Conjugated Linoleic Acid Promotes Human Adipocyte Insulin Resistance through NFκB-dependent Cytokine Production*

We previously demonstrated that trans-10, cis-12 conjugated linoleic acid (CLA) reduced the triglyceride content of human adipocytes by activating mitogen-activated protein kinase kinase/extracellular signal-related kinase (MEK/ERK) signaling via interleukins (IL) 6 and 8. However, the upstream mechanism is unknown. Here we show that CLA increased (≥6 h) the secretion of IL-6 and IL-8 in cultures containing both differentiated adipocytes and stromal vascular (SV) cells, non-differentiated SV cells, and adipose tissue explants. CLA isomer-specific induction of IL-6 and tumor necrosis factor-α was associated with the activation of nuclear factor κB (NFκB) as evidenced by 1) phosphorylation of IκBα, IκBα kinase, and NFκB p65, 2) IκBα degradation, and 3) nuclear translocation of NFκB. Pretreatment with selective NFκB inhibitors and the MEK/ERK inhibitor U0126 blocked CLA-mediated IL-6 gene expression. Trans-10, cis-12 CLA suppression of insulin-stimulated glucose uptake at 24 h was associated with decreased total and plasma membrane glucose transporter 4 proteins. Inhibition of NFκB activation or depletion of NFκB by RNA interference using small interfering NFκB p65 attenuated CLA suppression of glucose transporter 4 and peroxisome proliferator-activated receptor γ proteins and glucose uptake. Collectively, these data demonstrate for the first time that trans-10, cis-12 CLA promotes NFκB activation and subsequent induction of IL-6, which are at least in part responsible for trans-10, cis-12 CLA-mediated suppression of peroxisome proliferator-activated receptor γ target gene expression and insulin sensitivity in mature human adipocytes.

isomer (e.g. 80% cis-9, trans-11 CLA, 10% trans-10, cis-12 CLA). CLA is also available commercially as a dietary supplement for weight loss, with both isomers reported to be present at equal amounts (e.g. ϳ35% each). A great deal of attention has been centered on trans-10, cis-12 CLA due to its reported anti-obesity actions in animal models (for review, see Ref. 1) and some humans (for review, see Ref. 2). We have reported that trans-10, cis-12 CLA, but not cis-9, trans-11 CLA, inhibited human preadipocyte differentiation (3) and caused delipidation of newly differentiated human adipocytes (4). CLA isomer-specific delipidation of adipocytes was due largely to decreased glucose and fatty acid uptake and TG synthesis as opposed to increased oxidation. Interestingly, CLA suppression of glucose and fatty acid uptake was positively correlated with activation of mitogen-activated protein kinase kinase/extracellular signal-related kinase (MEK/ERK) and G protein-coupled receptor (GPCR) signaling and robust secretion of the proinflammatory cytokines interleukin-6 (IL-6) and IL-8 after 24 h of treatment. However, the underlying mechanism(s) by which trans-10, cis-12 CLA triggers cytokine production and impairs glucose and fatty acid uptake is unclear.
There is growing evidence linking inflammatory cytokines with the development of obesity, insulin resistance (5)(6)(7), and atherosclerosis (for review, see Ref. 8). Adipose tissue plays a central role in this relationship given its ability to both secrete cytokines and act as a substantial target for cytokines. A diverse array of cytokines such as tumor necrosis factor ␣ (TNF-␣), IL-6, and IL-8 have been positively associated with obesity and the development of insulin resistance in muscle and adipose tissue.
IL-6 expression can be induced by many transcription factors such as nuclear factor B (NFB), NF-IL6 (nuclear factor for IL-6 expression, a.k.a. CCAAT enhancer-binding protein ␤ (C/EBP␤)), activator protein-1, and cAMP response element (CRE)-binding protein (CREB) depending on cell type and stimulus (9,10). Activation of NFB is critical for fatty acid-induced IL-6 secretion and insulin resistance in myotubes in vitro (11,12) and muscle in vivo (13). Interestingly, several studies have shown that chronic exposure to IL-6 reduces adipogenic gene expression (14,15) and insulin signaling and glucose uptake in adipocytes (16).
NFB is a ubiquitous transcription factor regulating the expression of genes promoting inflammation and cell survival. NFB-dependent transcription normally begins with the phosphorylation of the inhibitory B proteins (IBs) and their subsequent degradation. IB proteins are key regulators of NFB, sequestering the NFB dimer in the cytoplasm. Phosphorylation of IB␣ by IB␣ kinase (IKK) triggers its polyubiquitination and proteosomal degradation, thereby releasing NFB, especially p50/p65, to the nucleus (for review, see Refs. 17 and 18). After phosphorylation by mitogen-activated protein kinases such as ERK1/2 and p38 or mitogen and stress-activated kinase 1 (MSK1) (for review, see Refs. 19 and 20), phospho-p50/p65 binds to the NFB response elements of target genes, thereby inducing their transcriptional activation (for review, see Refs. 17 and 21).
CLA TG-lowering actions in human adipocytes are consistent with NFB activation based on our data (3,4) showing that trans-10, cis-12 CLA 1) decreases adipogenic gene expression, 2) impairs glucose and fatty acid uptake and conversion to TG, and 3) increases the expression and secretion of several proinflammatory cytokines including IL-6 and IL-8. Therefore, the aim of the present study was to examine the extent to which NFB activation was essential for trans-10, cis-12 CLA-induced cytokine expression and impaired glucose uptake in primary human adipocytes. Here we demonstrate for the first time that trans-10, cis-12 CLA promotes NFB activation before inducing IL-6 expression and insulin resistance in cultures of newly differentiated human adipocytes. The essential role of NFB in mediating this effect of CLA was demonstrated by using either chemical inhibitors of NFB or RNA interference targeting NFB p65. Last, our data suggest that non-adipocytes or stromal vascular (SV) cells secrete a significant amount of cytokines in response to treatment with trans-10, cis-12 CLA.
Cell Culture of Primary Human SV Cells-Primary human SV cells were obtained from subcutaneous adipose tissues of non-obese (body mass index Ͻ30) females undergoing abdominoplasty with the consent from the Institutional Review Board at the University of North Carolina at Greensboro. Isolation, culture of SV cells from adipose tissue, and differentiation into adipocytes were performed as previously described (4) except for the condition that SV cells were pooled from 3ϳ5 independent human subjects for most experiments. Experimental treatment of cultures containing adipocytes began on ϳday 12 of differentiation unless otherwise indicated. For the preparation of cultures of non-differentiated SV cells, cells were seeded at ϳ80% confluent.
Plasma Membrane (PM) Isolation and Glut4 Levels-For the detection of Glut4 translocation, PM was fractionated according to Carvalho et al. (23). Briefly, cultures were washed once with TES buffer (20 mM Tris-HCl, 1 mM EDTA, 225 mM sucrose, pH 7.4 at 20°C) and then homogenized in ice-chilled TES using a 1-ml Dounce homogenizer. The homogenate was centrifuged at 16,000 ϫ g for 20 min at 4°C, and the solidified surface fat was removed. The resulting pellets (crude membrane) were resuspended in TES and layered on a 1.12 M sucrose cushion in 20 mM Tris-HCl, 1 mM EDTA and centrifuged at 100,000 ϫ g for 30 min. PMs at the interface were collected, resuspended in TES, and centrifuged at 100,000 ϫ g for 30 min. PM pellets were resuspended in TES and immunoblotted for Glut4. The abundance of Glut4 was quantified from exposed x-ray film using the KODAK image station 440 (Eastman Kodak Co.).
IL-6 and IL-8 Secretion-Differentiated cultures of adipocytes and cultures of non-differentiated SV cells were serum-starved for 20 h before fatty acid treatment. Fatty acid treatment was initiated by adding either BSA vehicle or 30 M trans-10, cis-12 CLA to the media directly. Conditioned media from above the cell monolayer was collected at 0, 3, 6, 12, or 24 h, centrifuged at 12,000 ϫ g for 20 min to remove cell debris, and kept at Ϫ80°C before analysis. Cultures were washed twice with ice-chilled Hanks' balanced salt solution, and cells were harvested and dedicated for protein determination. IL-6 and IL-8 secretion to the media was quantified using a commercial ELISA (R&D Systems) following the manufacturer's protocol and normalized to the protein content of the monolayer.
To measure IL-6 and IL-8 secretion from human subcutaneous adipose tissue explants, 500-mg pieces of adipose tissue obtained per subject were each minced into ϳ100 fragments as described by Fried et al. (24). Each minced explant was then preincubated in Dulbecco's modified Eagle's medium-Ham's F-12 (1:1) medium containing 100 units/ml penicillin, 100 units/ml streptomycin, 50 g/ml gentamicin, and 0.25 mg/ml amphotericin-B at 37°C overnight. Then explants were transferred to 10 ml of fresh media containing either BSA vehicle or 30 M trans-10, cis-12 CLA in culture tubes. Subsequently, conditioned media was collected after 8, 24, or 72 h of incubation with treatments and stored at Ϫ80°C until analysis. IL-6 and IL-8 secretion were quantified by ELISA.
Relative RT-PCR-Total RNA was isolated from the cultures using Tri Reagent (Molecular Research Center Inc., Cincinnati, OH) following the manufacturer's protocol. Relative (semiquantitative) RT-PCR was carried out using One-step RT-PCR kit (Qiagen Inc) as we described previously (3). Primer sets for adipose fatty acid-binding protein (aP2) were previously described (3). Primer sequences for IL-6 were (accession number NM_000600) sense (5Ј-CCAGCTATGAACTCCT-TCTC) and antisense (5Ј-GCTTGTTCCTCACATCTCTC), and running conditions for IL-6 were 26 cycles at 94°C for 30 s, 57°C for 30 s, and 72°C for 30 s. The running conditions for gene-specific primers for TNF-␣ (Ambion #5345) were 40 cycles at 94°C for 30 s, 57°C for 30 s, and 72°C for 1 min.
Immunofluorescence Microscopy and Phase Contrast Image-Cells were cultured on coverslips for immunofluorescence microscopy and stained as described previously (4,25). For phospho-IKK immunostaining, cells were permeabilized with 0.1% saponin and then incubated with a 1:10 dilution of rabbit-anti-phospho-IKK overnight at 4°C. After three vigorous washes, coverslips were incubated with 1:200 dilutions of FITC-conjugated anti-rabbit IgG for 1 h (Fig. 4C). For NFB p65 and aP2 double staining (Fig. 5), cells were initially incubated with a 1:10 dilution of NFB p65 for 12 h at 4°C followed by incubation with a 1:500 dilution of Cy3-conjugated anti-mouse IgG. After adequate washing, coverslips were incubated with a 1:200 dilution of aP2 for 2 h at room temperature followed by incubation with a 1:400 dilution of FITC-conjugated anti-rabbit IgG for 1 h. Fluorescent images were captured with a SPOT digital camera mounted on an Olympus BX160 fluorescence microscope. Cy3-labeled siGLO fluorescent, and phase contrast images were captured without fixation under the Olympus IX60 microscope equipped with a SPOT digital camera (Fig. 8B). For PPAR␥ immunostaining (Fig. 8E), transfected cells grown on coverslips were permeabilized with 0.1% Triton X-100 on ice for 10 min followed by incubating with a 1:10 dilution of mouse-anti-human PPAR␥ (sc7273) for 16 h at 4°C and a 1:200 dilution of FITC-conjugated anti-mouse IgG for 1 h. 1 g/ml Hoechst dye was used for nuclei staining (Fig. 8E).
Nuclear and Cytoplasmic Separation and Assessment of NFB p65 DNA Binding-Nuclear and cytosolic cellular fractions were prepared using a commercially available kit from Active Motif, with the following minor modifications. Cells were directly lifted in a 1ϫ hypotonic buffer, gently scraped from the plate, and then treated according to the manufacturer's recommendations. Trans-10, cis-12 CLA-treated nuclear extract was used to assess CLA induced-NFB p65 DNA binding by using the ELISA-based TransAM TM NFB family transcription factor assay kit (Active Motif) following the manufacturer's instructions.
Transfection with NFB p65 siRNA-Transfection of newly differentiated human adipocytes with NFB p65 siRNA was conducted on day 6 of differentiation in 35-mm cell culture plates without detaching the cells. Cells were seeded and differentiated as previously described. On ϳday 6 of differentiation, cultures containing fresh adipocyte media (AM-1, Zen Bio Inc., Research Triangle Park, NC) were supplemented with either 25 nM NFB p65 siRNA or non-targeting siRNA complexed with TransIT-TKO (6 l/ml), a non-lipid-based transfection reagent from Mirus Corp. Transfection reagent and undelivered siRNA were removed 24 h post-transfection by removing the media and washing the cells twice with Hanks' balanced salt solution. Transfection efficiency was examined by transfecting the cells with a fluorescent-tagged, RISCfree siRNA obtained from the company.
Statistical Analysis-Unless otherwise indicated, data are expressed as the means Ϯ S.E. Data were analyzed using one-way analysis of vari-ance followed by Student's t tests for each pair for multiple comparisons. Differences were considered significant if p Ͻ 0.05. All analyses were performed using JMP IN, Version 4.04 (SAS Institute; Cary, NC) software.

RESULTS
Trans-10, Cis-12 CLA Reduces Glucose Uptake and the Abundance of Glut4 and IRS-1-We previously demonstrated that trans-10, cis-12 CLA decreased glucose uptake and Glut4 gene expression after 24 or 72 h of treatment (4). However, we did not examine the time course of CLA-induced insulin resistance and the extent to which dysregulation of insulin signaling and Glut4 expression played a role in impaired glucose uptake. Therefore, we measured [2-3 H]deoxyglucose uptake, phosphorylation of downstream targets of insulin, and Glut4 abundance in cultures of differentiated human adipocytes. As shown in Fig. 1A, insulin-stimulated [2-3 H]deoxyglucose uptake was significantly lower in cultures treated with 30 M trans-10, cis-12 CLA for either 24 or 72 h compared with vehicle (BSA) controls. However, trans-10, cis-12 CLA had minimal effects on insulin signaling as determined by basal-and insulin-stimulated phosphorylation of insulin receptor substrate-1 (IRS-1) at Ser-307 or Tyr-891 and on Ser-473 of Akt/protein kinase B (Fig. 1B). Interestingly, trans-10, cis-12 CLA modestly decreased the abundance of total IRS-1 in the absence and presence of insulin as early as 24 h, but more robustly after 72 h (data not shown), which is similar to that reported for IL-6-induced insulin resistance in 3T3-L1 adipocytes (16). The basal levels of Glut4 were markedly lower in all three cellular fractions from cultures treated with trans-10, cis-12 CLA (Fig.  1C) compared with BSA controls, consistent with CLA-induced reductions in Glut4 gene expression (4). More importantly, the abundance of Glut4 in the PM fractions of cultures treated with both insulin and trans-10, cis-12 CLA was lower than in cultures treated with insulin and vehicle. The degree to which trans-10, cis-12 CLA decreased [2-3 H]deoxyglucose uptake after 24 h of treatment (ϳ30%, Fig. 1A) was relatively similar to the degree to which it attenuated Glut4 abundance in both membrane fractions (ϳ35%, Fig. 1C). Taken together, these data suggest that trans-10, cis-12 CLA decreases insulin-stimulated glucose uptake by decreasing intracellular pools of key proteins involved in insulin-stimulated glucose uptake (i.e. IRS-1, Glut4) rather than impairing insulin signaling per se.
Trans-10, Cis-12 CLA Induces Cytokine Secretion and/or Expression in SV Cells, Adipocytes, and Tissue Explants-Recent reports, including our own data using human primary cultures (4), suggest that adipocytes are molecular targets for IL-6, resulting in insulin resistance (16). Although we previously demonstrated that trans-10, cis-12 CLA robustly increased IL-6 and IL-8 protein secretion and gene expression after 24 h of treatment (4), we did not know how early this occurred, which cells in the cultures were secreting these cytokines, or if CLA induction of cytokine gene expression was isomer-specific. Therefore, we examined CLA-induced changes in cytokine secretion and gene expression over time. IL-6 and IL-8 secretion in cultures treated with 30 M trans-10, cis-12 CLA increased in a time-dependent manner in differentiated cultures of adipocytes ( Fig. 2A), whereas IL-6 and IL-8 secretion in control cultures increased only marginally with time ( Fig. 2A). Accumulations of IL-6 and IL-8 in the media of cultures treated with trans-10, cis-12 CLA were apparent after 6 h of treatment, reaching a plateau after 24 h ( Fig. 2A). Interestingly, trans-10, cis-12 CLA induced IL-6 and IL-8 secretion in non-differentiated cultures of SV cells (Fig.  2B) and adipose tissue explants (Fig. 2C) as well. In fact, IL-6 and IL-8 secretion was 10-and 7-fold higher, respectively, in the non-differentiated cultures of SV cells treated with CLA for 24 h (Fig. 2B) compared ; n ϭ 6) not sharing a common lowercase letter differ, p Ͻ 0.05. B, after treatment with either BSA or trans-10, cis-12 CLA for 24 h, cells were cultured in the absence (Ϫ) or presence (ϩ) of 100 nM insulin for 10 min, and total cell extracts were immunoblotted with phospho (P)-specific antibodies targeting p-IRS-1 (Ser-307), p-IRS-1 (Tyr-891), p-Akt (Ser-473), p-ERK1/2 (Thr-202,204), and total IRS-1. C, cultures were treated with either BSA vehicle or trans-10, cis-12 CLA for 24 h before incubation in the absence (Ϫ) or presence (ϩ) of 100 nM human insulin for 30 min. Total cell extracts were harvested, and crude and plasma membrane fractions were isolated by differential centrifugation. Each fraction was immunoblotted for Glut4 and caveolin-1 (Cav-1). Blots were quantified by densitometry, and the amount of Glut4 relative to caveolin-1 was expressed as a bar graph under each blot.
with differentiated cultures of adipocytes treated with CLA ( Fig. 2A). This supports our previous observation that SV cells are the predominant source of IL-6 and IL-8 secretion in our cultures treated with trans-10, cis-12 CLA (4).
To determine whether CLA-mediated IL-6 secretion was due to increased IL-6 gene expression, differentiated cultures of adipocytes were treated with either BSA vehicle or 30 M trans-10, cis-12 for 1, 3, 6, 12, or 24 h. As seen in Fig. 3A, trans-10, cis-12 induced IL-6 gene expression beginning at 3 h, which was consistent with the IL-6 protein secretion shown in Fig. 2A. Intriguingly, trans-10, cis-12 CLA treatment induced transient TNF-␣ gene expression, which peaked at 3 h. However, TNF-␣ protein secretion was not detectable (data not shown, Ͻ 0.5 pg/ml). The mRNA levels of IL-6 and IL-8 in adipose tissue explants treated with CLA were also higher compared with those receiving the BSA vehicle (data not shown).
To determine the extent to which CLA induction of IL-6 and TNF-␣ gene expression was isomer-specific, differentiated cultures of adipocytes were treated with either BSA vehicle, 30 M cis-9, trans-11 CLA, or 30 M trans-10, cis-12 CLA for 3 h, and IL-6 and TNF-␣ mRNA levels were measured. As a positive control, another set of cultures was treated with 100 ng/ml human recombinant TNF-␣ for 30 min. As seen in Fig.  3B, trans-10, cis-12 CLA, but not cis-9, trans-11 CLA or BSA, induced IL-6 and TNF-␣ gene expression in cultures of differentiated adipocytes (Fig. 3B) as well as in cultures of non-differentiated SV cells (data not shown). Collectively, these data demonstrate for the first time that cells isolated from human adipose tissue secrete proinflammatory cytokines in vitro when treated with trans-10, cis-12 CLA.
Trans-10, Cis-12 CLA Promotes NFB Activation-We hypothesized that trans-10, cis-12 CLA treatment would lead to NFB activation in cultures of newly differentiated adipocytes. This is supported by the following evidence. 1) Both IL-6 and TNF-␣ possess a nuclear factor B response element (21); 2) transient TNF-␣ gene expression is positively linked to NFB activation (26,27); 3) ERK1/2 phosphorylation of nuclear p50/p65 is important for its activation (20); 4) NFB-mediated IL-6 and TNF-␣ production are associated with insulin resistance in myotubes (11,12,28); 5) human adipose tissue is a target for IL-6mediated insulin resistance (16). To test this hypothesis, cultures were treated with either BSA vehicle or 30 M trans-10, cis-12 CLA, and IB␣ protein levels were measured by immunoblotting. In support of our hypothesis, IB␣ was lower in cultures treated for 3 h with trans-10, cis-12 CLA than in the controls (Fig. 4A), implying that NFB is activated by CLA. CLA-mediated NFB activation appeared to be isomerspecific, as evidenced by phosphorylation and subsequent degradation of IB␣ and ERK1/2 phosphorylation in trans-10, cis-12 CLA-treated cultures compared with cultures treated with cis-9, trans-11 CLA or BSA (Fig. 4B). In contrast to TNF-␣ treatment, trans-10, cis-12 CLA had no effect on IB␤ degradation.
To determine the specificity of CLA induced-NFB DNA binding among five NFB families (i.e. p50, p52, p65, c-Rel, Rel B) (for review, see Ref. 17), differentiated cultures of adipocytes were treated for 3 h with BSA vehicle or 30 M trans-10, cis-12 CLA. Subsequently, nuclear extracts were added to 96-well plates to which an oligonucleotide containing an NFB consensus binding site (5Ј-GGGACTTTCC-3Ј) had been immobilized as provided in the Trans-AM NFB Kit (Active Motif). Thereafter, the NFB complex bound to the oligonucleotide was detected by using an antibody that was specifically recognized by each NFB subunit. As shown in Fig. 4D, NFB p50-and p65-DNA binding were increased ϳ120 and 80%, respectively, in cultures treated with trans-10, cis-12 CLA compared with the BSA controls. Binding of DNA to the other NFB subunits due to CLA treatment was not detectable (data not shown). As shown in Fig. 4E, NFB p50 and p65 translocation from cytosol to nucleus were also increased in cultures treated with trans-10, cis-12 CLA compared with BSA-treated cultures. As expected, our positive control (100 ng/ml TNF-␣ for 1 h) increased nuclear accumulation of NFB p50 and p65. Furthermore, phosphorylation of nuclear NFB p65 (Ser-536) was higher in CLA-treated cultures compared with the BSA-treated cultures. Fractionation efficiency was verified using the nuclear protein nucleoporin and cytosolic protein GAPDH as markers. These data support our hypothesis that trans-10, cis-12 CLA promotes NFB activation and its translocation to the nucleus in differentiated cultures of human adipocytes.
These data raised the question as to the extent to which NFB activation occurs in adipocytes and SV cells. As our differentiated cell model is a heterogeneous model consisting of ϳ50% adipocytes and ϳ50% SV cells that do not differentiate into adipocytes, we used doubleimmunostaining of NFB p65 and aP2, which is expressed in adipocytes but not SV cells, to determine which cell type exhibited increased NFB activity in response to CLA. Supporting our data in Fig. 4E showing that trans-10, cis-12 CLA induced NFB translocation to nucleus, more NFB p65 (red) staining was found in nuclei of CLA-treated cultures, whereas most NFB p65 staining in the BSA controls was found in the cytosol (Fig. 5). The staining pattern of the TNF-␣-positive control was similar to that of the CLA treatment. As seen in the merged image in Fig.  5, nuclear NFB p65 appeared to be localized in both SV cells (e.g. aP2 negative cells staining red) and adipocytes (e.g. aP2 positive cells staining yellow) in CLA-treated cultures. These data suggest that trans-10, cis-12 CLA induces NFB activation both in human adipocytes and SV cells.
Inhibitors of NFB Block Trans-10, Cis-12 CLA-Induced IL-6 and TNF-␣ Gene Expression-To determine the extent to which CLA-induced cytokine gene expression is dependent on the activation of MEK/ ERK, GPCR, and/or NFB, the effects of selective chemical inhibitors of MEK/ERK, GPCR, and NFB activation on cytokine mRNA levels were investigated. The MEK/ERK inhibitor U0126 blocked CLA-induced IL-6 gene expression but not CLA induction of TNF-␣ (Fig. 6). Similarly, the GPCR-G i/o inhibitor pertussis toxin attenuated CLA-induced IL-6 gene expression without affecting CLA-induced TNF-␣ gene expression. Chemical inhibition of NFB activation was performed using 1) proteasome inhibitor I, which blocks proteasomal degradation of IB␣, 2) Bay11-7082, which inhibits IB␣ phosphorylation (29,30), and 3) kamebakaurin, which prevents NFB p50 DNA binding (Fig. 6). Chemical inhibition of NFB activation abolished CLA-inducible IL-6 and TNF-␣ gene expression. However, pretreatment with antioxidant N-acetylcysteine did not block CLA induction of either TNF-␣ or IL-6 (data not shown), suggesting CLA does not activate NFB or induce cytokines by producing prooxidants. These data show that CLA-induced IL-6 expression is dependent on both NFB and ERK1/2 activation. In contrast, CLA-induced TNF-␣ gene expression is dependent on NFB activation but not on the activation of ERK1/2 or GPCR.
Blocking IKK Complex Formation Reverses CLA Suppression of Glut4 and PPAR␥ Protein Levels-Given the reported antagonistic interaction between PPAR␥ and NFB (31), we investigated the role of NFB in mediating trans-10, cis-12 CLA suppression of insulin-stimulated glucose uptake via down-regulation of PPAR␥. Differentiated cultures of adipocytes were treated with either BSA vehicle or 30 M trans-10, cis-12 CLA for 24 h in the presence and absence of NEMO BP, a syn- thetic peptide that blocks the regulatory site of the active IKK complex, thereby inhibiting IB␣ degradation (32,33). As shown in Fig. 7, NEMO BP prevented or attenuated CLA-mediated suppression of PPAR␥ and Glut4 protein levels, respectively, without altering caveolin-1 protein expression. Collectively, these data demonstrate that trans-10, cis-12 CLA promotes NFB activation via the IKK-IB-NFB axis, thereby repressing the expression of PPAR␥ and its target genes that are required for insulin-stimulated glucose uptake and TG synthesis.
Depletion of NFB p65 by RNA Interference Attenuates CLA Suppression of Adipogenic Protein Expression and Glucose Uptake and Activation of ERK1/2-We demonstrated that the perturbation of NFB activation was sufficient to prevent or attenuate CLA suppression of adipogenic protein expression (Fig. 7). Using a second approach, selec-tive depletion of NFB p65 using siRNA was performed before treatment with trans-10, cis-12 CLA. As shown in Fig. 8A, control (nontargeting siRNA) or NFB p65 siRNA were introduced to the cultures of differentiating adipocytes on day 6. 72 h post-transfection, cultures were serum-starved for 20 h and treated with either BSA control or 30 M trans-10, cis-12 CLA for 24 h (Fig. 8A). To monitor the transfection efficiency, cultures were transfected with siGLO-RISC-free, a fluorescent labeled, non-targeting siRNA with impaired ability for RISC formation. Almost all cells (Ն90%) were positive to Cy3 fluorescence of siGLO 24 h post-transfection (Fig. 8B), indicating that both adipocytes and non-adipocytes were efficiently transfected (Fig. 8B). Specific depletion of NFB p65 was examined using immunoblotting as shown in Fig.  8C. We consistently obtained Ն50% NFB depletion. NFB p65 protein expression was severely blunted in the NFB p65 siRNA-transfected cultures compared with untreated or non-targeting control or lamin-siRNA-transfected cultures (Fig. 8C). In contrast, protein levels of 1) NFB p50, a heterodimeric partner of active NFB, 2) GAPDH, a constitutive cytoplasmic protein, 3) ␤-actin, a cytoskeleton protein, and 4) aP2, a specific adipocyte marker protein, were unchanged. siRNA-mediated specific knock-down was validated using lamin siRNA as a positive control (Fig. 8C).
To investigate the impact of CLA regulation on adipocytes under conditions of NFB depletion, we examined PPAR␥ expression and the activation of ERK1/2 and STAT3. As shown in Fig. 8D, depletion of NFB p65 modestly attenuated CLA suppression of PPAR␥2 (also confirmed by immunostaining in Fig. 8E) and its activation of ERK1/2 but had no marked effect on STAT3 phosphorylation. These data further support our hypothesis that NFB p65 plays a central role in CLAinduced suppression of adipogenic gene expression, activation of MEK/ ERK signaling, and subsequent secretion of cytokines.
Because CLA suppression of PPAR␥ was attenuated by knocking down NFB activity using siRNA, we hypothesized that silencing NFB would block or attenuate CLA suppression of insulin-stimulated glucose uptake demonstrated in Fig. 1. To test this hypothesis, we measured insulin-stimulated [2-3 H]deoxyglucose uptake in NFB-depleted cultures of adipocytes treated for 24 h with BSA vehicle or trans-10, cis-12 CLA. CLA suppression of glucose uptake was attenuated in siP65-transfected cultures compared with CLA-treated cultures not receiving siP65 (Fig. 9A). The degree of rescue of insulin-stimulated glucose uptake (ϳ45%) and Glut4 protein levels (ϳ55%) by siP65 for NFB in the CLA-treated group was nearly similar to the degree of knock-down achieved for NFB-p65 in this experiment (ϳ54%; densitometry not shown) (Fig. 9B). Collectively, these data demonstrate for the first time that specific depletion of NFB p65 in CLA-treated cultures partially rescued PPAR␥ and Glut4 protein levels and insulin-stimulated glucose uptake. This isomer-specific, CLA-mediated activation of cytokines and suppression of PPAR␥ and Glut4 protein levels shows striking similarity with other reports of cytokine-mediated insulin resistance reported in myotubes (11,28) and adipocytes (16).

DISCUSSION
We demonstrate in this article for the first time that trans-10, cis-12 CLA promotes NFB activation that induces cytokine production, leading to decreased glucose uptake in primary cultures of human adipocytes. Based upon these data and our previously published data (3,4,25), we propose in our working model (Fig. 10) that trans-10, cis-12 CLA first enters SV cells and activates a membrane protein or enters the cell by diffusion and is converted to a metabolite, which triggers a signal that activates NFB and ERK1/2, leading to cytokine synthesis, specifically TNF-␣, IL-6, and IL-8. These and possibly other cytokines activate their FIGURE 5. Trans-10, cis-12 CLA promotes NFB localization to the nucleus in adipocytes and non-adipocytes. Differentiated cultures of human adipocytes were treated with BSA vehicle or 30 M trans-10, cis-12 CLA for 24 h, and then cells were doublestained for NFB p65 (red, cy3 conjugated anti-mouse IgG) and aP2 (green, FITC-conjugated, anti-rabbit IgG). TNF-␣ (100 ng/ml for 1 h) treatment was used as a positive control. Results shown are representative of two separate experiments from pools of cells obtained from three to five different human subjects.  respective cell surface receptors on adipocytes, activating NFB and ERK1/2, leading to p65 and phosphorylated (P) ERK1/2 localization in the nucleus. P-ERK1/2 then phosphorylates specific nuclear transcription factors including p65, ELK-1, and PPAR␥. Together, these transcription factors acutely decrease PPAR␥ activity by phosphorylating PPAR␥ and/or by interfering with its ability to transactivate adipogenic target genes such as aP2, Glut4, fatty acid synthase, lipoprotein lipase, acetyl-CoA carboxylase, and stearoyl-CoA desaturase. Chronically, this leads to decreased PPAR␥ gene expression. Collectively, this causes decreased expression of adipogenic genes that promote glucose and fatty acid uptake and synthesis to TG, leading to insulin resistance and delipidation.
Using either chemical inhibition or depletion of NFB, we demonstrated that NFB activation was essential for CLA-induced IL-6, IL-8, and TNF-␣ production, impaired adipogenic gene expression and glucose uptake, and activation of ERK1/2 signaling. Based on our working model (Fig. 10) and data showing that NFB is activated by CLA in both SV cells and adipocytes (Fig. 5), we propose a differential role of NFB FIGURE 8. Specific depletion of NFB p65 attenuates CLA suppression of PPAR␥ and activation of MEK/ERK signaling in cultures of primary human adipocytes. A, experimental design of transfection protocol with siRNA NFB p65 in primary cultures of human adipocytes. trt, treatment. B, transfection efficiency was evaluated by transfection with siGLO, fluorescence (Cy3)-tagged non-targeting siRNA. C, knock-down specificity was analyzed using immunoblotting. Total cell extracts from either untreated (Unt) samples or samples transfected with non-targeting control siRNA (siCon), NFB p65 siRNA (siP65), and positive control Lamin siRNA (siLa) were immunoblotted with the antibodies targeting NFB p65, lamin C, NFB p50, GAPDH, actin, and aP2. D, the impact of CLA on NFB p65-depleted cultures were examined using immunoblotting. 72 h post-transfection with either non-targeting control siRNA or siP65, cultures were treated with BSA vehicle or a 30 M trans-10, cis-12 CLA for an additional 24 h. Total cell extracts were immunoblotted with antibodies targeting NFB p65, PPAR␥, Glut4, P-signal transducer and activators of transcription 3 (STAT3), P-ERK1/2, total-STAT3, and GAPDH. The blots for PPAR␥ and GAPDH were quantified by densitometry, and the amount of PPAR␥ relative to GAPDH was expressed as a bar graph under the blot. E, cultures were transfected and 72 h later treated with either BSA or 30 M trans-10, cis-12 CLA for 24 h and then immunostained for PPAR␥. Hoechst staining was conducted to identify nuclei. activation in SV cells versus adipocytes. In SV cells, we postulate that CLA activates NFB by an unidentified upstream signal. Upon activation and nuclear localization, ERK1/2 phosphorylates NFB-p65, which then induces cytokine production, as we demonstrated in non-differentiated SV cells (Fig. 2B) and in differentiated cultures of adipocytes ( Fig.  2A). Quantitatively, the non-differentiated SV cells (Fig. 2B) produced 10-and 7-fold more IL-6 and IL-8, respectively, in response to 24 h of treatment with trans-10, cis-12 CLA than did the differentiated cultures of adipocytes ( Fig. 2A). In agreement with these data, Harkins et al. (34) showed that 3T3-L1 preadipocytes stimulated with lipopolysaccharide express more IL-6 than adipocytes. Consistent with these data, IL-6, IL-8, and TNF-␣ possess a nuclear factor B response element (21). Furthermore, 1) cytokine secretion was preceded by IB␣ degradation (Figs. 4, A and B), IKK phosphorylation (Fig. 4C), increased p50 and p65 nuclear translocation (Figs. 4E and 5) and binding to a consensus NFB oligomer (Fig. 4D) compared with controls, and 2) selective chemical inhibitors of NFB acutely blocked CLA-mediated increases in IL-6 and TNF-␣ gene expression (Fig. 6).
Once these cytokines are secreted into the media, we propose that they bind to their cognate receptors on adipocytes, activating NFB. Upon activation and nuclear localization, ERK1/2 phosphorylates NFB and other transcription factors, which together suppress PPAR␥ activity leading to decreased adipogenic gene expression, glucose and fatty acid uptake, and TG synthesis (4). Consistent with these data, cytokines secreted from adipose tissue regulate glucose and lipid metabolism locally and peripherally (14,16,36,37). Several cytokines have been reported to signal through NFB (i.e. TNF-␣ via TNF-␣ receptors) and MEK/ERK (i.e. IL-6 via gp130-Janus kinase (JAK), IL-8 via GPCR-G i/o ).
In support of our working model proposing that trans-10, cis-12 CLA induces the activation of NFB in adipocytes, we found that selective inhibition of NFB using chemical inhibitors (Fig. 7) or depletion of NFB (Fig. 8, 9) attenuated CLA suppression of PPAR␥ and Glut4 levels and glucose uptake and activation of ERK1/2 signaling. IL-6 is a major paracrine regulator whose expression is controlled by NFB. Certain saturated fatty acids such as palmitate at high concentrations have been reported to activate NFB and IL-6, resulting in systemic insulin resistance in adipocytes (16), myotubes (12,28), and hepatocytes (39). We demonstrated that trans-10, cis-12 CLA activates the IKK-IB-NFB axis (Figs. 4 and 5), and NFB inhibitors block or attenuate CLA-induced IL-6 gene expression (Fig. 6), Glut4 protein levels (Figs. 7 and 9), and suppression of glucose uptake (Fig. 9), suggesting that the production of IL-6 is due to the CLA-mediated activation of NFB. To support our hypothesis, we found that differentiated human adipocytes treated with IL-6 (20 ng/ml) for 48 h had decreased Glut4 and adiponectin gene expression compared with controls. 3 Furthermore, we previously demonstrated that neutralization of IL-6 blocked CLA activation of ERK1/2 (4), supporting the important role of IL-6 in mediating CLA anti-adipogenic actions.
Unlike IL-6, TNF-␣ secretion to the media was not observed in CLA-treated human adipocytes (data not shown) even though there  ; n ϭ 4) not sharing a common lowercase letter differ, p Ͻ 0.05. B, immunoblotting for NFB p65, Glut4, and GAPDH were carried out as described in Fig. 8. The blots for Glut4 and GAPDH were quantified by densitometry, and the amount of Glut4 relative to GAPDH was expressed as a bar graph under the blot.
FIGURE 10. Working model; trans-10, cis-12 CLA reduces glucose and fatty acid uptake and TG synthesis via activation of NFB and ERK1/2 signaling and cytokine production. Upon entry into SV cells, trans-10, cis-12 CLA generates an unidentified signal, thereby activating NFB and ERK1/2. NFB p65/p50 then translocates to the nucleus where it is activated by P-ERK. Activated NFB initiates the transcription of specific cytokines (i.e. TNF-␣, IL-6, IL-8). These secreted cytokines in turn activate their cognate cell surface receptors on adipocytes, leading to NFB and ERK1/2 activation. NFB p65/p50 translocates into the nucleus, where P-ERK activates NFB p65 and other transcription factors (TF), leading to suppression of PPAR␥ activity and target gene expression including aP2, sterol-CoA desaturase (SCD), lipoprotein lipase (LPL), fatty acid (FA) synthase (FAS), Glut4, and perilipin (PLIN). CLA may also generate an unidentified signal in adipocytes that activates NFB and ERK1/2, which further suppresses PPAR␥ target gene expression. Together this leads to adipocyte delipidation via decreased glucose and fatty acid uptake and TG synthesis.
was a transient increase in the mRNA levels of TNF-␣ (Fig. 3). However, TNF-␣ may undergo altered processing in the plasma membrane (mTNF-␣) rather than being released in its soluble form. Xu et al. (40) reported that mTNF-␣ is capable of exerting diverse biological functions through cell contact-dependent signaling rather than through a receptor-mediated mechanism. Because bypassing the TNF-␣ receptor may result in activation of different mechanisms compared with the classic downstream signaling of TNF-␣ receptor, we cannot rule out the possibility that TNF-␣ plays an important role in mediating CLA induction of insulin resistance and delipidation (4) in adipocytes.
The term adipokines has been used to identify cytokines or chemokines originated from adipocytes (e.g. IL-6, IL-8, TNF-␣, leptin, resistin, adiponectin). However, this term is misleading based on data suggesting that nonfat cells from adipose tissue are the major site for cytokine production rather than adipocytes (37,41,42). These data are consistent with ours, suggesting that human SV cells or non-adipocytes have a greater capacity for cytokine production than adipocytes, at least in response to CLA. Considering the intimate communication between non-adipocytes and adipocytes within adipose tissue, this may explain why our primary human adipocytes cultures, which consist of ϳ50% SV cells and ϳ50% adipocytes, respond differently to CLA treatment compared with 3T3-L1 adipocytes, which are almost exclusively adipocytes. Studies are under way in our laboratory examining the effects of CLA on purified cultures of adipocytes devoid of SV cells to accurately address this issue.
We reported that trans-10, cis-12 CLA decreased insulin-stimulated glucose uptake in human differentiating preadipocytes (3) and newly differentiated adipocytes (4). Herein, we demonstrated that this CLAmediated suppression of insulin-stimulated glucose uptake (Fig. 1A) was due primarily to decreased levels of Glut4 (Fig. 1C), particularly plasma membrane Glut4, an indicator of insulin sensitivity. CLA had no effects on phosphorylation of IRS-1 (Ser-307 or Tyr-891) or Akt (Ser-437) (Fig. 1B), suggesting CLA has no significant effects on insulin signal transduction per se. Our hypothesis that CLA impaired insulin-stimulated glucose uptake via activation of NFB was confirmed by our NFB p65 siRNA transfection study, demonstrating that depletion of NFB attenuates CLA suppression of glucose uptake (Fig. 9). To our knowledge this is the first published report documenting depletion of NFB p65 using an RNA interference technique in primary human adipocytes. In addition, there is evidence that phosphorylation of IKK is a potential inhibitor of insulin-stimulated glucose uptake (43,44). Based on these reports, our data showing that trans-10, cis-12 CLA activates IKK (Fig.  4C) reinforces our working model, demonstrating that CLA promotes the IKK-IB-NFB cascade, which is critical for suppression of PPAR␥ and Glut4 proteins (Fig. 7).
Collectively, these data and our previously published data (3,4,25) demonstrate that trans-10, cis-12 CLA impairs glucose and fatty acid uptake in adipocytes, which is important for de novo TG synthesis and adipocyte hypertrophy. These data support human (45,46) and animal (47) studies showing that trans-10, cis-12 CLA, but not cis-9, trans-11 CLA, reduces adiposity. This isomer-specific effect of CLA on suppressing adipocyte glucose and fatty acid uptake may contribute to the hyperglycemia and hyperinsulinemia observed in obese humans (46) and the insulin resistance (48,49) and/or lipodystrophy (50) observed in some animal models supplemented with CLA. However, future clinical trials examining the isomerspecific effects of CLA on human adipose tissue metabolism, gene expression, and cell signaling are needed to validate this theory.
The intricate relationship between PPAR␥, NFB, and mitogen-activated protein kinase is not yet fully understood. P-ERK1/2 is an impor-tant regulator of PPAR␥ as demonstrated by its inhibition of adipogenesis (51). Furthermore, cytokine expression suppressed PPAR␥ activity in mesenchymal stem cells, showing direct interaction between PPAR␥ and NFB (52). Similar reports demonstrating that activation of NFB (27,31,53) and mitogen-activated protein kinase (22,38,54) hinders PPAR␥ DNA binding affinity, or transcriptional activation provides a potential mechanism by which trans-10, cis-12 CLA suppresses the expression of PPAR␥ target genes before that of PPAR␥ itself, leading to insulin resistance and delipidation (4). In our primary cultures of human adipocytes, trans-10, cis-12 CLA treatment activated NFB and ERK1/2. Furthermore, our results showed that chemical inhibition of NFB or targeted depletion of NFB p65 by siRNA not only attenuated CLA suppression of PPAR␥2 (Figs. 7 and 8, D and E) and Glut4 (Figs. 7 and 9B) but also prevented CLA-mediated ERK1/2 phosphorylation (Fig. 8D). Based on these findings, we hypothesize that mutual interactions between NFB and ERK1/2 are required for CLA to inhibit PPAR␥ activity acutely and PPAR␥ expression chronically. Studies are currently under way in our laboratory to test this hypothesis.
CLA suppression of insulin sensitivity and TG accumulation (3,4,25) in human adipocytes could be due to de-differentiation or apoptosis or blocking new differentiation. However, 30 M trans-10, cis-12 CLA does not cause apoptosis in our cultures based on the following evidence. 1) NFB activation antagonizes apoptosis and promotes cells survival (19); 2) caspase 3 is not activated by 30 M CLA for up to 72 h (unpublished data 3 ); 3) Hoechst and 4,6-diamidino-2-phenylindole staining indicate similar numbers of nuclei and no differences in nuclear condensation or fragmentation in CLA-treated cultures compared with controls ( Fig. 8E and Ref. 4); 4) protein and RNA levels are not reduced after CLA treatment; 5) CLA reduces the TG content of adipocytes without reducing the number of adipocytes (4,25). Similarly, CLA does not appear to cause de-differentiation per se because of the following evidence. 1) Adipocytes still contain small lipid droplets after 3 weeks of treatment even though CLA suppresses adipogenic gene expression and protein levels and impairs glucose uptake within 24 h of treatment (4); 2) CLA increases adipose differentiation-related protein and leptin expression (4,25), suggesting that CLA-treated cultures still contain adipocytes; 3) Pref-1 gene expression, which is abundant in our nondifferentiated SV cells, is low in CLA-treated cultures containing adipocytes, 3 suggesting these cells do not de-differentiate back to preadipocytes. Finally, we do not observe new differentiation of adipocytes after ϳday 6 of differentiation. Instead, the increased TG content of the cultures after day 6 is due primarily to increased size of the lipid droplets within adipocytes. Thus, because our experiments began on ϳday 12 of differentiation, CLA most likely does not impair the differentiation of new adipocytes. Instead, we postulate that CLA is causing delipidation, primarily by blocking de novo TG synthesis (3,4) and to a lesser extent by increasing lipolysis (25).
In summary, our in vitro data demonstrate that a physiological level of trans-10, cis-12 CLA activates NFBand ERK1/2-dependent cytokine production, which together suppress PPAR␥ and Glut4 levels and lead to impaired glucose uptake. Studies are currently under way examining 1) how CLA regulates PPAR␥ and the expression of its target genes, 2) the specific signaling role of SV cells and adipocytes in mediating the TG-lowering actions of CLA, and 3) the CLA-induced, upstream signal that activates NFB and ERK1/2.