to pre-treat cells for 1 h. IL-1β (5 ng/ml, PeproTech) and IFN-γ (50 ng/ml, PeproTech) were then added, and the cells were incubated for another 24 h. NF-κB expression was determined by the dual luciferase reporter (DLR) assay kits (Promega).

Data were expressed as the mean ± S.D. for the number (n) of animals in the group as indicated in table and figures. Repeated measures of analysis of variance were used to analyze the changes in blood glucose and other parameters. Compare value less than 0.05 was considered significant.

Alloxan specifically targets pancreatic beta cells, where it induces ROS, destroying the beta cells to cause diabetes. Mice administered 60 mg/kg, i.v. of alloxan became hyperglycemic after 3 days. The blood glucose reached 27.0 ± 1.2 mM (Table 1), a value within the acceptable diabetic range. Drugs were administered, i.g. starting on day 3 and continued daily for 6 days. On day 7, mice were sacrificed, and various assays were performed.

Glucose transport, which depends on insulin-stimulated translocation of glucose carriers within the cell membrane, is the rate-limiting step in carbohydrate metabolism of skeletal muscle 19. Glucose transporters mediate glucose transport across the cell membrane. GLUT4 is the predominant form in skeletal muscle 20. Diabetes is characterized by reduced insulin-mediated glucose uptake associated with reduced GLUT4 expression 21. In diabetic models, Andro and LA are both known to reduce blood glucose levels via upregulation of GLUT4 expression 1122. In the present study, the effect of AL-1 on GLUT4 content in the plasma membrane of isolated soleus muscles of diabetic mice was measured by western blot analysis. The protein level of GLUT4 in the soleus muscles of diabetic mice was only 49.5% that of the non-diabetic mice (Fig. 4; p < 0.05 compared with normal controls). Treatment of the diabetic mice with Andro (50 mg/kg) or AL-1 (80 mg/kg) for 6 days elevated GLUT4 protein levels to 94.6% and 84.7%, respectively, of that of the non-diabetic mice (Fig. 4; p < 0.05 compared with diabetic control). There was no significant difference between AL-1 and Andro treated group.

Alloxan produces ROS which contribute to destruction of pancreatic beta cells, leading to diabetes. The ability of AL-1 to protect RIN-m pancreatic cells from H₂O₂-induced oxidative damage was studied. The viability of RIN-m cells cultured 24 h with 500 μM H₂O₂ was reduced to 42.7 ± 11.1% (Fig. 5). Pretreatment of the H₂O₂-treated RIN-m cells with Andro, LA, AL-1 or a mixture of Andro and LA at 0.01, 0.1 and 1 μM 30 min prior to H₂O₂ exposure for 60 min, provided significant protection. The viabilities of cells at 24 h when incubated with 1 μM concentrations of Andro, LA, AL-1 or a mixture of Andro and LA was 59.7 ± 5.9%, 59.7 ± 4.4%, 64.3 ± 11% and 62.2 ± 10.6% respectively. AL-1 was more effective than either Andro or LA. At 0.1 μM, only LA and AL-1 provided a significant protective effect. The protective effect of AL-1 was concentration-dependent. The effect of the mixture of Andro and LA was not better than AL-1, demonstrating that AL-1 was more than a simple mixture of Andro and LA.

High concentrations of glucose stimulate ROS production both in vitro 23 and in vivo 2425. ROS subsequently impair cellular function and activate apoptosis signaling, leading to beta cell damage and death 26. To investigate the effect of AL-1 on glucose-induced ROS production in vitro, RIN-m cells were incubated in the presence of high concentrations of glucose, and the production of ROS was measured. Exposure of RIN-m cells to increasing concentrations of glucose (100–450 mg/dl) for 2 h increased ROS production in a concentration-dependent manner. Pretreatment of the cells with 1 μM of Andro, LA or AL-1 effectively quenched the production of increased ROS. AL-1 and LA were equally effective but more potent than Andro (Fig. 6a).

Glibenclamide treatment decreases hyperglycemia in alloxan-induced diabetic animals (Tab. 1) and protects beta cell mass from significant loss (Fig. 3f). However, the pancreatic beta cells of the glibenclamide-treated diabetic have reduced immunoreactive insulin (Fig. 3l). To understand these results, RIN-m cells were incubated with glibenclamide at increasing concentrations, and ROS production was measured. Glibenclamide dose-dependently increased ROS production (Fig. 6b), a finding previously reported 27. Iwakura et al.28 reported that viability of RIN-m cells was decreased in a dose-dependent manner by continuous exposure to glibenclamide at concentrations of 0.1 to 100 μM. When the cells were incubated in the presence of both 1 μM glibenclamide and 1 μM of Andro, LA or AL-1, the ROS induced by glibenclamide were almost completely eliminated (Fig. 6b).

Activation of NF-κB impairs the function of beta cells and contributes to cellular death 2930. A NF-κB reporter assay was used to investigate the effect of AL-1 on NF-κB activation. Cells were co-transfected with pNF-κB-luc and PRL-TK plasmids, pre-incubated with Andro, LA, AL-1 or vehicle followed by addition of IL-1β and IFN-γ. AL-1 at 0.1 and 1 μM significantly inhibited luciferase activity of the NF-κB reporter construct (Fig. 7; p < 0.01 compared with vehicle control). In fact, at 1 μM, AL-1 completely blocked IL-1β and IFN-γ-induced NF-κB activation. By contrast, Andro showed substantial NF-κB inhibition only at the highest concentration of 1 μM. AL-1 was at least 10-fold more potent than the parent compound Andro in this experiment.

Hidalgo et al. 31 reported that Andro at 5 and 50 μM significantly inhibited PAF-induced luciferase activity in a NF-κB reporter construct. Zhang and Frei 32 found that preincubation of human aortic endothelial cells for 48 h with LA (0.05–1 mM) inhibited TNF-α (10 U/ml)-induced NF-κB binding activity in a dose-dependent manner. In the presence of 0.5 mM LA, the TNF-α-induced NF-κB activation was inhibited by 81%. Thus, in the present experiment, a 1 μM concentration of LA may be too low to suppress NF-κB activation.