Resveratrol relieves gestational diabetes mellitus in mice through activating AMPK

Care and use of mice in this study abide the guidelines and protocol approved by the Institutional Animal Care and Use Committee of Shandong Provincial Hospital Affiliated to Shandong University. All efforts were exhausted to minimize unnecessary suffering of experimental mice.

Six to eight weeks old C57BL/KsJ^+/+ (wild type) and C57BL/KsJ ^db/+ (db/+) mice were purchased from Jackson Laboratories. Mice were housed in a room with controlled temperature (22 ± 2 °C), humidity (40–60 %) and light cycle (12/12 h light/dark). Mice were fed with the chow diet (Harlan Teklad) consisted of 29 % protein, 47 % carbohydrate and 17 % fat, and water. Female mice were randomly divided into three experimental groups (n = 18 per group): wild type, ad libitum-fed; db/+ pair-fed, food intake of the ad libitum-fed wild-type was measured daily, and the same amount of food was pair-fed to db/+ mice; db/+ pair-fed + RV, db/+ pair-fed mice were administered by oral gavage with resveratrol (RV; Sigma-Aldrich), solubilized in water, at the dose of 10 mg/kg body weight per day, throughout the entire span of the study. Wild type and db/+ pair-fed groups of mice were also orally gavaged with water only daily, throughout the entire span of the study, to serve as vehicle control. At 10–12 weeks of age, female mice were individually mated with males of the same genotype, and mating was confirmed by the presence of a copulatory plug the next morning, which was designated gestation day (GD) 0. Since not all females could become pregnant after mating, an initial number of more than 18 females per group were used, and eventually 18 pregnant females were used in each experimental group as described above.

Mice before pregnancy and on GD 10 were fasted for 6 h and injected intraperitoneally with glucose at 2 g/kg body weight. Blood samples were collected from the tail using capillary tubes at 0, 30 and 60 min after glucose administration.

Mice on GD 10 were fasted 6 h and injected intraperitoneally with insulin at 0.75 U/kg body weight. Blood samples were collected from the tail using capillary tubes at 0, 30 and 60 min after insulin administration.

Serum glucose, insulin levels and body weight were measured at GD 0, 10 and 20. Non-fasting blood samples were obtained via tail venipuncture, and serum glucose level was determined by glucometer (Lifescan Surestep). Plasma insulin levels were quantified by UltraSensitive Mouse Insulin ELISA kit (ALPCO Diagnostics). Body weight was measured on a top-loading balance (Fisher Scientific).

Livers were dissected using previously established method [19]. On GD 10 after completing all other measurements, 6 out of 18 mice in each group were sacrificed to harvest the liver, while the remaining 12 females were allowed to complete pregnancy, and their offspring were also sacrificed after birth to harvest the liver. Mice were anesthetized with 100 mg/kg ketamine, 10 mg/kg acepromazine and 100 mg/kg xylazine. The abdominal cavity was opened and the portal vein exposed. A bolus of insulin (10 U/kg body weight in 100 μL saline) was injected into the portal vein. The liver was immediately excised and frozen at −80 °C until further analysis.

Protein samples were extracted from liver homogenates in the presence of protease and phosphatase inhibitors. Protein concentration was quantified by BCA assay system (Bio-Rad, USA). Proteins were analyzed by 12 % SDS-PAGE and transferred to nitrocellulose membrane using semi-dry transfer unit (Bio-Rad). The membrane was then immersed in blocking buffer (PBS, 0.1 % Tween-20) with 5 % nonfat milk for 20 min and then incubated with appropriate primary antibody (1:1000 dilution in blocking buffer) overnight at 4 °C. After washes with blocking buffer, blots were incubated with appropriate HRP-conjugated secondary antibody (1:5000 dilution in blocking buffer; Life Technologies) for 20 min at room temperature, washed again with blocking buffer, and visualized using Luminata Forte Western HRP Substrate (EMD Millipore). AMPK (#2532), phosphor-AMPK (#2535), HDAC4 (#2072) and phosphor-HDAC4 (#3443) antibodies were purchased from Cell Signaling. Glucose-6-phosphatase (ab83690) and α-tubulin (ab125267) antibodies were purchased from Abcam.

Microsomes were prepared from frozen liver samples as previously described [19]. G6Pase activity was measured on intact microsomes with 0.5, 2.5, and 10 mM of glucose-6-phosphate (G6P) as previously described [19]. Total protein was determined by the Bradford method. G6Pase activity was normalized by subtracting nonspecific phosphatase activity, which was determined using paranitrophenylphosphate.

We first examined the effect of RV in the non-pregnant db/+ mice, which usually do not present diabetic symptoms as previously reported [19]. By employing a glucose tolerance test in all three groups of non-pregnant mice, we confirmed that glucose and insulin levels were similar in wild type and pair-fed db/+ mice, and RV didn’t cause any changes in either glucose or insulin levels (Fig. 1b and c).

Next we performed the same glucose tolerance test in all three groups of mice, during their pregnant state on GD 10. We found that, compared with wild type, pregnant db/+ females exhibited prominent glucose intolerance, as indicated by significantly elevated glucose levels after the initial glucose injection (Fig. 1d, 30 and 60 min), and RV treatment was able to largely alleviate this glucose intolerance in the pair-fed db/+ mice. In the same experiment, insulin levels of db/+ mice was significant lower than that of wild type throughout the test (Fig. 1e, 0, 30 and 60 min), explaining the glucose intolerance phenotype. Importantly, RV administration restored insulin response to an indiscernible level as the wild type control mice.

Compared to glucose tolerance test, which assesses the combined outcome of glucose disposal and insulin secretion, the insulin tolerance test is a more direct measurement on the ability of insulin to rapidly stimulate glucose disposal [8]. Therefore we next performed an acute insulin tolerance test on GD 10 in all three groups of mice (Fig. 1f). As expected, the pair-fed db/+ females exhibited significantly higher blood glucose levels than the wild type following the initial insulin injection, suggesting insulin resistance phenotype. Notably RV administration markedly reduced glucose levels in the pair-fed db/+ mice, to similar levels as the wild type. Taken together, the above results indicated that RV was able to extenuate acute glucose and insulin intolerance in pregnant db/+ mice.

We also measured the steady state levels of blood glucose and insulin levels throughout the pregnancy. We found that blood glucose levels of wild type females remained stable on GD 0, 10 and 20 (Fig. 2a). Whereas meanwhile, db/+ females showed significantly elevated blood glucose levels, indicating the typical hyperglycemia symptom of GDM. However, blood glucose levels of db/+ females receiving RV stayed at the same as the wild type group through the entire duration of pregnancy, indicating the alleviating effect on hyperglycemia by RV. In a similar manner, steady state insulin levels in db/+ mice were also significantly lower than wild type, which could be fully restored by RV administration (Fig. 2b). Maternal body weight gains in all three groups of mice were almost the same throughout the pregnancy (Fig. 2c).

As AMPK is likely to be a potential target of RV function in diabetes, we sacrificed 6 mice from each group at GD 10 and harvested their livers for immunoblot analysis. As shown in Fig. 3a, AMPK activation was found to be attenuated in db/+ pair-fed mice, contributing to higher HDAC4 activation, which in turn elevated G6Pase expression in the liver. Consistent with previous reports [15, 16], RV administration fully reversed the trend of activation profile of the related proteins, eventually reducing G6Pase levels to almost the same as in wild type (Fig. 3b).

One of the adverse effects of GDM is decreased fetal survival, presented as fewer number of fetuses [20, 21]. Moreover in the same db/+ GDM mouse model, body weight of offspring at birth was reported to increase by 5–8 % [7, 8]. Therefore we next investigated whether RV was able to improve fetal development of GDM female mice.

Offspring number at birth was counted for each females from wild type, db/+ pair-fed and db/+ pair-fed + RV groups (Fig. 4a). The 12 wild type female mice gave birth to 86 litters, whereas the 12 females from the db/+ pair-fed group gave birth to only 58 litters. In contrast, 82 litters were born by the 12 db/+ pair-fed + RV group of females. Furthermore we recorded the body weight of offspring at birth in the three groups (Fig. 4b), and found that, consistent with previous reports [7, 8], mean body weight of offspring born by db/+ pair-fed mothers was significantly higher than those by wild type mothers, whereas body weight of offspring from db/+ pair-fed + RV group was almost the same as wild type control. Differences in reproduction outcome among these groups was not attributed to any specific GDM genetic condition of the fetus, since all of their offspring genotype composition followed Mendelian ratio (Additional file 1: Table S1). These results clearly indicated RV was able to improve the reproductive outcome of GDM female mice.

We speculate the reason for the increased body weight at birth of offspring from db/+ mice could also be attributed to elevated liver G6Pase. Therefore we chose 12 offspring from each group of female mice, with the same genotype composition, and harvested their livers to assay for G6Pase activity (Fig. 5). As expected, G6Pase activity was significantly increased in the db/+ pair-fed offspring, which could be reduced by RV administration in the db/+ pair-fed + RV offspring. This result clearly indicated that RV also reduced the enzymatic capacity for glucose production in the fetus, likely through enhancing AMPK activation (Fig. 3).