The Endoplasmic Reticulum Chaperone Improves Insulin Resistance in Type 2 Diabetes

For Western blotting, anti-human ORP150 (1 μg/ml) (9), anti–insulin receptor substrate (IRS)-1 (Santa Cruz Biotechnology, Santa Cruz, CA), anti-phosphotyrosine (Calbiochem-Novabiochem, San Diego, CA), and anti–β-actin Ab IgG (Sigma Chemical, St. Louis, MO) were used. We used anti-human ORP150 IgG (5 μg/ml) (16) and anti-insulin IgG for immunohistochemical analysis. The construct of the rat insulin promoter was a kind gift from Dr. Richard D. Palmiter at the University of Washington. L6 cells were purchased from Japanese Collection of Research Bioresources (JCRB) Cell Bank (no. IFO50364; Osaka, Japan).

All procedures involving animals were approved by the Animal Care and Use Committee of Kanazawa University. ORP150 transgenic mice using a cytomegalovirus immediate early enhancer-chicken β-actin hybrid (CAG) promoter (16) were a kind gift from the HSP institute (17). ORP150 heterozygous knockout mice have been generated previously (18). Akita mice established from C57BL/6 mice were purchased from Japan SLC (Shizuoka, Japan). Transgenic mice using insulin promoter were generated at the Genome Information Research Center (Osaka University, Osaka, Japan) (10). The genotype of the mutant mice was determined by Southern blot analysis and PCR. These mice were crossed into the C57BL/6 background.

The levels of HbA_1c were measured from tail-vein blood using a DCA2000 analyzer (Bayer Medical, Tokyo, Japan) (19). The concentration of insulin was measured using an insulin enzyme-linked immunosorbent assay (ELISA) kit in accordance with the manufacturer’s instructions (Shibayagi, Shibukawa, Japan).

Intraperitoneal glucose tolerance tests (IPGTTs) and insulin tolerance tests (ITTs) were performed as described previously (20). Hyperinsulinemic-euglycemic clamp was performed as described previously (21,22). The rates of glucose appearance (R_a) and disappearance (R_d) were calculated according to Steele’s non–steady-state equations. Endogenous glucose production (EGP) was calculated as the difference between the tracer-derived R_a and exogenous infusion rates of glucose (GIRs) and tracer.

Animals that fasted overnight were anesthetized and injected with either saline or 5 IU human insulin via the inferior vena cava. The liver was obtained after 2 min and the skeletal muscle after 5 min. L6 cells were harvested after 5 min treatment with hydrogen peroxide. Protein samples from the liver, skeletal muscle, and L6 cells underwent immunoprecipitation for IRS-1 followed by Western blotting with an anti–IRS-1 antibody and anti-phosphorylated IRS-1 antibody.

At 6 weeks of age, C57BL/6 and Akita mice received an intraperitoneal injection of either vehicle or 30 mg/kg body wt α-lipoic acid over 5 days (23).

The formation of protein carbonyl groups was assessed by a OxyBlot protein oxidation detection kit (Integen) used in accordance with the manufacturer’s protocol.

Adenovirus coding for ORP150 in the sense (Ad/S-ORP150) or antisense (Ad/AS-ORP150) orientation and coding LacZ (AxCALacZ) have been generated previously (10). L6 cells were infected with an adenovirus at 100 multiplicities of infection (MOI) for 24 h as described previously.

Statistical analysis was performed by either an unpaired t test or an ANOVA followed by multiple comparison analysis using Newman-Kuehl’s equation. Where indicated, the data were analyzed by a two-way ANOVA followed by multiple contrast analysis. For nonparametric data, either a Kruskal-Wallis analysis or χ² analysis was applied.

ORP150 heterozygous mice (ORP150^−/+) were mated with heterozygous Akita mice (Ins2^WT/C96Y). Consistent with a previous report (16), ORP150 levels of ORP150^−/+ mice were reduced by ∼50% compared with those of their wild-type littermates (Fig. 1A), and similar results were obtained between ORP150^−/+Ins2^WT/C96Y and Ins2^WT/C96Y mice (data not shown). There was no significant difference in the body weights of ORP150^−/+Ins2^WT/C96Y and Ins2^WT/C96Y mice (Fig. 1B). The IPGTT showed there was no significant difference between ORP150^−/+ and their wild-type littermates (data not shown), whereas at 6 weeks the serum glucose levels of ORP150^−/+Ins2^WT/C96Y were significantly higher than those of Ins2^WT/C96Y mice (Fig. 1C). The measurement of HbA_1c showed a similar trend as IPGTT at 6 and 9 weeks; however, at 12 weeks there was no significant difference in HbA_1c levels between Ins2^WT/C96Y and ORP150^−/+Ins2^WT/C96Y mice (Fig. 1D). Contrary to our prediction, there was no significant difference in the insulin content of the pancreas of ORP150^−/+Ins2^WT/C96Y and Ins2^WT/C96Y mice (Fig. 1E).

ORP150 transgenic mice generated using the CAG promoter (ORP150^CAG) were mated with Ins2^WT/C96Y mice. The levels of ORP150 protein were significantly greater in ORP150^CAG compared with their wild-type littermates (Fig. 2A), and similar results were obtained between ORP150^CAGIns2^WT/C96Y and Ins2^WT/C96Y (data not shown). Consistent with a previous report (17), ORP150^CAGIns2^WT/C96Y had significantly lower body weights than Ins2^WT/C96Y mice (Fig. 2B). IPGTT showed that at 6 weeks, the glucose tolerance of ORP150^CAGIns2^WT/C96Y was greater than that of Ins2^WT/C96Y (Fig. 2C). HbA_1c showed a similar trend as IPGTT at 6 and 9 weeks; however, at 12 weeks there was no significant difference (Fig. 2D). There was no significant difference in the insulin content of the pancreas (Fig. 2E).

To determine whether overexpression of ORP150 in β-cells of the islets improves the glucose intolerance of Ins2^WT/C96Y, ORP150 transgenic mice were generated using the rat insulin promoter (ORP150^Ins), as described in research design and methods. Western blot analysis revealed that ORP150 levels in the pancreas of ORP150^Ins mice lines were significantly higher than those of their nontransgenic littermates. In comparison, there was no significant difference in the levels of ORP150 in the liver, skeletal muscle, white fat tissue, and brown fat tissue between the three lines of ORP150^Ins and their wild-type littermates (Fig. 3A). Furthermore, immunohistochemical analysis showed that ORP150 expression in the pancreas of ORP150^Ins was limited to the islets of Langerhans and overlapped with that of insulin (Figs. 3B and C), suggesting that, as predicted, the transgene was expressed in β-cells. There was no significant difference between the growth curves of Ins2^WT/C96Y and ORP150^InsIns2^WT/C96Y mice (Fig. 3D). Both glucose tolerance assessed by IPGTT (data not shown) and measurement of HbA_1c (Fig. 3E) were not significantly different between Ins2^WT/C96Y and ORP150^InsIns2^WT/C96Y mice. In addition, there was no significant difference in the pancreatic levels of insulin in Ins2^WT/C96Y and ORP150^InsIns2^WT/C96Y mice (data not shown). The results shown in Figs. 1 and 2 suggest that ORP150 improves the glucose intolerance of Akita mice but not the secretion of insulin.

To determine whether ORP150 expression was lessened by fasting (14), the pancreas (Fig. 4A), liver (Fig. 4B), and skeletal muscle (Fig. 4C) of C57BL/6 starved for indicated times were used for Western blotting with an anti-ORP150 antibody. In contrast to previous findings (14), there was no significant difference in the levels of ORP150 in the pancreatic tissue of control and starved mice, whereas starvation for 48 h significantly reduced the levels of ORP150 in the liver and skeletal muscle.

We assessed the levels of ORP150 in Akita mice. The levels of ORP150 transcript in the pancreas of Ins2^WT/C96Y were significantly higher than those of C57BL/6 (Figs. 4D and 4e). However, there was no significant difference in the levels of ORP150 protein between Ins2^WT/C96Y and C57BL/6 mice (Figs. 4F and G). Immunoblotting also revealed that the levels of ORP150 in the liver (Figs. 4H and I) and skeletal muscle (Figs. 4J and K) of Ins2^WT/C96Y were significantly greater compared with that of C57BL/6 mice at 6 weeks, whereas those of Ins2^WT/C96Y were significantly lower compared with those of C57BL/6 mice at 12 weeks.

To determine whether ORP150 is involved in the insulin sensitivity of Ins2^WT/C96Y, an ITT was performed. The ITT revealed that overexpression of ORP150 increased the insulin sensitivity of Ins2^WT/C96Y (Fig. 5A), whereas heterozygous disruption of ORP150 reduced insulin sensitivity (Fig. 5B). To further investigate these findings, ORP150^CAG and ORP150^−/+ mice underwent hyperinsulinemic-euglycemic clamp. As described previously (17), ORP150^CAG mice displayed the phenotype of myocardial degeneration and died of heart failure during the clamp test. The average GIR of 90–120 min in ORP150^−/+ was significantly lower than that of their littermates (Fig. 5C). EGP showed there was no significant difference between ORP150^−/+ and their wild-type littermates (data not shown). We assessed insulin signaling in Ins2^WT/C96Y, ORP150^CAGIns2^WT/C96Y, and ORP150^−/+Ins2^WT/C96Y mice. After overnight starvation and treatment with either saline or insulin, protein was extracted from the liver (Figs. 5D and E) and skeletal muscle (Figs. 4F and G) for immunoprecipitation with an anti–IRS-1 IgG, followed by Western blot analysis with an anti-phosphorylated IRS-1 IgG and anti–IRS-1. The levels of phosphorylated IRS-1 in ORP150^CAGIns2^WT/C96Y were increased compared with Ins2^WT/C96Y but were reduced in ORP150^−/+Ins2^WT/C96Y mice (Figs. 4E and G). These results indicate that ORP150 might play a role in the insulin sensitivity of the liver and skeletal muscle but not in the secretion of insulin.

We speculated that oxidative stress caused by hyperglycemia might induce ORP150 expression. To determine whether there was a relationship between oxidative stress and the induction of ORP150 in diabetes, Akita mice at 5 weeks were treated with α-lipoic acid (ALA), an antioxidant drug that improves insulin resistance (23). Northern blot analysis showed that ALA reduced the levels of ORP150 transcript in the liver (Figs. 6A and B) and skeletal muscle (Figs. 6C and D) of Akita mice. In addition, 10–40 μmol/l hydrogen peroxide induced ORP150 and GRP78 transcripts in L6 skeletal myoblast cells (Figs. 7A and B). In contrast, Chop transcripts were only induced by 100 μmol/l hydrogen peroxide (Figs. 7A and C). ORP150 and GRP78 transcript expression peaked 6–12 h after the treatment with hydrogen peroxide (data not shown). Immunoblot analysis showed the similar results as Northern blot analysis (data not shown). These data suggest that oxidative stress could induce ORP150 and GRP78. To further investigate the role of hydrogen peroxide in ER, L6 cells were treated with hydrogen peroxide for 24 h and then L6 cells were treated with 20 μmol/l hydrogen peroxide and tunicamycin (TM). TM induced ORP150 (Figs. 7D and E) and GRP78 (Figs. 7D and F), and hydrogen peroxide decreased their induction (Figs. 7D–F). These data are consistent with the in vivo data that levels of ORP150 in liver and muscle of Akita mice were decreased compared with those of wild-type mice at 12 weeks. Taken together, these data suggest that oxidative stress could induce ER chaperone; however, prolonged oxidative stress attenuates the induction of ER chaperone.

Furthermore, Western blotting using an anti-DNP (1-3 dinitrophenyl hydrazone) antibody revealed that overexpression of ORP150 reduced the levels of carbonylated protein in Akita mice (Fig. 8A) but that carbonylated protein was increased with heterozygous disruption of ORP150 (Fig. 8B). To further determine a possible role for ORP150, the levels of ORP150 were modulated in L6 myoblast cells using adenoviruses coding for ORP150 in the sense (Ad/S-ORP150) or antisense (Ad/AS-ORP150) orientation. Treatment of L6 cells with Ad/S-ORP150 significantly increased the levels of ORP150, whereas treatment with Ad/AS-ORP150 suppressed ORP150 levels (Fig. 8C). L6 cells infected with Ad/S-ORP150, Ad/AS-ORP150, or an adenovirus encoding LacZ (AxCALacZ) were treated with 20 μmol/l hydrogen peroxide and then used in Western blot analysis in conjunction with an anti-DNP antibody. The levels of oxidative protein in L6 cells treated with hydrogen peroxide were suppressed Ad/S-ORP150 but increased by Ad/AS-ORP150 (Fig. 8D). Using immunoprecipitation, insulin signals were further assessed in L6 cells treated with hydrogen peroxide. Ad/S-ORP150 consistently enhanced phosphorylation of IRS-1, but phosphorylation was suppressed by Ad/AS-ORP150 (Figs. 8E and F), suggesting that the ER chaperone might protect insulin signaling from oxidative stress in myoblast cells.

In this report, we demonstrated that ORP150 expression could be induced by hydrogen peroxide in myoblast cells and that overexpression of ORP150 using an adenovirus reduced oxidized protein. However, in a previous study, we showed that ORP150 had no influence on hydrogen peroxide–induced cell death (9,10). This discrepancy might arise from the different concentrations of hydrogen peroxide added to the cells. In previous reports, 100 μmol/l to 10 mmol/l hydrogen peroxide has been added to either 293 cells or primary neurons (9,10), whereas in this study, 10–40 μmol/l hydrogen peroxide induced ORP150 but produced no significant difference in cell growth and death (data not shown).

ORP150^CAGIns2^WT/C96Y showed lower body weight compared with the nontransgenic Akita (Ins2^WT/C96Y). It is possible that these differences in body weight might account for the differences in insulin sensitivity. To eliminate this possibility, we performed IPGTT and ITT using ORP150^CAG and weight-matched control. IPGTT showed no significant difference between ORP150^CAG and weight-matched control (see online appendix at http://diabetes.diabetesjournals.org). However, ITT showed that overexpression of ORP150 improved insulin sensitivity compared with weight-matched control (online appendix).

Using immunoblotting, we have shown that fasting reduces the levels of ORP150 in the liver and skeletal muscle (Figs. 4B and C). These data are consistent with a previous report showing that the expression of some ER chaperones, including ORP150, was lower in the liver of mice fed energy-restricted food compared with that of freely fed mice (24). In addition, we have shown that the levels of ORP150 in the liver and skeletal muscle of Akita mice were significantly greater compared with that of wild-type mice (Figs. 4H–K). Given that mutation of the insulin two gene is responsible for the phenotype of Akita mice, we expected hyperglycemia might secondarily induce ORP150 expression in the liver and skeletal muscle. This idea is consistent with our pilot data showing increased ORP150 expression in the liver and skeletal muscle of db/db diabetic mice or C57BL/6 mice treated with streptozotocin.

Although there is no significant difference in glucose levels during IPGTT between ORP150^−/+ and wild-type littermates (data not shown), hyperinsulinemic-euglycemic clamp revealed that insulin sensitivity of ORP150^−/+ was greater than that of wild-type littermates (Fig. 5C). As described in research design and methods, excess insulin was infused to maintain levels of glucose during clamp test, and insulin accelerates the process of mRNA translation (25) and results in increasing protein synthesis and ER stress. These reports led us to the idea that more ORP150 is required during clamp test than during IPGTT, and this is consistent with our pilot study that ORP150 levels in liver and muscle of db/db mice were greater than that of Akita mice (data not shown).

We have reported that ORP150 enhances the secretion of VEGF in wound healing and tumor formation (11,12), and, therefore, we expected that ORP150 might enhance the secretion of ACRP30/adiponectin, the expression of which correlates with insulin sensitivity (26), resulting in improved insulin resistance of Akita mice. To determine whether ORP150 increases the secretion of ACRP30, we assessed the levels of ACRP30 in serum and white/brown adipose tissues of Ins2^WT/C96Y, ORP150^−/+Ins2^WT/C96Y, and ORP150^CAGIns2^WT/C96Y by ELISA; however, no significant differences were shown between Ins2^WT/C96Y and ORP150^−/+Ins2^WT/C96Y or between Ins2^WT/C96Y and ORP150^CAGIns2^WT/C96Y(online appendix).

In contrast, the ER of proteins consumes oxidizing equivalents during the process of disulphide-bond formation (27,28), and ER stress in PERK^−/− or ATF4^−/− cells leads to the acute production of reactive oxygen species (ROS) through the accumulation of proteins oxidized by protein disulfide isomerase (29). These findings suggest that ER can reduce ROS in the liver and skeletal muscle, as well as improve insulin sensitivity in type 2 diabetes. As a molecular chaperone, ORP150 increases folding capacity and enhances protein secretion during ER stress (10,11,12,30). Consequently, ORP150 could enhance the role of the ER by reducing excess oxidizing equivalents and improving insulin sensitivity in type 2 diabetes.

Taken together, these data demonstrate that the ER chaperone ORP150 could remit insulin resistance caused by hyperglycemia by reducing oxidative stress and that it might be a novel therapeutic target to reduce the insulin resistance characteristic of type 2 diabetes.

We thank Dr. A. Koizumi for his kind advice about Akita mice; Dr. Y. Yamamoto (Kanazawa University), Dr. S. Kaneko (Kanazawa University), Dr. Y. Goriya (Nishino Hospital at Oyabe City, Japan), and Dr. H. Nakamura (Kyoto University) for their helpful suggestions; Dr. M. Okabe for generating the transgenic mice; and Y. Ichinoda for her excellent technical assistance.