Role of glucose-dependent insulinotropic polypeptide in adipose tissue inflammation of dipeptidylpeptidase 4-deficient rats

Authors

  • Shani Ben-Shlomo,

    1. The Research Center for Digestive Tract and Liver Diseases, The Department of Gastroenterology and Liver diseases, Tel-Aviv Sourasky Medical Center and the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel
    Search for more papers by this author
  • Isabel Zvibel,

    Corresponding author
    1. The Research Center for Digestive Tract and Liver Diseases, The Department of Gastroenterology and Liver diseases, Tel-Aviv Sourasky Medical Center and the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel
    Search for more papers by this author
  • Chen Varol,

    1. The Research Center for Digestive Tract and Liver Diseases, The Department of Gastroenterology and Liver diseases, Tel-Aviv Sourasky Medical Center and the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel
    Search for more papers by this author
  • Lior Spektor,

    1. The Research Center for Digestive Tract and Liver Diseases, The Department of Gastroenterology and Liver diseases, Tel-Aviv Sourasky Medical Center and the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel
    Search for more papers by this author
  • Amir Shlomai,

    1. The Research Center for Digestive Tract and Liver Diseases, The Department of Gastroenterology and Liver diseases, Tel-Aviv Sourasky Medical Center and the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel
    Search for more papers by this author
  • Erwin M. Santo,

    1. The Research Center for Digestive Tract and Liver Diseases, The Department of Gastroenterology and Liver diseases, Tel-Aviv Sourasky Medical Center and the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel
    Search for more papers by this author
  • Zamir Halpern,

    1. The Research Center for Digestive Tract and Liver Diseases, The Department of Gastroenterology and Liver diseases, Tel-Aviv Sourasky Medical Center and the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel
    Search for more papers by this author
  • Ran Oren,

    1. Gastroenterology Institute, Hadassah Medical Center, Jerusalem, Israel
    Search for more papers by this author
  • Sigal Fishman

    1. The Research Center for Digestive Tract and Liver Diseases, The Department of Gastroenterology and Liver diseases, Tel-Aviv Sourasky Medical Center and the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel
    Search for more papers by this author

  • Disclosure: The authors declare that there is no duality of interest associated with this manuscript.

  • Author contributions: SBS, IZ, SF, LS and CV contributed to the experimental design, data collection, analysis and manuscript preparation. SBS, IZ, SF, LS, CV, AS, EMS, ZH, RO contributed to the manuscript preparation, data analysis and intellectual input. All the authors have approved the final version of the article.

  • Funding agencies: The studies were supported by a grant from the Israeli Society for the Study of Liver Disease to S.F.

Abstract

Objectives

Dipeptidyl peptidase 4 (DPP4) inhibitors, used in obese diabetic patients, reduce inflammation in several models. The role of chronic DPP4-deficiency (DPP4-) in diet-induced obesity with respect to insulin sensitivity and adipose tissue inflammation was investigated.

Design and Methods

Insulin resistance was induced by 2 months high fat diet (HFD). In vitro effects of glucose-dependent insulinotropic polypeptide (GIP) were assessed in adipose tissue explants and stromal vascular fraction (SVF).

Results

HFD-fed DPP4-rats gained significantly more weight and visceral fat mass, yet were more insulin sensitive. Adipose tissue of DPP4- rats demonstrated increased adipocyte maturation and increased expression of enzymes involved in triglyceride uptake and synthesis, yet increased adiponectin mRNA, reduced mRNA of proinflammatory cytokines and reduced vascular adhesion molecules, suggesting reduced inflammation. In vitro and in vivo experiments explored the role of GIP in inducing this phenotype. Indeed, we demonstrated that GIP directly enhanced adiponectin expression in rat and human adipose tissue explants and in SVF. Lastly, GIP administration to normal or HFD-fed rats elevated serum adiponectin and improved their glucose tolerance test.

Conclusion

In a HFD model, DPP4-rats exhibited reduced adipose tissue inflammation and improved insulin resistance, which may be mediated in part by GIP induction of adiponectin.

Introduction

A large body of evidence has shown that development of insulin resistance in obese individuals is induced at least in part by the chronic inflammatory state of the adipose tissue [1]. It is well established that adipose tissue, in addition to its endocrine and metabolic functions, is a pivotal organ with immunoregulatory properties. In obesity, the adipose tissue becomes largely infiltrated with classically activated macrophages (M1) and T cell subpopulations, both of which secrete proinflammatory cytokines such as TNFα, IL-6, and iNOS and thus, contribute to insulin resistance through impairment of insulin signaling [1].

The enzyme dipeptidyl peptidase 4 is an exopeptidase present both in the serum and on the cell surface of many cell types, for instance endothelial cells, T cells, and monocytes [5]. DPP4 cleaves and inactivates numerous substrates, including the incretins glucose-dependent insulinotropic polypeptide (GIP) and glucagon like peptide-1 (GLP-1), as well as several other peptides and cytokines [6]. Notably, several recent studies have shown that DPP4 inhibitors exert anti-inflammatory effects on adipose tissue in diet-induced obesity (DIO), as evidenced by reduced levels of infiltrating macrophages and CD8+ T cells, as well as reduced levels of proinflammatory cytokines and reduced monocyte migration in response to chemotactic stimuli [7]. There are two possible mechanisms responsible for the anti-inflammatory effects of DPP4 inhibitors: direct inhibition of cell surface DPP4 may lead to attenuation of inflammatory signals or active elevated DPP4 substrates may provide the anti-inflammatory signals, the best candidates being the incretins [11].

The incretins, GIP and GLP-1, are glucose-lowering gut-derived peptides that account for ∼50-70% of the total insulin secreted after oral glucose administration. High fat diet (HFD) has been shown to regulate incretin levels in an opposite manner: it reduces secretion of GLP-1 and enhances levels of GIP [12]. The adipose tissue expresses both GIP and GLP-1 receptors, yet it is a well established target of GIP, in which GIP receptor (GIPR) levels increase with maturation of adipocytes [13]. GIP mediates several anabolic effects in adipose tissue, including stimulation of glucose uptake, increased expression of lipoprotein lipase (LPL) and of the lipogenesis enzyme fatty acid synthase (FAS) and reduction of glucagon-induced lipolysis, indicating a role for GIP in fat accumulation in adipose tissue [14, 15].

The role of GIP in insulin resistance in DIO models is still controversial. Some studies have shown deleterious metabolic effects of GIP [16], while other studies administering the DPP4-resistant GIP, (d-Ala2)-GIP, showed either no effect, or even resulted in reduced body weight and adipose tissue mass [19, 20]. Moreover, GIP-overexpression in transgenic mice leads to improved systemic metabolic phenotype and reduced adipose tissue inflammation [21].

Our studies have explored the long-term effect of lack of DPP4 in a HFD-induced insulin resistance model from metabolic and anti-inflammatory perspectives, especially in light of the extensive use of DPP4 inhibitors in obese diabetic patients. For this purpose, we have used the F344/DuCrg rat strain (DPP4-) with a point mutation in the DPP4 sequence, leading to rapid degradation of the protein. As a result, DPP4 is not present on the cell surface, but its substrates remain chronically elevated. Using a HFD-induced insulin resistance model in these rats, we have defined systemic metabolic parameters, as well as the phenotype of adipose tissue. In addition, we have investigated whether GIP may be the mediator of the improved metabolic and inflammatory phenotype observed in this model.

Methods

Animals

Young male F344/DuCrg DPP4- deficient rats (DPP4-) (obtained from the Liver Unit, Albert Einstein College of Medicine, Bronx, NY) and weight-matched wild type F344/jcl (WT) were fed ad libitum with regular rat chow (RC) (64% carbohydrate, 30% protein, and 6% fat). Male DPP4- and WT rats were fed for 1 week, 1 month, or 2 months a high fat diet (HFD) (PMI Nutrition International, Richmond, IN cat#AIN-76A Western diet, 44.5% carbohydrate, 15.5% protein, 40% fat). We administered two intraperitoneal (i.p.) GIP injections (10 μg kg−1 body weight (BW)), 3 h apart, to normal WT rats fed RC or twice daily i.p. GIP injections (10 μg kg−1 BW) for the last 2 weeks to 2 months HFD-fed WT rats. The studies were approved by the Animal Care and Use Committee of the Tel Aviv Sourasky Medical Center.

Patients

Visceral adipose tissue was obtained from patients who underwent oncology surgery and gave their informed consent previous to surgery. The protocol was approved by the local ethical committee (# 0209-10).

Intraperitoneal glucose tolerance test (IPGTT) and insulin tolerance test (ITT)

For the IPGTT, glucose was administered by i.p. injection at the dose of 1.5 g kg−1 BW after 5 h of fasting. Blood glucose was measured using the Freestyle Freedom blood glucometer (Abbott Diabetes Care, Alameda, CA) at time 0 and then 10, 20, 30, 60, 90, and 120 min after the i.p. glucose injection. For the ITT, following 2 h of fasting, insulin was administered by i.p. injection at the dose of 1.5 U kg−1 BW, then serum glucose was measured as described for the IPGTT.

Assays and analytical procedures

Serum active GLP-1 and adiponectin from serum and adiponectin from medium of adipose tissue explants were determined using rat-specific ELISA kits (Linco Research, St. Charles, MO), and insulin was determined using an RIA kit (DiaSorin, Stillwater, MN). Serum triglycerides (TG) and cholesterol were measured using a Hitachi 747 Automatic Analyzer.

Ex vivo adipose tissue explants

Epididymal fat from male rats, or human visceral adipose tissue, were weighted and cut into small pieces (2-3 mm3) with scissors into the wells of a 24-well plate containing M199 (Gibco) supplemented with 10% fetal bovine serum (FBS) and incubated for 1 h at 37°C. The medium was then removed and changed to M199 containing 0.5% bovine serum albumin (BSA) for 24 h. The explants were pretreated with 1 μmol l−1 of the GIP receptor antagonist GIP [7] (Sigma, St Louis, MO) for 1 h, then treated for 24 h with 1 nmol l−1 insulin and 100 nmol l−1 GIP (Preprotech, Israel). The explants were collected and immediately frozen in liquid nitrogen. The medium was collected and placed at −20°C until assayed for adiponectin.

Isolation of the SVF from adipose tissue

Epididymal fat pads from 2 month HFD-fed rats were cut into small pieces and digested in DMEM, 12.5 mM HEPES pH 7.4, 2% BSA and 10 mg collagenase type II for each mouse tissue (Sigma, St Louis, MO) for 20 min at 37°C into a shaking bath at 100 rpm. The digested tissue was filtered through a 250 μm nylon sieve and centrifuged at room temperature at 500g for 5 min. The pellet was washed and erythrocytes lysed with red blood cells lysis solution. The SVF cells were plated in six-well plates at a density of 0.5 × 106 cells/well in RPMI with 10%FBS. After 4 h, the medium was changed to serum-free RPMI and cells were treated for 24 h with 100 nmol l−1 GIP.

Western blots

Total protein from livers or adipose tissue was extracted by homogenization in ice-cold RIPA buffer (PBS, 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS, 1mM PMSF, aprotinin 5 μg ml−1, leupeptin 10 μg ml−1, pepstatin A 1 μg ml−1 and phosphatase inhibitors cocktail). Homogenates were centrifuged for 25 min at 13,000g, supernatants collected and extracts normalized to total protein content. Total protein from adipose tissue explants was extracted by incubation for 30 min on ice in lysis buffer (250 mmol l−1 sucrose, 5 mmol l−1 MgCl2, 10 mmol l−1 Tris pH 8.0, 0.5% Triton X-100, protease and phosphatase inhibitors). Proteins were separated by SDS-PAGE, blotted onto Hybond C extra and blots were blocked overnight in 5% milk. Blots were incubated with antibodies to pAkt, AMP-activated protein kinase (AMPK), pAMPK, pJNK1/2 (Cell Signaling Technology, Danvers, MA), JNK2, FAS, Akt, α-actinin (Santa Cruz Biotechnology, Santa Cruz, CA), then incubated with horseradish peroxidase-conjugated secondary antibody and subjected to chemiluminescent detection. Expression determined by densitometry was normalized to α-actinin expression, or in the case of phosphorylated proteins, to their unphosphorylated counterpart.

Quantitative real time RT-PCR

Total RNA was extracted from livers using the EZ-RNA kit (Biological Industries, Bet Haemek, Israel) and from adipose tissue using the RNeasy Lipid tissue kit (Qiagen) and 2.5 μg total RNA was reverse-transcribed using M-MLV (Promega, Madison, WI). Real time RT-PCR (qRT-PCR) was performed using the Absolute Blue QPCR SYBR green ROX mix (Thermo Fisher Scientific, Epsom, Surrey, UK) in the Corbett rotor light cycler (Corbett Robotics, Brisbane, Australia). The intron span primers were:

Phosphoenol pyruvate carboxykinase (PEPCK): 5′-AGCTGTGCCAGCCAGAGTAT-3′, 5′-ATGACCGTCTTGCTTTCGAT-3′; Glucose-6-phosphatase (G-6-P): 5′-GTGGGTCCTGGACACTGACT-3′, 5′-TTTCCACGAAAGATAGCGAGA-3′; FAS: 5′-GTGTCTGGGTGGGTGTGAGT-3′, 5′-GCTAGAGGAGCAGGCTGTGT-3′; lipoprotein lipase (LPL): 5′-CAGCAAAACCTTTGTGGTGA-3′, 5′-GAACCTGGCCACATCATTTC-3′; Stearoyl response element binding protein (SREBP-1c): 5′-CCTTACACACCCAGGTCCAG-3′ 5′-TCTTCATTGTGGCTCCTGTG-3′; SCD-1: 5′-GTATCGCCCCTACGACAAGA-3′, 5′-TCGATGAAGAACGTGGTGAA-3′; PPARγ: 5′-AAGAGCTGACCCAATGGTTG-3′, 5′-GCTTTATCCCCACAGACTCG-3′; TNFα: 5′-ACTCGAGTGACAAGCCCGTA-3′, 5′-CCTTGTCCCTTGAAGAGAACC-3′; IL-6: 5′-CGGAGAGGAGACTTCACAG-3′, 5′-ACAGTGCATCATCGCTGTTC-3′; IL-1β:5′-AGGACCCAAGCACCTTCTTT-3′, 5′-AGACAGCACGAGGCATTTTT-3′; leptin: 5′-TGACACCAAAACCCTCATCA-3′, 5′-TAGACTGCCAGGGTCTGGTC-3′; adiponectin: 5′-ACCCAAGGAAACTTGTGCAG-3′, 5′-CATCTCCTGGGTCACCCTTA-3′; plasminogen activator inhibitor (PAI-1): 5′-GCAGCACCCTTTGAAAAAGA-3′, 5′-GTCAGTCATGCCCAGCTTCT-3′; CCL7: 5′-TAAGGCTCCAGTCACCTGCT-3′, 5′-ATACCCTGCTTGGTCTGGAA-3′; Monocyte chemoattractant protein (MCP-1): 5′-TAGCATCCACGTGCTGTCTC-3′, 5′-TGCTGAAGTCCTTAGGGTTGA-3′; IL-10: 5′-AAGGACCAGCTGGACAACAT-3′, 5′-TCTCCCAGGGAATTCAAATG-3′; Arginase 1: 5′-CGAGGAGGGGTAGAGAAAGG-3, 5-ACAGACCGTGGGTTCTTCA C-3; Mannose Receptor 1: 5′-GATCGTACTTCCCTGCTTGC-3′, 5′-CGCACCCTCCATTTATCTGT-3′; CCL22: 5′-GCTCTCGTCCTTCTTGCTGT-3′, 5′-GCAGGACTTTGAGGTCCAGTA-3′; I-CAM 1: 5′-CAGACCCTGGAGATGGAGAA-3′, 5′-CCTGGGGGAAGTACTGTTCA-3′; V-CAM 1: 5′-TGCCGGAGTATACGAGTGTG-3′, 5′-CAATGGCGGGTATTACCAAG-3′; P-selection: 5′-AAGCCAACATGAAAGGATCG-3′, 5′-CCCTGCAATGTGAAACTGTG-3′; β-actin: 5′-GCTCTCTTCCAGCCTTCCTT-3′, 5′-CTTCTGCATCCTGTCAGCAA-3′; human adiponectin: 5′-CCTGGTGAGAAGGGTGAGAA-3′, 5′ CAATCCCACACTGAATGCTG-3′; human RPLP0: 5′-CCCGAGAAGACCTCCTTTTT-3, 5′-AGGAGAAGGGGGAGATGTTG-3′.

Immunohistochemistry

Paraffin-embedded epididymal fat tissue section (4 μm) were mounted on poly-L-lysine glass slides and deparaffinized with xylene and graduated ethyl alcohol, then rehydrated with Tris-buffered saline (TBS), pH 7.4. Before incubation with the primary antibody, sections were incubated twice for 5 min in a microwave oven in 10 mM citrate buffer. The sections were incubated for 1 h at room temperature with monoclonal antibodies to CD68 (Dako, Denmark), then the staining was developed using a Ventana Medical Systems kit (Tucson, AZ): slides were washed three times with TBS, incubated with a biotinylated secondary antibody, washed and labeled with peroxidase-conjugated streptavidin and the staining was visualized using DAB substrate.

Statistical analysis

The results are presented as means + SE of at least three separate experiments. Statistical significance was assessed using a paired two-tailed student t test with P value <0.05 considered significant.

Results

DPP4-rats displayed improved insulin sensitivity and serum lipid profile despite increased visceral fat

In light of several studies showing the beneficial effects of DPP4 inhibitors in inflammation [7], we have investigated the role of chronic DPP4 deficiency in HFD-induced insulin resistance. Although food intake was the same, DPP4-rats display increased body weight due to a significant increase in visceral fat (Table 1).

Table 1. Body and serum parameters of 2 months RC and HFD-fed WT and DPP4-rats
ParametersRCHFD
WTDPP4-WTDPP4-
  1. Data are averages ± STD, n = 4–8 from each group;
  2. aP < 0.05 DPP4- vs. WT on the same diet.
  3. bP < 0.05 HFD vs. RC (for both DPP4- and WT).
Body weight (g)233 ± 9.7247 ± 22357 ± 25b388 ± 26a, b
Food intake kcal/rat/week79 ± 1679 ± 1690 ± 791 ± 10
Liver weight (g)12 ± 1.511 ± 1.613.8 ± 2.515.7 ± 1.9b
Epididymal fat (g)3 ± 1.13.7 ± 1.310.0 ± 0.7b13.6 ± 1.4a, b
Perinephric fat (g)2 ± 12.5 ± 1.26 ± 2.6b10.0 ± 2.9a, b
% visceral fat/body weight2.1 ± 0.82.5 ± 0.94.6 ± 0.4b6.0 ± 0.9a, b
Serum glucose (mg dl−1)92 ± 1895 ± 4.591 ± 1097.8 ± 10
Serum insulin (μU ml−1)33.4 ± 6.245.1 ± 1640.2 ± 7.644.6 ± 8.3
Serum TG (mg dl−1)239 ± 46180 ± 22a1299 ± 525b666 ± 166a, b
Serum cholesterol (mg dl−1)72.6 ± 10.573.4 ± 6.1142.3 ± 30113.8 ± 17a
Active GLP-14.8 ± 0.215.7 ± 5.7a5 ± 1.68.8 ± 4.5

While the 2 month HFD regimen induces insulin resistance in WT rats, under the same diet, DPP4- show improved GTT (Figure 1A), as well as an improved response to insulin as measured by ITT (Figure 1B). This improved ITT in HFD-fed DPP4-rats is accompanied by increased hepatic insulin signaling as determined by a three -fold induction in pAkt expression (p = 0.03, Figure 1B). In addition, gluconeogenic enzymes G-6-P and PEPCK show a 2.3 and 5 fold-decrease respectively, in the livers of HFD-fed DPP4-rats compared to WT (P < 0.03, Figure 1C). Moreover, serum TG and cholesterol are reduced in DPP4-rats (Table 1).

Figure 1.

Two months HFD-fed DPP4-rats exhibit improved insulin sensitivity. (A) upper panel—Glucose tolerance test (GTT) was determined in WT and DPP4- rats fed either RC or HFD for 2 months. Results are expressed as blood glucose concentration (mg dl−1), lower panel—Area under the curve (AUC) was calculated based on the GTT curves. n = 8, results are expressed as means ± SE. *statistically significant compared to WT HFD P < 0.03. (B) Insulin tolerance test (ITT) was determined in WT and DPP4-rats fed HFD for 2 months. Results are expressed as blood glucose concentration (mg dl−1), lower panel—Area under the curve (AUC) was calculated based on the ITT curves. n = 4, results are expressed as means ± SE. *P < 0.05. (C) Representative Western blot showing expression of pAkt and Akt in livers of 2 months HFD-fed WT and DPP4-rats. Histogram below shows expression of pAkt normalized to Akt, n = 8, *P = 0.03. (D) Expression of mRNA for gluconeogenic genes in livers of 2 months HFD-fed WT and DPP4-rats. G6P and PEPCK mRNA expression was assessed by qRT-PCR and normalized to RPLP0. Results are expressed as means ± SE, n = 8 for each group, *P < 0.03.

Adipose tissue of DPP4-rats demonstrated increased adipocyte maturation and increased expression of enzymes regulating triglyceride uptake and synthesis

The intriguing observation that inhibition of DPP4 under HFD improves insulin sensitivity yet expands visceral fat tissue, has directed us to investigate the visceral fat phenotype with regard to maturation, lipogenesis and adipokine and cytokine profiles.

We have examined adipocyte maturation as assessed by adipocyte diameter and mRNA expression of several genes present in mature adipocytes. Indeed, DPP4-display larger adipocyte area after 2 months HFD, but not after 1 month HFD, compared to WT (Figure 2A). In addition, there is a three-fold increase in the mRNA levels of the transcription factor PPARγ, an adipocyte maturation marker, in the adipose tissue of DPP4-rats (Figure 2B). DPP4-adipose tissue also displays increased LPL and FAS mRNA levels, suggesting possible increased TG uptake and synthesis and explaining the increase in adipose tissue mass in DPP4-rats (Figure 2B). Interestingly, adipose tissue of DPP4- rats displays a 1.5- decrease in AMPK phosphorylation, an effect known to be mediated by GIP, as well as significantly higher pAkt and slightly increased FAS protein expression (Figure 2C).

Figure 2.

Adipose tissue from 2 months HFD-fed DPP4-rats has more mature adipocytes and increased expression of enzymes involved in TG uptake and lipogenesis compared to WT adipose tissue. (A) Morphology of H&E stained epididymal fat tissue from WT and DPP4- rats. Below: Adipocyte area assessed in WT and DPP4-rats fed RC, 1 and 2 months HFD. Results are means of the area+ SE of 30 adipocytes for each condition. *P < 0.05 vs. RC of same strain, #P < 0.05 vs. WT 2 month HFD. (B) mRNA expression for enzymes of TG uptake and lipogenic genes in adipose tissue of 2 months HFD-fed WT and DPP4-rats. Total RNA extracted from fat tissues was analyzed for mRNA expression of LPL, SREBP-1c, FAS, SCD-1 and PPARγ by qRT-PCR and normalized to β-actin. Results are expressed as means ± SE, n = 8 for each group, *P < 0.04. (C) Representative Western blot showing expression of pAkt, pAMPK, and FAS in adipose tissue of 2 months HFD-fed WT and DPP4-rats. Histograms below shows expression of pAkt normalized to Akt, pAMPK normalized to AMPK and FAS normalized to α-actinin. n = 8, *P = 0.003.

HFD fed DPP4-rats exhibited reduced adipose tissue inflammation, accompanied by increased adiponectin expression

We have found that HFD-fed DPP4- rats are more insulin-sensitive, despite increased VF accumulation. To elucidate the mechanisms responsible for this apparent paradox, we have examined the adipokine and cytokine profiles in the adipose tissue of HFD-fed DPP4- and WT rats.

As expected, we have found an increase in leptin mRNA levels that correlates with the increase in visceral fat (Figure 3A, P = 0.03). However, adiponectin mRNA levels are also significantly elevated in adipose tissue of DPP4- rats (Figure 3A, P = 0.01). Next, we have examined expression of specific markers for M1 proinflammatory macrophages and for M2 anti-inflammatory macrophages. Indeed, DPP4- adipose tissue expresses reduced mRNA levels of several proinflammatory cytokines, specific for M1 macrophages (TNFα, IL-1β, IL-6, PAI-1) (Figure 3B), as well as reduced expression of CCL7, a macrophage recruitment chemokine (Figure 3B). However, we have also found that DPP4-derived-adipose tissue shows lower expression of M2 markers, namely mannose receptor and CCL22, but similar levels of arginase (Figure. 3D). Furthermore, we have examined the total macrophage number in adipose tissue from WT and DPP4- rats by staining for the total macrophage marker CD68. We have found reduced numbers of CD68 positive cells in DPP4- adipose tissue (11cells/field+ 1.3), while adipose tissue from WT rats shows many more clusters positive for CD68 (23cells/field+2.1) (Figure 3E).

Figure 3.

Adipose tissue of 2 months HFD-fed DPP4- rats exhibits reduced inflammation. Expression of mRNA for adipokines (A) and proinflammatory cytokines (B) in 2 months HFD-fed WT and DPP4-rats. Total RNA extracted from epididymal adipose tissue was used for mRNA analysis of adiponectin, leptin, TNFα, IL-1β, IL-6, PAI1, CCL8, and MCP-1 by qRT-PCR and normalized to β-actin. Results are expressed as means ± SE, n = 8 for each group, *P < 0.05. (C) Representative Western blot showing expression of pJNK in adipose tissue of 2 months HFD-fed WT and DPP4-rats. Histogram below shows expression of pJNK normalized to JNK2, n = 8, *P = 0.05. (D) Expression of mRNA for M2 macrophages markers in 2 months HFD-fed WT and DPP4-rats. Total RNA extracted from fat tissue was used for mRNA analysis of CCL22, arginase and mannose R by qRT-PCR and normalized to β-actin. Results are expressed as means ± SE, n = 8 for each group, *P < 0.05. (E) Representative immunostaining using CD68 antibody a marker for total macrophage levels n = 4. (F) Expression of mRNA for adhesion molecules in 2 months HFD-fed WT and DPP4-rats. Total RNA extracted from fat tissue was used for mRNA analysis of VCAM and ICAM by qRT-PCR and normalized to β-actin. Results are expressed as means ± SE, n = 8 for each group, *P < 0.05.

Assessing expression of vascular adhesion molecules, we have observed significant decreases in mRNA expression of intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1) in adipose tissue from DPP4- rats (Figure 3F), suggesting that the reduction in macrophage number is due to reduced infiltration of M1 macrophages to the adipose tissue. In addition, DPP4- adipose tissue shows significantly lower JNK phosphorylation, indicating reduced stress and inflammation (Figure 3C).

Figure 4.

GIP increases insulin sensitivity and expression of lipogenic enzymes in adipose tissue explants. (A) Representative Western blot showing pAkt, Akt and FAS in adipose tissue explants treated with 100 nM GIP for the indicated time. (B) Representative Western blot showing pAkt and FAS in adipose tissue explants treated with 100nM GIP in the presence of 1 nM insulin for 24 h. Histograms below show expression of pAkt normalized to Akt and FAS normalized to α-actinin. Means of fold-induction versus control + SE of four separate experiments, *P = 0.03 versus control. (C) The effect of GIP treatment on mRNA expression for lipogenic and TG uptake enzymes in rat adipose tissue explants. Adipose tissue explants were treated for 24 h with 100 nM GIP in the presence of 1 nM insulin and mRNA expression for FAS and LPL was analyzed using qRT-PCR and normalized to β-actin expression. Data are fold-increase versus untreated cultures and shown as means ± SE of four separate experiments. *P = 0.02.

Collectively, our results show reduced expression of inflammatory markers in DPP4- deficient rats, accompanied by increased adiponectin expression.

GIP directly increased expression of enzymes involved in TG uptake and lipogenesis in rat adipose tissue explants

The phenotype we have observed in DPP4-deficient rats may be due either to a direct effect of DPP4 deficiency leading to reduced monocyte recruitment to adipose tissue [10] or due to elevated active forms of DPP4 substrates. Since we have noticed that the adipose tissue of DPP4-deficient rats displays an anabolic phenotype known to be induced by GIP [14, 15], we have assessed the possibility that GIP was the mediator of both the anabolic and the anti-inflammatory phenotype.

First of all, we have determined that the metabolic effect detected in adipose tissue of DPP4-rats can be induced by GIP, using ex vivo adipose tissue explants, in contrast to previous studies, where GIP signaling has been studied using the murine cell line 3T3L1 [13, 15, 18]. We have found that GIP enhances insulin signaling, as demonstrated both by increased pAkt and by increased FAS mRNA expression, a crucial enzyme involved in triglyceride synthesis (Figure 4A and 4B). Moreover, GIP increases LPL mRNA by 1.7-fold, suggesting a role for GIP in increased TG uptake (Figure 4C, P = 0.03). Interestingly, GIP treatment also enhances expression of its receptor, GIPR, in adipose tissue, suggesting that GIP increased adipocyte maturation, known to be accompanied by higher GIPR levels [13, 14] (Figure 4C).

Taken together, the results derived from in vitro studies with whole adipose tissue explants indicate that under excess nutrients, GIP potentiates insulin effects on adipose tissue maturation, mass expansion and TG uptake from the serum.

GIP increased adiponectin expression in vitro and in vivo

We have next investigated whether the induction of insulin sensitivity and anti-inflammatory effects in DPP4-adipose tissue is also possibly mediated by the elevated GIP levels. For this purpose, we have used ex vivo rat and human adipose tissue explants from RC-fed rats and non-obese operated patients, respectively, and treated them with GIP for 24 h in the presence of 1nM insulin. GIP directly increases adiponectin mRNA levels by ∼1.5-fold in both rat and human adipose tissue explants (Figure 5A and 5B). Furthermore, inhibition of GIP binding to its receptor using 1 μM of the GIP antagonist GIP [7] prevents the GIP-mediated increase in adiponectin levels (Figure 5A and 5B). After showing GIP induction of adiponectin mRNA in vitro, we have determined whether GIP could induce adiponectin in vivo. For this purpose, we have administrated two GIP injections 3 h apart to rats receiving regular chow and have assessed serum adiponectin levels 2 h after the last injection. We have observed a 20% statistically significant increase in adiponectin protein levels (Figure 5C).

Figure 5.

GIP induces adiponectin expression in vitro in rat and human adipose tissue explants and in vivo in normal rats. (A) Adipose tissue explants from RC-fed WT rats were treated with 100nM GIP and 1 μM of the GIP antagonist, GIP [7]. Expression of adiponectin mRNA was tested after 24 h treatment and normalized to expression of β-actin. Results are means + SE of four experiments, *P = 0.01. (B) Human adipose tissue explants were treated with 100 nM GIP and 1 μM GIP antagonist, GIP [7-30]. Expression of adiponectin mRNA was tested after 24-h treatment and normalized to expression of RPLP0. Results are means + SE of three experiments, *P = 0.01. (C) Serum adiponectin levels were measured in RC-fed WT rats that received two i.p. injections of GIP, 3-h apart. Serum was taken 2 h after the last injection, n = 4, *P = 0.02.

To test whether GIP also affects adiponectin levels in a HFD model, we have first determined adiponectin mRNA expression in vitro after GIP treatment. Intriguingly, GIP elevates adiponectin mRNA levels only in the SVF cells of the adipose tissue (Figure 6A). Moreover, treatment of SVF with GIP also reduces TNFα and PAI-1 expression and enhances expression of the GIPR (Figure 6C and 6D).

Figure 6.

GIP effect on adiponectin and inflammatory cytokines in ex vivo explants of adipose tissue and SVF from 2 month HFD-fed rats. Adiponectin mRNA expression in adipose tissue explants and SVF cells (A,B) respectively. Adipose tissue explants or isolated SVF cells from 2 months HFD-fed WT rats were treated for 24 h with 100 nM GIP in the presence of 1nM insulin and mRNA for adiponectin was analyzed using qRT-PCR and normalized to β-actin. Data are fold-increase versus untreated cultures and shown as means ± SE of four separate experiments, *P < 0.05. (C) Cytokine expression in SVF treated with GIP. Cells of the SVF from rats fed 2 months HFD were treated for 24 h with 100 nM GIP and mRNA expression for TNFα and PAI-1 was determined using qRT-PCR and normalized to β-actin. Data are fold-increase versus untreated cultures and shown as means ± SE of four separate experiments, *P < 0.05. (D) GIPR expression in SVF treated with GIP. The SVF cells from rats fed 2 months HFD were treated for 24 h with 100 nM GIP or 1 μM GIP antagonist, GIP [7-30] and GIPR mRNA expression was determined using qRT-PCR and normalized to β-actin. Data are fold-increase versus untreated cultures and shown as means ± SE of four separate experiments, *P < 0.05.

Finally, we have assessed the in vivo physiological effects of GIP administration to rats receiving HFD for 2 months. Indeed, twice daily injections of GIP for the last 2 weeks of the HFD feeding significantly increase serum adiponectin levels and improve the GTT of the treated rats (Figure 7A and 7B).

Figure 7.

In vivo injection of GIP to 2 months HFD-fed WT rats increases serum adiponectin and improves glucose homeostasis. (A) Serum adiponectin measured by ELISA in 2 months HFD-fed WT rats that received saline injections or GIP (10 μg kg−1 BW) twice daily for the last 2 weeks of experiment and serum was collected. n = 4, *P = 0.04. (B) GTT in 2 months HFD-fed WT rats that received saline injections or GIP (10 μg kg−1 BW) twice daily for the last 2 weeks of experiment. Results are expressed as blood glucose concentration (mg dl−1), lower panel—Area under the curve (AUC) was calculated based on the GTT curves. n = 4, results are expressed as means ± SE. *P = 0.04 versus WT HFD.

Discussion

Our study in a genetic model demonstrates that lack of DPP4 enzyme attenuates HFD-induced insulin resistance, reduces adipose tissue inflammation, reduces macrophage infiltration and enhances adiponectin, despite increased visceral fat mass. Of note, most of the previous studies using genetic models of DPP4 deficiency have only shown metabolic effects [25, 26]. In addition, for the first time, using combined in vivo and in vitro approaches, we have shown that GIP induces adiponectin expression in both human and rodent adipose tissues.

The anti-inflammatory effects of DPP4 inhibitors or deficiency may be explained either by direct effects of DPP4 inhibitors on immune cells or by effects mediated via increased levels of DPP4 substrates, such as the incretins.

The direct effects of DPP4 inhibitors have been demonstrated in human and rodent studies [7]. Use of DPP4 inhibitors in LDLR−/− mice prevents atherosclerosis development, improves metabolic parameters and downregulates monocyte infiltration to adipose tissue. The study demonstrates a direct effect of DPP4 inhibitors in attenuating monocyte migration and chemotaxis [10]. One of the mechanisms suggested is by DPP4 modulation of extracellular adenosine levels, since DPP4/ CD26 present on monocytes, macrophages and T cells forms a complex with the enzyme adenosine deaminase and potentiates its activity, inactivating adenosine [27].

One of substrates of DPP4, the incretin GIP, in addition to its metabolic role, has been shown to have anti-inflammatory effects in an atherosclerosis model [28, 29], therefore we have assumed that this is another mechanism by which DPP4 deficiency may reduce adipose tissue inflammation.

The role of GIP in HFD-induced insulin resistance is far from being elucidated and remains controversial. On one hand, GIP has been shown to have deleterious metabolic effects: targeted ablation of GIP-producing K cells reduces obesity and insulin resistance, GIPR−/− mice are resistant to obesity and GIP induces production of proinflammatory cytokines in GIPR-overexpressing 3T3L1 adipocytes [16]. On the other hand, transgenic GIP-over-expressing mice display an improved systemic metabolic phenotype, as well as reduced adipose tissue inflammation [22].

We suggest that the discrepancies between the studies showing deleterious effects of GIP [16, 17] and our results derive from the models employed, where the most significant difference is that the lack of GIP or its receptor precedes the development of obesity, while in our model, as well as in the other models [19, 22], GIP over-expression or administration occurs after the start of the obesity cascade. The in vitro studies cannot be used to extrapolate GIP effects on an adipocytes cell line [18] to the more intricate situation of adipose tissue in vivo, a tissue containing both adipocytes, but also preadipocytes and macrophages, interacting with each other during DIO.

Adipose tissue is a well known target for the anabolic effects of GIP, induction of TG uptake and lipogenesis via increased expression of LPL and FAS, respectively [14]. Our studies indicate that GIP may be responsible for the observed phenotype of DPP4- adipose tissue, since we demonstrate that GIP increases LPL and FAS expression in ex vivo adipose tissue explants. In addition, GIP improves insulin signaling in adipose tissue, as indicated by Akt phosphorylation, suggesting that this enhanced insulin signal is involved in the increased adipocyte differentiation and adipose tissue expansion seen in our model. Of note, despite presence of GLP-1R in adipose tissue, GLP-1 stimulates lipolysis [30], indicating that GLP-1 is not responsible for the metabolic effects observed in the DPP4-deficient rats.

We have further investigated the mechanism responsible for the reduced adipose tissue inflammation, which led us to the novel finding that adipose tissue of HFD-fed DPP4-rats exhibits increased adiponectin mRNA and that it is GIP that directly stimulates adiponectin expression in both human and rat adipose tissue explants. Adiponectin is an anti-inflammatory, insulin-sensitizing adipokine, whose levels are reduced in obesity [31]. Adiponectin transcription is positively regulated by PPARγ and negatively regulated by HFD and pro-inflammatory cytokines TNFα, IL-6 and IL-8, via increased endoplasmatic reticulum (ER) stress and activation of JNK [32]. Our results show that the adipose tissue of HFD-fed DPP4-rats has enhanced PPARγ, as well as reduced TNFα and pJNK, all of which may lead to the observed increase in adiponectin.

Normal adipose tissue contains mainly populations of resident, M2 anti-inflammatory type of macrophages, and during diet-induced obesity (DIO), the tissue is infiltrated by M1 proinflammatory type of macrophages [2]. Adiponectin has been shown to induce a switch from the classically proinflammatory M1 macrophages to anti-inflammatory M2 macrophages in adipose tissue [23]. Adiponectin is also known to reduce TNFα secretion, which may reduces macrophage infiltration into adipose tissue via reduced ICAM-1, VCAM-1 and E-selectin on endothelial cells [24]. Inflammatory M1 macrophages are characterized by expression of TNFα, IL-6 and PAI-1, while anti-inflammatory M2 macrophages express increased levels of CD163, CD206 (mannose receptor), arginase, IL-10 and CXCL22 [2, 3, 23]. Our results show reduced mRNA expression of both M1 and M2 cytokines in adipose tissue of HFD-fed DPP4-rats, suggesting that that in DPP4-deficient rats, there is reduced recruitment of macrophages rather than a M2 to M1 switch. In addition, supporting the reduced recruitment of M1 to adipose tissue is our data showing lower expression of macrophage chemoattractant CCL7 in the DPP4-adipose tissue.

Notably, in 2 months HFD adipose tissue, the target of GIP-enhanced adiponectin and reduced proinflammatory TNFα and PAI-1 is the SVF, in which we demonstrate expression of GIPR.

The proinflammatory cytokines produced by adipose tissue have been shown to be the key to the adipose tissue-mediated insulin resistance in obesity. Both TNFα and IL-6 are known inducers of insulin resistance in human patients and in animal models [33]. On the other hand, adiponectin has been found to directly affect hepatic insulin signaling and thereby improve insulin sensitivity [34]. In line with our results, studies using adiponectin knockout or adiponectin over-expressing transgenic mice demonstrate the anti-inflammatory role of adiponectin in adipose tissue, acting on both adipocytes and SVF [35, 36].

In conclusion, this study demonstrates a novel therapeutic potential resulting from DPP4 inhibition, mediated by GIP and resulting in increased adiponectin and reduced adipose tissue inflammatory state.

Acknowledgments

We want to thank Dr Alex Litvak, the head of our Animal Institute, for his help with our animal experiments.