Adipose Tissue Insulin Action and IL-6 Signaling …...Adipose Tissue Insulin Action and IL-6 Signaling after Exercise in Obese Mice REBECCA E. K. MACPHERSON, JASON S. HUBER, SCOTT

Adipose Tissue Insulin Action and IL-6Signaling after Exercise in Obese Mice

REBECCA E. K. MACPHERSON, JASON S. HUBER, SCOTT FRENDO-CUMBO, JEREMY A. SIMPSON,and DAVID C. WRIGHT

Department of Human Health and Nutritional Sciences, University of Guelph, Guelph, Ontario, CANADA

ABSTRACT

MACPHERSON, R. E. K., J. S. HUBER, S. FRENDO-CUMBO, J. A. SIMPSON, and D. C. WRIGHT. Adipose Tissue Insulin Action

and IL-6 Signaling after Exercise in Obese Mice.Med. Sci. Sports Exerc., Vol. 47, No. 10, pp. 2034–2042, 2015. Introduction: Adipose

tissue insulin action is impaired in obesity and is associated with inflammation, macrophage infiltration, and polarization toward a

proinflammatory phenotype. Acute exercise can reduce markers of adipose inflammation, including interleukin (IL) 6, in parallel with

improvements in insulin action; however, others have provided evidence that IL-6 has anti-inflammatory properties. Purpose: This study

aimed to examine the relation between IL-6 signaling, macrophage infiltration, and polarization and insulin action in inguinal fat after

acute exercise in obese, insulin-resistant mice. Methods: Male C57BL/6 mice were fed a low-fat diet (10% kcal lard) or a high-fat diet

(HFD, 60% kcal lard) for 7 wk and then underwent an acute bout of exercise (2-h treadmill running: 15 mIminj1, 5% incline). Results:

The HFD resulted in increased body mass, glucose intolerance, and attenuated insulin-induced AKT Thr308 phosphorylation in inguinal

fat. This was accompanied by increases in indices of macrophage infiltration (F4/80, CD68, and monocyte chemoattractant protein-1

expression) and polarization toward an M1 phenotype (increased expression of CD11c, CD11c/galactose-type C-type lectin 1, and

inducible nitric oxide synthase). Immunofluorescence imaging demonstrated increased F4/80- and CD11c-positive cells with the HFD.

Two hours after exercise, the insulin-induced activation of AKT Th308 phosphorylation was recovered in HFD mice. This was ac-

companied by an upregulation of IL-6 and IL-10 signaling, as demonstrated by increased expression of IL-6, IL-10, and SOCS3 as well

as STAT3 phosphorylation. Furthermore, acute exercise resulted in a shift toward reduction in M1 polarization indicated by a decrease in

the ratio of CD11c to galactose-type C-type lectin 1 mRNA as well as a decline in F4/80- and CD11c-positive cells. Conclusions: The

results suggest a link between exercise-induced increases in IL-6, reductions in indices of M1 macrophages, and increased IL-10, a

reputed anti-inflammatory cytokine with insulin-sensitizing properties. Key Words: INFLAMMATION, MACROPHAGE, ACUTE

EXERCISE, ADIPOSE TISSUE

Adipose tissue macrophage infiltration and the re-sulting inflammation have been shown to be causalevents in the pathogenesis of insulin resistance that

occurs with obesity (19). Adipose tissue macrophages liealong a continuum and orientate toward either a predomi-nantly proinflammatory (M1) or anti-inflammatory (M2)phenotype (19). M1 macrophages accumulate with obesityand secrete proinflammatory cytokines, such as tumor necro-sis factor alpha (TNF>), interleukin (IL) 6, inducible nitricoxide synthase (iNOS), and monocyte chemoattractant pro-tein (MCP) 1 (6,31) and have been directly linked to the ac-tivation of mitogen-activated protein kinases (MAPK) and the

development of insulin resistance (5,20,31). For example,TNF> is known to activate several serine kinases, such asJNK and p38 MAPK, which attenuate proximal steps in theinsulin signaling pathway (34). M2 macrophages are pre-dominant in the adipose tissue of healthy lean animals andproduce anti-inflammatory cytokines (19,20), such as IL-10(19,20), and are characterized by the presence of macrophagegalactose-type C-type lectin 1 (MGL-1) (19).

Exercise training is a useful tool to either prevent or treatobesity-related adipose tissue macrophage polarization, in-flammation, and insulin resistance (1,15). Specifically, ex-ercise training initiated at the onset of a high-fat diet (HFD)protects against increases in markers of adipose tissue in-flammation and indices of macrophage infiltration, such asIL-6, TNF>, and MCP-1 (14,15). Similarly, exercise train-ing that is initiated once obesity has already developed leadsto reductions in adipose tissue inflammation (36). However,these findings are complicated because exercise training re-sults in reductions in adipose tissue mass, which is typicallyshadowed by macrophage recruitment and inflammation(1,13–15). Because of the resultant fluctuations in adiposetissue mass with training, the underlying effects of exerciseper se on markers of adipose tissue macrophages and inflam-mation have not been clearly elucidated.

Address for correspondence: Rebecca MacPherson, Ph.D., Department ofHuman Health and Nutritional Sciences, 334 Animal Sciences Building,50 Stone Road, East Guelph, Ontario N1G 2W1, Canada; E-mail:[email protected] for publication October 2014.Accepted for publication March 2015.

0195-9131/15/4710-2034/0MEDICINE & SCIENCE IN SPORTS & EXERCISE�Copyright � 2015 by the American College of Sports Medicine

DOI: 10.1249/MSS.0000000000000660

2034

BASICSC

IENCES

Copyright © 2015 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.

An acute bout of exercise has previously been reported toinduce markers of inflammation in adipose tissue from leanhealthy rats (27). The underlying cause of this inflammationmay be the rapid increase in adipose tissue lipolysis andfatty acid release that occur with exercise (10). In agreementwith this supposition, recent work has demonstrated thatthe beta-3 adrenergic agonist (CL 316,243) rapidly inducesmarkers of inflammation in mouse adipose tissue in vivoparallel with large increases in adipose tissue lipolysis (3,23).Moreover, the induction of proinflammatory genes after CL316,243 treatment is abrogated when fatty acid release is at-tenuated (23). Together, these studies reveal an important roleof exercise-induced lipolysis in initiating inflammatory re-sponse in adipose tissues of lean mice.

In contrast to exercise in lean rats, it has recently beenreported that markers of inflammation such as IL-6, IL-1 beta,and TNF> as well as markers of M1 macrophage polarizationwere reduced in epididymal adipose tissues of obese rats 2 hafter 6 h of swimming exercise (24,25). The decrease in ad-ipose tissue IL-6 expression after exercise in obese rats, de-spite a shift in macrophage polarization and improvements inadipose tissue insulin action, is particularly interesting, as ithas been proposed that IL-6 plays a crucial role in the alter-native activation of macrophages in obesity (22).

Given these apparent discrepancies in the role of IL-6after exercise, the purpose of the current study was to re-examine the effects of an acute bout of exercise on markers ofinflammation and insulin action in the setting of preexistingobesity and insulin resistance. To examine this question, weused treadmill running, instead of swimming, because weviewed this as a more clinically relevant mode of exercise.Indices of inflammation, macrophage infiltration/polarization,and insulin signaling were examined in inguinal subcutaneousadipose tissue, a fat depot proven to play a significant role inthe regulation of whole body glucose homeostasis (33,35) andwhich responds more robustly to exercise than visceral adi-pose tissue depots (2). We hypothesized that an acute boutof exercise would induce IL-6, and this would be associatedwith reductions in markers of M1 macrophages and im-provements in insulin signaling.

METHODS

Materials. Molecular weight marker, reagents, and nitro-cellulose membranes for SDS-PAGE were purchased fromBio-Rad (Mississauga, Ontario, Canada). ECL Plus was aproduct of Amersham Pharmacia Biotech (Arlington Heights,IL). Horseradish peroxidase-conjugated donkey anti-rabbitand goat anti-mouse IgG secondary antibodies were fromJackson ImmunoResearch Laboratories (West Grove, PA).Antibodies against total AKT (CAT# 4685), phosphothreonine308AKT (CAT# 4056), total STAT3 (CAT# 8768S), phospho-STAT3 (CAT# 9138S), total JNK (CAT# 9252S), phospho-JNK (CAT# 4671S), total ERK (CAT# 4695S), phospho-ERK(CAT# 9101S), total p38 (CAT# 9212S), and phospho-p38 (CAT# 9211S) were purchased from Cell Signaling

(Danvers, MA); F4/80 (60343, 2 KgImLj1, FITC conju-gated), CD11c (33483, 2 KgImLj1), and Alexa Fluor 568(175716, 2 KgImLj1) were from Abcam (Toronto, Ontario,Canada). Random primers, SuperScript II Reverse Tran-scriptase, and dNTP were from Invitrogen (Burlington,Ontario, Canada). Taqman gene expression assays for mouseIL-6 (Mn00446190_m1), SOCS-3 (Mm00545913_s1), TNF>(Mn00443258_m1), MCP-1 (Mm00441242_m1), F4/80(Mm00802529_m1), IL-10 (Mm00439614_m1), CD68(Mm03047340_m1), iNOS (Mm00440502_m1), and GAPDH(4352932E) were from Applied Biosystems (Foster City, CA).All other reagents were from Sigma-Aldrich.

Animals and diet. Male C57BL/6 mice (Charles River)were fed a low-fat diet (LFD) (10% kcal from lard; ResearchDiets D12450J; n = 9) or an HFD (60% kcal from lard;Research Diets D12492; n = 18) ad libitum for 7 wk. In pre-liminary experiments, we found that this duration of high-fatfeeding induced adipose tissue insulin resistance and markersof macrophage infiltration. Animals were housed individually,had free access to water, and were maintained on a 12/12-hlight cycle. All protocols were approved by the University ofGuelph Animal Care Committee and met the guidelines of theCanadian Council on Animal Care.

Glucose tolerance. Intraperitoneal glucose tolerancetests were performed on fasted (6 h), nonanesthetized miceduring the last week of feeding. Glucose measures were ob-tained from tail vein blood using an automated glucometerat baseline and at 15, 30, 45, 60, 90, and 120 min after intra-peritoneal injection of glucose (2 gIkgj1 body mass).

Acute exercise protocol. To minimize differencesbetween groups and in stress, both LFD and HFD mice wereacclimated to motorized treadmill running during a 3-d pe-riod consisting of 15 minIdj1 of running at 15 mIminj1 and5% grade. Mice were maintained on their respective dietsduring this time. Mice were assigned into one of the fol-lowing three groups: sedentary LFD (LFD, n = 9), sedentaryHFD (HFD, n = 9), and exercised HFD (HFD + ex, n = 9).Seventy-two hours after the last day of acclimation, mice ranfor 120 min at 15 mIminj1, with an incline of 5%. The acutebout of exercise started at approximately 10:00 a.m., whichis the beginning of the light cycle. From a previous work inour laboratory (38), we have found that this duration andintensity of exercise are well tolerated; all mice in the ex-ercise treatment group completed the 2-h treadmill runningwithout issue. After the exercise bout, mice were placedback in their cages to recover for 2 h without food. Seden-tary low- and high-fat-fed mice had their food removed atthe same time as the high-fat-fed mice that exercised.

Insulin stimulation. Two hours after exercise, micewere anesthetized with a weight-adjusted bolus injection ofsodium pentobarbital (5 mg per 100-g body weight). The leftinguinal fat depot was dissected and immediately frozen inliquid nitrogen and stored at j80-C until further analysis.Mice were then injected in the peritoneal cavity with aweight-adjusted bolus of insulin (10.0 UIkgj1 BW). Thisdosage of insulin is typically used to measure responsiveness

EXERCISE-INDUCED IL-6 SIGNALING IN OBESE MICE Medicine & Science in Sports & Exercised 2035

BASIC

SCIEN

CES


of the insulin signaling pathway in vivo (17). Fifteen minutesafter injection, the contralateral fat depot was rapidly ex-cised and immediately frozen in liquid nitrogen and storedat j80-C. Sedentary LFD and HFD mice underwent thesame manipulations.

Real-time polymerase chain reaction. Total RNAwas extracted via Trizol reagent and reverse-transcribed intocDNA. Changes in mRNA expression were determined usingreal-time quantitative polymerase chain reaction (PCR), asdescribed in detail previously by our laboratory (3). Briefly,RNA was isolated from inguinal adipose tissue using anRNeasy kit according to the manufacturer_s instructions(RNeasy Kit, 74104; Quiagen). Quantity and purity wereassessed using a NanoDrop system (NanoDrop 2000 Spectro-photometer; Thermo Scientific). cDNA was synthesized fromtotal RNA (1ug) using SuperScript II Reverse Transcriptase,dNTP, and random primers (Invitrogen, Burlington, Ontario,Canada). Real-time PCR was carried out using a 7500 fastreal-time PCR system (Applied Biosystems). Samples wereloaded in triplicate using a 96-well plate layout. Each wellcontained a total volume of 20 KL comprised of 1-KL geneexpression assay, 1-KL cDNA template, 10-KL Taqman fastuniversal PCR master mix, and 8-KL RNase-free water.GAPDH was used as our housekeeping gene, and relativedifferences in gene expression between groups were deter-mined using the 2-$$CT method (18). PCR efficiency wassimilar between GAPDH and our genes of interest. Similarly,our experimental manipulations did not alter the expressionof our housekeeping gene (GAPDH).

Western blotting. Samples were homogenized (FastPrep�-24; MP Biomedicals, Santa Ana, CA) in cell lysis buffer (threevolumes) supplemented with phenylmethylsulfonyl fluorideand protease inhibitor cocktail (Sigma-Aldrich). Homoge-nized samples were centrifuged at 4-C (10 min at 5000g), theinfranatant was collected, and protein content was determinedusing a bicinchoninic acid assay (32). Samples were preparedto contain equal concentrations of protein that were then sep-arated on 10% SDS-PAGE gels. Protein was wet-transferredonto nitrocellulose membranes at 100 V per transfer unit.Membranes were blocked in Tris-buffered saline/0.1% tween20 (TBST) prepared with 5% nonfat dry milk for 1 h followedby overnight incubation at 4-C with the appropriate primaryantibody. After primary incubation, membranes were rinsedwith TBST and incubated with the appropriate horseradishperoxidase-conjugated secondary antibodies for 1 h at roomtemperature (Jackson ImmunoResearch Laboratories, WestGrove, PA). Phospho and total blots were run on duplicategels. Signals were detected using enhanced chemiluminesenceand were subsequently quantified by densitometry using aFluorChem HD imaging system (Alpha Innotech, SantaClara, CA). Phosphorylated proteins were expressed relativeto total.

Histological analysis. Inguinal adipose tissue was fixedin 10% neutral-buffered formalin (VWR,Mississauga, Ontario,Canada) dehydrated in xylene (Fisher Scientific) and em-bedded in paraffin. Five-micrometer sections were mounted

on 1.2-mm Superfrost slides stained with modified Harrishematoxylin and eosin stock with phloxine (Fisher Scientific)and imaged (Olympus FSX 100 light microscope; Olympus,Tokyo, Japan). Cells (approximately 100) from were sampledin each image to determine cross-sectional area (ImageJ soft-ware; National Institute of Mental Health, Bethesda, MD).

Immunofluorescence. Inguinal adipose tissue was em-bedded in Tissue-Tek optimal cutting temperature compoundatj80-C. The adipose tissue was then cut into 30-Km sectionsat j35-C and mounted on charged 1.2-mm Superfrost slides(Fisher Scientific). Frozen sections were then fixed in acetoneat j20-C. The slides were then incubated with anti-F4/80overnight at +4-C for the identification of macrophages,washed, then incubated with primary anti-CD11c and sec-ondary Alexa Fluor 568 to identify CD11c-positive cells.Images were acquired using Olympus FSX 100 light micro-scope and analyzed using Cell Sense software (Olympus,Tokyo, Japan). As sections were 30-Km thick, macrophageswere present at different focal planes when using light mi-croscopy, resulting in varying intensities of fluorescence signalfor macrophages. Cells that were F4/80+ and F4/80+CD11c+were counted per square centimeter in two images per animal.Image analysis was completed in a blinded fashion by anindividual experienced with this technique.

Statistics. Comparisons between LFD and HFD groupswere made using unpaired, two-tailed Student_s t-tests.Differences in insulin-induced Akt phosphorylation weredetermined with a two-way repeated-measures ANOVA fol-lowed by least significant difference post hoc analysis. Differ-ences in protein content, gene expression, and glucose overtime were determined using one-way ANOVA followed byleast significant difference post hoc analysis. In cases wheredata were not normally distributed, data were logarithmicallytransformed. Data are expressed as means T SEM with sig-nificance set at P G 0.05.

RESULTS

High-fat feeding increased body and adiposetissue mass and induced glucose intolerance. Micefed with the HFD for 7 wk gained more weight than the micefed the LFD (P G 0.05) (Fig. 1A and B). Similar to differ-ences in body mass, inguinal adipose tissue mass of the micefed with the HFD was greater than that of the LFD mice (P G0.05) (Fig. 1C). HFD mice cleared glucose less effectivelyafter an intraperitoneal glucose injection and had a higherglucose area under the curve compared with that of LFDmice (P G 0.05) (Fig. 1D and E). These results demonstratethat 7 wk of HFD resulted in obese, glucose-intolerant mice.

Acute exercise improved insulin signaling inde-pendent of reductions in MAPK signaling. To ex-amine the effects of an acute bout of exercise on adipose tissueinsulin action, we measured insulin-induced AKT phosphor-ylation 2 h after an acute bout of exercise. This time point waschosen because others have shown improvements in adiposetissue insulin signaling at this point after acute exercise in

http://www.acsm-msse.org2036 Official Journal of the American College of Sports Medicine

BASICSC

IENCES


obese rats (24). In LFD mice, insulin resulted in approxi-mately fourfold increase in AKT phosphorylation that wasabsent in sedentary HFD mice (Fig. 2A). Two hours after anacute bout of exercise, insulin-stimulated AKT phosphoryla-tion in inguinal adipose tissue was restored in HFD mice. Nochanges in basal phosphorylation or total AKT protein levelswere observed in any group. The ERK, p38, and JNK MAPKare involved in the cellular response to inflammatory cyto-kines (5), thus linking them to insulin resistance. Diet had noeffect on the phosphorylation of JNK, ERK, or p38. How-ever, 2 h after exercise, p38 phosphorylation was significantlyincreased (P G 0.05) (Fig. 2B).

Acute exercise increases indices of inflammationand IL-6 signaling in HFD mice. To characterize theeffects of exercise on markers of inflammation, we measuredthe mRNA expression of IL-6, SOCS3, TNF>, and MCP-1in inguinal adipose tissue. Sedentary HFD mice did notdemonstrate any changes in IL-6, TNF>, or SOCS3 mRNAexpression; however, MCP-1 mRNA expression was in-creased approximately twofold. Two hours after exercise ces-sation, expression of IL-6, SOCS3, and MCP-1 was increasedapproximately 10-fold (Fig. 3A). IL-6, among other cytokines,signals through the JAK/STAT signaling pathway. Similar tochanges in the expression of IL-6, SOSC3, and MCP-1, therewere no differences in STAT3 phosphorylation in inguinaladipose tissue from sedentary LFD and HFD mice. However,2 h after exercise, STAT3 phosphorylation was increasedapproximately twofold (Fig. 3B) (P G 0.05). This finding, inconjunction with the increase in SOCS3 mRNA expression, areputed transcriptional target of IL-6 (4,7), suggests that IL-6signaling is activated in inguinal adipose tissue from mice fedwith an HFD after exercise.

Acute exercise decreased indices of M1 macro-phages. Markers of macrophage infiltration, F4/80 andCD68, were significantly increased in inguinal adipose tissueafter HFD. Exercise had no effect on the expression of thesemacrophage markers (Fig. 4A). M1 macrophages can becharacterized by increased expression of iNOS (31). High-fatfeeding resulted in increased iNOS expression, and this wasreduced 2 h after exercise (Fig. 4B). M1 and M2 macro-phages can further be distinguished by the presence or ab-sence of the M1 marker, CD11c, and the M2 marker, MGL-1(19). The ratio of CD11c to MGL-1 gene expression is auseful measure of the relative polarization of macrophages(19). HFD led to marked increase in the mRNA expressionof CD11c, whereas MGL-1 expression was unaffected. Twohours after exercise, the expression of CD11c and the ratio ofCD11c to MGL-1 were significantly reduced in compari-son with HFD sedentary samples (Fig. 4B). To confirm thechanges in gene expression, we measured F4/80 and CD11cprotein using immunofluorescence. As shown in Figure 5, theconsumption of an HFD for 7 wk led to increases in inguinaladipose tissue fat cell size and F4/80-positive cells and anacute bout of exercise did not alter these end points. F4/80-and CD11c-positive cells were increased in inguinal adiposetissue from mice fed with an HFD. In keeping with our geneexpression data, the percentages of F4/80- and CD11c-positive cells were reduced 2 h after an acute bout of exercise.

Acute exercise increases the expression IL-10. Anti-inflammatory cytokines are expressed to a higher degree inadipose tissue from lean healthy mice in comparison with thatin obese mice, and previous work has demonstrated that about of acute exercise increases IL-10 and IL-4 expressionin adipose tissues of rats fed with an HFD (24). Furthermore,

FIGURE 1—Characterization of diet-induced obesity and glucose intolerance model. A. Body mass. B. Body mass gain over 7 wk of diet. C. Inguinalfat pad mass. D. Intraperitoneal glucose tolerance test (body weight, 2 gIkgj1). E. Blood glucose area under the curve. Data are presented as means TSEM of LFD (n = 9) and HFD (n = 18) (note that the HFD group includes both the sedentary and exercised mice). *P G 0.05.


BASIC

SCIEN

CES


both IL-4 and IL-10 have been demonstrated to induce al-ternative macrophage phenotype (12,19). In inguinal adiposetissue, IL-4 and IL-10 were not altered in sedentary HFDmice. However, 2 h after exercise, IL-10 was increased ap-proximately fivefold (Fig. 6).

DISCUSSION

Adipose tissue macrophages are known to be significantcontributors to inflammation in obesity and mediators of in-sulin resistance (11,19). Exercise training provides a potentialtherapeutic treatment to resolve adipose tissue inflammationand M1 macrophage activation; however, these results typi-cally happen with declines in adipose tissue mass occurringwith the training (1,13–15) and the effects of exercise per seremain relatively unstudied. The current investigation used anacute bout of exercise to characterize changes in the expressionof proinflammatory cytokines and markers of macrophage in-filtration as well as activation in mice with diet-induced obesity

and insulin resistance mice. Novel results from this study in-dicate that the recovery of insulin signaling in inguinal adiposetissue after an acute bout of exercise in obese mice is associ-ated with increases in markers of IL-6 and IL-10 signaling anddecreases in M1 macrophages.

Seven weeks of high-fat feeding resulted in attenuation ofinsulin-stimulated AKT phosphorylation in inguinal adiposetissue. This was associated with increases in macrophage in-filtration (increased F4/80 protein and mRNA expression ofF4/80 and CD68), M1 polarization (increased F4/80- andCD11c-positive cells as well as mRNA expression of iNOSand CD11c), and expression of MCP-1, a chemotaxic factorthought to play a role in macrophage infiltration (11,30). In-sulin signaling was recovered after the acute bout of exercise;however, this was accompanied by an increase in IL-6 sig-naling (SOCS3 expression (4) and STAT3 phosphorylation(7)). The specific mechanisms driving the increase in IL-6expression in adipose tissue with exercise are not known.However, recent evidence suggests that increases in fatty acid

FIGURE 2—Acute exercise improved insulin signaling independent of reductions in MAPK signaling. A. HFD-induced reductions in AKT phos-phorylation are rescued by an acute bout of exercise. B. Diet had no effect on the phosphorylation of JNK, ERK, or p38. Two hours after exercise, p38phosphorylation was significantly increased. Representative blots are shown beside (A) or below (B) the quantified data (LFD, n = 9; HFD, n = 9;HFD+ ex 2 h after exercise, n = 9; + indicates insulin stimulated conditions). Data are presented as means T SEM. *P G 0.05.


BASICSC

IENCES


release (23) and sphingosine kinase expression/activity (39)are involved in the regulation of IL-6 expression in adipo-cytes. Although we do not have direct evidence demonstrat-ing exercise-induced increases in adipose tissue lipolysis withour exercise model, the previously mentioned mechanismscould likely be involved in regulating the effects of exerciseon induction of IL-6.

The increase in IL-6 is an interesting finding, as IL-6 isgenerally viewed as a proinflammatory cytokine with in-creased expression in adipose tissue from obese, insulin-resistant animals (16). However, the inflammatory role of

IL-6 remains controversial and studies indicate that IL-6 mayinstead have positive immunometabolic effects (22). In sup-port of this, IL-6–deficient mice develop late-onset obesityand resistance to insulin (37). Moreover, when fed an HFD,IL-6–deficient mice develop enhanced inflammation and de-teriorations in glucose metabolism when compared with wild-type controls (21). Similarly, the myeloid-specific ablation ofthe IL-6 receptor leads to greater development of glucoseintolerance in mice fed with an HFD when compared withthat in wild-type controls, and this is associated with increasesin markers of M1 macrophages in insulin-sensitive tissues

FIGURE 3—Acute exercise increases IL-6 signaling in inguinal adipose tissue from obese mice. A. No effect of diet was observed on the expression ofIL-6, TNF>, or SOCS3, whereas MCP-1 expression was increased. Two hours after exercise, IL-6, SOCS3, and MCP-1 expression was increased. B.Acute exercise resulted in increased STAT3 phosphorylation two hours after exercise. Representative blots are shown beside the quantified data (B)(LFD, n = 9; HFD, n = 9; HFD+ ex 2 h after exercise, n = 9). Data are presented as means T SEM. * and # indicate significantly different values fromthose in other groups at P G 0.05.

FIGURE 4—Effects of diet and acute exercise on the expression of markers of macrophage infiltration and polarization. A. The HFD led to increasesin F4/80 and CD68 gene expression. B. High-fat feeding increased the expression of iNOS, CD11c, and the ratio of CD11c to MGL-1, with acuteexercise resulting in a decline in these endpoints (LFD, n = 9; HFD, n = 9; HFD+ ex 2 h after exercise, n = 9). Data are presented as means T SEM. * and# indicate significantly different values from those in other groups at P G 0.05.


BASIC

SCIEN

CES


(22). In keeping with these findings in the present study, wefound that an acute bout of exercise, parallel with increases inreputed markers of IL-6 signaling, i.e., SOCS3 and p-STAT3(4,7), led to reductions in F4/80- and CD11c-positive cells

as well as the ratio of CD11c to MGL-1 mRNA, providingevidence of a reduction in M1 macrophage polarization. Al-though the results from our immunohistological staining donot evaluate the absolute number of infiltrated macrophages,

FIGURE 5—Histological analysis of inguinal adipose tissue in LFD, HFD, and HFD + ex. A. Representative images of inguinal adipose tissue stainedwith hematoxylin and eosin, F4/80, and CD11c. B. Quantification of adipocyte cross-sectional area from hematoxylin and eosin stain (n = 5 per group,100 cells analyzed per animal). C. Quantification of F4/80-positive macrophages per square centimeter using immunofluorescence staining. D.Quantification of the percent of F4/80- and CD11c-positive macrophages per square centimeter using immunofluorescence staining. * indicatessignificance when compared with LFD (P G 0.05), and # indicates significance when compared with HFD (P G 0.05; n = 5 per group, two imagesanalyzed per animal for immunofluorescence).

FIGURE 6—Effects of diet and acute exercise on the expression of anti-inflammatory cytokines. There was no effect of high-fat feeding on theexpression of IL-4 or IL-10; however, 2 h after exercise, IL-10 expression was significantly increased (LFD, n = 9; HFD, n = 9; HFD+ ex 2 h afterexercise, n = 9). Data are presented as means T SEM. *P G 0.05.


BASICSC

IENCES


as can be done with flow cytometry, we believe that our re-sults provide sufficient evidence of relative changes in mac-rophage polarization.

In contrast to our findings, Oliveira et al. (24,25) foundthat the insulin-induced activation of AKT in epididymaladipose tissue was increased in obese rats 2 h after a singlebout of exercise; however, these changes were associatedwith reductions in the mRNA expression of inflammatoryfactors such as TNF>, IL-1A, and IL-6. The discrepancybetween the present study and that by Oliveira et al. may bedue to the mode of acute exercise (24,25). Oliveira et al.(24,25) used a strenuous 6-h swimming protocol, whereaswe used a 2-h treadmill running protocol that may beconsidered more physiologically relevant. Although the ad-ipose tissue depot examined could also play a role in thedifferences between studies, we find this to be unlikely be-cause IL-6, TNF>, and MCP-1 were also increased in theepididymal adipose tissue from mice fed with an HFD afterexercise (MacPherson et al. 2014; Effect of acute exerciseon epididymal adipose tissue from mice fed a high fat diet,unpublished findings).

In addition to the increased markers of IL-6 signaling, ourfindings demonstrate reductions in indices of M1 macro-phages, as demonstrated by the decline in the F4/80- andCD11c-positive cells after exercise. M1 macrophages areknown to be central mediators of obesity-induced inflam-mation and insulin resistance (19). Therefore, if an increasedM1 macrophage population leads to disruption of glucosehomeostasis, the ratio of M1 to M2 macrophages is likelyimportant for metabolic control. The significance of this ratiofor insulin sensitivity was previously demonstrated in a studyin which mice deficient in M2 macrophages displayed in-creased white adipose tissue inflammation and insulin re-sistance (8). Furthermore, ablation of CD11c-positive cellsresults in normalization of insulin sensitivity (26).

M2 macrophages express and secrete IL-10 to a greaterdegree than M1 macrophages (6). In cultured L6 myoblasts,IL-10 potentiates insulin action, as demonstrated by increases

in insulin-stimulated AKT phosphorylation, glucose transport,and GLUT4 translocation (28). Moreover, it has previouslybeen demonstrated that IL-10 improves glucose uptake in3T3-L1 adipocytes (9) and can activate PI3K-dependentpathways (40). In the current study, we have shown increasesin IL-10 expression that were paralleled by increases in thephosphorylation of p38 MAPK, a signaling enzyme previ-ously shown to control the expression of IL-10 (as reviewedby Saraiva and O_Garra [29]). Along with increased IL-10expression, the present study also found increases in thephosphorylation of STAT3, a marker of IL-10, in addition toIL-6 signaling. Based on these findings, it is our currentworking hypothesis that the exercise-induced increases inIL-6 may serve as a trigger to mediate a switch in macro-phage polarization, leading to a more anti-inflammatoryphenotype including increases in IL-10 expression and se-cretion. These changes could be linked to improved adi-pose tissue insulin action.

In summary, we have provided novel data demonstratingimprovements in inguinal adipose tissue insulin action inobese mice after a single bout of exercise that is paralleledby apparent increases in IL-6 and/or IL-10 signaling anddecreased M1 macrophages. These findings speak of thepowerful effects of exercise as a means of modulating adi-pose tissue immunometabolism independent of weight lossand further suggest that targeted interventions to stimulatethese pathways could be an effective approach to improveadipose tissue insulin action.

R. E. K. was funded by a postdoctoral fellowship from theAlzheimer_s Society. S. F. C. was funded by an Ontario graduatescholarship and subsequently by a masters postgraduate scholar-ship from the Natural Sciences and Engineering Research Councilof Canada. D. C. W. is a Tier II Canada Research chair. This workwas supported by an operating grant from the Canadian Institutesof Health Research to D. C. W. J. A. S. is a new investigator withthe Heart and Stroke Foundation of Ontario.

There are no conflicts of interest to disclose.The results of the present study do constitute endorsement by the

American College of Sports Medicine.

REFERENCES

1. Baynard T, Vieira-Potter VJ, Valentine RJ, Woods JA. Exercisetraining effects on inflammatory gene expression in white adiposetissue of young mice. Mediators Inflamm. 2012;2012:767953.

2. Bostrom P, Wu J, Jedrychowski MP, et al. A PGC1->-dependentmyokine that drives brown-fat-like development of white fat andthermogenesis. Nature. 2012;481(7382):463–8.

3. Castellani L, Root-Mccaig J, Frendo-Cumbo S, Beaudoin MS,Wright DC. Exercise training protects against an acute inflamma-tory insult in mouse epididymal adipose tissue. J Appl Physiol(1985). 2014;116(10):1272–80.

4. Fujimoto M, Naka T. Regulation of cytokine signaling by SOCSfamily molecules. Trends Immunol. 2003;24(12):659–66.

5. Fujishiro M, Gotoh Y, Katagiri H, et al. Three mitogen-activated protein kinases inhibit insulin signaling by differentmechanisms in 3T3-L1 adipocytes. Mol Endocrinol. 2003;17(3):487–97.

6. Gordon S. Alternative activation of macrophages. Nat Rev Immunol.2003;3(1):23–35.

7. Heinrich PC, Behrmann I, Haan S, Hermanns HM, Muller-Newen G,Schaper F. Principles of interleukin (IL)-6-type cytokine signallingand its regulation. Biochem J. 2003;374(Pt 1):1–20.

8. Hevener AL, Olefsky JM, Reichart D, et al. Macrophage PPARgamma is required for normal skeletal muscle and hepatic insulinsensitivity and full antidiabetic effects of thiazolidinediones. J ClinInvest. 2007;117(6):1658–69.

9. Hong EG, Ko HJ, Cho YR, et al. Interleukin-10 prevents diet-induced insulin resistance by attenuating macrophage and cytokineresponse in skeletal muscle. Diabetes. 2009;58(11):2525–35.

10. Jeppesen J, Kiens B. Regulation and limitations to fatty acid oxi-dation during exercise. J Physiol. 2012;590(Pt 5):1059–68.

11. Kanda H, Tateya S, Tamori Y, et al. MCP-1 contributes to mac-rophage infiltration into adipose tissue, insulin resistance, and he-patic steatosis in obesity. J Clin Invest. 2006;116(6):1494–505.

12. Kang K, Reilly SM, Karabacak V, et al. Adipocyte-derived Th2cytokines and myeloid PPARdelta regulate macrophage polariza-tion and insulin sensitivity. Cell Metab. 2008;7(6):485–95.


BASIC

SCIEN

CES


13. Kawanishi N, Mizokami T, Yano H, Suzuki K. Exercise attenuatesM1 macrophages and CD8+ T cells in the adipose tissue of obesemice. Med Sci Sports Exerc. 2013;45(9):1684–93.

14. Kawanishi N, Niihara H, Mizokami T, Yano H, Suzuki K. Exer-cise training attenuates adipose tissue fibrosis in diet-inducedobese mice. Biochem Biophys Res Commun. 2013;440(4):774–9.

15. Kawanishi N, Yano H, Yokogawa Y, Suzuki K. Exercise traininginhibits inflammation in adipose tissue via both suppression ofmacrophage infiltration and acceleration of phenotypic switchingfrom M1 to M2 macrophages in high-fat-diet-induced obese mice.Exerc Immunol Rev. 2010;16:105–18.

16. Kim S, Jin Y, Choi Y, Park T. Resveratrol exerts anti-obesityeffects via mechanisms involving down-regulation of adipogenicand inflammatory processes in mice. Biochem Pharmacol. 2011;81(11):1343–51.

17. Kumar A, Harris TE, Keller SR, Choi KM, Magnuson MA,Lawrence JC Jr. Muscle-specific deletion of rictor impairs insulin-stimulated glucose transport and enhances Basal glycogen synthaseactivity. Mol Cell Biol. 2008;28(1):61–70.

18. Livak KJ, Schmittgen TD. Analysis of relative gene expressiondata using real-time quantitative PCR and the 2(jDelta Delta C(T))method. Methods. 2001;25(4):402–8.

19. Lumeng CN, Bodzin JL, Saltiel AR. Obesity induces a phenotypicswitch in adipose tissue macrophage polarization. J Clin Invest.2007;117(1):175–84.

20. Lumeng CN, DelProposto JB, Westcott DJ, Saltiel AR. Phenotypicswitching of adipose tissue macrophages with obesity is generatedby spatiotemporal differences in macrophage subtypes. Diabetes.2008;57(12):3239–46.

21. Matthews VB, Allen TL, Risis S, et al. Interleukin-6-deficient micedevelop hepatic inflammation and systemic insulin resistance.Diabetologia. 2010;53(11):2431–41.

22. Mauer J, Chaurasia B, Goldau J, et al. Signaling by IL-6 promotesalternative activation of macrophages to limit endotoxemia andobesity-associated resistance to insulin. Nat Immunol. 2014;15(5):423–30.

23. Mottillo EP, Shen XJ, Granneman JG. beta3-adrenergic receptorinduction of adipocyte inflammation requires lipolytic activation ofstress kinases p38 and JNK. Biochimic Biophys Acta. 2010;1801(9):1048–55.

24. Oliveira AG, Araujo TG, Carvalho BM, et al. Acute exercise in-duces a phenotypic switch in adipose tissue macrophage polariza-tion in diet-induced obese rats. Obesity (Silver Spring). 2013;21(12):2545–56.

25. Oliveira AG, Carvalho BM, Tobar N, et al. Physical exercise re-duces circulating lipopolysaccharide and TLR4 activation andimproves insulin signaling in tissues of DIO rats. Diabetes. 2011;60(3):784–96.

26. Patsouris D, Li PP, Thapar D, Chapman J, Olefsky JM, Neels JG.Ablation of CD11c-positive cells normalizes insulin sensitivity inobese insulin resistant animals. Cell Metab. 2008;8(4):301–9.

27. Rosa Neto JC, Lira FS, Oyama LM, et al. Exhaustive exercisecauses an anti-inflammatory effect in skeletal muscle and a pro-inflammatory effect in adipose tissue in rats. Eur J Appl Physiol.2009;106(5):697–704.

28. Samokhvalov V, Bilan PJ, Schertzer JD, Antonescu CN, Klip A.Palmitate- and lipopolysaccharide-activated macrophages evoke con-trasting insulin responses in muscle cells. Am J Physiol EndocrinolMetab. 2009;296(1):E37–46.

29. Saraiva M, O_Garra A. The regulation of IL-10 production byimmune cells. Nat Rev Immunol. 2010;10(3):170–81.

30. Sartipy P, Loskutoff DJ. Monocyte chemoattractant protein 1 inobesity and insulin resistance. Proc Natl Acad Sci U S A. 2003;100(12):7265–70.

31. Shoelson SE, Herrero L, Naaz A. Obesity, inflammation, and in-sulin resistance. Gastroenterology. 2007;132(6):2169–80.

32. Smith PK, Krohn RI, Hermanson GT, et al. Measurement of pro-tein using bicinchoninic acid. Anal Biochem. 1985;150(1):76–85.

33. Stanford KI, Middelbeek RJ, Townsend KL, et al. A novel role forsubcutaneous adipose tissue in exercise-induced improvements inglucose homeostasis. Diabetes. 2015;64(6):2002–14.

34. Stephens JM, Lee J, Pilch PF. Tumor necrosis factor-alpha-induced insulin resistance in 3 T3-L1 adipocytes is accompaniedby a loss of insulin receptor substrate-1 and GLUT4 expressionwithout a loss of insulin receptor-mediated signal transduction.J Biol Chem. 1997;272(2):971–6.

35. Tran TT, Yamamoto Y, Gesta S, Kahn CR. Beneficial effects ofsubcutaneous fat transplantation on metabolism. Cell Metab. 2008;7(5):410–20.

36. Vieira VJ, Valentine RJ, Wilund KR, Antao N, Baynard T, WoodsJA. Effects of exercise and low-fat diet on adipose tissue inflam-mation and metabolic complications in obese mice. Am J PhysiolEndocrinol Metab. 2009;296(5):E1164–71.

37. Wallenius V, Wallenius K, Ahren B, et al. Interleukin-6-deficientmice develop mature-onset obesity. Nat Med. 2002;8(1):75–9.

38. Wan Z, Ritchie I, Beaudoin MS, Castellani L, Chan CB, WrightDC. IL-6 indirectly modulates the induction of glyceroneo-genic enzymes in adipose tissue during exercise. PLoS One. 2012;7(7):e41719.

39. Zhang W, Mottillo EP, Zhao J, et al. Adipocyte lipolysis-stimulated interleukin-6 production requires sphingosine kinase 1activity. J Biol Chem. 2014;289(46):32178–85.

40. Zhou JH, Broussard SR, Strle K, et al. IL-10 inhibits apoptosisof promyeloid cells by activating insulin receptor substrate-2and phosphatidylinositol 3¶-kinase. J Immunol. 2001;167(8):4436–42.


BASICSC

IENCES


Documents

Adipose Tissue Insulin Action and IL-6 Signaling …...Adipose Tissue Insulin Action and IL-6 Signaling after Exercise in Obese Mice REBECCA E. K. MACPHERSON, JASON S. HUBER, SCOTT