Marta Fabrizi,1 Valentina Marchetti,1 Maria Mavilio,1 Arianna Marino,1 Viviana Casagrande,1

Michele Cavalera,1 Josè Maria Moreno-Navarrete,2 Teresa Mezza,3 Gian Pio Sorice,3,4

Loredana Fiorentino,1 Rossella Menghini,1 Renato Lauro,1 Giovanni Monteleone,1 Andrea Giaccari,3,5

José Manuel Fernandez Real,2 and Massimo Federici1,6

IL-21 Is a Major NegativeRegulator of IRF4-DependentLipolysis Affecting Tregs inAdipose Tissue and SystemicInsulin SensitivityDiabetes 2014;63:2086–2096 | DOI: 10.2337/db13-0939

Obesity elicits immune cell infiltration of adipose tissueprovoking chronic low-grade inflammation. Regulatory Tcells (Tregs) are specifically reduced in adipose tissue ofobese animals. Since interleukin (IL)-21 plays an importantrole in inducing andmaintaining immune-mediated chronicinflammatory processes and negatively regulates Tregdifferentiation/activity, we hypothesized that it could playa role in obesity-induced insulin resistance. We found IL-21and IL-21R mRNA expression upregulated in adiposetissue of high-fat diet (HFD) wild-type (WT) mice and instromal vascular fraction from human obese subjects inparallel to macrophage and inflammatory markers. In-terestingly, a larger infiltration of Treg cells was seen in theadipose tissue of IL-21 knockout (KO) mice compared withWT animals fed both normal diet and HFD. In a context ofdiet-induced obesity, IL-21 KO mice, compared with WTanimals, exhibited lower body weight, improved insulinsensitivity, and decreased adipose and hepatic inflamma-tion. This metabolic phenotype is accompanied by a higherinduction of interferon regulatory factor 4 (IRF4), a tran-scriptional regulator of fasting lipolysis in adipose tissue.Our data suggest that IL-21 exerts negative regulation onIRF4 and Treg activity, developing andmaintaining adiposetissue inflammation in the obesity state.

Obesity-associated tissue inflammation is now recog-nized as a major cause of decreased insulin sensitivity(1,2). Obesity, insulin resistance, and type 2 diabetes areclosely associated with chronic inflammation character-ized by abnormal cytokine production, increased acute-phase reactants and other mediators, and activation ofa network of inflammatory signaling pathways (3,4). Ex-cessive triglyceride accumulation within adipocytes leadsto adipocyte hypertrophy and a dysregulation of adipo-kine secretory patterns. Adipocytes as well as cells of thestromal vascular fraction (SVF), including preadipocytes,fibroblasts, mesenchymal stem cells, and immune cells,contribute to the production of proinflammatory cytokinesin obesity (3–5), with a pivotal role played by macro-phages and T lymphocytes (6–8). In lean adipose tissue,T-helper (Th) type 2 cells produce anti-inflammatory cyto-kines such as interleukin (IL)-4, -10, and -13, which pro-mote alternative activated M2 macrophage polarization(9). M2 polarization is also induced by regulatory T cells(Tregs) and eosinophils via IL-4. Conversely, in obese ad-ipose tissue, investigators have observed an increase inthe number of Th1 cytokines, M1 polarized macrophages,mast cells, B cells, and CD8+ T cells, which contribute to

1Department of Systems Medicine, University of Rome “Tor Vergata,” Rome, Italy2University Department of Diabetes, Endocrinology and Nutrition, University Hos-pital of Girona “Dr. Josep Trueta,” Institut d’Investigació Biomédica de GironaIdibGi, and CIBER Fisiopatología de la Obesidad y Nutrición, Girona, Spain3Division of Endocrinology and Metabolic Diseases, Università Cattolica del SacroCuore, Rome, Italy4Diabetic Care Clinics, Associazione dei Cavalieri Italiani del Sovrano MilitareOrdine di Malta (ACI SMOM), Rome, Italy5Fondazione Don Gnocchi, Milan, Italy6Center for Atherosclerosis, Department of Medicine, Policlinico Tor Vergata,Rome, Italy

Corresponding author: Massimo Federici, federicm@uniroma2.it.

Received 15 June 2013 and accepted 8 January 2014.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db13-0939/-/DC1.

© 2014 by the American Diabetes Association. See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.

See accompanying article, p. 1838.

2086 Diabetes Volume 63, June 2014

PATHOPHYSIO

LOGY

insulin resistance and promote macrophage M1 accumula-tion and proinflammatory gene expression (9–13). Factorsorchestrating the switch between M1 and M2 are still un-defined. Loss of interferon regulatory factor 4 (IRF4) spe-cifically in the myeloid cells evoked a constitutive M1polarization in the adipose tissue, suggesting that IRF4 isa negative regulator of inflammation in diet-induced obesity,in part through regulation of macrophage polarization (14).Interestingly, IRF4 expression is nutritionally regulated bythe actions of insulin and FoxO1, playing a significant rolein the transcriptional regulation of lipid handling in adipo-cytes, promoting lipolysis, at least in part by inducing theexpression of the lipases adipose triglyceride lipase (ATGL;Pnpla2) and hormone-sensitive lipase (HSL; Lipe) (15).

Naturally occurring Treg cells are a unique subpopula-tion of CD4+ T cells specifically adapted to the suppres-sion of aberrant or excessive immune responses that areharmful to the host (16). Treg cells are abundant in visceraladipose tissue (VAT) and have a different T-cell receptorrepertoire compared with Treg cells in other tissues, sug-gesting that they might be activated via the recognition ofa fat tissue–specific antigen (13). A recent study reveals animportant role for VAT-specific natural Treg cells in thesuppression of obesity-associated inflammation in adiposetissue and consequently in reducing insulin resistance (17).The number of VAT Treg cells is strikingly and specificallyreduced in insulin-resistant models of obesity, and the cellsare characterized by the expression of the transcription fac-tor Foxp3 and the nuclear receptor peroxisome proliferator–activated receptor (PPAR)-g (17).

IL-21 is a member of the type-I cytokine family and issynthesized by a range of CD4+ Th cells, including Th1and Th17 cells, activated NKT cells, and T follicular helpercells (18–20). IL-21 biological functions are mediated viathe IL-21 receptor (IL-21R) and after activation of theJanus kinase (JAK) family protein tyrosine kinases JAK1and JAK3 and, subsequently, the activation of Stat1, Stat3and to a lesser degree Stat4, Stat5, and Stat6 (21–23).IL-21 expression in T cells can be regulated by IL-21 viaan autocrine positive-feedback loop, involving the activa-tion of STAT3 (24). This feedback loop is essential for thedevelopment of Th17 cells (25,26). IL-21–mediated T-cellactivation relies partly on its ability to inhibit the differen-tiation of inducible Tregs and to make T cells resistant tothe Treg-mediated immunosuppression (27,28).

Because IL-21 is known to exert negative effects onTreg activity, we hypothesized that it could play a role inobesity-induced insulin resistance.

RESEARCH DESIGN AND METHODS

Mouse Models and Metabolic AnalysisWild-type (WT) and IL-21 knockout (KO) (129S5-Il21tm1Lex)male mice, both on the same genetic background(C57BL/6J), were purchased from Lexicon Genetics, Inc.IL-21 KO mice are viable and do not exhibit any pheno-type. Mice were maintained in standard animal cagesunder specific pathogen–free conditions in the animal

facility at the University of Rome “Tor Vergata.” Micewere maintained under a strict 12-h light cycle (lights onat 7:00 A.M. and off at 7:00 P.M.), genotyped, and divided inseparate cages at the beginning of each experiment. For thediet-induced obesity model, individually caged mice fromall groups were fed a high-fat diet (HFD) (60% of caloriesfrom fat; Research Diets, New Brunswick, NJ) or normaldiet (ND) (10% calories from fat, GLP; Mucedola S.r.l.,Settimo Milanese, Italy) for 18 weeks after weaning as in-dicated. Metabolic testing procedures were performed aspreviously described (29,30).

Hormone and metabolite levels were measured usingcommercial kits: insulin (Mercodia), nonesterified fattyacid (NEFA) (Wako), glycerol (Sigma-Aldrich, St. Louis,MO), and glucagon (Uscn Life Science, Inc).

Evaluation of Peripheral Insulin Sensitivity (Clamp)After 12 weeks of HFD, we evaluated peripheral insulinsensitivity by the euglycemic-hyperinsulinemic clamp tech-nique. Surgery for the positioning of catheters wasperformed 3–5 days prior to the insulin clamp procedureas previously described (31,32), and then mice were housedin individual cages. The euglycemic-hyperinsulinemic clampwas performed in the awake state after a 6-h fast. At timezero, a primed continuous (18.0 mU $ kg21 $ min21,Actrapid 100 IU/mL; Novo Nordisk, Copenhagen, Den-mark) infusion of human insulin was started simulta-neously with a variable infusion of 20% dextrose in orderto maintain the plasma glucose concentration constant atits basal level (80–100 mg/dL). Fasting plasma glucose wasmeasured at time 0. Subsequently, blood samples (;2 mL)were taken from the tail vein at 10-min intervals for at least2 h to measure glucose concentration and adjust dextroseinfusion rates. Insulin sensitivity (rate of peripheral glucoseuptake [mg$kg21$min21]) was calculated from averageglucose concentrations and dextrose infusion rates duringthe last 30 min of the steady-state clamp period.

Analysis of Adipose and Hepatic TissueEpigonadal fat and liver were obtained from WT and IL-21KO mice; specimens were fixed in 10% paraformaldehydeand embedded in paraffin. Ten-micrometer consecutivesections were then mounted on slides and stained withhematoxylin-eosin. Adipose cell size and density werecalculated as previously described (33).

Isolation of Adipocytes and SVFVAT was subjected to collagenase digestion (1 mg/mLcollagenase type 1; Sigma-Aldrich) in Krebs-Ringer buffer,with shaking at 180 rpm for 30 min at 37°C. After diges-tion, adipocytes were allowed to separate by flotation andthe infranatant solution was centrifuged for 5 min at 300gto pellet the SVF. The adipocyte fraction was washed threetimes with the Krebs-Ringer buffer. Subsequently, RNA wasisolated from adipocytes and the SVF fractions and analyzedby real-time PCR. The profile of adiponectin mRNA expres-sion was used to test the purity of the isolated fractions.The SVF was analyzed by flow cytometry techniques.

diabetes.diabetesjournals.org Fabrizi and Associates 2087

Cell Culture3T3-L1 cells (American Type Culture Collection) werecultured in Dulbecco’s modified Eagle’s medium (DMEM)(Invitrogen) with 10% bovine calf serum (Invitrogen) in 5%CO2. Two days postconfluence, cells were exposed toDMEM 10% FBS (Invitrogen) with 1 mmol/L dexametha-sone (Sigma), 5 mg/mL insulin (Sigma), and 0.5 mmol/Lisobutylmethylxanthine (Sigma). After 2 days, cells weremaintained in medium containing FBS only. For IRF4 reg-ulation experiments, fully differentiated 3T3-L1 adipocyteswere incubated in serum-free DMEM containing 1% fattyacid–free BSA (Sigma) with isoproterenol (10 mmol/L;Sigma) and IL-21 (100 ng/mL; R&D) at the doses and timesindicated.

Western BlotPreparation of tissue lysates, quantification, and immuno-blot analysis were performed as previously described (33).

Antibodies to IRF4, actin, and total FoxO1 (Santa CruzBiotechnology), phospho-Ser473 Akt, total Akt, phospho-Ser256 FoxO1 (Cell Signaling Technology) were used.

Gene Expression Analysis by qRT-PCRTotal RNA was isolated and gene expression analysis wasperformed as previously described (34).

Flow Cytometry AnalysisCells from the SVF of adipose tissue were stained forsurface antigens CD4 and CD25 and for the intracellularFoxp3+ transcription factor by using a mouse regulatoryT-cell detection kit as directed by the manufacturer’s in-structions (Miltenyi Biotec, Bergisch Gladbach, Germany).For intracellular lipids, cells were stained with Nile red(1 mg/mL; Sigma-Aldrich). Samples were analyzed using aFACSCalibur flow cytometer (Becton Dickinson, Heidelberg,Germany) and FlowJo software.

Figure 1—Tregs are increased in SVF of IL-21 KO mice, and IL-21/IL-21R genes are increased in obese adipose tissue. A: Cells fromepigonadal fat SVF were stained and analyzed by flow cytometry. Tregs are defined as CD25+CD4+Foxp3+. Tregs from WT and IL-21 KOmice. Left: Representative dot plots. Right: Summary data. Numbers on dot plots indicate the percentage of cells in that gate for thatparticular experiment. (n = 4 per group, *P < 0.05, Student t test. Error bars represent the mean 6 SD.) B: mRNA expression of IL-21 andIL-21R in epigonadal adipose tissue of WT mice fed ND or HFD. (**P < 0.01.) C: mRNA expression of adiponectin and (consideredadipocyte markers) IL-21 and IL-21R in adipocyte fraction (AF) and SVF from epigonadal adipose tissue of WT mice fed ND or HFD.(n = 6 per group, *P < 0.05, Student t test. Error bars represent the mean 6 SD.) D: mRNA expression of IL-21 and IL-21R in 3T3-L1 cellsunder basal and starving condition. (n = 3 per group, *P < 0.05, Student t test. Error bars represent the mean 6 SD.) a.u., arbitrary units.

2088 IL-21 and IRF4-Dependent Lipolysis Diabetes Volume 63, June 2014

Human StudyA total of 207 adipose tissue samples (112 visceral and 95subcutaneous) were collected at the Endocrinology Serviceof Hospital Universitari of Girona “Dr. Josep Trueta” froma group of Caucasian subjects with BMI between 20 and 58kg/m2. All subjects reviewed that their body weight hadbeen stable for at least 3 months before the study andgave written informed consent after the purpose, nature,and potential risks of the study were explained to them.

Adipose tissue samples were obtained from subcutane-ous and visceral depots during elective surgical procedures(cholecystectomy, surgery of abdominal hernia, and gastricbypass surgery), washed, fragmented, and immediatelyflash frozen in liquid nitrogen before being stored at280°C and used for gene expression analysis.

Statistical AnalysisResults of the experimental studies are expressed asmeans 6 SD. Statistical analyses were performed usingthe unpaired Student t test as indicated. Values of P ,0.05 were considered statistically significant.

RESULTS

IL-21/IL-21R and Treg Cells Are Increased in ObeseAdipose TissueA recent study demonstrated that Treg cells with a uniquephenotype were highly enriched in the abdominal fat of leanmice but were strikingly and specifically reduced at this sitein insulin-resistant models of obesity (13). It has also beenreported that IL-21 can counteract the immune-suppressiveproperties of Tregs in several types of tissues both in vitroand in vivo (27,35). In order to determine whether this IL-21effect on Treg cells could be extended to those residing in theadipose tissue, we firstly studied the amount of Tregs,marked as CD4+CD25+Foxp3+, in the SVF of IL-21 KOmice, finding a significant increase compared with WT litter-mates (Fig. 1A). Next, in visceral (epigonadal) adipose tissueof WT mice fed an HFD for 16 weeks compared with WT fedan ND, we found a significant increase of both IL-21 and IL-21R mRNA expression (Fig. 1B). More in detail, we observedan augment of IL-21 and IL-21R expression both in adipo-cyte fraction and SVF, although this was not significant

Figure 2—Metabolic effect of IL-21 deficiency on diet-induced obesity. At 6 weeks of age, WT and IL-21 KO mice were fed an ND or HFD for18 weeks. A: Body weight curves and blood glucose level of WT and IL-21 KO mice fed ND in the fasting conditions. B: Fasting body weightcurves and fasting glucose levels during HFD. C: IPGTT, intraperitoneal insulin tolerance test (IPITT), and relative area under the curve (AUC).(n = 7 per group, *P < 0.05, **P < 0.01, ***P < 0.001, Student t test.) D: Insulin levels measured during IPGTT. (n = 5 per group, *P < 0.001,Student t test. Error bars represent the mean6 SD.) E: Glucose uptake in WT and IL-21 KOmice fed HFD. Glucose infusion rates (GIR) duringthe euglycemic-hyperinsulinemic clamp. (n = 4 WT vs. 6 KO, **P < 0.0025, Student t test. Error bars represent the mean 6 SD.)

diabetes.diabetesjournals.org Fabrizi and Associates 2089

(Fig. 1C). We also found IL-21 and IL-21R gene expression infully differentiated 3T3-L1 adipocytes (Fig. 1D). This sug-gested an association between increased IL-21/IL-21R signal-ing and the progression of obesity.

Metabolic Effect of Diet-Induced Obesity on IL-21 KOMiceNext, in order to understand the effects of this cytokine ondiet-induced obesity, we conducted a complete metaboliccharacterization of IL-21 KO mice under ND and HFDconditions. IL-21 KO mice fed ND did not show differencesin body weight or fasting plasma glucose levels comparedwith WT mice (Fig. 2A). We fed 6- to 7-week-old WT andIL-21 KO mice in a context of HFD for 18 weeks. Fastingand fed body weight and fasting glycemia were comparableat the beginning of treatment, but their curves significantlydiverged from week 5 to the end of our observation at week18 (Fig. 2B). Intraperitoneal glucose tolerance test (IPGTT),intraperitoneal insulin tolerance test, and serum insulinlevels suggested that metabolic control was improved, on

an HFD, by IL-21 deficiency (Fig. 2C and D). The relieffrom diet-induced insulin resistance was finally confirmedthrough the measurement of peripheral (skeletal muscle)insulin sensitivity by the euglycemic-hyperinsulinemicclamp (Fig. 2E).

Effect of IL-21 KO on Adipose Tissue Morphology andFunction During Diet-Induced ObesityAfterward, we conducted a morphological and molecularcharacterization of adipose tissue in order to betterunderstand the mechanism by which IL-21 deficiencyprotects from metabolic injury caused by diet-inducedobesity. IL-21 KO mice fed an ND did not show sig-nificant differences in adipose tissue morphometry com-pared with WT mice littermates (Supplementary Fig. 1A).On the other hand, the IL-21 KO–reduced body weight,observed with HFD, was characterized by lower adiposityassociated with decreased fat pad mass (Fig. 3A), reducedadipocyte size, and a higher density of smaller adipocytes(Fig. 3B).

Figure 3—Adipose tissue structure and molecular characterization in IL-21 KO during diet-induced obesity. Epigonadal adipose tissue ofIL-21 KO and WT mice after 18 weeks of HFD. A: Representative images of WT and IL-21 KO mice and respective fat pad after 18 weeksof HFD. B: Representative sections of adipose tissue stained with hematoxylin-eosin, mean adipocyte area, and frequency distribution ofadipocyte area in IL-21 KO mice vs. WT littermates. (n = 6 per group, ***P < 0.001, Student t test. Data are means 6 SD.) C: WATexpression of genes involved in inflammation, metabolism, and mitochondrial biogenesis. Expression of mRNA was determined by real-time PCR and normalized to b-actin. (n = 6 per group, *P< 0.05 and **P< 0.01, Student t test. Error bars represent the mean6 SD.) D: Aktphosphorylation (p) in refeeding condition was analyzed by Western blot. (n = 6 per group, *P< 0.05, Student t test. Data are means6 SD.)A representative image of four mice per group is shown. a.u., arbitrary units.

2090 IL-21 and IRF4-Dependent Lipolysis Diabetes Volume 63, June 2014

Gene expression analysis of adipose tissue revealedsignificantly reduced levels of macrophage markers suchas F4/80 and CD68 in IL-21 KO mice. This may indicatea low grade of infiltration of proinflammatory macro-phages. On the other hand, we have found increased levelsof YM1 and Mgl2 mRNA, suggesting a high presence ofalternatively activated macrophages M2 (Fig. 3C). Next, weanalyzed genes involved in the regulation of glucose/lipidmetabolism and mitochondrial function, finding signifi-cantly increased levels of adiponectin, FoxO1, SOCS3,Sirt1, ERRa, and Nrf1 (Fig. 3C). The improved metabolicstate of IL-21 KO mice was also supported by increasedphosphorylation of Ser473 Akt in the refeeding condition(Fig. 3D).

Expression of IRF4 in Adipose Tissue From IL-21 KOand WT MiceIn adipose tissue, fasting induces IRF4-dependent lypol-isis, and insulin, during refeeding, inhibits its expres-sion via AKT/FoxO1. In adipose tissue of IL-21 KO mice,despite higher Akt phosphorylation (Fig. 3D), we observed

significantly increased expression of IRF4 in the refeedingstate at both mRNA and protein levels (Fig. 4A). In thefasting state, we found increased expression of IRF4 onlyin adipose tissue from IL-21 KO mice fed an HFD com-pared with WT. Expression of IRF4 targets pnpla2 and lipeconfirmed increased expression of both lipolytic genes, par-ticularly in the refeeding state (Fig. 4A). To control thatthis nutritional effect was specific for adipose tissue, wemeasured IRF4 expression in spleen from the same mice,finding no differences (Fig. 4C). Analysis of NEFA and glyc-erol in fasting sera confirmed a trend to increased lipolysisin IL-21 KO compared with WT littermates during both NDand HFD (Fig. 4D).

Effect of IL-21 on IRF4 Expression in 3T3-L1Adipocytes and in SVFsIRF4 mRNA expression rose significantly in 3T3-L1adipocytes treated with isoproterenol for 2 h and decreasedwhen adipocytes were pretreated with IL-21. Accordingly,we found decreased mRNA levels of IRF4 targets when thetreatment was extended to 4 h (Fig. 5A).

Figure 4—Lipolysis regulation in IL-21 KO adipose tissue. A: Expression of IRF4, Lipe, and Pnpla2mRNA in fasted and refed WT and IL-21fed ND or HFD. (*P < 0.05, n = 6 per group.) B: Protein expression of IRF4 in IL-21 KO in both fasting and refeeding conditions in ND andHFD mice. (*P < 0.05, **P < 0.01.) C: Spleen mRNA expression of IRF4 in ND and HFD mice. (n = 6 per group.) D: Fasting NEFA andglycerol serum levels. (*P < 0.05, **P < 0.01. n = 6 per group.)

diabetes.diabetesjournals.org Fabrizi and Associates 2091

A recent study demonstrates that IRF4 promotes M2polarization of adipose tissue macrophages (14). In SVFsfrom IL-21 KO adipose tissue, we found significant in-creased IRF4 and M2, markers of mRNA expression(Fig. 5B). M2 macrophages and Treg cells are known toprevalently use fatty acids for ATP generation to maintaintheir functions (36). Recently, PPAR-g was highlightedas a crucial molecular orchestrator of VAT Tregs andM2 macrophage accumulation, phenotypes, and functions(17). We found increased levels of PPAR-g mRNA in SVFof IL-21 KO HFD compared with WT. The profile of adi-ponectin mRNA expression provides evidence of the pu-rity of the fraction preparations (Fig. 5B).

Effect of IL-21 Deficiency on VAT Tregs During HFDTo determine whether the degree of infiltration of Tregsin the adipose tissue in our KO model could be influencedby a treatment inducing obesity treatment, we quantifiedby flow cytometry the number of Treg cells in SVFs of thetwo experimental groups at the end of 18 weeks of HFD.As well as for the animal KO fed an ND, the amount ofTregs present in the adipose tissue of HFD IL-21 KO mice

was significantly higher than the equivalent WT animals(Fig. 6A). Comparative analysis of Tregs in SVF of IL-21KO and WT fed an ND or HFD confirmed that the obesitycondition reduced Tregs in SVF of WT mice. It is inter-esting to note that IL-21 KO animals fed HFD maintaineda high number of Tregs comparable with that of WT fedND (Fig. 6B). Of note, loss of IL-21 is associated withincreased lipolysis and increased Tregs and M2 macro-phage markers, suggesting that a state in which IL-21 isreduced is possibly associated with increased lipid uptakefrom anti-inflammatory cells such as Tregs and M2 mac-rophages. Interestingly, we found increased lipid contentin Tregs from IL-21 KO compared with WT (Fig. 6C).

Reduced Liver Steatosis in IL-21 KO Subjected toObesity ChallengeThe overall improvement of glucose tolerance was asso-ciated with absence of liver steatosis in IL-21 KOcompared with WT mice during HFD (Fig. 7A), with wasassociated with reduced inflammatory markers such asF4/80 and CD68 and an unexpected mild increase in glu-coneogenic enzymes such as Pck1 and G6pc (Fig. 7B).

Figure 5—Regulation of IRF4 in 3T3-L1 adipocytes and adipose fraction (AF) and SVF. 3T3-L1 adipocytes were incubated in serum-freeDMEM with isoproterenol and IL-21 as indicated. A: IRF4, Lipe, and Pnpla2 mRNA expression after 2 or 4 h of treatments. (^P < 0.05,^^^P < 0.001 isoproterenol vs. control. *P < 0.05, **P < 0.01 isoproterenol vs. isoproterenol plus IL-21.) B: mRNA expression of IRF4,F4/80, YM1 Arg1, and Mgl2 PPAR-g in SVF from WT and IL-21 KO mice. Adiponectin was used as a marker for adipose fraction. (n = 5per group, *P < 0.05. Student t test. Error bars represent the mean 6 SD.) hrs, hours.

2092 IL-21 and IRF4-Dependent Lipolysis Diabetes Volume 63, June 2014

Consistently, we found reduced Ser256 phosphorylation ofFoxO1 in the fasting IL-21 KO liver (Fig. 7C). IL-21 KOmice during HFD revealed a marked tendency to lowerfasting glucose compared with WT littermates, with nodifferences in glucagon levels (Supplementary Fig. 2A).This suggests that the slight increase in gluconeogenicenzymes is a reactive response to maintain glucose atphysiological levels. The concept of reactive response isalso supported by the intraperitoneal pyruvate tolerancetest showing increased glucose levels in IL-21 KO mice fedHFD (Fig. 7D). Furthermore, we analyzed Pck1 and G6pcexpression also in livers from HFD refed mice and at theend of euglycemic-hyperinsulinemic clamp (Supplemen-tary Fig. 2B); overall, the data suggest that during fastingor intense glucose uptake from the muscle, the absence ofIL-21 increases Pck1 expression, possibly to compensatefor lower peripheral glucose level.

IL-21R Expression in Adipose Tissue From HumanSubjects With Obesity and Glucose IntoleranceTo explore the involvement of IL-21 effects on humanadipose tissue inflammation, we analyzed its expressionin adipose tissue biopsies from patients with differentdegrees of obesity (Supplementary Table 1). We founda significant negative correlation between IL-21R andthe CD206-to-CD68 ratio expression; this ratio is knownto be higher in subjects with less body fat and lower

fasting glucose concentrations (37). Moreover, PPAR-gwas also found to be significantly and negatively corre-lated to IL-21R expression in subcutaneous adipose tissueof obese subjects (Table 1). On the other hand, we founda significant positive correlation between IL-21R and tu-mor necrosis factor-a both in visceral and in subcutane-ous adipose tissue (Table 1), indicating an involvement ofIL-21 signaling in the development or persistence of ad-ipose tissue inflammation.

DISCUSSION

IRF4 expression is highly restricted to immune cells andadipose tissue and is more abundant in mature adipocytes(38). IRF4 is nutritionally regulated by the action of in-sulin and FoxO1 and plays a significant role in the tran-scriptional regulation of lipid handling in adipocytes,promoting lipolysis. Interestingly, we found a strong re-lationship between IRF4 and ADRP gene expression, a li-polytic gene, in human VAT (r = 0.47, P , 0.0001, datanot shown). During fasting, in adipocytes, mRNA andprotein levels of IRF4 rise dramatically with subsequentdownregulation after refeeding. IRF4 promotes lipolysisat least in part by inducing the expression of the lipasesATGL and HSL (15). We measured, in adipose tissue of IL-21 KO mice fed an ND, high levels of expression of thetranscription factor IRF4 and its targets lipe and pnpla2,particularly in the refeeding conditions. Interestingly

Figure 6—Effect of IL-21 deficiency on VAT Tregs during diet-induced obesity. At 6 weeks of age, IL-21 KO mice and WT littermateswere fed HFD for 18 weeks. Cells from the epigonadal fat SVF were stained and analyzed by flow cytometry. Tregs are defined asCD25+CD4+Foxp3+. A: Left, representative dot plots; right, summary data (for fraction of CD4+ live cells). Dot plot numbers indicatethe percentage of cells in that gate for that particular experiment. (n = 4 per group, **P < 0.01. Student t test. Error bars represent themean 6 SD.) B: Comparative analysis of Tregs in SVF from IL-21 KO and WT fed ND and HFD for 18 weeks. (*P < 0.05.) C: Cells wereisolated from epigonadal VAT SVFs of WT and IL-21 KO mice fed HFD and stained for CD4+, CD25+, Foxp3+, and Nile red. (n = 5 pergroup, *P < 0.05. Student t test. Error bars represent the mean 6 SD.)

diabetes.diabetesjournals.org Fabrizi and Associates 2093

IRF4 expression is repressed in whole VAT of three dif-ferent rodent models of obesity, a hyperinsulinemicstate (15). This may seem paradoxical given that obesityis associated with insulin resistance and mice lackinginsulin receptors in fat display elevated Irf4 expression(15). We found large induction of IRF4 and its targetsalso in HFD IL-21 KO adipose tissue both in the fastingand in the refeeding state. Therefore, it remains possiblethat other factors, such as IL-21, dominate the control ofIrf4 gene expression in the context of obesity. Indeed,our in vitro studies confirm a role for IL-21 in reducingIRF4 and its target levels during lipolysis. Despite thehigh levels of IRF4, we unexpectedly found only mildlyelevated NEFA levels in fasting sera of IL-21 KO animals.The reason could be attributed to the abundance of M2macrophages and Tregs residing in the adipose tissue ofthese animals, immune populations with a strong abilityto capture and oxidize fatty acids released by adipocytes(17,36). IRF4 is a well-known player in a variety of

immune activities, including Tregs function and the de-velopment of inflammatory Th17 cells, and is absolutelyrequired for the autocrine production of IL-21 in Th17cells (39). Recently, IL-21 has emerged as a key cytokinefor the maintenance of the mucosal immune system ho-meostasis by modulating the balance between Tregs andproinflammatory Th17 cells (28). Interestingly, the fre-quencies of Tregs and Th17 cells often show an inverserelationship, as their differentiation processes are alsocounterbalanced (40). It is worthy of noting that the invivo presence of Th17 T cells in adipose tissue undernormal chow conditions or an HFD has not yet beenextensively reported. Few recent observations demon-strate that diet-induced obesity predisposes to IL-6–dependent Th17 expansion in adipose tissue (41). Giventhat IL-21 induces and amplifies Th17 development in-dependently of IL-6 (39), it is possible that IL-21 is se-creted in specific phases of adipose tissue expansion,eventually exacerbating early disease progression.

Figure 7—IL-21 deficiency protects from hepatic steatosis during diet-induced obesity. A: IL-21 KO does not show macrovescicularsteatosis during diet-induced obesity. B: Expression of metabolic and inflammatory genes. Expression of mRNA was determined byreal-time PCR and normalized to b-actin. (n = 6 per group, *P < 0.05, Student t test. Error bars represent the mean 6 SD.) C: Ser473

Akt phosphorylation (p) in refeeding condition and Ser256 FoxO1 phosphorylation in fasting condition were analyzed by Western blot. (n = 6per group, *P < 0.05, Student t test. Data are means 6 SD.) A representative image of four and three mice per group is shown. a.u.,arbitrary unit. D: Intraperitoneal pyruvate tolerance test. (n = 6 per group.)

2094 IL-21 and IRF4-Dependent Lipolysis Diabetes Volume 63, June 2014

Tregs with a unique phenotype were highly enrichedin the abdominal fat of lean mice, but their numberswere strikingly and specifically reduced at this site ininsulin-resistant models of obesity (11). Recent studiesreveal an important role for VAT-specific natural Tregsin the suppression of obesity-associated inflammation inVAT and consequently in reducing insulin resistance.The number of VAT Tregs decreases with obesity, anda boost in the number of these cells in obese mice canimprove insulin sensitivity (11,13). Tregs expressingFoxp3 can secrete anti-inflammatory signals such as IL-10and transforming growth factor b, inhibit macrophagemigration, and induce M2-like macrophage differenti-ation (11). IL-21–mediated T-cell activation relies partlyon its ability to inhibit the differentiation of Tregs andto make T cells resistant to the Treg-mediated im-munosuppression (27,28). In SVF of IL-21 KO adiposetissue, we found a significant increase in the number ofresident Tregs in both mice fed an ND than in those re-ceiving an HFD. This may suggest the possibility that IL-21 regulates Treg number and differentiation even in theadipose tissue. What causes the decrease in Treg fractionin abdominal adipose tissue during obesity is still unde-fined; nevertheless, recent reports indicate a possible rolefor the hyperleptinemic state characterizing obesity tomodulate Treg number and activity (42,43). Thus, a hypoth-esis that needs further exploitation in appropriate modelsis that leptin and IL-21 share some biological function andcooperate in causing Treg dramatic reduction in adiposetissue during obesity development. Since the adipose tissue

of IL-21 KO mice is characterized by a lower degree ofmacrophage infiltration and increased expression of M2polarization antigens (YM1, Mgl2), it is intriguing to hy-pothesize that a IL-21/Tregs axis might regulate the balancebetween macrophage M2 polarization and M1 infiltrationin the context of obese adipose tissue. A recent study showsthat VAT-resident Tregs and M2 macrophages specificallyexpress PPAR-g, an important factor controlling their accu-mulation, phenotype, and function (17,36). Consistently,we measured significant increased levels of PPAR-g andM2 markers in SVF of IL-21 KO mice. In this context, inhumans we found a negative association between IL-21Rand PPAR-g gene expression in SAT, suggesting that IL21signaling runs in parallel to PPAR-g. A newly discoveredproperty was that in VAT, but not in lymphoid tissue, Tregscan take up lipids—an ability not shared by conventionalT cells residing at the same site (17). It is also known thatM2 macrophages and Tregs prevalently use fatty acids forATP generation to maintain their functions (36). In IL-21KO mice, we observed a high induction of IRF4-relatedlipolysis and at the same time increased lipid uptake byTregs. This highlights the possibility that IL-21 regulatesTreg activity in adipose tissue.

In conclusion, we relate for the first time the IL-21/IL-21R dyad to IRF4-dependent regulation of lipolysis andreduction of Tregs in adipose tissue. We hypothesize thatIL-21 is a crucial player in this context, since we found anincrease in mRNA levels of IL-21 and IL-21R in adiposetissue of obese animals and obese human subjects com-pared with their lean controls. Our data suggest thatpreventing IL-21 signaling might counteract obesity andthe consequent metabolic defects in an experimentalmodel—a finding with potential therapeutic implicationsin human subjects with metabolic syndrome and type 2diabetes.

Funding. This study was funded in part by Fondazione Roma 2008, Euro-pean Foundation for the Study of Diabetes/Lilly 2012, AIRC 2012 Project IG13163, FP7-Health-241913-FLORINASH, FP7-Health-EURHYTHDIA, and PRIN2012 (all to M.Fe.). T.M. is the recipient of the Albert Reynolds Travel Fellow-ship from the European Association for the Study of Diabetes and FellowshipPrize from Società Italiana di Diabetologia. G.P.S. is the recipient of a fellow-ship from Laboratori Guidotti, Pisa, Italy. A.G. has received support by grantsfrom Università Cattolica del Sacro Cuore (Fondi Ateneo Linea D.3.2 Sin-drome Metabolica); from the Italian Ministry of Education, Universities andResearch (PRIN 2010JS3PMZ_011); and from Fondazione Don Gnocchi,Milan, Italy.Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. M.Fa. performed experiments, analyzed data,drafted the manuscript, and wrote the final version of the manuscript. V.M.performed experiments, analyzed data, and reviewed the manuscript. M.M.,A.M., V.C., M.C., T.M., G.P.S., and A.G. performed experiments and analyzeddata. J.M.M.-N. performed experiments. L.F. and R.M. contributed to the dis-cussion and drafted the manuscript. R.L. and G.M. contributed to the discussionand edited the manuscript. J.M.F.R. analyzed data and reviewed the manu-script. M.Fe. drafted the manuscript and wrote the final version of the manu-script. M.Fe. is the guarantor of this work and, as such, had full access to all

Table 1—IL-21R is positively correlated with inflammatoryfactors in human adipose tissue

VAT (n = 112) SAT (n = 95)

r P r P

Age (years) 20.11 0.2 20.11 0.2

BMI (kg/m2) 0.07 0.4 0.22 0.03

Fat mass (%) 0.10 0.3 0.18 0.08

Fasting glucose (mg/dL) 0.08 0.4 20.01 0.9

Total cholesterol (mg/dL) 20.021 0.8 20.20 0.06

HDL cholesterol (mg/dL) 20.07 0.4 20.18 0.08

Fasting triglycerides(mg/dL) 0.20 0.04 0.5 0.6

PPARg (R.U.) 0.10 0.4 20.31 0.01

FASN (R.U.) 20.25 0.01 20.19 0.08

ACC1 (R.U.) 20.16 0.08 20.25 0.02

TNFa (R.U.) 0.47 ,0.0001 0.24 0.02

CD206/CD68 (R.U.) 20.17 0.1 20.26 0.02

Bivariate correlation between IL-21R and anthropometric, clini-cal parameters as well as adipose tissue gene expression inhuman adipose tissue biopsies (n = 207). Expression of mRNAwas determined by real-time PCR and normalized to cyclophilinA (peptidylprolyl isomerase A). Bivariate correlation was per-formed using nonparametric (Spearman) tests. R.U., relativeunits of gene expression; SAT, subcutaneous adipose tissue.

diabetes.diabetesjournals.org Fabrizi and Associates 2095

the data in the study and takes responsibility for the integrity of the data andthe accuracy of the data analysis.

References1. Osborn O, Olefsky JM. The cellular and signaling networks linking theimmune system and metabolism in disease. Nat Med 2012;18:363–3742. Xu H, Barnes GT, Yang Q, et al. Chronic inflammation in fat plays a crucialrole in the development of obesity-related insulin resistance. J Clin Invest 2003;112:1821–18303. Gregor MF, Hotamisligil GS. Inflammatory mechanisms in obesity. Annu RevImmunol 2011;29:415–4454. Kanneganti TD, Dixit VD. Immunological complications of obesity. Nat Im-munol 2012;13:707–7125. Fernández-Real JM, Pickup JC. Innate immunity, insulin resistance and type 2diabetes. Diabetologia 2012;55:273–2786. Chawla A, Nguyen KD, Goh YP. Macrophage-mediated inflammation inmetabolic disease. Nat Rev Immunol 2011;11:738–7497. Oh DY, Morinaga H, Talukdar S, Bae EJ, Olefsky JM. Increased macrophagemigration into adipose tissue in obese mice. Diabetes 2012;61:346–3548. Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, Ferrante AWJr. Obesity is associated with macrophage accumulation in adipose tissue. J ClinInvest 2003;112:1796–18089. Lumeng CN, Bodzin JL, Saltiel AR. Obesity induces a phenotypic switch inadipose tissue macrophage polarization. J Clin Invest 2007;117:175–18410. Kosteli A, Sugaru E, Haemmerle G, et al. Weight loss and lipolysis promotea dynamic immune response in murine adipose tissue. J Clin Invest 2010;120:3466–347911. Winer S, Chan Y, Paltser G, et al. Normalization of obesity-associated insulinresistance through immunotherapy. Nat Med 2009;15:921–92912. Nishimura S, Manabe I, Nagasaki M, et al. CD8+ effector T cells contributeto macrophage recruitment and adipose tissue inflammation in obesity. Nat Med2009;15:914–92013. Feuerer M, Herrero L, Cipolletta D, et al. Lean, but not obese, fat is enrichedfor a unique population of regulatory T cells that affect metabolic parameters. NatMed 2009;15:930–93914. Eguchi J, Kong X, Tenta M, Wang X, Kang S, Rosen ED. Interferon regulatoryfactor 4 regulates obesity-induced inflammation through regulation of adiposetissue macrophage polarization. Diabetes 2013;62:3394–340315. Eguchi J, Wang X, Yu S, et al. Transcriptional control of adipose lipidhandling by IRF4. Cell Metab 2011;13:249–25916. Sakaguchi S, Yamaguchi T, Nomura T, Ono M, Regulatory T. RegulatoryT cells and immune tolerance. Cell 2008;133:775–78717. Cipolletta D, Feuerer M, Li A, et al. PPAR-g is a major driver of the accu-mulation and phenotype of adipose tissue Treg cells. Nature 2012;486:549–55318. Monteleone G, Pallone F, Macdonald TT. Interleukin-21 as a new therapeutictarget for immune-mediated diseases. Trends Pharmacol Sci 2009;30:441–44719. Parrish-Novak JD, Dillon SR, Nelson A, et al. Interleukin 21 and its receptorare involved in NK cell expansion and regulation of lymphocyte function. Nature2000;408:57–6320. Spolski R, Leonard WJ, Nelson A, et al. Interleukin-21: basic biology andimplications for cancer and autoimmunity. Annu Rev Immunol 2008;26:57–7921. Mehta DS, Wurster AL, Grusby MJ. Biology of IL-21 and the IL-21 receptor.Immunol Rev 2004;202:84–9522. Parrish-Novak JD, Foster DC, Holly RD, Clegg CH. Interleukin-21 and the IL-21 receptor: novel effectors of NK and T cell responses. J Leukoc Biol 2002;72:856–863

23. Asao H, Okuyama C, Kumaki S, et al. Cutting edge: the common gamma-chain is an indispensable subunit of the IL-21 receptor complex. J Immunol2001;167:1–524. Caprioli F, Sarra M, Caruso R, et al. Autocrine regulation of IL-21 productionin human T lymphocytes. J Immunol 2008;180:1800–180725. Nurieva R, Yang XO, Martinez G, et al. Essential autocrine regulation by IL-21 in the generation of inflammatory T cells. Nature 2007;448:480–48326. Wei L, Laurence A, Elias KM, O’Shea JJ. IL-21 is produced by Th17 cellsand drives IL-17 production in a STAT3-dependent manner. J Biol Chem 2007;282:34605–3461027. Peluso I, Fantini MC, Fina D, et al. IL-21 counteracts the regulatory T cell-mediated suppression of human CD4+ T lymphocytes. J Immunol 2007;178:732–73928. Fantini MC, Rizzo A, Fina D, et al. IL-21 regulates experimental colitis bymodulating the balance between Treg and Th17 cells. Eur J Immunol 2007;37:3155–316329. Menghini R, Casagrande V, Menini S, et al. TIMP3 overexpression inmacrophages protects from insulin resistance, adipose inflammation, and non-alcoholic fatty liver disease in mice. Diabetes 2012;61:454–46230. Serino M, Menghini R, Fiorentino L, et al. Mice heterozygous for tumornecrosis factor-alpha converting enzyme are protected from obesity-inducedinsulin resistance and diabetes. Diabetes 2007;56:2541–254631. Kiechl S, Wittmann J, Giaccari A, et al. Blockade of receptor activator ofnuclear factor-kB (RANKL) signaling improves hepatic insulin resistance andprevents development of diabetes mellitus. Nat Med 2013;19:358–36332. Giaccari A, Morviducci L, Pastore L, et al. Relative contribution of glyco-genolysis and gluconeogenesis to hepatic glucose production in control and di-abetic rats. A re-examination in the presence of euglycaemia. Diabetologia 1998;41:307–31433. Marino A, Menghini R, Fabrizi M, et al. ITCH deficiency protects from diet-induced obesity. Diabetes. 29 October 2013 [Epub ahead of print]34. Fiorentino L, Cavalera M, Menini S, et al. Loss of TIMP3 underlies di-abetic nephropathy via FoxO1/STAT1 interplay. EMBO Mol Med 2013;5:441–45535. Clough LE, Wang CJ, Schmidt EM, et al. Release from regulatory T cell-mediated suppression during the onset of tissue-specific autoimmunity is as-sociated with elevated IL-21. J Immunol 2008;180:5393–540136. Odegaard JI, Chawla A. Alternative macrophage activation and metabolism.Annu Rev Pathol 2011;6:275–29737. Moreno-Navarrete JM, Ortega F, Gómez-Serrano M, et al. The MRC1/CD68ratio is positively associated with adipose tissue lipogenesis and with musclemitochondrial gene expression in humans. PLoS ONE 2013;8:e7081038. Eguchi J, Yan QW, Schones DE, et al. Interferon regulatory factors aretranscriptional regulators of adipogenesis. Cell Metab 2008;7:86–9439. Huber M, Brüstle A, Reinhard K, et al. IRF4 is essential for IL-21-mediatedinduction, amplification, and stabilization of the Th17 phenotype. Proc Natl AcadSci USA 2008;105:20846–2085140. Lee BC, Lee J. Cellular and molecular players in adipose tissue in-flammation in the development of obesity-induced insulin resistance. BiochimBiophys Acta 2013;pii:S0925-4439(13)00179-841. Winer S, Paltser G, Chan Y, et al. Obesity predisposes to Th17 bias. Eur JImmunol 2009;39:2629–263542. De Rosa V, Procaccini C, Calì G, et al. A key role of leptin in the control ofregulatory T cell proliferation. Immunity 2007;26:241–25543. Matarese G, Procaccini C, De Rosa V, Horvath TL, La Cava A. RegulatoryT cells in obesity: the leptin connection. Trends Mol Med 2010;16:247–256

2096 IL-21 and IRF4-Dependent Lipolysis Diabetes Volume 63, June 2014