Inhibition of Gastric Inhibitory Polypeptide Receptor Signaling in Adipose Tissue Reduces Insulin Resistance and Hepatic Steatosis in High-Fat Diet–Fed Mice

Obesity induces insulin resistance characterized by fasting and postprandial hyperinsulinemia to maintain euglycemia (1). It is one of the major risk factors in the progression to type 2 diabetes (2) and cardiovascular disease (3). Therefore, reduction of obesity can normalize hyperinsulinemia and ameliorate the progression of type 2 diabetes and arteriosclerosis.

Gastric inhibitory polypeptide/glucose-dependent insulinotropic polypeptide (GIP) is an incretin secreted from enteroendocrine K cells in response to meal ingestion (4–6) and potentiates insulin secretion through the GIP receptor (GIPR) expressed in pancreatic β-cells (7,8). High-fat diet (HFD) induces hypersecretion of GIP from enteroendocrine K cells (9,10), and we previously reported increased insulin secretion in response to GIP from β-cells in HFD-fed obese mice compared with lean mice fed a normal-fat diet (11), indicating that GIP plays a critical role in postprandial hyperinsulinemia under conditions of HFD-induced obesity. GIPR is also expressed in various extrapancreatic tissues, including the gastrointestinal tract, adipose tissue, heart, inner layers of the adrenal cortex, and several brain regions (12). In vitro studies have shown that GIP directly induces energy accumulation in adipose tissue by increasing lipoprotein lipase (LPL) expression through CREB and the CREB-regulated transcription coactivator 2 (TORC2)-mediated pathway and by increasing LPL enzyme activity and plasma membrane GLUT4 expression through an Akt-mediated pathway (13–15). Thus, there is a strong connection between GIP and obesity under HFD-fed conditions (16). Inhibition of GIPR signaling using systematic GIPR knockout (GIPR^−/−) mice, a GIP antagonist, and GIP immunoneutralization ameliorates obesity under HFD-fed conditions, suggesting that the condition might be due to lack of both direct and indirect GIP actions on adipose tissue (17–21). However, the importance of the direct effects of GIPR signaling on adipose tissue in vivo remains unclear.

In the current study, we generated adipose tissue-specific GIPR knockout (GIPR^adipo−/−) mice to evaluate whether the GIP receptor expressed in adipose tissue plays an important role in HFD-induced obesity and insulin resistance in vivo.

Targeting vector constructs were designed as long-loxP-GIPR(exon2–7)-FRT-neomycin (Neo)-FRT-loxP-short cassettes using the mouse B6N BAC clone (Fig. 1A). The targeting vector was injected into embryonic stem cells from C57BL/6 mice, and a Neo-resistant strain was established. Floxed GIPR (GIPR^fl/fl) mice were obtained from intercrosses with CAG-FLP transgenic mice. Adipocyte protein 2–Cre (Ap2-Cre) transgenic mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Heterozygous (GIPR^adipo+/−) mice were established by breeding Ap2-Cre mice with GIPR^fl/fl mice. Homozygous (GIPR^adipo−/−) mice were obtained by intercrossing with GIPR^adipo+/− mice. GIPR^fl/fl mice were used as wild-type (WT) mice in the current study. loxP and Cre genotyping were done with the specific primers listed in Supplementary Table 1. Mice were weaned at 4 weeks of age and fed a control fat diet (CFD; 10% fat by energy) or HFD (60% fat by energy) (Research Diets, Inc., New Brunswick, NJ) for 15 weeks. Animal care and procedures were approved by the Kyoto University Animal Care Committee (MedKyo 15298).

Total RNAs were extracted using TRIzol reagent (Invitrogen, Grand Island, NY). For cDNA synthesis, 1 μg total RNA was reverse transcribed using a PrimeScript RT reagent kit (Takara Bio, Shiga, Japan). Ex Taq (Takara Bio) was applied for RT-PCR. The control sample was prepared in the absence of tissue cDNA. SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA) was applied for quantitative real-time PCR using an ABI StepOnePlus Real-Time PCR System (Applied Biosystems). The signals of the products were standardized against GAPDH or β-actin signals in each sample. Primer pairs for PCR are listed in Supplementary Table 1. Microarray analysis was performed using SurePrint G3 Mouse Gene Expression 8 × 60 K (Agilent Technologies, Santa Clara, CA).

Insulin, C-peptide, total GIP, interleukin 6 (IL-6), adiponectin, and leptin levels were measured by insulin ELISA kit (Shibayagi, Gunma, Japan), C-peptide ELISA kit (Shibayagi), total GIP ELISA kit (Millipore, Billerica, MA), IL-6 ELISA kit (R&D Systems, Minneapolis, MN), adiponectin ELISA kit (Otsuka Pharmaceutical, Tokyo, Japan), and leptin ELISA kit (Shibayagi), respectively.

After 16 h of fasting, oral glucose tolerance tests (OGTTs) (1 g/kg body weight) were performed. Blood glucose levels were measured by the glucose oxidase method (Sanwa Kagaku Kenkyusho, Nagoya, Japan). The HOMA of insulin resistance (HOMA-IR) was calculated (22). For the insulin tolerance test (ITT), human regular insulin (Novo Nordisk, Copenhagen, Denmark) at a dose of 0.75 units/kg body weight (for CFD-fed mice) or 1 unit/kg body weight (for HFD-fed mice) was injected subcutaneously after 2 h of fasting. For the pyruvate tolerance test (PTT), pyruvate at a dose of 1 g/kg body weight was injected subcutaneously to HFD-fed mice after 16 h of fasting.

Fasting plasma samples were taken from the mice after 21 weeks of the HFD. LDL-cholesterol (C), HDL-C, triglyceride (TG), and nonesterified fatty acids (NEFA) levels were measured by LDL-C assay kit (Sekisui Medical Co., Ltd., Tokyo, Japan), HDL-C assay kit (Sekisui Medical Co., Ltd), TG assay kit (Wako, Osaka, Japan), and NEFA assay kit (Eiken Chemical Co., Ltd., Tokyo, Japan). Hepatic lipids were extracted as described previously (23), and TG, and total cholesterol levels were measured using commercial kits (Sekisui Medical Co., Ltd.). Hepatic lipid content was defined as per gram of the liver tissue weight.

Concentrations of O₂ and CO₂ were measured using an Alco System model 2000 (Alco System, Chiba, Japan), and VO₂, VCO₂, energy expenditure, and fat oxidation were calculated (24). The locomotor activity of the mice was measured using an automated activity counter (NSAS01; Neuroscience, Tokyo, Japan).

Mice were anesthetized with pentobarbital, fixed in a chamber, and transaxially scanned using a La Theta (LCT-100M) experimental animal computed tomography (CT) system (Hitachi Aloka Medical, Ltd., Tokyo, Japan). CT scan was performed from xiphisternum to sacrum. Contiguous 1-mm slice images of the trunk were used for quantitative assessment of visceral fat, subcutaneous fat, lean mass (visceral mass without visceral fat), liver volume, and fat content in the liver using La Theta 1.00 software (25). Liver density levels were measured by CT scan, and fat content in the liver was calculated from density data of fat (100% fat) and muscle (0% fat).

Liver and visceral fat samples were fixed in 10% formalin buffer, embedded in paraffin, and sectioned at 3 µm. The paraffin sections of the liver were stained with hematoxylin and eosin. Images were taken using a microscope with the BZ-8100 system (KEYENCE Corp., Osaka, Japan). The sections of visceral fat were blocked with 3% BSA and incubated overnight at 4°C with a monoclonal rabbit anti–perilipin-2 antibody (LSBio, Seattle, WA) and afterward with a secondary antibody at room temperature for 1 h. After immunostaining, the mean adipocyte size (surface areas of 15 representative adipocytes per mouse) was analyzed by BZ Analyzer software (KEYENCE Corp.).

Frozen tissue extracts were minced and homogenized in cold radioimmunoprecipitation assay buffer consisting of 50 mmol/L Tris-HCl (pH 7.4), 150 mmol/L NaCl, 1% NP-40, 0.1% SDS, 1 mmol/L EDTA, phosphatase inhibitor cocktail (Sigma-Aldrich, St. Louis, MO), and protease inhibitor cocktail (Roche, Basel, Switzerland), and then centrifuged at 12,000g for 5 min. Equal amounts of proteins were subjected to immunoblot analysis. Anti-Akt and anti-phosphorylated-Akt (Ser473) antibodies were used (Cell Signaling, Danvers, MA).

3T3-L1 cells (Japanese Collection of Research Bioresources Cell Bank, Osaka, Japan) was induced by culture in 10% FBS/DMEM containing 10 µg/mL insulin, 0.5 μmol/L isobutyl methylxanthine, and 2.5 μmol/L dexamethasone for 2 days. After another 2 days of incubation in 10% FBS/DMEM with 10 µg/mL insulin, the medium was changed every 2 days. On day 9 of differentiation, 3T3-L1 adipocytes were pretreated with 100 nmol/L human GIP (Peptide Institute Inc., Osaka, Japan) and 2.5 μmol/L dipeptidyl peptidase 4 inhibitor K579 (Wako, Osaka, Japan) for 3 h and then treated with 100 nmol/L human GIP and 2.5 μmol/L dipeptidyl peptidase 4 inhibitor for 24 h in the presence or absence of mouse tumor necrosis factor-α (TNF-α) (5 or 10 ng/mL; R&D Systems).

All data are expressed as the mean ± SE. Data were analyzed by Student t test and one-way ANOVA with the Bonferroni post hoc test. Statistical analyses were performed using IBM SPSS Statistics 20.0 software (IBM, Armonk, NY). A significant difference was considered at P < 0.05.

The gene encoding mouse GIPR contains 14 exons (11). Cre-mediated recombination occurred between exon 2, in which there is the start codon, and exon 7 in adipose tissue because of the presence of both the floxed GIPR gene and Ap2-Cre transgene (Fig. 1A). GIPR is expressed in various tissues, including pancreatic β-cells, gastrointestinal tract, adipose tissue, and several brain regions (12). RT-PCR was performed using specific primers for amplification of the region between exon 5 and exon 6. In WT mice, GIPR mRNA was detected in subcutaneous fat tissue and islets but not in liver and skeletal muscle (Fig. 1B). The expression levels of GIPR mRNA in visceral and subcutaneous fat were dramatically decreased by 90% in GIPR^adipo−/− mice compared with those in WT mice (Fig. 1C). The expression levels in islet, brain, and duodenum were not significantly different between the two groups. These results indicate that the Cre-loxP system specifically operated in the presence of Cre protein in the adipose tissues of GIPR^adipo−/− mice.

In the CFD-fed condition, body weight gain was not different between WT and GIPR^adipo−/− mice (Fig. 1D). OGTT and ITT were performed after 15 weeks of CFD feeding. There were no statistical differences in blood glucose levels (Fig. 1E), insulin levels (Fig. 1F), and total GIP levels (Fig. 1G) between WT and GIPR^adipo−/− mice. ITT data showed no statistical difference between the two groups (Fig. 1H).

Under HFD feeding, GIPR^adipo−/− mice showed significantly lower body weight gain compared with that of WT mice (Fig. 2A). Blood glucose levels and total GIP levels from ad libitum fed mice were not different between the two groups (Fig. 2B). However, insulin levels were significantly lower in GIPR^adipo−/− mice than those in WT mice after 5 weeks of HFD feeding (Fig. 2B). We then measured blood glucose, insulin, C-peptide, and total GIP levels in another ad libitum fed cohort (Fig. 2C). After 12 weeks of HFD feeding, body weight was significantly reduced in GIPR^adipo−/− mice compared with that in WT mice. Blood glucose and total GIP levels were not different between the two groups. C-peptide levels and insulin levels were significantly reduced in GIPR^adipo−/− mice compared with those in WT mice. The difference in food intake between the two groups was not significant (Fig. 2D).

To determine the effect of GIPR deficiency in adipocytes on glucose tolerance, an OGTT was performed after 15 weeks of HFD feeding. Blood glucose levels tended to be lower in GIPR^adipo−/− mice than those in WT mice (Fig. 2E). Moreover, insulin levels were significantly lower at 0 and 120 min in GIPR^adipo−/− mice than those in WT mice (Fig. 2F). However, area under the curve (AUC) of glucose and the AUC of insulin were not significantly different between the two groups. Total GIP levels and the AUC of GIP were not different between the two groups (Fig. 2G). Fasting TG levels were significantly lower in GIPR^adipo−/− mice than those in WT mice (Table 1). LDL-C, HDL-C, and NEFA levels were not different between the two groups.

The values for HOMA-IR, which is an index of insulin resistance under fasting conditions, were significantly lower in GIPR^adipo−/− mice than those in WT mice (Fig. 2H). ITT data showed that blood glucose levels were significantly decreased in GIPR^adipo−/− mice compared with those in WT mice (Fig. 2I). Glucose levels during PTT were significantly lower in HFD-fed GIPR^adipo−/− mice than those in HFD-fed WT mice (Fig. 2J). These results suggest that GIPR deficiency in adipose tissue reduces hepatic insulin resistance under HFD condition. Akt phosphorylation after insulin treatment in adipose, liver, and skeletal muscle tissues was evaluated with Western blot analysis. Akt phosphorylation was increased in all of these tissues of GIPR^adipo−/− mice compared with those of WT mice (Fig. 2K).

Differences in energy expenditure, and fat oxidation of both dark and light phases between the two groups were not significant (Fig. 3A). Locomotor activity was also not significantly different in the two groups (Fig. 3B).

We next assessed whether the difference in body weight between WT and GIPR^adipo−/− mice was related to morphological alterations of adipose tissue. The differences in subcutaneous and visceral fat volumes between the two groups were not significant (Fig. 3C and E). Furthermore, we compared adipocyte size by immunostaining of perilipin-2, a protein expressed on the surface of lipid droplets. Adipocyte size was not significantly different between the two groups (Fig. 3D).

The lean mass volume was estimated by CT analysis and was significantly lower in GIPR^adipo−/− mice compared with that in WT mice (Fig. 3C). We then evaluated the liver in WT and GIPR^adipo−/− mice. The liver volume and fat content measured by CT scan analysis were significantly lower in GIPR^adipo−/− mice than in WT mice (Fig. 3F). A significant difference was also observed in liver weights between the two groups (Fig. 3G). Histological analysis of the liver showed that the size and number of lipid droplets were smaller in GIPR^adipo−/− mice than those in WT mice (Fig. 3H). Moreover, TG content in the liver was significantly lower in GIPR^adipo−/− mice than that in WT mice, but total cholesterol content was not different (Fig. 3I). These results indicate that GIPR deficiency in adipose tissue reduces hepatic steatosis under HFD-fed conditions.

Deficiency of GIPR in adipose tissue reduced insulin resistance (Fig. 2G and H) and hepatic steatosis (Fig. 3F–I) in HFD-fed mice, although GIPR mRNA was not detected in the liver or skeletal muscle by RT-PCR (Fig. 1B). To identify the molecules, we performed microarray analysis of adipose tissue from HFD-fed WT and GIPR^adipo−/− mice and found that IL-6 expression was decreased in the adipose tissue of GIPR^adipo−/− mice (WT [n = 3] 67.6 ± 20.9 vs. GIPR^adipo−/− mice [n = 3] 29.7 ± 9.4; P < 0.05). We then examined the mRNA expression levels of inflammatory cytokines, including IL-6, in the adipose tissue of WT and GIPR^adipo−/− mice by quantitative real-time PCR. Expression levels of TNF-α, monocyte chemoattractant protein 1, and IL-1β mRNA were not different between the two groups under HFD-fed conditions. However, IL-6 mRNA expression levels were significantly lower in HFD-fed GIPR^adipo−/− mice than those in HFD-fed WT mice (Fig. 4A). Plasma IL-6 levels were significantly lower in HFD-fed GIPR^adipo−/− mice than those in HFD-fed WT mice (Fig. 4B). Expression levels of adiponectin and leptin mRNAs were not significantly different between the two groups (Fig. 4C). The plasma levels of adiponectin and leptin were also not different between the two groups under HFD-fed conditions (Fig. 4D).

A previous study reported that CREB enhances IL-6 expression induced by nuclear factor (NF)-κB through TNF-α receptor signaling (26). GIP is well known to increase intracellular cAMP through GIPR and induce CREB activation in adipocytes (27). We next investigated the effects of GIP on IL-6 mRNA expression and production in differentiated 3T3-L1 adipocytes. IL-6 mRNA expression was not enhanced by GIP stimulation (Fig. 4E). Production of IL-6 was increased significantly in the cells after GIP treatment, but the increase in IL-6 levels was very small (Fig. 4F). However, GIP significantly enhanced IL-6 mRNA expression (Fig. 4G) and IL-6 production (Fig. 4H) in 3T3- L1 adipocytes in the presence of TNF-α, indicating that GIP enhances TNF-α–induced IL-6 expression and production in 3T3-L1 adipocytes.

Suppressor of cytokine signaling (SOCS) is known to be induced by cytokine receptors via Janus kinase/STAT3 signaling and to suppress intracellular cytokine signaling (28). Recent studies have reported that SOCS1 and SOCS3 inhibit insulin signaling (28,29). On the one hand, the expression levels of SOCS1 mRNA under HFD-fed conditions tended to be lower in adipose and liver tissues of GIPR^adipo−/− mice than in those of WT mice, but the difference was not significant (Fig. 5A). On the other hand, expression levels of SOCS3 mRNA were significantly decreased in adipose and liver tissues of GIPR^adipo−/− mice compared with those in WT mice (Fig. 5B). We next examined SOCS1 and SOCS3 mRNA expression levels in differentiated 3T3-L1 adipocytes. GIP did not increase SOCS1 mRNA expression in the absence or presence of TNF-α (Fig. 5C and E). SOCS3 mRNA expression was significantly increased in the cells after GIP treatment in the absence of TNF-α, but the increase in SOCS3 levels was very small (Fig. 5D). However, GIP remarkably enhanced SOCS3 mRNA expression in 3T3-L1 adipocytes in the presence of TNF-α (Fig. 5F).

SOCS1 and SOCS3 also have been reported to enhance promoter activity of SREBP-1c, a key regulator of fatty acid synthesis in the liver (30). The mRNA expression levels of SREBP-1c were significantly lower in the liver of HFD-fed GIPR^adipo−/− mice compared with those of WT mice, but the expression levels of carbohydrate-responsive element-binding protein (ChREBP) were not (Fig. 5G). The mRNA expression levels of stearoyl-CoA desaturase 1, which is activated by SREBP-1c and induces fatty acid synthesis, were also significantly decreased in the liver of GIPR^adipo−/− mice compared with those of WT mice, without a reduction in the mRNA expression levels of ChREBP (Fig. 5G and H). The mRNA expression levels of acetyl-CoA carboxylase, which is also a fat synthesis enzyme, tended to be lower in GIPR^adipo−/− mice, but the difference was not significant.

In the current study, we used GIPR^adipo−/− mice to investigate direct GIP action in adipose tissue in vivo. GIPR^adipo−/− mice showed lower body weight, reduced insulin resistance, and improved hepatic steatosis compared with WT mice under the HFD-fed condition.

We previously showed that fat accumulation in adipose tissue is inhibited in GIPR^−/− mice compared with WT mice under the HFD-fed condition (17). GIP potentiates insulin secretion from β-cells, and hypersecretion of insulin is known to induce obesity (31). GIPR is also expressed in adipose tissue; however, the importance of the direct effects of GIPR signaling on adipose tissue in vivo remains unclear. In the current study, fat accumulation in adipose tissue did not differ between HFD-fed WT and HFD-fed GIPR^adipo−/− mice, suggesting that GIPR in adipose tissue is not involved in fat accumulation in vivo and that the indirect effect of insulin hypersecretion through GIPR on β-cells plays a critical role in HFD-induced obesity. However, GIP has been reported to stimulate fat and glucose accumulation in differentiated 3T3-L1 cells and human adipocytes in vitro (13–15). Because insulin sensitivity is increased in GIPR^adipo−/− mice in our study, the lack of apparent decrease in the fat volume of these GIPR^adipo−/− mice is possibly a result of the counteracting effects of enhanced insulin signaling and inhibition of GIP signaling in adipose tissue.

Although the differences in food intake, locomotor activity, and energy expenditure were not significant, body weight gain was significantly lower in GIPR^adipo−/− mice than that in WT mice. It is possible that this effect is due to reduced absorption of calories from the intestine in GIPR^adipo−/− mice or to limits of the methods used to detect differences in body weight between WT and GIPR^adipo−/− mice. However, CT scan analysis revealed the lean body mass was significantly lower in GIPR^adipo−/− mice than in WT mice, and the liver weight and fat content in the liver were decreased in GIPR^adipo−/− mice. The contribution of reduced liver weight to the overall reduction in body weight of GIPR^adipo−/− mice compared with WT mice under the HFD-fed condition was ∼45%. Thus, the reduction in body weight is partly due to the loss of liver weight in GIPR^adipo−/− mice. GIPR^−/− mice showed reduced hepatic steatosis compared with WT mice under the HFD-fed condition (32). In addition, body weight was increased GIPR^−/− mice with an Ap2 promoter-GIPR transgene compared with the control GIPR^−/− mice, which was not due to an increase in fat volume but rather to an increase in lean mass (33). Although liver morphology and liver weight were not evaluated in the study, GIPR signaling in adipose tissue may affect lean body mass such as liver. However, which tissues other than liver contribute to the difference in body weight between WT and GIPR^adipo−/− mice under HFD-fed condition remains unclear.

Key molecules regulated by GIPR signaling in adipose tissue might be related to HFD- induced insulin resistance and hepatic steatosis, because insulin resistance and hepatic steatosis were reduced in HFD-fed GIPR^adipo−/− mice. Microarray analysis of adipose tissues revealed that expression of the inflammatory cytokine IL-6 was particularly decreased in adipose tissue of GIPR^adipo−/− mice. IL-6 is involved in inflammation and in regulation of metabolic processes (34,35). Circulating levels of IL-6 are positively associated with obesity in humans (36), and adipose tissue–derived IL-6 has a critical role in induction of insulin resistance and hepatic steatosis (37,38). Previous in vitro studies have reported that GIP directly increases IL-6 expression levels in adipose tissue (39,40). That CREB activated by the β-adrenergic receptor enhances IL-6 expression induced by NF-κB through TNF-α receptor signaling in astrocytes has also been reported (26). GIP also activates CREB in adipose tissue (27). In the current study using differentiated 3T3-L1 adipocytes, GIP-induced IL-6 expression and production were significantly augmented by TNF-α in a dose-dependent manner, suggesting that GIP enhances IL-6 mRNA expression in adipose tissue in the presence of inflammatory cytokines such as TNF-α. IL-6 is secreted from macrophage as well as adipose tissue (41). A recent study reported that GIPR is expressed in macrophages and that pharmacological doses of GIP prevent macrophage-induced arteriosclerosis (42).

The GIPR^adipo−/− mice used in this study were generated from floxed GIPR mice and Ap2-Cre mice. Ap2-Cre mice have been widely used for deletion of floxed sequences in adipose tissue. Ap2 is expressed in macrophage and in adipose tissue (43). When we evaluated GIPR mRNA expression in isolated macrophages, GIPR mRNA expression by RT-PCR was not detected (data not shown). In addition, we evaluated IL-6 mRNA expression in isolated macrophages from WT and GIPR^adipo−/− mice under the HFD-fed condition. The difference in the expression levels of IL-6 mRNA between the two groups was not significant (IL-6 mRNA/GAPDH mRNA: 2.70 ± 1.00 [WT] vs. 1.523 ± 0.50 [GIPR^adipo−/−], n = 5–7; P = 0.37). Therefore, it is unlikely that GIPR in macrophages contributes to the phenotype of GIPR^adipo−/− mice, which have physiological plasma GIP levels.

In this study, we showed that GIP increases IL-6 mRNA expression and production in the presence of TNF-α using adipocytes differentiated from cell line 3T3-L1, but we did not evaluate them using the adipose tissues isolated from WT and GIPR^adipo−/− mice. To more clearly show the effect of GIP on IL-6 expression and release in vivo, we need to evaluate a primary culture of adipocytes isolated from WT and GIPR^adipo−/− mice.

SOCS protein is induced through some cytokine receptors, such as IL-6, TNF-α, and Toll-like receptors, and attenuates cytokine signal transduction (44,45). Seven SOCS proteins have been identified (28), and SOCS1 and SOCS3 were recently reported to negatively regulate the insulin-signaling pathway in ways such as by competition for insulin receptor substrate binding sites on the insulin receptor, inhibition of insulin receptor kinase activity, and proteasomal degradation of insulin receptor substrate (28,29,46). In particular, SOCS3 signaling is located downstream of the IL-6 receptor (45). Our in vitro experiments in differentiated 3T3-L1 adipocytes showed that GIP-induced mRNA expression of SOCS3 in the presence of TNF-α and that IL-6 expression and production were increased remarkably under this condition. In addition, expression levels of SOCS3 mRNA were decreased in adipose and liver tissues of GIPR^adipo−/− mice. These results suggest that a decrease in SOCS3 expression might increase Akt phosphorylation after insulin receptor stimulation in adipose and liver tissues. However, SOCS3 mRNA expression in skeletal muscle was not different between the two groups of mice, although Akt phosphorylation was increased in the skeletal muscle of GIPR^adipo−/− mice. Further study is needed to clarify the mechanism of increase in insulin-induced Akt phosphorylation in the skeletal muscle of GIPR^adipo−/− mice. SOCS1 and SOCS3 are also reported to bind to the gene promoter of SREBP-1c, which has critical roles in fat synthesis of the liver and increases SREBP-1c mRNA expression (47,48). Inhibition of the expression of SOCS1 or SOCS3 in db/db mice improves their hepatic steatosis via SREBP-1c expression (30). In the current study, SREBP-1c mRNA and SOCS3 mRNA expression was decreased in the liver of HFD-fed GIPR^adipo−/− mice. These results suggest that the reduction of SOCS3 expression levels might improve hepatic steatosis through reduction of SREBP-1c.

In this study, 60% HFD feeding for 15 weeks increased plasma GIP levels, body weight, and insulin resistance more than CFD feeding not only in WT mice but also in GIPR^adipo−/− mice, suggesting that insulin hypersecretion through GIPR on β-cells plays a critical role in HFD-induced obesity. Although 60% HFD feeding for 15 weeks is not physiological, dietary fat is a strong stimulant of GIP secretion in human (49,50). We previously showed that GIP secretion after glucose ingestion is positively correlated with insulin resistance, BMI, and insulin secretion in healthy subjects (51), implicating the role of insulin hypersecretion through GIP in HFD-induced obesity in human.

Although GIPR function in human adipose tissue is unclear in vivo, in vitro study shows that GIP induces IL-6 mRNA expression and potentiates IL-6 secretion in the presence of TNF-α in human adipocytes (39). These results are consistent with our data showing that IL-6 mRNA expression is decreased in the adipocytes of HFD-fed GIPR^adipo−/− mice. Thus, GIPR^adipo−/− mice might be useful for studying the role of GIPR expressed in adipose tissue in HFD-induced obesity and insulin resistance in vivo. These data suggest that inhibition of GIP signaling may have potential as a therapeutic target in the treatment of HFD-induced insulin resistance and hepatic steatosis.

In conclusion, GIPR signaling in adipose tissue plays an important role in HFD-induced insulin resistance and hepatic steatosis in vivo, which may involve IL-6 signaling.

Acknowledgments. The authors thank Shoichi Asano, Yasuo Oguri, and Toshihiro Nakamura (Department of Diabetes, Endocrinology and Nutrition, Graduate School of Medicine, Kyoto University) for technical support.

Funding. This study was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology (MEXT); Japan Society for the Promotion of Science; Ministry of Health, Labour, and Welfare; Ministry of Agriculture, Forestry and Fisheries; Japan Diabetes Foundation; Japan Association for Diabetes Education and Care; Merck Sharp & Dohme (MSD); Novo Nordisk Pharma; and Banyu Life Science Foundation International.

Duality of Interest. N.I. receives fees from Takeda Pharmaceutical Co., Ltd.; Mitsubishi Tanabe Pharma Corporation; MSD; Sanofi; Novartis Pharma; Dainippon Sumitomo Pharma; Kyowa Hakko Kirin Co., Ltd.; Eli Lily Japan; Shiratori Pharmaceutical; Roche Diagnostics; Japan Tobacco; Nippon Boehringer Ingelheim Co., Ltd.; Astellas Pharma Inc.; Daiichi Sankyo Company, Ltd.; Ono Pharmaceutical Co., Ltd.; and Taisho Toyama Pharmaceutical Co., Ltd. outside the submitted work. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. E.J. planned the study, researched data, contributed to discussions, and wrote, reviewed, and edited the manuscript. N.H. planned the study, contributed to discussions, and reviewed and edited the manuscript. S.Y., T.F., and D.T. researched data and contributed to discussions. K.I., A.S., K.Sh., T.H., K.Su., and A.H. contributed to discussions. N.I. contributed to discussions and reviewed and edited the manuscript. All authors reviewed the results and approved the final version of the manuscript. N.I. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented at the Keystone Symposium on Diabetes, Kyoto, Japan, 25–29 October 2015; the Incretin 2015 Conference, Vancouver, Canada, 29 July to 1 August 2015; the 50th European Association for the Study of Diabetes, Vienna, Austria, 15–19 September 2014; and at the 5th Annual Meeting of the Asian Association for the Study of Diabetes, Seoul, Korea, 6–9 November 2013.