Development of insulin resistance and obesity in mice overexpressing cellular glutathione peroxidase James P. McClung*, Carol A. Roneker*, Weipeng Mu*, Donald J. Lisk*, Paul Langlais†, Feng Liu†, and Xin Gen Lei*‡ *Department of Animal Science, Cornell University, Ithaca, NY 14853; and †Department of Pharmacology, University of Texas Health Science Center, 7703 Floyd Curl Drive, San Antonio, TX 78229 Edited by Robert J. Cousins, University of Florida, Gainesville, FL, and approved April 29, 2004 (received for review December 5, 2003)
Insulin resistance, a hallmark of type 2 diabetes, is associated with oxidative stress. However, the role of reactive oxygen species or specific antioxidant enzymes in its development has not been tested under physiological conditions. The objective of our study was to investigate the impact of overexpression of glutathione peroxidase 1 (GPX1), an intracellular selenoprotein that reduces hydrogen peroxide (H2O2) in vivo, on glucose metabolism and insulin function. The GPX1-overexpressing (OE) and WT male mice (n ⴝ 80) were fed a selenium-adequate diet (0.4 mg兾kg) from 8 to 24 weeks of age. Compared with the WT, the OE mice developed (P < 0.05) hyperglycemia (117 vs. 149 mg兾dl), hyperinsulinemia (419 vs. 1,350 pg兾ml), and elevated plasma leptin (5 vs. 16 ng兾ml) at 24 weeks of age. Meanwhile, these mice were heavier (37 vs. 27 g, P < 0.001) and fatter (37% vs. 17% fat, P < 0.01) than the WT mice. At 30 – 60 min after an insulin challenge, the OE mice had 25% less (P < 0.05) of a decrease in blood glucose than the WT mice. Their insulin resistance was associated with a 30 –70% reduction (P < 0.05) in the insulin-stimulated phosphorylations of insulin receptor (-subunit) in liver and Akt (Ser473 and Thr308) in liver and soleus muscle. Here we report the development of insulin resistance in mammals with elevated expression of an antioxidant enzyme and suggest that increased GPX1 activity may interfere with insulin function by overquenching intracellular reactive oxygen species required for insulin sensitizing.
T
ype 2 diabetes is one of the fastest growing and most costly disorders worldwide. Because insulin resistance precedes the onset of type 2 diabetes, it is generally considered to be an important factor in the pathogenesis of the disease (1). Although there is substantial evidence that hyperglycemia results in the generation of reactive oxygen species (ROS) and increased oxidant stress in the late complications of diabetes (2), the role of ROS in the development of insulin resistance (3, 4), especially at the whole-body level, remains virtually unknown. Selenium (Se) is an essential micronutrient that functions mainly through Se-dependent proteins. Glutathione peroxidase 1 (GPX1, EC 1.11.19) is the first identified (5) and the most abundant selenoprotein (6) in mammals. The enzyme is expressed in both cytosol and mitochondria (7) and uses glutathione to reduce H2O2 and organic hydroperoxides. Overexpression or knockout of GPX1 renders mice resistant or susceptible to acute oxidative stress (8, 9) but has no effect on the expression of other Se-dependent or antioxidant enzymes (10). Although an insulin-mimetic property of Se has been shown in isolated adipocytes (11) or in streptozotocin-induced diabetic rodents (12, 13), the role of any specific selenoprotein in the development of insulin resistance has not been studied in humans or animals. Insulin imparts a vital role in glucose metabolism and homeostasis, and its action is mediated through a tightly regulated signaling cascade (14). When insulin binds the insulin receptor, a heterotetrameric glycoprotein with two extracellular ␣subunits (135 kDa) as the binding site and two transmembrane -subunits (95 kDa) as the tyrosine kinase, the receptor undergoes autophosphorylation that increases the kinase activity. The activated insulin receptor phosphorylates insulin receptor sub8852– 8857 兩 PNAS 兩 June 15, 2004 兩 vol. 101 兩 no. 24
strate proteins, leading to activation of phosphatidylinositol (PI) 3-kinase (15). Activation of PI3-kinase can enlist PI3-kinasedependent kinase and the serine兾threonine kinase Akt from the cytoplasm to the plasma membrane. This event causes conformational changes in Akt, which allows PI3-kinase-dependent kinase to phosphorylate Thr308 and Ser473 in the protein. The activation of Akt leads to the activation of glycogen synthesis and the translocation of glucose transporter 4 to the cell membrane for glucose transport (16). Any alteration in phosphorylation or dephosphorylation of these signal proteins, from the insulin receptor to many downstream signal proteins such as Akt, may affect insulin function. In fact, insulin resistance is associated with impaired Akt activation (17), and insulin responsiveness can be improved by knockout of protein tyrosine phosphatase 1B, which dephosphorylates the insulin receptor (18). Although ROS such as H2O2 are known to inhibit phosphatases, their role in insulin signaling has been controversial (19–22). This is largely due to the difficulty in specifically altering or measuring intracellular ROS concentrations. Because overexpression or knockout of antioxidant enzymes such as GPX1 in mice modulates intra- or subcellular levels of ROS (7), these transgenic animals may be used to clarify the physiological role of ROS in insulin action. During our oxidant stress research, we observed that old GPX1 overexpression mice (OE mice) became heavier than the WT mice. Therefore, this study was conducted to determine the effect of GPX1 overexpression on blood glucose, plasma insulin and leptin, body weight and composition, insulin responsiveness, and signaling in mice from 8 to 24 weeks of age. Strikingly, we found that the OE mice developed hyperglycemia, hyperinsulinemia, and mild insulin resistance at 24 weeks of age. This abnormality was associated with attenuations in the insulinmediated phosphorylation of insulin receptor and Akt in liver and兾or muscle. Materials and Methods Chemicals and Antibodies. All chemicals were purchased from
Sigma unless indicated otherwise. Insulin (Humulin R) was purchased from Eli Lilly. Antibody against insulin receptor (subunit, C-19) was from Santa Cruz Biotechnology; antibody against phosphotyrosine (4G10) was from Upstate Biotechnology (Lake Placid, NY); and antibodies against AKT, phosphoAKT Ser473, and phospho-AKT Thr308 were from Cell Signaling Technology (Beverly, MA).
Transgenic Mice. The OE and the WT breeding mice were
provided by B. H. L. Chua (East Tennessee State University, Johnson City). Both genotypes were originally developed by Y. S. Ho (Wayne State University, Detroit) and bred on the same genetic background (129兾SVJ ⫻ C57BL兾6). All experimental
This paper was submitted directly (Track II) to the PNAS office. Abbreviations: ROS, reactive oxygen species; Se, selenium; GPX1, glutathione peroxidase 1; PI, phosphatidylinositol; OE mice, GPX1 overexpression mice. ‡To
whom correspondence should be addressed. E-mail:
[email protected].
© 2004 by The National Academy of Sciences of the USA
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0308096101
mice used in this study (n ⫽ 80) were bred in our mouse facility and were 8-week-old males at the start of the experiment. All mice were fed a Se-adequate diet (0.4 mg兾kg) (6), given free access to feed and distilled water, and housed in shoebox cages in a constant temperature (22°C) animal room with a 12-h light兾dark cycle. All experiments were approved by the Institutional Animal Care and Use Committee at Cornell University and conducted in accordance with National Institutes of Health guidelines for animal care. genotype) were fasted overnight for 8 h before determinations of whole-blood glucose, plasma insulin, and plasma leptin at various time points. Whole-blood glucose was determined by clipping tails and using the Glucometer Elite system (Bayer, Elkhart, IN) with Ascensia Elite blood glucose test strips as described by the manufacturer. For determination of insulin or leptin, plasma samples were prepared from whole blood that was collected by heart puncture using a heparinized syringe after mice were anesthetized with carbon dioxide. Plasma insulin was determined by using a rat insulin ELISA kit with mouse insulin as a standard (Crystal Chem, Downers Grove, IL). Plasma leptin was determined by using an ELISA kit with mouse leptin as a standard (Crystal Chem).
Body Composition. Body composition of mice (n ⫽ 4 per genotype,
24 weeks old) was determined by K. J. Ellis (Body Composition Laboratory, U.S. Department of Agriculture兾Agricultural Research Service Children’s Nutrition Research Center and Department of Pediatrics, Baylor College of Medicine, Houston), using dual-energy x-ray absorptiometry. Insulin Sensitivity. Blood glucose reductions after insulin challenge were used to determine insulin sensitivity of WT and OE mice (n ⫽ 8–13 per genotype) at 8 and 24 weeks of age. Both groups were fasted overnight (8 h) and then given an i.p. injection of 0.25 unit of insulin per kg of body weight. Wholeblood glucose was determined at 0, 30, 60, 120, and 240 min after the insulin injection, and the data were presented as the relative percentage of the initial levels. Antioxidant Enzyme Activity. At the end of the experiment, mice (24 weeks of age) were killed to collect liver, soleus muscle, and plasma samples for enzyme activity assays (6). Before the assays, tissue samples (n ⫽ 3–5 for each genotype) were thawed on ice and homogenized (Polytron PT3100; Brinkman Instruments, Littau, Switzerland) in selected buffers. Activities of the Sedependent GPX1, GPX3 (plasma GPX), and GPX4 (phospholipids hydroperoxide GPX), and the Se-independent glutathione S-transferase were measured as described (6). Se-dependent thioredoxin reductase activity was determined by using the NADPH-dependent reduction of 5,5⬘-dithiobis-(2-nitrobenzoic acid) method (23). The Cu,Zn-superoxide dismutase and Mnsuperoxide dismutase activities were measured by using a watersoluble formazan dye kit (Dojindo Molecular Technologies, Gaithersburg, MD). Protein was determined as described by Lowry et al. (24). Western Blotting, Immunoprecipitation, and Insulin Signaling. Liver and muscle samples used for Western blot analysis were homogenized in buffer A [100 mM Tris, pH 7.4兾250 mM sucrose兾protease inhibitor mixture (1 mM sodium pyrophosphate兾1 mM sodium orthovanadate兾10 g of leupeptin per ml兾10 g of aprotinin per ml兾1 M microcystin兾1 mM PMSF兾10 mM sodium fluoride)]. The samples used for immunoprecipitation were homogenized in buffer B (50 mM Hepes, pH 7.6兾100 mM sodium chloride兾1% Triton X-100兾5 mM EDTA兾protease inhibitor mixture as in buffer A). Tissue homogenates were centrifuged at 14,000 ⫻ g for 10 min at McClung et al.
Fig. 1. Effect of GPX1 overexpression on GPX1 activity in liver and muscle of mice (24 weeks of age). Values are means ⫾ SE (n ⫽ 5), and the asterisk indicates a difference (P ⬍ 0.05) between the two genotypes.
4°C. Western blotting was performed as described (25), by using a chemiluminescent method (Supersignal West Pico kit; Pierce) for signal detection. The relative density of the protein bands was quantified by using the Alpha-Imager 2200 system (Alpha Innotech, San Leandro, CA). For determination of hepatic insulin receptor (-subunit) phosphorylation, mice (24 weeks of age, n ⫽ 6 per genotype) were fasted overnight (8 h) and were injected (i.p.) with insulin (5 units兾kg). Mice were killed 3 min after the injection and liver samples were excised, mixed with buffer B, and homogenized immediately on ice. The immunoprecipitation was performed as described by Cheng et al. (25). Briefly, supernatants (0.2 mg of protein) from liver homogenates were precleared by incubation with 5 l of rabbit serum and 30 l of protein A-Sepharose beads (50% slurry) for 30 min at 4°C. After centrifugation at 3,000 ⫻ g for 1 min, the supernatants were incubated with the antibody against insulin receptor -subunit overnight at 4°C with rotation. After the preparation was incubated with 50 l of protein A beads for 30 min at 4°C with rotation, the beads were washed four times with 1 ml of buffer B on ice. Buffer B (20 l) and 5⫻ SDS buffer (20 l) were then added to the washed beads and used for Western blot analysis with the antibody against phosphotyrosine (4G10). For the determination of Akt protein and phosphorylation, liver and soleus muscle samples were collected 8 min after mice (24 weeks of age, n ⫽ 6–8 per genotype) were injected (i.p.) with 10 units of insulin per kg (this insulin dose and exposure time was chosen for reliable signals). The tissue samples were homogenized in buffer A and subjected to Western blot analysis with the appropriate antibody as described above. Statistical Analysis. Data were analyzed by using SAS (release 6.11, SAS Institute, Cary, NC), and Student’s t test was used for mean comparisons.
Results Hyperglycemia, Elevated Body Weight, and Obesity in OE Mice. As
shown in Fig. 1, GPX1 activity in liver and soleus muscle of OE mice was 21% and 3-fold greater (P ⬍ 0.05) than that of WT mice, respectively. However, these two genotypes had similar activities of GPX3, GPX4, thioredoxin reductase, glutathione S-transferase, Cu,Zn-superoxide dismutase, and Mn-superoxide dismutase in plasma, liver, or muscle (Table 1). Blood glucose concentrations were similar between the OE and the WT mice initially (8 weeks of age), but became 27% greater (P ⬍ 0.05) in the OE mice than the WT mice at 24 weeks of age (Fig. 2A). The OE mice were 37% heavier (37 vs. 27 g, P ⬍ 0.001) than the WT mice at 24 weeks of age (Fig. 2B). The dual-energy x-ray absorptiometry of whole body indicated that fat deposition was responsible for the increased body weight in the OE mice, because those mice had 37% body fat, compared with 17% in WT mice (P ⬍ 0.05, Fig. 2C). There was no difference in body protein or mineral content between the two groups of mice (data not shown). PNAS 兩 June 15, 2004 兩 vol. 101 兩 no. 24 兩 8853
APPLIED BIOLOGICAL SCIENCES
Blood Glucose, Plasma Insulin, and Plasma Leptin. Mice (n ⫽ 4–6 per
Table 1. Effect of GPX1 overexpression on activities of other selenium-dependent or antioxidant enzymes in liver, muscle, and plasma of mice (24 weeks of age) Enzyme
WT
OE
Plasma GSH peroxidase (GPX3), nmol GSH oxidized per min per mg of protein 35.0 ⫾ 3.3 36.7 ⫾ 1.7 Phospholipid hydroperoxide GPX (GPX4), nmol GSH oxidized per min per mg of protein Liver 11.3 ⫾ 2.2 15.9 ⫾ 2.8 Muscle 6.3 ⫾ 3.3 8.6 ⫾ 2.4 Thioredoxin reductase, nmol 5⬘-thionitrobenoic acid formed per min per mg of protein Liver 34.8 ⫾ 3.7 38.4 ⫾ 2.0 Muscle 20.2 ⫾ 3.5 18.9 ⫾ 4.7 GSH S-transferase, nmol S-2,4-dintrophenylglutathione formed per min per mg of protein Liver 4,038 ⫾ 424 3,510 ⫾ 229 Muscle 906 ⫾ 230 778 ⫾ 69 Cu, Zn-Superoxide dismutase, 50% formazan dye formation rate inhibition per mg protein Liver 1,672 ⫾ 62 1,424 ⫾ 55 Muscle 445 ⫾ 47 404 ⫾ 24 Mn-Superoxide dismutase, 50% formazan dye formation rate inhibition per mg protein Liver 8.2 ⫾ 0.3 8.7 ⫾ 0.4 Muscle 12.9 ⫾ 1.5 11.1 ⫾ 0.5 Values are means ⫾ SE (n ⫽ 3–5). GSH, glutathione.
Hyperinsulinemia and Elevated Plasma Leptin in OE Mice. Initial (8
weeks old) plasma insulin and leptin concentrations were not significantly different between the two genotypes (Fig. 3). However, the OE mice had elevated plasma insulin and leptin concentrations as high as 3.2-fold (P ⬍ 0.05) of those of the WT mice at 24 weeks of age (insulin: 1,350 vs. 419 pg兾ml; leptin: 16 vs. 5 ng兾ml).
Insulin Resistance in OE Mice. Whole-body sensitivity to insulin, measured by the relative blood glucose reductions in response to an insulin challenge, was not different between the two genotypes at 8 weeks of age (Fig. 4A). However, the OE mice demonstrated insulin resistance compared with the WT mice at 24 weeks of age (Fig. 4B). After the insulin injection, the reductions of whole-blood glucose, relative to their initial levels, were 25% less (P ⬍ 0.05) in the OE mice than in the WT mice at 30 and 60 min. Diminished Phosphorylations of Insulin Receptor and Akt in OE Mice.
To elucidate the mechanism by which overexpression of GPX1 led to insulin resistance, we examined insulin signaling in two insulin-sensitive tissues, liver and skeletal muscle, of the WT and OE mice. As shown in Fig. 5A, the overexpression of GPX1 had no effect on hepatic insulin receptor -subunit protein levels. However, the insulin-stimulated phosphorylation of the insulin receptor was attenuated by ⬇70% (P ⬍ 0.05, n ⫽ 6 for each genotype) in the OE mice, compared with the WT mice (Fig. 5B). Likewise, the overexpression of GPX1 had no significant effect on the Akt protein levels in liver or soleus muscle (Fig. 6 A and B). However, the insulin-induced phosphorylations of Akt at Ser473 and Thr308 in both liver and muscle were 30–55% lower (P ⬍ 0.05, n ⫽ 6–8 for each genotype) in the OE mice than in the WT mice (Fig. 6). Discussion The most interesting finding from the present study is the development of insulin resistance in mice with elevated expres8854 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0308096101
Fig. 2. Effect of GPX1 overexpression on whole-blood glucose (A), body weight (B), and body fat (C) of mice. Whole-blood glucose was measured in mice fasted overnight by using tail blood. Body fat accretion in mice (24 weeks of age) was determined by dual-energy x-ray absorptiometry. Values are means ⫾ SE (n ⫽ 4 – 6 per genotype), and the asterisks indicate differences (P ⬍ 0.05) between the two genotypes.
sion of a major antioxidant selenoprotein, GPX1. Compared with the WT mice, the OE mice showed hyperglycemia, hyperinsulinemia, elevated body fat accretion and plasma leptin, and reduced insulin sensitivity at 24 weeks of age. The malady matches signs typical of insulin resistance. It is striking that overexpression of GPX1, an effective intracellular enzyme against ROS (8, 9, 25), does not improve, but reduces, body insulin sensitivity in mice. Our finding is strongly supported by a recent human study. Chen et al. (26) have observed a significantly (P ⬍ 0.01) positive association between increases in erythrocyte GPX1 activity and levels of insulin resistance in normal pregnant women from 16 weeks of pregnancy (entry of study) to the third trimester. That association was also highly (P ⬍ 0.05) related to dietary fat intake and ethnic differences. The concurrence of GPX1 activity increase and insulin resistance during pregnancy provides not only a strong physiological basis, but also clinical relevance for our data. Because we report that insulin resistance is related to the overexpression of an antioxidant enzyme in mice, our results are apparently contrary to the popular notion that incidences of insulin resistance and diabetes inversely correlate with tissue antioxidant enzyme activities (13, 27, 28). In addition, a causative role of ROS has been suggested in the onset of insulin resistance or type 2 diabetes (3, 4), implying possible benefit of supplemental antioxidants in improving insulin function. However, McClung et al.
these notions have been largely derived from clinical investigations or studies using conventional animal models. The complexity of clinical data and the limitation of those models, along with the use of gross status of antioxidant enzymes in whole tissues or extracellular fluids in these studies, might have contributed to the discrepancy with our results. Another challenge to our findings is the insulin-mimetic effect of Se (29). Since the initial illustration of that property of Se in isolated rat adipocytes (11), a number of studies have shown the ability of selenate to initiate the translocation of glucose transporters to the membrane surface (30) and to induce phosphorylation of the insulin receptor (31) or insulin receptor substrate 1 (32). Because overexpression of GPX1,
Fig. 4. Effect of GPX1 overexpression on insulin sensitivity of mice at 8 (A) and 24 (B) weeks of age. After fasted mice (n ⫽ 8 –13 per genotype) were given an injection (i.p.) of insulin (0.25 unit兾kg), whole-blood glucose was measured at the indicated time points by using the Glucometer (Bayer). Values are expressed as the percentage of the initial glucose levels, and the asterisks indicate differences (P ⬍ 0.05) between genotypes within a given time point.
McClung et al.
the most abundant biochemically functional form of body Se (6), does not duplicate the insulin-mimetic effect of Se in the present study, that effect of Se is unlikely mediated by selenoproteins, but rather by the element itself. This view is supported by a recent rat study in which only selenate, but not selenite, was effective in that regard (33). The key question related to our findings is how overexpression of GPX1 in mice induced insulin resistance. We believe that the most plausible mechanism could be the effect of GPX1 overexpression on intracellular H2O2 tone. Although the general role of H2O2, if any, in insulin signaling is still controversial, H2O2 can undoubtedly modulate the insulin-induced phosphorylation of the insulin receptor -subunit (19) and Akt (21). It has been shown that insulin stimulation generates a burst of H2O2 in hepatoma and adipose cells that is associated with a reversible oxidative inhibition of overall cellular protein–tyrosine phosphatase activity (20). Thus, maintenance of normal activity of protein–tyrosine phosphatase is critical to the regulation of reversible tyrosine phosphorylation in the insulin signaling cascade and, hence, insulin sensitivity. Meanwhile, a recent study has suggested a role of ROS, including H2O2, in the activation of Akt through phosphorylation of Ser473 (34). Apparently, normal or minimal levels of intracellular ROS or H2O2 are required for sensitizing insulin signaling. Overexpression of GPX1 may accelerate the quenching of the intracellular H2O2 burst after insulin stimulation, resulting in less inhibition of protein–tyrosine phosphatase activity and, subsequently, attenuated phosphorylation of insulin receptor (Fig. 5B). Presumably, the diminished intracellular H2O2 concentrations in the OE mice might also be responsible for their reduced phosphorylation of Akt in liver and soleus muscle (Fig. 6). PNAS 兩 June 15, 2004 兩 vol. 101 兩 no. 24 兩 8855
APPLIED BIOLOGICAL SCIENCES
Fig. 3. Effect of GPX1 overexpression on plasma insulin (A) and leptin concentrations (B) of mice at 8 and 24 weeks of age. Both hormones were measured by ELISA in mice fasted overnight. Values are means ⫾ SE (n ⫽ 4 – 6 per genotype), and the asterisks indicate differences (P ⬍ 0.05) between the two genotypes at a given time point.
Fig. 5. Effect of GPX1 overexpression on hepatic insulin receptor (-subunit) (IR) protein (A), the insulin-stimulated phosphorylation of hepatic insulin receptor (-subunit) (P-IR) (B), and the relative quantification of phosphorylated hepatic insulin receptor (-subunit) (C). Fasted mice (24 weeks of age) were given an injection of saline (S) or insulin (I; 5.0 units兾kg of body weight), and liver samples were collected 3 min after the injection. The Western blot analysis of insulin receptor protein was performed by separating freshly prepared tissue homogenates (50 g of protein per lane) on an SDS兾PAGE gel (10%) and probing with an antibody against insulin receptor (-subunit, C-19; Santa Cruz Biotechnology). Phosphorylation of hepatic insulin receptor was detected by SDS兾PAGE (10% gel) using the 4G10 antibody (Upstate Biotechnology) after the immunoprecipitation of tissue homogenate (200 g protein per sample) with the above-described antibody against insulin receptor. The blots are representative of three independent experiments. The relative density values of phosphorylated hepatic insulin receptor (-subunit) are means of six mice for each genotype (n ⫽ 6) from the three experiments, and the asterisk indicates a significant genotype difference (P ⬍ 0.05).
Fig. 6. Effect of GPX1 overexpression on Akt protein (AKT), phospho-Akt (Thr308) (P-308), and phospho-Akt (Ser473) (P-473) in liver (A) and soleus muscle (B) and the relative quantification of P-308 and P-473 in both tissues (C). Fasted mice (24 weeks of age) were given an injection of saline (S) or insulin (I; 10 units兾kg of body weight), and tissue samples were collected 8 min after the injection. The Western blot analyses were performed by separating freshly prepared tissue homogenates (30 g of protein per lane) on an SDS兾PAGE gel (10%) and probing with appropriate antibodies (Cell Signaling Technology). The blots are representative of three independent experiments. The relative density values of phosphorylated Akt are means of six to eight mice for each genotype (n ⫽ 6 – 8) from the three experiments, and the asterisk indicates a significant genotype difference (P ⬍ 0.05).
A possible impact of GPX1 overexpression on mitochondrial ROS and oxidative phosphorylation (35) might also contribute to insulin resistance in the OE mice. In certain tissues such as liver and kidney (7), GPX1 is expressed in both cytosol and mitochondria. In fact, we detected a greater GPX1 activity in the liver mitochondria of the OE mice than that of the WT mice (117 vs. 197 units兾mg protein, n ⫽ 7, P ⬍ 0.05). Probably because of the low abundance of GPX1 (7) in muscle, we were unable to detect GPX1 protein or activity in the mitochondrial fractions prepared (7, 36) from the very limited amount of soleus muscle of individual mice. However, others have clearly shown the existence of GPX1 in the muscle mitochondria (36). As the main site of cellular energy, mitochondria produce much of the cellular H2O2 via Mn-superoxide dismutase (37). Thus, altered GPX1 expression in mitochondria may affect its ROS tone and oxidative phosphorylation (7). Although it is unclear to us how mitochondrial GPX1 or ROS directly modulates insulin function, impaired oxidative phosphorylation in liver and muscle mitochondria has been shown to be linked with insulin resistance or type 2 diabetes in humans (38, 39). In addition, elevated GPX1 activity in mitochondria may reduce its release of H2O2 to cytosol (7), and then desensitizes insulin signaling as discussed above. Although inherent diabetic phenotypes have been shown to be associated with the creation of transgenic mice (40), it is highly unlikely the cause of insulin resistance in the OE mice of the present study. This is because we found hyperglycemia and obesity in another group of OE mice derived from a different founder line (data not shown). Likewise, insulin resistance in the OE mice cannot be attributed to other compensatory changes caused by GPX1 overexpression, because we found no differences in activities of other Se-dependent or antioxidant enzymes in liver or muscle between the OE and WT mice. In addition, these two genotypes showed no differences in hepatic F28856 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0308096101
isoprostanes (a sensitive in vivo biomarker of lipid peroxidation), plasma glutathione兾glutathione disulfide, or stress-related signal proteins (mitogen-activated protein kinase p-38 and c-Jun Nterminal kinase) in liver or muscle (data not shown). Hyperphage can also be ruled out as the cause of insulin resistance in OE mice. We found no significant differences in daily food intake between the OE and WT mice (4.2 vs. 4.0 g, n ⫽ 8, P ⫽ 0.65) from three independent observation periods (weeks). However, it will be fascinating to find out whether caloric restriction can affect the development of insulin resistance in the OE mice. The combined attenuation in the insulin-mediated phosphorylation of insulin receptor in liver and Akt in both tissues of the OE mice is rather unique. Disruptions in hepatic glucose metabolism impart an important role in the development of hyperglycemia and type 2 diabetes (41). Liver dysfunction may cause hyperglycemia and insulin resistance, and mice lacking hepatic insulin receptors develop hyperglycemia, hyperinsulinemia, glucose intolerance, and insulin resistance (42). Insulin sensitivity and obesity can be improved by knockout of protein tyrosine phosphatase-1B that dephosphorylates the insulin receptor (18). Thus, diminished hepatic insulin receptor phosphorylation in the OE mice might be a major contributor to their insulin resistance and obesity, because we did not find that change in muscle (data not shown). Because the PI-dependent serine–threonine protein kinase Akt functions as an essential intermediate in the signaling pathway by which insulin controls glucose uptake and hepatic gluconeogeneis (43), insulin resistance has been associated with impaired Akt activation and decreased glucose transport (17). In fact, mice lacking the Akt2 protein are insulin resistant and develop a diabetes-like syndrome (44). Thus, the diminished insulin-mediated phosphorylation of Akt at Ser473 and Thr308 in liver and soleus muscle of OE mice likely contributed to the development of insulin resistance. Meanwhile, the effect of McClung et al.
GPX1 overexpression on insulin receptor and Akt phosphorylations implies a critical role of normal or minimal intracellular ROS in sensitizing insulin signaling. Illustrating the impacts of GPX1 overexpression on glucose metabolism and insulin function has multiple physiological and pharmacological implications. First, it provides us with an etiological factor for insulin resistance, and argues against the general belief in the beneficial effects of antioxidant enzymes on insulin function that may not reflect the actual role of any specific enzyme such as GPX1 at the intra- or subcellular level. Second, the opposite impact of GPX1 overexpression to that of selenate on insulin function indicates that the insulin mimetic effect of Se is probably caused by the element itself, not selenoproteins, at least not GPX1. Third, alterations of GPX1 allele frequencies have been identified and associated with human cancers (45, 46). As discussed above, elevated erythrocyte GPX1 activity is associated with insulin resistance, dietary
fat intake, and ethnic difference in pregnant women (26). If the impact of GPX1 expression on insulin resistance is indeed mediated via cytosolic and兾or mitochondrial ROS, variations in other antioxidant compounds or enzymes in human populations (47) may produce similar outcomes. Effects of GPX1 overexpression on the phosphorylation of insulin receptor and Akt may shed light on new drug developments to treat insulin resistance. Finally, the phenotype of the OE mice, including hyperglycemia, hyperinsulinemia, obesity, and reduced insulin sensitivity, offers a desirable model for the study of antioxidant balance, insulin resistance, and type 2 diabetes (40, 48–50).
1. Stumvoll, M. & Haring, H. (2001) Horm. Res. 55, 3–13. 2. Packer, L., Kraemer, K. & Rimbach, G. (2001) Nutrition 17, 888–895. 3. Rosen, P., Nawroth, P. P., King, G., Moller, W., Tritschler, H. J. & Packer, L. (2001) Diabetes Metab. Res. Rev. 17, 189–212. 4. Evans, J. L., Goldfine, I. D., Maddux, B. A. & Grodsky, G. M. (2002) Endocrine Rev. 23, 599–622. 5. Rotruck, J. T., Pope, A. L., Ganther, H. E., Swanson, A. B., Hafeman, D. G. & Hoekstra, W. G. (1973) Science 179, 588–590. 6. Cheng, W. H., Ho, Y. S., Ross, D. A., Valentine, B. A., Combs, G. F., Jr., & Lei, X. G. (1997) J. Nutr. 127, 1445–1450. 7. Esposito, L. A., Kokoszka, J. E., Waymire, K. G., Cottrell, B., MacGregor, G. R. & Wallace, D. C. (2000) Free Radical Biol. Med. 28, 754–766. 8. Cheng, W. H., Ho, Y. S., Valentine, B. A., Ross, D. A., Combs, G. F., Jr., & Lei, X. G. (1998) J. Nutr. 128, 1070–1076. 9. Cheng, W. H., Fu, Y. X., Porres, J. M., Ross, D. A. & Lei, X. G. (1999) FASEB J. 13, 1467–1475. 10. Lei, X. G. (2002) Methods Enzymol. 347, 213–225. 11. Ezaki, O. (1990) J. Biol. Chem. 265, 1124–1128. 12. Becker, D. J., Reul, B., Ozcelikay, A. T., Buchet, J. P., Henquin, J. C. & Brichard, S. M. (1996) Diabetologia 39, 3–11. 13. Mukherjee, B., Anbazhagan, S., Roy, A., Ghosh, R. & Chatterjee, M. (1998) Biomed. Pharmacother. 52, 89–95. 14. O’Brien, R. M. & Granner, D. K. (1996) Physiol. Rev. 76, 1109–1161. 15. Dupont, J. & LeRoith, D. (2001) Horm. Res. 55, 22–26. 16. Tremblay, F., Lavigne, C., Jacques, H. & Marette, A. (2001) Diabetes 50, 1901–1910. 17. Tomas, E., Lin, Y. S., Dagher, Z., Saha, A., Luo, Z., Ido, Y. & Ruderman, N. B. (2002) Ann. N.Y. Acad. Sci. 967, 43–51. 18. Elchebly, M., Payette, P., Michaliszyn, E., Cromlish, W., Collins, S., Loy, A. L., Normandin, D., Cheng, A., Himms-Hagen, J., Chan, C. C., et al. (1999) Science 283, 1544–1548. 19. Hansen, L. L., Ikeda, Y., Olsen, G. S., Busch, A. K. & Mosthaf, L. (1999) J. Biol. Chem. 274, 25078–25084. 20. Mahadev, K., Zilbering, A., Zhu, L. & Goldstein, B. J. (2001) J. Biol. Chem. 276, 21938–21942. 21. Gardner, C. D., Eguchi, S., Reynolds, C. M., Eguchi, K., Frank, G. D. & Motley, E. D. (2003) Exp. Biol. Med. 228, 836–842. 22. Potashnik, R., Bloch-Damti, A., Bashan, N. & Rudich, A. (2003) Diabetologia 46, 639–648. 23. Holmgren, A. & Bjornstedt, M. (1995) Methods Enzymol. 252, 199–208. 24. Lowry, O. H., Rosenbrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193, 265–275. 25. Cheng, W. H., Zheng, X., Quimby, F. R., Roneker, C. A. & Lei, X. G. (2003) Biochem. J. 370, 927–934.
26. Chen, X., Scholl, T. O., Leskiw, M. J., Donaldson, M. R. & Stein, T. P. (2003) J. Clin. Endocrinol. Metab. 88, 5963–5968. 27. McNeill, J. H., Delgatty, H. L. M. & Battell, M. L. (1991) Diabetes 40, 1675–1678. 28. Douillet, C., Bost, M., Accominotti, M., Borson-Chazot, F. & Ciavatti, M. (1998) Lipids 33, 393–399. 29. Stapleton, S. R. (2000) Cell. Mol. Life Sci. 57, 1874–1879. 30. Furnsinn, C., Englisch, R., Ebner, K., Nowotny, P., Vogle, C. & Waldhausl, W. (1996) Life Sci. 59, 1989–2000. 31. Pillay, T. S. & Makgbo, M. W. (1992) FEBS Lett. 308, 38–42. 32. Stapleton, S. R., Garlock, G., Foellmi-Adams, L. & Kletzein, R. F. (1997) Biochim. Biophys. Acta 1355, 259–269. 33. Mueller, A. S., Pallauf, J. & Rafael, J. (2003) J. Nutr. Biochem. 14, 637–647. 34. Esposito, F., Chirico, G., Gesualdi, N. M., Posadas, I., Ammendola, R., Russo, T., Cirino, G. & Cimino, F. (2003) J. Biol. Chem. 278, 20828–20834. 35. Boirie, Y. (2003) Trends Endocrinol. Metab. 14, 393–394. 36. Esposito, L. A., Melov, S., Panov, A., Cottrell, B. A. & Wallace, D. C. (1999) Proc. Natl. Acad. Sci. USA 96, 4820–4825. 37. Wallace, D. C. (1999) Science 283, 1482–1488. 38. Petersen, K. F., Befroy, D., Dufour, S., Dziura, J., Ariyan, C., Rothman, D. L., DiPietro, L., Cline, G. W. & Shulman, G. I. (2003) Science 300, 1140–1142. 39. Petersen, K. F., Dufour, S., Befroy, D., Garcia, R. & Shulman, G. I. (2004) N. Engl. J. Med. 350, 664–671. 40. Leiter, E. H. (2002) Diabetologia 45, 296–308. 41. Riu, E., Ferre, T., Hidalgo, A., Mas, A., Franckhauser, S., Otaegui, P. & Bosch, F. (2003) FASEB J. 17, 1715–1717. 42. Michael, M. D., Kulkarni, R. N., Postic, C., Previs, S. F., Shulman, G. I., Magnuson, M. A. & Kahn, C. R. (2000) Mol. Cell 6, 87–97. 43. Whiteman, E. L., Cho, H. & Birnbaum, M. J. (2002) Trends Endocrinol. Metab. 10, 444–451. 44. Cho, H., Mu, J., Kim, J. K., Thorvaldsen, J. L., Chu, Q., Crenshaw, E. B., III, Kaestner, K. H., Bartolomei, M. S., Shulman, G. I. & Birnbaum, M. J. (2001) Science 292, 1728–1731. 45. Moscow, J. A., Schmidt, L., Ingram, D. T., Guarra, J., Johnson, B. & Cowan, K. H. (1994) Carcinogenesis 15, 2769–2773. 46. Hu, Y. J. & Diamond, A. M. (2003) Cancer Res. 15, 3347–3351. 47. Forsberg, L., de Faire, U. & Morgenstern, R. (2001) Arch. Biochem. Biophys. 389, 84–93. 48. Joshi, R. L., Lamothe, B., Bucchini, D. & Jami, J. (1997) FEBS Lett. 20, 99–103. 49. Kim, J. H., Nishina, P. M. & Naggert, J. K. (1998) J. Basic Clin. Physiol. Pharmacol. 9, 325–345. 50. Mauvais-Jarvis, F. & Kahn, C. R. (2000) Diabetes Metab. 26, 433–448.
McClung et al.
PNAS 兩 June 15, 2004 兩 vol. 101 兩 no. 24 兩 8857
APPLIED BIOLOGICAL SCIENCES
We thank B. H. L. Chua and Y. S. Ho for the donation of transgenic mouse breeders for this study, J. H. Zhu for help in Western analysis, and K. J. Ellis for body composition analysis. This work was supported by National Institutes of Health Grant DK53018 (to X.G.L.) and in part by National Institutes of Health Grant DK52933 (to F.L.).
