Contents lists available at ScienceDirect

Can J Diabetes xxx (2014) 1e8

Canadian Journal of Diabetesjournal homepage:

www.canadianjournalofdiabetes.com

Original Research

Macrophages Inhibit Insulin Signalling in Adipocytes: Roleof Inducible Nitric Oxide Synthase and Nitric Oxide

Alemu Fite MS, PhD a,*, Abdul Badi Abou-Samra MD, PhD a,b, Berhane Seyoum MD, MPHb

aWayne State University School of Medicine, Detroit, Michigan, USAbHamad Medical Corporation, Doha, Qatar

a r t i c l e i n f o

Article history:Received 18 December 2013Received in revised form7 February 2014Accepted 20 February 2014Available online xxx

Keywords:Akt/AS160glucose consumptioninflammationinsulin resistancemacrophage infiltration

Mots clés :Akt et AS160consommation de glucoseinflammationinsulinorésistanceinfiltration macrophagique

* Address for correspondence: Alemu Fite, MS, PhMichigan 48201, USA.

E-mail address: afite@med.wayne.edu

1499-2671/$ e see front matter � 2014 Canadian Diahttp://dx.doi.org/10.1016/j.jcjd.2014.02.023

a b s t r a c t

Objectives: The interaction of immune cells with adipocytes within the adipose tissues in obese personswith diabetes mellitus may play a role in insulin resistance. We examined in vitro whether nitric oxide(NO) and inducible nitric oxide synthase (iNOS) play a role in impaired insulin signalling in adipocytesexposed to activated macrophages.Methods: We used a co-culture system in which Raw264.7 macrophages were plated over differentiated,low passage 3T3-L1 cells (dif3T3) at a cell density ratio of 1:2. Inflammation was induced by a challengewith bacterial lipopolysaccharide.Results: Significantly (p<0.001) enhanced iNOS expression and NO synthesis was observed in activatedco-cultures. In the co-cultures as compared with Raw264.7 cells alone, iNOS protein was induced up to11-fold above background, and NO release was significantly (p<0.001) increased up to 2.8-fold.Co-culturing dif3T3 and Raw264.7 cells as compared to dif3T3 alone reduced insulin-induced Aktphosphorylation by 50% and AS160 phosphorylation by 42%. This was correlated with reduced glucoseconsumption when dif3T3 was exposed to 1,3-morpholinosydnonimine. Adiponectin, GLUT4 andAS160 mRNA were reduced by 4-fold, 5-fold and 2-fold, respectively, in co-cultures as compared todif3T3 alone. On the contrary, GLUT1 mRNA levels were increased by 2-fold in co-cultures as comparedto dif3T3. NG-monomethyl-L-arginine abolished NO production with modest reversal of Akt/AS160phosphorylation.Conclusions: This study demonstrated a potential association between iNOS/NO-mediated inflammationand insulin resistance.

� 2014 Canadian Diabetes Association

r é s u m é

Objectifs : L’interaction entre les cellules immunitaires et les adipocytes dans les tissus adipeux chez les per-sonnesobèses souffrantdediabète sucrépeut jouerun rôle sur l’insulinorésistance.Nousavonsexaminé invitrosi l’oxyde nitrique (NO) et l’oxyde nitrique synthase inductible (iNOS) joue un rôle dans l’altération de la sig-nalisation de l’insuline des adipocytes exposés à l’activation des macrophages.Méthodes : Nous avons utilisé un système de coculture dans lequel les macrophages Raw264.7 ont étéensemencés dans des plaques sur des cellules 3T3-L1 différenciées à faible passage (dif3T3) selon unratio de densité cellulaire de 1:2. L’inflammation a été induite par une stimulation avec le lipopolysac-charide bactérien.Résultats : L’amélioration significative (p < 0,001) de l’expression de l’iNOS et de la synthèse du NO aété observée dans les cocultures activées. Dans les cocultures comparativement aux cellules Raw264.7seules, la protéine iNOS a été induite jusqu’à 11 fois au-dessus de la normale, et la libération du NO aété significativement (p < 0,001) augmentée jusqu’à 2,8 fois. La coculture de la dif3T3 seule a réduit de50 % la phospholyration de l’Akt induite par l’insuline et de 42 % la phosphorylation de l’AS160. Cela aété corrélé à la réduction de la consommation de glucose lorsque la dif3T3 était exposée au3-morpholinosydnonimine (ou SIN-1). L’adiponectine, l’ARNm de GLUT4 et d’AS160 ont été respec-tivement réduits de 4 fois, 5 fois et 2 fois dans les cocultures comparativement à la dif3T3 seule. Àl’inverse, les concentrations de l’ARNm de GLUT1 ont été augmentées de 2 fois dans les cocultures

D, 421 East Canfield, Detroit,

betes Association

A. Fite et al. / Can J Diabetes xxx (2014) 1e82

comparativement à la dif3T3. La G-monométhyl-L-arginine a aboli la production de NO par lerenversement modeste de la phosphorylation de l’Akt et de l’AS160.Conclusions : Cette étude a démontré un lien potentiel entre l’inflammation médiée par l’iNOS et le NO, etl’insulinorésistance.

� 2014 Canadian Diabetes Association

Introduction

Obesity is widely considered a state of chronic low-gradeinflammation that contributes to insulin resistance and type 2diabetes (1), but there is no clear understanding of the precisecause of insulin resistance. Insulin resistance in obese persons isassociated with adipose tissue macrophages infiltration (2,3),which is speculated to be through nitric oxide (NO) generation,lipolysis and free fatty acids activation of TLR4 complex (4,5).Secretion of proinflammatory adipokines and chemokinesthrough lipolysis may further potentiate inflammation (6).Macrophages express inducible nitric oxide synthase (iNOS),which catalyzes the synthesis of NO; the latter may impairinsulin-stimulated PI3K/Akt activation (7e9) and play a role inobesity-related insulin resistance. Mice with deletion of iNOS areprotected from high fat diet-induced insulin resistance anddecreased Akt activation (10). Conversely, NO has been reported tostimulate glucose transport through GLUT4 translocation (11,12).Clinical studies show differential circulating and skeletalmuscle NO levels depending on the adiposity of type 2 diabetespatients (13).

We have recently observed that co-culturing adipocyteswith myocytes impairs insulin signalling through increasedinterleukin-6 expression and secretion (14). Here, we examinethe effects of co-culturing adipocytes with macrophages, iNOSexpression and NO release on insulin signalling and glucoseuptake. The 3T3-L1 cells are well-established alternative modelsystems, and when adequately differentiated (dif3T3) can providea physiological model mimicking adipocytes (15), and these lipid-laden adipocytes parallel the in vivo upregulation of transcriptionfactors and the formation of characteristic fat depots (16).Raw264.7 cells are also known to be similar to bone marrowmacrophages in expression of CD14 and F4/80 and bacteriallipopolysaccharide (LPS)einduced NO production and represent acommon point in the monocyte-macrophage differentiationpathway (17). Direct coculture of dif3T3 with Raw264.7 mayrepresent adipose tissue infiltration by macrophages. In thepresent study, we used the Raw264.7 and dif3T3 coculture modeland the pharmacological manipulation of NO production with3-morpholinosydnonimine (SIN-1) for NO generation orNG-monomethyl-L-arginine (LNMMA) for inhibition of iNOStogether with proinflammatory activation by LPS.

Methods

Cell culture design

Raw264.7 cell line and murine macrophage and 3T3-L1preadipocyte (pre3T3) cell lines, originally derived from murinefibroblasts, were purchased from the American Type CultureCollection (ATCC). Raw264.7 cells were grown and maintained inDulbecco’s Modified Eagle’s medium (DMEM) supplemented with1% penicillin/streptomycin (growth medium) and 10% fetal bovineserum (FBS). For use in experiments, cells were seeded in multiple-well plates and grown to confluence as recommended by ATCC.The pre3T3 cells were differentiated into dif3T3 according to apreviously described method (18). Briefly, 2 days postconfluence,

pre3T3 cells were induced in DMEM supplemented with 10% FBS,250 nmol/L dexamethasone, 25 nmol/L insulin and 0.5 mmol/Lisobutyl methylxanthine for 3 to 5 days. Seven days after inductionof differentiation, at which time more than 95% of cells wereobserved to show multiples of fat droplets, the dif3T3 cells wereused in co-culture with Raw264.7 cells. For co-culture experiments,the Raw264.7 cells were trypsinized and 5�105 cells were plated ineach well of 6-well plates containing approximately 1�106 dif3T3cells 7 days after induction of differentiation (i.e. cellular ratio of1:2, respectively). All cell culture medium and supplements werefrom Invitrogen (Carlsbad, CA, USA).

Protein extraction and immunoblotting

The antibody phospho-Akt (Ser 474) was purchased fromAbcam; iNOS, pAS160 and alpha-tubulin were all from CellSignaling (Beverly, MA). Chemical reagents LPS, SIN-1, LNMMAand wortmannin were from Sigma. The isozyme selective Akt1/2kinase inhibitor (iAkt) was from Sigma and used in cells at a doseof 5 mM. Cells treated with LPS, SIN-1, LNMMA and iAkt whereappropriate and control cells were harvested and washed withphosphate-buffered saline; total cell proteins were extracted byradio-immunoprecipitation assay buffer containing a proteaseinhibitor cocktail, antiphosphatases I and II (Sigma-Aldrich,St Louis, MO). Protein concentrationwasmeasured with BCA assayreagents (Thermo Scientific Pierce, Rockford, IL). Protein lysateswere mixed with sample buffer (20% glycerol, 4% sodium dodecylsulfate [SDS], 10% 2-mercaptoethanol, 0.05% bromophenol blueand 1.25 M Tris-HCl, pH 6.8). Cell lysate proteins from treated andcontrol cells were resolved using SDS-PAGE in 10% to 12% poly-acrylamide. For a given Western analysis, equivalent proteinloading controls and pre-stained molecular weight markers(Thermo Fisher, Rockford, IL) were included. After electrophoresis,the proteins were transferred onto PVDF membrane (Amersham,Pittsburgh, PA) in a buffer containing 25 mMTris-HCl, 192 mMglycine, 20% methanol and 0.01% SDS (pH 8.5) using Bio-Radtransblot semidry apparatus at 14V for 1.5 hours. Residualprotein-binding sites on the membranes were blocked by incu-bation for 1 hour in Tris-buffered saline and Tween 20 buffer (20mMTris-HCl, pH 7.6, 250 mM NaCl, 0.05% Tween-20) containing5% nonfat dry milk (Bio-Rad, Hercules, CA). The membranes wereincubated with primary antibody overnight at 4

�C. After washing

with Tris-buffered saline and Tween 20 buffer, a secondary antibody(anti-immunoglobulin G conjugated with horseradish peroxidase)was added for 1 hour. Finally, the protein bands were visualized byautoradiography using an enhanced chemiluminescence ThermoFisher detection system. Blot spot densities were adjusted for totalprotein loading indicated by alpha-tubulin.

RNA extraction and complementary DNA synthesis

Total RNA extraction and purification was carried out using aPromega kit following manufacturer’s instructions (Promega,Madison, WI). Briefly, the total RNA extraction was processed byadsorbing total nucleic acids to the column matrix; adsorbedDNA was hydrolyzed with DNase enzymatic digestion buffer andremoved by further washing. The purified RNA was eluted into

A. Fite et al. / Can J Diabetes xxx (2014) 1e8 3

nuclease-free water and stored at �80�C. The RNA concentrationand purity were determined using optical density ratios at260 nm and 280 nm. The mRNA was converted into comple-mentary DNA (cDNA) using the Promega first strand reversetranscription kit in a reaction mixture consisting of MgCl2, dNTPmixture, recombinant RNase inhibitor, Moloney murine leukemiavirus reverse transcriptase, and random hexamer primers. ThecDNA was diluted and either stored at �20�C or immediatelyused for polymerase chain reaction (PCR) analysis.

Primers design

Primer Premier for Windows version 5.0 (Premier BiosoftInternational, Palo Alto, CA, USA) was used for searching, align-ing, editing and handling primers. The primers were designed tospan 1 or more exon-exon junctions to avoid interference fromgenomic DNA as confirmed by single product melt-curves foreach primer set. The designed primers were synthesized atInvitrogen Life Technologies (Carlsbad, CA, USA). Primers usedwere beta actin; CGTTGACATCCGTAAAGAC forward primer,and GAAGGTGGACAGTGAGGC reverse primer; adiponectin50 CATTATGACGGCAGCACT forward, GCAGATGGAGGAGGACAreverse; GLUT4 50 CTGGCACTTCCACTGAAC 30 forward 50

GAGACTGATGCGCTCTAAC 30 reverse; AS160 50 CTGCTGGACAAAGAGGTAG 30 forward, 50 CATCATATCCGGTCGGTAC 30 reverse; andGLUT1 50 GTATCCTGTTGCCCTTCTG 30, 50 CAGGTCTCGGGTCACATCG 30.

Real-time PCR

Quantitative real-time PCR used the ABI StepOnePlus Systemdetection system (Applied Biosystems, New York) and the SYBRGreen fluorophore (SABiosciences, Valencia, CA). The PCR cyclecondition was 10 minutes at 95�C followed by 40 cycles at 94�C for15 seconds, 62�C for 30 seconds and 72�C for 30 seconds. Cyclethreshold (CT ) was the preset threshold fluorescence reached forquantitative analyses and normalization against the values for thehouse-keeping gene, beta-actin. Relative quantification wascomputed as follows: values of each gene and value of the house-keeping gene DCT ¼ CT (target gene) � the CT (housekeepinggene). Finally, the relative gene expression was transformed bycalculating as 2�ðSDCT�CDCT Þwhere SDCT � CDCT is the differencebetween the sample (SDCT ) and the control (CDCT ).

Glucose consumption

Glucose consumption by dif3T3 control and those treated withSIN-1, LNMMA and insulin was determined from treatmentmedium, DMEM (4.5 mM glucose) supplemented with 1%penicillin/streptomycin and 10% FBS. Test medium (50 mL) wastaken for glucose assay and calculated for net glucose consumptionaccording to the method described by Blake et al (19). Briefly,exposure media was appropriately diluted, mixed with glucoseassay reagents and optical density read at 550 nm.

Nitric oxide determination

Nitric oxide measurement was conducted in terms of the stableendproduct, nitrite (NO2�) using the standard Griess reagent assayprotocol (20). The Griess reagent consisted of 2% sulphanilamidein 5% O-phosphoric acid and 0.2% N-(1-naphthyl)-ethyl-enediaminedihyrochloride (Sigma) in distilled water. Test mediawere analyzed with the Griess reagent mixed in equal volumes,and absorbance readings were obtained at lambda of 540 nm.Finally, values were expressed in terms of fold change againstcontrols.

Statistical analysis

Statistical analysis was performed using IBM SPSS Statistics 20(IBM Corporation, Armonk, NY). One-way analysis of variance wasused to compare means of treatment groups applying Tukey’smultiple comparisons, considering p<0.05 as significant. Levels ofsignificance are indicated in the figures and text as appropriate.

Results

Expression of iNOS and insulin signalling

Figure 1 depicts a Western blot analysis for pAkt (pSer473),pAS160 (pThr642) and iNOS induction in dif3T3 and Raw264.7 cellscultured separately or together. Co-culture of Raw264.7 and dif3T3cells showed a significant reduction in Akt and AS160 phosphory-lations by 50% and 42%, respectively. Robust induction of iNOS wasobserved in the dif3T3/Raw264.7 co-culture. Visual observationand densitometric comparisons of immunoblots showed thatinsulin-induced pAkt S473 and pAS160 T642 were significantly(p<0.001) decreased in dif3T3/Raw264.7 co-cultures. In this study,we have not observed a detectable expression of pAkt/pAS160 by asmuch as 1�106 Raw264.7 cells, whereas iNOS induction requiredthe presence of Raw246.7 cells. However, a ratio of as little as 1:10of Raw264.7 cells to dif3T3 was sufficient to significantly increaseiNOS induction and decrease Akt phosphorylation (data not shown)in the co-cultures, whereas a ratio of 1:2 had a profound effect.Figure 2 depicts supporting evidence of exposure of dif3T3 toactivated Raw264.7 cells in the presence of studied factors (iAkt,SIN-1 and LNMMA). To show whether Raw264.7 cells elicit insulin-induced Akt activation, 2 different cell counts (105 and 106 cells)were assessed (Figure 2, R). No Akt phosphorylation was detectedin Raw264.7 cells. Treatment of dif3T3 cells with SIN-1 as well asdirect co-culturing resulted in decreased pAkt/pAS160. Treatmentwith LNMMA slightly restored pAkt/pAS160 in LPS activateddif3T3/Raw264.7 co-culture.

NO release and Akt signalling

In response to iNOS expression, robust NO secretion byRaw264.7/dif3T3 as a function of exposure time to LPS and celldensity was evident. Figure 3A through 3D shows NO production byRaw264.7 cells when co-cultured with dif3T3 cells (Figure 3A and3B), when treated with LPS (Figure 3A, 3B and 3D) and with SIN-1or LNMMA (Figure 3B). Raw264.7 cells at high seeding densityproduced measurable quantity of NO within 6 hours of culture(Figure 3D). At a lower seeding density, such as 100,000 cells/mL,NO release was detectable after 24 hours of incubation. For a givencell density treated with LPS, Raw264.7 cells produced significantlyhigher levels of NO in co-culture with dif3T3 (Figure 3A). Condi-tioned media from dif3T3 also enhanced NO production byRaw264.7 cells in a concentration-dependent manner (Figure 3C).Regardless of cell type or treatment, SIN-1 liberated high levelsof NO whereas LNMMA effectively suppressed LPS-induced NOproduction by Raw264.7 cells or their co-culture with dif3T3(Figure 3B). Spontaneous decomposition of SIN-1 forming a SIN-1cation-radical intermediate has been reported that seeminglyliberates an pH dependent NO (21). We have not measured theexact pH levels of the media, but NO release from SIN-1 is mainlyobserved during the first 6 hours in culture. Raw264.7 cells pre-treatment with LNMMA significantly suppressed LPS-induced NOproduction. Insulin-stimulated Akt and AS160 phosphorylationwasreduced in the co-cultures (Figures 1 and 2). Compared tountreated cocultures, SIN-1-treated cocultures showed a decreasedprotein phosphorylation, whereas LNMMA-treated cells showed aslightly increased protein phosphorylation.

Figure 1. Akt or AS160 phosphorylation is decreased in response to co-culturing Raw264.7 and differentiated 3T3-L1 (dif3T3) cells. The 3T3-L1 cells were grown to confluence anddifferentiated, as detailed inMethods. Seven days after differentiation, dif3T3 cells were co-culturedwith Raw264.7 at a 2:1 ratio, respectively. (A) Cells were challengedwith 1 mg/mLbacterial lipopolysaccharide (LPS [24 hours]), harvested for protein analysis using Western blotting and probed for phospho-Akt S473 (pAkt), phospho-AS160 T642 (pAS160) andinducible nitric oxide synthase (iNOS) using alpha-tubulin as a loading control. Densitometric valueswere pooled and corrected for endogenous protein control and background values.Densitometric comparisons of (B) pAkt S473, (C) pAS160 T642 and (D) iNOS for the co-culture between dif3T3 and Raw264.7 and monoculture controls are shown without (light-hatched bars) or with (dark-hatched bars) treatment of insulin. Mean�SEM values were computed for at least 3 independent experiments. *p<0.001 ANOVA, co-culture values vs.dif3T3 alone. Ins, insulin.

A. Fite et al. / Can J Diabetes xxx (2014) 1e84

Glucose consumption

To correlate alterations in Akt and AS160 phosphorylation withglucose consumption, we measured insulin-stimulated glucoseconsumption in dif3T3 cells exposed to the NO donor (Figure 3E).Baseline glucose consumption was significantly (p<0.001)decreased by 90% in dif3T3 cells exposed to SIN-1 (p¼0.029). Incontrast, baseline glucose consumption by dif3T3 was furtherdecreased in treatment with SIN-1 plus insulin (p<0.001).Regardless of baseline differences, insulin hadmore profound effectin the presence of either SIN-1 relative to untreated dif3T3 cells.This finding was consistent with the effect of SIN-1 or dif3T3/Raw264.7 co-culture on Akt and AS160 phosphorylation (Figure 2).

Macrophages downregulate adipocyte genes involved in insulinsignalling

The effect of co-culture of dif3T3 and Raw264.7 cells on geneexpressionwasexamined inRaw264.7 anddif3T3 cells co-cultured at1:2 ratio and compared with the individual cells cultured separately(Figure 4). Quantitative real-time PCR showed that the mRNAs ofadiponectin, GLUT4, GLUT1 and AS160 were more abundantd7500-fold, 5500-fold, 2.5-fold and 7.5-fold, respectivelydin dif3T3 cellsthan inRaw264.7 cells. The presenceofmacrophages in the coculturewith dif3T3 reduced mRNA expression of adiponectin, GLUT4 andAS160by 4-fold, 5-fold and 2-fold, respectively. GLUT1 transcriptwasincreased by 2-fold in the presence of Raw264.7 cells in coculturewith dif3T3. To minimize the dilution effect of the different cells, CTvalues were uniformly corrected for beta-actin as endogenous con-trol and Raw264.7 mRNA as sample control. Both cells were frommurine origin, and it is assumed that beta-actin is expressed in bothdif3T3 and Raw264.7 cells with equal copy numbers. The level ofbeta-actin amplicons from dif3T3 and Raw264.7 cells had compara-ble CT values (data not shown).

Discussion

A co-culture system was developed in our attempt to addressthe effects of paracrine interactions between adipocytes andimmune cells on activation of proteins involved in insulin sig-nalling and glucose consumption. Previously, we reported the linkof the direct cell to cell communications between myocytes andadipocytes as a possible explanation for obesity-related impairedinsulin signalling through defects in activation of Akt (14). Wehave further observed that co-culture of macrophages withadipocytes activates cells through NO synthesis. We found thatRaw264.7 macrophages when activated with LPS produce NOdepending on LPS concentration, exposure time and cell density.Raw264.7 cells were also activated when co-cultured with dif3T3cells or incubated with conditioned medium from dif3T3 cells.This observation is important as central obesity is followed byinfiltration and activation of macrophages, which may enhanceinflammatory response through production of NO. However,dif3T3 cells showed no NO production, neither to LPS exposurenor to conditionedmedium from Raw264.7 cells (data not shown).It is thus reasonable to assume that macrophages are the impor-tant source of NO in the Raw264.7/dif3T3 co-culture system usedin this study. However, precaution is needed not to conclude thatunidentified humeral factors from dif3T3 would not synergize theNO synthesis by Raw264.7 or vice versa.

In this study, we showed an inhibition of pAkt and pAS160 by co-culturing dif3T3 cells with Raw264.7 cells. Assuming pAkt/pAS160activation was exclusively from dif3T3, as much as 33% reduction ofthe protein activation would be accounted for by a physical dilutionof protein from the co-cultures. We believe that a significant sup-pression of protein phosphorylation resulted from interaction ofdif3T3 and Raw264.7 cells, suggesting the effect of inflammation-activated macrophage blunting of insulin action in adipocytes.Other investigators showed upregulation of macrophage-derived

Figure 2. Nitric oxide (NO) exposure had a modifying effect on phospho-Akt (pAkt) or phospho-AS160. The 3T3-L1 cells were grown to confluence and differentiated intoadipocytes (dif3T3), as detailed in Methods. Seven days after differentiation, dif3T3 was co-cultured with Raw264.7 at a 2:1 ratio, for 24 hours, pretreated for 2 hours with NO donor3-morpholinosydnonimine (SIN-1) or inducible nitric oxide synthase (iNOS) inhibitor (NG-methyl-L-arginine acetate [LNMMA]) and challenged with bacterial lipopolysaccharide(LPS) for a further 24 hours. Cell lysates were subjected to protein analysis using Western blot and probed against phospho-Akt (pAkt) or phospho-AS160 (pAS160) using alpha-tubulin as a loading control. Akt kinase inhibitor (iAkt), an isozyme selective inhibitor, was used as a positive control for suppressing Akt activation. No insulin-induced pAkt wasobserved in Raw264.7 cells (R) at 2 different cell densities, 105 cells per well (0.1 mn) and 106 per well (1 mn). The SIN-1 decreased and LNMMA increased pAkt/pAS160 inco-cultured (TþR) of dif3T3 (T) and Raw264.7 (R). Ins, insulin.

A. Fite et al. / Can J Diabetes xxx (2014) 1e8 5

proinflammatory cytokines and adipocyte-derived fatty acids tomediate inflammation, but it is not knownwhether that may inhibitinsulin signalling (22). Akt phosphorylation is a key enzyme modi-fication involved in insulin signalling, whereas AS160 is a key Aktsubstrate regulating GLUT4 translocation in insulin-sensitive 3T3-L1adipocytes (23) and L6 myoblasts (24). The AS160 links the impor-tant downstream insulin-signalling pathway, not yet well charac-terized, through Rabs, small G proteins that play critical roles invesicle formation, movement and fusion (23). The result of this studysuggests that overproduction of NOmay inhibit Akt phosphorylationat S473 and AS160 phosphorylation at T642 residues, and may serveas a potential mechanism for the inhibition of insulin signalling bymacrophage activation.

The role played by NO in insulin signalling is complexbecause of diverse reactive nitrogen intermediates produced under

physiological conditions. Although NO was reported to stimulateglucose consumption in some studies (11,25), high concentration ofNO can inhibit glucose consumption (26). Our data, at odds withother studies, showed an inhibition of glucose consumption in adi-pocytes. The discrepancy may be explained by the difference in NOconcentration, subjecting cells in the present study to a supra-physiological level of NO. It should also be noted that the highlyvariable oxidation states of nitrogen from reactive nitrogenintermediates might contribute to the variable effects. A highconcentration of NO synthesized by iNOS very rapidly reacts withsuperoxide free radicals to form a highly reactive and relativelymorestable anion peroxynitrite (ONOO-), which leads to DNA damage andmodification of proteins (27,28). Post-translational nitrosylation ofthiol groups inactivates certain enzymes, such as Akt and insulinreceptor substrate-1, the key insulin signalling proteins (29,30).

Figure 3. The co-culture of differentiated 3T3-L1 (dif3T3) and Raw264.7 cells potentiates nitric oxide (NO) synthesis and NO reduces glucose consumption by dif3T3.In a different set of experiments, dif3T3 cells were subjected (A and B) to co-culture with Raw264.7, (A, B and D) to bacterial lipopolysaccharide (LPS) and (B) to3-morpholinosydnonimine (SIN-1) and NG-methyl-L-arginine acetate (LNMMA). Co-cultures and monocultures at shown cell densities were 2 hours pretreated (B) with the NOdonor (5 mM SIN-1) and iNOS inhibitor (100 mM LNMMA) and challenged with 1 mg/mL LPS for a further 24 hours (A and B). (A, light bars ¼without LPS, dark barks ¼with LPS; B,light bars ¼ SIN-1, dark bars ¼ LNMMA.) (C) The NO release by confluent Raw264.7 cells was determined after treatment with serially diluted 24 hours dif3T3 conditioned media(CM). (D) The NO release into culture media after 24 hours of treatment was determined using the standard Griess reagent assay protocol. (E) The level of insulin-stimulatedglucose (Glc) consumption in the media by dif3T3 monoculture treated with SIN-1 was determined relative to untreated dif3T3 glucose consumption at baseline, using theglucose oxidase enzymatic method. Values are mean�SEM of at least 3 independent experiments. *p < 0.05, ANOVA vs. untreated dif3T3. xp < 0.05, ANOVA vs. untreated Raw264.7. Ins, insulin.

A. Fite et al. / Can J Diabetes xxx (2014) 1e86

In addition, we observed a pronounced inhibition of glucoseconsumption by SIN-1, more than the level indicated by the cor-responding Akt activation. This finding might indicate differentpathways for how NOmay inhibit or enhance glucose consumptionthat is noninsulin dependent. Our observation that GLUT1 mRNAis increased in co-culture, representing an insulin-independentglucose transport, supports the latter hypothesis. In contrast, adi-ponectin (insulin sensitizer), GLUT4 (insulin-dependent glucosetransporter) and AS160 (substrate linking insulin signalling andglucose transport) were all downregulated. If it is accurate thatactivated macrophages inhibit glucose transport through thepotentiated NO synthesis, this observation warrant further studiesof the noninsulin-dependent mechanism of NO action in glucosetransport.

The use of co-culture models is important to identify molecularmechanisms involved in the physiological interaction of inflam-matory cells infiltrating insulin target cells. The main limitation ofthis study is that these data alone would not imply causality.Future studies are certainly needed using ex vivo human tissuesand in vivo animal models and immunohistochemical tools toaddress the role played by NO on exact cross talk between mac-rophages in the context of the target adipose tissue. Identifyinghumoral factors released by macrophages is needed to elucidatethe mechanisms by which they alter adipocyte insulin resistance.Overall, the results of this study shed some light on the paracrinalinteractions betweenmacrophages and adipocytes leading to highlevels of NO synthesis, inhibition of insulin signalling and glucoseconsumption.

Figure 4. The macrophages downregulate adipocyte genes involved in insulin signalling. Differentiated 3T3-L1 (dif3T3) cells were co-cultured with Raw264.7 cells for 24 hours at acell density ratio of 1:2, respectively. Cells were harvested for mRNA extraction and cDNA synthesis. Quantitative real-time polymerase chain reactionwas performed. The thresholdcycle (CT) value of each gene was corrected for the housekeeping gene (beta-actin) and relative gene expression was computed for (A) adiponectin (AdipQ), (B) phospho-AS160(AS160), (C) GLUT4 and (D) GLUT1. Means are of at least 3 independent experiments. ANOVA shows the following multiple comparisons: ap<0.001, relative mRNA expressionby dif3T3 vs. Raw264.7; bp<0.001, relative gene expression by dif3T3/Raw264.7 co-culture vs. dif3T3 alone.

A. Fite et al. / Can J Diabetes xxx (2014) 1e8 7

Acknowledgements

This work was supported by the Department of InternalMedicine, Wayne State University, School of Medicine.

Author Disclosures

No conflicts of interest.

References

1. Blasi E, Mathieson BJ, Varesio L. Selective immortalization of murine macro-phages from fresh bone marrow by a raf/myc recombinant murine retrovirus.Nature 1985;318:667e70.

2. Ortega Martinez de Victoria E, Xu X, Koska J, et al. Macrophage content insubcutaneous adipose tissue. Diabetes 2009;58:385e93.

3. Furuhashi M, Fucho R, Gorgun CZ, et al. Adipocyte/macrophage fatty acid-binding proteins contribute to metabolic deterioration through actions inboth macrophages and adipocytes in mice. J Clin Invest 2008;118:2640e50.

4. Suganami T, Ogawa Y. Adipose tissue macrophages: Their role in adipose tissueremodeling. J Leukoc Biol 2010;88:33e9.

5. Sun K, Kusminski CM, Scherer PE. Adipose tissue remodeling and obesity. J ClinInvest 2011;121:2094e101.

6. Nielsen S, Guo Z, Johnson CM, et al. Splanchnic lipolysis in human obesity. J ClinInvest 2004;113:1582e8.

7. Perreault M, Marette A. Targeted disruption of inducible nitric oxide synthaseprotects against obesity-linked insulin resistance in muscle. Nat Med 2001;7:1138e43.

8. Pilon GV, Charbonneau A, White PJ, et al. Endotoxin mediated-iNOS inductioncauses insulin resistance via ONOO- induced tyrosine nitration of IRS-1 inskeletal muscle. PLoS ONE 2010;5:e15912.

9. Bedard S, Marcotte B, Marette A. Cytokines modulate glucose transport inskeletal muscle by inducing the expression of inducible nitric oxide synthase.Biochem J 1997;325:487e93.

10. Charbonneau A, Marette A. Inducible nitric oxide synthase induction underlieslipid-induced hepatic insulin resistance in mice: Potential role of tyrosinenitration of insulin signaling proteins. Diabetes 2010;59:861e71.

11. Tanaka T, Nakatani K, Morioka K, et al. Nitric oxide stimulates glucose transportthrough insulin-independent GLUT4 translocation in 3T3-L1 adipocytes. EurJ Endocrinol 2003;149:61e7.

12. Balon TW, Nadler JL. Evidence that nitric oxide increases glucose transport inskeletal muscle. J Appl Physiol 1997;82:359e63.

13. Krause M, Rodrigues-Krause J, O’Hagan C, et al. Differential nitric oxidelevels in the blood and skeletal muscle of type 2 diabetic subjects may beconsequence of adiposity: A preliminary study. Metabolism 2012;61:1528e37.

14. Seyoum B, Fite A, Abou-Samra AB. Effects of 3T3 adipocytes on interleukin-6expression and insulin signaling in L6 skeletal muscle cells. Biochem BiophysRes Commun 2011;410:13e8.

15. Green H, Kehinde O. An established preadipose cell line and its differentiationin culture II. Factors affecting the adipose conversion. Cell 1975;5:19e27.

16. Rosen ED, Walkey CJ, Puigserver P, et al. Transcriptional regulation of adipo-genesis. Genes Dev 2000;14:1293e307.

17. Berghaus LJ, Moore JN, Hurley DJ, et al. Innate immune responses of pri-mary murine macrophage-lineage cells and RAW 264.7 cells to ligands ofToll-like receptors 2, 3, and 4. Comp Immunol Microbiol Infect Dis 2010;33:443e54.

18. Green H, Meuth M. An established pre-adipose cell line and its differentiationin culture. Cell 1974;3:127e33.

19. Blake DA, McLean NV. A colorimetric assay for the measurement of D-glucoseconsumption by cultured cells. Anal Biochem 1989;177:156e60.

20. Mizutani A, Maki H, Torii Y, et al. Ascorbate-dependent enhancement ofnitric oxide formation in activated macrophages. Nitric Oxide 1998;2:235e41.

21. Shirai K, Okada T, Konishi K, et al. Bicarbonate plays a critical role in thegeneration of cytotoxicity during SIN-1 decomposition in culture medium.Oxid Med Cell Longev 2012;2012:15.

22. Suganami T, Nishida J, Ogawa Y. A paracrine loop between adipocytes andmacrophages aggravates inflammatory changes: Role of free fatty acidsand tumor necrosis factor a. Arterioscler Thromb Vasc Biol 2005;25:2062e8.

23. Sano H, Kane S, Sano E, et al. Insulin-stimulated phosphorylation of a RabGTPase-activating protein regulates GLUT4 translocation. J Biol Chem 2003;278:14599e602.

24. Thong FSL, Dugani CB, Klip A. Turning signals on and off: GLUT4 traffic in theinsulin-signaling highway. Physiology 2005;20:271e84.

25. Higaki Y, Hirshman MF, Fujii N, et al. Nitric oxide increases glucose uptakethrough a mechanism that is distinct from the insulin and contraction path-ways in rat skeletal muscle. Diabetes 2001;50:241e7.

A. Fite et al. / Can J Diabetes xxx (2014) 1e88

26. McGrowdera D, Ragoobirsingh D, Brown P. Modulation of glucose uptake inadipose tissue by nitric oxide-generating compounds. J Biosci 2006;31:347e54.

27. Forrester MT, Stamler JS. A classification scheme for redox-based modificationsof proteins. Am J Respir Cell Mol Biol 2007;36:135e7.

28. Szabo C, Ischiropoulos H, Radi R. Peroxynitrite: Biochemistry, pathophysiologyand development of therapeutics. Nat Rev Drug Discov 2007;6:662e80.

29. Sugita H, Fujimoto M, Yasukawa T, et al. Inducible nitric-oxide synthase and NOdonor induce insulin receptor substrate-1 degradation in skeletal muscle cells.J Biol Chem 2005;280:14203e11.

30. Yasukawa T, Tokunaga E, Ota H, et al. S-nitrosylation-dependent inacti-vation of Akt/protein kinase B in insulin resistance. J Biol Chem 2005;280:7511e8.