https://orcid.org/0000-0002-4908-1572
Abstract
The hypothalamus plays a critical role in controlling energy balance. High-fat diet (HFD) feeding increases the gene expression of proinflammatory mediators and decreases insulin actions in the hypothalamus. Here, we show that a gut-derived hormone, glucose-dependent insulinotropic polypeptide (GIP), whose levels are elevated during diet-induced obesity, promotes and mediates hypothalamic inflammation and insulin resistance during HFD-induced obesity. Unbiased ribonucleic acid sequencing of GIP-stimulated hypothalami revealed that hypothalamic pathways most affected by intracerebroventricular (ICV) GIP stimulation were related to inflammatory-related responses. Subsequent analysis demonstrated that GIP administered either peripherally or centrally, increased proinflammatory-related factors such as Il-6 and Socs3 in the hypothalamus, but not in the cortex of C57BL/6J male mice. Consistently, hypothalamic activation of IκB kinase-β inflammatory signaling was induced by ICV GIP. Further, hypothalamic levels of proinflammatory cytokines and Socs3 were significantly reduced by an antagonistic GIP receptor (GIPR) antibody and by GIPR deficiency. Additionally, centrally administered GIP reduced anorectic actions of insulin in the brain and diminished insulin-induced phosphorylation of Protein kinase B and Glycogen synthase kinase 3β in the hypothalamus. Collectively, these findings reveal a previously unrecognized role for brain GIP signaling in diet-induced inflammation and insulin resistance in the hypothalamus.
The hypothalamus is a critical regulator of whole-body metabolism. High-fat diet (HFD)-induced obesity is characterized by hypothalamic inflammation that often clusters with other hypothalamic hallmarks of obesity, including diminished hypothalamic sensitivity to the key metabolic hormones insulin and leptin (1-5). Consumption of a HFD rapidly and chronically increases proinflammatory cytokines and chemokines in the hypothalamus and aberrantly activates inflammatory-related signaling. Forced activation of these pathways impairs the action of insulin and leptin in the hypothalamus; in contrast, inhibition of the pathways protects animals from the effects of HFD-induced hypothalamic inflammation, including insulin and leptin resistance and adiposity (6-8). Accordingly, HFD-induced hypothalamic inflammation eventually leads to metabolic disarrangements including adiposity, a disordered glucose and lipid balance and cognitive dysfunction (1-4, 9, 10). Despite the tremendous progress made toward our understanding of the role of inflammation in obesity, a key question remains as to what promotes hypothalamic inflammation in HFD-induced obesity.
Glucose-dependent insulinotropic polypeptide (GIP, also known as gastric inhibitory polypeptide) is a 42 amino acid polypeptide produced and secreted by K cells in the upper small intestine (11-14). GIP is an incretin hormone that stimulates the secretion of insulin from beta cells (11-14). In addition to its pancreatic actions, accumulating evidence suggests that GIP also plays a role in energy balance. Genetic ablation of GIP or its receptor in mice confers long-term metabolic protection from diet-induced obesity and insulin resistance (15-20). The protection from dietary obesity was commonly observed with various experimental approaches: genetic ablation of GIP-producing K cells, vaccination against GIP, infusions of a monoclonal antibody that antagonizes GIP, and neutralizing antibodies against GIP receptor (GIPR) (21-25). Supporting this notion, serum concentrations of GIP are increased in obese animals and humans (26-31). Further, genome-wide association studies have identified multiple GIPR loci associated with body mass index and visceral fat accumulation, indicating a link between GIP and obesity (32-36). In addition, GIP agonism also exhibits an antiobesity effect. Clinical and preclinical studies have shown that GIPR agonism alone or in combination with Glucagon-like peptide-1 (GLP-1) is beneficial to food intake and body weight in obesity (21, 24, 37-40). Thus, there is the dichotomy residing within the GIP field as to whether to agonize or antagonize for metabolic benefits, and it is of interest to elucidate the biological basis underlying the metabolic effects of GIP agonism and antagonism.
Recent studies have begun to elucidate the role and mechanism of GIP signaling in the brain in the context of obesity. GIPR is expressed in the brain, including in multiple hypothalamic nuclei that regulate whole-body energy homeostasis (25, 41-43). Central administration of GIP diminished neural leptin actions, thereby causing neural leptin resistance (25). In contrast, acute inhibition of GIPR using a naturalizing antibody against GIPR markedly improved body weight and adiposity by enhancing the action of neural leptin (25). In diet-induced obesity, increased GIP signaling in the brain diminishes neural leptin action, and is a key driver of leptin resistance. Leptin resistance often coexists with other hypothalamic hallmarks of obesity such as hypothalamic inflammation and insulin resistance in dietary obesity. However, a potential hypothalamic link between GIP, inflammation, and insulin resistance remains unclear.
A growing body of evidence supports the role of GIP in peripheral inflammatory responses. Treatment of cultured and primary adipocytes with GIP resulted in an increased expression of proinflammatory cytokines and chemokines in vitro (19, 44-48). Peripheral in vivo infusion of GIP in animals and humans increased adipokines and proinflammatory cytokines in adipocytes (26, 47, 49). Higher plasma GIP levels are associated with increased proinflammatory gene expression profiles in obese humans (50). A genetic study has also supported the role of GIP in adipose inflammation by demonstrating that adipocyte-specific deletion of GIPR decreased proinflammatory cytokine interleukin (IL)-6 (19). In contrast, there is evidence for the opposing role of GIP in inflammation. The overexpression of GIP or the chronic infusion of a long-lasting GIP analog reduced adipose tissue macrophage infiltration and proinflammatory cytokine expression in adipose tissues in HFD-induced obese mice (51). Consistently, GIPR deficiency in bone marrow myeloid cells was shown to increase the expression of the proinflammatory alarmin S100A8 in adipose tissue and promoted diet-induced insulin resistance (52). These previous studies clearly suggest that GIPR signaling plays a role in regulating peripheral inflammation in a tissue/cell-specific manner. Here, we explore whether central GIP signaling mediates hypothalamic inflammation in dietary obesity.
Materials and Methods
Mice
Male mice were used for all experiments. C57BL/6J male mice (000664) and ob/ob male mice (000632) were obtained from the Jackson Laboratory. Gipr knockout male mice (Riken BRC 00782) were provided by Riken BRC (53). All mice were maintained on a 12:12 hour light–dark cycle condition (lights on 06:00-18:00) and temperature-controlled environment with ad libitum access to water and normal diet (Pico Lab 5V5R) or HFD (60% kcal fat; Research diet, D12492). For peripheral injection studies, C57/BL6 male mice were fasted overnight and administered a single intraperitoneal dose of GIP or its derivative peptide at 11 am of the following day. Two hours later, tissues were collected to perform biochemical assays. Care of all animals and procedures conformed to the Guide for Care and Use of Laboratory Animals of the US National Institutes of Health and were approved by the Institutional Animal Care and Use Committee of Baylor College of Medicine (AN-6076).
RNA sequencing and bioinformatic analyses
C57BL/6J male mice were fasted overnight and received a single intraperitoneal bolus injection of GIP1-42 (30 pmol/mouse) or saline at 11 am on the following day. The mice were further maintained in the same cage without foods for 4 hours. Total ribonucleic acid (RNA) was isolated using the same method described in “Total RNA Extraction and Quantitative Real-time PCR.” The Genomic and RNA Profiling Core first conducted Sample Quality checks using the NanoDrop spectrophotometer and Agilent Bioanalyzer 2100. We then used Illumina TruSeq RNA library preparation protocol. Briefly, a double-stranded DNA library was created using 250 ng of total RNA (measured by picogreen) plus ERCC Spike-in Mixes, preparing the fragments for hybridization onto a flow cell. First, cDNA was created using the fragmented 3′ poly(A) selected portion of total RNA and random primers. During second strand synthesis, dTTP is replaced with dUTP, which quenches the second strand during amplification, thereby achieving strand specificity. Libraries were created from the cDNA by first blunt ending the fragments, attaching an adenosine to the 3′ end and finally ligating unique adapters to the ends (for more information on this process, see below). The ligated products were then amplified using 15 cycles of polymerase chain reaction (PCR). The resulting libraries were quantified using the NanoDrop spectrophotometer and fragment size assessed with the Agilent Bioanalyzer. A qPCR quantitation was performed on the libraries to determine the concentration of adapter-ligated fragments using Applied Biosystems ViiA7 Real-Time PCR System and a KAPA Library Quant Kit. All samples were pooled equimolar, requantitated by qPCR, and reassessed on the Bioanalyzer. Using the pooled concentration from the qPCR assay, the library pool was loaded onto two separate rapid run flow cells at a concentration of 29 pM for on-board cluster generation and sequencing on the HiSeq 2500, at a read length of 100 bp, paired-end. For mapping and analyzing gene expression, the pair-ended reads were mapped to the mouse genome (UCSC mm10) using STAR (54) with NCBI RefSeq genes as the reference. STAR generated read count for each gene, which was used as the measurement for gene expression. EdgeR (55) was used to analyze the gene-based read counts to detect differentially expressed genes between the groups of interest. The false discovery rate (FDR) of the differentially expressed genes was estimated using Benjamini and Hochberg method. FDR < 0.05 was considered statistically significant. The enrichments of GO (gene ontology) and KEGG terms in the up- or downregulated genes were analyzed using GOstats (56).
Total protein extraction and western blot analysis
Proteins were extracted by homogenizing samples in lysis buffer (25 mM TrisHCl pH 7.4, 150 mM NaCl, 1% NP-40, 1 mM ethylenediamine tetra-acetate, 5% glycerol [87787 and 87788 Pierce IP Lysis Buffer, Thermo Fisher Scientific]), with protease and phosphatase inhibitor cocktails (1:100, 78442, Thermo Fisher Scientific). Equal amounts of the samples were separated by sodium dodecyl sulfate- polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane by electroblotting. The following primary antibodies were used for western blot assays: anti-SOCS-3 (1:600, Abcam, ab16030 (57)), anti-phospho-IKKb (1:1,000, Cell Signaling Technology, 2697 (58)), anti-phospho-Protein kinase B (AKT) (1:1,000, Cell Signaling Technology, 4060 (59)), anti-AKT (1:1,000, Cell Signaling Technology, 2920 (60)), anti-phospho-Glycogen synthase kinase (GSK)-3β (1:1,000, Cell Signaling Technology, 9323 (61)), anti-GSK-3β (1:1,000, Cell Signaling Technology, 9315 (62)), and anti-β-actin (1:10 000, Sigma, A3853 (63)). After incubation with primary antibodies for 48 to 72 hours at 4°C, the membranes were incubated in the secondary antibody conjugated to a fluorescent entity: IRDye 680RD goat antirabbit immunoglobulin (Ig)G (LI-COR Biosciences (64)) and/or IRDye 800CW goat antimouse IgG (LI-COR Biosciences (65)) with gentle agitation for 1 hour at room temperature. The Odyssey IR imaging system (LI-COR Biosciences) was used to measure the fluorescent intensity.
Total RNA extraction and quantitative real-time PCR
Tissues were quickly removed and frozen in liquid nitrogen and kept in −80°C until further processing. Total RNA was extracted using the RNeasy Lipid Tissue Kit (QIAGEN Sciences). cDNA was generated by iScript RT Supermix (Bio-Rad Laboratories) and used with SsoAdvanced Universal SYBR Green Supermix (Bio-Rad Laboratories) for quantitative real-time PCR analysis. qPCR assays were performed using a CFX384 Touch Real-Time PCR Detection System (Bio-Rad Laboratories). Normalized mRNA levels were expressed in arbitrary units obtained by dividing the averaged, efficiency-corrected values for sample mRNA expression by that for cyclophilin mRNA expression for each sample. The resulting values were expressed as fold change above average control levels. The primer sequences were as follows: SOCS-3 (F-CACCTGGACTCCTATGAGAA AGTG and R-GAGCATCATACTGATCCAGGAACT), PTP1B (F-GGAACAGGTACCGAGATGTCA and R-AGTCATTA TCTTCCTGATGCAATT), TCPTP (F-AGGGCTTCCTTCT AAGG and R-GTTTCATCTGCTGCACCTTCTGAG), IL-6 (F-TAGTCCTTCCTACCCCAATTTCC and R-TTGG TCCTTAGCCACTCCTTC), IL-1β (F-GCAACTGTTCCT GAACTCAACT and R-ATCTTTTGGGGTCCGTCAACT), tumor necrosis factor-α (F-CCCTCACACTCAGATCATCTTCT and R-GCTACGACGTGGGCTACAG) or cyclophilin (F-TGGAGAGCACCAAGACAGACA and R-TGCCGGAGT CGACAATGAT).
Cannula implantation and treatments
ICV cannulation (a 26-gauge single stainless steel guide cannula (C315GS-5-SPC, Plastics One, Roanoke, VA, USA)) was performed as previously described (25, 66). After ICV cannulation, the mice were housed singly and given at least 1 week to recover. For acute ICV infusion studies, male mice were fasted overnight and received a single bolus injection of the GIP species (30 pmol/mouse, Phoenix Pharmaceuticals or Tocris Bioscience, 30 pmol/mouse) or vehicle (Saline, Sigma-Aldrich) at 11 am of the following day. Two hours later, tissues were collected. For the coadministration of GIP and ESI-05, C57BL/6J mice were centrally injected with ESI-05 (0.2 nmol/mouse, Axxora, BLG-M092-05) at 5 pm coinciding with the initiation of fasting. GIP plus ESI-05 were injected at 11 am on the following day. Two hours later, the tissues were collected. For chronic ICV treatments, C57/BL6 male mice were administered with a daily bolus ICV injection of 30 pmol GIP at noon for 4 days. Two hours after the last injection of GIP coinciding with the initiation of fasting, tissues were collected for western blot analysis for pIKKα/β. Gipg013 (1 μg, MedImmune) or control IgG (Sigma-Aldrich, I4506) was centrally administered (ICV, 1 µg, every other day at 11 am for 15 days) to HFD-fed mice (20 weeks of HFD feeding).
Central insulin sensitivity test
After ICV cannulation, mice were singly housed and acclimatized for 1 week before the study. Age- and body weight–matched cohorts were used. For cellular insulin sensitivity test, lean (normal chow-fed) C57/BL6 male mice were fasted overnight and administered a single dose of the GIP (ICV, 30 pmol/mouse) at 9 am. Three hours later, insulin (5 mU, Humulin-R 100 U/mL, Lilly) was ICV administered. The hypothalamus was collected 20 minutes after a bolus injection of insulin or vehicle and then subjected to western blot analysis using anti-pAKT, AKT, pGSK3β, and GSK3β antibodies. To measure food intake, lean (normal chow-fed) C57/BL6 male mice were fasted overnight and administered a single dose of the GIP (ICV, 30 pmol/mouse) at 9 am. At 12 pm, insulin (5 mU) was ICV administered coinciding with the initiation of refeeding. Cumulative food intake was measured 24 hours after the insulin injection.
Statistics
The data are presented as mean ± standard error of the mean (SEM). Statistical analyses were performed using GraphPad Prism for a 2-tailed unpaired Student’s t-test, or 2- or 2-way analysis of variance (ANOVA) followed by post hoc Tukey’s, or Sidak’s tests. P < .05 was considered to be statistically significant. Sample sizes were designed to be sufficient to enable statistical significance to be accurately determined. No samples or animals were excluded from analysis, with the exception of exclusions due to technical errors. For in vivo studies, mice were randomly assigned to groups. The investigators were not blinded in the studies. Appropriate statistical analyses were applied, assuming a normal sample distribution, as specified in the figure legends. No estimate of variance was made between each group.
Results
Centrally administered GIP induces hypothalamic inflammation in mice
To gain insight into the putative cellular functions downstream of GIPR signaling in the hypothalamus, we conducted an unbiased RNA-seq analysis to explore GIP-induced transcriptomic signatures in the hypothalamus. Adult lean C57BL/6J mice on normal chow were surgically implanted with an ICV cannula in the lateral ventricle and were subsequently administered with a single ICV bolus injection of native GIP peptide (GIP1-42, 30 pmol) or control vehicle. The hypothalami of the mice with or without ICV GIP infusion were collected, mRNA was extracted, and subjected to whole-transcriptome RNA sequencing (Illumina HiSeqTM 2500).
A total of 331 genes were found to be differentially regulated by GIP stimulation in the mediobasal hypothalamus (Benjamini–Hochberg FDR < 0.05, Fig. 1A). Overall, 290 genes were upregulated, and the majority of the transcripts among the top 50 most upregulated genes by GIP stimulation (FDR < 5 × 10−4, fold change >4.0) are linked to inflammatory-related genes such as cytokines and chemokines (Fig. 1B). The most highly upregulated 50 genes include cytokines (Interleukin (Il)12b, Il1a, Il1b, Il1rn, Il6) and chemokines (chemokine (C-C motif) ligand (Ccl) 25, Ccl12, Ccl2, Ccl4, Ccl5, Ccl7, Cxcl1, Cxcl10, Cxcl11, Cxcl2, Cxcl9). Consistent with previous studies (19, 25), Socs3, a common negative regulator of leptin and insulin (67), was also increased in GIP-stimulated hypothalami (Fig. 1A and 1B). Further, GO and gene set enrichment analysis of the differentially expressed genes also indicated inflammatory-related signaling, including inflammatory responses, immune responses, Toll-like receptor signaling, and chemokine signaling, as the most significantly represented molecular pathways (Fig. 1C and 1D). Thus, upregulation of these genes represents the capability of GIP to selectively promote a hypothalamic inflammatory gene program, which is a major pathological signature associated with obesity.
Centrally administered GIP activates hypothalamic inflammatory program. (A) Heat map of upregulated transcripts in the hypothalamus of lean (normal chow-fed) C57BL/6J mice receiving a single ICV bolus injection of GIP (30 pmol, 4 hours) vs control saline (n = 4). (B) Top 50 hypothalamic genes that are the most upregulated in response to centrally infused GIP. False discovery rate (FDR) q-value cutoff is 0.001. Inflammatory-related genes are shown in red. (C) Gene ontology (GO) term enrichment for hypothalamic genes upregulated after the central GIP injection. (D) Gene Set Enrichment Analysis (GSEA) showing enrichment of inflammatory-related pathways in the hypothalamus of mice receiving ICV GIP infusion compared to those with vehicle injection (n = 4).
GIP induces proinflammatory cytokines and Socs3 in the hypothalamus but not in the cortex in a leptin-independent manner
To validate the results of the RNA-seq analysis, we used independent samples and performed quantitative reverse transcription PCR (qPCR). Subsequent qPCR analysis indeed confirmed marked increases in the expression of proinflammatory cytokines (Il-6, Il1β, and Tnfα) and Socs3 by GIP (Fig. 2A) in the hypothalamus. Interestingly, GIP-induced elevation of proinflammatory cytokines (Il-6, Il1β, and Tnfα) and Socs3 was not observed in the cortex (Fig. 2A), suggesting that GIP selectively increases hypothalamic proinflammatory cytokines. In agreement with GIP induction of hypothalamic inflammation, we observed that centrally administered GIP (30 pmol) increased the phosphorylation of IκB kinase (IKK) IKKα/β, a marker of cellular inflammation (Fig. 2B). Since GIP is produced in the periphery, that is, in the gut K cells, we next examined whether peripherally administered GIP was able to induce hypothalamic proinflammatory cytokines and Socs3. The peripheral injection of a stable long-acting GIP mimetic ([D-Ala2]-GIP, 60 pmol) or native GIP (300 pmol) resulted in increased hypothalamic expression of mRNAs encoding proinflammatory cytokines (Il-6, Il1β, and Tnfα) and Socs3 (Fig. 2C). Peripheral administered GIP did not increase hypothalamic expression of protein tyrosine phosphatases (Ptp1b and Tcptp), which inhibit insulin and leptin signaling (Fig. 2C). We tested whether GIP-induced proinflammatory signal is mediated through leptin action by using mice that lack leptin (ob/ob mice), because GIP negatively controls leptin actions in the hypothalamus (25), ICV-administered GIP (30 pmol) clearly increased proinflammatory cytokines (Il-6, Il1β, and Tnfα) and Socs3 in the hypothalamus (Fig. 2D) but not in the cortex (Fig. 2D) of ob/ob mice, suggesting that these inductions occur in a leptin-independent manner.
![GIP induces proinflammatory cytokines and cellular inhibitors of insulin and leptin in the hypothalamus. (A) ICV-administered GIP increased mRNA expression of proinflammatory cytokines (Il-6, Il1b, and Tnfα) in the hypothalamus (n = 12–17) of lean (normal chow-fed) C57BL/6J mice. (B) ICV-administered GIP induced SOCS3 protein and phosphorylation of IKKα/β in the hypothalamus. Left, Western blot images of hypothalamic SOCS3, phosphorylated IKKα/β and β-actin. Right, quantification of SOCS3 protein and phosphorylated IKKα/β. n = 9-11 (C) Relative mRNA expression of the indicated genes in the hypothalamus of GIP (300 pmol)- or [D-Ala2]-GIP (60 pmol for 2 hours)-treated (via IP) lean (normal chow-fed) C57BL/6J mice (n = 9–10). Data represent the mean ± SEM of 2 different experiments. (D) Relative mRNA expression of the indicated genes in the hypothalamus or the cortex of GIP-treated (ICV) ob/ob mice. Ob/ob mice received a single bolus injection (GIP1-42, 30 pmol) followed 2 hours later by tissue collection (n = 5-6). Values are presented as the mean ± SEM; *P < .05, **P < .01 based on Student’s t-test.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/endo/161/9/10.1210_endocr_bqaa102/4/m_endo_161_9_bqaa102_f2.jpeg?Expires=1747868747&Signature=QmxsYqqSdworU~RiaYiyt86sDjJ6-IVi0qiS~6n504njrQ8GK5tlp8s5JKF5sRNxIitnjTWyWicG~436yzwSp2sJWm2RFtqPSLG07izYKmCvXFd4-GuTYOjFmqI89QjphEK~TX9Ti8TqS6nt6PfohjvV42Wdsn9b6Chf5Io2idV5W5VZQJ5gURuvEEc8yfm0YjWRhjLU-Y-qyWNFQG7kFdyYK~XOYyUR-ExBhBEfehAllIZoJ6xNEYtvF7FIdjo3mdUp2pva-~l92Rlyyvrw6UhbDwC3yykc1CfZ0XqmjLyNgFHjaVubL64I256jrq2wol~8QLeDFDPXs77ZF-9GNw__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
GIP induces proinflammatory cytokines and cellular inhibitors of insulin and leptin in the hypothalamus. (A) ICV-administered GIP increased mRNA expression of proinflammatory cytokines (Il-6, Il1b, and Tnfα) in the hypothalamus (n = 12–17) of lean (normal chow-fed) C57BL/6J mice. (B) ICV-administered GIP induced SOCS3 protein and phosphorylation of IKKα/β in the hypothalamus. Left, Western blot images of hypothalamic SOCS3, phosphorylated IKKα/β and β-actin. Right, quantification of SOCS3 protein and phosphorylated IKKα/β. n = 9-11 (C) Relative mRNA expression of the indicated genes in the hypothalamus of GIP (300 pmol)- or [D-Ala2]-GIP (60 pmol for 2 hours)-treated (via IP) lean (normal chow-fed) C57BL/6J mice (n = 9–10). Data represent the mean ± SEM of 2 different experiments. (D) Relative mRNA expression of the indicated genes in the hypothalamus or the cortex of GIP-treated (ICV) ob/ob mice. Ob/ob mice received a single bolus injection (GIP1-42, 30 pmol) followed 2 hours later by tissue collection (n = 5-6). Values are presented as the mean ± SEM; *P < .05, **P < .01 based on Student’s t-test.
GIP is partially required for diet-induced hypothalamic inflammation
HFD feeding causes inflammation in the hypothalamus, and a HFD elevates serum GIP levels. Given that GIP increases hypothalamic pro-inflammatory signals, we directly determined whether GIP signaling is required for diet-induced hypothalamic inflammation by employing a well-established neutralizing antibody against GIPR (Gipg013, Ref (68)) that selectively inhibits the cellular signaling and action of GIPR (25). As previously reported (69), we confirmed that prolonged HFD feeding (20 weeks) increased the mRNA level of Il-6 in the hypothalamus (Fig. 3A). Gipg013 was infused into the brain of diet-induced obese mice to acutely inhibit brain GIPR signaling, and we found that centrally administered Gipg013 significantly reduced Il-6 level in the hypothalamus (Fig. 3A). We further confirmed this observation by assessing hypothalamic levels of proinflammatory cytokines in Gipr-deficient mice (GIPR KO) (15, 53) and control mice. Both were fed on either normal chow or a HFD diet for 4 weeks. As shown in Fig. 3B, GIPR deficiency decreased mRNA levels of pro-inflammatory cytokines and Socs3 in the hypothalamus compared to those in control mice. The hypothalamic expression levels of protein tyrosine phosphatases (Ptp1B and Tcptp) were not different between the groups under a HFD condition. Collectively, these results demonstrate that GIP sufficiently and selectively induces proinflammatory cytokines and Socs3 in the hypothalamus.
GIPR is required for HFD-induced hypothalamic inflammation. (A) Relative mRNA expression of IL-6 and IL-1β in the hypothalamus of Gipg013 (for 15 days)- or vehicle-treated (via ICV) male C57BL/6J mice fed on either normal chow (NC) or HFD starting at 4 weeks of age for 20 weeks. Data are from 2 different experiments (n = 9–21). (B) Relative mRNA expression of the indicated genes in the hypothalamus of GIPR KO mice or control mice fed on NC or HFD starting at 4 weeks of age for 4 weeks (n = 10). (C) Relative mRNA expression of the indicated genes in the hypothalamus of GIP-treated (30 pmol for 2 hours) C57BL/6J mice with or without ESI-05 (0.2 nmol) (via ICV) (n = 5-6). Mice were centrally injected with ESI-05 at 5 pm coinciding with the initiation of fasting. GIP plus ESI-05 were injected at 11 am of the following day. Two hours later, the tissues were collected and the indicated mRNAs were examined. Values are presented as the mean ± SEM; *P < .05, **P < .01, ***P < .001, ****P < .0001; 1-way ANOVA followed by Tukey’s multiple comparison tests. **P < .01 in t-test in (A).
Epac signaling is involved in GIP-dependent inflammatory cytokine production
As neural actions of GIPR were reported to be mediated through exchange protein activated by cyclic AMP (Epac) signaling (25), we investigated whether Epac signaling is involved in the process of GIP-dependent pro-inflammatory cytokine induction by centrally administering GIP with or without the EPAC-specific inhibitor ESI-05 (70). ESI-05 treatment completely abolished the GIP-induced hypothalamic induction of pro-inflammatory cytokines and Socs3 expression (Fig. 3C). Thus, Epac signaling is required to elicit GIP-dependent hypothalamic pro-inflammatory induction.
GIP inhibits insulin actions in the hypothalamus
Since hypothalamic inflammation has been proposed to limit the cellular actions of insulin, we examined whether GIP affects hypothalamic insulin signaling by delivering GIP, insulin and their combination into the brain of lean C57BL/6J mice. ICV injection of GIP markedly impaired hypothalamic cellular insulin signaling, as demonstrated by diminished insulin-dependent phosphorylation of AKT at Ser-473 and its downstream target glycogen synthase kinase 3β (GSK-3β) at Ser-9 in the hypothalamus of lean C57BL/6J mice (Fig. 4A). We further tested whether GIP negatively regulates the anorectic action of insulin in the brain. Along with impaired brain insulin signaling, GIP inhibited the insulin-induced anorectic response (Fig. 4B). These data clearly demonstrate that GIP inhibits neural insulin actions in the brain.
Centrally administered GIP diminishes cellular and anorectic actions of insulin in the brain. (A) Centrally injected GIP causes hypothalamic insulin resistance. Insulin (5 mU, 20 minutes) was administered as a single bolus ICV injection to lean (normal chow fed) C57BL/6J mice that received GIP (via ICV bolus injection) 3 hours prior to insulin stimulation. Western blot (left) and its quantification (middle and right) of hypothalamic AKT and GSK-3β phosphorylation (n = 3-4). (B) GIP (30 pmol) was ICV-injected 3 hours before insulin injection (5 mU, ICV) to overnight-fasting mice, and food intake was then measured 24 hours after insulin injection (n = 6-7). Values are presented as the mean ± SEM. *P < .05, **P < .01, ***P < .001, ****P < .0001 compared with control mice based on 2-way ANOVA followed by Sidak’s multiple comparisons tests.
Discussion
The present study demonstrates, for the first time, that GIP mediates diet-induced hypothalamic induction of proinflammatory cytokines. The unbiased RNA-seq analysis clearly suggested that inflammatory-related signaling pathways were the most influenced by centrally administered GIP. Furthermore, GIP also impaired insulin actions in the brain by diminishing the cell signaling and anorectic actions of insulin. Our findings reveal the role of GIP signaling in hypothalamic inflammation and insulin resistance in HFD-induced obesity.
HFD-induced hypothalamic inflammation has emerged as a potential pathogenic factor of obesity. Some extrinsic factors have been proposed to drive and/or promote hypothalamic inflammation during obesity. For example, the direct infusion of glucose or long-chain saturated fatty acids into the brain of rodents promotes hypothalamic inflammation (6, 71). In addition, HFD feeding causes hypothalamic accumulation of various fatty acid species that are reported to induce inflammation and insulin resistance (71, 72). Thus, nutrient components per se are thought to be potential drivers of the hypothalamic inflammatory process. Evidence also suggests a role for pro-inflammatory cytokines and chemokines in mediating hypothalamic inflammation (8, 69, 73, 74). Our findings point to the role of GIP for hypothalamic induction of proinflammatory cytokines, and add GIP to the long list of candidate mediators of hypothalamic inflammation.
A true mediator of hypothalamic inflammation should fulfill, at least, the following criteria: (1) the receptor of the mediator must be expressed in the hypothalamus, (2) it induces chronic inflammation, and (3) its inhibition alleviates hypothalamic inflammation. Indeed, our data suggest that GIP is able to increase inflammatory signals, and its inhibition reduces hypothalamic inflammation. Consistent with its hypothalamic role, the GIP receptor is widely expressed by the brain, including in hypothalamic nuclei that regulate energy balance (25, 41-43). Recently, knock-in transgenic mice expressing a Cre recombinase-dependent fluorescent reporter driven by the endogenous Gipr promoter were generated and were used to determine the precise expression pattern of Gipr in hypothalamic nuclei; expression in the paraventricular, dorsomedial, and arcuate nuclei of the hypothalamus was demonstrated in the mice (41). Consistently, RNAscope analysis revealed the presence of Gipr in areas of human hypothalamus, including the paraventricular nucleus, dorsomedial nucleus and the lateral hypothalamic nucleus (41). Thus, GIPR is clearly expressed by the hypothalamus.
GIP is mainly expressed and produced by K cells in the upper gut (11-14), and thus peripherally secreted GIP must reach the brain to induce hypothalamic inflammation. Several lines of evidence support this model. Peripherally administered GIP peptides and analogs have been reported to elicit central effects in rodents (75-77), implying that GIP may cross the blood–brain barrier. In addition, peripherally injected GIP raised the GIP levels in cerebrospinal fluid from the cisterna magna in mice (25). While earlier studies reported the presence of GIP in the brain (78), some of the subsequent studies failed to confirm the expression of GIP in the brain (25, 79). Thus, whether GIP is expressed and produced in the brain at biologically meaningful levels is still debated. Nevertheless, the accumulation of recent evidence strongly supports the idea that peripherally secreted GIP is able to reach the brain.
Our findings are congruent with and extend the prior observations demonstrating that GIP increases the production of proinflammatory cytokines and chemokines and decreases insulin sensitivity in adipocytes (19, 26, 44-50). In addition to proinflammatory cytokines, GIP induces hypothalamic SOCS3 mRNA and protein expression. SOCS3 is a negative regulator of cellular insulin signaling in various tissues (67, 80). In response to HFD feeding, hypothalamic SOCS3 is known to be induced by multiple mechanisms, including IL-6 family cytokine-STAT3 signaling (81), Toll-like receptor4-MyD88 signaling (7), IKKβ/NF-κB signaling (6), and Epac-Rap1 signaling (66, 82). Consistent with a prior genetic study (25), we pharmacologically demonstrate that GIP induces SOCS3 via Epac-Rap1 signaling, revealing a novel pathway linking GIPR to SOCS3 via EPAC.
It is also important to note that GIPR agonism can either increase or decrease the peripheral inflammation, perhaps depending on the types and the doses of the GIP peptide derivatives used. While many studies have shown that GIP derivatives increase peripheral inflammation (19, 26, 44-48), some studies showed that the long-acting [D-Ala2]-GIP reduces adipose inflammation in diet-induced obese mice when peripherally administered at high doses (51, 52). Our study demonstrated that hypothalamic inflammation was induced by peripherally administered native GIP1-42 or [D-Ala2]-GIP peptide that raised blood concentrations to 923.4 ± 406.9 pg/mL, which is thought to be at the high end of the physiological range (25). While the exact reason for this discrepancy remains unclear, it could be possible that GIPR action might be diminished by chronic activation of the GIPR via the long-lasting GIPR peptide, because GIPR can become desensitized by chronic GIP stimulation (83). In addition to inflammation, there is compelling data in the literature that both agonism and antagonism of the GIP receptor can lead to weight loss. More surprisingly, both strategies seem to potentiate the antiobesity effect of GLP-1 in combination studies. There are many biological processes involved in this regulation, with food intake reduction and nausea being a key component. The ongoing clinical trials with the GIPR antibody antagonist (21) will show if the nausea profile match that observed from a dual GIP and GLP-1 receptor agonist LY3298176 (37). In this study, we explored additional effects of the GIPR signaling in die-induced obesity to help elucidate the global understanding, without categorizing GIP as obesogenic based on genetic studies. Nevertheless, given the mixed effects of GIP on hypothalamic inflammation and energy balance in the current and previous studies (15, 19, 21, 24, 38, 40, 48, 52), the physiological and pharmacological role in GIP in obesity requires further evaluation.
The mode of the inflammatory response to GIP might be cell/tissue dependent. In the periphery, loss of GIPR in adipocytes improves local and systemic inflammation (19, 48). In contrast, GIPR deficiency in myeloid seems to promote adipocyte inflammation (52). In the brain, GIPR is expressed not only by neural cells, including GABAergic neurons and glutamatergic neurons, but also by nonneural cells, such as oligodendrocytes, mural cells, vascular and leptomeningeal cells, and ependymocytes (41). We also detected GIPR mRNA in an isolated microglia fraction (data not shown). Since the complex hypothalamic inflammatory processes require the coordinated participation of multiple different cell types, such as neurons, astrocytes, microglia, and tanycytes, in response to HFD feeding (1, 84), future work is warranted to assess what cell types mediate the central actions of GIPR. In this regard, the cell type-specific deletion of GIPR would be a powerful tool to dissect the roles of GIPR in distinct neural and nonneural populations.
It is widely acknowledged that depending on the duration and the degree of inflammation in the hypothalamus, opposite metabolic outputs can be observed (85). The acute hypothalamic inflammatory response is reported to promote negative energy balance (86). Consistently, the genetic deletion of proinflammatory cytokines or their receptors has been reported to render mice more prone to dietary obesity (87-92). The exact roles of cytokines in obesity are contradictory, and further studies are needed. The observed lack of effect of the native, intact GIP peptide on body weight and food intake may arise from the complexity of hypothalamic inflammation.
This present study has only considered the context of HFD-induced hypothalamic inflammation. However, we find that GIPR deficiency decreases hypothalamic expression of cytokines and SOCS3 under normal and HFD conditions, implying that GIPR may determine hypothalamic inflammation regardless of diets and thus its action may not be confined to HFD-dependent context. Indeed, hypothalamic inflammation is induced by and casually related to other pathophysiological conditions (3). In particular, hypothalamic inflammation occurs in aged mice and is indicated by increased gene expression of proinflammatory cytokines and inflammatory-related signaling (93). Given that GIP promotes the hypothalamic induction of proinflammatory cytokines, it is intriguing to speculate that GIP might have a role in hypothalamic inflammation in aging. Intriguingly, GIPR deficiency was recently reported to significantly extend the average and maximal lifespan of mice (94). In addition, serum GIP levels are elevated in aged mice (data not shown). Whether GIP signaling in the brain plays a role in hypothalamic inflammation in the aging processes is one of the interesting questions to be addressed in future studies.
Abbreviations
- AKT
Protein kinase B
- ANOVA
analysis of variance
- FDR
false discovery rate
- GIP
glucose-dependent insulinotropic polypeptide
- GO
gene ontology
- GSK
glycogen synthase kinase
- HFD
high-fat diet
- ICV
intracerebroventricular
- IL
interleukin
- RNA
ribonucleic acid
- SEM
standard error of the mean
Acknowledgments
Financial Support: This work was supported by USDA ARS 6250-51000-055, AHA-14BGIA20460080, NIH-P30-DK079638 and NIH R01DK104901 to M.F., AHA-15POST22500012 and the Uehara Memorial Foundation 201340214 to K.K. This project was also supported in part by the Genomic and RNA Profiling Core at Baylor College of Medicine with funding from Center Core Grant (P30) Digestive Disease Center Support Grant (NIDDK-DK56338) and P30 Cancer Center Support Grant (NCI-CA125123).
Author Contributions: Y.F., K.K., H.-Y.L., Q.M., and M.F. designed the research studies. Y.F., K.K., and H.-Y.L. conducted experiments. Y.F., K.K., H.-Y.L., Q.M., T.S., Y.F., and M.F. analyzed and interpreted the results. R.P., T.S., and Y.X. provided reagents. P.R. contributed reagents and intellectually assisted with Gipg013 studies. Y.F., K.K., H.-Y.L., Q.M., and M.F. wrote the manuscript. We gratefully acknowledge Mr. Firoz Vohra and Dr. Marta Fiorotto for the CLAMS analyses, and Ms. Zainab Mabizari, Ms. Amy Ng and Dr. Ana De la Puente Gomez for technical assistance.
Disclosure Summary: The authors have declared that no conflict of interest exists. P.R. is employed by AstraZeneca.
Additional Information
Data Availability: All data generated or analyzed during this study are included in this published article or in the data repositories listed in references.
References
Author notes
Yukiko Fu, Kentaro Kaneko, and Hsiao-Yun Lin contributed equally to this work.
