1-Autophagy Activity is Up-Regulated in Adipose Tissue

Autophagy Activity Is Up-Regulated in Adipose Tissueof Obese Individuals and Modulates ProinflammatoryCytokine Expression

H. J. Jansen,* P. van Essen,* T. Koenen, L. A. B. Joosten, M. G. Netea,C. J. Tack, and R. Stienstra

Department of Medicine, Radboud University Nijmegen Medical Centre and Nijmegen Institute forInfection, Inflammation, and Immunity, 6525 GA, Nijmegen, The Netherlands

Autophagy, an evolutionary conserved process aimed at recycling damaged organelles and proteinaggregates in the cell, also modulates proinflammatory cytokine production in peripheral bloodmononuclear cells. Because adipose tissue inflammation accompanied by elevated levels of pro-inflammatory cytokines is characteristic for the development of obesity, we hypothesized thatmodulation of autophagy alters adipose tissue inflammatory gene expression and secretion. Wetested our hypothesis using ex vivo and in vivo studies of human and mouse adipose tissue. Levelsof the autophagy marker LC3 were elevated in sc adipose tissue of obese vs. lean human subjectsand positively correlated to both systemic insulin resistance and morphological characteristics ofadipose tissue inflammation. Similarly, autophagic activity levels were increased in adipose tissueof obese and insulin resistant animals as compared with lean mice. Inhibition of autophagy by3-methylalanine in human and mouse adipose tissue explants led to a significant increase in IL-1�,IL-6, and IL-8 mRNA expression and protein secretion. Noticeably, the enhancement in IL-1�, IL-6,and keratinocyte-derived chemoattractant (KC) by inhibition of autophagy was more robust in thepresence of obesity. Similar results were obtained by blocking autophagy using small interferingRNA targeted to ATG7 in human Simpson-Golabi-Behmel syndrome adipocytes. Our results dem-onstrate that autophagy activity is up-regulated in the adipose tissue of obese individuals andinhibition of autophagy enhances proinflammatory gene expression both in adipocytes and ad-ipose tissue explants. Autophagy may function to dampen inflammatory gene expression andthereby limit excessive inflammation in adipose tissue during obesity. (Endocrinology 153:5866–5874, 2012)

Autophagy is a homeostatic mechanism functioning asa disposal system that degrades large intracellular or-

ganelles or protein aggregates to change cellular structureduring differentiation or to generate essential nutrients intimes of energy deprivation (1, 2). When autophagy isactivated, a double-membraned vesicle, named the au-tophagosome, engulfs these components and fuses withlysosomes to complete degradation (3). The elongationand shape of the autophagosome is controlled by multipleautophagy-related proteins (ATG proteins) that include

ATG5, ATG7, ATG8, ATG12, and ATG16L1. The mam-malian homolog of yeast ATG8, LC3, is a commonly usedas a marker of autophagy because this protein is conju-gated with the lipid phosphatidylethanolamine on activa-tion of autophagy resulting in LC3-II formation.

A growing body of evidence suggests that autophagynot only monitors cellular energy balance but is also im-portant in the regulation of apoptosis and protectionagainst infection with pathogens (4–6). Indeed, au-tophagy has been closely linked to control of innate and

ISSN Print 0013-7227 ISSN Online 1945-7170Printed in U.S.A.Copyright © 2012 by The Endocrine Societydoi: 10.1210/en.2012-1625 Received June 12, 2012. Accepted September 24, 2012.First Published Online November 1, 2012

* P.v.E. and H.J.J. contributed equally to this manuscript.

Abbreviations: ATG, Autophagy-related proteins; BMI, body mass index; CLS, crown-likestructure; eWAT, epididymal WAT; KC, keratinocyte-derived chemoattractant; 3MA,3-methyladenine; NOB-1, Nin1 binding protein; NOD2, nucleotide-binding oligomeriza-tion domain containing 2; SAT, sc adipose tissue; SGBS, Simpson-Golabi-Behmel syn-drome; siRNA, small interference RNA; TBS, Tris-buffered saline; VAT, visceral adiposetissue; WAT, white adipose tissue.

G E N E R A L E N D O C R I N O L O G Y

5866 endo.endojournals.org Endocrinology, December 2012, 153(12):5866–5874

adaptive immune responses in host defense in part by reg-ulation of cytokine production (7). Cytokines includingIL-1�, IL-18, TNF�, and IL-6 are known to be regulatedby autophagy (8). For example, macrophages derivedfrom ATG16L1-deficient mice produced higher levels ofIL-1� (9), whereas mice with a conditional deletion ofAtg7 in the intestinal epithelium showed an enhancedmRNA expression of IL-1� (10). Additional studies usingcells from human origin demonstrated that the inhibitionof autophagy led to the up-regulated production of IL-1�

(11, 12).In addition to its role in host defense partly through

control of cytokine production, autophagy has beenshown to regulate fat accumulation within adipocytes.Animals lacking the autophagy-related proteins ATG5and ATG7 are characterized by a limited capacity of whiteadipose tissue (WAT) to store triacylglycerols and thusdisplay a robust reduction in WAT mass compared withwild-type animals (13–15), suggesting that autophagy isessential for normal adipogenesis. Interestingly, au-tophagy also appears to contribute to the development ofobesity as adipose tissue of obese human individuals ischaracterized by an enhancement in autophagic activity(16, 17).

In parallel with the development of obesity, a chronicinflammation in the adipose tissue develops that is partlyinitiated by adipocytes releasing chemokines, proinflam-matory cytokines, and adipokines, resulting in the infil-tration of immune cells into adipose tissue (18, 19). Var-ious cytokines are known to be released or produced byinflamed adipose tissue of obese individuals includingIL-1� (20–23), IL-6, TNF�, and IL-18 (22) and contributeto the development of obesity-induced insulin resistance.Despite the importance of proinflammatory cytokines inthe development of insulin resistance, the regulation andpotential triggers inducing their production are onlypartly understood. In line with its enhanced activity inadipose tissue during obesity and its role in the regulationof inflammatory cytokine secretion, we hypothesized thatautophagy modulates adipose tissue inflammation. Ourresults show that inhibition of autophagic activity en-hances both gene expression and protein secretion of theproinflammatory cytokines IL-1�, IL-6, and IL-8 in adi-pocytes and adipose tissue of both humans and mice.

Materials and Methods

Subcutaneous adipose tissue biopsies from leanand obese individuals

Adipose tissue samples were obtained from human subjectsthrough advertisements in local newspapers. We included over-

weight [body mass index (BMI) 27–35 kg/m2; 25% male sub-jects] and normal (BMI 20–25 kg/m2; 35% male subjects) sub-jects between 40 and 70 yr old. Subcutaneous adipose tissuebiopsies were obtained under local anesthesia by needle biopsies6–10 cm lateral to the umbilicus. Samples were taken after anovernight fast. Various biochemical measurements includingplasma triacylglycerols, cholesterol, glucose, and insulin weredone using standard laboratory methods.

In vitro and ex vivo experiments with human andmurine adipose tissue

Intact adipose tissue fragments from sc adipose tissue (SAT)or visceral adipose tissue (VAT), obtained during surgery fromhealthy individuals 30–70 yr old with a BMI between 20 and 33kg/m2, were used to study the effects of autophagy. Adiposetissue fragments were directly cultured in DMEM supplementedwith 10% fetal calf serum with or without 3-methyladenine(3MA; autophagy inhibitor; Sigma-Aldrich, St. Louis, MO).Minced adipose tissue was digested using collagenase (Sigma-Aldrich) at a concentration of 5 mg/ml dissolved in DMEM.Tissues were incubated for 45 min at 37 C and were subsequentlyfiltered through a 250-�M nylon mesh filter. After centrifugationat 200 rpm for 10 min, the floating cells were collected as adi-pocytes and the pelleted cells as stromal vascular cells. The studyprotocols are approved by the University of Nijmegen EthicalCommittee, and all participants gave written informed consent.Epididymal white adipose tissue (eWAT) obtained from leanC57/BL6 male mice and obese leptin deficient (Lepob) male miceon a C57/BL6 background (25) was similarly treated with 3MA.After 24 h of incubation with or without 3MA, cytokine secre-tion and gene expression levels were determined. The study pro-tocol was approved by the Animal Experimentation Committeeof the University Medical Centre of Nijmegen.

Small interference RNA (siRNA)To specifically suppress autophagy-related protein 7 (ATG7)

expression in differentiated adipocytes, human Simpson-Golabi-Behmel syndrome (SGBS) preadipocyte cells were differentiatedtoward mature adipocytes using a standard adipogenic protocol(26). After 12 d of differentiation, cells were transfected (X-tremeGENE siRNA transfection reagent; Roche, Basel, Switzer-land) with siRNA against ATG7 (Thermo Scientific, Waltham,MA). As a nonspecific control, scrambled siRNA (Thermo Sci-entific) was used. After 96 h of incubation, gene and proteinexpression were analyzed in cells and supernatant.

Cytokine measurementsConcentrations of mouse IL-1� were determined by specific

RIA (detection limit is 20 pg/ml) as previously described (27).Mouse IL-6 and keratinocyte-derived chemoattractant (KC)concentrations were measured by commercial ELISA kits (Invit-rogen, Carlsbad, CA; detection limits 16 pg/ml; and R&D Sys-tems, Minneapolis, MN; detection limits 16 pg/ml, respectively),according to the instructions of the manufacturer. Bio-activeIL-1 was measured using the Nin1 binding protein (NOB-1) as-say, which consists of mouse cells derived from the EL-4 line,which produce IL-2 in response to concentrations of IL-1 as lowas 1 pg/ml. NOB-1 is not responsive to TNF-�, TNF-�, inter-feron-� , and lipopolysaccharide (28). IL-2 was detected using amouse IL-2 ELISA kit (R&D Systems; detection limits 16 pg/ml).

Endocrinology, December 2012, 153(12):5866–5874 endo.endojournals.org 5867

Human IL-6 and IL-8 concentrations were measured using com-mercially available ELISA kits (R&D Systems).

Western blotsFor Western blotting, SGBS cells or approximately 30 mg of

SAT or VAT were lysed in 100 �l of lysis buffer [50 mM Tris (pH7.4), 150 mM NaCl, 2 mM EDTA, 1% Nonidet P-40, 50 mM NaF,and 0.25% sodium deoxycholate with phosstop phosphataseinhibitor cocktail tablet (Roche)] and complete, EDTA-free pro-tease inhibitor cocktail tablet (Roche). The homogenate was fro-zen and then thawed and centrifuged at 4 C for 10 min at15,000 � g, and the supernatant was used for Western blotanalysis. Protein concentrations of the 30% lysates were deter-mined using a bicinchoninic assay protein assay (Thermo FisherScientific, Rockford, IL). Samples of approximately 20 �g pro-tein were subjected to SDS-PAGE using 12 and 16% polyacryl-amide gels at a voltage of 70–120 V. After SDS-PAGE, proteinswere transferred to nitrocellulose membrane (0.2 mm), and themembrane was blocked with 5% (wt/vol) milk powder in Tris-buffered saline (TBS)/Tween 20 for 1 h at room temperature,followed by incubation overnight at 4 C with a LC3 antibody(Novus Biochemicals, Littleton, CO) in 5% (wt/vol) milk pow-der/TBS/Tween 20 or with an actin antibody (Sigma-Aldrich, St.Louis, MO) in 5% milk powder in TBS/Tween 20. After over-night incubation, the blots were incubated with horseradish per-oxidase-conjugated swine antirabbit antibody at a dilution of1:5000 in 5% (wt/vol) milk powder in TBS/Tween 20 for 1 h atroom temperature and subsequently developed with enhancedchemiluminescence (GE Healthcare, Pittsburgh, PA) or ultrasen-sitive enhanced enhanced chemiluminescence (Thermo Scien-tific), according to the manufacturer’s instructions. The intensityof the bands on the Western blots was assessed by Image Labstatistical software (Bio-Rad Laboratories, Hercules, CA).

Real-time PCRTotal RNA purification of SGBS cells, SAT, or VAT was done

using TRIzol Reagent (Invitrogen) according to the manufactur-er’s instructions. Isolated RNA was subsequently transcribedinto cDNA using iScript cDNA synthesis kit (Bio-Rad Labora-tories) followed by quantitative PCR using SYBR Green (AppliedBiosystems, Foster City, CA). The following primers were used:for human IL-1� forward, 5�-GCCCTAAACAGATGAAGT-GCTC-3� and reverse, 5�-GAACCAGCATCTTCCTCAG-3�;for human IL-6 forward, 5�-AACCTGAACCTTCCAAA-GATGG-3� and reverse, 5�-TCTGGCTTGTTCCTCACTACT-3�; for 149 human IL-8 forward, 5�-ACTGAGAGTGATT-GAGAGTGGAC-3� and reverse, 5�-AACCCTCTGCACCC-AGTTTTC-3�; and for human ATG7 forward, 5�-CAGTTT-GCCCCTTTTAGTAGTGC-3� and reverse, 5�-CTTAATGTC-CTTGGGAGCTTCA-3�. For mouse primers the following wereused: IL-1� forward, 5�-GCAACTGTTCCTGAACTCA-ACT-3� and reverse, 5�-ATCTTTTGGGGTCCGTCAACT-3�;for mouse IL-6 forward, 5�-CAAGTCGGAGGCTTAATTA-CACATG-3� and reverse, 5�-ATTGCCATTGCACAACTCT-TTTCT-3�; for mouse IL-8 (KC) forward, 5�-TGGC-TGGGATTCACCTCAA-3� and reverse, 5�-GAGTGTGGC-TATGACTTCGGTTT-3�; for mouse ATG7 forward, 5�-CCTTCGCGGACCTAAAGAAGT-3�and reverse, 5�-CCCG-GATTAGAGGGATGCTC-3�. All primer pairs were tested forefficiency using standard curves and were between 95 and 105%.Additionally, a melt curve analysis was included in every run toascertain the formation of one single specific PCR product. Thehousekeeping genes �2-microglobulin and 36B4 were selected ashuman and mouse housekeeping genes.

Gene expression data obtained from the human samples wascorrected for expression of the housekeeping genes �2-micro-globulin, for which the forward primer, 5�-ATGAGTATGCCT-

FIG. 1. Levels of the autophagy marker LC3 determined by Western blot in human and mouse adipose tissue. A, Representative Western blot ofLC3 in human lean (BMI �22 kg/m2) and obese subjects (BMI �32 kg/m2). LC3 levels in epididymal adipose tissue from C57BL6 wt mice andleptin deficient (Lep ob) mice (C). LC3-II, the actin ratio that reflects the level of autophagic activity in adipose tissue, is calculated by quantificationof Western blot intensity in human subjects (n � 33) (B) and wild-type (n � 5) vs. leptin deficient (Lepob) mice (n � 4) (D). Data are presented asmeans � SEM. *, P � 0.05; ***, P � 0.001.

5868 Jansen et al. Autophagy, Adipose Tissue, and Obesity Endocrinology, December 2012, 153(12):5866–5874

GCCGTGTG-3�, and reverse primer, 5�-CCAAATGCG-GCATCTTCAAAC-3�, were used. For mouse samples, forwardprimer, 5�-AGCGCGTCCTGGCATTGTGTGG-3�, and reverseprimer, 5�-GGGCAGCAGTGGTGGCAGCAGC-3�, were usedto detect 36B4 expression levels, and gene expression resultswere corrected using 36B4 values.

Statistical analysesDifferences in cytokine production capacity and data on

mRNA expression levels between groups were analyzed usingthe ANOVA with the least significant difference post hoc test todetect statistical differences between groups and treatment (SATvs. VAT, lean vs. obese, control vs. 3MA treatment) after vari-ables were log transformed. Data on immunoblot intensity val-ues were statistically analyzed using Student’s t tests. Differenceswere considered statistically significant at P � 0.05.

Results

Autophagy is up-regulated in adipose tissueduring obesity

To assess the correlation between autophagy and obe-sity, SAT was obtained from healthy individuals (n � 33)that varied in BMI from 19 to 40 kg/m2 [SupplementalTable 1 (patient characteristics), published on The Endo-crine Society’s Journals Online web site at http://endo.en-

dojournals.org] and were analyzed by Western blot to de-tect LC3-II levels (Fig. 1A; a representative Western blot isshown). The mammalian homolog of yeast ATG8, LC3, isa commonly used marker of autophagy because this pro-tein is conjugated with the lipid phosphatidylethano-lamine upon activation of autophagy resulting in LC3-IIformation. The autophagy marker LC3-II (microtubuleassociated protein light chain 3) was more abundantlypresent in obese individuals (average BMI �32 kg/m2)compared with lean subjects (average BMI �22 kg/m2)(Fig. 1B). Similarly, LC3 levels were profoundly higher inadipose tissue of obese animals lacking leptin (Lepob),compared with lean wild-type mice (Fig. 1, C and D). Thissuggests that autophagy is activated in adipose tissue dur-ing obesity and confirms previous work by Kovsan et al.(16). In line with these results, a positive correlation wasobserved between LC3-II levels and BMI (r � 0.43; P �

0.01, Fig. 2A). Furthermore, the positive correlation be-tween HOMA-IR and adipose tissue levels of LC3-II levels(r � 0.44; P � 0.02, Fig. 2B) revealed that insulin sensi-tivity negatively correlated with the autophagic activitylevels. Because obesity and insulin resistance is associatedwith the development of adipose tissue inflammation thatis characterized by the influx of macrophages, we set outto correlate macrophage influx with LC3-II levels in the sc

FIG. 2. Correlation between autophagic activity, obesity, insulin sensitivity, and adipose tissue morphology. Spearman correlations analysisbetween autophagic LC3-II levels and BMI (A), HOMA-IR (B), and influx of macrophages within the adipose tissue (C) are shown. HOMA-IRdenotes the homeostasis model assessment insulin resistance. HOMA-IR was calculated by the following formula: HOMA-IR � fasting plasmainsulin (microinternational units per milliliter) � fasting plasma glucose (millimoles per liter)/22.5. LC3-II activity levels were determined in subjectswith or without the presence of CLSs in adipose tissue (D).


biopsies obtained from our study subjects. We observed apositive correlation between the influx of macrophagesand LC3-II levels in SAT (r � 0.38; P � 0.03, Fig. 2C).Additionally, aggregation of macrophages in proinflam-matory, crown-like structures (CLSs) that surround dyingadipocytes was accompanied by significantly higherLC3-II levels compared with adipose tissue from individ-uals that lacked CLSs (Fig. 2D).

Autophagy regulates expression ofproinflammatory cytokines in human adiposetissue ex vivo

To study the role of autophagy in the production andsecretion of proinflammatory cytokines by the adiposetissue, human SAT and VAT explants were cultured inmedium containing the autophagy inhibitor 3MA or PBSfor 24 h. The concentration of 3MA (10 mM) usedthroughout the experiments was in accordance with pre-vious studies (29). 3MA clearly reduced autophagic ac-tivity in adipose tissue according to lower LC3-II proteinlevels (Fig. 3A). 3MA treatment resulted in a significantincrease of IL-1�, IL-6, and IL-8 protein secretion by bothSAT and VAT (Fig. 3B). In line with the cytokine secretionprofile, mRNA levels of IL-1�, IL-6, and IL-8 gene ex-pression were increased in SAT and VAT tissue explants

exposed to 3MA (Fig. 3C). Secretionlevels of leptin were unchanged, sug-gesting that inhibition of autophagy by3MA does not modulate the general en-docrine function of adipose tissue (Sup-plemental Fig. 1).

Autophagy regulates expressionof proinflammatory cytokines inmouse adipose tissue explants

To confirm the effects of inhibitionof autophagy in mouse fat tissue,eWAT derived from obese leptin defi-cient (Lepob) mice and lean wild-typemice was treated with 3MA. Inhibitionof autophagy in lean mice led to a sig-nificant increase of proinflammatoryprotein secretion. Interestingly, the in-duction in IL-1�, IL-6, and KC wasmore robust in obese animals (Fig. 4A).Most likely, the enhanced secretion is aresult of elevated gene expression levels(Fig. 4B). Indeed, 3MA treatment led toa 20-fold increase in adipogenic IL-1�

gene expression in obese animals,whereas inhibition of autophagy in leanmice resulted in a 3-fold induction ingene expression. Similar results were

seen for IL-6 and IL-8 gene expression levels (Fig. 4B).These data imply that autophagy somehow may act tolimit excessive inflammatory gene expression in adiposetissue during obesity as evidenced by the exacerbated cy-tokine expression upon autophagic inhibition in adiposetissue from obese and insulin resistant animals. Notice-ably, 3MA treatment did not appear to affect the generalsecretory capacity of the WAT explants because leptinsecretion levels did not differ upon autophagic inhibition(Supplemental Fig. 2).

Autophagy inhibition by siRNA ATG7 in humanmature adipocytes increases cytokine production

Because adipose tissue is composed of both adipocytesand stromal vascular cells, we set out to characterize geneexpression levels of various autophagy-related genes inboth cell populations. Quantitative PCR analysis revealedthat numerous autophagy-related genes are expressed inboth adipocytes and nonadipocyte cells part of the adiposetissue (Fig. 5A and Supplemental Fig. 3). Expression of theautophagy machinery in both the stromal vascular frac-tion and adipocytes suggest that various cell types maycontribute to the autophagy-mediated inhibition of in-flammation. To learn more about the contribution of adi-

FIG. 3. Protein levels and gene expression of cytokines in human SAT and VAT treated with3MA for 24 h. Western blot analysis of LC3-II levels in human VAT explants were conductedafter treatment with 3MA for 24 h (A). Secretion levels of bioactive IL-1 (indirectly by IL-2production of NOB-1 assay), IL-6, and IL-8 by human SAT and VAT. The secretion levels areexpressed as the relative change in picograms per milliliter per milligram (cytokine/volume ofsupernatant/weight of explant). Mean values of cytokine expression from SAT control sampleshave been set as 100% (B). Ctrl, Control. Gene expressions of IL-1�, IL-6, and IL-8 in SAT andVAT (C) are shown. Control samples were set as 1. All experiments have been repeated threetimes or more. Data are presented as means � SEM. *, P � 0.05; **, P � 0.01; ***, P �0.001.


pocyte-specific autophagy to control of inflammation,ATG7 expression, an essential protein in the assembly ofautophagosomes by controlling the vesicle elongation andpredominantly expressed in adipocytes, was blocked bysiRNA treatment in fully differentiated human SGBS adi-pocytes. siRNA treatment led to a 65% reduction inATG7expression (Fig. 5B). The reduction in ATG7 expres-sion promoted a significant increase in both gene expressionand protein levels of IL-1�, IL-6, and IL-8 (Fig. 5, C and D).These data confirm the involvement of autophagy in con-trolling the inflammatory trait of adipocytes.

Discussion

Autophagy has been shown to affect many cellular pro-cesses including inflammation, oxidative stress, and in-

nate and acquired immune response (30). Our studyprovides evidence that autophagy modulates the in-flammatory status of adipose tissue by controlling theproduction of proinflammatory cytokines including IL-1�, IL-6, and IL-8. Inhibition of autophagy leads to anincrease in gene expression and secretion of proinflam-matory cytokines by the adipose tissue.

Previous work has shown that insulin inhibits au-tophagic action (31), and this regulatory pathway maypartly explain the enhanced levels of LC3 within adi-pose tissue of obese individuals characterized by insulinresistance. Whereas autophagic inhibition in adipocytesand total adipose tissue explants led to both an enhance-ment in inflammatory gene expression and secretion ofcytokines, it is tempting to speculate that changes in

FIG. 4. Protein levels and gene expression of IL-1�, IL-6, and KC in murine eWAT explants treated with 3MA for 24 h. Cytokine was secreted insupernatant by eWAT explants from lean C57BL6 mice and obese leptin deficient (Lep ob ) mice (IL-1�, IL-6, and KC). The secretion levels areexpressed as the relative change in picograms per milliliter per milligram (cytokine/volume of supernatant/weight of explant). Mean values ofcytokine expression from lean control samples have been set as 100% (A). Relative gene expressions of IL-1�, IL-6, and KC in eWAT from leanmice and obese leptin-deficient (Lep ob ) mice (B). Control (Ctrl) samples were set as 1. Experiments have been repeated three times or more. Dataare presented as means � SEM. *, P � 0.05; **, P � 0.01; ***, P � 0.001.

FIG. 5. Effect of siRNA ATG7 on protein and gene expression of cytokines in human differentiated SGBS adipocytes. Relative gene expression ofATG7 in the stromal vascular fraction (SVF) and mature adipocytes (MAT) in human WAT (A). Relative gene expression of ATG7 after siRNAtreatment (B) is shown. Cytokine concentrations of IL-1 (indirectly by IL-2 production of NOB-1 assay), IL-6, and IL-8 secreted in the supernatant bySGBS cells after siRNA treatment (C) are also shown. Relative gene expression levels of IL-1�, IL-6, and IL-8 in SGBS cells after siRNA treatment (D)is shown. Control samples were set as 1. Data are presented as means � SEM. *, P � 0.05; ***, P � 0.001 (n � 3).


autophagy in adipocytes mainly contribute to the ob-served effects in line with a previous study (31). How-ever, inhibition of autophagy is also known to promoteinflammation in macrophages that are present in adi-pose tissue and profoundly affects its inflammatory sta-tus. Although it is challenging to identify the responsiblecell type in our experimental setup using adipose tissueexplants and treatment with 3MA may not effectivelyinhibit autophagy in all adipose tissue cell types, a closeinteraction between various cell types including adi-pocytes and macrophages most likely determines theinflammatory-modulating properties of autophagy inadipose tissue.

It is unclear at this point whether the inhibition of au-tophagy in adipose tissue worsens insulin sensitivity via itseffects on inflammatory cytokine production. In line withprevious work demonstrating the deleterious effects of in-flammation on insulin sensitivity (32, 33), one may hy-pothesize that the enhancement of autophagy in adiposetissue during obesity may serve to limit inflammation andprevent the further worsening of insulin resistance. Thishypothesis is supported by our observation that inhibitionof autophagy in obese adipose tissue severely enhances aproinflammatory response compared with lean adiposetissue. Thus, obesity-associated inflammation in adiposetissue seems to up-regulate autophagy to mitigate the pro-duction of proinflammatory cytokines. As such, au-tophagy activity seems a consequence, rather than a cause,of obesity-induced adipose tissue inflammation. It re-mains to be explored whether the (partial) defects in au-tophagy may contribute to the inflammatory process inobesity.

Our findings may appear somewhat counterintuitivebecause the absence of autophagy has been shown to pro-tect against obesity and insulin resistance in mice (14, 15).Indeed, adipocyte-specific ATG7 knockout mice exhibitincreased insulin sensitivity and are protected againsthigh-fat diet-induced obesity (13–15). However, the func-tion of autophagy during the development of adipocytesmay be quite different compared with its role in expandingadipose tissue in obese subjects. In line with our results, aprevious study demonstrated that obese individuals hadhigher levels of autophagy in their adipose tissue com-pared with lean subjects (16), suggesting that autophagymay actually chaperone fat mass expansion during obe-sity. Because LC3-II levels were closely correlated withCD68 influx and the presence of CLSs in adipose tissue,autophagy activation may actually be aimed at counter-balancing inflammation. Because variation in adipose tis-sue LC3 levels was observed in obese subjects, it might beworthwhile to further characterize the inflammatory sta-tus in future studies. Indeed, one may hypothesize that

metabolically healthy obese, which are characterized by alower inflammatory status of the adipose tissue (34), havea lower level of LC3-II in adipose tissue because the ne-cessity of the inflammatory dampening action of au-tophagy is diminished.

The exact mechanism underlying enhanced inflamma-tory gene expression upon autophagy inhibition in adi-pose tissue remains to be identified. One possible mech-anism may involve the intracellular degradation of pro-IL-1� by a process involving autophagy that would resultin lower secretion levels of IL-1� (35). Hence, inhibition ofautophagy would prevent intracellular degradation ofpro-IL-1� and lead to enhanced production of IL-1� byadipose tissue. Inhibition of autophagy may also promoteactivation of the pyrin domain containing 3 (NLRP3) in-flammasome in a reactive oxygen species-dependent man-ner and would thus produce more mature IL-1� due tomore caspase-1 cleavage (36, 37). Alternatively, nucle-otide-binding oligomerization domain containing 2(NOD2), which recognizes small peptides present in thebacterial wall, is important in regulating IL-1� gene ex-pression because it functions by either binding toATG16L1 to induce autophagy or by activating the ERK/nuclear factor-�B signaling pathway that induces IL-1�

gene expression. Autophagy inhibition may down-regu-late ATG16L1 expression, which could lead to moreNOD2-dependent stimulation of ERK/nuclear factor-�B.This hypothesis is supported by a recent study (12) re-porting that production of IL-1� by macrophages fromindividuals with the T300A variant of ATG16L1 is en-hanced upon stimulation of the NOD2 receptor. Finally,mitochondrial autophagy has been linked to insulin resis-tance (38) and may somehow regulate inflammatory geneexpression as well. Future studies should be aimed at de-ciphering the exact molecular mechanisms that facilitateproinflammatory responses upon inhibition of au-tophagic activity in adipose tissue on the development ofinflammation.

In summary, inhibition of autophagy stimulates proin-flammatory gene expression levels in both human and mu-rine WAT. In adipose tissue of obese mice, the inflamma-tion-promoting properties of autophagy inhibitionoutclass it effects in lean animals, even after correction fortotal fat mass. Therefore, we hypothesize that autophagyfunctions as a mechanism to control proinflammatorygene expression in adipose tissue to prevent chronic in-flammation. Altogether, this study provides new insidesshowing that autophagy affects the inflammatory status ofthe adipocyte and adipose tissue.


Acknowledgments

P.v.E., H.J.J., and R.S. researched the data and wrote the article.R.S., H.J.J., L.A.B.J., M.G.N., and C.J.T. planned the experi-ments, contributed to the discussion, and reviewed the article.

Address all correspondence and requests for reprints to: Dr.Henry Jansen, M.D., Radboud University Medical CentreNijmegen, Department of General Internal Medicine, 6500 HBNijmegen, The Netherlands. E-mail: [email protected].

R.S. was supported by a Ruby grant from the Dutch DiabetesResearch Foundation. M.G.N. was supported by a Vici grant ofThe Netherlands Organization for Scientific Research.

Disclosure Summary: The authors have no conflicts ofinterest.

References

1. Mizushima N, Levine B 2010 Autophagy in mammalian develop-ment and differentiation. Nat Cell Biol 12:823–830

2. Reggiori F, Klionsky DJ 2005 Autophagosomes: biogenesis fromscratch? Curr Opin Cell Biol 17:415–422

3. Yoshimori T, Noda T 2008 Toward unraveling membrane biogen-esis in mammalian autophagy. Curr Opin Cell Biol 20:401–407

4. Delgado M, Singh S, De Haro S, Master S, Ponpuak M, Dinkins C,Ornatowski W, Vergne I, Deretic V 2009 Autophagy and patternrecognition receptors in innate immunity3. Immunol Rev 227:189–202

5. Gannagé M, Münz C 2009 Autophagy in MHC class II presentationof endogenous antigens. Curr Top Microbiol Immunol 335:123–140

6. Scarlatti F, Granata R, Meijer AJ, Codogno P 2009 Does autophagyhave a license to kill mammalian cells? Cell Death Differ 16:12–20

7. Kuballa P, Nolte WM, Castoreno AB, Xavier RJ 2012 Autophagyand the immune system. Annu Rev Immunol 30:611–646

8. Harris J 2011 Autophagy and cytokines. Cytokine 56:140–1449. Saitoh T, Fujita N, Jang MH, Uematsu S, Yang BG, Satoh T, Omori

H, Noda T, Yamamoto N, Komatsu M, Tanaka K, Kawai T, Tsu-jimura T, Takeuchi O, Yoshimori T, Akira S 2008 Loss of the au-tophagy protein Atg16L1 enhances endotoxin-induced IL-1� pro-duction. Nature 456:264–268.

10. Fujishima Y, Nishiumi S, Masuda A, Inoue J, Nguyen NM, Irino Y,Komatsu M, Tanaka K, Kutsumi H, Azuma T, Yoshida M 2011Autophagy in the intestinal epithelium reduces endotoxin-inducedinflammatory responses by inhibiting NF-�B activation. ArchBiochem Bisphys 506:223–235.

11. Crisan TO, Plantinga TS, van d, V, Farcas MF, Stoffels M, KullbergBJ, van der Meer JW, Joosten LA, Netea MG 2011 Inflammasome-independent modulation of cytokine response by autophagy in hu-man cells. PLoS One 6:e18666

12. Plantinga TS, Crisan TO, Oosting M, van de Veerdonk FL, de JongDJ, Philpott DJ, van der Meer JW, Girardin SE, Joosten LA, NeteaMG 2011 Crohn’s disease-associated ATG16L1 polymorphismmodulates pro-inflammatory cytokine responses selectively uponactivation of NOD2. Gut 60:1229–1235

13. Goldman S, Zhang Y, Jin S 2010 Autophagy and adipogenesis: im-plications in obesity and type II diabetes. Autophagy 6:179–181

14. Singh R, Xiang Y, Wang Y, Baikati K, Cuervo AM, Luu YK, TangY, Pessin JE, Schwartz GJ, Czaja MJ 2009 Autophagy regulatesadipose mass and differentiation in mice. J Clin Invest 119:3329–3339

15. Zhang Y, Goldman S, Baerga R, Zhao Y, Komatsu M, Jin S 2009Adipose-specific deletion of autophagy-related gene 7 (atg7) in mice

reveals a role in adipogenesis. Proc Natl Acad Sci USA 106:19860–19865

16. Kovsan J, Blüher M, Tarnovscki T, Klöting N, Kirshtein B, MadarL, Shai I, Golan R, Harman-Boehm I, Schön MR, Greenberg AS,Elazar Z, Bashan N, Rudich A 2011 Altered autophagy in humanadipose tissues in obesity. J Clin Endocrinol Metab 96:E268–E277

17. Ost A, Svensson K, Ruishalme I, Brännmark C, Franck N, Krook H,Sandström P, Kjolhede P, Strålfors P 2010 Attenuated mTOR sig-naling and enhanced autophagy in adipocytes from obese patientswith type 2 diabetes. Mol Med 16:235–246

18. Ouchi N, Parker JL, Lugus JJ, Walsh K 2011 Adipokines in inflam-mation and metabolic disease. Nat Rev Immunol 11:85–97

19. Schenk S, Saberi M, Olefsky JM 2008 Insulin sensitivity: modulationby nutrients and inflammation. J Clin Invest 118:2992–3002

20. Lagathu C, Yvan-Charvet L, Bastard JP, Maachi M, Quignard-Bou-langé A, Capeau J, Caron M 2006 Long-term treatment with inter-leukin-1� induces insulin resistance in murine and human adi-pocytes. Diabetologia 49:2162–2173

21. McGillicuddy FC, Harford KA, Reynolds CM, Oliver E, ClaessensM, Mills KH, Roche HM 2011 Lack of interleukin-1 receptor I(IL-1RI) protects mice from high-fat diet-induced adipose tissue in-flammation coincident with improved glucose homeostasis. Diabe-tes 60:1688–1698

22. Jager J, Grémeaux T, Cormont M, Le Marchand-Brustel Y, Tanti JF2007 Interleukin-1�-induced insulin resistance in adipocytesthrough down-regulation of insulin receptor substrate-1 expression.Endocrinology 148:241–251

23. Nov O, Kohl A, Lewis EC, Bashan N, Dvir I, Ben-Shlomo S, FishmanS, Wueest S, Konrad D, Rudich A 2010 Interleukin-1� may mediateinsulin resistance in liver-derived cells in response to adipocyte in-flammation. Endocrinology 151:4247–4256

24. Duvnjak L, Duvnjak M 2009 The metabolic syndrome—an ongoingstory. J Physiol Pharmacol 60(Suppl 7):19–24

25. Drel VR, Mashtalir N, Ilnytska O, Shin J, Li F, Lyzogubov VV,Obrosova IG 2006 The leptin-deficient (ob/ob) mouse: a new animalmodel of peripheral neuropathy of type 2 diabetes and obesity. Di-abetes 55:3335–3343

26. Wabitsch M, Brenner RE, Melzner I, Braun M, Möller P, Heinze E,Debatin KM, Hauner H 2001 Characterization of a human prea-dipocyte cell strain with high capacity for adipose differentiation. IntJ Obes Relat Metab Disord 25:8–15

27. Netea MG, Demacker PN, Kullberg BJ, Boerman OC, Ver-schueren I, Stalenhoef AF, van der Meer JW 1996 Low-densitylipoprotein receptor-deficient mice are protected against lethalendotoxemia and severe gram-negative infections. J Clin Invest97:1366 –1372

28. Gearing AJ, Bird CR, Bristow A, Poole S, Thorpe R 1987 A simplesensitive bioassay for interleukin-1 which is unresponsive to 10(3)U/ml of interleukin-2. J Immunol Methods 99:7–11

29. Ireland JM, Unanue ER 2011 Autophagy in antigen-presenting cellsresults in presentation of citrullinated peptides to CD4 T cells. J ExpMed 208:2625–2632

30. Levine B, Mizushima N, Virgin HW 2011 Autophagy in immunityand inflammation. Nature 469:323–335

31. Yoshizaki T, Kusunoki C, Kondo M, Yasuda M, Kume S, MorinoK, Sekine O, Ugi S, Uzu T, Nishio Y, Kashiwagi A, Maegawa H2012Autophagy regulates inflammation in adipocytes. Biochem BiophysRes Commun 417:352–357

32. Farb MG, Bigornia S, Mott M, Tanriverdi K, Morin KM, Freed-man JE, Joseph L, Hess DT, Apovian CM, Vita JA, Gokce N 2011Reduced adipose tissue inflammation represents an intermediatecardiometabolic phenotype in obesity. J Am Coll Cardiol 58:232–237

33. Lê KA, Mahurkar S, Alderete TL, Hasson RE, Adam TC, Kim JS,Beale E, Xie C, Greenberg AS, Allayee H, Goran MI 2011 Subcu-taneous adipose tissue macrophage infiltration is associated with


hepatic and visceral fat deposition, hyperinsulinemia, and stimula-tion of NF-�B stress pathway. Diabetes 60:2802–2809

34. Barbarroja N, López-Pedrera R, Mayas MD, García-Fuentes E,Garrido-Sánchez L, Macias-González M, El Bekay R, Vidal-Puig A,Tinahones FJ 2010 The obese healthy paradox: is inflammation theanswer? Biochem J 430:141–149

35. Harris J, Hartman M, Roche C, Zeng SG, O’Shea A, Sharp FA, LambeEM, Creagh EM, Golenbock DT, Tschopp J, Kornfeld H, FitzgeraldKA, Lavelle EC 2011 Autophagy controls IL-1� secretion by targetingpro-IL-1� for degradation. J Biol Chem 286:9587–9597

36. Nakahira K, Haspel JA, Rathinam VA, Lee SJ, Dolinay T, LamHC, Englert JA, Rabinovitch M, Cernadas M, Kim HP, FitzgeraldKA, Ryter SW, Choi AM 2011 Autophagy proteins regulate in-nate immune responses by inhibiting the release of mitochondrialDNA mediated by the NALP3 inflammasome. Nat Immunol 12:222–230

37. Zhou R, Yazdi AS, Menu P, Tschopp J 2011 A role for mitochondriain NLRP3 inflammasome activation. Nature 469:221–225

38. Jung HS, Lee MS 2010 Role of autophagy in diabetes and mito-chondria. Ann NY Acad Sci 1201:79–83


Documents

1-Autophagy Activity is Up-Regulated in Adipose Tissue