Nonalcoholic fatty liver disease and cirrhosis are strongly associated with insulin resistance and glucose intolerance. To date, the influence of metformin on glycogen synthesis in the liver is controversial. Limited studies have evaluated the effect of metformin on hepatic insulin signaling pathway in vivo. In this study, an insulin-resistant rat model of nonalcoholic steatohepatitis and cirrhosis was developed by high-fat and high-sucrose diet feeding in combination with subcutaneous injection of carbon tetrachloride. Liver tissues of the model rats were featured with severe steatosis and cirrhosis, accompanied by impaired liver function and antioxidant capacity. The glucose tolerance was impaired, and the index of insulin resistance was increased significantly compared with the control. The content of hepatic glycogen was dramatically decreased. The expression of insulin receptor β (IRβ); phosphorylations of IRβ, insulin receptor substrate 2 (IRS2), and Akt; and activities of phosphatidylinositol 3-kinase (PI3K) and glycogen synthase (GS) in the liver were significantly decreased, whereas the activities of glycogen synthase kinase 3α (GSK3α) and glycogen phosphorylase a (GPa) were increased. Metformin treatment remarkably improved liver function, alleviated lipid peroxidation and histological damages of the liver, and ameliorated glucose intolerance and insulin resistance. Metfromin also significantly upregulated the expression of IRβ; increased the phosphorylations of IRβ, IRS2, and Akt; increased the activities of PI3K and GS; and decreased GSK3α and GPa activities. In conclusion, our study suggests that metformin upregulates IRβ expression and the downstream IRS2/PI3K/Akt signaling transduction, therefore, to increase hepatic glycogen storage and improve insulin resistance. These actions may be attributed to the improved liver histological alterations by metformin.
The liver is a primary target organ for insulin and plays a critical role in glucose homeostasis. It maintains a balance between the uptake and storage of glucose by regulating glycogenesis and the production of glucose by glycogenolysis and gluconeogenesis (Radziuk & Pye 2001). Nonalcoholic fatty liver disease (NAFLD) is a spectrum of diseases ranging from simple steatosis to nonalcoholic steatohepatitis (NASH), which may progress to cirrhosis. NAFLD is strongly associated with insulin resistance, and up to 50% of NAFLD patients have type 2 diabetes (Dyson & Day 2014). A significant proportion of patients with cirrhosis ranging between 35 and 80% may be insulin resistant, and 20–30% of patients may be clinically diabetic (Ahmadieh & Azar 2014).
In the liver, the insulin receptor substrate 2 (IRS2)/phosphatidylinositol 3-kinase (PI3K)/protein kinase B (Akt) signaling transduction pathway plays a pivotal role in modulating the glucose metabolic actions of insulin (Pessin & Saltiel 2000). Insulin receptor (IR) is a membrane protein composed of two extracellular α-subunits and two transmembrane β-subunits, which contain intracellular tyrosine kinase domains. Binding of insulin to the α-subunit increases the intrinsic tyrosine kinase activity and initiates autophosphorylation of the β-subunit, triggering a cascade of phosphorylation and activation of essential downstream targets, including IRS2, PI3K, and Akt (Valverde et al. 2003, Ide et al. 2004). The activation of Akt causes the phosphorylation and inactivation of glycogen synthase kinase 3α (GSK3α). GSK3 is the first identified Akt substrate and refers to two isoforms, GSK3α and GSK3β. GSK3α is mainly involved in the process of glycogen synthesis, whereas GSK3β exhibits more complex biological functions (Beurel et al. 2015). The inactivation of GSK3α induces the activation of glycogen synthase (GS). PI3K/Akt activation also leads to the conversion of glycogen phosphorylase (GP) from the active state (GPa) to the inactive state (GPb). The activation of GS and the depletion of GPa finally result in the stimulation of glycogen synthesis (Agius 2015).
Metformin, a potent biguanide hypoglycemic agent, has been clinically used for over five decades and is considered the first-line oral therapy for type 2 diabetes. Plenty of studies demonstrate that metformin improves glucose metabolism by suppressing hepatic glucose production mainly through inhibition of gluconeogenesis (Viollet et al. 2012). A number of potential mechanisms of decreasing hepatic gluconeogenesis by metformin have been proposed, including inhibition of mitochondrial respiratory chain activity, activation of AMP-activated protein kinase (AMPK), decrease in hepatic energy state, and inhibition of cAMP signaling through phosphorylation of CREB-binding protein (He et al. 2009, Cao et al. 2014, Madiraju et al. 2014). However, the precise molecular targets of metformin action still remain elusive. Few studies have examined the effect of metformin on hepatic glycogen synthesis, and the results are controversial. Increased hepatic glycogen content in response to metformin was detected in high-fat-fed, insulin-resistant rat model and in hepatoma cells. Also it has been reported that metformin impaired glycogen synthesis in isolated rat hepatocytes (Purrello et al. 1988, Mithieux et al. 2002, Otto et al. 2003). No in vivo data on the influence of metformin on IRS2/PI3K/Akt signaling pathway is available.
In this study, an insulin resistance-associated rat model of NASH and cirrhosis was developed and treated with metformin. The glycogen content in the liver was determined. The expression, phosphorylation status, and the activity of the key functional molecules in IRS2/PI3K/Akt signaling pathway were detected. Our aim was to evaluate the influence of metformin on hepatic glycogen storage and insulin signaling transduction pathway.
Materials and methods
Animals and chemicals
Male Spraque–Dawley (SD) rats were obtained from Sino-British SIPPR/BK Lab Animal Ltd, Shanghai, China (license no. SCXK 2008-0016) and reared at the Experimental Animal Research Center, Zhejiang Chinese Medical University, Hangzhou, China. Animals were maintained in the specific pathogen-free facility in accordance with the Institutional Animal Care Guidelines. The protocol was reviewed and approved by the Committee on Animal Research and Ethics of Zhejiang Chinese Medical University (authorization number: ZSLL-2012-058). All efforts were made to minimize suffering. Animals were killed under sodium pentobarbital anesthesia in a separate room and terminally bled via ventral aorta puncture.
Carbon tetrachloride (CCl4) (289116) was purchased from Sigma-Aldrich. Metformin was purchased from Sino-American Shanghai Squibb Pharmaceuticals (Shanghai, China). Enzyme-linked immunosorbent assay (ELISA) kit (50R-E.1637) for determining insulin levels was obtained from Biovalue (Shanghai, China). Activity assay kits for PI3K (GMS50058.2), GSK3α (GMS50161.2), GS (GMS50500.2), and GPa (GMS50092.4) and glycogen assay kit (GMS70085.2) were purchased from Genmed Scientifics (Shanghai, China). Malondialdehyde (MDA; A003-1), superoxide dismutase (SOD; A001-2), reduced glutathione (GSH; A006-1), and glutathione peroxidase (GSH-Px; A005) assay kits were purchased from Nanjing Jiancheng Biotechnology Institute, Nanjing, China. Bicinchoninic acid (BCA) protein assay kit (23227) was purchased from Pierce. Mouse anti-IRβ monoclonal antibody (ab69508) was obtained from Abcam. Rabbit anti-IRS2 (3089), rabbit anti-phospho Akt (Ser473) (4060), and rabbit anti-IRβ monoclonal antibody (3025) were purchased from Cell Signaling Technology. Mouse anti-phosphotyrosine (PY) monoclonal antibody (05-321) was obtained from Millipore. Rabbit anti-Akt (sc-8312) and Protein A/G plus-agarose beads (sc-2003) were obtained from Santa Cruz Biotechnology.
Developing insulin resistance-associated rat model of NASH and cirrhosis
Thirty male SD rats weighing 180–200 g were equally randomized into three groups: control group (n=10), model group (n=10), and metformin-treated group (n=10). Cirrhosis was induced by high-fat and high-sucrose diet (HFHSD) feeding in combination with subcutaneous injection of CCl4. The rats in the model group and metformin-treated group were fed with HFHSD, which consist of 74.25% standard chow, 10% sucrose, 0.5% cholesterol, 5% egg yolk powder, 10% lard, and 0.25% sodium cholate. Meanwhile, these rats received a single injection of 40% CCl4 dissolved in olive oil at 5mL/kg and then 3mL/kg twice every week for 12 weeks. After 4 weeks of CCl4 exposure, the rats in the metformin-treated group additionally received intragastric administration of metformin twice a day at a dose of 200mg/kg/day. The control rats were fed a standard chow, and received olive oil subcutaneously and normal saline intragastrically. The standard laboratory chow contains 20.5% protein, 4.62% fat, 52.5% nitrogen-free extract, and 4.35% fiber. Food and water were available ad libitum throughout the study.
Measurement of serum biochemical parameters
Blood samples without anticoagulant from the inferior vena cava were centrifuged at 2000g for 10min to collect the serum. The activities of serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), and alkaline phosphatase (ALP), and the levels of serum albumin (Alb), total bilirubin (TBil), and total cholesterol (TC) were detected using an Olympus AU5400 automatic biochemical analyzer.
Oxidative stress biomarker and antioxidant profile
Liver samples were homogenized in 0.1M ice-cold phosphate buffer saline (pH 7.4). The homogenates were centrifuged at 14,000g for 10min at 4°C. The supernatant was used for measuring MDA, SOD, GSH, and GSH-Px using commercial kits according to the manufacturer’s protocols.
Hematoxylin and eosin, Sirius red, and periodic acid–Schiff staining
Liver specimens were fixed in 4% paraformaldehyde and dehydrated in a graded alcohol series. After embedded in paraffin, the tissues were sectioned at 4μm in thickness and stained with hematoxylin and eosin (H&E). For evaluation of hepatic collagen deposition, Sirius red staining was performed and a red color staining was considered where the collagen deposited. Glycogen storage was detected by periodic acid-Schiff (PAS) staining staining as described elsewhere (Saitoh et al. 2010).
After deparaffinization and rehydration, microwave antigen retrieval was performed on the sections as per a standard protocol before peroxidase quenching with 3% hydrogen peroxide. Slides were incubated at 4°C overnight with mouse anti-IRβ monoclonal antibody (1:500), followed by secondary biotynilated goat anti-mouse IgG and streptavidin–peroxidase at 37°C for 30min each. The reaction was revealed with 3,3′-diaminobenzidine tetrahydrochloride. Nuclei were stained with hematoxylin. For negative controls, the primary antibody was replaced with phosphate buffer saline.
Oral glucose tolerance test
At the end of the 12-week experiment, the rats were fasted overnight and fed with d-glucose by gavage at a dose of 3.0g/kg body weight. Blood samples were obtained by tail bleeding for determination of glucose concentrations before and at 30, 60, 90, and 120min after the glucose load. The glucose concentration was determined using a Bayer Contour TS blood glucose meter with matched test strips.
Determination of serum insulin concentration
Serum insulin concentration was determined using a competitive ELISA kit following the manufacturer’s instructions. The index of homeostasis model assessment of insulin resistance (HOMA-IR) was calculated according to the following formula: (fasting serum glucose × fasting serum insulin)/22.5 (Zhao et al. 2010).
Measurement of hepatic glycogen
The content of hepatic glycogen was determined by a commercially available assay kit based on the anthrone reagent method (Jung et al. 2004). In brief, glycogen was precipitated with anhydrous ethanol and dehydrated by 98% H2SO4 in order to generate furfural derivatives, which may produce a blue compound after reacting with anthrone. The optical density of the blue compound was measured at 620nm using an Agilent 8453 UV/Visible spectrophotometer (Agilent Technologies). Glycogen content of the sample was determined according to the standard curve generated by standard glucose solutions.
Western blot analysis
Tissues were homogenized in ice-cold RIPA buffer (50mM Tris, pH 7.4, 150mM NaCl, 1% Nonidet P-40, 0.1% SDS, 1mM EDTA, 1mM PMSF, 1mM Na3VO4, 25mM NaF, 10μg/mL leupeptin, 10μg/mL aprotinin). The supernatants were collected by centrifugation at 10,000g for 10min at 4°C, and protein concentrations were determined by BCA protein assay. Equal amounts (30–50μg) of protein were loaded on 10% SDS–polyacrylamide gel, separated, and transferred onto polyvinylidene difluoride membranes. The blots were probed with rabbit anti-phospho Akt (Ser473) (1:1000) and rabbit anti-Akt (1:500), followed by HRP-conjugated goat anti-rabbit IgG (1:10,000). Signals were visualized by enhanced chemiluminescence detection.
Liver tissues were homogenized 1:10w/v in ice-cold lysis buffer (20mM Tris, pH 7.5, 150mM NaCl, 1% Triton X-100) supplemented with protease and phosphatase inhibitors. Homogenates were centrifuged at 10,000g for 10min to remove cellular debris, and supernatants were precleared with Protein A/G plus-agarose beads. Two micrograms of rabbit anti-IRβ monoclonal antibody or rabbit anti-IRS2 antibody was incubated with 50μL Protein A/G beads overnight at 4°C with constant mixing. Antibody-coated beads were washed 3 times with lysis buffer to remove uncoupled antibody. The supernatants (2000μg protein) were added to the antibody-coated beads and incubated overnight at 4°C with constant mixing. The beads were washed with lysis buffer, resuspended in Laemmli reducing sample buffer, and pelleted by a brief centrifugation. The supernatants were loaded onto 10% SDS–polyacrylamide gel. Phosphorylated IRβ and IRS2 were detected by Western blotting using anti-PY antibody (1:1000).
Activities of hepatic enzymes
The activities of PI3K, GSK3α, and GS were determined by measuring NADH oxidation in the pyruvate kinase–lactate dehydrogenase coupling system following the instruction from the manufacturer’s protocols. One unit of activity is defined as the quantity of enzyme that catalyzed the oxidation of 1μmol NADH to NAD+ per min at 30°C and pH 7.5 after incubation with different substrates, which were phosphatidylinositol 4,5-bisphosphate for PI3K, RRAAEELDSRAGSPQL for GSK3α, and uridine diphosphate glucose for GS, respectively. NADH was determined by measuring spectrophotometric absorbance at 340nm. Protein content in the enzymatic extracts was determined by the Bradford protein assay. The activity was expressed as μmol NADH/min per mg protein. GPa activity was determined by evaluating the reduction of NADP+ to NADPH in a system containing phosphoglucomutase and glucose-6-phosphate dehydrogenase and expressed as μmol NADP/min per mg protein. Control experiments without substrate and with heat-denatured liver tissue extract were also performed. All of the samples were assessed in triplicate.
Data are expressed as mean±s.d. or median values (IQR) as appropriate. All boxplots show median values (horizontal lines inside the boxes), quartiles (box boundaries), maximum and minimum values (whiskers), and outliers (dots). Statistical Package for the Social Sciences (SPSS version 20.0) software was used for the statistical analysis. Comparison of continuous variables was performed using one-way ANOVA or the nonparametric analysis of variance Kruskal–Wallis test as appropriate. All subgroup analyses were performed using Bonferroni’s correction to adjust for multiple comparisons. A P-value of <0.05 was considered to be statistically significant.
Metformin alleviates liver injury in rats of NASH and cirrhosis
At the end of the 12-week experiment, all animals in the model group developed steatohepatitis and cirrhosis, and one rat died during the experiment. Macroscopic evaluation of the livers from the control rats showed red, smooth, and shiny liver tissues, whereas the livers from the model rats displayed greasy appearances with a variety of small nodules on the surfaces. The sick appearance of the liver was largely improved by the treatment of metformin (Fig. 1A). There were no significant differences in body weight among the three groups (Fig. 1B). Animals in the model group exhibited increased liver weight and spleen weight. Metformin treatment to the cirrhosis rats reduced both the liver weight and the spleen weight (Fig. 1C).
Metformin also significantly decreased serum concentrations of Tbil, ALP, and TC, and increased Alb level (Fig. 1D, E, F, G, H and I). For investigation of histopathological changes of the liver, H&E and Sirius red staining were performed with representative photomicrographs shown in Fig. 2A and B. The control rats displayed the normal liver architectures. In the model group, the structure of normal hepatic lobules was severely destroyed, featured with extensive collagen deposition, fibrous septa, and cirrhotic nodules as well as with massive fatty degeneration. The lesions were significantly alleviated by metformin. A marked decrease in the severity of fat accumulation and collagen deposition was observed in the liver tissues of the metformin-treated rats.
Metformin suppresses hepatic oxidative stress
As illustrated in Fig. 2C, the level of MDA, a marker of lipid peroxidation, in the livers of the model rats was increased significantly up to 3.13-fold over control (P<0.001). MDA level of the metformin-treated group was significantly reduced by 60.73% compared with the model group (P<0.001).
To evaluate the antioxidant capacity of the liver, the activities of SOD and GSH-Px and the level of GSH in the liver homogenates were measured. Compared with the control group, the activities of SOD and GSH-Px, as well as the GSH level, were significantly decreased in the model group (P=0.004, P=0.003, and P=0.009, respectively). Metformin significantly restored the activities of SOD (P=0.007) and GSH-Px (P=0.034) and increased the content of GSH (P=0.036) (Fig. 2D, E and F).
Metformin ameliorates glucose intolerance and insulin resistance in rats of NASH and cirrhosis
OGTT was performed to evaluate the status of glucose tolerance. As illustrated in Fig. 3A, there were no significant differences in fasting serum glucose level among the three groups. After the oral glucose load, the control rats showed a rapid elevation of blood glucose, reaching a peak level of 8.9±0.2mmol/L at 60min. The blood glucose levels of the cirrhosis model rats after the glucose load were significantly higher than those of the control rats and exhibited a maximal level of 11.3±0.8mmol/L. The impaired glucose tolerance was significantly improved by metformin treatment (Fig. 3B).
The fasting serum insulin concentration of the model rats was higher than that of the control rats (71.36±16.51mIU/L vs 34.98±18.02mIU/L, P=0.005). Metformin remarkably reduced the fasting serum insulin level to 35.43±20.59mIU/L (P=0.001) (Fig. 3C). Similarly, HOMA-IR was 9.25±4.17 in the control, 16.59±4.37 in the model, and 8.37±4.78 in the metformin-treated group (Control vs Model P=0.008 and Model vs Metformin-treated group P=0.002, respectively) (Fig. 3D).
Metformin increases glycogen storage in the liver
As metformin improved glucose intolerance and insulin resistance, we next evaluated the glycogen content in the liver. As illustrated in Fig. 4A, the glycogen content of the control rats was 64.30 (54.33, 67.35)μg/mg liver tissue. Rats in the model group exhibited remarkably decreased hepatic glycogen of 29.51 (24.99, 39.13)μg/mg. The glycogen of the metformin-treated rats was significantly increased to 53.93 (44.02, 60.54)μg/mg.
Representative images of PAS staining were shown in Fig. 4B. The livers of the control rats displayed high PAS staining intensity, which reflected the number of glycogen particles, in the cytoplasm of hepatocytes, whereas the staining intensity in the model group was dramatically attenuated. In the area of severe collagen and fat accumulation, the positive staining was barely observed. The staining intensity of glycogen in the livers of the metformin-treated rats was much stronger than that in the model rats.
Metformin upregulates hepatic IRβ expression
IR consists of two extracellular α-subunits and two transmembrane β-subunits. Figure 5A shows the immunohistochemical staining of IRβ in the liver. The normal livers displayed intensive brown staining of IRβ in the cytoplasm, whereas tissues from the model rats had little positive staining. Significantly increased IRβ staining in the metformin-treated group was observed compared with the model group. The results were consistent with the data of western blotting, which showed that the expression of IRβ in the livers of the model rats was 43.26% of control level (P<0.001). Metformin caused a significant increase in IRβ expression up to 2.02-fold over the model group (P<0.001) (Fig. 5B and F). No significant differences were detected in the expression of IRS2 among the three groups (Fig. 5C and G).
Metformin increases phosphorylations of IRβ, IRS2, and Akt
By using anti-PY antibody, IRβ and IRS2 were immunoprecipitated for the detection of tyrosine phosphorylation. As shown in Fig. 5B, C, D and E, tyrosine phosphorylations of IRβ and IRS2 in the model group were significantly decreased to 33.22 and 35.08% of control level, respectively (both P<0.001). Metformin treatment increased IRβ and IRS2 phosphorylation to 2.73-fold and 2.30-fold over the model group, respectively (P<0.001 and P=0.002).
As shown in Fig. 6A and B, the model rats exhibited 2.89-fold higher Akt expression than the control rats. Metformin caused a considerable reduction in Akt level by 51.87%. By contrast, Ser473phosphorylation of Akt in the model rats was significantly decreased to 52.21% of control level. In the metformin-treated group, Ser473-Akt was significantly increased and was 2.09-fold over the model group.
Metformin causes an increase in activities of PI3K and GS, whereas it causes a decrease in activities of GSK3α and GPa
As illustrated in Fig. 6C, D, E and F, significantly decreased activities of PI3K and GS were detected in the liver tissues of the model rats (P=0.012 and P<0.001, respectively). In the metformin-treated group, the activities of PI3K and GS were higher than that of the model group (P=0.048 and P<0.001, respectively). By contrast, the activities of GSK3α and GPa were significantly increased up to 2.07-fold and 1.81-fold over control (P=0.008 and P=0.002), respectively. Metformin treatment caused a 52.3% reduction in GSK3α activity (P=0.004) and 33.2% reduction in GPa activity (P=0.007).
To date, the influence of metformin on hepatic glycogen synthesis has been still controversial (Radziuk et al. 2003). By developing an insulin-resistant rat model of NASH and cirrhosis, this study has provided evidence that metformin increases hepatic glycogen storage. This action of metformin is attributed to the increased expression and phosphorylation of IR and the augmentation of downstream IRS2/PI3K/Akt signaling transduction pathway. To the best of our knowledge, this is for the first time indicating that metformin upregulates hepatic IR expression and the downstream IRS2/PI3K/Akt pathway in vivo.
Animal models are essential research tools for investigating the pathophysiology of hepatic insulin resistance. Dietary-induced NAFLD rodent models, which vary in diet compositions, duration of feeding, animal strain, sex, and age, are the most widely used according to the previous studies (Kucera & Cervinkova 2014). A variety of genetic models of NAFLD are more expensive and less available (Qi et al. 2005, Nagarajan et al. 2012). These models failed to show high reproducibility with regard to the development of insulin resistance, and few of them showed the progress from NASH to fibrosis even cirrhosis (Larter & Yeh 2008). During our preliminary experiments, rats fed with HFHSD for 12 weeks failed to acquire insulin resistance based on HOMA-IR assessment, whereas increased HOMA-IR was detected in those fed with HFHSD for 24 weeks. As 24-week feeding is time consuming, we attempt to create a new insulin-resistant rat model of liver disease in a shorter period of time. CCl4 is associated with increased lipid peroxidation, impaired hepatic antioxidant activity, and necrosis of hepatocytes, because it is metabolized into trichloromethyl radical (Weber et al. 2003). Available study indicated that hepatic glycogen content was considerably reduced in the rat model of cirrhosis induced by intraperitoneal injection of CCl4 for 19 weeks. However, increased early insulin secretion might compensate adequately for hepatic insulin resistance (Arai et al. 2010). CCl4 usually caused a dramatic reduction in body weight and food intake. Thus, we designed a rat model, not only presenting steatohepatitis and cirrhosis but also exhibiting insulin resistance, by HFHSD feeding in combination with subcutaneous injection of CCl4 for 12 weeks. More importantly, the remarkable weight loss and low food intake that happened in cirrhosis rats exposed to CCl4 alone were not observed. The histological observations were accompanied by increased serum TC level and dropped albumin level, whereas there was no change in ALT and AST levels, indicating that liver function was severe deteriorated. We suggest that the successful development of insulin resistance may depend on the severity of liver function injury and histopathological damage. The rats fed by HFHSD in combination with CCl4 exposure appear more susceptible to the development of insulin resistance.
Our model rats were treated with metformin at a clinically relevant concentration, and marked beneficial effects were observed. Apparently, the macroscopic appearance of steatotic and cirrhotic liver was considerably improved. Liver weight and spleen weight were decreased. Liver histological damages were significantly alleviated with serum level of TC decreased and Alb level increased. Glucose intolerance and insulin resistance were effectively improved. Moreover, metformin treatment significantly increased hepatic glycogen storage. To date, the preponderance of the literature has suggested that the major therapeutic action of metformin in the liver is on suppressing gluconeogenesis, whereas its influence on glycogenesis is likely quite small (Radziuk et al. 2003). Available data also demonstrated that metformin impaired glycogen synthesis in isolated rat hepatocytes (Otto et al. 2003). However, increased hepatic glycogen content in response to metformin was detected in high-fat-fed insulin-resistant rat model and in hepatoma cells (Purrello et al. 1988, Mithieux et al. 2002). These conflicting results may be attributable to the differences in research models and methodology.
Limited studies have evaluated the effect of metformin on hepatic insulin signaling transduction pathway. In chronic insulin-exposed HepG2 cells, the reduction in phosphorylation of IRβ and IRS2 was reversed by metformin (Yuan et al. 2003). Gunton and coworkers suggested that metformin caused an increase in tyrosine phosphorylation of the β-subunit in whole IR partially purified from primary human hepatocytes as well as in Huh7 human hepatoma cells (Gunton et al. 2003). These data were all based on in vitro studies. In our insulin-resistant rat model of NASH and cirrhosis, subdued IRS2/PI3K/Akt signaling pathway in the liver was detected. The expression of IRβ; phosphorylations of IRβ, IRS2, and Akt; and activities of PI3K and GS were significantly reduced, whereas the activities of GSK3α and GPa were increased. These molecular events were significantly reversed by metformin treatment. The most important novel finding in this study was that metformin upregulates IR expression in the steatotic and cirrhotic liver. Although increased phosphorylation of IRβ by metformin in vitro has been reported, no literature can be found showing that metformin directly affects the numbers of IR in the liver. We consider the increased activation of IRS2/PI3K/Akt signaling pathway as a consequence of the increased hepatic IR expression.
To date, considerable clinical studies have investigated the efficacy of metformin on NAFLD. It aids weight loss, improves biochemical and metabolic parameters, and reduces insulin resistance in patients with NAFLD (Haukeland et al. 2009, Zheng et al. 2015). The findings of Zhang and coworkers indicated that continuation of metformin improves the survival of NASH-related cirrhotic patients with diabetes by reducing the risk of death by 57% (Zhang et al. 2014). However, the majority of clinical evidence denied the effectiveness of this drug in ameliorating liver histological damages. Thus, metformin is not recommended as a specific treatment for NAFLD in clinical practice guidelines (Chalasani et al. 2012). By contrast, the data based on animal studies have shown that metformin improves hepatic steatosis, liver inflammation, and to a certain degree fibrosis (Lin et al. 2000, Woo et al. 2014, Tripathi et al. 2015). These results are consistent with the data of this study. In our model rats of NASH and cirrhosis, the liver damages, including steatosis and collagen deposition, were significantly improved by metformin. Interestingly, available evidence indicated that metformin alleviated liver fibrosis in CCl4 cirrhotic rats, but not in cirrhotic rats induced by common bile duct ligation (Tripathi et al. 2015). Published data also suggested that metformin prevented hepatic steatosis by regulating the expression of adipose differentiation-related protein (Liu et al. 2014). AMPK, which can be activated by metformin, is originally discovered by its ability to inhibit fatty acid and cholesterol synthesis (Carlson & Kim 1973). This study indicated that metformin suppressed the oxidative stress and restored the antioxidant capacity of the liver, which may result in improved hepatic lipid metabolism and decreased steatosis. Thus, it is reasonable for us to speculate that metformin may slow the progression of fibrosis by attenuating steatohepatitis. The improvement of liver histological alterations by metformin potentially leads to the upregulation of hepatic IR expression.
In summary, this study supports the theory that metformin improves lipid peroxidation and histological damages of the liver, alleviates glucose intolerance and insulin resistance, and increases hepatic glycogen storage. Importantly, we present novel evidence that metformin increases IR expression in the liver, resulting in the enhancement of the downstream IRS2/PI3K/Akt signaling pathway. The data are derived from in vivo studies based on animal models and enriches the existing literatures on mechanisms by which metformin improves hepatic glucose metabolism.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
This work was supported by the National Natural Science Foundation of China (81102707) and Medical Science and Technology Project of Zhejiang Province (2013KYA165).
The authors thank Professor Lv Hua for his assistance in statistical analysis.
CaoJMengSChangEBeckwith-FickasKXiongLColeRNRadovickSWondisfordFEHeL2014Low concentrations of metformin suppress glucose production in hepatocytes through AMP-activated protein kinase (AMPK). Journal of Biological Chemistry28920435–20446. (doi:10.1074/jbc.M114.567271)
ChalasaniNYounossiZLavineJEDiehlAMBruntEMCusiKCharltonMSanyalAJ2012The diagnosis and management of non-alcoholic fatty liver disease: practice Guideline by the American Association for the Study of Liver Diseases, American College of Gastroenterology, and the American Gastroenterological Association. Hepatology552005–2023. (doi:10.1002/hep.25762)
HaukelandJWKonopskiZEggesbøHBvon VolkmannHLRaschpichlerGBjøroKHaalandTLøbergEMBirkelandK2009Metformin in patients with non-alcoholic fatty liver disease: a randomized, controlled trial. Scandinavian Journal of Gastroenterology44853–860. (doi:10.1080/00365520902845268)
TripathiDMEriceELafozEGarcia-CalderoHSarinSKBoschJGracia-SanchoJGarcia-PaganJC2015Metformin reduces hepatic resistance and portal pressure in cirrhotic rats. American Journal of Physiology: Gastrointestinal and Liver Physiology309G301–G309. (doi:10.1152/ajpgi.00010.2015)