Mechanism Of TNF-α-induced Hepatic Insulin Resistance: The Role Of NADPH Oxidase 3-derived Reactive Oxygen Species

Objectives:There is emerging evidence that oxidative stress contributes to insulin resistance and diabetes. It is considered that TNF-a-induced insulin resistance is associated with generation of reactive oxygen species (ROS). However, little is known about the cellular mechanism of ROS generation in TNF-a-induced hepatic insulin resistance and the link between ROS and hepatic insulin resistance. This study aims at investigating mechanism of TNF-a-induced hepatic insulin resistance. We observed the oxidative stress and insulin resistance in hepatic tissue of established insulin resistance model-Sprague dawley (SD) rats. Moreover, we analyzed ROS level, NADPH oxidase 3 (NOX3) expression, glycogen content and insulin signal transduction in HepG2 cells treated with TNF-α, in turn demonstrated the key role of NOX3-derived ROS in TNF-a-induced hepatic insulin resistance.Methods:Four-week-old male SD rats (6 rats) were fed a high-fat diet containing 20% fat and 20% sucrose for 12 weeks to induce insulin resistance. Plasma TNF-a level and insulin level was detected by ELISA and radioimmunoassay respectively. Liver intracellular glycogen content was measured using a glycogen assay kit. ROS generation in liver tissues was assessed by DHE fluorescence. The mRNA expression profile of NOX isoforms and subunits in liver tissues was analyzed by RT-PCR. Cultured HepG2 cells were treated for 4 days with 4ng/ml TNF-a to become a cellular model of insulin resistance. ROS level was determined by DCF-DA and FACS. The expression of NOX isoforms and subunits such as P22PHOX, P47PHOX, P67PHOX and Racl in non-treated and TNF-a-treated HepG2 was analyzed by RT-PCR and Western blot. The subcellular localization of P47PHOX was investigated by confocal immunofluoresence microscopy. The siRNA targeting NOX3 (siRNA-NOX3) was transiently transfected into HepG2 cells. Western blot was used to analyze the phosphorylation of signal molecules such as JNK, IRS1, AKT and GSK.Results:After 12 weeks, blood glucose content was increased but still maintained in normal level in rats fed with high-fat diet. However, insulin sensitivity decreased obviously. Rats under high-fat diet showed elevated plasma TNF-a level. Hepatic glycogen content in rats fed with high-fat diet was significantly decreased, confirming insulin resistance. ROS measurement in hepatic tissue slices showed enhanced ROS production in rats fed a high-fat diet, raising the possibility that ROS may be the link between TNF-a and insulin resistance. Importantly, the expression of NOX3 in liver was enhanced in response to high fat diet in vivo. Moreover, in cultured HepG2 cells, TNF-a has been suggested to induce insulin resistance, as assessed by their capacity to accumulate glycogen in the presence of insulin. ROS levels were dose-and time-dependently increased by exposure of cells to TNF-a. To identify the source of ROS generated in response to TNF-a, we studied the effects of different inhibitors of ROS-generating systems:DPI (NADPH oxidase,5μmol/l), L-NAME (Nitric oxide synthases,100μmol/l), Rotenone (Mitochondrial respiratory chain,2μmol/l) and Oxypurinol (Xanthine oxidase,100μmol/l). The results indicate that only DPI inhibites the generation of ROS in response to TNF-a (4 ng/ml,4 days), raising the possibility that a NOX enzyme could be the source of ROS. We therefore analyzed the expression of NOX isoforms and subunits in non-treated and TNF-a-treated HepG2 cells. Using RT-PCR, we found expression of NOX3, but not NOX1, NOX2, NOX4 and NOX5 in non-treated and TNF-a-treated HepG2, indicating that TNF-a-induced ROS might be derived from NOX3. We also found P22PHOX,P47PHOX, P67PHOX and Racl. We next examined the effect of TNF-a treatment on expression and activity of NOX3 in HepG2. Interestingly, TNF-a upregulated the expression of NOX3, but not P22PHOX,P47PHOX, P67PHOX and Racl. It is considered that PKC regulates the activation of NOX3. PMA, an activator of PKC, induced a strong ROS generation, but PKC inhibitor hypericin inhibited TNF-a-induced ROS production. In addition, translocation of P47PHOX may play a role in the activation of NOX3. The results show that P47PHOX had a cytoplasmic distribution in unstimulated HepG2 cells, but a plasma membrane distribution in response to stimulation of TNF-a. Furthermore, membrane proteins were extracted to confirm TNF-a-induced membrane distribution of P47 PHOX. Apocycin, an antioxidant/NOX inhibitor, prevented TNF-a-induced P47PHOX translocation. Apocynin also prevented increased ROS level in response to TNF-a. These results demonstrate that TNF-a enhanced the production of ROS through increased NOX3 protein level and PKC-dependent NOX3 activation, presumably involving P47PHOX translocation. To further explore the role of NOX3 in TNF-a-induced increased ROS generation and decreased glycogen level in hepatocytes, we transiently transfected siRNA targeting NOX3 (NOX3 siRNA) into HepG2 cells. Analysis by real-time PCR and Western blot indicated that the amount of NOX3 mRNA and protein was markedly down-regulated both in control and in TNF-a-treated cells, while control siRNA had no effect. NOX3 siRNA led to an approximately 50% decrease of basal ROS generation, but completely abolished TNF-a-induced ROS generation. It appears that NOX3 is the predominant source of ROS induced by TNF-a. We next investigated the role of NOX3 in the regulation of intracellular glycogen content. NOX3 siRNA did not impact glycogen level under control condition, but completely prevented the TNF-a-induced decrease of glycogen. These data demonstrate that NOX3-derived ROS are essential mediators of the TNF-a-induced decreased glycogen content in hepatocytes. Finally, we investigated downstream pathways translating elevated ROS level into decreased glycogen content. Our results show that JNK was activated in response to TNF-a and that effect was reversed by NOX3 down-regulation. In parallel with increased phosphorylation of JNK, TNF-a down-regulated level of IRS1 and stimulated its inhibitory phosphorylation on Ser307. Importantly, the decreased IRS1 protein level and the increased Ser307 phosphorylation was reversed by NOX3 siRNA, to a similar extend as observed with SP600125 (JNK inhibitor). Moreover, pretreatment with SP600125 can reverse the decrease of cellular glycogen content induced by TNF-a. The TNF-α-induced impaired phosphorylation of AKT and GSK was also rescued by siRNA mediated NOX3 reduction. Thus, NOX3-derived ROS play a key role in TNF-a signaling towards insulin resistance.Conclusions:Our results indicated that 1) rats fed with high-fat diet displayed increased plasma TNF-αlevel, decreased insulin sensitivity, enhanced ROS level and NOX3 expression but reduced glycogen content in liver; 2) TNF-αenhanced the production of ROS through increased NOX3 protein level and PKC-dependent NOX3 activation, presumably involving P47PHOX translocation; 3) TNF-αinduced insulin resistance in HepG2 cells via JNK/IRS1/AKT/GSK pathway; 4) pretreatment with JNK inhibitor SP600125 and NOX3 down-regulation by NOX3 siRNA can reverse TNF-a-induced the decrease of cellular glycogen content in HepG2 cells. In conclusion, the effects of TNF-a on hepatic insulin resistance appear to be, at least in part, mediated by NOX3-derived ROS through a JNK pathway.