Fatty acid-induced induction of Toll-like receptor-4/nuclear factor-κB pathway in adipocytes links nutritional signalling with innate immunity

Abstract

To study the effects of fatty acids and the involvement of the Toll-like receptor-4/nuclear factor-κB (TLR-4/NF-κB) pathway with respect to the secretion of adipokines from adipocytes 3T3-L1 adipocytes were stimulated with increasing doses of fatty acids. The secretion of adiponectin, resistin and monocyte chemoattractant protein-1 (MCP-1) was measured by enzyme-linked immunosorbent assay. The NF-κB p65 nuclear translocation and TLR-4 expression were investigated by Western blot. The effects mediated by NF-κB were tested using a specific NF-κB-inhibitor and TLR-4-induced effects were analysed with a neutralizing TLR-4 antibody. Binding of ¹⁴C-labelled fatty acids to TLR-4/MD-2 was investigated using a FLAG-tagged extracellular part of TLR-4 fused to full-length MD-2 via a linker (lipopolysaccharide-Trap). The messenger RNA (mRNA) expression of adipokines in abdominal adipose tissue of rats fed a standard chow or a high-fat diet was investigated by reverse transcription–polymerase chain reaction. The TLR-4 is induced during adipocyte differentiation and its expression is enhanced following fatty acid stimulation. The stimulatory effects of stearic and palmitic acids on MCP-1 secretion and of palmitoleic acid on resistin secretion are mediated via NF-κB. The stimulatory effects of stearic, palmitic and palmitoleic acids on resistin secretion and the stimulatory effect of stearic acid on MCP-1 secretion are mediated via TLR-4. Fatty acid-mediated effects are caused by an endogenous ligand because fatty acids were shown not to bind directly to TLR-4/MD-2. Adipose tissue mRNA expression and serum levels of adipokines did not differ in rats fed a high-fat diet. These data provide a new molecular mechanism by which fatty acids can link nutrition with innate immunity.

Introduction

Obesity is regarded as a chronic and low-grade state of inflammation.^1,2 Adiponectin^3–5 and resistin^6–8 are currently discussed as key metabolic regulators in obesity, insulin resistance, type 2 diabetes mellitus and the metabolic syndrome. Whereas adiponectin usually exerts anti-inflammatory and antidiabetic effects, resistin has been controversially discussed as a proinflammatory and prodiabetic mediator. Whereas resistin is almost exclusively expressed in white adipose tissue in mice, human resistin expression is low in adipocytes and high in monocytes, lung and bone marrow.⁷ Experimental data show a high level of variability dependent on the primary cells or cell lines used and because contradictory data exist between the human and the murine system,⁸ data are sometimes difficult to interpret. Monocyte chemoattractant protein 1 (MCP-1) is mainly secreted by adipose tissue macrophages; however, adipocytes themselves also release considerable amounts of MCP-1.⁹ Importantly, adipose tissue MCP-1 has recently been recognized as a main chemokine that causes adipose tissue infiltration by monocytes/macrophages in the context of obesity and insulin resistance.¹⁰ Type 2 diabetes mellitus, insulin resistance and themetabolic syndrome are characterized by increased levels of systemic fatty acids, by an increased lipolysis of visceral adipose tissue and by an accumulation of MCP-1-attracted monocytes within the adipose tissue. Fatty acids might therefore have pronounced and direct effects on adipocytes. In contrast to the systemic effects of diets enriched with fatty acids, there are no experimental data available that systematically address the direct (i.e. paracrine/autocrine) effects of fatty acids on the adipocytic secretion of adipokines and chemokines. Recently, Shi et al. demonstrated that nutritional fatty acids can activate the Toll-like receptor-4 (TLR-4) signalling in monocytes and adipocytes.¹¹ Moreover, mice lacking TLR-4 are protected against insulin resistance induced by a high-fat diet. These data seem to support the hypothesis that TLR-4 links innate immunity and fatty acid-induced insulin resistance.

Therefore, it was the aim of the present study

1. To investigate systematically the direct effects of fatty acids on the secretion of adiponectin, resistin and MCP-1 from differentiated mature 3T3-L1 adipocytes.

2. To study the involvement of the TLR-4/nuclear factor-κB (NF-κB) pathway in the regulation of fatty-acid-induced adipokine and chemokine secretion.

3. To compare these data with the in vivo effects of a diet rich in fatty acid on adipokine expression in the visceral adipose tissue of rats.

To study potential class effects of fatty acids, a panel of five saturated, mono- and polyunsaturated C16 and C18 fatty acids was used for stimulation experiments.

Materials and methods

Adipocyte cell culture

3T3-L1-preadipocytes were cultured in a 10% CO₂ atmosphere at 37° in Dulbeccos modified Eagles’ medium (DMEM; BioWhittaker, Verviers, Belgium) supplemented with 10% newborn calf serum (Sigma Biosciences, Deisenhofen, Germany) and penicillin/streptomycin (GIBCO BRL, Berlin, Germany). At confluence, cells were differentiated into adipocytes by treating them with DMEM/F12/glutamate medium supplemented with 0·5 mm 3-isobutyl-methyl-xanthine, 10⁻⁷ m corticosterone, 10⁻⁶ m insulin, 200 μm ascorbate, 2 μg/ml transferrin, 1 μm biotin, 17 μm panthothenate and 300 mg/l Pedersen-fetuin^12,13 for 5 days. Thereafter, the cells were exposed to DMEM/F12/glutamate medium with 10⁻⁹ m insulin until they reached the fully differentiated phenotype,^14–18 this was controlled by observing the cells using light microscopy for the existence of a more rounded cell shape and the typical appearance of extensive accumulation of lipid droplets.

Stimulation experiments using fatty acids and measurement of adipokine and MCP-1 secretion

Cells were washed with phosphate-buffered saline (PBS) and incubated under serum-free culture conditions. The following fatty acids were used with the appropriate non-toxic concentrations ranging within the physiological range. These concentrations were tested in screening experiments on 3T3-L1 adipocytes and toxicity was excluded by measuring lactate dehydrogenase activity in the supernatants: C16 and C18 saturated fatty acids: palmitic acid (C16; 10, 100 μm); stearic acid (C18; 10, 100 μm); C16 and C18 monounsaturated fatty acids: palmitoleic acid (C16:1, cis-9; 1, 10 μm); oleic acid (C18:1, cis-9; 1, 10 μm); C18 polyunsaturated fatty acid: linoleic acid (18:2, n-6; 1, 10 μm). Fatty acids were dissolved (200 mm) in ethanol at 70° and then complexed 1 : 10 with 10% bovine serum albumin at 55° (20 mm) for 10 min. The formation of albumin complexes is important for reducing the possible cell toxicity of fatty acids. For controlling cell viability and for data normalization, the lactate dehydrogenase (LDH) assay (Roche, Mannheim, Germany) and total protein concentration (bicinchonic acid assay; Interchim, Montlucon, France) was measured. For each experimental group, six independent experiments were performed (stimulation by fatty acids was 24 hr) and adipokines and MCP-1 were measured in duplicate by enzyme-linked immunosorbent assay (ELISA). Data were normalized to total protein content and are given as ng/ml/24 hr or pg/ml/24 hr to describe the secretion rate. Adiponectin, resistin and MCP-1 were measured in cell culture supernatants using ELISA (all from R&D Systems Europe, Abingdon, UK). As a specific NF-κB inhibitor, 6-amino-4-(phenoxyphenylethylamino)quinazoline (Calbiochem, Darmstadt, Germany) was used (n= 3) in a non-toxic concentration of 22 nm (50% inhibitory concentration = 11 nm). A neutralizing and specific TLR-4 antibody was used to block TLR-4-mediated effects (anti-TLR-4/MD-2, mouse MAB; Axxora, Lörrach, Germany). The antibody (isotype rat immunoglobulin G2a; clone MTS510) recognizes and neutralizes the TLR-4/MD-2 complex on Ba/F3 cells and is routinely used for blocking TLR-4-mediated effects.

Western blot analysis

Equal amounts of total protein were submitted to gel electrophoresis after a 1 : 1 dilution with Laemmli loading buffer. Sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) was performed following standard procedures. Proteins were transferred to a Fluotrans Transfer Membrane (Pall Corp., Portsmouth, UK). Transfer was confirmed by Ponceau Red staining of the membrane. The NF-κB p65 primary antibody (rabbit; Cell Signaling, Boston, MA) was used at a 1 : 1000 dilution, the secondary horseradish peroxidase-coupled anti-rabbit antibody (Dianova, Hamburg, Germany) at a 1 : 5000 dilution in a 5% non-fat dried milk/PBS suspension. Detection of the immune complexes was carried out with the enhanced chemiluminescence Western blot detection system (Amersham Corp., Freiburg, Germany). Antibodies against TLR-4 (Glaxo-Wellcome, Stevenage, UK) were raised in rabbits and affinity purified as described elsewhere.¹⁹

Monitoring of quantitative gene expression by real-time reverse transcription–polymerase chain reaction

A 2-μg sample of total RNA was reverse transcribed using the Promega Reverse Transcription System (Promega, Madison, WI) in a volume of 40 μl; 2 μl complementary DNA was amplified in glass capillaries (LightCycler) using polymerase chain reaction (PCR) primers specific for rat adiponectin, resistin, MCP-1 and β-actin. The reaction conditions were as follows: 2 μl complementary DNA, 2 μl 10× LightCycler-Fast Start DNA Master SYBR Green I (Roche Diagnostics GmbH, Mannheim, Germany), 2·4 μl MgCl₂ (25 mm), 1 μl of each primer (5 pmol/μl), in a total volume of 20 μl. Amplification in the LightCycler capillaries was for 45 cycles with initial incubation of 10 min at 95° for activation of the Taq polymerase. Cycling parameters were 10 seconds at 95°, 10 seconds at 63° and 15 seconds at 72°. Fluorescence was monitored at 83° for CORS-26 and 85° for β-actin. For quantification of the results obtained by real-time PCR the standard curve method was used. Primer sequences: resistin-universe: atcaagacttcagctccctactg; resistin-reverse: 5′-cagtgacacattgtatcctcacg-3′; MCP-1-universe: 5′-cagaaaccagccaactctca-3′; MCP-1-reverse: 5′-tctcttgagcttggtgacaaata-3′; adiponectin-universe: 5′-gcccagtcatgaagggatta-3′; adiponectin-reverse: 5′-cccttaggaccaagaacacct-3′; β-actin-universe: 5′-ttgtaaccaactgggacgatatgg-3′; β-actin-reverse: 5′-cttgatcttcatggtgctagg-3′.

Immunoprecipitation

To determine whether fatty acids can bind directly to TLR-4, the binding of ¹⁴C-labelled stearic and oleic acids to a soluble fusion complex consisting of the FLAG-tagged extracellular part of TLR-4 fused to full-length MD-2 via a flexible linker was measured. This protein was shown to bind and neutralize lipopolysaccharide (LPS).²⁰ Flag-TLR-4/MD-2 fusion protein (LPS-Trap) containing Schneider S2 supernatants were incubated with biotinylated LPS (0·5 μg/ml) (Invivogen, San Diego, USA) and with or without 100× molar excess of stearic or oleic acid, respectively, for 1 hr at room temperature, followed by immunoprecipitation with 20 μl streptavidin–Sepharose (GE Healthcare Bioscience AB, Uppsala, Sweden). This excess of fatty acids should effectively compete with LPS if a specific binding of fatty acids occurs. The precipitate was washed three times with PBS, boiled with Laemmli loading buffer and subjected to SDS–PAGE. After blotting, fusion proteins were detected by Anti-Flag-horseradish peroxidase (Sigma, Taufkirchen, Germany) and enhanced chemiluminescence substrate. To exclude possible LPS contamination, the LAL Chromogenic Endpoint Assay (Hycult biotechnology, Uden, the Netherlands) was routinely used.

Fatty acid binding assay

Flag-TLR-4/MD-2 fusion protein or Flag-interleukin-1-receptor a as control were immunoprecipitated by Anti-Flag M2 Agarose (Sigma) at 4° overnight. The precipitate was incubated with 1 μm or 10 μm albumin-coupled ¹⁴C-labelled stearic or oleic acid, respectively, with or without a 100× molar excess of unlabelled albumin-coupled fatty acids in Tris-buffered saline with 20% fetal calf serum for 2 hr at room temperature. The agarose gel was washed three times with Tris-buffered saline and subsequently resuspended in PBS with 2% SDS containing 5 mm ethylenediaminetetraacetic acid to ensure detachment of bound substances from the agarose gel. Gel suspension was directly mixed with Lumasafe Plus scintillation cocktail (Lumac, Groningen, the Netherlands) and radioactivity was measured on a 1600R liquid scintillation analyser (Canberra-Packard, Schwadorf, Austria). The % specific binding was calculated as: (bound ¹⁴C – bound ¹⁴C with excess unlabelled fatty acid)/total input of ¹⁴C. No binding to the control protein was detected.

Experimental animals

Six-week-old, male Wistar rats (Charles River, Sulzfeld, Germany) were kept individually with free access to water and subjected to either a standard chow or a high-fat diet as described previously.²¹ Diets were ordered in pellet form from Altromin (Lage, Germany). Animals were kept on a 12 : 12-hr light : dark cycle. All animal procedures were approved by the local animal rights committee. After 3 days of acclimatization, rats had access to either a standard chow (fat content 11% of energy) or a high-fat diet (fat content 42% of energy). The fat portion of the high-fat diet consists of 45% saturated fatty acids (mainly palmitic acid and stearic acid), 40% monounsaturated fatty acids (mainly oleic acid) and 15% polyunsaturated fatty acids (mainly linoleic acid). Weight gain and food intake were monitored once a week. Data on final body weight or general characteristics of these experimental animals have been reported previously.²¹ After 12 weeks, the animals were killed after an overnight fast (16 hr) and abdominal adipose tissue samples were collected for immediate mRNA preparation following standard protocols. Simultaneously, serum aliquots were sampled for measurement of adiponectin, resistin and MCP-1 using ELISA (R&D Systems Europe).

Statistics

Means were compared by the Mann–Whitney U-test. A P-value below 0·05 (two-tailed) was considered to be statistically significant.

Results

Adiponectin secretion upon stimulation with fatty acids

Palmitic acid

Adiponectin secretion increased significantly from 1·6 ±0·1 up to 2·7 ± 0·2 ng/ml/24 hr (P= 0·001) and 2·5 ±0·2 ng/ml/24 hr (P= 0·007) after stimulation with 10 and 100 μm palmitic acid, respectively (Fig. 1a). The higher dose of 100 μm could not further enhance adiponectin secretion.

Figure 1.

Effects of fatty acids on the secretion of adiponectin from differentiated 3T3-L1 adipocytes. Six wells were used for each experimental group. After data normalization for total protein content, the secretion of adiponectin is given in ng/ml/24 hr. Box plots show median (thick black line), lower and upper quartiles (box) and lower and upper extreme values (thin vertical lines). The dose effects of palmitic acid (a), stearic acid (b), oleic acid (c), palmitoleic acid (d) and linoleic acid (e) were investigated.

Stearic acid

Stearic acid at 10 μm stimulated adiponectin secretion from 1·6 ± 0·1 up to 2·8 ± 0·2 ng/ml/24 hr (P= 0·02) (Fig. 1b). There was a further increase in adiponectin secretion when using 100 μm stearic acid (3·2 ± 0·1 ng/ml/24 hr). However, this trend did not reach statistical significance when compared to stimulation using 10 μm stearic acid.

Oleic acid

Adiponectin secretion increased from a basal value of 1·6 ± 0·1 up to 2·6 ± 0·1 ng/ml/24 hr (P< 0·001) upon stimulation with 1 mm oleic acid (Fig. 1c). The further increase to 3·1 ± 0·3 ng/ml/24 hr following stimulation with 10 μm oleic acid did not reach statistical significance.

Palmitoleic acid

Both 1 and 10 μm palmitoleic acid stimulated adiponectin secretion from 1·6 ± 0·1 up to 2·6 ± 0·2 ng/ml/24 hr (P=0·003) and 2·4 ± 0·1 ng/ml/24 hr (P< 0·001) without any significant difference between the two doses (Fig. 1d).

Linoleic acid

Linoleic acid increased adiponectin secretion stepwise from 1·6 ± 0·1 up to 2·6 ± 0·1 ng/ml/24 hr at 1 μm (P< 0·001) and further up to 2·9 ± 0·1 ng/ml/24 hr at 10 μm (P< 0·001) (Fig. 1e).

Resistin secretion upon stimulation with fatty acids

Palmitic acid

Stimulation with palmitic acid increased resistin secretion from 384 ± 23 up to 591 ± 54 ng/ml/24 hr at 10 μm (P= 0·005) (Fig. 2a). The higher dose of 100 μm palmitic acid was not effective in elevating resistin secretion (388 ± 40 ng/ml/24 hr).

Figure 2.

Effects of fatty acids on the secretion of resistin from differentiated 3T3-L1 adipocytes. Six wells were used for each experimental group. After data normalization for total protein content, the secretion of adiponectin is given in ng/ml/24 hr. Box plots show median (thick black line), lower and upper quartiles (box) and lower and upper extreme values (thin vertical lines). The dose effects of palmitic acid (a), stearic acid (b), oleic acid (c), palmitoleic acid (d) and linoleic acid (e) were investigated.

Stearic acid

Both 10 and 100 μm stearic acid enhanced resistin secretion significantly from 384 ± 23 up to 604 ± 63 ng/ml/24 hr (P= 0·008) and 629 ± 41 ng/ml/24 hr (P< 0·001), respectively (Fig. 2b).

Oleic acid

Basal resistin secretion (384 ± 23 ng/ml) was significantly enhanced by 1 μm oleic acid (620 ± 34 ng/ml/24 hr; P< 0·001) and by 10 μm oleic acid (501 ± 30 ng/ml/24 hr; P< 0·001) (Fig. 2c).

Palmitoleic acid

Whereas 1 μm palmitoleic acid could stimulate resistin secretion from 384 ± 23 up to 655 ± 48 ng/ml/24 hr (P< 0·001), the higher dose of 10 μm was not effective (Fig. 2d).

Linoleic acid

Linoleic acid increased resistin secretion from 384 ± 23 up to 709 ± 36 ng/ml/24 hr (P< 0·001) at a dose of 1 μm, whereas the higher dose of 10 μm was ineffective (Fig. 2e).

MCP-1 secretion upon stimulation with fatty acids

Stearic acid

Basal MCP-1 secretion was 174 ± 20 pg/ml/24 hr. Whereas stimulation with 10 μm stearic acid caused a significant inhibition of MCP-1 secretion down to 71 ± 12 pg/ml/24 hr (P= 0·001) (Fig. 3a), the high dose of 100 μm stearic acid strongly induced MCP-1 secretion up to 573 ± 39 pg/ml/24 hr (P< 0·001).

Figure 3.

Effects of fatty acids on the secretion of monocyte chemoattractant protein-1 (MCP-1) from differentiated 3T3-L1 adipocytes. Six wells were used for each experimental group. After data normalization for total protein content, the secretion of MCP-1 (n= 6) is given in pg/ml/24 hr. Box plots show median (thick black line), lower and upper quartiles (box) and lower and upper extreme values (thin vertical lines). The dose effects of palmitic acid, stearic acid (a), oleic acid, palmitoleic acid (b) and linoleic acid (c) were investigated.

Palmitic acid

Palmitic acid in the lower doses of 10 μm did not influence MCP-1 secretion (118 ± 17 pg/ml/24 hr); however, 100 μm palmitic acid strongly stimulated MCP-1 secretion up to 648 ± 70 pg/ml/24 hr (P< 0·001) (Fig. 3a).

Oleic acid

Oleic acid at a dose of 1 μm strongly inhibited MCP-1 secretion from 174 ± 20 down to 35 ± 4 pg/ml/24 hr (P< 0·001) (Fig 3b). Furthermore, the higher dose of 10 μm completely blocked MCP-1 secretion to undetectable levels (P< 0·001).

Palmitoleic acid

Similarly, 1 μm palmitoleic acid reduced MCP-1 secretion from 174 ± 20 down to 25 ± 6 pg/ml/24 hr (P< 0·001) (Fig. 3b). However, higher doses of palmitoleic acid (10 μm) had no impact on MCP-1 secretion (210 ± 23 pg/ml/24 hr; P= 0·2).

Linoleic acid

Linoleic acid at a dose of 1 μm inhibited MCP-1 secretion from 174 ± 20 down to 32 ± 5 pg/ml/24 hr (P< 0·001), whereas the higher dose of 10 μm was not effective (113 ± 32 pg/ml/24 hr; P= 0·1) (Fig. 3c).

Effects of NF-κB inhibition on the fatty acid-induced secretion of proinflammatory resistin and MCP-1

Resistin and MCP-1 are proinflammatory mediators and NF-κB potentially mediates the fatty acid-induced effects on the secretion of MCP-1 and resistin so we repeated the experiments (n= 3) using stearic acid (100 μm) and palmitic acid (100 μm) in the presence and absence of the soluble and cell-permeable NF-κB inhibitor 6-amino-4-(phenoxyphenylethylamino)quinazoline (22 nm).

The palmitoleic acid induced stimulation of resistin secretion (Fig. 4a) from 255 ± 59 ng/ml/24 hr under basal conditions to 688 ± 79 ng/ml/24 hr after 1 μm palmitoleic acid (P= 0·009) could be significantly reduced to 94 ± 3 ng/ml 24 hr (P= 0·002) by NF-κB inhibition (P= 0·001).

Figure 4.

The role of nuclear factor-κB (NF-κB) in the fatty acid-induced secretion of resistin and monocyte chemoattractant protein-1 (MCP-1) from differentiated 3T3-L1 adipocytes. (a) Effects of NF-κB inhibition on the basal and palmitoleic acid-induced secretion of resistin. Three wells were used for each experimental group. After data normalization for total protein content, the secretion of resistin (n= 3) is given in pg/ml/24 hr. Box plots show median (thick black line), lower and upper quartiles (box) and lower and upper extreme values (thin vertical lines). (b) Effects of NF-κB inhibition on the basal and stearic acid-induced and palmitic acid-induced secretion of MCP-1. Three wells were used for each experimental group. After data normalization for total protein content, the secretion of MCP-1 (n= 3) is given in pg/ml/24 hr. Box plots show median (thick black line), lower and upper quartiles (box) and lower and upper extreme values (thin vertical lines). (c) Western blot analysis using nuclear extracts prepared from 3T3-L1 adipocytes. When compared to controls (absence of fatty acids), stearic acid, palmitic acid and palmitoleic acid increase NF-κB p65 nuclear translocation. As a loading control, the Coomassie stained gel is shown.

The basal secretion of MCP-1 (Fig. 4b) was 109 ± 13 pg/ml/24 hr and this basal secretion could be suppressed by NF-κB inhibition to 26 ± 11 pg/ml/24 hr (P= 0·008). Both the stearic acid-induced (346 ± 22 pg/ml/24 hr; P= 0·001 versus control) and the palmitic acid-induced (206 ± 70 pg/ml/24 hr; P= 0·036 versus control) secretion of MCP-1 was significantly inhibited by NF-κB inhibition down to 217 ± 15 pg/ml/24 hr (P= 0·008) in the case of stearic acid and down to 70 ± 5 pg/ml/24 hr (P= 0·009) in the case of palmitic acid.

Table 1 summarizes the effects of fatty acids on adipokine and chemokine secretion from adipocytes.

Table 1.

Summary of fatty acid-induced effects on the adipocytic secretion of adiponectin, resistin and monocyte chemoattractant protein-1 (MCP-1)

FA classification	FA	FA dose (μm)	Adiponectin	Resistin	MCP-1
C16 saturated	Stearic acid	10	↑	↑	↓
	Stearic acid	100	↑	↑2	↑↑12
C18 saturated	Palmitic acid	10	↑	↑2	↔
	Palmitic acid	100	↑	↔	↑↑1
C18 monounsaturated	Oleic acid	1	↑	↑	↓
	Oleic acid	10	↑	↑	↓↓
C16 monounsaturated	Palmitoleic acid	1	↑	↑12	↓↓
	Palmitoleic acid	10	↑	↔	↔
C18 polyunsaturated	Linoleic acid	1	↑	↑	↓
	Linoleic acid	10	↑	↔	↔

FA = fatty acid, ↔ no change, ↑ up-regulation, ↓ down-regulation; ↑↑ strong up-regulation, ↓↓ strong down-regulation.

¹Effect mediated via nuclear factor-κB

²effect mediated via Toll-like receptor-4.

Effects of fatty acids on nuclear NF-κB p65 translocation in adipocytes

Nuclear extracts were prepared from 3T3-L1 adipocytes and Western blot analysis was performed to investigate whether fatty acids are able to increase nuclear translocation of NF-κB p65 in adipocytes. As shown in Fig. 4(c), stearic acid, palmitic acid and palmitoleic acid strongly increased the nuclear translocation of NF-κB p65.

Effects of antibody-mediated TLR-4 blockade on the fatty acid-induced secretion of resistin and MCP-1 from differentiated 3T3-L1 adipocytes

To investigate whether fatty acid-induced effects on adipokine secretion are mediated via TLR-4, experiments were performed using a neutralizing TLR-4 antibody at increasing doses.

Figure 5(a) summarizes the effects of TLR-4 blockade on the secretion of resistin induced by stearic acid, palmitic acid, and palmitoleic acid from adipocytes. The TLR-4 blockade significantly inhibited the fatty acid-induced effects on resistin secretion in a dose-dependent manner.

Figure 5.

Effects of antibody-mediated Toll-like receptor-4 (TLR-4) blockade on the fatty acid-induced secretion of resistin and monocyte chemoattractant protein-1 (MCP-1) from differentiated 3T3-L1 adipocytes. Three wells were used for each experimental group. After data normalization for total protein content, the secretion of resistin and MCP-1 (n= 3 wells/group) is given in ng/ml/24 hr and pg/ml/24 hr, respectively. Box plots show median (thick black line), lower and upper quartiles (box) and lower and upper extreme values (thin vertical lines). Three doses (0·1, 1 and 5 μg per well) of a neutralizing TLR-4 antibody (°) were used to block TLR-4-induced signalling. (a) Effects of antibody-mediated TLR-4 blockade on the fatty acid-induced secretion of resistin. *Significant when compared to control; °Significant when compared to fatty acid stimulation without TLR-4 blockade. (b) Effects of antibody-mediated TLR-4 blockade on the fatty acid-induced effects on the secretion of MCP-1. *Significant when compared to control; °Significant when compared to fatty acid stimulation without TLR-4 blockade.

Figure 5(b) summarizes the effects of TLR-4 blockade on the effects of stearic acid, palmitic acid, and palmitoleic acid on the secretion of MCP-1 from adipocytes. The TLR-4 blockade significantly inhibited the stearic acid-induced stimulation of MCP-1 secretion, whereas it had no impact on the inhibitory effects of palmitoleic acid on MCP-1 secretion. Palmitic acid had no effect on MCP-1 secretion, either alone or after TLR-4 blockade.

TLR-4 is induced during adipocyte differentiation and after stimulation with fatty acids

Since TLR-4 was shown to mediate fatty acid-induced effects on adipokine secretion, we aimed to investigate the protein expression profile of TLR-4 during adipocyte differentiation and in response to fatty acids. As demonstrated by Western blot analysis, TLR-4 is almost absent in preadipocytes but is induced during adipocyte differentiation at day 4, day 7 and day 9 of differentiation (Fig. 6). Stearic acid (100 μm), palmitic acid (100 μm) and palmitoleic acid (1 μm) increased the amount of TLR-4 to levels comparable to those observed with monocytic THP-1 cells.

Figure 6.

Analysis of Toll-like receptor-4 (TLR-4) expression by Western blot using extracts prepared from 3T3-L1 adipocytes during differentiation and after stimulation with fatty acids. The TLR-4 is almost absent in preadipocytes and is induced during adipocyte differentiation. Fatty acids slightly increase the amount of TLR-4 with palmitoleic acid showing the most pronounced effect. After stimulation with palmitoleic acid, adipocytes have about equal amounts of TLR-4 compared to monocytic THP-1 cells, which were used as a positive control. Lane 1 = 3T3-L1 preadipocytes, lane 2 = adipocytes after 4 days of differentiation, lane 3 = adipocytes after 7 days of differentiation, lane 4 = mature adipocytes after 9 days of differentiation, lane 5 = mature adipocytes stimulated with 100 μm stearic acid (24 h), lane 6 = mature adipocytes stimulated with 100 μm palmitic acid (24 hr), lane 7 = mature adipocytes stimulated with 1 μm palmitoleic acid (24 hr), lane 8 = THP-1 cells (monocytes), lane 9 = negative control (phosphate-buffered saline).

Investigation of fatty acid binding to TLR-4/MD-2 using a FLAG-tagged extracellular part of TLR-4 fused to full-length MD-2 (LPS-Trap)

To determine whether fatty acids can bind directly to TLR-4, the binding of ¹⁴C-labelled stearic and oleic acids to a soluble fusion complex consisting of the FLAG-tagged extracellular part of TLR-4 fused to full-length MD-2 via a flexible linker (LPS-Trap) was measured. This protein was shown earlier by our group to bind and neutralize LPS.²⁰ The LPS-Trap was incubated with various concentrations of labelled fatty acids with and without an excess of unlabelled acids to determine non-specific binding. The complex was then immunoprecipitated with anti-FLAG antibody M2 and washed; bound radioactivity was then measured in a scintillation counter. As shown in Fig. 7, (bottom panel) no specific binding was detected. The complex of TLR-4 and MD-2 forms a large hydrophobic pocket which can accommodate acyl chains.²² We therefore reasoned that fatty acids could compete for this pocket with LPS. The LPS-Trap was therefore incubated with biotin-labelled LPS with and without a large excess of stearic or oleic acid. The LPS/LPS-Trap complex was then pulled down with streptavidin–agarose and Western-blotted, and the blot was probed with anti-FLAG antibody. Figure 7 (top panel) shows that the binding of LPS to LPS-Trap was not inhibited. We conclude from these two experiments that these fatty acids do not interact with TLR-4/MD-2 heterodimer.

Figure 7.

Investigation of fatty acid binding to Toll-like receptor-4 (TLR-4)/MD-2 by using a FLAG-tagged extracellular part of TLR-4 fused to full-length MD-2 [lipopolysaccharide (LPS)-Trap]. The binding of ¹⁴C-labelled stearic and oleic acids to a soluble fusion complex consisting of the FLAG-tagged extracellular part of TLR-4 fused to full-length MD-2 via a flexible linker (LPS-Trap) was investigated by immunoprecipitation (upper panel) and by fatty acid binding studies (lower panel). The LPS-Trap was incubated with an excess of C¹⁴-labelled fatty acids or with LPS alone. The complex was immunoprecipitated with anti-FLAG antibody M2. LPS-Trap was incubated with biotin-labelled LPS with and without a 100 × molar excess of stearic or oleic acid, a dose that should compete effectively with LPS if specific fatty acid binding occurs. The LPS/LPS-Trap complex was then pulled down with streptavidin–agarose, Western-blotted, and the blot was probed with anti-FLAG antibody. Upper panel: Neither stearic acid nor oleic acid can bind to the TLR-4/MD-2 heterodimer. Lower panel: There was no specific binding of C¹⁴-labelled fatty acids to the TLR-4/MD-2 complex. IP= immunoprecipitation, WB= Western blotting.

Serum levels and expression of adiponectin, resistin and MCP-1 in abdominal adipose tissue of rats fed a standard chow and a high-fat diet

To investigate the systemic effects of a high-fat diet on the adipose tissue expression of adiponectin, resistin and MCP-1, mRNA was prepared from abdominal adipose tissue samples of rats fed either a standard chow (n= 6) or a high-fat diet (n= 6) and used for quantitative LightCycler real-time PCR analysis. The detailed morphological and biochemical alterations commonly seen in these experimental animals were published previously.²¹ As summarized in Table 2, the high-fat diet did not influence the mRNA expression of adiponectin, resistin and MCP-1 when compared to animals fed a standard chow. Additionally, serum levels were measured and found to be similar in these animals (Table 2).

Table 2.

Serum levels and adipose tissue messenger RNA expression of adiponectin, resistin and monocyte chemoattractant protein-1 (MCP-1) of rats fed a standard chow (n= 6) or a high fat diet (n= 6)

	Standard chow (n= 6)	High-fat diet (n= 6)	P
Fat tissue adiponectin (adiponectin mRNA /β-actin mRNA; mean value ± SEM)	0·95 ± 0·1	0·88 ± 0·1	0·7
Fat tissue resistin (resistin mRNA/β-actin mRNA; mean value ± SEM)	1·75 ± 0·4	1·2 ± 0·2	0·3
Fat tissue MCP-1 (MCP-1 mRNA/β-actin mRNA; mean value ± SEM)	1·19 ± 0·1	1·77 ± 0·39	0·2
Serum adiponectin (ng/ml) (mean value ± SEM)	5833 ± 506	6450 ± 413	0·4
Serum MCP-1 (pg/ml) (mean value ± SEM)	24·2 ± 4·8	22·4 ± 4·1	0·7
Serum resistin (ng/ml) (mean value ± SEM)	28·7 ± 2·5	30·2 ± 2·9	0·7

Messenger RNA (mRNA) was prepared from abdominal adipose tissue and analysed for β-actin and adipokine mRNA expression using quantitative LightCycler real-time reverse transcription–polymerase chain reaction. The adipokine/β-actin ratio is given as mean value ± standard error of the mean (SEM). The serum levels of adiponectin, resistin and MCP-1 were measured by enzyme-linked immunosorbent assay.

Discussion

Table 1 summarizes the dose effects and class effects of fatty acids on the adipocytic secretion of adiponectin, resistin and MCP-1. Both dose and class of fatty acids were noted to influence adipokine and MCP-1 secretion differentially. A high-fat diet neither alters adipokine and MCP-1 mRNA expression in abdominal adipose tissue nor changes the respective serum levels of rodents, so systemic and local (paracrine) effects of fatty acids on adipocytes/adipose tissue have to be interpreted separately. The present data argue for a paracrine or autocrine role of fatty acids in adipocyte biology.

Although there are numerous studies²³ in rodents and humans investigating the effects of diet, exercise, and weight gain/weight loss on plasma adipokines, serum adiponectin levels or adipose tissue adiponectin gene expression, there are no studies available that systematically investigate the direct effects of different classes of fatty acids on the secretion of adiponectin from mature adipocytes. Intake of diets rich in eicosapentaenoic acid and docosahexaenoic acid increases plasma adiponectin levels in mice fed a high-fat diet.²⁴ One study reported on isomer-specific effects of conjugated linoleic acid on adiponectin secretion from 3T3-L1 adipocytes.²⁵ In this study, adiponectin secretion was increased by cis-9,trans-11-conjugated linoleic acid but not by linoleic acid or trans-10,cis-12-conjugated linoleic acid. However, another study reported a decrease of serum and adipose tissue adiponectin levels in response to dietary fish oil and conjugated linoleic acid in mice.²⁶ As far as adiponectin was concerned, all of the investigated fatty acids did increase adiponectin secretion in the present study. The mechanism behind this up-regulation remains unclear.

Recombinant resistin stimulates triglyceride lipolysis in human adipocytes and simultaneously accelerates triglyceride lipolysis and fatty acid re-esterification in mouse adipose tissue explants.²⁷ Based on this, if resistin accelerates the fatty acid/triglyceride futile cycle, fatty acids might directly influence adipocyte resistin secretion. Since a drug-induced reduction of plasma fatty acid concentration by acipimox does not alter serum resistin levels,²⁸ we decided to study the direct effects of fatty acids on adipocyte resistin secretion. Resistin secretion was found to be up-regulated by fatty acids in our study with pronounced effects mediated by palmitoleic and linoleic acids. Very high doses of palmitic acid, palmitoleic acid and linoleic acid were not effective in altering resistin secretion. A study by Juan et al.²⁹ on a rat model reported that diet-induced insulin resistance leads to reduced expression of the adipocyte resistin gene. Moreover, they found a direct and inhibitory effect of free fatty acids on resistin gene expression in rat adipocytes of animals fed normal chow.²⁹ In their study, the authors used oleic acid and a combination of oleic with linoleic acid in very high doses of 150 μm, which might exert toxic effects on cells. In the present study we used physiological concentrations of oleic and linoleic acid (1–10 μm) and could demonstrate an up-regulation of resistin secretion. Rea and Donnelly ³⁰ stimulated 3T3-L1 adipocytes with high doses of oleic acid (100 μm) and found a down-regulation of resistin mRNA expression. However, this high dose of oleic acid might be toxic to cells. In a physiological range of 1–10 μm we could clearly demonstrate an increased resistin secretion which is a more suitable and stable parameter than the mRNA expression level. From a pathophysiological point of view, an induction of the prodiabetic and proinflammatory resistin by fatty acids might in part explain the mechanism of fatty acid-induced systemic insulin resistance at least in mice.

In the context of obesity, adipose tissue becomes infiltrated by significant amounts of monocytes/macrophages.^1,2,10 This process leads to a chronic inflammation within the adipose tissue that might cause insulin resistance and an altered adipokine expression profile. Suganami et al.³¹ suggested the existence of a paracrine loop between adipocytes and macrophages that aggravates inflammatory processes. Coculture of differentiated 3T3-L1 adipocytes with RAW264 macrophages results in an up-regulation of macrophage tumour necrosis factor-α secretion and a down-regulation of adipocytic adiponectin secretion. In addition, this paracrine loop is also composed of adipocyte-derived fatty acids such as palmitate, because palmitate stimulates tumour necrosis factor-α secretion from macrophages. Proinflammatory chemokines such as MCP-1 are not only produced by monocytes/macrophages but also by adipocytes: Therefore, it was our aim to investigate whether mature 3T3-L1 adipocytes do secrete significant amounts of MCP-1 and whether this basal secretion of MCP-1 can be influenced by fatty acids.

Mature 3T3-L1 adipocytes do secrete stable and significant amounts of MCP-1 protein that can easily be detected by commercially available ELISA. These results suggest that not only adipose tissue-derived macrophages but also adipocytes themselves produce and secrete high amounts of MCP-1. This provides a potential pathophysiological explanation as to how monocytes are directed from the bloodstream into the adipose tissue, e.g. in the context of obesity or insulin resistance. Indeed, a recent study³² proved this hypothesis by generating mice overexpressing MCP-1 in adipose tissue using the adipocyte aP2 promoter (aP2-MCP-1 mice). These mice were insulin resistant and had a higher adipose tissue macrophage accumulation than control mice. Since insulin resistance is accompanied by elevated plasma free fatty acid concentrations, fatty acid-induced induction of MCP-1 secretion might explain adipose tissue macrophage accumulation in the context of the metabolic syndrome. However, there are class effects of fatty acids on adipocytic MCP-1 secretion. High doses of the saturated fatty acids stearic acid and palmitic acid strongly enhance MCP-1 secretion. In contrast, low doses of stearic acid and linoleic acid suppress MCP-1 secretion. Oleic acid has the strongest inhibitory effect on MCP-1 secretion. Further studies are needed to clarify the molecular mechanism of these dose and class effects. The physiological consequences of the observed dose effects remain unclear. However, in the physiological in vivo system, serum levels of fatty acids are changing dramatically dependent on the prandial state, the general state of energy supply and the diet. Therefore, dose-dependent effects of fatty acids are of physiological relevance.

The intracellular signal transduction cascade mediating these fatty acid-induced effects described here might be complex. However, two important mechanisms could be clarified in this study. Since NF-κB inhibition strongly antagonizes the proinflammatory effects of palmitic acid and stearic acid, and because these fatty acids enhance the nuclear translocation of NF-κB p65, the stimulatory effect of stearic acid and palmitic acid on the secretion of proinflammatory resistin and MCP-1 is clearly mediated by NF-κB. Most interestingly, the TLR-4 is expressed on adipocytes^33,34 and it was shown to be activated by fatty acids on monocytes^31,35 in a recent study. In the present study, the stimulatory effects of stearic, palmitic and palmitoleic acids on resistin secretion and the stimulatory effect of stearic acid on MCP-1 secretion are mediated via TLR-4, as demonstrated by specific TLR-4 blockade. Accordingly, the role of the adipose tissue as an immunological organ linking nutrition/metabolism with immune function has been suggested recently by our group.³⁶

Most importantly, we hypothesized that fatty acids such as oleic acid and stearic acid might interact directly with the heterodimeric TLR-4/MD-2 complex. As a complex, TLR-4 and MD-2 form a large hydrophobic pocket that can accommodate acyl chains.²² It is known that lauric acid, for example, is one of the lipid components of LPS and lauric acid (a medium chain fatty acid) can mediate inflammatory responses similar to that of LPS. However, as demonstrated by coimmunoprecipitation experiments and fatty acid binding studies using the LPS-trap, we could exclude direct binding of these fatty acids to the heterodimeric TR-4/MD-2 complex. Moreover, we could also exclude the possibility that fatty acids inhibit the binding of LPS to the TLR-4/MD-2 complex. The results suggest that fatty acid-induced effects are possibly mediated by an as yet unknown endogenous ligand that is released by adipocytes upon fatty acid stimulation.

Rats fed a high-fat diet over a relatively short period of 12 weeks neither showed a significant up-regulation of adipokine mRNA expression in total adipose tissue nor an increase in adipokine serum levels, so the observed effects of the fatty acids are the result of direct effects on adipocytes and suggest a paracrine or autocrine role for fatty acids in adipocyte biology. However, we cannot exclude the possibility that the disparity between in vitro and in vivo results is the result of specific culture conditions or might be a consequence of something unique about the 3T3-L1 cell line. Future work has to extend the key findings of these experiments on human adipocytes isolated from different adipose depots.

This is the first study to systematically investigate the effects of different classes of fatty acids on adipokine and MCP-1 secretion from differentiated adipocytes with respect to TLR-4 and NF-κB signalling.

1. Different classes of fatty acids are capable of inducing adiponectin secretion.

2. Fatty acids can induce resistin secretion with palmitoleic and linoleic acid showing the most pronounced effects.

3. Concerning MCP-1 secretion, fatty acids exert class-specific and differential effects. Palmitic acid and stearic acid effectively stimulate MCP-1 secretion by activating the NF-κB pathway.

4. The stimulatory effects of stearic, palmitic and palmitoleic acids on resistin secretion and the stimulatory effect of stearic acid on MCP-1 secretion are mediated via TLR-4. A direct binding of fatty acids to TLR-4/MD-2 could be excluded.

5. Since adipose tissue mRNA expression and serum levels of adipokines did not differ in rats fed a high-fat diet, the observed effects argue for an autocrine or paracrine role of fatty acids in causing local adipose tissue inflammation as a prerequisite for an adipose tissue-directed accumulation of macrophages.

6. The present data suggest a molecular mechanism by which fatty acids could link nutrition with innate immunity and future basic research and experimental animal studies are necessary to clarify the exact mechanisms.