Gastric inhibitory polypeptide (GIP) is an incretin released from enteroendocrine K cells in response to nutrient intake, especially fat. GIP is one of the contributing factors inducing fat accumulation that results in obesity. A recent study shows that fatty acid-binding protein 5 (FABP5) is expressed in murine K cells and is involved in fat-induced GIP secretion. We investigated the mechanism of fat-induced GIP secretion and the impact of FABP5-related GIP response on diet-induced obesity (DIO). Single oral administration of glucose and fat resulted in a 40% reduction of GIP response to fat but not to glucose in whole body FABP5-knockout (FABP5−/−) mice, with no change in K cell count or GIP content in K cells. In an ex vivo experiment using isolated upper small intestine, oleic acid induced only a slight increase in GIP release, which was markedly enhanced by coadministration of bile and oleic acid together with attenuated GIP response in the FABP5−/− sample. FABP5−/− mice exhibited a 24% reduction in body weight gain and body fat mass under a high-fat diet compared with wild-type (FABP5+/+) mice; the difference was not observed between GIP-GFP homozygous knock-in (GIPgfp/gfp)-FABP5+/+ mice and GIPgfp/gfp-FABP5−/− mice, in which GIP is genetically deleted. These results demonstrate that bile efficiently amplifies fat-induced GIP secretion and that FABP5 contributes to the development of DIO in a GIP-dependent manner.
- gastric inhibitory polypeptide
- fatty acid-binding protein 5
obesity has been recognized as a worldwide problem, especially for developing insulin resistance and increasing the risk of type 2 diabetes (11). The increased adipose tissue associated with obesity secretes various kinds of cytokines, adipokines, and lipokines (2) and triggers systemic inflammatory response, resulting in the disturbance of metabolic homeostasis (9).
Gastric inhibitory polypeptide (GIP) is one of the incretins, peptide hormones released from the gastrointestinal tract into circulation in response to nutrient ingestion, that potentiates glucose-stimulated insulin secretion (5, 16, 35). Dietary lipid is a very strong stimulant of GIP secretion. For example, the peak value of GIP increase in response to a high-fat meal (450 kcal containing 33.3% of fat) is three times higher than that in the 75-g oral glucose tolerance test (OGTT) in human subjects, suggesting that fat content in a mixed meal strongly stimulates GIP secretion (45). GIP binds to the GIP receptor (GIPR) on the surface of pancreatic β-cells to stimulate insulin secretion (34). GIP is considered to increase the volume of adipose tissue by two major pathways: directly by binding to GIPR located on the adipocytes (15, 39) and indirectly by accelerating fat deposition and expansion of fat depots by increasing insulin secretion from β-cells (32). Studies of GIPR knockout (Gipr−/−) mice (27) describe GIP as an obesity-promoting factor in high-fat diet (HFD) conditions and show that deletion of GIPR signaling causes resistance to diet-induced obesity (DIO) (26). Additionally, we have reported that partial reduction of GIP alleviates obesity and lessens the degree of insulin resistance without exacerbating glucose tolerance under HFD conditions (28). These findings suggest that regulation of GIP secretion, especially after fat intake, is a promising therapeutic approach to obesity and type 2 diabetes. However, the precise mechanism of GIP secretion has remained unclear mainly because of the inability to isolate GIP-producing enteroendocrine K cells from intestinal epithelium. Recently, we generated GIP-green fluorescent protein (GFP) knock-in mice in which K cells are labeled by enhanced GFP (EGFP). On the basis of microarray analysis of K cells isolated from GIP-GFP knockin heterozygous (GIPgfp/+) mice, we demonstrate that transcriptional regulatory factor X6 (Rfx6) is expressed exclusively in K cells, and that Rfx6 is involved in GIP hypersecretion in DIO (43).
Very recently, It was reported that fatty acid-binding protein 5 (FABP5) is expressed in K cells and is involved in fat-induced GIP secretion (38). FABP has been known as an intracellular chaperon that transports long-chain fatty acid (LCFA) into various organelles (42). Human enterocytes are known to coexpress FABP1 and FABP2, which are thought to play distinct functional roles in fatty acid metabolism (17). Variation in the FABP2 gene in Pima Indians (6) is reported to alter fatty acid transport in the reconstitution study using human cell lines (1). FABP4 and FABP5 have been reported to be expressed in adipocytes and to play a critical role in regulating fat accumulation and progression of insulin resistance. FABP4/FABP5 double-knockout (FABP4−/−/FABP5−/−) mice have exhibited a dramatic phenotype of increased insulin sensitivity and inhibited progression of atherosclerosis (23). FABP5 single-knockout (FABP5−/−) mice have also showed reduced body weight gain on HFD compared with wild-type (FABP5+/+) mice, albeit with a milder phenotype than that of FABP4−/−/FABP5−/− mice (24). Given that FABP5 expressed in GIP-producing K cells is involved in fat-induced GIP secretion (38) and that FABP5−/− mice are resistant to the progression of DIO (24), we hypothesized that FABP5 contributes to the development of DIO via GIP secretion. In addition, because FABPs are known as intracellular proteins, we thought there might be uptake of fatty acid into K cells via transport proteins and/or bile, as is the case in intestinal epithelial cells.
In this study, we examined the mechanism of fat-induced GIP secretion, focusing on the pathway through which ingested fatty acid accesses intracellularly expressed FABP5 in the K cell and exploring the impact of the FABP5-related GIP response on DIO by comparing the degree of obesity in FABP5+/+ and FABP5−/− mice with that in GIP-GFP knock-in homozygous (GIPgfp/gfp)-FABP5+/+ mice and GIPgfp/gfp-FABP5−/− mice, in which GIP is genetically deleted.
MATERIALS AND METHODS
All animals used in this study were C57BL/6 mice maintained under conditions of a 12:12-h light-dark cycle, with free access to water and food unless otherwise indicated. All mice used in individual experiments were age-matched at the beginning of the experiments. FABP5−/− mice (31) and glucagon-GFP knock-in mice (Gcggfp/+) (8) were generously provided by Dr. Y. Owada of Yamaguchi University and Dr. Y. Hayashi of Nagoya University, respectively. GIPgfp/+ mice and GIPgfp/gfp mice were generated as described previously (28). FABP5+/+ and FABP5−/− mice (n = 4) were fed a control/fat diet (CFD; 10% fat, 20% protein, and 70% carbohydrate by energy) or HFD (60% fat, 20% protein, and 20% carbohydrate by energy) (Research Diets, New Brunswick, NJ) for 8 wk from the age of 10 wk. GIPgfp/gfp-FABP5+/+ and GIPgfp/gfp-FABP5−/− mice (n = 4) were fed a HFD for 8 wk. Body weight change of the mice was recorded weekly. Under HFD feeding, food intake was assessed using metabolic cages. Measurement of body fat composition was performed after the recording of body weight by a previously described method (28). In brief, after the mice were anesthetized with an intraperitoneal injection of pentobarbital sodium (30 mg/kg), total body fat, subcutaneous fat, and visceral fat were measured by a computed tomography (CT) scan (A La Theta LCT-100; Hitachi Aloka, Tokyo, Japan). The obtained images were analyzed by A La Theta software, version 1.00. After 8 wk of HFD feeding, the mice were euthanized by cervical dislocation. The white (visceral) adipose tissue was harvested, frozen immediately in liquid nitrogen, and stored at −80°C until further use. All animal protocols were approved by the Animal Care Committee of Kyoto University.
Isolation of K cells from mouse intestinal epithelium.
The procedure of isolation of K cells from murine intestinal epithelium was described previously (43). Briefly, the upper small intestine of a 10-wk-old mouse was removed and cut into several round pieces and tied on one side with a thread. The pouch-like intestine was injected with Hanks' balanced salt solution. The intestinal epithelium was digested by collagenase P (Roche Diagnostics, Mannheim, Germany). Afterward, it was centrifuged at 180 g for 5 min, resuspended in PBS twice, and filtered with a cell strainer (352340, Falcon Cell Strainer; BD Biosciences). The opening size of the strainer was 40 μm. GFP-positive cells in the intestinal epithelium were isolated using a BD FACSAria flow cytometer (BD Biosciences). Sorted cells were collected into vials containing media.
Total RNA extraction from sorted K cells and visceral adipose tissue was performed with a PicoPure RNA isolation kit (Applied Biosystems, Alameda, CA) and RNeasy Lipid Tissue Mini Kit (Qiagen, Hilden, Germany), respectively. Complementary DNA was prepared by reverse transcriptase (Invitrogen) with an oligo(dT) primer (Invitrogen). Messenger RNA (mRNA) levels were measured by quantitative (q)RT-PCR using the ABI PRISM 7000 Sequence Detection System (AB StepOne Plus Real Time PCR; Applied Biosystems, Alameda, CA). PCR analyses were carried out using oligonucleotide primers. SYBR Green PCR master mix (Applied Biosystems) was prepared for PCR runs. Thermal cycling conditions were denaturated at 95°C for 10 min, followed by 50 cycles at 95°C for 15 s and 60°C for 1 min. COOH- and NH2-terminal primers of target molecules were designed as shown in Table 1. To quantify the results obtained by qPCR, we used the 2−ΔΔCT method for K cells and standard curve method for adipose tissue. In the standard curve method, ribosomal protein S18 was used as an internal control.
Ten-week-old mouse upper small intestine samples were fixed in Bouin's solution and transferred into 70% ethanol before being processed through paraffin. To examine the localization of FABP5 in the upper small intestine, rehydrated paraffin sections were incubated overnight at 4°C with primary mouse anti-GFP antibody (sc-9996, 1:100; Santa Cruz Biotechnology) and rabbit anti-FABP5 antibody (kindly provided by Dr. Y. Owada). The sections were incubated for 1 h at room temperature with secondary antibody. Details of secondary antibodies are as follows: Alexa Fluor 546 goat anti-rabbit IgG (H + L; Life Technologies Japan, Tokyo, Japan) for anti-FABP5 antibody and Alexa Fluor 488 goat anti-mouse (Life Technologies Japan) for anti-GFP antibody. Images were taken using fluorescent microscopy with a BZ-8100 system (Keyence, Osaka, Japan).
Measurement of GIP content in isolated K cells and number of K cells.
For measurement of GIP content in K cells, we isolated K cells from GIPgfp/+-FABP5+/+ and GIPgfp/+-FABP5−/− mice fed CFD, as described above. Five thousand cells per mouse were collected, and samples were extracted with 0.1 M HCl. GIP content was measured using an ELISA kit (Millipore, Billerica, MA). The number of GFP-positive cells was counted through the procedure of sorting and adjusted by the number of total isolated enterocytes (n = 3).
OGTT and oral lard tolerance test.
Following overnight fasting (16 h, standardized between mice), we gave 2 g/kg glucose and 10 ml/kg of lard (Megmilk Snow Brand, Sapporo, Japan) orally to FABP5+/+ (n = 4) and FABP5−/− (n = 4) mice (8–10 wk old) using a gavage tube. Macronutrient of lard (9.41 kcal/g) was as follows: 41.2% oleic acid, 23.8% palmitic acid, and 13.5% stearic acid. To collect blood samples for the oral glucose tolerance and oral lard tolerance tests (OLTT), we extracted ∼45 μl/bleed and a total volume of ∼180 μl/mouse. Body weights of mice were ∼25 g. Repeated collection was performed on alternating eyes each time. Blood was extracted by performing two collections, separated by a 2-wk interval, to obtain a sufficient volume for the measurement of glucose and glucagon-like peptide-1 (GLP-1; 1st collection) and insulin and GIP (2nd collection) by ELISA. Total sera needed to perform all ELISAs were 45 μl for each collecting point. An anesthetic agent was not used in this test because of its possible effects on measured values, especially on incretins via diminished bowel movement. During the test, blood samples were obtained into heparinized microcapillary tubes from the orbital sinus of the mice at 0, 30, 60, and 120 min after administration. Blood glucose levels were measured by the glucose oxidase method (Sanwa Kagaku Kenkyusho, Nagoya, Japan). After collection, blood samples were kept on ice and then centrifuged (3,000 rotations/min for 10 min at 4°C), and serum was separated. The serum samples were used fresh or kept at −80°C until further processing. Insulin, total GLP-1, and total GIP levels were measured by the ELISA kit as follows: insulin kit (Shibayagi, Shibukawa, Gumma, Japan), total GLP-1 kit (Meso Scale Discovery, Rockville, MD), and total GIP kit (Millipore, Billerica, MA), respectively. The area under the curve of total GIP (AUC-GIP) was calculated by the trapezoidal rule. All measurements were performed by single assay.
Common bile duct ligation.
All surgical procedures were performed utilizing clean techniques. Mice 10–15 wk of age (n = 4 for each group) were subjected to bile duct ligation (BDL). BDL was performed as described previously (36). After mice were anesthetized with intraperitoneal injection of pentobarbital sodium, a median abdominal incision was made and the common bile duct identified. The duct was dissected carefully under the microscope and doubly ligated with a sterilized 4-0 silk blade (Akiyama Medical, Tokyo, Japan) and cut across close to the duodenum upstream of the pancreaticobiliary ducts. In the sham operation group, the operative method was the same except for the ligation and cutting steps. The abdominal incision was closed in two layers. After the operation, the mice were kept warm until recovery was confirmed and fasted overnight with free access to water. On the 2nd day, OGTT and OLTT were performed, followed by the collection of blood samples. Intestinal transit rate after BDL was measured by the following steps: after overnight fasting, the mice (n = 3 for each group) were given lard containing 5 vol/vol% of Oil Red O (Wako Pure Chemical Industries, Tokyo, Japan) orally. After 20 min the mice were euthanized by cervical dislocation, and the entire gastrointestinal tract was excised and dissected medially, and the distance from the pylorus to the front of the stained lard and to the ileocecal junction was measured. The rate of transit was determined as [(distance to Oil Red front)/(length of small intestine)] × 100 (%).
Ex vivo experiments using isolated upper small intestine.
The upper half portion of the small intestine was harvested from FABP5+/+ and FABP5−/− mice (n = 4 for each group) after 16 h of food withdrawal. Samples were rinsed in PBS, shredded into small pieces with scissors, and then incubated in conditioned media as follows: 5.5 mM glucose Dulbecco's modified Eagle's medium (DMEM) as a control, 100 μM oleic acid in 5.5 mM DMEM, 4 vol/vol% of bile in 5.5 mM DMEM, and 100 μM oleic acid plus 4 vol/vol% of bile in 5.5 mM DMEM. After 15 min of incubation, all samples were collected and centrifuged (10,000 g) for 5 min. The supernatant was collected in Amicon Ultra 0.5-ml centrifugal filters (Millipore) and concentrated to 10 times, following the manufacturer's instructions. GIP levels were measured by ELISA kit, as described above.
Insulin tolerance test.
Following 8 wk of HFD feeding, the mice (n = 4 for each group) were fasted 4–6 h before the start of the experiment; 0.5 U/kg of regular insulin (Novolin R; Novo Nordisk, Bagsvaerd, Denmark) was administered intraperitoneally. Blood samples were drawn from the tail using heparinized microcapillary tubes at 0, 15, 30, and 60 min after insulin administration. Blood glucose levels were measured by the glucose oxidase method (Sanwa Kagaku Kenkyusho).
All results are expressed as means ± SE. Statistical analyses were performed using ANOVA with Tukey test, and P values <0.05 were considered statistically significant.
Expression of FABP5 in enteroendocrine K cells.
FABP5 mRNA expression levels in GFP-positive cells isolated from GIPgfp/+mice (K cells) showed a ninefold increase compared with GFP-negative cells (Fig. 1A). Immunohistochemical staining of the upper small intestines of GIPgfp/+ mice showed that FABP5 and EGFP proteins are expressed in the same cell in GFP-positive cells (Fig. 1B), and >90% of GFP-positive cells are FABP5 positive (data not shown). In Gcggfp/+ mice, FABP5 and EGFP proteins, which are expected to be expressed in GLP-1-producing L cells, are not expressed in the same cell (Fig. 1B). qRT-PCR revealed that FABP1, -2, and -5 are expressed in K cells and that there is no significant change in expression levels of FABP1 and -2 between FABP5+/+ mice and FABP5−/− mice (Fig. 1C). Expression of FABP3, -4, -6, -7, and -9 was not detected in K cells (data not shown). There was no significant difference between FABP5+/+ mice and FABP5−/− mice in GIP content in K cells (Fig. 1D) or K cell count (Fig. 1E).
Single oral administration of glucose and lard.
In OGTT, there was no significant change in blood glucose levels (Fig. 2A), plasma levels of insulin (Fig. 2B), GLP-1 (Fig. 2C), or GIP (Fig. 2D) between FABP5+/+ mice and FABP5−/− mice. Regarding OLTT, FABP5−/− mice exhibit a significant decrease in GIP secretion compared with FABP5+/+ mice (Fig. 2H). AUC-GIP in FABP5−/− mice is decreased to ∼60% of that in FABP5+/+ mice (data not shown). There is no significant difference in blood glucose levels (Fig. 2E), plasma levels of insulin (Fig. 2F), or GLP-1 (Fig. 2G) between FABP5+/+ mice and FABP5−/− mice.
Involvement of bile in acute fat-induced GIP secretion.
qRT-PCR showed very low expression levels of fatty acid transport protein (FATP) isoforms in K cells, and there was no significant difference in the expression levels of FATPs and CD36 (data not shown), an integral membrane protein expressed in intestinal epithelial cells essential for LCFA uptake in the upper small intestine, between FABP5+/+ mice and FABP5−/− mice (Fig. 3A). We found no significant difference in blood glucose levels or GIP levels between mice subjected to common BDL (BDL mice) and mice subjected to sham operation (sham mice) during OGTT (Fig. 3B). In OLTT, BDL mice exhibited a significant decrease in GIP secretion compared with sham mice (Fig. 3C), although there was no significant difference in the levels of blood glucose between BDL and sham mice (Fig. 3, B and C). To exclude the possibility that postoperative impairment of GIP secretion was due to diminished peristalsis, we measured the transition length of administered lard. In both groups, lard reached more than one-half the length of the duodenum 20 min after administration, and there was no difference in intestinal transit rate between BDL and sham mice (Fig. 3D), suggesting that impaired GIP secretion is due to the lack of bile. In an ex vivo experiment using isolated upper small intestine, we found a very small increase in GIP release after 15 min of incubation with oleic acid; we were unable to determine the statistical significance because three of the four samples with FABP5 deficiency were not measurable (Fig. 3E). By contrast, there were marked increases in GIP secretion from the samples incubated with oleic acid and bile, and the FABP5−/− samples showed a 3.4-fold decrease compared with that in FABP5+/+ samples (Fig. 3E).
FABP5−/− mice are resistant to HFD-induced obesity.
Whereas no difference in body weight was observed between FABP5+/+ mice fed CFD (FABP5+/+ CFD) and FABP5−/− mice fed CFD (FABP5−/− CFD), FABP5−/− mice fed HFD (FABP5−/− HFD) gained less body weight compared with FABP5+/+ mice fed HFD (FABP5+/+ HFD) during 8 wk of feeding (Fig. 4A). These data are consistent with an earlier study (24). Whole body CT scan revealed that visceral and subcutaneous fat mass in FABP5−/− HFD were significantly reduced compared with FABP5+/+ HFD mice, whereas those in FABP5+/+ CFD and FABP5−/− CFD were comparable (Fig. 4, B and C).
FABP5 contributes to the development of DIO in a GIP-dependent manner.
The differences observed in body weight between FABP5+/+ and FABP5−/− mice under HFD feeding were reproducible (Fig. 5A). By contrast, there was no significant difference in body weight on HFD between GIPgfp/gfp-FABP5+/+ and GIPgfp/gfp-FABP5−/− mice, in which GIP is genetically deleted (Fig. 5A). Measurement of body fat mass using whole body CT scan revealed that visceral and subcutaneous fat mass in FABP5−/− mice were significantly reduced compared with FABP5+/+ mice, whereas those in GIPgfp/gfp-FABP5+/+ and GIPgfp/gfp-FABP5−/− mice were comparable (Fig. 5, B and C). Food intake was similar in the four groups on HFD (Fig. 5D). Assessment of insulin resistance by insulin tolerance test showed a decrease in glucose levels in FABP5−/− mice compared with FABP5+/+ mice, whereas no difference in glucose level was observed between GIPgfp/gfp-FABP5+/+ and GIPgfp/gfp-FABP5−/− mice (Fig. 5E). There was no significant difference in mRNA expression levels of FABP4 and GIPR in white (visceral) adipose tissue after 8 wk of HFD feeding (22 wk old) between FABP5+/+ and FABP5−/− mice or between GIPgfp/gfp-FABP5+/+ and GIPgfp/gfp-FABP5−/− mice, respectively (Fig. 5F).
Ever since FABP5 was first identified as a novel keratinocyte protein highly upregulated in psoriatic skin (22), broad and diverse roles of FABP5 such as neurogenesis (25), protective effect from lipotoxic injury (20), carcinoma cell growth (12), mammary tumorigenesis (19), and keratinocyte differentiation (4) have been elucidated. Maeda et al. (24) found that deletion of FABP5 in mice resulted in reduction in body weight gain and improvement of insulin sensitivity under HFD. They emphasized the function of FABP5 in adipocytes on the grounds that adipocytes isolated from FABP5−/− mice exhibited enhanced insulin sensitivity and that mice with adipocyte-specific overexpression of FABP5 displayed increased systemic insulin resistance. However, the possibility that FABP5 expressed in other cell types is crucial for the phenotype of FABP5−/− mice cannot be excluded given the wide range of FABP5 expression in other tissues, including skin (31), lung (30.7), and brain (21).
Very recently, Sommer and Mostoslavsky (38) reported that FABP5 is expressed in GIP-producing K cells and is involved in fat-induced GIP secretion. From microarray analysis of K cells, we obtained data similar to their report using RNA-seq, and we confirmed the expression of FABP5 in K cells using our own mouse model (Fig. 1, A and B). Furthermore, immunohistochemistry of the upper small intestine from Gcggfp/+ mice reveals that FABP5 is not expressed in L cells (Fig. 1B), a type of enteroendocrine cell that secretes another incretin, GLP-1. In addition, we evaluated the expression levels of FABP isoforms in K cells. There was no compensatory increase in FABP isoforms expressed in K cells of FABP5−/− mice compared with those of FABP5+/+ mice, suggesting that the proposed function of FABP5 in the K cell is independent of other FABPs (Fig. 1C). We also found no significant change in GIP content in isolated K cells (Fig. 1D) or in the number of K cells sorted from the upper small intestine (Fig. 1E) between FABP5+/+ mice and FABP5−/− mice, indicating that FABP5 is not engaged in either GIP biosynthesis or K cell differentiation, as has been shown in keratocytes (29). Together, these data support the notion that FABP5 might produce its effect on the secretory pathway of GIP in matured K cells.
In OLTT, but not in OGTT, FABP5−/− mice showed a significant decrease in GIP response compared with that of FABP5+/+ mice (Fig. 2H). Blood glucose levels (Fig. 2, A and E), plasma levels of insulin (Fig. 2, B and F), and GLP-1 (Fig. 2, C and G) after oral glucose and lard administration were not significantly different between FABP5+/+ and FABP5−/− mice. In addition, an ex vivo experiment using isolated upper small intestine revealed that GIP release after 15 min of incubation with oleic acid plus bile was significantly lower in the FABP5−/− sample compared with that in the FABP5+/+ sample (Fig. 3E), as will be discussed later in detail. These data clearly show that FABP5 expressed in K cells is associated with GIP secretion in response to fat ingestion, but not to glucose, without affecting glucose tolerance. Because the insulinotropic effect of GIP is dependent on the glucose level, it is supposed that GIP did not exert its insulinotropic effect due to small elevation of glucose from baseline in oral lard administration. It was demonstrated recently that circulating levels of GIP in FABP5−/− mice are significantly decreased compared with that in FABP5+/+ mice not only after oil administration but also in the fasting state (38). The reason for this discrepancy is unknown. It may be due to the difference in experimental conditions.
Given that FABPs are known to function as an intracellular lipid chaperon, we speculated that there should be some mechanism by which fatty acid permeates the cell membrane of K cells for FABP5-associated GIP secretion in response to fat ingestion. FABP5 is reported to have high binding capacity for long-chain and saturated fatty acids (37). With regard to the mechanism of LCFA uptake into K cells, a proposed pathway is protein-mediated transport (14, 41). Nevertheless, qRT-PCR revealed relatively low expression of FATP isoforms and CD36 compared with FABP isoforms (Fig. 4A). Furthermore, no significant change was observed in any FATP isoform or CD36 associated with genetic deletion of FABP5. We also examined the significance of bile-aided incorporation of LCFA into K cells, as is the case for absorptive epithelial cells in the small intestine. We evaluated GIP secretion in OGTT and OLTT in BDL mice and sham mice. Whereas GIP secretion during OGTT remained unaffected by BDL (Fig. 3B), the secretion in OLTT was seriously diminished in BDL-mice (Fig. 3C). Dissociation of the effect of BDL on GIP secretion in OGTT and OLTT suggests that bile is the crucial factor in GIP secretion in response to fat but not glucose. In ex vivo experiments using the isolated upper small intestine, culture in control medium + oleic acid (100 μM) (OLA) induced only a slight increase in GIP release, which was markedly enhanced by coadministration of bile and OLA (Fig. 3E). On the other hand, bile per se did not induce GIP release. This finding was reinforced by performing oral administration of bile extracted from other mice to fasted BDL mice, in which we found no elevation of serum GIP level (data not shown). These observations are consistent with previous reports mentioning the importance of bile in fatty acid-induced GIP secretion in human patients with disturbed bile secretion into the upper small intestine (10, 46). The amplifying effect of bile on fat-induced GIP secretion enabled us to determine the significant reduction in fatty acid-induced GIP release from FABP5-deficient intestinal samples in an ex vivo experiment using isolated upper small intestine.
Bile acids are classically known for their physiological function as emulsifiers, facilitating absorption of fat in gut (18). However, growing evidence has indicated that bile acids have diverse physiological functions as metabolic regulators (40). Recently, it has been shown that bile acids exert their effects through nuclear receptor farnesoid X receptor and transmembrane G protein-coupled receptor 5 (TGR5) (33). Furthermore, TGR5 is reported to be expressed in intestinal L cells, and TGR5 agonist promotes GLP-1 secretion (13). We confirmed that TGR5 mRNA expression levels in K cells are extremely low (data not known), in accordance with our data showing that bile per se did not stimulate GIP secretion either in BDL mice or in an ex vivo experiment using isolated upper small intestine. However, based on our data, it is clear only that bile is an essential amplifying factor of fat-induced GIP secretion, and the contribution of G protein-coupled receptors or FATPs expressed in K cell membrane to FABP5-associated GIP secretion cannot be excluded simply because expression levels are low; further examination is required to clarify the detailed mechanism of GIP secretion in response to fat ingestion.
Finally, we have evaluated the chronic effect of FABP5 deletion on body weight regulation in the presence and absence of GIP secretion. We confirm that FABP5−/− mice fed HFD exhibit a significant decrease in body weight and fat accumulation compared with FABP5+/+ mice, similarly to the study by Maeda et al. (24). The fact that body weight progression was comparable between GIPgfp/gfp-FABP5+/+ and GIPgfp/gfp-FABP5−/− illustrates that FABP5 regulation of body weight and composition occurs in a GIP-dependent manner. There have been several reports noting a compensatory increase in FABP5 in adipose tissue in the absence of FABP4 (3, 44). However, in general, when several FABP isoforms are expressed in the same organ, deletion of one isoform does not necessarily lead to change in the expression of the others, as with FABP1 and -2 in small intestinal mucosa (17), for example. Our results did not show any significant difference in FABP4 mRNA expression in adipocytes under HFD feeding between FABP5+/+ and FABP5−/− mice or between GIPgfp/gfp-FABP5+/+ and GIPgfp/gfp-FABP5−/− mice (Fig. 5E), indicating that the function of FABP5 in adipose tissue is independent of FABP4 with regard to the status of obesity. We also found no change in GIPR expression in association with FABP5 deletion, suggesting that GIP sensitivity of adipocytes is maintained in FABP5−/− mice. The reason why GIPR and FABP4 expression levels were significantly higher in GIPgfp/gfp mice compared with GIP+/+ mice requires further investigation. However, the possibility remains that some other metabolic-related genes located downstream of GIP signaling in adipocytes are altered by FABP5 deletion and play some role in fat deposition and/or improvement in insulin resistance in FABP5−/− mice.
In conclusion, our data in this study contain several novel findings not shown in previous reports. First, bile efficiently amplifies GIP secretion mediated by FABP5 in response to single administration of fatty acid. Second, FABP5 is associated with GIP secretion induced by fatty acid but is not involved in glucose-induced GIP secretion. And especially important is that FABP5 contributes to the development of diet-induced obesity in a GIP-dependent manner. Because partial reduction of GIP has been reported to alleviate obesity and lessen the degree of insulin resistance under the HFD condition without exacerbating glucose tolerance, these findings suggest a potentially promising therapeutic approach to prevent obesity and subsequent insulin resistance.
This study was supported by scientific research grants from the Ministry of Education, Culture, Sports, Science, and Technology, Japan, the Ministry of Health, Labor, and Welfare, Japan, and the Integration Research for Agriculture and Interdisciplinary Fields, Japan. N. Inagaki also received a clinical commission/joint research grant from Merck Sharp & Dohme (MSD) and Eli Lilly Japan.
This study was also supported by the Japan Diabetes Foundation, the Japan Association for Diabetes Education and Care, the Suzuken Memorial Foundation, Novo Nordisk Pharma, Shiratori Pharmaceutical, and Roche Diagnostics as well as a scholarship grant from MSD, Japan Tobacco, Nippon Boehringer Ingelheim, Takeda, Dainippon Sumitomo Pharma, Astellas Pharma, Daiichi-Sankyo, and Mitsubishi Tanabe Pharma.
N. Inagaki served as a medical advisor for Takeda, Taisho Pharmaceutical, GlaxoSmithKline, and Mitsubishi Tanabe Pharma, and lectured for MSD, Sanofi, Novartis Pharma, Dainippon Sumitomo Pharma, Kyowa Kirin, and Mitsubishi Tanabe Pharma and received payment for services. No other potential conflicts of interest relevant to this article are reported.
K. Shibue, S.Y., N.H., and N.I. conception and design of research; K. Shibue and K. Suzuki performed experiments; K. Shibue analyzed data; K. Shibue, S.Y., N.H., A.H., K. Suzuki, E.J., K.I., D.N., T.H., Y.H., Y.A., Y.O., R.T., and N.I. interpreted results of experiments; K. Shibue prepared figures; K. Shibue and S.Y. drafted manuscript; K. Shibue, S.Y., N.H., and N.I. approved final version of manuscript; K. Shibue, S.Y., and N.I. edited and revised manuscript.
We thank Shoichi Asano from the Department of Diabetes, Endocrinology, and Nutrition, Graduate School of Medicine, Kyoto University, for technical support regarding the study.
- Copyright © 2015 the American Physiological Society