ATP synthase β subunit (ATPSβ) had been previously shown to play an important role in controlling ATP synthesis in pancreatic β-cells. This study aimed to investigate the role of ATPSβ in regulation of hepatic ATP content and glucose metabolism in diabetic mice. ATPSβ expression and ATP content were both reduced in the livers of type 1 and type 2 diabetic mice. Hepatic overexpression of ATPSβ elevated cellular ATP content and ameliorated hyperglycemia of streptozocin-induced diabetic mice and db/db mice. ATPSβ overexpression increased phosphorylated Akt (pAkt) levels and reduced PEPCK and G6pase expression levels in the livers. Consistently, ATPSβ overexpression repressed hepatic glucose production in db/db mice. In cultured hepatocytes, ATPSβ overexpression increased intracellular and extracellular ATP content, elevated the cytosolic free calcium level, and activated Akt independent of insulin. The ATPSβ-induced increase in cytosolic free calcium and pAkt levels was attenuated by inhibition of P2 receptors. Notably, inhibition of calmodulin (CaM) completely abolished ATPSβ-induced Akt activation in liver cells. Inhibition of P2 receptors or CaM blocked ATPSβ-induced nuclear exclusion of forkhead box O1 in liver cells. In conclusion, a decrease in hepatic ATPSβ expression in the liver, leading to the attenuation of ATP-P2 receptor-CaM-Akt pathway, may play an important role in the progression of diabetes.
Introduction
In past decades, diabetes became one of the major diseases threatening our health, with an estimated global prevalence of 6.4% in 2010 (1). The liver is the key tissue that releases glucose into the circulation in fasting status, and an increase in hepatic glucose production due to insulin deficiency or insulin resistance is the central event in the development and progression of type 1 or type 2 diabetes (2). Moreover, the liver is also one of the key tissues for lipid metabolism.
ATP serves as an energy molecule as well as a signal molecule in many cells (3). We previously showed that leucine upregulates the ATP synthase β (ATPSβ) subunit to enhance ATP synthesis and insulin secretion in rat islets, type 2 diabetic human islets, and Rattus insulin-1 (INS-1) cells. Overexpression or knockdown of ATPSβ increases or reduces cellular ATP levels in INS-1 cells (4,5). Overall, these findings suggest that ATPSβ plays a crucial role in controlling ATP synthesis.
ATP content was significantly reduced in the livers of streptozocin (STZ)-induced type 1 diabetic mice, high-fat diet (HFD)-induced type 2 diabetic mice, and a methionine- and choline-deficient diet-induced nonalcoholic steatohepatitis rats (6–8). Hepatic ATP content was also decreased in insulin-resistant and type 2 diabetic patients (9,10). A decrease in the hepatic ATP-to-adenosine 5´-monophosphate (AMP) ratio stimulates food intake by signals via the vagus nerve to the brain, which may lead to obesity and insulin resistance in humans (11). In obese healthy subjects, hepatic ATP content was inversely related to BMI, decreasing steadily with increasing BMI (12). Hepatic knockdown of PEPCK, one of the key gluconeogenic enzymes, ameliorated hyperglycemia and insulin resistance with increased hepatic ATP content in db/db mice (13). Overall, all of these studies revealed that a decrease in ATP content is associated with an increase in glucose production in the livers of diabetic animals and humans. Given the crucial role of ATPSβ in controlling mitochondrial ATP synthesis (4,5), we hypothesized that its expression in the liver may be deregulated and associated with increased hepatic glucose production under diabetic conditions.
In this study, we report that ATPSβ expression is reduced and correlated with ATP content in the livers of type 1 and type 2 diabetic mice. Hepatic overexpression of ATPSβ increased cellular ATP content and suppressed gluconeogenesis, leading to the amelioration hyperglycemia of type 1 and type 2 diabetic mice.
Research Design and Methods
Construction of Adenovirus Expressing Rattus ATPSβ
Adenovirus (Ad) expressing Rattus ATPSβ was constructed using the ATPSβ cDNA coding sequence cloned from INS-1 cells in our previous study (4) with a 6xHis tag inserted in the N-terminus.
Overexpression of ATPSβ in the Livers of Diabetic Mice
To assess the effect of an HFD on hepatic ATPSβ expression, 8-week-old C57BL/6 mice were fed on a 45% HFD or a control diet (Diet #MD 45% fat and Diet #MD 10% fat, Medicience Ltd., Yangzhou, China) for 12 weeks (14). The mRNA and protein expression of ATPSβ or other genes in the livers was analyzed by real-time PCR and Western blotting assays.
For overexpression of ATPSβ in the liver, 8- to 12-week-old female db/db mice on a BKS background were chosen for the experiment. As previously described, 1.0 × 109 plaque-forming unit Ad-ATPSβ or Ad-green fluorescent protein (GFP) were injected into the db/db mice via tail vein (15,16). At the third and seventh day after virus injection, oral glucose tolerance tests (OGTTs) were performed. On the eighth day, the fed animals were killed for experimental analysis.
Male C57BL/6 mice (8 weeks old) were injected intraperitoneally for 7 consecutive days with STZ (50 mg/kg, Sigma-Aldrich) freshly dissolved in citrate buffer (pH 4.2). Control mice were injected with citrate buffer. Mice with a fasting blood glucose (FBG) level >16.4 mmol/L at 30 days after the final STZ injection were chosen for experiments. The mice were treated with Ad-ATPSβ or Ad-GFP as were the db/db mice. FBG was monitored at 3 days after the viral injection. On the fourth day, the fed animals were killed for experimental analysis. Serum was collected for determination of insulin, triglyceride, and total cholesterol levels, and the liver was taken for biochemical analysis. All procedures were approved by the Peking University Health Science Center Institutional Animal Care and Use Committee.
OGTTs
Mice were fasted for 6 h (8:00 a.m. to 2:00 p.m.) in a cage with fresh bedding before OGTT. Mice were orally administered with glucose at the dose of 3 g/kg, and blood glucose levels were monitored at 0, 15, 30, 60, 90 and 120 min after the glucose load using a Freestyle brand glucometer (Roche) by blood collected from the tail (15). The blood glucose concentrations at 0 min were determined as FBG. Blood samples were taken at 0, 15, 30, and 60 min from the tail vein for detecting insulin levels. Plasma insulin levels were measured using the Rat/Mouse Insulin ELISA kit (Millipore) (16).
Pyruvate Tolerance Tests
The protocol for pyruvate tolerance tests (PTT) was detailed elsewhere (17). In brief, db/db mice were infected with Ad-ATPSβ or Ad-GFP for 7 days, fasted for 16 h, and injected intraperitoneally with pyruvate (0.75 mg/kg body weight in saline). Blood glucose levels were measured from the tail vein at indicated times using a Freestyle glucometer.
Insulin Tolerance Tests
Immunoprecipitation and Immunoblot Assays
The immunoprecipitation was performed using Protein G Agarose beads, as described previously (20). Liver protein (200 μg) was precipitated with anti-ubiquitin or IgG, and separated by 10–12% SDS gel. Immunoblot was performed, and the membrane was developed with enhanced chemiluminescence.
Cell Culture
Human hepatocellular carcinoma (HepG2) cells were purchased from American Type Culture Collection (Rockville, MD) and cultured at 37°C in 5% CO2 and 95% air in Dulbecco’s modified Eagle’s medium. The cells were infected with 20 multiplicity of infection of Ad-GFP or Ad-ATPSβ for 48 h. For insulin stimulation, the cells were serum-starved for 12 h, followed by treatment with 100 nmol/L insulin (Novo Nordisk) for 5 min and then lysed in fresh Roth Lysis buffer. For inhibition of phosphatidylinositide 3-kinase (PI3K), the infected cells were treated with indicated concentrations of wortmannin (inhibitor of PI3K), PIK75 (inhibitor of p110α), or TGX221 (inhibitor of p110β) for 30 min before being lysed. The infected cells were treated with indicated concentrations of PPADS (antagonist of P2 receptor), U73122 (inhibitor of IP3R), and carboxypeptidase Z (CPZ; inhibitor of calmodulin [CaM]), for 30 min to 1 h before experimental assay. The lysate was centrifuged at 13,000 rpm at 4°C for 10 min, and 20–100 μg total proteins underwent Western blotting assay immediately.
Primary Hepatocyte Culture
Primary mouse hepatocytes were cultured as previously described (4). In brief, 5- to 6-week-old male C57BL/6 mice were anesthetized with 10% chloral hydrate, and a catheter was placed in the inferior vena cava. The liver was perfused with 1 mL heparin (320 μ/mL), 40 mL solution I (Krebs’s solution and 0.1 mmol/L EGTA), and 30 mL solution II (Krebs’s solution, 2.74 mmol/L CaCl2, and 0.05% collagenase I), respectively. The perfused liver was passed though a 400-μm screening size filter by flushing with RPMI 1640 medium. The hepatocytes were collected by centrifuge at 50g for 2 min. Hepatocytes were resuspended with RPMI 1640 medium and plated in six-well plates for experiments after three washes with RPMI 1640.
Immunoblotting
Cells or liver tissues were lysed in fresh Roth lysis buffer (50 mmol/L HEPES, 150 mmol/L NaCl, 1% Triton X-100, 5 mmol/L EDTA, 5 mmol/L EGTA, 20 mmol/L sodium pyrophosphate, 20 mmol/L NaF, 0.2 mg/mL phenylmethylsulfonyl fluoride, 0.01 mg/mL leupeptin, and 0.01 mg/mL aprotinin, pH 7.4). The lysate was centrifuged at 4°C at 13,000 rpm for 10 min. The protein concentration in the supernatant was determined by bicinchoninic acid assay, and 20 to 200 μg total protein were separated by 12–15% SDS-PAGE. Proteins in the gel were transferred to the Hybond-C Extra membrane (Amersham Biosciences) at 120 V for 2 h at 4°C. The membrane was washed once with Tris-buffered saline with Tween (TBST; 1.2 g/L Tris Base, 5.84 g/L NaCl, and 0.1% Tween-20, pH 7.5) before blocking in TBST containing 1% BSA at room temperature for 1 h. The membrane was incubated in 1:200–1:1,000 primary antibodies at 4°C overnight. The membrane was washed five times with TBST and incubated in 1:5,000 peroxidase-conjugated secondary antibodies at room temperature for 1 h before washing as above, and then developed with ECL. After immunoblotting assays of target proteins, the membrane was stripped with 0.2 N NaOH and reprobed for EIF5 or another housekeeping protein as loading control. Anti-ATPSβ, phosphorylated (p)Akt, Akt, forkhead box class O1 (FOXO1), Fas, and sterol responsive element binding protein (SREBP1) and its mature form of SREBP1 antibodies were from Abcam. Anti-ATPSα, 6His-tag, G6Pase, ubiquitin, EIF5, and β-actin antibodies were from Santa Cruz Biotechnology. Ant-ATPSγ and ATPSd antibodies were from BioWorld. Anti-COXIV antibodies were from Invitrogen (Medical & Biological Laboratories). Peroxidase-conjugated secondary antibodies were from Santa Cruz Biotechnology. Activation of Akt was evaluated by analyzing phosphorylation at Ser473 site.
Quantitative Real-Time PCR Assay
Total RNA of tissues and cells were extracted with RNApure High-purity Total RNA Rapid Extraction Kit (BioTeke Corporation). Quantification of target gene expression was performed by quantitative real-time PCR. The complementary DNA was synthesized with the use of RevertAid First Strand cDNA Synthesis Kit (Fermentas K1622). Target gene mRNA level was normalized to that of β-actin in the same sample, as detailed previously (21). Each sample was measured in duplicate or triplicate in each experiment. The melting curve for each PCR product was analyzed to ensure the specificity of the amplification product. All of the primer sequences for the real-time PCR assays are listed in Supplementary Table 1.
ATP Content Determination
ATP content was assessed by the bioluminescence method using ATP-Lite Assay Kit (Vigorous Biotechnology Beijing Co., Ltd.). For hepatic ATP content determination, 1 mL lysis buffer was added per 100 mg frozen tissue. For cellular ATP content determination, 300 μL lysis buffer was added to the 35-mm dish. For extracellular ATP content determination, HepG2 cell culture medium was collected. The lysate ATP content or the ATP content in the medium was measured by luminometric determination of the luciferin-luciferase reaction. For determination of relative ATP level in the cells, the ATP content values (nmol) were first normalized to the protein mount (nmol/mg protein) in the same sample and then normalized to the control values. For determination of relative ATP level in the medium, the absolute concentration was determined and normalized to the control value.
Cellular Calcium Measurement
HepG2 cells seeded on coverslips were loaded with 1 μmol/L Fura-2 acetoxymethyl ester (Molecular Probes) for 10 min and then imaged under an Olympus IX71 fluorescence microscope. The emission intensities under 340 and 380 nm illumination were recorded every 1 s, with the ratio of the emission densities (F340-to-F380) reflecting the intracellular free calcium concentration (5). For basal calcium concentration measurement, infected HepG2 cells were treated with indicated concentrations of PPADS and suramin for 1 h.
Mitochondria Isolation
Mitochondria were isolated from liver tissue or HepG2 cells using the Mitochondria/Cytosol Fractionation Kit (Pierce) according to the manufacturer’s protocol. In brief, liver tissue or HepG2 cells were homogenized in Mito-Cyto extraction buffer provided by the kit, and the lysate was centrifuged twice at 800g for 5 min to pellet the nucleus and cell debris. The supernatant was collected and centrifuged at 10,000g for 10 min to pellet mitochondria, which was washed three times with lysis buffer to clean contaminated cytoplasmic proteins (22).
Immunofluorescence
To detect the translocation of FOXO1, HepG2 cells were infected with Ad-GFP or Ad-ATPSβ for 48 h and treated with of PPADS (50 μmol/L) and CPZ (100 μmol/L) for 1 h before been fixed with 4% paraformaldehyde for 15 min and permeabilized with 0.5% Triton X-100 for 10 min. Nonspecific binding was blocked in 1% BSA at room temperature for 10 min, followed by incubation with anti-FOXO1 antibody overnight at 4°C. The slides were then incubated with Alexa Fluor 594 antibodies (1:200). Images were obtained using confocal microscopy.
Statistical Analysis
Data are presented as mean ± SEM. Statistical significance of differences between groups was analyzed by unpaired Student t test or by one-way ANOVA when more than two groups were compared.
Results
ATPSβ Expression Was Reduced in the Livers of db/db and HFD-Fed Diabetic Mice
To determine the potential role of ATPSβ in the progression of hepatic insulin resistance and type 2 diabetes, its expression in the livers of db/db and HFD-fed diabetic mice was analyzed. The mRNA (Fig. 1A) and protein (Fig. 1B) levels of ATPSβ were significantly reduced in the livers of db/db mice. In contrast, the mRNA levels of the α, γ, δ, and ε subunits of the F1 domain, and the b, c, and d subunits of the F0 domain were not significantly different between db/db and db/m mouse livers (Supplementary Fig. 1A). The protein levels of the α and γ subunits of the F1 domain, and the d subunits of the F0 domain were confirmed to remain unchanged between db/db and db/m mouse livers (Supplementary Fig. 1B). Hepatic ATP content was also reduced in db/db mice compared with db/m mice (Fig. 1C). The mRNA and protein levels of ATPSβ and the ATP content were also reduced in the livers of HFD-fed diabetic mice (Fig. 1D–F). Similarly, the mRNA levels of the α, γ, δ, ε, b, c, and d subunits of ATP synthase were not significantly different between HFD and normal chow mouse livers (Supplementary Fig. 1C). The protein levels of the α, γ, and d subunits were also confirmed to remain unchanged in HFD-fed diabetic mouse livers (Supplementary Fig. 1D).
Hepatic Overexpression of ATPSβ Ameliorated Hyperglycemia in db/db Mice
The efficacy of Ad-ATPSβ treatment on the cellular ATPSβ protein level was firstly determined in HepG2 cells (Supplementary Fig. 2). Then, ATPSβ was overexpressed in the livers of db/db mice via tail vein injection of Ad-ATPSβ to evaluate its effect on ATP content and glucose metabolism. At 3 days after the virus injection, Ad-ATPSβ–treated mice exhibited significant improvement of fasting hyperglycemia and glucose intolerance compared with Ad-GFP–treated mice (Fig. 2A and B). At 7 days after the virus injection, fasting hyperglycemia and glucose intolerance were markedly attenuated in Ad-ATPSβ–treated mice compared with Ad-GFP–treated mice (Fig. 2C). The FBG of mice at 0, 3, and 7 days after the viral injection is shown in Fig. 2D. At day 8 after the virus injection, the serum insulin levels in Ad-ATPSβ–treated mice were significantly lower than those in control mice (Fig. 2E). ATPSβ overexpression also reduced hepatic lipid deposition in db/db mice (Supplementary Fig. 3).
ATPSβ Overexpression Activated Akt and Repressed Gluconeogenic Genes in the Livers of db/db Mice
Ad-ATPSβ treatment increased the ATPSβ protein level in the livers of db/db mice (Fig. 3A). Moreover, overexpressed ATPSβ was predominantly located in mitochondria (Fig. 3B). In contrast, the injection of Ad-ATPSβ had little effect on the mRNA and protein levels of ATPSβ in skeletal muscle and the pancreas, indicating the specificity of ATPSβ overexpression in the liver (Supplementary Fig. 4). ATPSβ overexpression elevated cellular ATP content, increased phosphorylated Akt (pAkt) levels, and reduced the protein level of gluconeogenic gene G6Pase in the livers of db/db mice (Fig. 3C–E). The mRNA levels of PEPCK and G6Pase in the livers were significantly repressed after ATPSβ overexpression (Fig. 3F). In vivo, OGTT, insulin secretion, ITT, and PTT assays indicated there was no significant difference between saline- and Ad-GFP–treated mice (Supplementary Figs. 5 and 6). Moreover, ATPSβ overexpression failed to affect insulin secretion in db/db mice (Supplementary Fig. 5D and E). In vitro, Ad-GFP infection also failed to significantly affect Akt activation in HepG2 cells (Supplementary Fig. 5F). These findings suggested that the virus itself, at the dose used in the current study, has little effect on glucose metabolism and Akt activation in mouse liver cells. ITT assays indicated that global insulin sensitivity was improved by hepatic overexpression of ATPSβ (Supplementary Fig. 6A and B). The PTT assays revealed that hepatic glucose production was repressed after ATPSβ overexpression (Supplementary Fig. 6C and D). Overall, these findings were consistent with the decrease in gluconeogenic genes and the attenuation of hyperglycemia in db/db mice after hepatic ATPSβ overexpression.
ATPSβ Activated PI3K-Akt Signaling Pathway Through the P2 Receptor
Because Akt plays a crucial role in controlling glucose metabolism, the effect of ATPSβ on Akt activation was further analyzed in HepG2 cells and primary cultured mouse hepatocytes. In HepG2 cells, immunofluorescence staining and immunoblotting assays indicated that Ad-ATPSβ treatment increased the ATPSβ protein level in the mitochondria (Supplementary Figs. 7 and 8). ATPSβ overexpression significantly elevated intracellular and extracellular ATP levels in HepG2 cells (Fig. 4A) and in primary cultured mouse hepatocytes (Fig. 4B). In HepG2 cells infected by Ad-GFP, insulin potently stimulated Akt activation (Fig. 4C). Ad-ATPSβ treatment markedly elevated pAkt levels (Fig. 4C) compared with Ad-GFP–treated cells without insulin stimulation. ATPSβ overexpression also augmented insulin-stimulated Akt activation in HepG2 cells (Fig. 4C). Similarly, ATPSβ also induced Akt activation in primary cultured mouse hepatocytes in the absence of insulin stimulation (Fig. 4D). ATPSβ-induced Akt activation was completely blocked by the PI3K inhibitor wortmannin in HepG2 cells in the absence of insulin stimulation (Fig. 4E). Inhibition of the PI3K p110α catalytic subunit blocked ATPSβ-induced Akt activation by ∼70–80%, whereas inhibition of the PI3K p110β catalytic subunit had little effect on ATPSβ-induced Akt activation in the absence of insulin (Fig. 4E). Overall, these results suggested that ATPSβ-induced Akt activation requires the activity of the PI3K p110α catalytic subunit but is independent of insulin.
ATP can be released to function as a signal molecule in various cell types (23,24). In HepG2 cells, ATPSβ-induced Akt activation was significantly attenuated by the ATP receptor P2 receptor antagonist PPADS (Fig. 5A). Pretreatment with suramin, another P2 receptor antagonist, also attenuated ATPSβ-induced Akt activation (Supplementary Fig. 9A). Furthermore, inhibition of the P2 receptor downstream molecule phospholipase C (PLC) using U73122 also attenuated ATPSβ-induced Akt activation (Fig. 5B). ATPSβ overexpression elevated the cytosolic free calcium level, which was inhibited by PPADS and suramin in HepG2 cells (Fig. 5C). Because an increase in cytosolic free calcium level has been shown to activate the PI3K/Akt signaling axis through CaM in several cell types (25,26), we further analyzed whether ATPSβ induced Akt activation by the calcium-CaM signaling pathway. The results indicated that depletion of extracellular calcium partially attenuated (Supplementary Fig. 9B), whereas inhibition of CaM using CPZ completely abolished, ATPSβ-induced Akt activation in HepG2 cells (Fig. 5D).
Inhibition of P2 Receptor or CaM Blocked ATPSβ-Induced Translocation of FOXO1
Activation of Akt plays a crucial role in suppressing gluconeogenesis by phosphorylating and inactivating FOXO1, the key transcriptor controlling the transcription of gluconeogenic genes PEPCK and G6Pase (27). As expected, ATPSβ overexpression significantly promoted nuclear exclusion of FOXO1 with the activation of Akt in HepG2 cells (Fig. 6). In support, inhibition of P2 receptors or CaM blocked ATPSβ-promoted nuclear exclusion of FOXO1 with the attenuation of Akt activation (Fig. 6).
Hepatic Overexpression of ATPSβ Ameliorated Hyperglycemia of STZ-Induced Type 1 Diabetic Mice
In the livers of STZ-induced type 1 diabetic mice, the mRNA and protein levels of ATPSβ expression was also significantly reduced with a decrease in ATP content (Fig. 7A–C). Hepatic overexpression of ATPSβ for 3 days elevated cellular ATP content and ameliorated fasting hyperglycemia in STZ-induced type 1 diabetic mice (Fig. 7C–E). ATPSβ overexpression increased pAkt levels and reduced mRNA and protein levels of PEKCK and G6Pase in the livers of STZ-treated mice, suggesting that hepatic gluconeogenesis was inhibited (Fig. 7F–H).
Discussion
In the livers of type 1 and type 2 diabetic mice, ATPSβ expression was reduced and positively correlated with ATP content. Hepatic overexpression of ATPSβ elevated cellular ATP content, suppressed hepatic glucose production, and improved hyperglycemia, hyperinsulinemia, insulin resistance, and fatty liver in db/db mice. Hepatic overexpression of ATPSβ also elevated cellular ATP and ameliorated hyperglycemia in STZ-induced diabetic mice. ATPSβ overexpression increased pAkt levels and repressed the expression of gluconeogenic genes in the livers of these diabetic mice. These findings suggested that inhibition of hepatic glucose production would attenuate hyperglycemia and global insulin resistance in type 2 diabetic mice. Similarly, a previous report showed that silencing of hepatic PEPCK suppressed hepatic glucose production and improved global insulin resistance in db/db mice (13). In vitro, ATPSβ overexpression increased intracellular and extracellular ATP levels in HepG2 cells and primary cultured mouse hepatocytes. In support of our previous (4,5) and current findings that ATPSβ plays a crucial role in controlling ATP synthesis, hydrogen peroxide induces ATPSβ expression and increases cellular ATP level in melanocytes (28). Proteomic analysis revealed that ATPSβ expression was reduced, with a decrease in ATP content in mouse livers after ischemia-reperfusion treatment (29). No change of other subunits of ATP synthase was found in mouse livers injured by ischemia-reperfusion in the same study (29).
Extracellular ATP can activate the PI3K-Akt pathway through P2 receptors in several cell types. Ishikawa et al. (30) found that pannexin 3 releases ATP into extracellular space, which activates the PI3K/Akt pathway through P2 receptors to promote the proliferation of osteoblasts. Electrical pulses can stimulate skeletal muscle cells to release ATP, which enhances glucose uptake by activation of the PI3K-Akt pathway through P2 receptors (31). Extracellular ATP also activates the PI3K-Akt pathway to attenuate ischemia-induced apoptosis of human endothelial cells (32). P2 receptors were divided into P2X receptors and P2Y receptors. P2X receptors are ligand-gated ion channels that are permeable for calcium (33), whereas P2Y receptors are G protein–coupled receptors that stimulate PLC to increase IP3, resulting in calcium release from internal storage (23). Liver cells and HepG2 cells have been shown to release ATP. P2 receptors are also constitutively expressed in liver cells and HepG2 cells (24,34). Exposure to exogenous ATP has been reported to elevate the cytosolic free calcium level in HepG2 cells and primary cultured mouse hepatocytes (35). PPADS is a specific inhibitor for the P2X subtype, whereas suramin is considered to be a broad-spectrum P2 receptor inhibitor (36). PPADS and suramin both significantly attenuated the ATPSβ-induced increase in the cytosolic free calcium level and Akt activation in HepG2 cells. Depletion of extracellular calcium also attenuated ATPSβ-induced Akt activation. Furthermore, U73122, an inhibitor of the P2Y receptor downstream molecule PLC, also attenuated ATPSβ-induced Akt activation in HepG2 cells. These results revealed the involvement of both P2X and P2Y receptors in ATPSβ-induced increase in cytosolic free calcium level and Akt activation in liver cells.
Cytosolic free calcium level is the crucial second messenger in the ATP/P2 receptor–signaling pathway. An increased cytosolic free calcium level activates CaM, which can activate PI3K by direct association with its p85 regulatory subunit, leading to activation of Akt (37). Importantly, inhibition of CaM using CPZ completely abolished ATPSβ-induced Akt activation in HepG2 cells. ATPSβ-induced Akt activation requires the activity of the PI3K p110α catalytic subunit but is independent of insulin signaling. These findings were consistent with recent discoveries that p110α predominantly mediated Akt activation in liver cells. Deletion of p110α blunted Akt activation in response to insulin, whereas deletion of p110β had little effect on insulin-stimulated Akt activation in the liver (38). Furthermore, ATPSβ overexpression reduced hepatic and serum triglyceride content with little effect on liver and serum cholesterol levels in db/db mice. The protein levels of key lipogenic genes FAS and mature SREBP1 were reduced in the livers of db/db mice after ATPSβ overexpression (Supplementary Fig. 3D). FOXO1 is a crucial transcriptor controlling the transcription of gluconeogenic enzymes PEPCK and G6Pase (27). Insulin-stimulated activation of Akt can phosphorylate FOXO1 and promote its translocation from the nucleus to the cytoplasm, suppressing the transcription of gluconeogenic genes and hepatic gluconeogenesis (27). ATPSβ overexpression promotes nuclear exclusion of FOXO1, which can be inhibited by antagonisms of the P2 receptor or CaM. Overall, these findings revealed that ATPSβ overexpression elevated ATP synthesis and secretion in liver cells. Released ATP activates both P2X and P2Y receptors to increase the cytosolic free calcium level, which activates the CaM-PI3K-Akt signaling axis to suppress gluconeogenesis independent of insulin. However, it should be noted that elevated intracellular ATP levels will inhibit the activity of the ATP-sensitive K+ channel and increase the influx of extracellular calcium through the L-type voltage-gated calcium channel (39,40), which may also contribute to ATPSβ-induced increase in free cytosolic calcium and Akt activation. Interestingly, although we found that ATPSβ overexpression for 7 days significantly reduced the mRNA level of PEPCK in db/db mouse livers, its protein level remained unchanged (Fig. 3D–F, Supplementary Fig. 10A–C). The decrease in the PEPCK mRNA level was confirmed by two different sets of primers according to the mRNA sequence of PEPCK (Supplementary Fig. 10C). At 3 days after ATPSβ overexpression, the mRNA and protein levels of PEPCK were both significantly reduced in the livers of db/db mice (data not shown) and STZ-induced diabetic mice (Fig. 7). A coimmunoprecipitation assay revealed that the ubiquitination of the PEPCK protein was reduced at 7 days after ATPSβ overexpression, suggesting that the ubiquitin-mediated degradation of PEPCK protein might decrease (Supplementary Fig. 10D). A decrease in PEPCK protein degradation might be a protective mechanism against persistent suppression of hepatic glucose production induced by long-term ATPSβ overexpression.
That excessive accumulation of free fatty acids (FFAs), in particular, saturated FFAs such as palmitate, plays an important role in the progression of insulin resistance in various tissues has been intensively studied (41). Højlund et al. (42,43) found that the ATPSβ protein level is reduced in skeletal muscle of insulin-resistant and type 2 diabetic patients. In vivo, FFAs induced insulin resistance in skeletal muscle cells with a decrease in ATP generation (44). Rooibos ameliorates palmitate-induced insulin resistance in C2C12 muscle cells with increased cellular ATP levels (45). In support, inhibition of ATP synthase using oligomycin reduces cellular ATP levels and induces insulin resistance in cultured human myotubes (46). In the current study, we found that palmitate downregulated ATPSβ and induced insulin resistance in HepG2 cells, which was reversed by overexpression of ATPSβ (Supplementary Fig. 11). Palmitate has been shown to induce insulin resistance by stimulating reactive oxygen species (ROS) production in hepatocytes (47). We further found that ATPSβ overexpression reduced palmitate-induced ROS production in HepG2 cells (Supplementary Fig. 12). ATPSβ also tended to reduce ROS production in basal conditions in HepG2 cells (Supplementary Fig. 12). In a lipid stressed condition, an increase in electron transfer on the respiratory chain as well as a decrease in ATPSβ expression may contribute to ROS overproduction and oxidative stress, which is one of the main causes of hepatic insulin resistance (47,48). These findings suggest that repression of ATPSβ expression, leading to a decrease in ATP synthesis and a possible increase in ROS production, might be a novel mechanism for lipid-induced insulin resistance in liver and muscle cells.
Activation of AMP-activated protein kinase (AMPK) exerts beneficial effects on amelioration of hyperglycemia (49). The activity of AMPK is modulated by various factors such as an increase in the AMP-to-ATP ratio, adiponectin, and metformin (49). The antidiabetic drug metformin can target mitochondrial complex I to inhibit ATP production (50,51), leading to the activation of AMPK (52), which at least partially mediates metformin’s hypoglycemic effects. However, some deleterious effects of AMPK activation, including inhibition of insulin secretion (53), induction of pancreatic β-cell, and renal medullary interstitial cell apoptosis (54,55), have also been reported. Given the role of ATP as the key energy storage molecule and an important signaling molecule (30–32), activation of ATPSβ-ATP-P2 receptor signaling may have some unique advantages than activation of metformin-AMPK signaling in the treatment of type 2 diabetes.
In summary, the current study demonstrated that ATPSβ expression is positively correlated with ATP content in the livers of diabetic mice. A decrease in ATPSβ expression in the liver, leading to the reduction of ATP content and attenuation of the P2 receptor-PI3K-Akt signaling pathway, may play an important role in the progression of diabetes. Upregulating hepatic ATPSβ might be an attractive method for the treatment of type 1 and type 2 diabetes (Fig. 8).
Article Information
Funding. This study was supported by grants from the Ministry of Science and Technology (2012CB517504/2011ZX09102), the Natural Science Foundation of China (81170791/30870905/81200625/81322011), and the Beijing Natural Science Foundation (7122107).
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Author Contributions. C.W., Z.C., S.L., Y.Z., S.J., and Y.M. researched data and contributed to discussion. C.W. wrote the manuscript. J.L. and Y.C. reviewed and edited the manuscript. C.W., Y.G., and J.Y. designed the study and revised and edited the manuscript. J.Y. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.