Plasma 5′-AMP (pAMP) is elevated in mouse models of type 2 diabetes. However, the metabolic regulatory role of adenine nucleotides in type 2 diabetes remains unclear.
Adenine nucleotides and their metabolites in plasma and liver were examined by HPLC. 1H NMR-based metabolomics analysis was performed to explore the changes of metabolites in mouse models of type 2 diabetes. Na+/K+ ATPase and Na+/H+ exchanger activity were measured in response to adenine nucleotide metabolites. Human recombinant protein tyrosine phosphatase 1B (PTP1B) was used for enzyme kinetic assays. Protein binding assays were performed with microscale thermophoresis. The intracellular pH of hepatocyte AML12 cell lines was measured using the BCECF-AM method. We also analysed pAMP levels in participants with type 2 diabetes.
Elevation of pAMP was a universal phenomenon in all mouse models of type 2 diabetes including db/db vs lean mice (13.9 ± 2.3 μmol/l vs 3.7 ± 0.9 μmol/l; p < 0.01), ob/ob vs lean mice (9.1 ± 2.0 μmol/l vs 3.9 ± 1.2 μmol/l; p < 0.01) and high-fat diet/streptozotocin-induced vs wild-type mice (6.6 ± 1.5 μmol/l vs 4.1 ± 0.9 μmol/l; p < 0.05); this elevation was required for the occurrence of hyperglycaemia in obese mice. 1H NMR-based metabolomics study following HPLC analysis revealed that the metabolite profile in wild-type mice treated with 5′-AMP was similar to that in db/db diabetic mice, especially the accumulation of a large quantity of ATP and its metabolites. The glucose-lowering drug metformin reduced the severity of hyperglycaemia both in 5′-AMP-induced wild-type mice and db/db mice. Metformin decreased the accumulation of liver ATP but not its metabolites in these hyperglycaemic mice. ATP and metformin reciprocally change cellular pH homeostasis in liver, causing opposite shifts in liver activity of PTP1B, a key negative regulator of insulin signalling. Furthermore, pAMP levels were also elevated in individuals with type 2 diabetes (45.2 ± 22.7 nmol/l vs 3.1 ± 1.9 nmol/l; p < 0.01).
These results reveal an emerging role for adenine nucleotide in the regulation of hyperglycaemia and provide a potential therapeutic target in obesity and type 2 diabetes.
Endogenous glucose production, including hepatic glycogenolysis and gluconeogenesis, is essential for survival during starvation . On the other hand, excessive glucose production contributes to hyperglycaemia in individuals with diabetes . In type 2 diabetes, hyperglycaemia occurs due to insulin insensitivity in main metabolic organs such as liver and skeletal muscle . Obesity is a major risk factor for type 2 diabetes . Mice deficient in leptin and leptin receptors (ob/ob and db/db) exhibit severe, spontaneous, early-onset obesity, with a body weight three times that of wild-type mice , and display type 2 diabetes-like characteristics such as insulin insensitivity, glucose intolerance and hyperinsulinaemia . Both db/db and ob/ob mice develop hyperglycaemia after the age of 4 weeks [7, 8]. However, with increasing age, blood glucose levels of ob/ob mice become nearly normal even in the presence of obesity and severe insulin resistance [8, 9]. Similarly, most high-fat diet (HFD)-induced obese mice maintain non-diabetic glucose levels and present insulin resistance only . In humans, although most people who develop type 2 diabetes are overweight or obese, most obese individuals do not become diabetic even though they present with a high degree of insulin resistance .
The failure of pancreatic beta cells to compensate for insulin resistance is thought to release the suppression of gluconeogenesis in the liver and cause hyperglycaemia . However, it is not possible to determine separately the effect of hyperglycaemia and dyslipidaemia on beta cell differentiation status in diabetic mice or in pathological samples from human donors with diabetes . Whether additional factors are acquired secondary to changes in the metabolic milieu associated with the progression to type 2 diabetes remains elusive. It is well known that insulin receptor tyrosine kinase is directly involved in intracellular insulin signalling transduction  and that protein tyrosine phosphatase 1B (PTP1B) negatively regulates insulin receptor phosphorylation . It has been observed that the activity of insulin receptor tyrosine kinase does not change in db/db diabetic mice . Inhibition of PTP1B is demonstrated to improve insulin sensitivity in obese and insulin-resistant ob/ob mice and reduce the hyperglycaemia in obese db/db mice .
Metformin is used extensively to treat type 2 diabetes. It acts through multiple mechanisms such as increasing phosphorylation of insulin receptors, stimulating translocation of GLUT4 to the cell membrane, decreasing gluconeogenesis in hepatocytes and activation of AMP-activated protein kinase (AMPK) [17,18,19,20]. Other research indicates metformin inhibits the respiratory chain that restrains hepatic gluconeogenesis .
Our previous observation indicates that plasma 5′-AMP (pAMP) was elevated in plasma of obese db/db diabetic mice and that exogenous 5′-AMP caused hyperglycaemia-like type 2 diabetes in wild-type mice . We are interested in understanding the role of adenine nucleotides in metabolic regulation, insulin resistance and hyperglycaemia occurring in type 2 diabetes. In this study, we have investigated the role of adenine nucleotides in regulating the activity of PTP1B in type 2 diabetes. Our studies provide insight into the participation of adenine nucleotides in obese diabetic mice.
Mice were obtained from SLACCAS (Shanghai, China). Male C57BL/6, C57BL/6 ob/ob, C57BL/Ks db/db mice and their homozygous lean littermates (+/+) were used. To create diet-induced insulin resistance models, 4-week-old male C57BL/6 mice were fed an HFD. HFD-fed mice were injected intraperitoneally with streptozotocin (STZ) to induce diabetes. All procedures were approved by the Animal Care and Use Committee at Nanjing University of Science and Technology (NJUST-AC-05-1217). See electronic supplementary material (ESM) Methods for further details.
Injection of 5′-AMP and metformin and ITT
Male 8-week-old C57BL/6 mice were injected intraperitoneally with 5′-AMP at 09:00. Metformin hydrochloride was administered by gavage. Mice were killed 1 h after 5′-AMP treatment. For ITT, mice were fasted and injected intraperitoneally with insulin. See ESM Methods for details.
Tissue and blood sampling
Mice were euthanised, blood was collected and then tissues were removed and freeze-clamped in liquid nitrogen. Erythrocytes and plasma were separated. See ESM Methods for further details.
HPLC analysis of nucleotides and metabolites
Mouse livers were homogenised using ice-cold 0.4 mol/l perchloric acid. Samples were analysed using HPLC (Waters 1525 System; Millipore, Bedford, MA, USA) on a reversed-phase C18 column as described previously [23,24,25]. Pure nucleotides and metabolites were used to identify the peaks and obtain the calibration curves. Quantification was based on the area under the peaks.
1H NMR-based metabolomics analysis of tissue extraction
Tissues were extracted using ice-cold 50% acetonitrile. The spectra were acquired on a Bruker AV 500-MHz spectrometer (Bruker, Karlsruhe, Germany). The integrated area was normalised. Unsupervised principal component analysis methods were performed. According to the integrated areas of metabolites, the increased or decreased factors were calculated. See ESM Methods for details.
Na+/K+ ATPase assay for hepatic plasma membrane
Liver plasma membrane was prepared as described previously . Na+/K+ ATPase assay was performed as previously described, with modification . The reaction was started by adding ATP. Released inorganic phosphate was measured. Na+/K+ ATPase activity was calculated by comparison with the ATPase activity inhibited by ouabain. See ESM Methods for details.
Quantitative real-time PCR
Total RNA from mouse liver was isolated and quantitative real-time PCR was used to determine relative mRNA of mouse Pepck (also known as Pck1) and G6pase (also known as G6pc). See ESM Methods for details; primers are listed in ESM Table 1.
Glucose-6-phosphatase and PEPCK activity measurement
Treatment of db/db mice with polyethylene glycol-modified adenosine deaminase
Polyethylene glycol-modified adenosine deaminase (PEG-ADA) was prepared with recombinant adenosine deaminase and mPEG-succinimidyl propionate (mPEG-SPA; Chemgen Pharma, Beijing, China). Obese db/db mice were injected with PEG-ADA prior to experiments. See ESM Methods for details.
Preparation of metformin
Metformin hydrochloride (20 g) was dissolved in 1 litre of methanol containing 2% vol./vol. water. The solution was then passed through a column containing anion exchange resin Duolite A101D (Jinjinle Chemical Co., Changsha, China). The column was then rinsed with 1 litre of methanol containing 2% vol./vol. water. The elute solution was evaporated to dryness under reduced pressure at 40°C to obtain metformin, which was used immediately.
PTP1B activity and kinetic analysis
PTP1B activity in liver extract was determined as described previously, with modification . Liver was homogenised in cold PBS. The supernatant fractions were diluted and reaction was started by adding p-nitrophenyl phosphate. For kinetic analysis, recombinant PTP1B was used. See ESM Methods for details.
Microscale thermophoresis binding analyses
Microscale thermophoresis (MST) analysis was performed using the MST device (Monolith NT.115) as described previously with modification . Data were treated with NT analysis software (see ESM Methods for further details).
Intracellular pH measurement
The measurement of intracellular pH (pHi) was performed as previously described, with modification . AML12 cells were used and stained with Hanks’ buffered saline solution (HBSS) solution containing 1 μmol/l (2',7'-bis-[carboxyethyl]-5-[and-6]-carboxyfluorescein)-tetraacetoxymethyl ester (BCECF-AM) for 20 min at room temperature. Then the cells were perfused with HBSS. pHi measurements were performed at 37°C using a fluorescence microscope (Nikon, Tokyo, Japan). BCECF fluorescence was excited at 490 nm and 440 nm and emitted fluorescence was measured at 530 nm. The F490/F440 emission ratio was converted to pH scale according to the nigericin technique. See ESM Methods for details.
Na+/H+ exchanger activity
Na+/H+ exchanger activity was analysed in bicarbonate-free solution. After equilibrating, AML12 cells were acidified by adding propionate. The recovery rate of pHi was calculated from data obtained during the initial recovery period. See ESM Methods for details.
Intracellular sodium determination
AML12 cells were treated with 5′-AMP. After removal of medium, cells were washed and lysed. Intracellular sodium was estimated using atomic absorption spectrometry. See ESM Methods for details.
Vacuolar-type (H+) ATPase activity
The microsome fraction was prepared from mouse liver by differential centrifugation. Hydrolysis of ATP by microsome was analysed. Released inorganic phosphate was measured. V-ATPase activity was defined as the ATPase activity inhibited by bafilomycin A1. See ESM Methods for details.
Twenty-five individuals with type 2 diabetes and 21 healthy individuals (control group) were recruited. Blood samples were obtained in the early morning through the Second Hospital of Nanjing. Upon arrival, body mass and height measurements were taken to determine BMI, and blood samples were obtained using venepuncture. All blood samples were obtained after informed consent was obtained from both the diabetic and healthy individuals for their participation in the study. HbA1c was determined with the gold standard ion-exchange method (Tosoh G8 HPLC Analyzer, Tosoh Bioscience, San Francisco, CA, USA). Blood glucose was determined using a OneTouch Ultra blood glucose meter (Life-Scan, Milpitas, CA, USA). Plasma insulin was analysed using the Roche E170 chemiluminescent immunoassay (Roche Diagnostics). Plasma was deproteinised using ice cold 0.4 mol/l perchloric acid and then analysed using HPLC. Demographics are shown in Table 1. All procedures were approved by the Institutional Review Board of Nanjing University of Science and Technology.
Data are presented as means ± SD. Statistical analysis was performed by Student’s t test, one-way ANOVA or two-way ANOVA followed by Tukey’s post hoc test. Significance was defined as p < 0.05.
Elevated pAMP as a biomarker for type 2 diabetes in mouse models
To investigate whether elevated pAMP is an independent event in obese db/db diabetic mice, we measured pAMP levels in different mouse models of insulin resistance. In accordance with findings from a previous study, pAMP increased in db/db vs lean mice (13.9 ± 2.3 μmol/l vs 3.7 ± 0.9 μmol/l; p < 0.01; Fig. 1a). The increase in pAMP was also observed in diabetic ob/ob vs lean mice (9.1 ± 2.0 μmol/l vs 3.9 ± 1.2 μmol/l; p < 0.01; Fig. 1b), HFD-STZ mice vs wild-type mice (6.6 ± 1.5 μmol/l vs 4.1 ± 0.9 μmol/l; p < 0.05; Fig. 1c) and wild-type mice treated with 5′-AMP vs saline (154 mmol/l NaCl) (15.1 ± 2.1 μmol/l vs 4.4 ± 1.5 μmol/l; p < 0.01; Fig. 1d). Then we analysed the pAMP level in non-diabetic mice with insulin resistance. The results showed no difference in pAMP level between non-diabetic ob/ob and non-diabetic HFD mice compared with control mice (Fig. 1e,f). To understand the role of increased pAMP in metabolic regulation in type 2 diabetes, a survey of 1H NMR-based liver metabolomic data was performed in an obese db/db diabetic mouse model for insulin resistance and metabolic disorders (ESM Fig. 1). The plots generated from profiles of extracts of the insulin-sensitive liver in lean and db/db obese mice revealed significant separation of metabolites between the two groups (Fig. 1g,h), suggesting that these phenotypes exhibit significantly different metabolic patterns of known and unknown metabolites (ESM Table 2). Interestingly, when we injected 5′-AMP into the wild-type mice, the profiles of liver extracts in these mice exhibited similar metabolomic shifts to those observed in the db/db mice (Fig. 1i,j; ESM Table 2). A heat map of the 38 indicated metabolites in at least one of three pair-wise comparisons is shown in Fig. 1k. Pearson r analysis showed strong correlations between biological replicates in each group. Metabolic pathway analysis of the differential metabolite clusters revealed that livers of 5′-AMP-treated and db/db mice are enriched in metabolite accumulations including amine acid and nucleotide metabolism. Together, these results suggest that elevated pAMP is a common marker and plays a potential role in hyperglycaemia in mouse models of type 2 diabetes.
The accumulation of liver ATP and adenine nucleotide metabolites is required for hyperglycaemia in obese mice
HPLC analysis for the detection of water-soluble liver chemicals confirmed that, relative to lean mice, db/db mice exhibited significantly elevated ATP and adenine nucleotide metabolites (ANMs), including adenosine, hypoxanthine, xanthine and uric acid in the livers (Fig. 2a,b). Similarly, ATP and ANMs were also increased in the liver of 5′-AMP-treated mice compared with saline-treated control mice (Fig. 2c,d). These results indicate that abnormal adenine nucleotide metabolism occurs and liver ATP and ANMs accumulate in obese diabetic db/db mice. Next, we investigated whether abnormal adenine nucleotide metabolism occurs in other mouse models of diabetes. As observed in db/db diabetic mice, HPLC analysis showed obvious elevation of ATP and ANMs in the liver of the obese diabetic ob/ob mice (ESM Table 3), but not in non-diabetic obese ob/ob mice even in the presence of severe insulin resistance. In another high-fat STZ-induced type 2 diabetes model, we found elevated ATP and ANMs in the liver of hyperglycaemic mice but no accumulation of ATP and ANMs in liver in mice with HFD-induced obesity (ESM Table 3); the latter mice exhibited insulin resistance but had normal blood glucose levels. Interestingly, adenosine increased in the liver of all obese mice and ATP and ANMs were only elevated in the liver of hyperglycaemic mice. To investigate whether the accumulation of ANMs and ATP contributed to hyperglycaemia, we injected 5′-AMP into the wild-type mice and examined the time course of induction of hyperglycaemia and the changes in liver concentrations of ATP and ANMs. Injection of 5′-AMP induced hyperglycaemia in 0.5–1 h and the blood glucose levels returned to normal after 2 h (Fig. 2e). HPLC analysis revealed that ANMs and ATP levels in liver increased at 1 h and returned to baseline levels at 2 h after 5′-AMP injection (Fig. 2f,g). These observations suggest that the accumulation of ATP and ANMs in liver is required for the occurrence of hyperglycaemia in obese mice. Next, we investigated the potential relationship between cellular ATP and ANMs. Figure 2h–k shows the in vitro effects of ANMs on Na+/K+ ATPase activity in plasma membrane in mouse liver. Adenosine had no effect on Na+/K+ ATPase activity but ANMs such as hypoxanthine, xanthine and uric acid markedly inhibited Na+/K+ ATPase activity in a dose-dependent manner. In most animal cells, Na+/K+ ATPase is responsible for about one-fifth to two-thirds of the cell’s energy expenditure . We reasoned that impairment of Na+/K+ ATPase activity by ANMs should decrease ATP utilisation and cause a cellular ATP elevation in liver.
Metformin lowers 5′-AMP induced hyperglycaemia and decreases cellular ATP level
Metformin, a widely used glucose-lowering drug, markedly alleviated 5′-AMP induced hyperglycaemia (Fig. 3a,b). Our analysis showed that metformin decreased 5′-AMP-induced G6Pase and Pepck mRNA expression levels (Fig. 3c). Metformin also reduced 5′-AMP-related increases in G6Pase and PEPCK activity (Fig. 3d). Moreover, metformin had no effect on the activity of Na+/K+ ATPase (ESM Fig. 2). HPLC analysis revealed that metformin decreased 5′-AMP-induced ATP elevation but failed to change 5′-AMP-induced ANM accumulation in the liver (Fig. 3e,f). Similar results were also observed in the metformin-treated db/db mice (Fig. 3g,h). These results suggest that elevated liver ATP plays a potential role in inducing hyperglycaemia.
ATP is an activator of PTP1B
One key question arising from these studies is the biological significance of this abnormal purinergic metabolism in mice. Perhaps increased ATP level in the liver of obese mice is an important signal for energy utilisation related to glucose metabolism. Our previous work demonstrates that adenosine is an ATP-competitive substrate inhibitor in the phosphorylation of insulin receptors in vitro . We tested whether decreasing cellular adenosine levels alleviated hyperglycaemia in obese db/db mice. PEG-ADA was injected intraperitoneally into db/db mice to decrease liver adenosine levels. As expected, PEG-ADA obviously decreased cellular adenosine level in mouse livers (ESM Fig. 3). However, PEG-ADA failed to lower blood glucose level in db/db mice (ESM Fig. 3), suggesting that adenosine is not a key regulator of blood glucose in vivo. An unexpected finding was that ATP dose-dependently enhanced the activity of PTP1B, a known negative regulator of the insulin signalling pathway (Fig. 4a). To rule out the possibility that this was a non-specific effect that could be triggered by any nucleotide, we added various concentrations of AMP, adenosine and uric acid to mouse liver extracts but none of these molecules increased PTP1B activity (Fig. 4b). A kinetics study of the stimulatory effects of ATP on recombinant PTP1B activity was performed over the substrate concentration range of 0.25–2 mmol/l in the presence of 0.2 or 0.4 mmol/l of ATP. Lineweaver–Burk double-reciprocal plot analysis revealed that ATP increased the maximal PTP1B velocity and decreased the Km value. The Ki value (activation constant) obtained from the Dixon plot was 0.65 mmol/l (Fig. 4c–e). These studies demonstrate that ATP is an activator of PTP1B and can function as a negative feedback regulator of the insulin signalling pathway.
Metformin is an inhibitor of PTP1B
We next investigated whether the stimulatory effect of ATP on PTP1B activity can be inhibited by glucose-lowering drugs such as metformin. Because metformin must be transported into the cell as cationic metformin , we first assayed PTP1B activity with cationic metformin in vitro, and found that cationic metformin could effectively inhibit the ATP-stimulated activity of recombinant PTP1B (Fig. 4f,g). Cationic metformin also decreased the activity of the PTP1B in mouse liver extracts in a dose-dependent manner (Fig. 4h). In contrast to the effect of ATP, cationic metformin decreased the maximal PTP1B velocity and increased its Km value based on Lineweaver–Burk double-reciprocal plot analysis (Fig. 4i–k). The Ki value (inhibition constant) obtained from the Dixon plot was 0.31 mmol/l. These results indicate that cationic metformin is an inhibitor of PTP1B.
A reciprocal regulation of PTP1B activity by ATP and metformin
To further understand the effects of ATP and metformin on PTP1B activity, we investigated the interaction between PTP1B and ATP or metformin by MST assay. The fluorescence intensity of labelled PTP1B detected by the MST reader did not vary until micromolar levels of ATP or cationic metformin were present (Fig. 5a–d). This suggests a lower affinity between ATP and PTP1B. Similarly, metformin displayed a lower affinity for PTP1B. The sigmoidal binding curve observed in the presence of ATP was different from that obtained with metformin. The different shape of the fit curves can be related to pH-dependent thermophoresis type . Next, we observed that the binding curve obtained at acidic pH values was equal to that following addition of ATP and that the binding curve obtained at basic pH values was similar to that following addition of metformin (Fig. 5e–h). These data suggest that ATP and metformin might influence pH value, resulting in the variation in fluorescence intensity under the MST reader. Indeed, a precise pH assay revealed that ATP and metformin caused a reciprocal change in the pH of assay buffers (Fig. 5i,j) and that PTP1B activity was strongly dependent on pH (Fig. 5k).
Extracellular nucleotides and metformin regulate pHi homeostasis in hepatocytes
It is known that uptake of metformin hydrochloride into hepatocytes is catalysed by the organic cation transporter-1 (OCT1) . Being positively charged, the cation metformin accumulates in cells . We investigated whether exogenous metformin hydrochloride and nucleotides induced intracellular pH changes in cultured hepatocytes. After addition of metformin hydrochloride and ATP, hepatic pHi was measured from green (BCECF) images taken at 490 nm and 440 nm excitation wavelengths. Under control conditions, the typical effect of extracellular 5′-AMP on pHi in hepatocytes is illustrated in Fig. 6a. Exposure to 5′-AMP resulted in a rapid decrease of 0.45 ± 0.08 pH units; lower concentrations of 5'-AMP (10 μmol/l and 100 μmol/l) also reduced intracellular pH to different degrees (ESM Fig. 4). Extracellular metformin hydrochloride caused pHi to rise (Fig. 6b). Cytoplasmic pH was stable at 7.2–7.4 under the control experimental conditions but promptly became acidic (6.7–6.9) after 5′-AMP treatment. On the contrary, treating cells with metformin hydrochloride increased the cytoplasmic pH to 7.5–7.8. Representative wide-field fluorescence images of the BCECF green fluorescence excited at two wavelengths, together with a pseudocoloured ratio image map of cell pH, are shown in Fig. 6c. Next, we found that ANMs, including hypoxanthine, xanthine and uric acid, significantly reduced Na+/H+ exchanger activity in AML12 cells (ESM Fig. 5), indicating that the ANMs caused deregulation of pHi. A decreased intracellular Na+ level was observed in 5′-AMP-treated cells (ESM Fig. 6). The ANMs had no effects on liver vacuolar H+ ATPase activity (ESM Fig. 7). Together, our studies suggest that accumulated ANMs inhibit Na+/H+ exchanger activity and that ATP and metformin, as proton donors and receptors, reciprocally regulate pHi homeostasis in liver cells (Fig. 6d).
pAMP was elevated in people with type 2 diabetes
To further investigate the potential association between elevated pAMP and type 2 diabetes, we analysed the pAMP level in individuals with type 2 diabetes. The clinical characteristics of the participants are shown in Table 1. Twenty-five participants with type 2 diabetes and 21 non-diabetic healthy control individuals participated in the study. The control group had similar age and sex distribution to the diabetes patients. Diabetic individuals had higher BMI and HbA1c. As shown in Fig. 7a, although the absolute level of pAMP in humans was far less than that in mice, it was still significantly elevated in diabetic individuals (45.2 ± 22.7 nmol/l) compared with non-diabetic individuals (3.1 ± 1.9 nmol/l). Fasting plasma glucose and insulin concentrations were significantly increased in the diabetic individuals (Fig. 7b,c). Together, these observations confirm that elevated pAMP is a universal marker in type 2 diabetes.
Our results show that ATP and ANM accumulation in liver cells promote hyperglycaemia in obesity-related type 2 diabetes. Sustained hyperglycaemia impairs insulin-stimulated glucose utilisation in the skeletal muscle, a phenomenon that is referred to clinically as glucose toxicity. Although insulin resistance is thought to arise during the early stage of type 2 diabetes, this process often remains undetectable. In contrast to hyperglycaemia, insulin resistance is a well-evolved protective mechanism during acute illness that helps conserve the brain’s glucose supply by preventing muscles from taking up excessive glucose . Thus, a better understanding of the pathological causes of hyperglycaemia in type 2 diabetes is essential for more rational and appropriate treatment of diabetic individuals.
Release of NEFA from adipose tissue and increased levels of ANMs such as uric acid are important hallmarks of insulin resistance and type 2 diabetes [38, 39]. NEFA-induced nucleotide release from vascular endothelial cells is important for accumulation of ANMs in organs such as liver and skeletal muscle . Inhibiting ATP release by drugs such as glibenclamide (known as glyburide in the USA and Canada) improves insulin resistance and decreases blood glucose levels in type 2 diabetes . It is reported that insulin resistance develops within hours of an acute increase in plasma NEFA levels in humans , and hepatic inflammation has been suggested to be a consequence (rather than a cause) of insulin resistance . Thus, elevated levels of plasma nucleotides could serve as an early signal for diabetes. Different from diabetic mice, non-diabetic obese mice did not display a significant change in pAMP level. Non-diabetic ob/ob mice had higher NEFA levels compared with lean mice but the levels were much lower than in diabetic db/db mice . Thus, levels of nucleotides released by NEFA in these obese mice might be less than in diabetic mice. Ecto-5′-nucleotidase in circulation converts nucleotides to adenosine and is rapidly transported into cells . The intracellular effect of plasma nucleotides is likely mediated through increased cellular adenosine levels. Indeed, adenosine is an inhibitor of insulin signalling . However, elevation of adenosine alone causes insulin resistance but not hyperglycaemia, as observed in ob/ob and HFD-fed mice. Similarly, a low dose of 5′-AMP induced insulin resistance rather than hyperglycaemia. With increased plasma nucleotides, liver accumulates ANMs and hyperglycaemia occurs. ANMs are independent inhibitors of Na+/K+ ATPase; this inhibition inevitably reduces ATP utilisation, resulting in elevation of hepatic ATP. Cellular ATP elevations were also observed in diabetic or 5′-AMP-treated mouse kidney but not skeletal muscle (ESM Fig. 8). On the other hand, ANMs also inhibited activity of the Na+/H+ exchanger, which is crucial in maintaining homeostasis of pHi and sodium . Thus, ANM accumulation could cause a deregulation of pHi.
Coincidentally, pH is an important regulatory factor for PTP1B , which negatively regulates insulin signalling by dephosphorylating tyrosine residues on insulin receptors and IRS proteins . Loss of PTP1B function in mice is associated with increased insulin sensitivity . In our liver cell extracts or recombinant PTP1B enzyme assay system, ATP enhanced PTP1B activity in a dose-dependent manner; metformin (but not metformin hydrochloride) caused a dose-dependent inhibition of PTP1B activity. ATP and metformin acted as an activator and an inhibitor of PTP1B, respectively, causing a reciprocal change in the pH of PTP1B assay buffer. This phenomenon was further proved by binding assay. Metformin, an oral glucose-lowering agent that is widely used in the management of non-insulin-dependent diabetes mellitus , significantly improved nucleotide-induced insulin resistance and hyperglycaemia. Metformin is primarily excreted unchanged in the urine, with negligible metabolism , and is a substrate for OCT1 , which is responsible for the hepatic uptake of metformin. Thus, metformin hydrochloride must be transported into cells in the form of cationic metformin. Indeed, the cation metformin is observed to accumulate in cells . Metformin is a stronger base than most other basic drugs . Therefore, it is believed that hepatic uptake of cationic metformin is an essential step in elevating hepatic cell pH, as well as the occurrence of PTP1B-inhibiting effects such as improved insulin sensitivity.
Thus, our studies uncover a novel key step in the pathogenesis of diabetes-related excessive hepatic glucose production and reveal an emerging role for adenine nucleotides that permits hyperglycaemia to occur in obese mice. As illustrated in Fig. 6d, NEFA-related increase in intracellular ANM accumulation inhibits Na+/K+ ATPase and Na+/H+ exchanger activity, resulting in cellular ATP elevation and pHi deregulation. Cellular ATP, as an H+ donor, directly causes cytoplasmic acidification and activates the insulin negative regulatory pathway of PTP1B, increasing hepatic glucose production. A causal role of cellular pH homeostasis is shown by the observation that metformin-elevated hepatic pHi substantially increases insulin sensitivity. More effective pH regulatory agents under development for use in treating obesity-induced insulin resistance may prove worthy of investigation.
Data are available upon request from the corresponding author.
AMP-activated protein kinase
Adenine nucleotide metabolite
Organic cation transporter-1
Polyethylene glycol-modified adenosine deaminase
- pHi :
Protein tyrosine phosphatase 1B
Cahill GF Jr (2012) Starvation in man. N Engl J Med 282:668–675
DeFronzo RA, Ferrannini E, Simonson DC (1989) Fasting hyperglycemia in non-insulin-dependent diabetes mellitus: contributions of excessive hepatic glucose production and impaired tissue glucose uptake. Metabolism 38(4):387–395. https://doi.org/10.1016/0026-0495(89)90129-7
Kraegen EW, Clark PW, Jenkins AB, Daley EA, Chisholm DJ, Storlien LH (1991) Development of muscle insulin resistance after liver insulin resistance in high-fat-fed rats. Diabetes 40(11):1397–1403. https://doi.org/10.2337/diab.40.11.1397
Kahn SE, Hull RL, Utzschneider KM (2006) Mechanisms linking obesity to insulin resistance and type 2 diabetes. Nature 444(7121):840–846. https://doi.org/10.1038/nature05482
Halaas JL, Gajiwala KS, Maffei M et al (1995) Weight-reducing effects of the plasma protein encoded by the obese gene. Science 269(5223):543–546. https://doi.org/10.1126/science.7624777
Srinivasan K, Ramarao P (2007) Animal models in type 2 diabetes research: an overview. Indian J Med Res 125:451–472
Aasum E, Hafstad AD, Severson DL, Larsen TS (2003) Age-dependent changes in metabolism, contractile function, and ischemic sensitivity in hearts from db/db mice. Diabetes 52(2):434–441. https://doi.org/10.2337/diabetes.52.2.434
Hotamisligil GS (1999) Mechanisms of TNF-α-induced insulin resistance. Diabetes 107:119–125
Coleman DL (1978) Obese and diabetes: two mutant genes causing diabetes-obesity syndromes in mice. Diabetologia 14(3):141–148. https://doi.org/10.1007/BF00429772
Tahara A, Matsuyama-Yokono A, Shibasaki M (2011) Effects of antidiabetic drugs in high-fat diet and streptozotocin-nicotinamide-induced type 2 diabetic mice. Eur J Pharmacol 655(1-3):108–116. https://doi.org/10.1016/j.ejphar.2011.01.015
Taylor SI, Accili D, Imai Y (1994) Insulin resistance or insulin deficiency. Which is the primary cause of NIDDM? Diabetes 43(6):735–740. https://doi.org/10.2337/diab.43.6.735
White MG, Shaw JA, Taylor R (1978) Type 2 diabetes: the pathologic basis of reversible β-cell dysfunction. Diabetes Care 39:2080–2088
Kashyap SR, Defronzo RA (2007) The insulin resistance syndrome: physiological considerations. Diabetes Vasc Dis Res 4(1):13–19. https://doi.org/10.3132/dvdr.2007.001
Egawa K, Maegawa H, Shimizu S et al (2001) Protein-tyrosine phosphatase-1B negatively regulates insulin signaling in l6 myocytes and Fao hepatoma cells. J Biol Chem 276(13):10207–10211. https://doi.org/10.1074/jbc.M009489200
Vicario P, Brady EJ, Slater EE, Saperstein R (1987) Insulin receptor tyrosine kinase activity is unaltered in ob/ob and db/db mouse skeletal muscle membranes. Life Sci 41(10):1233–1241. https://doi.org/10.1016/0024-3205(87)90201-3
Zinker BA, Rondinone CM, Trevillyan JM et al (2002) PTP1B antisense oligonucleotide lowers PTP1B protein, normalizes blood glucose, and improves insulin sensitivity in diabetic mice. Proc Natl Acad Sci U S A 99(17):11357–11362. https://doi.org/10.1073/pnas.142298199
Argaud D, Roth H, Wiernsperger N, Leverve XM (1993) Metformin decreases gluconeogenesis by enhancing the pyruvate kinase flux in isolated rat hepatocytes. Eur J Biochem 213(3):1341–1348. https://doi.org/10.1111/j.1432-1033.1993.tb17886.x
Gunton JE, Delhanty PJ, Takahashi S, Baxter RC (2003) Metformin rapidly increases insulin receptor activation in human liver and signals preferentially through insulin-receptor substrate-2. J Clin Endocrinol Metab 88(3):1323–1332. https://doi.org/10.1210/jc.2002-021394
Lee JO, Lee SK, Jung JH et al (2011) Metformin induces Rab4 through AMPK and modulates GLUT4 translocation in skeletal muscle cells. J Cell Physiol 226(4):974–981. https://doi.org/10.1002/jcp.22410
Zhou G, Myers R, Li Y et al (2001) Role of AMP-activated protein kinase in mechanism of metformin action. J Clin Invest 108(8):1167–1174. https://doi.org/10.1172/JCI13505
Owen MR, Doran E, Halestrap AP (2000) Evidence that metformin exerts its anti-diabetic effects through inhibition of complex 1 of the mitochondrial respiratory chain. Biochem J 34:607–614
Zhang Y, Wang Z, Zhao Y et al (2012) The plasma 5'-AMP acts as a potential upstream regulator of hyperglycemia in type 2 diabetic mice. Am J Physiol Endocrinol Metab 302(3):E325–E333. https://doi.org/10.1152/ajpendo.00424.2011
Knudsen TB, Winters RS, Otey SK et al (1992) Effects of (R)-deoxycoformycin (pentostatin) on intrauterine nucleoside catabolism and embryo viability in the pregnant mouse. Teratology 45(1):91–103. https://doi.org/10.1002/tera.1420450109
Ramsey KM, Yoshino J, Brace CS et al (2009) Circadian clock feedback cycle through NAMPT-mediated NAD+ biosynthesis. Science 324(5927):651–654. https://doi.org/10.1126/science.1171641
Smolenski RT, Lachno DR, Ledingham SJ, Yacoub MH (1990) Determination of sixteen nucleotides, nucleosides and bases using high-performance liquid chromatography and its application to the study of purine metabolism in hearts for transplantation. J Chromatogr 527:414–420. https://doi.org/10.1016/S0378-4347(00)82125-8
Santra A, Maiti A, Das S, Lahiri S, Charkaborty SK, Mazumder DN (2000) Hepatic damage caused by chronic arsenic toxicity in experimental animals. J Toxicol Clin Toxicol 38(4):395–405. https://doi.org/10.1081/CLT-100100949
Tsakiris S, Deliconstantinos G (1984) Influence of phosphatidylserine on (Na++K+)-stimulated ATPase and acetylcholinesterase activities of dog brain synaptosomal plasma membranes. Biochem J 220(1):301–307. https://doi.org/10.1042/bj2200301
Minassian C, Daniele N, Bordet JC, Zitoun C, Mithieux G (1995) Liver glucose-6-phosphatase activity is inhibited by refeeding in rats. J Nutr 125(11):2727–2732. https://doi.org/10.1093/jn/125.11.2727
Petrescu I, Bojan O, Saied M, Bârzu O, Schmidt F, Kühnle HF (1979) Determination of phosphoenolpyruvate carboxykinase activity with deoxyguanosine 5′-diphosphate as nucleotide substrate. Anal Biochem 96(2):279–281. https://doi.org/10.1016/0003-2697(79)90582-7
Wang CD, Teng BS, He YM et al (2012) Effect of a novel proteoglycan PTP1B inhibitor from Ganoderma lucidum on the amelioration of hyperglycaemia and dyslipidaemia in db/db mice. Br J Nutr 108(11):2014–2025. https://doi.org/10.1017/S0007114512000153
Timofeeva OA, Chasovskikh S, Lonskaya I et al (2012) Mechanisms of unphosphorylated STAT3 transcription factor binding to DNA. J Biol Chem 287(17):14192–14200. https://doi.org/10.1074/jbc.M111.323899
Humez S, Monet M, van Coppenolle F, Delcourt P, Prevarskaya N (2004) The role of intracellular pH in cell growth arrest induced by ATP. Am J Physiol Cell Physiol 287(6):C1733–C1746. https://doi.org/10.1152/ajpcell.00578.2003
Howarth C, Gleeson P, Attwell D (2012) Updated energy budgets for neural computation in the neocortex and cerebellum. J Cereb Blood Flow Metab 32(7):1222–1232. https://doi.org/10.1038/jcbfm.2012.35
Xia L, Wang Z, Zhang Y et al (2015) Reciprocal regulation of insulin and plasma 5'-AMP in glucose homeostasis in mice. J Endocrinol 224(3):225–234. https://doi.org/10.1530/JOE-14-0501
Wang DS, Jonker JW, Kato Y, Kusuhara H, Schinkel AH, Sugiyama Y (2002) Involvement of organic cation transporter 1 in hepatic and intestinal distribution of metformin. J Pharmacol Exp Ther 302(2):510–515. https://doi.org/10.1124/jpet.102.034140
Sénéchal F, L’Enfant M, Domon JM et al (2015) Tuning of pectin methylesterification: pectin methylesterase inhibitor 7 modulates the processive activity of co-expressed pectin methylesterase 3 in a pH-dependent manner. J Biol Chem 290(38):23320–23335. https://doi.org/10.1074/jbc.M115.639534
Wang G (2014) Raison dʼêtre of insulin resistance: the adjustable threshold hypothesis. J R Soc Interface 11(101):20140892–20140892. https://doi.org/10.1098/rsif.2014.0892
Herman JB, Goldbourt U (1982) Uric acid and diabetes: observations in a population study. Lancet 320(8292):240–243. https://doi.org/10.1016/S0140-6736(82)90324-5
Dehghan A, van Hoek M, Sijbrands EJ, Hofman A, Witteman JC (2008) High serum uric acid as a novel risk factor for type 2 diabetes. Diabetes Care 31(2):361–362. https://doi.org/10.2337/dc07-1276
Rains SG, Wilson GA, Richmond W, Elkeles RS (1998) The effect of glibenclamide and metformin on serum lipoproteins in type 2 diabetes. Diabet Med 5:653–658
Roden M, Price TB, Perseghin G et al (1996) Mechanism of free fatty acid-induced insulin resistance in humans. J Clin Invest 97(12):2859–2865. https://doi.org/10.1172/JCI118742
Funke A, Schreurs M, Aparicio-Vergara M et al (2014) Cholesterol-induced hepatic inflammation does not contribute to the development of insulin resistance in male LDL receptor knockout mice. Atherosclerosis 232(2):390–396. https://doi.org/10.1016/j.atherosclerosis.2013.11.074
Nowicki M, Kosacka J, Serke H, Blüher M, Spanel-Borowski K (2012) Altered sciatic nerve fiber morphology and endoneural microvessels in mouse models relevant for obesity, peripheral diabetic polyneuropathy, and the metabolic syndrome. J Neurosci Res 90(1):122–131. https://doi.org/10.1002/jnr.22728
Zimmermann H (1992) 5'-Nucleotidase: molecular structure and functional aspects. Biochem J 285(2):345–365. https://doi.org/10.1042/bj2850345
Grinstein S, Rothstein A (1986) Mechanisms of regulation of the Na+/H+ exchanger. J Membr Biol 90(1):1–12. https://doi.org/10.1007/BF01869680
Peters GH, Branner S, Møller KB, Andersen JN, Møller NP (2003) Enzyme kinetic characterization of protein tyrosine phosphatases. Biochimie 85(5):527–534. https://doi.org/10.1016/S0300-9084(03)00036-1
Klaman LD, Boss O, Peroni OD et al (2000) Increased energy expenditure, decreased adiposity, and tissue-specific insulin sensitivity in protein-tyrosine phosphatase 1B-deficient mice. Mol Cell Biol 20(15):5479–5489. https://doi.org/10.1128/MCB.20.15.5479-5489.2000
Stumvoll M, Nurjhan N, Perriello G, Dailey G, Gerich JE (1995) Metabolic effects of metformin in non-insulin-dependent diabetes mellitus. N Engl J Med 333(9):550–554. https://doi.org/10.1056/NEJM199508313330903
Scheen AJ (1996) Clinical pharmacokinetics of metformin. Clin Pharmacokinet 30(5):359–371. https://doi.org/10.2165/00003088-199630050-00003
This work was supported by a grant from the National Natural Science Foundation of China (nos 31671220, 31861163004).
The authors declare that there is no duality of interest associated with this manuscript.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
About this article
Cite this article
Yang, X., Zhao, Y., Sun, Q. et al. Adenine nucleotide-mediated regulation of hepatic PTP1B activity in mouse models of type 2 diabetes. Diabetologia 62, 2106–2117 (2019). https://doi.org/10.1007/s00125-019-04971-1