SIRT1 Exerts Anti-Inflammatory Effects and Improves Insulin Sensitivity in Adipocytes^{▿ †}

Materials.

The hemagglutinin (HA) (in the first exofacial loop)-GLUT4-e green fluorescent protein (GFP) (at the carboxyl terminus) constructs was a generous gift from T. E. McGraw (Weill Cornell Medical College, New York, NY). Adenovirus (Ad) with SIRT1 constructs was kindly gifted by Pere Puigserver (Harvard Medical School, Boston, MA). Anti-insulin receptor antibody, anti-Akt1/2 antibody, anti-NF-κB antibody (for chromatin immunoprecipitation [ChIP]), anti-SIRT1 antibody, anti-iNOS antibody, anti-TRAF2 antibody, horseradish peroxidase-linked anti-goat antibody, and small interfering RNA (siRNA) against NF-κB were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Anti-phospho-IRS-1 (Ser307) antibody, anti-IRS1 antibody, and anti-PTP1B antibody were from Upstate Biotechnology, Inc. (Lake Placid, NY). Anti-phospho-insulin receptor (Tyr1146) antibody, anti-phospho-Akt (Ser 473) antibody, anti-phospho-ERK (Thr202/Tyr204) antibody, anti-ERK antibody, anti-NF-κB antibody, anti-phospho-JNK (Thr183/Tyr185) antibody, and anti-JNK antibody were from Cell Signaling (Beverly, MA). Horseradish peroxidase-linked anti-rabbit and anti-mouse antibodies, sheep immunoglobulin G, rhodamine-conjugated anti-rabbit antibody, and Cy3-conjugated secondary antibody were obtained from Jackson Immunoresearch Laboratories, Inc. (West Grove, PA). Anti-phosphotyrosine antibody was from Transduction Laboratory (Lexington, KY). Anti-HA-11 antibody was from Covance (Princeton, NJ). Anti-acetyl-NF-κB (Lysine 310) antibody was from Abcam (Cambridge, MA). Recombinant mouse TNF-α was from R&D systems (Minneapolis, MN). Dulbecco modified Eagle medium and fetal calf serum were obtained from Life Technologies, Inc. (Grand Island, NY). 2-[³H]deoxyglucose or l-[³H]glucose were from ICN (Costa Mesa, CA). All other reagents and chemicals were purchased from Sigma Chemical Co. (St. Louis, MO).

Cell treatment and transient transfection.

3T3-L1 cells were cultured and differentiated as described previously (35). For preparation of whole-cell lysates for immunoblotting experiments, 3T3-L1 adipocytes were stimulated with or without 0.5 or 17 nM insulin at 37°C for various periods as indicated in the figures. Differentiated 3T3-L1 adipocytes were transiently transfected by electroporation, as previously described (34). For Ad infection, 3T3-L1 adipocytes were transduced at a multiplicity of infection (PFU/cell) as indicated in the figures for 16 h with the recombinant Ad encoding GFP or the wild-type (WT) SIRT1 as described previously (21, 34).

Western blotting.

Western blotting experiments were performed as described previously (34). Cells were lysed in a cold solubilizing buffer containing 40 mM Tris, 1 mM EGTA, 100 mM NaCl, 1 mM MgCl₂, 1% Nonidet P-40, 10% glycerol, 1 mM Na₃VO₄, 1 mM phenymethylsulfonyl fluoride, and 20 mM NaF (pH 7.5). The soluble fractions were boiled with Laemmli sample buffer containing 100 mM dithiothreitol and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting with the indicated antibodies.

On phosphorylation and total protein blots, the phosphorylation blotted membrane was stripped in Restore Western blot stripping buffer (Pierce, Rockford, IL) and reprobed with an anti-target protein antibody. If the stripped and reblotted gels did not work adequately, due to different nonspecific bands or uneven protein blots (mirror image due to insufficient strip), then the same samples were loaded at the same amount of protein per lane on separate gels, and one gel was blotted for the phospho-protein and the other for total protein. This latter approach was used in the Western blots of Akt, ERK, and JNK.

Isolation of adipose tissues from mice.

Mice were on the C57BL/6 background. We fed male mice (aged 5 to 6 months) on a high-fat diet (HFD), containing 40% fat by weight (Harlan Teklad Custom Diets), for 16 weeks. The epididymal fat was collected from mice as described previously (14), and Western blotting was performed with the indicated antibodies. For SRT1720 treatment, 16-week-old male mice were placed on a 60% HFD or a 60% HFD that was supplemented with SRT1720 (HFD+SRT1720, 100 mg/kg; Dyets, Inc.). We have previously demonstrated that this dose of SRT1720 improves insulin sensitivity in obese mice and rats (19). After 7 weeks, the mice were anesthetized in the fed state, and the epididymal fat pad was excised, rinsed in saline, and frozen in liquid nitrogen. The body weight at the end of the study was not different between the two groups (43.5 ± 3.2 versus 40.6 ± 2.4 g, HFD versus HFD+SRT1720, P > 0.05). Mouse procedures conformed to Guide for Care and Use of Laboratory Animals of the U.S. National Institutes of Health and were approved by the Animal Subjects Committee of the University of California, San Diego.

SVF isolation.

Epididymal fat pads were excised from mice fed normal chow (NC) or HFD, weighed, rinsed three times in phosphate-buffered saline, and then minced. Tissue suspensions were centrifuged at 500 × g for 5 min, and then treated with collagenase (1 mg/ml) for 30 min at 37°C with shaking. The cell suspensions were filtered through a 100-μm-pore-size filter and centrifuged at 500 × g for 5 min. Stromal vascular fraction (SVF) pellets were then analyzed for RNA.

RNA interference.

The duplexes of siRNA, targeting SIRT1 mRNA (target sequences: 469, 5′-TTAGTGAGGAGTCCATCGG-3′; 1167, 5′-AAATGTCTCCACGAACAGC-3′) and a negative control (scrambled sequence) were purchased from Dharmacon Research, Inc. (Lafayette, CO). The target sequences against SIRT1 were chosen by an online search program (Thermo Scientific), and the absence of homology to any other gene was confirmed by a BLAST search (National Center for Biotechnology Information, National Institutes of Health). On day 7 postdifferentiation, 3T3-L1 adipocytes were electroporated with a siRNA by using the GenePulser XCell (Bio-Rad). Electroporated cells were incubated for 48 h at 37°C prior to assays.

RNA isolation and reverse transcription-PCR (RT-PCR).

Total RNA was isolated and purified from treated cells using TRIzol according to the manufacturer's instructions (Invitrogen, Carlsbad, CA). First-strand cDNA synthesis and quantitative real-time PCR was performed as previous described (12). The ArrayPlate assay was performed by High Throughput Genomics, Inc. (Tucson, AZ) (17). We performed separate control experiments to ensure that the efficiencies of target and reference amplification were equal. The specificity of the PCR amplification was verified by melting-curve analysis of the final products using Opticon 3 software (Bio-Rad). The oligonucleotide primers or target sequence used are available upon request.

2-Deoxyglucose (2-DOG) uptake.

The procedure for evaluating glucose transport was performed as previously described (35). Glucose uptake was determined after the addition of 5 μl of substrate. 2-[³H]deoxyglucose or l-[³H]glucose (0.1 μCi; final concentration, 0.01 mM) was added to provide a concentration at which cell membrane transport is rate limiting.

Membrane targeting kinetics of HA-GLUT4-eGFP.

An assay of the membrane targeting kinetics of HA-GLUT4-eGFP was performed as previously described (34). In transfected cells, the intensities of the GFP (reflecting the total cellular HA-GLUT4-GFP level) and Cy3 (reflecting the cell surface HA-GLUT4-GFP) signals were quantified, and background GFP and Cy3 fluorescent emissions were subtracted. The Cy3 fluorescence intensity for each HA-GLUT4-GFP-expressing cell was divided by the GFP fluorescence intensity to determine the fraction of tagged GLUT4 transporter at the membrane (arbitrary units). Fluorescence quantification was previously described (34).

NF-κB acetylation assays.

3T3-L1 adipocytes from different experiment conditions were stimulated with 10 ng of TNF-α/ml for 24 h. Cells were lysed in a cold solubilizing buffer described in Western blotting section with 5 mM nicotinamide. The soluble fractions were boiled with Laemmli sample buffer and analyzed by SDS-PAGE and immunoblotting with anti-acetyl-NF-κB antibody or anti-NF-κB antibody.

ChIP assay.

The ChIP assay was performed by using a ChIP assay kit form Upstate Biotechnology, Inc., with some modifications. Briefly, ∼10⁶ differentiated 3T3-L1 adipocytes, electroporated with control or SIRT1 siRNA, were cross-linked for 10 min by adding formaldehyde directly to the tissue culture medium to a final concentration of 1%. Cross-linked cells were scraped, pelleted by centrifugation, and resuspended in SDS lysis buffer. After centrifugation of sonicated chromatin solution, the supernatant was divided into aliquots for 10-fold dilution in ChIP dilution buffer and precleared with protein A-agarose containing salmon sperm DNA for 30 min. Anti-NF-κB antibody was added, followed by incubation for 18 h at 4°C, followed in turn by incubation with protein A-agarose for 1 h. The precipitates were washed, and chromatin complexes were eluted. After reversal of the cross-linking, the DNA was purified by using a QIAquick PCR purification kit from Qiagen (Germantown, MD), and input control or ChIP samples were used as a template for PCR using the primer sets for regions containing NF-κB binding sites. The oligonucleotide primers used are available upon request.

Preparation of SRT1720 and SRT2530.

The compounds SRT1720 (19) and SRT2530 were synthesized at Sirtris, a GSK company.

SIRT1 protein expression, purification, and mass spectrometry assay.

Human SIRT1 was expressed from the pSIRT1FL vector which places expression under the control of the T7 promoter. The protein was expressed in Escherichia coli BL21Start (DE3) (Invitrogen) as an N-terminal fusion to a His₆ affinity tag. The expressed protein was purified by Ni²⁺-chelate chromatography, followed by size exclusion chromatography and ion exchange. The resulting protein was typically >95%, pure as assessed by SDS-PAGE analysis in 25 mM Tris-HCl (pH 8.0) and 2 mM TCEP [tris(2-carboxyethyl) phosphine]. The mass spectrometry assay was performed as previously described (19), using a peptide [Ac-Glu-Glu-Lys(biotin)-Gly-Gln-Ser-Thr-Ser-Ser-His-Lys(Ac)-Nle-Ser-Thr-Glu-Gly-Lys(MR121 or TAMRA)-Glu-Glu-NH2] derived from the sequence of p53. The mass spectrometry assay was conducted using the following: 0.5 μM peptide substrate, 120 μM βNAD⁺, 10 nM SIRT1, and reaction buffer (50 mM Tris-acetate [pH 8], 137 mM NaCl₂,7 mM KCl, 1 mM MgCl₂, 5 mM dithiothreitol, 0.05% bovine serum albumin). Reactions were incubated for 25 min at 25°C. Test compounds were added to the reaction or the vehicle control, dimethyl sulfoxide (DMSO). After the incubation with SIRT1, 10% formic acid with 50 mM nicotinamide was added to stop the reaction. The mass spectrometry analysis was performed by BioTrove, Inc. (Woburn, MA). Determination of the mass of the substrate peptide allows for precise determination of the degree of acetylation (i.e., substrate) compared to deacetylated peptide (product).

Statistics.

The values are expressed as mean ± the standard error (SE). Scheffe's multiple comparison test was used to determine the significance of any differences among more than three groups. A P value of <0.05 was considered significant.

SIRT1 is downregulated by HFD in adipose tissue.

SIRT1 is upregulated by caloric restriction and likely to be a nutrient sensor. Accordingly, we measured in vivo SIRT1 protein expression in adipose tissue from 16-week HFD-fed mice compared to NC-fed mice. HFD mice are obese and insulin resistant (18, 25), and Fig. 1A shows a dramatic reduction of adipose tissue SIRT1 protein content. We also performed quantitative PCR to measure SIRT1 mRNA expression in the adipocyte and SVF of adipose tissue. Consistent with the protein content, SIRT1 mRNA was decreased in the adipocyte fraction, with a smaller decrease in the SVF component from HFD mice (Fig. 1B).

Knockdown of SIRT1 by RNAi inhibits insulin-stimulated glucose transport in 3T3-L1 adipocytes.

The findings presented in Fig. 1 raises the question as to the potential role of SIRT1 in insulin action. To assess this, we utilized siRNA against SIRT1 in 3T3-L1 adipocytes. Two siRNAs against SIRT1 sequences were tested, and Fig. 2A and B show the results. At 48 h after electroporation of siRNAs, we observed a marked decrease in SIRT1 mRNA in cells electroporated with siSIRT1 (469 and 1167) compared to cells treated with a scrambled siRNA (lane ctrl) (Fig. 2A). We also assessed protein expression and, similar to the mRNA results, electroporation of siSIRT1 led to a comparable decrease in SIRT1 protein (Fig. 2B), with siSIRT (SIRT1 1167) being most effective. In contrast, the expression levels of actin were unaffected (Fig. 2A and B).

To test whether SIRT1 affects glucose metabolism, we measured 2-DOG uptake in 3T3-L1 adipocytes with or without SIRT1 depletion. SIRT1 knockdown, led to a 25 to 40% decrease in insulin-stimulated glucose transport compared to control, scrambled siRNA electroporated cells (Fig. 2C). Treatment with siSIRT1 469, which had approximately half the effect to knock down SIRT1 compared to siSIRT1 1167 (Fig. 2A and B), led to a 10 to 15% decrease in insulin-stimulated glucose uptake (Fig. 2C). Thus, the efficiency of SIRT1 knockdown correlated with the inhibition of glucose uptake, suggesting that the effect is coupled to SIRT1 expression level.

We also measured GLUT4 translocation. 3T3-L1 adipocytes were electroporated with siSIRT1, followed by the detection of insulin-stimulated GLUT4 translocation by measuring movement of HA-GLUT4-eGFP to the cell surface, as previously described (34). As seen in Fig. 2D, these results demonstrate inhibition of GLUT4 translocation by siSIRT1.

SIRT1 knockdown leads to increased IRS-1 serine phosphorylation and decreased insulin signaling.

To assess the mechanisms for the decreased insulin-stimulated glucose uptake and GLUT4 translocation, we measured the effect of SIRT1 knockdown on the early steps of insulin signaling. Compared to control 3T3-L1 adipocytes, knockdown of SIRT1 had no effect on tyrosine phosphorylation of the insulin receptor (Fig. 2E). In contrast, SIRT1 knockdown led to a decrease in IRS-1 tyrosine phosphorylation, which was most evident at the submaximal dose of insulin (0.5 nM). Consistent with this, downstream signaling was also impaired in the knockdown cells, as seen by decreased phosphorylation of Akt and ERK (Fig. 2E).

IRS-1 serine (Ser)/threonine phosphorylation has been increasingly recognized as a negative regulator of IRS signaling (38), and Ser307 has been the most intensively studied and is a well-known target of JNK and inhibitor of κB kinase (IKKβ) (1, 8). Since SIRT1 knockdown can activate proinflammatory pathways in macrophages (T. Yoshizaki et al., unpublished data), we hypothesized that SIRT1 knockdown in adipocytes could enhance inflammatory pathway activity, leading to increased Ser phosphorylation of IRS-1. Consistent with this idea, Fig. 2F shows that SIRT1 knockdown led to an increase in JNK phosphorylation, which was accompanied by increased IRS-1 Ser 307 phosphorylation. This suggests that SIRT1 knockdown increases the JNK/IRS1 Ser 307 pathway and decreases tyrosine phosphorylation of IRS-1, which then inhibits downstream insulin signaling to glucose transport.

SIRT1 represses of proinflammatory gene expression.

Recent reports suggest a close link between insulin resistance and chronic inflammation (2, 4, 13, 25, 36), and the data in Fig. 2 indicate that SIRT1 may play a role in adipocyte inflammatory responses. To further assess this, we quantified the transcript levels of several inflammatory pathway components. As seen in Fig. 3, SIRT1 knockdown led to a significant increase in mRNA expression for TNF-α (P < 0.05), interleukin-6 (P < 0.05), MCP-1 (P < 0.05), JNK (P < 0.01), KC (P < 0.01), VCAM-I (P < 0.05), MMP9 (P < 0.05), CRP (P < 0.05), COX2 (P < 0.01), and interleukin-1β (P < 0.01), indicating that SIRT1 can broadly suppress inflammatory gene expression.

SIRT1 deacetylates NF-κB and inhibits NF-κB binding to NF-κB target gene promoters.

To test the molecular mechanisms of these SIRT1 effects, we assessed the acetylation state of NF-κB, a known SIRT1 target (33). We utilized anti-acetyl-NF-κB antibody, which recognizes NF-κB acetylated at lysine 310. Western blots with this antibody showed that SIRT1 knockdown led to increased NF-κB acetylation (Fig. 4A). To evaluate whether this NF-κB acetylation affected transcriptional responses, ChIP assays were conducted (Fig. 4B). Chromatin extracts from cells electroporated with control or SIRT1 siRNA were immunoprecipitated with NF-κB antibody, and the DNA sequences encompassing the regions containing the NF-κB binding sites in the MMP9 and MCP1 promoters (Fig. 4B) were amplified by real-time quantitative PCR. Treatment of the cells with TNF-α strongly induced the association of NF-κB with the MMP9 and MCP1 promoters (Fig. 4B), and SIRT1 knockdown significantly increased this effect. These data indicate that the increased inflammatory responses in SIRT1 knockdown cells is due, at least in part, to hyperacetylated NF-κB-induced proinflammatory gene expression.

NF-κB knockdown improves insulin signaling.

Electroporation of siRNA against NF-κB led to a 80 to 90% decrease in NF-κB protein (Fig. 5A bottom). In contrast to SIRT1 depletion, NF-κB knockdown caused a decrease in JNK phosphorylation, with less IRS-1 serine 307 phosphorylation, accompanied by an increase in IRS-1 tyrosine phosphorylation, Akt and ERK phosphorylation (Fig. 5A), and glucose uptake (Fig. 5B). Concomitantly with this, as seen in Fig. 5C, NF-κB depletion inhibited mRNA expression of several inflammatory genes.

SIRT1 activation is required for anti-inflammatory effects.

We next examined the effect of SIRT1 overexpression in 3T3-L1 adipocytes. The infection of Ad WT SIRT1 increased SIRT1 expression (see Fig. S1A in the supplemental material). However, as seen in Fig. S1B in the supplemental material, overexpression of WT SIRT1 had no effect on insulin-stimulated glucose uptake. Consistent with this, phosphorylation of the insulin receptor, IRS-1, Akt and ERK were equal in both control Ad GFP and Ad WT SIRT1-infected cells (see Fig. S1C in the supplemental material). Furthermore, overexpression of WT SIRT1 did not affect serine phosphorylation of IRS-1, JNK phosphorylation (see Fig. S1D in the supplemental material), or NF-κB acetylation (see Fig. S1E in the supplemental material), suggesting that expression of SIRT1 beyond the normal endogenous level is without effect, unless SIRT1 is activated. This point is clearly demonstrated in Fig. S1F to L in the supplemental material, which shows that treatment of the Ad WT SIRT1-infected cells with a highly potent, selective SIRT1 activator, SRT2530 (see Fig. S1F in the supplemental material) or SRT1720 (19), led to a marked decrease in JNK phosphorylation, with a corresponding decrease in IRS-1 serine 307 phosphorylation (see Fig. S1G and J in the supplemental material), accompanied by an increase in the early steps of insulin signaling (IRS-1 tyrosine phosphorylation, Akt and ERK phosphorylation) (see Fig. S1H and K in the supplemental material) with enhanced glucose transport (see Fig. S1I and L in the supplemental material). These compounds are highly selective for activation of SIRT1, and the glucose-lowering and in vivo insulin-sensitizing effects of SRT1720 have been recently published (19).

Anti-inflammatory effects of SIRT1 activator in adipose tissue from HFD-fed mice.

We have assessed SIRT1 protein levels in adipose tissue from mice placed on a HFD with or without SRT1720 treatment (Fig. 6A and B). The ratio of acetylated NF-κB (p65) to total p65 was significantly decreased by SIRT1720 treatment (Fig. 6C, P < 0.001). Furthermore, the protein expression levels of NF-κB target genes, JNK (Fig. 6D, P = 0.004), iNOS (NOS2) (Fig. 6E, P = 0.031), and TNF-α receptor-associated factor 2 (Fig. 6F, P = 0.019) were decreased by SRT1720 treatment. No change in SIRT1 protein expression was observed in adipose tissue after SIRT1 activator treatment. These in vivo findings are consistent with the results in 3T3-L1 adipocytes, showing that SIRT1 activation inhibits proinflammatory gene expression via deacetylation of NF-κB.

SIRT1 activator increases insulin action in nontransduced cells.

On the basis of the ability of SIRT1 to repress inflammatory components in 3T3-L1 adipocytes, we next sought to determine whether a SIRT1 activator could produce similar effects in nontransfected cells expressing endogenous levels of SIRT1. In these studies, we used the selective SIRT1 activators, SRT2530 and SRT1720 (19). Both compounds clearly increased SIRT1 enzymatic activity for SRT2530: effective concentration required to increase enzyme activity by 50% = 60 nM (see Fig. S1F in the supplemental material) (19). Treatment with these activators significantly increased insulin-stimulated glucose uptake (Fig. 7A and see Fig. S2A in the supplemental material), as well as GLUT4 translocation to the plasma membrane (Fig. 7B and see Fig. S2B in the supplemental material). In addition, pretreatment with the activators led to an increase in insulin-stimulated tyrosine phosphorylation of IRS-1, Akt, and ERK (Fig. 7C and see Fig. S2C in the supplemental material), as well as decreased JNK phosphorylation and serine 307 phosphorylation of IRS-1 (Fig. 7D and see Fig. S2D in the supplemental material). In contrast to SIRT1 knockdown, treatment with these agents significantly decreased TNF-α-stimulated NF-κB acetylation (Fig. 7E). These data further show that SIRT1 activation inhibits inflammatory responses and increases cellular insulin sensitivity. In SIRT1 knockdown cells, pretreatment with SRT2530 had no effect on insulin-stimulated glucose uptake (Fig. 7F) and JNK, IRS1 (tyrosine and serine), Akt, and ERK phosphorylation (Fig. 7G), indicating that the SIRT1 activator modulates insulin actions via effects on SIRT1. Furthermore, in NF-κB knockdown cells, the SIRT1 activator was without effects on insulin signaling (Fig. 7H). This suggests that activation of SIRT1 inhibits inflammatory responses and improves insulin action, at least in part, by inhibition of NF-κB activity.

SIRT1 activation rescues TNF-α-induced insulin resistance.

Since TNF-α is a major proinflammatory cytokine that can cause decreased systemic insulin sensitivity, we determined whether TNF-α-induced insulin resistance could be rescued by SIRT1 activation. To accomplish this, we measured the effect of TNF-α on glucose transport in adipocytes that were pretreated with or without SRT2530. Figure 8A shows that exposure to TNF-α for 48 h significantly decreased insulin responsiveness by 29%, whereas pretreatment with the activator partially restored insulin-stimulated glucose uptake by 56% (P < 0.01) in TNF-α-treated cells. Protein tyrosine phosphatase 1B (PTP1B) has been recognized as a negative regulator of insulin sensitivity (7), and recent reports suggest that TNF-α can increase PTP1B expression (37) and that SIRT1 can decrease it (26). To test whether the effects of SRT2530 on TNF-α-induced insulin resistance depend on PTP1B, we measured PTP1B expression. Treatment with TNF-α for 48 h did not affect PTP1B protein content, and there was also no effect of pretreatment of SRT2530 (Fig. 8B).

SIRT1 activation has no effect in salicylate-treated adipocytes.

Salicylate is a well-known anti-inflammatory molecule, and it has been shown that pretreatment of 3T3-L1 adipocytes with salicylate prevented TNF-α-induced JNK and IRS-1 (serine) phosphorylation, and restored insulin-stimulated glucose uptake (9). To determine whether salicylate and SIRT1 activation could have additive effects, we cotreated cells with both agents. Consistent with previous reports, pretreatment with salicylate reduced TNF-α-induced IRS-1 Ser 307 phosphorylation, along with decreased JNK phosphorylation (Fig. 9A). Salicylate treatment also improved insulin-stimulated IRS-1 tyrosine phosphorylation, Akt, and ERK phosphorylation, as well as glucose uptake in TNF-α-treated cells. In all of these experiments, SRT2530 had no further effect (Fig. 9).