Transmembrane tumor necrosis factor-alpha sensitizes adipocytes to insulin

Molecular and Cellular Endocrinology 406 (2015) 78–86

Contents lists available at ScienceDirect

Molecular and Cellular Endocrinology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / m c e

Transmembrane tumor necrosis factor-alpha sensitizes adipocytes to insulin Wenjing Zhou a,1, Peng Yang a,1, Li Liu a, Shan Zheng a, Qingling Zeng a,c, Huifang Liang a, Yazhen Zhu a, Zunyue Zhang a, Jing Wang a, Bingjiao Yin a, Feili Gong a, Yiping Wu b, Zhuoya Li a,* a

Department of Immunology, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430030, China Department of Plastic Surgery, Tongji Hospital, Huazhong University of Science and Technology, Wuhan 430030, China c Department of Hematology & Endocrinology, Fifth Hospital of Wuhan, Wuhan 430071, China b

A R T I C L E

I N F O

Article history: Received 10 June 2014 Received in revised form 7 February 2015 Accepted 22 February 2015 Available online 25 February 2015 Keywords: Transmembrane TNF-α Adipocyte Insulin resistance IL-6 Adiponectin

A B S T R A C T

Transmembrane TNF-α (tmTNF-α) acts both as a ligand, delivering ‘forward signaling’ via TNFR, and as a receptor, transducing ‘reverse signaling’. The contradiction of available data regarding the effect of tmTNF-α on insulin resistance may be due to imbalance in both signals. Here, we demonstrated that high glucoseinduced impairment of insulin-stimulated glucose uptake by 3T3-L1 adipocytes was concomitant with decreased tmTNF-α expression and increased soluble TNF-α (sTNF-α) secretion. However, when TACE was inhibited, preventing the conversion of tmTNF-α to sTNF-α, this insulin resistance was partially reversed, indicating a salutary role of tmTNF-α. Treatment of 3T3-L1 adipocytes with exogenous tmTNF-α promoted insulin-induced phosphorylation of IRS-1 and Akt, facilitated GLUT4 expression and membrane translocation, and increased glucose uptake while addition of sTNF-α resulted in the opposite effect. Furthermore, tmTNF-α downregulated the production of IL-6 and MCP-1 via NF-κB inactivation, as silencing of A20, an inhibitor for NF-κB, by siRNA, abolished this effect of tmTNF-α. However, tmTNF-α upregulated adiponectin expression through the PPAR-γ pathway, as inhibition of PPAR-γ by GW9662 abrogated both tmTNF-α-induced adiponectin transcription and glucose uptake. Our data suggest that tmTNF-α functions as an insulin sensitizer via forward signaling. © 2015 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Transmembrane TNF-α (tmTNF-α) is a type II transmembrane protein. Its extracellular segment is cleaved by membrane-bound metalloproteases, chiefly TNF-α-converting enzyme (TACE), releasing soluble TNF-α (sTNF-α). Both forms of TNF-α exert their biological functions via binding to TNF receptors (TNFRs) (Black et al., 1997). Although sTNF-α has been widely recognized as a link between adiposity and insulin resistance, a few studies have shown that tmTNF-α is also bioactive in adipocytes and is involved in obesityrelated insulin resistance. Xu et al. demonstrated that tmTNF-α inhibits adipocyte differentiation in vitro (Xu et al., 1999), and that its expression is significantly increased in the adipose tissue in different rodent obesity models as well as in obese humans (Xu et al., 2002b). Suppressing ectodomain shedding of tmTNF-α by

* Corresponding author. Department of Immunology, Tongji Medical College, Huazhong University of Science and Technology, 13 Hangkong Road, Wuhan, Hubei 430030, China. Tel.: +86 27 83692611; fax: +86 27 83693500. E-mail address: [email protected] (Z. Li). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.mce.2015.02.023 0303-7207/© 2015 Elsevier Ireland Ltd. All rights reserved.

the TACE inhibitor KB-R7785 shows an antidiabetic effect (Morimoto et al., 1997). Similarly, mice heterozygous for TACE (Tace+/−) resulting in increased expression of tmTNF-α were relatively protected from obesity and insulin resistance (Serino et al., 2007). These data suggest that tmTNF-α, unlike sTNF-α, may promote insulin sensitivity. However, it has also been reported that adipocyte specific expression of noncleavable tmTNF-α impaired local insulin sensitivity and decreased whole body adipose mass in a transgenic mouse model (Xu et al., 2002a). When this noncleavable tmTNF-α mutant was expressed in multiple different organs, it led to increased weight gain and adipose tissue mass of mice fed a high-fat diet (Voros et al., 2004). The discrepancy of the reported roles of tmTNF-α in insulin sensitivity may be associated with the balance of bidirectional signaling of tmTNF-α. tmTNF-α as a ligand delivers ‘forward signaling’ via TNFR to target cell or, as a receptor, transduces ‘reverse signaling’ to its expressing cell. Since both tmTNF-α and TNFR are substrates of TACE (Scheller et al., 2011), inhibition of TACE not only increases expression of tmTNF-α and TNFR on the cell surface, but also inhibits release of these soluble molecules. tmTNF-α binds TNFR and transduces bidirectional signals simultaneously via tmTNF-α and TNFR respectively. In noncleavable tmTNF-α transgenic mice, TNFR but

W. Zhou et al./Molecular and Cellular Endocrinology 406 (2015) 78–86

not tmTNF-α can still be cleaved by TACE that is upregulated in obesity (Xu et al., 2002b). These increased soluble TNFRs bind to tmTNF-α to deliver reverse signaling, meanwhile these soluble molecules can competitively decrease interaction of tmTNF-α and cellsurface TNFRs to disrupt forward signaling. If this were the case, the beneficial effect of TACE inhibitor would be attributed to the forward signaling of tmTNF-α. Furthermore, none of those experimental systems was sufficient to distinguish the actions of tmTNF-α via forward signaling from those via reverse signaling. Obesity is considered to be a chronic low-grade inflammatory state that results in insulin resistance (Emanuela et al., 2012). Because we previously showed that exogenous tmTNF-α inhibits NF-κB activation (a pathway associated with inflammation) in the neutrophil-like cell line HL-60 (Chen et al., 2011), we hypothesized that tmTNF-α may play a role in sensitizing insulinresponse via its forward signaling. In the present study, we directly treated 3T3-L1 and human primary adipocytes with exogenous tmTNF-α and found that tmTNF-α inhibited the production of proinflammatory adipokines and promoted the release of antiinflammatory adipokine, increasing insulin sensitivity of adipocytes via forward signaling. 2. Materials and methods 2.1. Preadipocyte isolation and adipogenic differentiation Human preadipocytes were isolated from the freshly excised abdominal subcutaneous adipose tissue of seven healthy women (aged 25–42 years, with BMI of 21.8 ± 2.6 kg/m2) after liposuction or abdominoplasties at the Department of Plastic Surgery, Tongji Hospital. This study was conducted in accordance with the guidelines of the local ethics committee. Adipose tissue was minced and digested by 1 mg/ml type II collagenase (Grand Island, NY) at 37 °C for 60 min under constant shaking. After filtration through a doublelayered sterile gauze, the filtrates were centrifuged at 1000 rpm for 10 min. After lysis of erythrocytes, the cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM)/F-12 with 10% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin and 1% Fungizone. Once confluence of human preadipocytes was reached, adipogenic differentiation was induced for 3 days in OriCellTM Adipogenic Differentiation Medium A, and then in matched Medium B (Cyagen, Goleta, CA) for another 24 h. This was repeated 3 times. The cells were then maintained in Medium B for an additional 7–9 days until accumulated visible lipid droplets emerged. 3T3-L1 murine fibroblast cell line was grown to confluence in DMEM supplemented with 10% FBS at 37°C. Adipogenic differentiation of 3T3-L1 was induced by exposure to 0.5 mM IBMX (3-isobutyl-1-methylxanthine, Sigma-Aldrich, St. Louis, MO), 1 μM dexamethasone, and 10 μg/ml insulin for 2 days, and then to 10 μg/ml insulin alone for additional 2 days. Thereafter, the cells were maintained in the medium containing 10% FBS until >90% of the cells showing accumulated lipid vacuoles in the cytoplasm by staining with Oil Red O. 2.2. Stimulation of adipocytes with both forms of TNF-α Fully differentiated 3T3-L1 adipocytes or primary human adipocytes as target cells were treated for 24 h with 20 ng/ml sTNF-α (Peprotech, Rocky Hill, NJ) or tmTNF-α expressed at a high level by Raji cells as effector cells (Zhang et al., 2008), a malignant B-cell line had been fixed with 4% paraformaldehyde for 30 min at room temperature (RT), at an effector/target (E/T) ratio of 10:1. The untreated adipocytes served as a control. For identification of the specific actions of tmTNF-α, 4% paraformaldehyde-fixed Raji cells were treated with an anti-TNF-α antibody (BD Pharmingen, San Jose, CA) for 1 h to neutralize tmTNF-α prior to addition to the target

79

adipocytes. For glucose uptake assay or test of insulin signaling, 100 nM of insulin was added to the cells and incubated for 30 min after TNF-α-stimulation. 2.3. Glucose uptake assay 3T3-L1 adipocytes or primary human adipocytes were treated for 24 h with different concentrations of glucose and 1 μM insulin, or the two forms of TNF-α. After re-equilibration in serum- and glucose-free media for 2 h, 1 × 106 adipocytes were incubated for 30 min with or without 100 nM of insulin in glucose-free Krebs– Ringer Hepes (KRH) buffer. Glucose uptake was detected by the addition of 2-[1,2-3H]-deoxy-D-glucose (0.5 μCi/ml) for 10 min, followed by measurement of the radioactivity of the cell lysates after solubilization in 0.1 M NaOH, by a liquid scintillation counter (Perkin Elmer). Specific uptake was obtained by subtracting nonspecific deoxyglucose uptake that was determined in the presence of 20 μM cytochalasin B, from each of the resultant values (Zhou et al., 2010). 2.4. Western blot analysis Adipocytes were harvested and lysed in ice-cold lysis buffer (20 mM Tris pH 7.5, 150 mM NaCl, 1% Triton X-100) and a protease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO). The total protein was obtained after centrifugation at 12,000 rpm for 20 min at 4 °C. For detection of GLUT4 translocation, 3T3-L1 adipocytes were stimulated with insulin for 30 min. The plasma membrane and cytoplasmic proteins were extracted and fractionated using the ProteoJETTM Membrane Protein Extraction kit (Fermentas, Shenzhen, China) according to the manufacturer’s recommended protocol. For determination of NF-κB p65 nuclear translocation, the cytosolic and nuclear proteins from cells were separated and isolated by NuclearCytosol Extraction Kit (Applygen Technologies Inc, Beijing, China) following the manufacturer’s instructions. Fifty micrograms of total, membranous, cytoplasmic or nuclear protein was electrophoresed on a polyarcrylamide gel and transferred to polyvinylidene difluoride membranes by electroblotting. The membranes were blocked for 2 h at RT with 5% skim milk in PBS containing 0.1% Tween-20 and then probed overnight at 4 °C with primary antibodies including anti-GLUT4 and anti-PPAR-γ (Millipore, Billerica, MA), anti-IκB-α, anti-p-Tyr-IRS-1, anti-IRS-1, antiAkt, anti-TNF-α, anti-Lamin B1, anti-caveolin-1 and anti-β-actin (Santa Cruz, CA), anti-p-Akt (Cell Signaling Technology, Beverly, MA) and anti-NF-κB p65 (Epitomics, Burlingame, CA), followed by corresponding horseradish peroxidase-conjugated anti-rabbit IgG or anti-mouse IgG secondary antibody (Pierce, Rockford, IL) at RT for 1 h. Immunoreactive bands were visualized using the enhanced chemiluminescence kit (Pierce, Rockford, IL) and the Kodak Image Station 4000 MM (East-man Kodak Co., Rochester, NY). 2.5. Flow cytometry After stimulation, 3T3-L1 adipocytes (2 × 105 cells/well) were incubated for 1 h at 4 °C with an antibody specific to murine TNF-α (Abcam, Cambridge, MA), followed by a 45 min-incubation at 4 °C with fluorescein isothiocyanate (FITC)-conjugated secondary antibody. Surface expression of tmTNF-α was analyzed on a FACS Calibur 440 E flow cytometer (Becton Dickinson, San Jose, CA). 2.6. ELISA for adipokines Commercial ELISA kits were used to detect sTNF-α, IL-6, MCP-1 (eBioscience, San Diego, CA) and adiponectin (LINCO, St. Charles, MO) in the supernatants of cultured 3T3-L1 adipocytes according to the manufacturers’ protocols.

80

W. Zhou et al./Molecular and Cellular Endocrinology 406 (2015) 78–86

Table 1 Primer sequences of mouse-specific gene for qPCR. Gene

Ref Seq ID

Reaction efficiency

Primer sequence

IL-6

NM_031168.1

95.23%

MCP-1

NM_011333.3

99.71%

Adiponectin

NM_009605.4

98.68%

A20

NM_001166402.1

96.72%

β-Actin

NM_007393.3

97.03%

Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse

2.7. Fatty acid assay The levels of free fatty acid (FFA) in the supernatant of 3T3-L1 adipocytes were detected by a Cu-colorimetric method using an ultrasensitive assay kit (Applygen, Beijing, China) according to the manufacturer’s recommended protocol. The absorbance was measured spectrophotometrically at 540 nm. 2.8. RNA interference and transfection Three siA20 oligonucleotides and a scrambled control siRNA were designed and synthesized by RiboBio (Guangzhou, China). The most efficient siA20 oligonucleotide (5′-CUUUGAAUGUGCAGCAUAA-3′ with 3′-dTdT) was chosen by screening with real-time qPCR and Western blot for A20 expression. On day 6 after induction of differentiation, 3T3-L1 adipocytes were transfected with 100 nM siRNA using LipofectamineTM 2000 (Invitrogen, Eugene, OR) reagent according to the manufacturer’s instructions. At 48 h after transfection, the cells were treated for 24 h with the two forms of TNF-α, and then harvested for RNA or protein extraction.

AGTTGCCTTCTTGGGACTGA CAGAATTGCCATTGCACAAC CCACTCACCTGCTGCTACTCAT TGGTGATCCTCTTGTAGCTCTCC TGTTCCTCTTAATCCTGCCCA CCAACCTGCACAAGTTCCCT GACCAAGTTGTCCCATTC TTCCTCAGGCTTTGTATTT CATCCGTAAAGACCTCTATGCCAAC ATGGAGCCACCGATCCACA

adipocytes in the presence of 15 mM glucose, 25 mM glucose markedly decreased the response of adipocytes to insulin with a significant reduction of IRS-1 expression (Fig. 1A). Meanwhile tmTNF-α expressed by adipocytes was almost completely proteolytically processed into sTNF-α in the presence of 25 mM glucose. In contrast, in the presence of 15 mM glucose 3T3-L1 adipocytes expressed tmTNF-α at high level but secreted sTNF-α at basal level (Fig. 1B and D). Although total 26 kD tmTNF-α production in cell lysate was stimulated by both 15 mM and 25 mM glucose (Fig. 1C), the evidence that only elevated release of sTNF-α from adipocytes was observed in 25 mM glucose (Fig. 1D) strongly indicated an obvious activation of TACE in this condition. When we used a specific antibody to neutralize TACE, a proteinase responsible for the cleavage of tmTNF-α into sTNF-α, 25 mM glucose-induced proteolytic processing of tmTNF-α was significantly suppressed, resulting in increased tmTNF-α expression and decreased sTNF-α release (Fig. 1E and F). The high glucose-induced insulin resistance could be partially reversed by TACE antibody, as manifested by the enhancement of cellular glucose uptake in response to insulin (Fig. 1G). These results point out a different action between tmTNF-α and sTNF-α in insulin resistance.

2.9. Real-time quantitative PCR analysis Total RNA was extracted from 3T3-L1 adipocytes using TRIzol reagents (Invitrogen). Two micrograms of the total RNA was reversely transcribed into first-strand cDNA using the TransScript FirstStrand cDNA Synthesis SuperMix (TransGen, Beijing, China). cDNA was amplified with gene-specific forward and reverse primers (Table 1) in a volume of 20 μl UltraSYBR Mixture (with ROX) (Cowin Biotech, Beijing, China) using an Mx3000P Real-Time PCR System (Stratagene). The reactions were performed in triplicate for 5 min at 95 °C, followed by 15 s at 95 °C, 20 s at 58 °C and 20 s at 72 °C for 40 cycles. Results were analyzed with Stratagene Mx3000 software using the 2−ΔΔCt method and normalized with β-actin. 2.10. Statistics Statistical analyses were performed by one-way or two-way ANOVA. Data are presented as mean ± SD. Differences with P values <0.05 were considered statistically significant. 3. Results 3.1. tmTNF-α expression is decreased by high glucose-induced insulin resistance To determine the relationship of tmTNF-α expression and insulin resistance in adipocytes, we treated fully differentiated 3T3-L1 adipocytes for 24 h with different concentrations of glucose and 1 μM insulin (Ling et al., 2012), and detected the sensitivity of the cells to insulin and the production of both forms of TNF-α. We found that, in contrast to maximal insulin-stimulated glucose uptake by 3T3-L1

3.2. Exogenous tmTNF-α increases insulin sensitivity by directly improving insulin signaling and enhancing the expression and membrane translocation of GLUT4 To further test whether tmTNF-α increased insulin sensitivity via forward signaling, exogenous tmTNF-α, as well as sTNF-α, was used to stimulate 3T3-L1 adipocytes for 24 h. Insulin alone stimulated cellular glucose uptake by 3.7-fold, as compared with spontaneous uptake (basal), however, this insulin-induced uptake was significantly suppressed by sTNF-α, but markedly enhanced by tmTNF-α (Fig. 2A), although the basal values were not affected by either sTNF-α or tmTNF-α (data not shown). The effect of tmTNF-α could be almost completely blocked by a specific neutralizing antibody. In addition, similar opposite action between the both forms of TNF-α on insulin sensitivity was also observed in primary human adipocytes (Fig. 2B). We then checked the effect of tmTNF-α on the insulin signaling pathway of 3T3-L1 adipocytes. In contrast to the inhibitory effect of sTNF-α, tmTNF-α evidently increased insulin-induced tyrosine phosphorylation of IRS-1 (Fig. 2C) and its downstream molecule Akt phosphorylation (Fig. 2D), but IRS-1 production remained unchanged (Fig. 2C). As the production and translocation of GLUT4 from the cytoplasm to the plasma membrane are directly correlated with glucose uptake, we observed the effect of both forms of TNF-α on this glucose transporter by Western blot. tmTNF-α, in contrast to sTNF-α, significantly promoted GLUT4 production (Fig. 2E). Furthermore, the insulin-induced redistribution of GLUT4, namely translocation of GLUT4 from the cytoplasm to the plasma membrane (by 3-fold), was precluded by sTNF-α, but significantly facilitated by tmTNF-α (Fig. 2F). All of these were consistent with the improvement of insulin-induced glucose uptake by tmTNF-α.

W. Zhou et al./Molecular and Cellular Endocrinology 406 (2015) 78–86

81

Fig. 1. tmTNF-α expression is decreased by high glucose-induced insulin resistance. Fully differentiated 3T3-L1 adipocytes were incubated with different concentrations of glucose and 1 μM insulin in the absence (A–D) or presence (E–G) of anti-TACE antibody (160 ng/ml) for 24 h. (A and G) To detect insulin response, cells were re-equilibrated in serum- and glucose-free media for 2 h and [3H]-2-deoxyglucose uptake was assessed after insulin (100 nM) stimulation for 30 min. Spontaneous cellular glucose uptake (basal) was measured in insulin-untreated cells. IRS-1 production (A, upper) was tested by Western blot. (B and E) tmTNF-α expression on the surface of the adipocytes was detected by flow cytometry. FACS image is representative (upper) and the quantitative analysis of three independent experiments is shown in histogram (lower). (C) 26 kD tmTNF-α production in cell lysate was detected by Western blot. (D and F) sTNF-α in the culture supernatants of 3T3-L1 adipocytes was measured by ELISA. Data represent mean ± SD of at least three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 vs. 5 mM glucose (A, B and D) or vs. 15 mM glucose (E–G); #p < 0.05, ##p < 0.01, ###p < 0.001 vs. 15 mM glucose (B and D) or vs. 25 mM glucose alone (E–G).

3.3. Exogenous tmTNF-α downregulates insulin-resisting adipokines, but upregulates insulin-sensitizing adipokine without affecting lipolysis Since the proinflammatory adipokines IL-6 and MCP-1 promote insulin resistance (Kanda et al., 2006; Rotter et al., 2003), while the anti-inflammatory adipokine, such as adiponectin, increases insulin sensitivity (Kadowaki et al., 2006), we hypothesized that tmTNF-α

may affect insulin sensitivity by regulating the production of these adipokines. As expected, tmTNF-α led to downregulation of IL-6 (Fig. 3A and B) and MCP-1 (Fig. 3C and D), but upregulation of adiponectin (Fig. 3E and F), at both mRNA and protein levels in adipocytes. These regulatory effects of tmTNF-α could be totally blocked by anti-TNF-α antibody. In contrast to tmTNF-α, sTNF-α had an opposite effect on these insulin-resisting and -sensitizing molecules.

82

W. Zhou et al./Molecular and Cellular Endocrinology 406 (2015) 78–86

Fig. 2. tmTNF-α-induced increase in insulin sensitivity and insulin signaling. Fully differentiated adipocytes were treated for 24 h with sTNF-α (20 ng/ml) or tmTNF-α on the surface of fixed Raji cells at an E/T ratio of 10:1. [3H]-2-Deoxyglucose uptake by 3T3-L1 (A) or primary human adipocytes (B) was determined after insulin (100 nM) stimulation for 30 min. Spontaneous cellular glucose uptake was detected in insulin-untreated adipocytes. For neutralization of tmTNF-α, fixed Raji cells were treated for 1 h with a TNF-α specific antibody prior to addition to adipocytes. (C–F) After a 24 h-stimulation with the both forms of TNF-α, followed by a 30 min-incubation with 100 nM insulin, total, cytoplasmic or membraneous protein was isolated from 3T3-L1 adipocytes. Immunoblot analysis of total protein was performed with anti-IRS-1, anti-pTyr-IRS-1, anti-pAkt, anti-Akt and anti-β-actin antibodies on the same membrane (C and D). The expression (E) and translocation (F) of GLUT4 were detected in total, cytoplasmic or membraneous protein by immunoblot analysis with anti-GLUT4, anti-β-actin and anti-caveolin-1 antibodies respectively. Densitometric analysis of immunoblots from (C) to (F) was shown in corresponding histograms. Data represent mean ± SD of at least three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 vs. insulin control.

It has been well known that sTNF-α increases adipocyte lipolysis and promotes release of free fatty acid (FFA) which is also implicated in the etiology of obesity-induced insulin resistance (Dresner et al., 1999; Green et al., 1994; Hauner et al., 1995). We tested whether tmTNF-α was involved in lipolysis using an ultrasensitive assay kit. In contrast to marked promotion of FFA release by sTNF-α, tmTNF-α did not affect lipolysis, as the release of FFA remained unchanged (Fig. 3G). The results suggest that tmTNF-α, unlike sTNF-α, failed to influence adipocyte lipolysis. 3.4. Exogenous tmTNF-α decreases production of pro-inflammatory adipokines through inhibition of NF-κB activation As NF-κB pathway is considered as a link between inflammation and insulin resistance (Shoelson et al., 2003), and IL-6 and MCP-1 are NF-κB target genes (Shoelson et al., 2007), we examined whether tmTNF-α has an effect on NF-κB activation in adipocytes and thus affects the production of these proinflammatory adipokines. We found that tmTNF-α suppressed NF-κB activation in 3T3-L1

adipocytes, showing an apparent decrease in the degradation of IκB-α (Fig. 4A) and in the nuclear translocation of NF-κB p65 from the cytoplasm (Fig. 4B). Conversely, sTNF-α induced NF-κB activation, displaying promotion of IκB-α degradation (Fig. 4A) and nuclear accumulation of NF-κB p65 (Fig. 4B). To further explore the molecular mechanism by which tmTNF-α suppressed NF-κB activation, we tested whether tmTNF-α has an effect on A20, an inhibitor for NF-κB activation (Lee et al., 2000). We found that tmTNF-α, like sTNF-α (by 2.3-fold), enhanced A20 transcription by 1.8-fold (Fig. 4C). Conversely, when A20 gene expression was downregulated using a specific siRNA (Fig. 4D), the positive effect of the both forms of TNF-α on A20 could be totally abolished (Fig. 4E). Interestingly, this abolishment completely blocked tmTNF-α-induced cytosolic accumulation of IκB-α, indicating that NF-κB activity was recovered (Fig. 4F). As a result, the transcription of two NF-κB target genes IL-6 (Fig. 4G) and MCP-1 (Fig. 4H) was also recovered from tmTNF-α-induced inhibition and restored to the basal level. These data suggest that the inhibitory effect of tmTNF-α on the production of proinflammatory adipokines IL-6

W. Zhou et al./Molecular and Cellular Endocrinology 406 (2015) 78–86

83

Fig. 3. tmTNF-α decreases production of IL-6 and MCP-1, but increases expression of adiponectin without affecting the lipolysis. 3T3-L1 adipocytes were treated with or without sTNF-α (20 ng/ml) or tmTNF-α on fixed Raji cells at an E/T ratio of 10:1 for 24 h. For neutralization of tmTNF-α, fixed Raji cells were treated for 1 h with a TNF-α specific antibody prior to addition to adipocytes. (A, C and E) Total RNA was isolated and reversely transcribed into cDNA. mRNA levels of IL-6, MCP-1 and adiponectin were determined by real-time quantitative PCR and normalized to β-actin. (B, D and F) Release of IL-6, MCP-1 and adiponectin into the culture supernatants was detected in duplicate by ELISA. (G) FFA in the supernatants of adipocytes was tested by an ultrasensitive assay kit. All data represent mean ± SD of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 vs. control.

and MCP-1 is attributed to suppressing NF-κB activation and that tmTNF-α-induced A20 expression is likely to be involved in this suppressive effect. 3.5. Exogenous tmTNF-α increases insulin sensitivity by PPAR-γ pathway The peroxisome proliferator-activated receptor γ (PPAR-γ) has been reported to be a master regulator for adiponectin expression (Iwaki et al., 2003), therefore, we detected the effect of tmTNF-α on the expression of PPAR-γ. As shown in Fig. 5A, tmTNF-α increased the expression of PPAR-γ by 2.3-fold. In addition, tmTNFα-upregulated gene expression of adiponectin was blocked by the addition of GW9662, an irreversible antagonist of PPAR-γ, although GW9662 alone failed to affect the basal level of adiponectin transcription (Fig. 5B). As a result, the insulin-sensitizing activity of tmTNF-α was abolished by GW9662, manifested as failure to promote Akt phosphorylation (Fig. 5C) and to enhance insulinstimulated glucose uptake (Fig. 5D). These results strongly suggest that the improvement of insulin sensitivity by tmTNF-α is closely associated with its induction of adiponectin via PPAR-γ pathway. 4. Discussion Our results show that high glucose-induced insulin resistance was accompanied by increased tmTNF-α processing, while inhibition of this ectodomain shedding by neutralization of TACE partially reversed high glucose-induced inhibition of insulin-stimulated glucose uptake by 3T3-L1 adipocytes. This is consistent with the

results from TACE inhibition (Morimoto et al., 1997) or haploinsufficiency (Serino et al., 2007), although the inducers for insulin resistance vary among these experimental systems. High glucose has been shown to induce release of sTNF-α (Esposito et al., 2002; Guha et al., 2000) and impair insulin signaling in adipocytes when combined with high level of insulin (Buren et al., 2003; Ling et al., 2012; Renstrom et al., 2007). It is believed that the improvement of insulin sensitivity by TACE inhibition is mainly attributed to failure of sTNF-α release due to suppressed tmTNF-α processing. The increased tmTNF-α would be involved in the improvement of insulin sensitivity. However, tmTNF-α binds membrane-bound TNFR and delivers forward signaling via TNFR and reverse signaling via itself simultaneously. Our previous studies demonstrated that reinforced expression of tmTNF-α that lacks extracellular domain, and thus only transduces reverse signaling, renders tumor cells resistant to apoptosis, while ectopic expression of tmTNF-α that lacks intracellular domain, and thereby only mediates forward signaling, confers sensitivity to apoptosis on tumor cells (Yan et al., 2009). It is possible that tmTNF-α may exert different bioactivities toward adipocytes via forward and reverse signaling. Because our previous study showed that NF-κB, a pathway involved in insulin resistance, can be inhibited by forward signaling but activated by reverse signaling of tmTNF-α (Zhang et al., 2008), we presumed that tmTNF-α may play a role in sensitizing adipocytes to insulin via forward signaling. Therefore, we observed the influence of exogenous tmTNF-α as a ligand on insulin sensitivity of adipocytes. Indeed, our results for the first time demonstrated that tmTNF-α significantly increased insulin-stimulated glucose uptake by both 3T3-L1 and primary adipocytes. In opposition to sTNF-α, tmTNF-α

84

W. Zhou et al./Molecular and Cellular Endocrinology 406 (2015) 78–86

Fig. 4. The inhibitory effect of tmTNF-α on IL-6 and MCP-1 takes place via the suppression of NF-κB activation. 3T3-L1 adipocytes were treated with or without sTNF-α (20 ng/ml) or tmTNF-α on fixed Raji cells at an E/T ratio of 10:1 for 24 h. For neutralization of tmTNF-α, fixed Raji cells were treated for 1 h with a TNF-α specific antibody prior to addition to adipocytes. Total, cytosolic or nuclear protein was isolated. IκB-α level in total protein (A) and the translocation of NF-κB p65 from cytoplasm to nucleus (B) were detected by immunoblot analysis. β-Actin or Lamin-B1 was used as a loading control for cytoplasmic or nuclear protein, respectively. (C and D lower) Total RNA was isolated and reversely transcribed into cDNA. A20 mRNA was determined by real-time qPCR. (D–H) 3T3-L1 adipocytes were lipofected for 48 h with either 100 nM scrambled or A20 siRNA on day 6 of differentiation induction, (E–H) followed by a 24 h-treatment with both forms of TNF-α as described earlier. Levels of protein A20 (D upper) and IκB-α (F) affected by siA20 were analyzed by Western blot. (A, B, and F) Immunoblots (upper) were representative, and densitometric data (lower) for the relative levels of these proteins represent means ± SD of at least three independent experiments. The effect of A20 siRNA on transcripts of A20 (E), IL-6 (G) and MCP-1 (H) in adipocytes in response to both forms of TNF-α were determined by real-time qPCR and normalized to β-actin. Data represent mean ± SD of three independent experiments. *p < 0.05, **p < 0.01. ***p < 0.001 vs. control (A–C) or vs. scrambled siRNA alone (D–H); ▲p < 0.05, ▲▲p < 0.01, ▲▲▲p < 0.001 vs. siA20 alone (E–H); ##p < 0.01, ###p < 0.001 vs. corresponding scrambled siRNA in a pair of group (E–H).

directly promoted insulin-induced insulin receptor substrate-1 (IRS-1) tyrosine phosphorylation and Akt phosphorylation, GLUT4 expression and its membrane translocation. Since insulin-stimulated phosphatidylinositol 3-kinase (PI3K) pathway is mainly responsible for the GLUT4 redistribution (Hausdorff et al., 1999), it is likely that exogenous tmTNF-α is bound to TNFRs on adipocytes, and

enhances insulin-stimulated IRS-1/Akt pathway, thus promoting GLUT4 membrane translocation and glucose uptake of adipocytes. Another molecular mechanism for insulin-sensitizing activity of tmTNF-α was downregulation of proinflammatory adipokines IL-6 and MCP-1, a key factor for recruiting macrophages, at both mRNA and protein levels in 3T3-L1 adipocytes. IL-6 and macrophage

W. Zhou et al./Molecular and Cellular Endocrinology 406 (2015) 78–86

85

Fig. 5. tmTNF-α increased insulin sensitivity through upregulation of adiponectin via PPAR-γ. 3T3-L1 adipocytes were treated with or without sTNF-α (20 ng/ml) or tmTNF-α on fixed Raji cells at an E/T ratio of 10:1 for 24 h. For neutralization of tmTNF-α, fixed Raji cells were treated for 1 h with a TNF-α specific antibody prior to addition to adipocytes. 3T3L1 adipocytes were incubated with GW9662 (10 μM) for 12 h prior to stimulation with tmTNF-α on fixed Raji cells (at an E/T ratio of 10:1) for 24 h (B–D). (A) PPAR-γ was tested in total cellular protein by immunoblot analysis. Upper panels show representative immunoblots, and lower histograms show densitometric analysis of PPAR-γ levels normalized to β-actin from three independent experiments. (B) After stimulation of 3T3-L1 adipocytes as earlier, total RNA was isolated and reversely transcribed into cDNA. Adiponectin mRNA was determined by real-time qPCR and normalized to β-actin. (C) Phosphorylation of Akt was analyzed by Western blot. (D) [3H]-2-Deoxyglucose uptake was assessed following stimulation of adipocytes for 30 min with or without (basal) 100 nM insulin. All quantitative data represent mean ± SD of at least three independent experiments. *p < 0.05, ***p < 0.001 vs. control (A and B) or vs. insulin control (D); ##p < 0.01, ###p < 0.001 vs. tmTNF-α alone (B) or vs. insulin + tmTNF-α (D).

infiltration into adipose tissue are involved in obesity-related inflammation and development of insulin resistance (Weisberg et al., 2003; Xu et al., 2003). In line with our results, Matthias Canault et al. reported that exclusive expression of tmTNF-α in mice reduces infiltration of macrophages and inflammatory response in early lipid lesions of the aortic sinus (Canault et al., 2004). As these two adipokines are NF-κB-targeted genes (Shoelson et al., 2007), we assumed that the inhibitory function of tmTNF-α might be related to NF-κB pathway. Our hypothesis was supported by the following observations: (i) Treatment with tmTNF-α inhibited the degradation of IκB and nuclear translocation of NF-κB p65 in 3T3L1 adipocytes, which is consistent with our previous studies (Xin et al., 2006; Zhang et al., 2008). (ii) Although both forms of TNF-α could promote transcription of A20, silencing of A20 totally blocked the inhibitory effect of tmTNF-α, but enhanced the activatory effect of sTNF-α on IκB degradation and the both target genes’ transcription. A20, an ubiquitin-modifying enzyme, functions in interfering with sTNF-α-mediated signaling to NF-κB (Beyaert et al., 2000). A20deficiency induces persistent IKK activity and subsequent IκB-α degradation in embryonic fibroblasts stimulated with sTNF-α (Lee et al., 2000). Since A20 is itself a target of NF-κB, sTNF-α-induced NF-κB activation promotes A20 induction, which in turn restricts sTNF-α-mediated NF-κB activation. It is not surprising that silencing of A20 enhanced sTNF-α signaling to NF-κB. Conversely, tmTNF-α is likely to suppress NF-κB by promoting A20 expression, and thus inhibiting proinflammatory adipokines and affecting insulin sensitivity. Furthermore, in contrast to sTNF-α-induced release of FFA, an insulin-resistant molecule (Boden et al., 1994), tmTNF-α failed to influence lipolysis in adipocytes. This may be of benefit to obesityrelated insulin resistance. Adiponectin is one of the principal insulin-sensitizing adipokines (Kadowaki et al., 2006). Our results showed that tmTNF-α, unlike sTNF-α, significantly increased this sensitizer production via PPARγ. Consistently, Xu et al. demonstrated a 2-fold increase in the serum levels of adiponectin in the adipose tissue-restricted tmTNF-α transgenic mice, with some demonstrating improved systemic insulin sensitivity (Xu et al., 2002a). It is possible that noncleavable tmTNF-α expressed by visceral fat cells may exert bioactivity toward proximal tissue via forward signaling to affect systemic insulin sensitivity. Furthermore, our evidence demonstrated that tmTNF-α not only increased PPAR-γ expression, but also enhanced its activity on adiponectin transcription and insulin sensitivity, because treatment

with the specific PPAR-γ antagonist GW9662 blocked tmTNF-αmediated upregulation of adiponectin and insulin-stimulated glucose uptake. Since PPAR-γ binds to NF-κB p65, and promotes export of p65 from the nucleus and thereby prevents NF-κB activation (Kelly et al., 2004), this may also contribute to tmTNF-α-induced inhibition of proinflammatory adipokine production via suppressing NFκB, accordingly increasing insulin sensitivity of adipocytes. Our data indicate a beneficial regulatory effect of tmTNF-α via forward signaling on insulin sensitivity of adipocytes. This is an important reason to explain why adipocytes that mainly expressed tmTNF-α could maintain insulin response in the presence of 15 mM glucose. However, this membrane cytokine can also act as a receptor instead of ligand, delivering reverse signaling (Mitoma et al., 2005; Waetzig et al., 2005) and may exert bioactivities in adipocytes differently. To further clarify the effect of tmTNF-α-mediated bidirectional signaling on insulin response may be helpful in identifying the correct target for drugs designed to treat obesity, type 2 diabetes and insulin resistance-related diseases. 5. Conclusions Our findings demonstrate that tmTNF-α sensitizes adipocytes to insulin by improving insulin signaling, downregulating proinflammatory adipokine secretion as a result of NF-κB inactivation, and upregulating adiponectin expression owing to activation of the PPAR-γ pathway, indicating that tmTNF-α acts as an insulin sensitizer via its forward signaling. Acknowledgments This work was supported by the National Natural Science Foundation of China (30971397), National Program on Key Basic Research Project (973 Program, 2013CB530505) and National Science Foundation for Young Scholars of China (81202300). We thank Dr. Kentner Singleton for English editing. References Beyaert, R., Heyninck, K., Van Huffel, S., 2000. A20 and A20-binding proteins as cellular inhibitors of nuclear factor-kappa B-dependent gene expression and apoptosis. Biochem. Pharmacol. 60, 1143–1151.

86

W. Zhou et al./Molecular and Cellular Endocrinology 406 (2015) 78–86

Black, R.A., Rauch, C.T., Kozlosky, C.J., Peschon, J.J., Slack, J.L., Wolfson, M.F., et al., 1997. A metalloproteinase disintegrin that releases tumour-necrosis factor-alpha from cells. Nature 385, 729–733. Boden, G., Chen, X., Ruiz, J., White, J.V., Rossetti, L., 1994. Mechanisms of fatty acid-induced inhibition of glucose uptake. J. Clin. Invest. 93, 2438–2446. Buren, J., Liu, H.X., Lauritz, J., Eriksson, J.W., 2003. High glucose and insulin in combination cause insulin receptor substrate-1 and -2 depletion and protein kinase B desensitisation in primary cultured rat adipocytes: possible implications for insulin resistance in type 2 diabetes. Eur. J. Endocrinol. 148, 157–167. Canault, M., Peiretti, F., Mueller, C., Kopp, F., Morange, P., Rihs, S., et al., 2004. Exclusive expression of transmembrane TNF-alpha in mice reduces the inflammatory response in early lipid lesions of aortic sinus. Atherosclerosis 172, 211–218. Chen, H., Xiao, L., Zhang, H., Liu, N., Liu, T., Liu, L., et al., 2011. The involvement of beta-actin in the signaling of transmembrane TNF-alpha-mediated cytotoxicity. J. Leukoc. Biol. 89, 917–926. Dresner, A., Laurent, D., Marcucci, M., Griffin, M.E., Dufour, S., Cline, G.W., et al., 1999. Effects of free fatty acids on glucose transport and IRS-1-associated phosphatidylinositol 3-kinase activity. J. Clin. Invest. 103, 253–259. Emanuela, F., Grazia, M., Marco de, R., Maria Paola, L., Giorgio, F., Marco, B., 2012. Inflammation as a link between obesity and metabolic syndrome. J. Nutr. Metab. 2012, 476380. Esposito, K., Nappo, F., Marfella, R., Giugliano, G., Giugliano, F., Ciotola, M., et al., 2002. Inflammatory cytokine concentrations are acutely increased by hyperglycemia in humans: role of oxidative stress. Circulation 106, 2067–2072. Green, A., Dobias, S.B., Walters, D.J., Brasier, A.R., 1994. Tumor necrosis factor increases the rate of lipolysis in primary cultures of adipocytes without altering levels of hormone-sensitive lipase. Endocrinology 134, 2581–2588. Guha, M., Bai, W., Nadler, J.L., Natarajan, R., 2000. Molecular mechanisms of tumor necrosis factor alpha gene expression in monocytic cells via hyperglycemiainduced oxidant stress-dependent and -independent pathways. J. Biol. Chem. 275, 17728–17739. Hauner, H., Petruschke, T., Russ, M., Rohrig, K., Eckel, J., 1995. Effects of tumour necrosis factor alpha (TNF alpha) on glucose transport and lipid metabolism of newlydifferentiated human fat cells in cell culture. Diabetologia 38, 764–771. Hausdorff, S.F., Fingar, D.C., Morioka, K., Garza, L.A., Whiteman, E.L., Summers, S.A., et al., 1999. Identification of Wortmannin-sensitive targets in 3T3-L1 adipocytes. Dissociation of insulin-stimulated glucose uptake and glut4 translocation. J. Biol. Chem. 274, 24677–24684. Iwaki, M., Matsuda, M., Maeda, N., Funahashi, T., Matsuzawa, Y., Makishima, M., et al., 2003. Induction of adiponectin, a fat-derived antidiabetic and antiatherogenic factor, by nuclear receptors. Diabetes 52, 1655–1663. Kadowaki, T., Yamauchi, T., Kubota, N., Hara, K., Ueki, K., Tobe, K., 2006. Adiponectin and adiponectin receptors in insulin resistance, diabetes, and the metabolic syndrome. J. Clin. Invest. 116, 1784–1792. Kanda, H., Tateya, S., Tamori, Y., Kotani, K., Hiasa, K., Kitazawa, R., et al., 2006. MCP-1 contributes to macrophage infiltration into adipose tissue, insulin resistance, and hepatic steatosis in obesity. J. Clin. Invest. 116, 1494–1505. Kelly, D., Campbell, J.I., King, T.P., Grant, G., Jansson, E.A., Coutts, A.G., et al., 2004. Commensal anaerobic gut bacteria attenuate inflammation by regulating nuclear-cytoplasmic shuttling of PPAR-gamma and RelA. Nat. Immunol. 5, 104–112. Lee, E.G., Boone, D.L., Chai, S., Libby, S.L., Chien, M., Lodolce, J.P., et al., 2000. Failure to regulate TNF-induced NF-kappaB and cell death responses in A20-deficient mice. Science 289, 2350–2354. Ling, H.Y., Hu, B., Hu, X.B., Zhong, J., Feng, S.D., Qin, L., et al., 2012. miRNA-21 reverses high glucose and high insulin induced insulin resistance in 3T3-L1 adipocytes through targeting phosphatase and tensin homologue. Exp. Clin. Endocrinol. Diabetes 120, 553–559. Mitoma, H., Horiuchi, T., Hatta, N., Tsukamoto, H., Harashima, S., Kikuchi, Y., et al., 2005. Infliximab induces potent anti-inflammatory responses by outside-to-inside signals through transmembrane TNF-alpha. Gastroenterology 128, 376–392.

Morimoto, Y., Nishikawa, K., Ohashi, M., 1997. KB-R7785, a novel matrix metalloproteinase inhibitor, exerts its antidiabetic effect by inhibiting tumor necrosis factor-alpha production. Life Sci. 61, 795–803. Renstrom, F., Buren, J., Svensson, M., Eriksson, J.W., 2007. Insulin resistance induced by high glucose and high insulin precedes insulin receptor substrate 1 protein depletion in human adipocytes. Metabolism 56, 190–198. Rotter, V., Nagaev, I., Smith, U., 2003. Interleukin-6 (IL-6) induces insulin resistance in 3T3-L1 adipocytes and is, like IL-8 and tumor necrosis factor-alpha, overexpressed in human fat cells from insulin-resistant subjects. J. Biol. Chem. 278, 45777–45784. Scheller, J., Chalaris, A., Garbers, C., Rose-John, S., 2011. ADAM17: a molecular switch to control inflammation and tissue regeneration. Trends Immunol. 32, 380–387. Serino, M., Menghini, R., Fiorentino, L., Amoruso, R., Mauriello, A., Lauro, D., et al., 2007. Mice heterozygous for tumor necrosis factor-alpha converting enzyme are protected from obesity-induced insulin resistance and diabetes. Diabetes 56, 2541–2546. Shoelson, S.E., Lee, J., Yuan, M., 2003. Inflammation and the IKK beta/I kappa B/NF-kappa B axis in obesity- and diet-induced insulin resistance. Int. J. Obes. Relat. Metab. Disord. 27 (Suppl. 3), S49–S52. Shoelson, S.E., Herrero, L., Naaz, A., 2007. Obesity, inflammation, and insulin resistance. Gastroenterology 132, 2169–2180. Voros, G., Maquoi, E., Collen, D., Lijnen, H.R., 2004. Influence of membrane-bound tumor necrosis factor (TNF)-alpha on obesity and glucose metabolism. J. Thromb. Haemost. 2, 507–513. Waetzig, G.H., Rosenstiel, P., Arlt, A., Till, A., Brautigam, K., Schafer, H., et al., 2005. Soluble tumor necrosis factor (TNF) receptor-1 induces apoptosis via reverse TNF signaling and autocrine transforming growth factor-beta1. FASEB J. 19, 91–93. Weisberg, S.P., McCann, D., Desai, M., Rosenbaum, M., Leibel, R.L., Ferrante, A.W., Jr., 2003. Obesity is associated with macrophage accumulation in adipose tissue. J. Clin. Invest. 112, 1796–1808. Xin, L., Wang, J., Zhang, H., Shi, W., Yu, M., Li, Q., et al., 2006. Dual regulation of soluble tumor necrosis factor-alpha induced activation of human monocytic cells via modulating transmembrane TNF-alpha-mediated ‘reverse signaling. Int. J. Mol. Med. 18, 885–892. Xu, H., Sethi, J.K., Hotamisligil, G.S., 1999. Transmembrane tumor necrosis factor (TNF)-alpha inhibits adipocyte differentiation by selectively activating TNF receptor 1. J. Biol. Chem. 274, 26287–26295. Xu, H., Hirosumi, J., Uysal, K.T., Guler, A.D., Hotamisligil, G.S., 2002a. Exclusive action of transmembrane TNF alpha in adipose tissue leads to reduced adipose mass and local but not systemic insulin resistance. Endocrinology 143, 1502– 1511. Xu, H., Uysal, K.T., Becherer, J.D., Arner, P., Hotamisligil, G.S., 2002b. Altered tumor necrosis factor-alpha (TNF-alpha) processing in adipocytes and increased expression of transmembrane TNF-alpha in obesity. Diabetes 51, 1876– 1883. Xu, H., Barnes, G.T., Yang, Q., Tan, G., Yang, D., Chou, C.J., et al., 2003. Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J. Clin. Invest. 112, 1821–1830. Yan, D., Qin, N., Zhang, H., Liu, T., Yu, M., Jiang, X., et al., 2009. Expression of TNF-alpha leader sequence renders MCF-7 tumor cells resistant to the cytotoxicity of soluble TNF-alpha. Breast Cancer Res. Treat. 116, 91–102. Zhang, H., Yan, D., Shi, X., Liang, H., Pang, Y., Qin, N., et al., 2008. Transmembrane TNF-alpha mediates “forward” and “reverse” signaling, inducing cell death or survival via the NF-kappaB pathway in Raji Burkitt lymphoma cells. J. Leukoc. Biol. 84, 789–797. Zhou, Q.L., Jiang, Z.Y., Mabardy, A.S., Del Campo, C.M., Lambright, D.G., Holik, J., et al., 2010. A novel pleckstrin homology domain-containing protein enhances insulin-stimulated Akt phosphorylation and GLUT4 translocation in adipocytes. J. Biol. Chem. 285, 27581–27589.