Visfatin is induced by peroxisome proliferator-activated receptor gamma in human macrophages

Abstract

Obesity is a low-grade chronic inflammatory disease associated with an increased number of macrophages (adipose tissue macrophages) in adipose tissue. Within the adipose tissue, adipose tissue macrophages are the major source of visfatin/pre-B-cell colony-enhancing factor/nicotinamide phosphoribosyl transferase. The nuclear receptor peroxisome proliferator-activated receptor gamma (PPARgamma) exerts anti-inflammatory effects in macrophages by inhibiting cytokine production and enhancing alternative differentiation. In this study, we investigated whether PPARgamma modulates visfatin expression in murine (bone marrow-derived macrophage) and human (primary human resting macrophage, classical macrophage, alternative macrophage or adipose tissue macrophage) macrophage models and pre-adipocyte-derived adipocytes. We show that synthetic PPARgamma ligands increase visfatin gene expression in a PPARgamma-dependent manner in primary human resting macrophages and in adipose tissue macrophages, but not in adipocytes. The threefold increase of visfatin mRNA was paralleled by an increase of protein expression (30%) and secretion (30%). Electrophoretic mobility shift assay experiments and transient transfection assays indicated that PPARgamma induces visfatin promoter activity in human macrophages by binding to a DR1-PPARgamma response element. Finally, we show that PPARgamma ligands increase NAD(+) production in primary human macrophages and that this regulation is dampened in the presence of visfatin small interfering RNA or by the visfatin-specific inhibitor FK866. Taken together, our results suggest that PPARgamma regulates the expression of visfatin in macrophages, leading to increased levels of NAD(+).

Figure 1. PPARγ agonists regulate visfatin gene expression in human macrophages and foam cells in a PPARγ dependent manner

Primary human macrophages were incubated or not (control) with (A) GW1929 (600 nM) RSG (100 nM), for indicated time points, or (B) with GW1929 (300, 600 and 3000 nM) or RSG (50, 100 and 1000 nM) for 24 hours, or (C) transformed into foam cells by AcLDL (50μg/mL) loading before treatment with PPARγ ligands. (D), Human visceral ATM where treated with GW1929 (600 nM) during 24 hours. (E), Primary human monocytes were differentiated in macrophages in the presence or absence of GW1929 (600 nM), T0070907 (1μM) or both, added at the beginning of the differentiation. Primary human macrophages were infected with recombinant adenovirus AdGFP or AdPPARγ and treated with RSG (100 nM) for 24 hours. Visfatin (F), CD36 (G) and FABP4 (H) mRNA were analyzed by quantitative PCR and normalized to cyclophilin mRNA. Results are representative of those obtained from 3 independent macrophage preparations and are expressed relative to the levels in untreated cells set as 1. Each bar is the mean value ± SD of triplicate determinations. Statistically significant differences between treatments and controls are indicated (t test; *P<0.05; **P<0.01; ***P<0.001).

Figure 2. PPARγ agonists do not regulate visfatin gene expression in murine macrophages or human adipocytes

(A, B), Murine bone marrow-derived macrophages (BMDM) were incubated or not (control) in the presence of PPARγ ligands GW1929 (1.2 μM) or rosiglitazone (1μM). (C,D), human mature adipocytes derived from the differentiation of preadipocytes in vitro were incubated or not (control) in the presence of PPARγ ligands GW1929 (600 nM). CD36 (A, B) and visfatin (C, D) mRNA was analyzed by quantitative PCR and normalized to cyclophilin mRNA. Results are representative of at least 3 independent macrophage preparations and are expressed relative to the levels in untreated cells set as 1. Each bar is the mean value ± SD of triplicate determinations. Statistically significant differences between treatments and controls are indicated (t test; *P<0.05; ***P<0.001).

Figure 3. PPARγ binds to and activates a PPRE in the human visfatin gene promoter

(A) EMSA were performed using the end-labeled DR1-consensus-PPRE (lanes 1–2) or DR1- visfatin-PPREwt oligonucleotide in the presence of unprogrammed reticulocyte lysate or in vitro translated hPPARγ and hRXRα (lanes 3–5). Competition experiments were performed in the presence of excess of cold unlabeled wild type (wt) (lanes 6–11) or mutated (mut) DR1-visfatin-PPRE oligonucleotides (lanes 12–17). Supershift assays were performed using a anti-human PPARγ antibody (lane 18). (B) Primary human macrophages were transfected with the indicated reporter constructs (DR1-visfatin PPRE)₆ or (DR1-consensus PPRE)₆, in the presence of pSG5 empty vector or pSG5-PPARγ. Cells were treated or not (Control) with GW1929 (600nM) and luciferase activity was measured. Statistically significant differences are indicated (pSG5 vs pSG5-PPARγ §§ p< 0.01, §§§ p< 0.001; control vs GW1929 *p< 0.05, **p< 0.01).

Figure 4. PPARγ agonists induce visfatin gene expression in M1, M2 and adipose tissue macrophages

(A), Primary human monocytes were differentiated to resting macrophages (RM), treated for 24 hours with GW1929 (600 nM). Where indicated, RM were activated to M1 macrophages with recombinant human TNFα (5 ng/ml), recombinant human IL-1β (5 ng/ml) for 4 hours or LPS (100 ng/ml) for 1 hour after GW1929 treatment. (B), Primary human monocytes were differentiated in RM or M2 macrophages in the presence of IL-4 (15 ng/ml) and the PPARγ agonist GW1929 (600 nM) was added or not during differentiation process. Visfatin mRNA was analyzed by quantitative PCR and normalized to cyclophilin mRNA. Results are representative of those obtained from 5 independent macrophage preparations and are expressed relative to the levels in untreated cells set as 1. Each bar is the mean value ± SD of triplicate determinations. Statistically significant differences between treatments and controls are indicated (control vs PPARγ agonists *p< 0.05, ***p< 0.001; control vs cytokines ^§p<0.05, ^§§p<0.01).

Figure 5. PPARγ regulates visfatin protein expression and secretion in primary human macrophages

Primary human macrophages were treated or not (control) with GW1929 (600 nM) for 24 hours. (A), Intracellular visfatin and β-actin protein expression was analyzed by western blot and relative signal intensities were quantified using Quantity One Software. Results are representative of 4 independent macrophage preparations and are expressed relative to the levels in untreated cells set as 1. (B), Visfatin protein secretion was quantified in macrophage supernatant by ELISA. Results are representative of 3 independent macrophage preparations. Each bar is the mean value ± SD of triplicate determinations. Statistically significant differences between treatments and controls are indicated (t test; *P<0.01; ***P<0.001).

Figure 6. PPARγ activation affects intracellular NAD concentrations in primary human macrophages

Primary human macrophages were transfected or not with non-silencing control or silencing siRNA against human visfatin (A) or treated or not with the visfatin inhibitor FK866 (100nM) (B) or infected or not with PPARγ-expressing (AdPPARγ) or GFP (AdGFP) adenovirus (C) and subsequently treated with GW1929 (600 nM), RSG (100 nM) or DMSO during 24 hours. Cells were lysed in NAD extraction buffer and NAD+ concentrations were measured by an enzymatic cycling reaction assay and normalized to protein levels and are expressed in percentage, the control non-stimulated cells being expressed as 100%. Results are representative of those obtained from 3 independent macrophage preparations. Values are means ± SD of triplicates. Statistically significant differences are indicated (t test; control vs PPARγ agonists *p< 0.05, **p< 0.01; scrambled vs siRNA visfatin or vehicle vs FK866 ^§p<0.05, ^§§p<0.05; AdGFP +PPARγ agonists vs AdPPARγ +PPARγ agonists ^§p<0.05).