G-protein coupled receptor 120 (GPR120) has been shown to act as an omega-3 unsaturated fatty acid sensor and is involved in insulin secretion. However, the underlying mechanism in pancreatic β cells remains unclear. To explore the potential link between GPR120 and β-cell function, its agonists docosahexaenoic acid (DHA) and GSK137647A were used in palmitic acid (PA)-induced pancreatic β-cell dysfunction, coupled with GPR120 knockdown (KD) in MIN6 cells and GPR120 knockout (KO) mice to identify the underlying signaling pathways. In vitro and ex vivo treatments of MIN6 cells and islets isolated from wild-type (WT) mice with DHA and GSK137647A restored pancreatic duodenal homeobox-1 (PDX1) expression levels and β-cell function via inhibiting PA-induced elevation of proinflammatory chemokines and activation of nuclear factor κB, c-Jun amino (N)-terminal kinases1/2 and p38MAPK signaling pathways. On the contrary, these GPR120 agonism-mediated protective effects were abolished in GPR120 KD cells and islets isolated from GPR120 KO mice. Furthermore, GPR120 KO mice displayed glucose intolerance and insulin resistance relative to WT littermates, and β-cell functional related genes were decreased while inflammation was exacerbated in islets with increased macrophages in pancreas from GPR120 KO mice. DHA and GSK137647A supplementation ameliorated glucose tolerance and insulin sensitivity, as well as improved Pdx1 expression and islet inflammation in diet-induced obese WT mice, but not in GPR120 KO mice. These findings indicate that GPR120 activation is protective against lipotoxicity-induced pancreatic β-cell dysfunction, via the mediation of PDX1 expression and inhibition of islet inflammation, and that GPR120 activation may serve as a preventative and therapeutic target for obesity and diabetes.
Pancreatic β cells are key players in the regulation of glucose homeostasis, and β-cell dysfunction is closely associated with the pathogenesis and progression of diabetes. Pancreatic duodenal homeobox-1 (PDX1) has been demonstrated to have a central role in the maintenance of islet β-cell function and survival [1,2]. Mimicking observations seen in human, Pdx1 deficiency in rodent models has been shown to cause β-cell dysfunction, thus contributing to diabetes [3,4]. Non-selective Pdx1 deletion resulted in pancreatic agenesis [3,4], whereas selective Pdx1 gene knockout in islets led to a dramatic reduction in the number of β cells . Mice with β-cell specific PDX1 inactivation or PDX1-haploinsufficiency exhibited severe hyperglycemia and developed diabetes due to β-cell death with loss of β-cell mass and decreased insulin secretion [6–8]. PDX1 has also been known as insulin promoter factor 1, which plays an important role in the regulation of insulin biosynthesis and secretion , and thus inactivation of PDX1 reduces insulin expression . Under diabetic conditions, the gene expression levels of Insulin, Pdx1 and Mafa were decreased, while overexpressing PDX1 in β cells rescued Insulin and Mafa expression, thus improving β-cell function . In addition, chronic hyperglycemia and hyperlipidemia were also attributable to β-cell dysfunction via reducing PDX1 expression and insulin levels [3,11,12].
In the last decade, it has become clear that islet inflammation is a critical contributor to obesity and diabetes . In type 2 diabetic patients, islet-associated immune cells were increased versus nondiabetic controls . Moreover, increased numbers of monocytes/macrophages were also observed within islets from db/db mice, streptozotocin-injected mice, high-fat-fed mice and GK rats [14,15]. IL-6 family cytokines were elevated in pancreatic islets isolated from those rodent models of obesity and diabetes, whereas circulating chemokine KC/CXCL1 was increased in high-fat-fed animals. Moreover, ex vivo treatment of type 2 diabetic milieu (glucose plus palmitate) in both human and mouse islets showed that increased IL-6, chemokine CXCL1 and granulocyte colony-stimulating factor were released, whereas palmitate alone could also induce the production of CXCL1 . Islet inflammation is usually triggered by high plasma levels of glucose, saturated free fatty acids and proinflammatory chemokines/cytokines due to over-nutrition and obesity [16,17]. As such, intensification of inflammation conduces to the development of hyperglycemia, insulin resistance and, ultimately, pancreatic islet dysfunction.
In 2010, G-protein coupled receptor 120 (GPR120, a.k.a. free fatty acid 4 receptor, FFAR4) was found to act as an omega-3 free fatty acid sensor, contributing to anti-inflammatory responses in RAW264.7 cells and primary mouse macrophages . Subsequently, it was reported that dysfunction of GPR120 in adipose tissues led to obesity in both human and mouse . GPR120 had been demonstrated to be expressed in human islets where its expression correlates with insulin secretion and its activation prevents lipid-induced islet apoptosis . In addition, GPR120 was found to inhibit somatostatin secretion from mouse islet delta cells , promote glucagon secretion from islet α cells and increase insulin secretion from rodent islets, INS-1E cells and MIN6 cells [22,23]. Our recent studies have demonstrated that activation of GPR120 with the polyunsaturated fatty acid docosahexaenoic acid (DHA) or selective GPR120 agonist GSK137647A produced a surge of insulin release . Given that the natural ligand DHA stimulates other G-protein coupled receptors, such as GPR40 , the GPR120-selective agonist GSK137647A was used to confirm the specificity of GPR120 activation-mediated effects in the present study.
The expression of GPR120 has been identified in many tissues, but its expression and function in pancreatic islets still remain ambiguous and, more importantly, the underlying mechanism-of-action whereby GPR120 regulates β-cell function has yet to be determined. In this regard, a recent systems genetics approach has reported that diabetic human islets were observed to have reduced GPR120 and PDX1 expression in association with increased HbA1c levels, comparing with normal controls; in addition, the expression of PDX1 in islets showed a direct correlation with insulin secretion . In light of these prior findings, we aimed to explore the potential link between GPR120 and PDX1 in islet functions during inflammatory responses.
Materials and methods
Mice were obtained from the Laboratory Animal Services Centre of The Chinese University of Hong Kong and our experimental animal protocols were approved by the Animal Experimental Ethics Committee of The Chinese University of Hong Kong (Ref. #16-220-MIS). In the present study, we generated a GPR120-knock-out (KO)/LacZ knock-in (KI) mouse model. Briefly, the sperms of GPR120 KO/KI C57 mice were purchased from KOMP Repository, in which exon 2 of GPR120 had been replaced by a β-galactosidase (β-Gal) expression cassette (LacZ). Heterozygous mice with C57BL/6J background were intercrossed to generate GPR120 KO mice and wild-type (WT) littermates. Male mice used for subsequent study were fed a normal chow diet (ND) (Select Rodent Diet 50 IFI6F Auto Diet, Totoro) or high-fat diet (HFD) (60% of Kcal from fat; Harlan Laboratories, U.S.A.), given water ad libitum, and maintained on a 12-h light–dark cycle. Mice were genotyped by polymerase chain reaction (PCR). The primer sequences for genotyping are: Gpr120, forward: 5′-AGACTACCGACTCTTCCGCA-3′; reverse: 5′-GAGTTGGCAAACGTGAAGGC-3′; LacZ, forward: 5′-TTCGGCTATGACTGGGCACAACAG-3′; reverse: 5′-TACTTTCTCGGCAGGAGCAAGGTG-3′; Gadph, forward: 5′-GCACAGTCAAGGCCGAGAAT-3′; reverse: 5′-GCCTTCTCCATGGTGGTGAA-3′. The PCR products were analyzed by electrophoresis in 1.5% agarose gels.
GPR120 agonists and supplementation
The natural GPR120 agonist DHA (Cayman, U.S.A.) and selective GPR120 agonist GSK137647A (Tocris, U.S.A.) were dissolved in 100% ethanol and dimethyl sulfoxide, respectively, to 100 mM. Palmitic acid (PA) was dissolved in 50% ethanol to 100 mM before and combined at a 1:9 volume ratio with 10% free fatty acid-free bovine serum albumin (BSA; Sigma, U.S.A.), as described previously [23,25]. Nine-week GPR120 KO and WT mice were fed with HFD for 10 weeks, and then given DHA (40 mg/kg) in phosphate-buffered saline (PBS), GSK137647A (20 mg/kg) in PBS, or plain PBS intraperitoneally every 2 days for 4 weeks.
MIN6 cell culture and transfection
MIN6 cells were purchased from AddexBio and cultured in Dulbecco’s modified eagle medium, containing 25 mM glucose with 15% fetal bovine serum (Thermo Fisher, U.S.A.) in a warm (37°C) incubator with 5% CO2. MIN6 cells (passage 16–26) were treated with 30 µM DHA or 10 µM GSK137647A, with/without 500 µM PA as indicated in the results. Small interfering RNAs (siRNAs) for mouse GPR120 and scrambled siRNA were chemically synthesized (Invitrogen) and transfected into MIN6 cells with lipofectamine RNAi Max (Invitrogen) for 48 h according to the manufacturer’s instructions. Then, cells were treated with PA, DHA or GSK137647A as indicated. The sequences of the GPR120 and scrambled control (Scr) siRNAs used are listed in Supplementary Table S1. The efficiency of siRNA-GPR120 was assessed with real-time PCR and immunoblotting.
Pancreatic islet isolation
Pancreatic islets were isolated from mice by injection of collagenase  and processed using previously described methods [27–29]. Isolated islets were cultured overnight in RPMI-1640 medium (GIBCO) containing 5.6 mM glucose and 10% FBS under 5% CO2 at 37°C, and then exposed to experimental stimuli as indicated in the results.
Glucose-stimulated insulin secretion assay and insulin content measurement
The Glucose-stimulated insulin secretion (GSIS) assay of β-cell function was performed in MIN6 cells or isolated islets (15 islets per group). Cells or islets were pre-incubated with Kreb’s Ringer bicarbonate buffer (KRBB) containing 2.8 mM glucose for 1 h, and then incubated with fresh KRBB containing 2.8 mM glucose for 1 h, followed by 22.0 mM glucose for an additional 1 h with or without the supplementation of DHA or GSK137647A. After incubation, buffers were collected and released insulin was measured with a high-sensitive mouse insulin immunoassay kit (The University of Hong Kong), as we described previously . Intracellular insulin content was measured and normalized to total protein using a bicinchoninic acid assay (Bio-Rad), as we described previously .
In vivo glucose homeostasis
Fasting blood glucose and insulin levels were measured as described in our previous reports [28,29]. After mice were fasted for 5 h with a free access to water, the blood glucose levels were measured with a glucometer (Bayer Corporation, Tarrytown, U.S.A.) and 20 μl of blood samples were collected from the tail vein for insulin measurement with an ultrasensitive mouse insulin immunoassay kit (The University of Hong Kong). Insulin resistance was assessed with the homeostatic model assessment of insulin resistance (HOMA-IR). To assess the glucose homeostasis in mice, the intraperitoneal glucose tolerance tests (GTTs) and insulin tolerance tests (ITTs) were performed as described previously [28,29]. Briefly, blood glucose levels of mice with fasting of 5 h were measured at 0, 15, 30, 60, 90 and 120 min after intraperitoneal glucose administration (1.5 g/kg body weight). For ITT, the mice were fasted for 5 h and then challenged with an intraperitoneal injection of insulin (0.75 unit/kg body weight). Area under the curve values for blood glucose were calculated.
Given that GPR120 is highly expressed in murine colon [21,31], we chose the colon as a positive control for GPR120 expression. Freshly dissected mouse colons and pancreas were embedded in O.C.T compound (Tissue Tek, Sakura, Japan). Five-micron-thick sections were prepared with a cryomicrotome (Leica, Microsystmes GhbH) at −20°C, collected immediately onto slides, and fixed in 4% paraformaldehyde for 30 min. After three rinses in PBS, the slides were incubated in 3% BSA/PBS for 1 h, and then incubated with appropriate primary antibodies, such as anti-beta Galactosidase/LacZ (ab9361, Abcam), anti-Insulin (ab7842, Abcam), anti-Glucagon (#14-9743-82, Invitrogen), anti-Somatostatin (sc-74556, Santa Cruz), Alexa Fluor® 647 anti-mouse F4/80 (Biolegend) at 4°C overnight. After washing with PBS, the slides were incubated with fluorescent-conjugated secondary antibodies (Invitrogen-Alexa) for 1 h. Nuclei were then stained with 4′,6-diamidino-2-phenylindole (DAPI; Life technologies) for 5 min. After further washing with PBS, slides were cover-slipped with VectaShield mounting medium (Vector Laboratories). Fluorescent signals were examined with a confocal inverted microscope (Olympus FV1200).
Western blotting analysis and CXCL1 measurement
Total proteins were extracted from cells or islets with CytoBuster extraction reagent (Novagen, Billerica, MA). Protein samples were subjected to electrophoresis on 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis for 2.5 h at 80 V. The separated proteins were transferred to 2μm PVDF membranes (Bio-Rad). Blotted membranes were blocked with 5% BSA (w/v) for 1 h, probed with appropriate primary antibodies overnight at 4°C, and then incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 h. Protein signaling bands were revealed by chemiluminescense (Amersham Biosciences) on autoradiography film (RX; Fuji Film). The intensity of the protein bands was quantified in ImageJ. The protein bands were normalized to β-actin or corresponding total protein as indicated. Antibodies are listed as below: Goat PDX1, Rabbit MCP1/CCL2, Mouse β-Actin, Rabbit JNK (Santa Cruz, U.S.A.), Rabbit GPR120 (GeneTex, U.S.A.), Rabbit p-NF-κB p65, p-JNK1/2, p-p38 MAPK, NF-κB p65, JNK1/2, p38 MAPK (Cell Signaling Technology, U.S.A.). CXCL1 levels in isolated islets and plasma were detected using mouse CXCL1 (KC) ELISA assay kit (Invitrogen) according to the protocol.
RT-PCR and real-time PCR analyses
MIN6 cells or isolated islets were lysed with TRIzol reagent (Takara Bio, Japan) and total mRNA was extracted according to the manufacturer’s protocol. cDNA synthesis was performed with a PrimeScript RT Master Mix kit (Takara Bio). mRNA levels were quantified by quantitative Real-time PCR performed in a ViiA 7 Real-Time PCR System (Applied Biosystems Life Technologies, U.S.A.) with SYBR Green mix according to the system protocol. The target gene primers employed are listed in Supplementary Table S1. Relative gene expression was analyzed by the comparative threshold cycle method (2−ΔΔCT) and normalized to Gapdh.
All data are presented as means ± SEM. Group data were compared between two groups with unpaired t-test. Multiple-group comparisons were performed using one-way analysis of variance (ANOVA) followed by Dunnett’s test or Tukey’s, or two-way ANOVA followed by Bonferroni post hoc tests. For all comparisons, P<0.05 was considered as statistically significant. Statistical analyses were performed in GraphPad Prism 5.
GPR120 mediates insulinotropic effects through regulation of PDX1 expression via inhibiting PA-induced inflammatory responses in vitro
To explore the potential GPR120-mediated islet function and inflammation, we employed palmitic acid (PA) as a stimulus to mimic lipotoxicity-induced β-cell dysfunction and inflammation. Our results showed that the mRNA expression of Insulin and Pdx1 were significantly decreased after PA treatment for 24 h (Figure 1A,B), whereas the expression levels of the proinflammatory chemokines C-X-C motif ligand 1 (Cxcl1) and C-C motif ligand 2 (Ccl2, a.k.a. monocyte chemotactic protein 1, Mcp1) were increased in MIN6 β cells (Figure 1C–E). Similar results were observed in islets isolated from 12-week-old C57BL/6J male mice (Figure 1F–J). Meanwhile, GPR120 activation with DHA or GSK137647A restored the expression levels of Insulin and PDX1, and diminished the expression of Cxcl1 and CCL2 levels in MIN6 cells (Figure 1A–E), as well as those expression in isolated islets (Figure 1F–J) under PA-induced lipotoxicity-like conditions.
GPR120 activation mediated β-cell functional genes expression and PA-induced inflammation
To further examine whether GPR120 is indispensable for the protection of β-cell function under PA-induced inflammation, we performed an siRNA-based knockdown approach (Supplementary Figure S1A,B). When GPR120 was knocked down (KD) in MIN6 cells (Figure 2A), GSIS results showed that there was no significant difference between Scr-group and KD-group with PA, DHA or vehicle treatments under low-glucose conditions (Figure 2B). Under high-glucose conditions, treatment with DHA or GSK137647A failed to restore the PA-impaired insulin secretion in GPR120-silenced cells, relative to Scr-transfected groups (Figure 2B). In addition, insulin secretion in GPR120 KD cells was significantly lower versus those of scrambled groups in the presence of GSK137647A agonist under high-glucose conditions (Figure 2B). To further confirm the GPR120-mediated insulinotropic effects at the protein levels, Western blot results revealed that the inhibitory effect of PA on the expression of PDX1 was abolished by DHA or GSK137647A treatment in the control cells, while the stimulatory effect on the expression of CCL2 was also diminished (Figure 2C). However, the beneficial effects of DHA or GSK137647A treatment were abolished in MIN6 cells with GPR120 KD (Figure 2C). Similarly, the elevated level of CXCL1 secretion upon PA treatment was also inhibited by co-treatment with DHA or GSK137647A in the control MIN6 cells, not in the GPR120 KD cells (Figure 2D).
GPR120 protection of β-cell function through regulation of PDX1 expression and inhibition of PA-induced inflammatory activation in MIN6 cells
To explore the potential mechanism involved, our preliminary data showed that PA treatment induced the phosphorylation of nuclear factor-κB (NF-κB) p65, c-Jun amino (N)-terminal kinases1/2 (JNK) and p38 mitogen-activated protein kinase (MAPK) in a time-dependent manner (Supplementary Figure S1C). Next, we found that DHA or GSK137647A significantly inhibited PA-induced activation of NF-κB, JNK and p38 signaling pathways in Scr-transfected MIN6 cells; however, these suppressive effects were abrogated by the silencing of GPR120 (Figure 2E). We also examined the effects of DHA and GSK137647A treatments alone, and our results showed that there was no significant difference or negative effects on the expression levels of PDX1, CCL2 and CXCL1 as well as the signaling pathways involved (Figure 2C–E).
The modulatory effects of GPR120 activation on PDX1 expression and inflammation are abolished in GPR120 KO mouse islets
To further explore the regulatory role of GPR120, knockout mice in which exon 2 of GPR120 had been replaced by β-Gal expression cassette were generated. First, we examined the expression and localization of GPR120 in pancreatic islets. Gpr120 mRNA was found to be expressed in isolated islets from WT (+/+) and heterozygous (+/-) mice, as well as in MIN6 cells, but not in islets from GPR120 KO mice (Figure 3A). GPR120 KO, but not WT mouse colon sections were confirmed to be positively blue for LacZ/β-Gal stain, and β-Gal was also found to be positive in pancreatic islets (data not shown). Immunostaining examinations further revealed the localization of LacZ in the colon (Figure 3B) and different pancreatic islet cell types (Figure 3C–E) from GPR120 KO mice. Importantly, the co-localization of LacZ with insulin in GPR120 KO mouse pancreas suggested that GPR120 was expressed in islet β cells (Figure 3C). In addition, LacZ was also co-localized with somatostatin (Figure 3D), implying that GPR120 was also expressed in islet delta cells. However, LacZ was rarely co-localized with glucagon-producing α cells (Figure 3E).
Expression and localization of GPR120 in pancreatic islets
On the other hand, Western blot analysis also confirmed that GPR120 was expressed in WT mouse colon and islets, but not in GPR120 KO mouse islets (Figure 4A). In consistent with previous in vitro results, GPR120 agonists-mediated protective effects on PA-reduced PDX1 expression were abolished in islets isolated from 12-week-old GPR120 KO mice, comparing with WT littermates (Figure 4B). In addition, DHA and GSK137647A-treated islets from GPR120 KO mice did not suppress the expression of Cxcl1 and Ccl2 (Figure 4C) as well as failed to inhibit the phosphorylation of NF-κB, JNK and p38 signaling pathways (Figure 4D), which were brought by PA treatment. Moreover, we also detected the expression of PDX1, CCL2 and CXCL1 as well as the signaling activation with DHA or GSK137647A treatment alone in islets, and results showed that there was no significant difference or harmful effect relative to vehicle-treated groups (Supplementary Figure S2A,B).
GPR120 activation-mediated protective effects on PDX1 expression via inhibition of islet inflammation
In vivo functionality of Gpr120 gene in glucose homeostasis
To investigate the function of GPR120 in glucose homeostasis, related parameters were examined. The body weight of GPR120 KO mice was similar to those of WT littermates (Figure 5A), while the fasting blood glucose and insulin levels of the GPR120 KO mice were significantly higher than that of WT mice age-dependently (Figure 5B,C). HOMA-IR analysis revealed that insulin resistance was developed in GPR120 KO mice comparing with WT mice, particularly at the age of 24 weeks (Figure 5D). Moreover, 24-week-old GPR120 KO mice displayed impairment in glucose tolerance (Figure 5E), GSIS (Figure 5F) and insulin sensitivity (Figure 5G), as evidenced by the results of GTT and ITT.
Beneficial effects of Gpr120 gene on glucose homeostasis in mice on normal diet
Effects of GPR120 deletion on β-cell dysfunction and islet inflammation
Islets from the 24-week-old mice were isolated and stimulated with KRBB containing 2.8 or 22.0 mM glucose in vitro. Results showed that insulin secretion was decreased from GPR120 KO mouse islets, compared with those from WT mouse islets, notably observed in high glucose conditions (Figure 6A), while there was no significant difference in insulin content between WT and KO mouse islets (Figure 6B). We then examined the basal mRNA expression of Insulin and Pdx1 as well as PDX1 protein levels in islets. Results showed that their expression levels were significantly reduced in islets isolated from GPR120 KO mice, compared with those in islets from WT mouse (Figure 6C–E).
Deletion of GPR120-induced islet dysfunction and inflammation
On the other hand, the islets isolated from 24-week-old GPR120 KO mice showed elevation in the levels of CXCL1 and CCL2, comparing with WT mice (Figure 6F–I), which were positively associated with enhanced phosphorylation of NF-κB, JNK and p38 MAPK signals (Figure 6J). These results were consistent with our preliminary data on the expression of insulin-related genes that were decreased, and on the expression of inflammation-related genes that were increased age-dependently in isolated islets from 12- and 24-week-old GPR120 KO mice, compared with those islets from WT mice (Supplementary Figure S3A,B). Furthermore, the number of F4/80 positive cells was significantly increased in the pancreas of GPR120 KO mice (Figure 6K,L), indicating that macrophage infiltration in pancreatic islets was increased in GPR120 KO mice. Consistently, the plasma levels of CXCL1 were also detected to be elevated in GPR120 KO mice (Figure 6M). This convergence of our data supports the notion that GPR120 deficiency results in islet dysfunction, probably via the mediation of inflammation.
GPR120 mediates anti-diabetic effects in diet-induced obese mice
To further explore the translational relevance of GPR120 activation, agonism with DHA and GSK137647A was performed in vivo supplementation in HFD-fed mice. The body weight of HFD-fed WT mice with supplementation was moderately decreased without significant difference (Figure 7A). Comparing with normal diet (ND) condition, HFD induced severely impaired glucose tolerance and insulin resistance in both types of mice (Supplementary Figure S4A,B). In addition, the expression of Pdx1 was decreased in islets isolated from HFD-fed mice compared with ND-fed mice (Supplementary Figure S4C), while the expression of the proinflammatory genes, Cxcl1 and Ccl2, was increased in mice islets after HFD treatment (Supplementary Figure S4D,E). However, supplementation with DHA (blue line) and GSK137647A) (yellow line) ameliorated fat diet-induced glucose intolerance and insulin insensitivity in WT mice (black line), but not GPR120 KO mice (red line), as demonstrated by Figure 7B,C. Moreover, there was no significant difference between supplementation and no supplementation of GSK137647A in HFD-fed GPR120 KO mice (Figure 7B,C). Furthermore, the expression of Pdx1 was detected to be increased only in the islets isolated from WT mice supplemented with DHA or GSK137647A on HFD (Figure 7D). Meanwhile, the expression of islet inflammatory genes, such as Cxcl1 (Figure 7E), Ccl2 (Figure 7F), Tnfα (Figure 7G) and Il1β (Figure 7H), was suppressed in obese WT mice supplemented with DHA or GSK137647A, especially compared with GPR120 KO mouse islet groups (Figure 7E–H). These results indicated the therapeutic benefits of GPR120 activation with agonists in obesity and islet inflammation, at least partially.
In vivo studies of GPR120 activation-mediated anti-diabetic effects in diet-induced obese mice
GPR120 has been reported to enhance islet insulin secretion while disruption of GPR120 was shown to result in a 50% reduction of insulinotropic effects in human islets . Recently, we have demonstrated that GPR120 activation with short-term exposure of DHA or GSK137647A augmented insulin release in rodents via activation of Akt signaling pathway and calcium (Ca2+) channels . To further explore the potential GPR120 activation-mediated protective mechanism against lipotoxicity-induced β-cell dysfunction and inflammation, we sought to examine the potential link between GPR120 and PDX1, the latter being indispensable for insulin transcription and β-cell function in mature islets. In the present study, our study findings point toward GPR120 having an important role in protecting the biological function of β cells, via mediation of regulation of PDX1 expression and inhibition of islet inflammation.
The precise islet-cell-type distribution of GPR120 expression was still controversial. Stone et al.  described the expression of GPR120 in delta cells and in a small proportion of α cells, but not in β cells. Conversely, Moran et al.  observed the expression of GPR120 in islet β cells. Our findings of GPR120 mRNA expression in mouse islets and MIN6 cells are in agreement with previous reports [22,23,32,33]. By means of using immunolabeling and staining techniques in the present study, GPR120 was found to be mainly expressed in the pancreatic β cells and delta cells. Regardless of these observations, we were to focus on the protective mechanism by which GPR120 mediates islet function and inflammation.
PA/palmitate has been previously reported to mimic lipotoxicity-induced β-cell dysfunction in both human and rodent in in vitro or in vivo studies [17,34–36]. In this context, we found that prolonged exposure to PA resulted in decreased expression of insulin and PDX1 levels, whereas treatment with DHA and GSK137647A could restore PA-downregulated expression of PDX1 and insulin in MIN6 cells as well as in isolated islets. When GPR120 was knocked down in MIN6 cells, PDX1 production could not be restored by treatment with DHA or GSK137647A. Moreover, neither DHA nor GSK137647A could improve GSIS after GPR120 silencing (Figure 2B). Similar results were obtained in our analysis of isolated islets from GPR120 KO mice treated with PA in the presence or absence of DHA/GSK137647A, relative to those from WT littermates. These data implicate that GPR120-mediated protective effects on lipotoxicity-induced pancreatic β-cell dysfunction are likely to be involved in the regulation of PDX1 expression.
Over time, chronic hyperglycemia and hyperlipidemia induce oxidative stress and endoplasmic reticulum stress, leading to reduced insulin biosynthesis and deterioration of β-cell function. This reduction in insulin expression has been found to be closely associated with decreased expression or inactivation of PDX1 [12,37]. It was previously reported that PA might induce JNK activation, resulting in nucleocytoplasmic translocation of PDX1, and preventing its binding with the Insulin promotor, thereby leading to reduced Insulin transcription . Indeed, overexpression of JNK has been shown to suppress the transcription and secretion of insulin, whereas suppression of JNK activity was suggested to protect β cells from oxidative stress .
Our finding of enhanced phosphorylation of NF-κB and reduced expression of PDX1 in GPR120 KO mouse islets (Figure 6E,J) are related to a previous report, showing that the toll-like receptor 4 (TLR4)/NF-κB signal cascade contributes to lipopolysaccharide-induced reduction of PDX1 expression in islets . In this context, Eguchi et al.  demonstrated that PA/palmitate reduced PDX1 and insulin expression via the TLR4 pathway activation, thus inducing the binding of transforming growth factor-β activated kinase 1 (TAK1) and TAK1 binding protein 1 (TAB1), and finally activating the downstream NF-κB and JNK signaling. The resultant signaling cascade subsequently triggered the expression of chemokines and cytokines to recruit M1-type proinflammatory macrophages into mouse islets, ultimately leading to β-cell dysfunction . Conversely, GPR120 activation has been reported to recruit β-arrestin2 (β-arr), and the GPR120–β-arr complex interacts with TAB1, blocking its interaction with TAK1, thereby inhibiting subsequent NF-κB and JNK activation . In addition, anti-inflammatory effects mediated by GPR120 have been confirmed in macrophages, neurons and adipocytes [18,41,42].
In light of these findings, it is plausible to propose that GPR120 activation may also suppress PA-induced inflammation in islets. Our results demonstrate that treatment with DHA/GSK137647A suppresses PA-induced expression of the proinflammatory chemokines, such as CXCL1 and CCL2, through inhibition of the phosphorylation of NF-κB, JNK and p38MAPK signaling pathways in both MIN6 cells and isolated WT mouse islets (Figures 2 and 3). Our observations of the anti-inflammatory action under GPR120 KD and KO conditions further suggest that GPR120 activation suppresses PA-induced inflammatory responses via inhibition of NF-κB, JNK and p38 signaling cascades, and thus preserve PDX1 production as well as β-cell function. In addition, we also examined the expression of PDX1, CCL2 and CXCL1 as well as the signaling activation with DHA or GSK137647A treatment alone in MIN6 cells and islets. Results showed that there was no significant difference or harmful effect relative to vehicle-treated groups; these findings further support for the hypothesis that activation of GRP120 with agonists is protective against lipotoxicity-induced β-cell dysfunction and inflammation. Moreover, the GPR120 KO mice exhibited higher expression levels of inflammatory genes in islets comparing with WT mice, which is due, at least in part, to enhanced activation of NF-κB, JNK and p38 signals (Figure 6J). Furthermore, the deficiency of GPR120 leads to infiltration of macrophage in the pancreas (Figure 6K). Altogether, it is plausible to postulate that GPR120 protects β-cell function, probably via the mediation of anti-inflammatory effects.
On the other hand, our in vivo experiments demonstrated the importance of GPR120 in regulating glucose homeostasis. The finding of similar body weight between GPR120 KO mice and WT mice on ND or HFD is consistent with previous studies [19,43,44]; however, there was a report of GPR120 KO mice with weight gain versus WT mice on HFD . We found that the levels of fasting blood glucose and insulin of GPR120 KO mice were higher than those of WT mice on ND as mice aged (Figure 5B,C). In addition, impaired glucose tolerance and insulin sensitivity were observed in 24-week-old GPR120 KO mice (Figure 5E–G). Basal levels of insulin and PDX1 in islets isolated from 24-week GPR120 KO mice were reduced compared with islets isolated from WT littermates; meanwhile, enhanced inflammatory responses were also observed in GPR120 KO mice islets (Figure 6C–J), indicating that GPR120 may be crucial for long-term metabolic health.
Severe glucose intolerance and insulin insensitivity were found in HFD-fed mice, which were likely due to HFD-induced obesity and tissue inflammation, as reported previously [42,45–48]. DHA has been shown repeatedly to improve insulin secretion in obesity [23,32,49]. It was also reported that omega-3 fatty acid-supplementation improved HFD-impaired glucose metabolism, and supplementation with compound A, a GPR120-selective agonist, also enhanced glucose tolerance and insulin sensitivity in HFD-fed WT, but not in GPR120 KO mice . In our in vivo studies, there were improvements in glucose tolerance and insulin sensitivity observed in HFD-induced obese WT mice following DHA or GSK137647A supplementation, but not in GPR120 KO mice (Figure 7B,C). The results on WT mice fed on GSK137647A and supplemented HFD with slightly decreased body weight suggest the potent effect of body weight reduction via GPR120 activation in obesity (Figure 7A). Moreover, DHA or GSK137647A supplementation rescued Pdx1 expression and attenuated islet inflammatory responses in diet-induced obese WT mice (Figure 7D–H).
In conclusion, activation of GPR120 mediates the expression of PDX1 and protects pancreatic islet function via inhibition of CXCL1/CCL2 and NF-κB/JNK/p38 MAPK as well as macrophage infiltration. In light of our present findings and others, we propose a mechanism-of-action whereby GPR120 agonism increases insulin secretion and ameliorates anti-inflammatory effects in pancreatic β cells (Supplementary Figure S5). If confirmed, GPR120 agonism mediated protective effects may have potential pharmacotherapeutic implications in obesity and diabetes that warrants to be further explored.
The present study demonstrates that GPR120 activation is protective against lipotoxicity-impaired physiological function of pancreatic β cells, probably via modulation of PDX1 expression and inhibition of islet inflammation.
Activation of GPR120 rescues the expression levels of PDX1 and inhibits the elevation of CXCL1 and CCL2, as well as NF-κB, JNK and p38MAPK signaling during PA-induced pancreatic β-cell dysfunction and inflammation.
GPR120 supplementation with DHA and GSK137647A ameliorates high-fat-diet-induced glucose intolerance and insulin resistance in vivo, at least partially, through the increase of Pdx1 expression and decrease of islet inflammation in obese WT mice.
These findings provide a scientific basis for GPR120 agonists being considered a preventative and therapeutic strategy for obesity and obesity-related diseases.
Y.W. designed and performed experiments, analyzed and interpreted the data, and drafted the manuscript. T.X. and D.Z. performed experiments and analyzed data. P.S.L. conceived and designed the experiments, analyzed and interpreted the data, and revised the manuscript. All authors approved the manuscript. P.S.L. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
This work was fully supported by the General Research Fund of The Research Grants Council of the Hong Kong Special Administrative Region, China (Ref. No.: CUHK14107415), awarded to PS Leung.
The authors declare that there are no competing interests associated with the manuscript.
chemokine C-C motif ligand 2
chemokine C-X-C motif ligand 1
free fatty acid receptor 4
G-protein coupled receptor 40
G-protein coupled receptor 120
- GPR120 KD
- GPR120 KO
glucose-stimulated insulin secretion
c-Jun amino (N)-terminal kinases1/2
musculoaponeurotic fibrosarcoma oncogene family protein A
mitogen-activated protein kinase
neurogenic differentiation 1
- NF-κB p65
nuclear factor κB p65
pancreas and duodenal homeobox gene 1
tumor necrosis factor α